12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235123612371238123912401241124212431244124512461247124812491250125112521253125412551256125712581259126012611262126312641265126612671268126912701271127212731274127512761277127812791280128112821283128412851286128712881289129012911292129312941295129612971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355135613571358135913601361136213631364136513661367136813691370137113721373137413751376137713781379138013811382138313841385138613871388138913901391139213931394139513961397139813991400140114021403140414051406140714081409141014111412141314141415141614171418141914201421142214231424142514261427142814291430143114321433143414351436143714381439144014411442144314441445144614471448144914501451145214531454145514561457145814591460146114621463146414651466146714681469147014711472147314741475147614771478147914801481148214831484148514861487148814891490149114921493149414951496149714981499150015011502150315041505150615071508150915101511151215131514151515161517151815191520152115221523152415251526152715281529153015311532153315341535153615371538153915401541154215431544154515461547154815491550155115521553155415551556155715581559156015611562156315641565156615671568156915701571157215731574157515761577157815791580158115821583158415851586158715881589159015911592159315941595159615971598159916001601160216031604160516061607160816091610161116121613161416151616161716181619162016211622162316241625162616271628162916301631163216331634163516361637163816391640164116421643164416451646164716481649165016511652165316541655165616571658165916601661166216631664166516661667166816691670167116721673167416751676167716781679168016811682168316841685168616871688168916901691169216931694169516961697169816991700170117021703170417051706170717081709171017111712171317141715171617171718171917201721172217231724172517261727172817291730173117321733173417351736173717381739174017411742174317441745174617471748174917501751175217531754175517561757175817591760176117621763176417651766176717681769177017711772177317741775177617771778177917801781178217831784178517861787178817891790179117921793179417951796179717981799180018011802180318041805180618071808180918101811181218131814181518161817181818191820182118221823182418251826182718281829183018311832183318341835183618371838183918401841184218431844184518461847184818491850185118521853185418551856185718581859186018611862186318641865186618671868186918701871187218731874187518761877187818791880188118821883188418851886188718881889189018911892189318941895189618971898189919001901190219031904190519061907190819091910191119121913191419151916191719181919192019211922192319241925192619271928192919301931193219331934193519361937193819391940194119421943194419451946194719481949195019511952195319541955195619571958195919601961196219631964196519661967196819691970197119721973197419751976197719781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015201620172018201920202021202220232024202520262027202820292030203120322033203420352036203720382039204020412042204320442045204620472048204920502051205220532054205520562057205820592060206120622063206420652066206720682069207020712072207320742075207620772078207920802081208220832084208520862087208820892090209120922093209420952096209720982099210021012102210321042105210621072108210921102111211221132114211521162117211821192120212121222123212421252126212721282129213021312132213321342135213621372138213921402141214221432144214521462147214821492150215121522153215421552156215721582159216021612162216321642165216621672168216921702171217221732174217521762177217821792180218121822183218421852186218721882189219021912192219321942195219621972198219922002201220222032204220522062207220822092210221122122213221422152216221722182219222022212222222322242225222622272228222922302231223222332234223522362237223822392240224122422243224422452246224722482249225022512252225322542255225622572258225922602261226222632264226522662267226822692270227122722273227422752276227722782279228022812282228322842285228622872288228922902291229222932294229522962297229822992300230123022303230423052306230723082309231023112312231323142315231623172318231923202321232223232324232523262327232823292330233123322333233423352336233723382339234023412342234323442345234623472348234923502351235223532354235523562357235823592360236123622363236423652366236723682369237023712372237323742375237623772378237923802381238223832384238523862387238823892390239123922393239423952396239723982399240024012402240324042405240624072408240924102411241224132414241524162417241824192420242124222423242424252426242724282429243024312432243324342435243624372438243924402441244224432444244524462447244824492450245124522453245424552456245724582459246024612462246324642465246624672468246924702471247224732474247524762477247824792480248124822483248424852486248724882489249024912492249324942495249624972498249925002501250225032504250525062507250825092510251125122513251425152516251725182519252025212522252325242525252625272528252925302531253225332534253525362537253825392540254125422543254425452546254725482549255025512552255325542555255625572558255925602561256225632564256525662567256825692570257125722573257425752576257725782579258025812582258325842585258625872588258925902591259225932594259525962597259825992600260126022603260426052606260726082609261026112612261326142615261626172618261926202621262226232624262526262627262826292630263126322633263426352636263726382639264026412642264326442645264626472648264926502651265226532654265526562657265826592660266126622663266426652666266726682669267026712672267326742675267626772678267926802681268226832684268526862687268826892690269126922693269426952696269726982699270027012702270327042705270627072708270927102711271227132714271527162717271827192720272127222723272427252726272727282729273027312732273327342735273627372738273927402741274227432744274527462747274827492750275127522753275427552756275727582759276027612762276327642765276627672768276927702771277227732774277527762777277827792780278127822783278427852786278727882789279027912792279327942795279627972798279928002801280228032804280528062807280828092810281128122813281428152816281728182819282028212822282328242825282628272828282928302831283228332834283528362837283828392840284128422843284428452846284728482849285028512852285328542855285628572858285928602861286228632864286528662867286828692870287128722873287428752876287728782879288028812882288328842885288628872888288928902891289228932894289528962897289828992900290129022903290429052906290729082909291029112912291329142915291629172918291929202921292229232924292529262927292829292930293129322933293429352936293729382939294029412942294329442945294629472948294929502951295229532954295529562957295829592960296129622963296429652966296729682969297029712972297329742975297629772978297929802981298229832984298529862987298829892990299129922993299429952996299729982999300030013002300330043005300630073008300930103011301230133014301530163017301830193020302130223023302430253026302730283029303030313032303330343035303630373038303930403041304230433044304530463047304830493050305130523053305430553056305730583059306030613062306330643065306630673068306930703071307230733074307530763077307830793080308130823083308430853086308730883089309030913092309330943095309630973098309931003101310231033104310531063107310831093110311131123113311431153116311731183119312031213122312331243125312631273128312931303131313231333134313531363137313831393140314131423143314431453146314731483149315031513152315331543155315631573158315931603161316231633164316531663167316831693170317131723173317431753176317731783179318031813182318331843185318631873188318931903191319231933194319531963197319831993200320132023203320432053206320732083209321032113212321332143215321632173218321932203221322232233224322532263227322832293230323132323233323432353236323732383239324032413242324332443245324632473248324932503251325232533254325532563257325832593260326132623263326432653266326732683269327032713272327332743275327632773278327932803281328232833284328532863287328832893290329132923293329432953296329732983299330033013302330333043305330633073308330933103311331233133314331533163317331833193320332133223323332433253326332733283329333033313332333333343335333633373338333933403341334233433344334533463347334833493350335133523353335433553356335733583359336033613362336333643365336633673368336933703371337233733374337533763377337833793380338133823383338433853386338733883389339033913392339333943395339633973398339934003401340234033404340534063407340834093410341134123413341434153416341734183419342034213422342334243425342634273428342934303431343234333434343534363437343834393440344134423443344434453446344734483449345034513452345334543455345634573458345934603461346234633464346534663467346834693470347134723473347434753476347734783479348034813482348334843485348634873488348934903491349234933494349534963497349834993500350135023503350435053506350735083509351035113512351335143515351635173518351935203521352235233524352535263527352835293530353135323533353435353536353735383539354035413542354335443545354635473548354935503551355235533554355535563557355835593560356135623563356435653566356735683569357035713572357335743575357635773578357935803581358235833584358535863587358835893590359135923593359435953596359735983599360036013602360336043605360636073608360936103611361236133614361536163617361836193620362136223623362436253626362736283629363036313632363336343635363636373638363936403641364236433644364536463647364836493650365136523653365436553656365736583659366036613662366336643665366636673668366936703671367236733674367536763677367836793680368136823683368436853686368736883689369036913692369336943695369636973698369937003701370237033704370537063707370837093710371137123713371437153716371737183719372037213722372337243725372637273728372937303731373237333734373537363737373837393740374137423743374437453746374737483749375037513752375337543755375637573758375937603761376237633764376537663767376837693770377137723773377437753776377737783779378037813782378337843785378637873788378937903791379237933794379537963797379837993800380138023803380438053806380738083809381038113812381338143815381638173818381938203821382238233824382538263827382838293830383138323833383438353836383738383839384038413842384338443845384638473848384938503851385238533854385538563857385838593860386138623863386438653866386738683869387038713872387338743875387638773878387938803881388238833884388538863887388838893890389138923893389438953896389738983899390039013902390339043905390639073908390939103911391239133914391539163917391839193920392139223923392439253926392739283929393039313932393339343935393639373938393939403941394239433944394539463947394839493950395139523953395439553956395739583959396039613962396339643965396639673968396939703971397239733974397539763977397839793980398139823983398439853986398739883989399039913992399339943995399639973998399940004001400240034004400540064007400840094010401140124013401440154016401740184019402040214022402340244025402640274028402940304031403240334034403540364037403840394040404140424043404440454046404740484049405040514052405340544055405640574058405940604061406240634064406540664067406840694070407140724073407440754076407740784079408040814082408340844085408640874088408940904091409240934094409540964097409840994100410141024103410441054106410741084109411041114112411341144115411641174118411941204121412241234124412541264127412841294130413141324133413441354136413741384139414041414142414341444145414641474148414941504151415241534154415541564157415841594160416141624163416441654166416741684169417041714172417341744175417641774178417941804181418241834184418541864187418841894190419141924193419441954196419741984199420042014202420342044205420642074208420942104211421242134214421542164217421842194220422142224223422442254226422742284229423042314232423342344235423642374238423942404241424242434244424542464247424842494250425142524253425442554256425742584259426042614262426342644265426642674268426942704271427242734274427542764277427842794280428142824283428442854286428742884289429042914292429342944295429642974298429943004301430243034304430543064307430843094310431143124313431443154316431743184319432043214322432343244325432643274328432943304331433243334334433543364337433843394340434143424343434443454346434743484349435043514352435343544355435643574358435943604361436243634364436543664367436843694370437143724373437443754376437743784379438043814382438343844385438643874388438943904391439243934394439543964397439843994400440144024403440444054406440744084409441044114412441344144415441644174418441944204421442244234424442544264427442844294430443144324433443444354436443744384439444044414442444344444445444644474448444944504451445244534454445544564457445844594460446144624463446444654466446744684469447044714472447344744475447644774478447944804481448244834484448544864487448844894490449144924493449444954496449744984499450045014502450345044505450645074508450945104511451245134514451545164517451845194520452145224523452445254526452745284529453045314532453345344535453645374538453945404541454245434544454545464547454845494550455145524553455445554556455745584559456045614562456345644565456645674568456945704571457245734574457545764577457845794580458145824583458445854586458745884589459045914592459345944595459645974598459946004601460246034604460546064607460846094610461146124613461446154616461746184619462046214622462346244625462646274628462946304631463246334634463546364637463846394640464146424643464446454646464746484649465046514652465346544655465646574658465946604661466246634664466546664667466846694670467146724673467446754676467746784679468046814682468346844685468646874688468946904691469246934694469546964697469846994700470147024703470447054706470747084709471047114712471347144715471647174718471947204721472247234724472547264727472847294730473147324733473447354736473747384739474047414742474347444745474647474748474947504751475247534754475547564757475847594760476147624763476447654766476747684769477047714772477347744775477647774778477947804781478247834784478547864787478847894790479147924793479447954796479747984799480048014802480348044805480648074808480948104811481248134814481548164817481848194820482148224823482448254826482748284829483048314832483348344835483648374838483948404841484248434844484548464847484848494850485148524853485448554856485748584859486048614862486348644865486648674868486948704871487248734874487548764877487848794880488148824883488448854886488748884889489048914892489348944895489648974898489949004901490249034904490549064907490849094910491149124913491449154916491749184919492049214922492349244925492649274928492949304931493249334934493549364937493849394940494149424943494449454946494749484949495049514952495349544955495649574958495949604961496249634964496549664967496849694970497149724973497449754976497749784979498049814982498349844985498649874988498949904991499249934994499549964997499849995000500150025003500450055006500750085009501050115012501350145015501650175018501950205021502250235024502550265027502850295030503150325033503450355036503750385039504050415042504350445045504650475048504950505051505250535054505550565057505850595060506150625063506450655066506750685069507050715072507350745075507650775078507950805081508250835084508550865087508850895090509150925093509450955096509750985099510051015102510351045105510651075108510951105111511251135114511551165117511851195120512151225123512451255126512751285129513051315132513351345135513651375138513951405141514251435144514551465147514851495150515151525153515451555156515751585159516051615162516351645165516651675168516951705171517251735174517551765177517851795180518151825183518451855186518751885189519051915192519351945195519651975198519952005201520252035204520552065207520852095210521152125213521452155216521752185219522052215222522352245225522652275228522952305231523252335234523552365237523852395240524152425243524452455246524752485249525052515252525352545255525652575258525952605261526252635264526552665267526852695270527152725273527452755276527752785279528052815282528352845285528652875288528952905291529252935294529552965297529852995300530153025303530453055306530753085309531053115312531353145315531653175318531953205321532253235324532553265327532853295330533153325333533453355336533753385339534053415342534353445345534653475348534953505351535253535354535553565357535853595360536153625363536453655366536753685369537053715372537353745375537653775378537953805381538253835384538553865387538853895390539153925393539453955396539753985399540054015402540354045405540654075408540954105411541254135414541554165417541854195420542154225423542454255426542754285429543054315432543354345435543654375438543954405441544254435444544554465447544854495450545154525453545454555456545754585459546054615462546354645465546654675468546954705471547254735474547554765477547854795480548154825483548454855486548754885489549054915492549354945495549654975498549955005501550255035504550555065507550855095510551155125513551455155516551755185519552055215522552355245525552655275528552955305531553255335534553555365537553855395540554155425543554455455546554755485549555055515552555355545555555655575558555955605561556255635564556555665567556855695570557155725573557455755576557755785579558055815582558355845585558655875588558955905591559255935594559555965597559855995600560156025603560456055606560756085609561056115612561356145615561656175618561956205621562256235624562556265627562856295630563156325633563456355636563756385639564056415642564356445645564656475648564956505651565256535654565556565657565856595660566156625663566456655666566756685669567056715672567356745675567656775678567956805681568256835684568556865687568856895690569156925693569456955696569756985699570057015702570357045705570657075708570957105711571257135714571557165717571857195720572157225723572457255726572757285729573057315732573357345735573657375738573957405741574257435744574557465747574857495750575157525753575457555756575757585759576057615762576357645765576657675768576957705771577257735774577557765777577857795780578157825783578457855786578757885789579057915792579357945795579657975798579958005801580258035804580558065807580858095810581158125813581458155816581758185819582058215822582358245825582658275828582958305831583258335834583558365837583858395840584158425843584458455846584758485849585058515852585358545855585658575858585958605861586258635864586558665867586858695870587158725873587458755876587758785879588058815882588358845885588658875888588958905891589258935894589558965897589858995900590159025903590459055906590759085909591059115912591359145915591659175918591959205921592259235924592559265927592859295930593159325933593459355936593759385939594059415942594359445945594659475948594959505951595259535954595559565957595859595960596159625963596459655966596759685969597059715972597359745975597659775978597959805981598259835984598559865987598859895990599159925993599459955996599759985999600060016002600360046005600660076008600960106011601260136014601560166017601860196020602160226023602460256026602760286029603060316032603360346035603660376038603960406041604260436044604560466047604860496050605160526053605460556056605760586059606060616062606360646065606660676068606960706071607260736074607560766077607860796080608160826083608460856086608760886089609060916092609360946095609660976098609961006101610261036104610561066107610861096110611161126113611461156116611761186119612061216122612361246125612661276128612961306131613261336134613561366137613861396140614161426143614461456146614761486149615061516152615361546155615661576158615961606161616261636164616561666167616861696170617161726173617461756176617761786179618061816182618361846185618661876188618961906191619261936194619561966197619861996200620162026203620462056206620762086209621062116212621362146215621662176218621962206221622262236224622562266227622862296230623162326233623462356236623762386239624062416242624362446245624662476248624962506251625262536254625562566257625862596260626162626263626462656266626762686269627062716272627362746275627662776278627962806281628262836284628562866287628862896290629162926293629462956296629762986299630063016302630363046305630663076308630963106311631263136314631563166317631863196320632163226323632463256326632763286329633063316332633363346335633663376338633963406341634263436344634563466347634863496350635163526353635463556356635763586359636063616362636363646365636663676368636963706371637263736374637563766377637863796380638163826383638463856386638763886389639063916392639363946395639663976398639964006401640264036404640564066407640864096410641164126413641464156416641764186419642064216422642364246425642664276428642964306431643264336434643564366437643864396440644164426443644464456446644764486449645064516452645364546455645664576458645964606461646264636464646564666467646864696470647164726473647464756476647764786479648064816482648364846485648664876488648964906491649264936494649564966497649864996500650165026503650465056506650765086509651065116512651365146515651665176518651965206521652265236524652565266527652865296530653165326533653465356536653765386539654065416542654365446545654665476548654965506551655265536554655565566557655865596560656165626563656465656566656765686569657065716572657365746575657665776578657965806581658265836584658565866587658865896590659165926593659465956596659765986599660066016602660366046605660666076608660966106611661266136614661566166617661866196620662166226623662466256626662766286629663066316632663366346635663666376638663966406641664266436644664566466647664866496650665166526653665466556656665766586659666066616662666366646665666666676668666966706671667266736674667566766677667866796680668166826683668466856686668766886689669066916692669366946695669666976698669967006701670267036704670567066707670867096710671167126713671467156716671767186719672067216722672367246725672667276728672967306731673267336734673567366737673867396740674167426743674467456746674767486749675067516752675367546755675667576758675967606761676267636764676567666767676867696770677167726773677467756776677767786779678067816782678367846785678667876788678967906791679267936794679567966797679867996800680168026803680468056806680768086809681068116812681368146815681668176818681968206821682268236824682568266827682868296830683168326833683468356836683768386839684068416842684368446845684668476848684968506851685268536854685568566857685868596860686168626863686468656866686768686869687068716872687368746875687668776878687968806881688268836884688568866887688868896890689168926893689468956896689768986899690069016902690369046905690669076908690969106911691269136914691569166917691869196920692169226923692469256926692769286929693069316932693369346935693669376938693969406941694269436944694569466947694869496950695169526953695469556956695769586959696069616962696369646965696669676968696969706971697269736974697569766977697869796980698169826983698469856986698769886989699069916992699369946995699669976998699970007001700270037004700570067007700870097010701170127013701470157016701770187019702070217022702370247025702670277028702970307031703270337034703570367037703870397040704170427043704470457046704770487049705070517052705370547055705670577058705970607061706270637064706570667067706870697070707170727073707470757076707770787079708070817082708370847085708670877088708970907091709270937094709570967097709870997100710171027103710471057106710771087109711071117112711371147115711671177118711971207121712271237124712571267127712871297130713171327133713471357136713771387139714071417142714371447145714671477148714971507151715271537154715571567157715871597160716171627163716471657166716771687169717071717172717371747175717671777178717971807181718271837184718571867187718871897190719171927193719471957196719771987199720072017202720372047205720672077208720972107211721272137214721572167217721872197220722172227223722472257226722772287229723072317232723372347235723672377238723972407241724272437244724572467247724872497250725172527253725472557256725772587259726072617262726372647265726672677268726972707271727272737274727572767277727872797280728172827283728472857286728772887289729072917292729372947295729672977298729973007301730273037304730573067307730873097310731173127313731473157316731773187319732073217322732373247325732673277328732973307331733273337334733573367337733873397340734173427343734473457346734773487349735073517352735373547355735673577358735973607361736273637364736573667367736873697370737173727373737473757376737773787379738073817382738373847385738673877388738973907391739273937394739573967397739873997400740174027403740474057406740774087409741074117412741374147415741674177418741974207421742274237424742574267427742874297430743174327433743474357436743774387439744074417442744374447445744674477448744974507451745274537454745574567457745874597460746174627463746474657466746774687469747074717472747374747475747674777478747974807481748274837484748574867487748874897490749174927493749474957496749774987499750075017502750375047505750675077508750975107511751275137514751575167517751875197520752175227523752475257526752775287529753075317532753375347535753675377538753975407541754275437544754575467547754875497550755175527553755475557556755775587559756075617562756375647565756675677568756975707571757275737574757575767577757875797580758175827583758475857586758775887589759075917592759375947595759675977598759976007601760276037604760576067607760876097610761176127613761476157616761776187619762076217622762376247625762676277628762976307631763276337634763576367637763876397640764176427643764476457646764776487649765076517652765376547655765676577658765976607661766276637664766576667667766876697670767176727673767476757676767776787679768076817682768376847685768676877688768976907691769276937694769576967697769876997700770177027703770477057706770777087709771077117712771377147715771677177718771977207721772277237724772577267727772877297730773177327733773477357736773777387739774077417742774377447745774677477748774977507751775277537754775577567757775877597760776177627763776477657766776777687769777077717772777377747775777677777778777977807781778277837784778577867787778877897790779177927793779477957796779777987799780078017802780378047805780678077808780978107811781278137814781578167817781878197820782178227823782478257826782778287829783078317832783378347835783678377838783978407841784278437844784578467847784878497850785178527853785478557856785778587859786078617862786378647865786678677868786978707871787278737874787578767877787878797880788178827883788478857886788778887889789078917892789378947895789678977898789979007901790279037904790579067907790879097910791179127913791479157916791779187919792079217922792379247925792679277928792979307931793279337934793579367937793879397940794179427943794479457946794779487949795079517952795379547955795679577958795979607961796279637964796579667967796879697970797179727973797479757976797779787979798079817982798379847985798679877988798979907991799279937994799579967997799879998000800180028003800480058006800780088009801080118012801380148015801680178018801980208021802280238024802580268027802880298030803180328033803480358036803780388039804080418042804380448045804680478048804980508051805280538054805580568057805880598060806180628063806480658066806780688069807080718072807380748075807680778078807980808081808280838084808580868087808880898090809180928093809480958096809780988099810081018102810381048105810681078108810981108111811281138114811581168117811881198120812181228123812481258126812781288129813081318132813381348135813681378138813981408141814281438144814581468147814881498150815181528153815481558156815781588159816081618162816381648165816681678168816981708171817281738174817581768177817881798180818181828183818481858186818781888189819081918192819381948195819681978198819982008201820282038204820582068207820882098210821182128213821482158216821782188219822082218222822382248225822682278228822982308231823282338234823582368237823882398240824182428243824482458246824782488249825082518252825382548255825682578258825982608261826282638264826582668267826882698270827182728273827482758276827782788279828082818282828382848285828682878288828982908291829282938294829582968297829882998300830183028303830483058306830783088309831083118312831383148315831683178318831983208321832283238324832583268327832883298330833183328333833483358336833783388339834083418342834383448345834683478348834983508351835283538354835583568357835883598360836183628363836483658366836783688369837083718372837383748375837683778378837983808381838283838384838583868387838883898390839183928393839483958396839783988399840084018402840384048405840684078408840984108411841284138414841584168417841884198420842184228423842484258426842784288429843084318432843384348435843684378438843984408441844284438444844584468447844884498450845184528453845484558456845784588459846084618462846384648465846684678468846984708471847284738474847584768477847884798480848184828483848484858486848784888489849084918492849384948495849684978498849985008501850285038504850585068507850885098510851185128513851485158516851785188519852085218522852385248525852685278528852985308531853285338534853585368537853885398540854185428543854485458546854785488549855085518552855385548555855685578558855985608561856285638564856585668567856885698570857185728573857485758576857785788579858085818582858385848585858685878588858985908591859285938594859585968597859885998600860186028603860486058606860786088609861086118612861386148615861686178618861986208621862286238624862586268627862886298630863186328633863486358636863786388639864086418642864386448645864686478648864986508651865286538654865586568657865886598660866186628663866486658666866786688669867086718672867386748675867686778678867986808681868286838684868586868687868886898690869186928693869486958696869786988699870087018702870387048705870687078708870987108711871287138714871587168717871887198720872187228723872487258726872787288729873087318732873387348735873687378738873987408741874287438744874587468747874887498750875187528753875487558756875787588759876087618762876387648765876687678768876987708771877287738774877587768777877887798780878187828783878487858786878787888789879087918792879387948795879687978798879988008801880288038804880588068807880888098810881188128813881488158816881788188819882088218822882388248825882688278828882988308831883288338834883588368837883888398840884188428843884488458846884788488849885088518852885388548855885688578858885988608861886288638864886588668867886888698870887188728873887488758876887788788879888088818882888388848885888688878888888988908891889288938894889588968897889888998900890189028903890489058906890789088909891089118912891389148915891689178918891989208921892289238924892589268927892889298930893189328933893489358936893789388939894089418942894389448945894689478948894989508951895289538954895589568957895889598960896189628963896489658966896789688969897089718972897389748975897689778978897989808981898289838984898589868987898889898990899189928993899489958996899789988999900090019002900390049005900690079008900990109011901290139014901590169017901890199020902190229023902490259026902790289029903090319032903390349035903690379038903990409041904290439044904590469047904890499050905190529053905490559056905790589059906090619062906390649065906690679068906990709071907290739074907590769077907890799080908190829083908490859086908790889089909090919092909390949095909690979098909991009101910291039104910591069107910891099110911191129113911491159116911791189119912091219122912391249125912691279128912991309131913291339134913591369137913891399140914191429143914491459146914791489149915091519152915391549155915691579158915991609161916291639164916591669167916891699170917191729173917491759176917791789179918091819182918391849185918691879188918991909191919291939194919591969197919891999200920192029203920492059206920792089209921092119212921392149215921692179218921992209221922292239224922592269227922892299230923192329233923492359236923792389239924092419242924392449245924692479248924992509251925292539254925592569257925892599260926192629263926492659266926792689269927092719272927392749275927692779278927992809281928292839284928592869287928892899290929192929293929492959296929792989299930093019302930393049305930693079308930993109311931293139314931593169317931893199320932193229323932493259326932793289329933093319332933393349335933693379338933993409341934293439344934593469347934893499350935193529353935493559356935793589359936093619362936393649365936693679368936993709371937293739374937593769377937893799380938193829383938493859386938793889389939093919392939393949395939693979398939994009401940294039404940594069407940894099410941194129413941494159416941794189419942094219422942394249425942694279428942994309431943294339434943594369437943894399440944194429443944494459446944794489449945094519452945394549455945694579458945994609461946294639464946594669467946894699470947194729473947494759476947794789479948094819482948394849485948694879488948994909491949294939494949594969497949894999500950195029503950495059506950795089509951095119512951395149515951695179518951995209521952295239524952595269527952895299530953195329533953495359536953795389539954095419542954395449545954695479548954995509551955295539554955595569557955895599560956195629563956495659566956795689569957095719572957395749575957695779578957995809581958295839584958595869587958895899590959195929593959495959596959795989599960096019602960396049605960696079608960996109611961296139614961596169617961896199620962196229623962496259626962796289629963096319632963396349635963696379638963996409641964296439644964596469647964896499650965196529653965496559656965796589659966096619662966396649665966696679668966996709671967296739674967596769677967896799680968196829683968496859686968796889689969096919692969396949695969696979698969997009701970297039704970597069707970897099710971197129713971497159716971797189719972097219722972397249725972697279728972997309731973297339734973597369737973897399740974197429743974497459746974797489749975097519752975397549755975697579758975997609761976297639764976597669767976897699770977197729773977497759776977797789779978097819782978397849785978697879788978997909791979297939794979597969797979897999800980198029803980498059806980798089809981098119812981398149815981698179818981998209821982298239824982598269827982898299830983198329833983498359836983798389839984098419842984398449845984698479848984998509851985298539854985598569857985898599860986198629863986498659866986798689869987098719872987398749875987698779878987998809881988298839884988598869887988898899890989198929893989498959896989798989899990099019902990399049905990699079908990999109911991299139914991599169917991899199920992199229923992499259926992799289929993099319932993399349935993699379938993999409941994299439944994599469947994899499950995199529953995499559956995799589959996099619962996399649965996699679968996999709971997299739974997599769977997899799980998199829983998499859986998799889989999099919992999399949995999699979998999910000100011000210003100041000510006100071000810009100101001110012100131001410015100161001710018100191002010021100221002310024100251002610027100281002910030100311003210033100341003510036100371003810039100401004110042100431004410045100461004710048100491005010051100521005310054100551005610057100581005910060100611006210063100641006510066100671006810069100701007110072100731007410075100761007710078100791008010081100821008310084100851008610087100881008910090100911009210093100941009510096100971009810099101001010110102101031010410105101061010710108101091011010111101121011310114101151011610117101181011910120101211012210123101241012510126101271012810129101301013110132101331013410135101361013710138101391014010141101421014310144101451014610147101481014910150101511015210153101541015510156101571015810159101601016110162101631016410165101661016710168101691017010171101721017310174101751017610177101781017910180101811018210183101841018510186101871018810189101901019110192101931019410195101961019710198101991020010201102021020310204102051020610207102081020910210102111021210213102141021510216102171021810219102201022110222102231022410225102261022710228102291023010231102321023310234102351023610237102381023910240102411024210243102441024510246102471024810249102501025110252102531025410255102561025710258102591026010261102621026310264102651026610267102681026910270102711027210273102741027510276102771027810279102801028110282102831028410285102861028710288102891029010291102921029310294102951029610297102981029910300103011030210303103041030510306103071030810309103101031110312103131031410315103161031710318103191032010321103221032310324103251032610327103281032910330103311033210333103341033510336103371033810339103401034110342103431034410345103461034710348103491035010351103521035310354103551035610357103581035910360103611036210363103641036510366103671036810369103701037110372103731037410375103761037710378103791038010381103821038310384103851038610387103881038910390103911039210393103941039510396103971039810399104001040110402104031040410405104061040710408104091041010411104121041310414104151041610417104181041910420104211042210423104241042510426104271042810429104301043110432104331043410435104361043710438104391044010441104421044310444104451044610447104481044910450104511045210453104541045510456104571045810459104601046110462104631046410465104661046710468104691047010471104721047310474104751047610477104781047910480104811048210483104841048510486104871048810489104901049110492104931049410495104961049710498104991050010501105021050310504105051050610507105081050910510105111051210513105141051510516105171051810519105201052110522105231052410525105261052710528105291053010531105321053310534105351053610537105381053910540105411054210543105441054510546105471054810549105501055110552105531055410555105561055710558105591056010561105621056310564105651056610567105681056910570105711057210573105741057510576105771057810579105801058110582105831058410585105861058710588105891059010591105921059310594105951059610597105981059910600106011060210603106041060510606106071060810609106101061110612106131061410615106161061710618106191062010621106221062310624106251062610627106281062910630106311063210633106341063510636106371063810639106401064110642106431064410645106461064710648106491065010651106521065310654106551065610657106581065910660106611066210663106641066510666106671066810669106701067110672106731067410675106761067710678106791068010681106821068310684106851068610687106881068910690106911069210693106941069510696106971069810699107001070110702107031070410705107061070710708107091071010711107121071310714107151071610717107181071910720107211072210723107241072510726107271072810729107301073110732107331073410735107361073710738107391074010741107421074310744107451074610747107481074910750107511075210753107541075510756107571075810759107601076110762107631076410765107661076710768107691077010771107721077310774107751077610777107781077910780107811078210783107841078510786107871078810789107901079110792107931079410795107961079710798107991080010801108021080310804108051080610807108081080910810108111081210813108141081510816108171081810819108201082110822108231082410825108261082710828108291083010831108321083310834108351083610837108381083910840108411084210843108441084510846108471084810849108501085110852108531085410855108561085710858108591086010861108621086310864108651086610867108681086910870108711087210873108741087510876108771087810879108801088110882108831088410885108861088710888108891089010891108921089310894108951089610897108981089910900109011090210903109041090510906109071090810909109101091110912109131091410915109161091710918109191092010921109221092310924109251092610927109281092910930109311093210933109341093510936109371093810939109401094110942109431094410945109461094710948109491095010951109521095310954109551095610957109581095910960109611096210963109641096510966109671096810969109701097110972109731097410975109761097710978109791098010981109821098310984109851098610987109881098910990109911099210993109941099510996109971099810999110001100111002110031100411005110061100711008110091101011011110121101311014110151101611017110181101911020110211102211023110241102511026110271102811029110301103111032110331103411035110361103711038110391104011041110421104311044110451104611047110481104911050110511105211053110541105511056110571105811059110601106111062110631106411065110661106711068110691107011071110721107311074110751107611077110781107911080110811108211083110841108511086110871108811089110901109111092110931109411095110961109711098110991110011101111021110311104111051110611107111081110911110111111111211113111141111511116111171111811119111201112111122111231112411125111261112711128111291113011131111321113311134111351113611137111381113911140111411114211143111441114511146111471114811149111501115111152111531115411155111561115711158111591116011161111621116311164111651116611167111681116911170111711117211173111741117511176111771117811179111801118111182111831118411185111861118711188111891119011191111921119311194111951119611197111981119911200112011120211203112041120511206112071120811209112101121111212112131121411215112161121711218112191122011221112221122311224112251122611227112281122911230112311123211233112341123511236112371123811239112401124111242112431124411245112461124711248112491125011251112521125311254112551125611257112581125911260112611126211263112641126511266112671126811269112701127111272112731127411275112761127711278112791128011281112821128311284112851128611287112881128911290112911129211293112941129511296112971129811299113001130111302113031130411305113061130711308113091131011311113121131311314113151131611317113181131911320113211132211323113241132511326113271132811329113301133111332113331133411335113361133711338113391134011341113421134311344113451134611347113481134911350113511135211353113541135511356113571135811359113601136111362113631136411365113661136711368113691137011371113721137311374113751137611377113781137911380113811138211383113841138511386113871138811389113901139111392113931139411395113961139711398113991140011401114021140311404114051140611407114081140911410114111141211413114141141511416114171141811419114201142111422114231142411425114261142711428114291143011431114321143311434114351143611437114381143911440114411144211443114441144511446114471144811449114501145111452114531145411455114561145711458114591146011461114621146311464114651146611467114681146911470114711147211473114741147511476114771147811479114801148111482114831148411485114861148711488114891149011491114921149311494114951149611497114981149911500115011150211503115041150511506115071150811509115101151111512115131151411515115161151711518115191152011521115221152311524115251152611527115281152911530115311153211533115341153511536115371153811539115401154111542115431154411545115461154711548115491155011551115521155311554115551155611557115581155911560115611156211563115641156511566115671156811569115701157111572115731157411575115761157711578115791158011581115821158311584115851158611587115881158911590115911159211593115941159511596115971159811599116001160111602116031160411605116061160711608116091161011611116121161311614116151161611617116181161911620116211162211623116241162511626116271162811629116301163111632116331163411635116361163711638116391164011641116421164311644116451164611647116481164911650116511165211653116541165511656116571165811659116601166111662116631166411665116661166711668116691167011671116721167311674116751167611677116781167911680116811168211683116841168511686116871168811689116901169111692116931169411695116961169711698116991170011701117021170311704117051170611707117081170911710117111171211713117141171511716117171171811719117201172111722117231172411725117261172711728117291173011731117321173311734117351173611737117381173911740117411174211743117441174511746117471174811749117501175111752117531175411755117561175711758117591176011761117621176311764117651176611767117681176911770117711177211773117741177511776117771177811779117801178111782117831178411785117861178711788117891179011791117921179311794117951179611797117981179911800118011180211803118041180511806118071180811809118101181111812118131181411815118161181711818118191182011821118221182311824118251182611827118281182911830118311183211833118341183511836118371183811839118401184111842118431184411845118461184711848118491185011851118521185311854118551185611857118581185911860118611186211863118641186511866118671186811869118701187111872118731187411875118761187711878118791188011881118821188311884118851188611887118881188911890118911189211893118941189511896118971189811899119001190111902119031190411905119061190711908119091191011911119121191311914119151191611917119181191911920119211192211923119241192511926119271192811929119301193111932119331193411935119361193711938119391194011941119421194311944119451194611947119481194911950119511195211953119541195511956119571195811959119601196111962119631196411965119661196711968119691197011971119721197311974119751197611977119781197911980119811198211983119841198511986119871198811989119901199111992119931199411995119961199711998119991200012001120021200312004120051200612007120081200912010120111201212013120141201512016120171201812019120201202112022120231202412025120261202712028120291203012031120321203312034120351203612037120381203912040120411204212043120441204512046120471204812049120501205112052120531205412055120561205712058120591206012061120621206312064120651206612067120681206912070120711207212073120741207512076120771207812079120801208112082120831208412085120861208712088120891209012091120921209312094120951209612097120981209912100121011210212103121041210512106121071210812109121101211112112121131211412115121161211712118121191212012121121221212312124121251212612127121281212912130121311213212133121341213512136121371213812139121401214112142121431214412145121461214712148121491215012151121521215312154121551215612157121581215912160121611216212163121641216512166121671216812169121701217112172121731217412175121761217712178121791218012181121821218312184121851218612187121881218912190121911219212193121941219512196121971219812199122001220112202122031220412205122061220712208122091221012211122121221312214122151221612217122181221912220122211222212223122241222512226122271222812229122301223112232122331223412235122361223712238122391224012241122421224312244122451224612247122481224912250122511225212253122541225512256122571225812259122601226112262122631226412265122661226712268122691227012271122721227312274122751227612277122781227912280122811228212283122841228512286122871228812289122901229112292122931229412295122961229712298122991230012301123021230312304123051230612307123081230912310123111231212313123141231512316123171231812319123201232112322123231232412325123261232712328123291233012331123321233312334123351233612337123381233912340123411234212343123441234512346123471234812349123501235112352123531235412355123561235712358123591236012361123621236312364123651236612367123681236912370123711237212373123741237512376123771237812379123801238112382123831238412385123861238712388123891239012391123921239312394123951239612397123981239912400124011240212403124041240512406124071240812409124101241112412124131241412415124161241712418124191242012421124221242312424124251242612427124281242912430124311243212433124341243512436124371243812439124401244112442124431244412445124461244712448124491245012451124521245312454124551245612457124581245912460124611246212463124641246512466124671246812469124701247112472124731247412475124761247712478124791248012481124821248312484124851248612487124881248912490124911249212493124941249512496124971249812499125001250112502125031250412505125061250712508125091251012511125121251312514125151251612517125181251912520125211252212523125241252512526125271252812529125301253112532125331253412535125361253712538125391254012541125421254312544125451254612547125481254912550125511255212553125541255512556125571255812559125601256112562125631256412565125661256712568125691257012571125721257312574125751257612577125781257912580125811258212583125841258512586125871258812589125901259112592125931259412595125961259712598125991260012601126021260312604126051260612607126081260912610126111261212613126141261512616126171261812619126201262112622126231262412625126261262712628126291263012631126321263312634126351263612637126381263912640126411264212643126441264512646126471264812649126501265112652126531265412655126561265712658126591266012661126621266312664126651266612667126681266912670126711267212673126741267512676126771267812679126801268112682126831268412685126861268712688126891269012691126921269312694126951269612697126981269912700127011270212703127041270512706127071270812709127101271112712127131271412715127161271712718127191272012721127221272312724127251272612727127281272912730127311273212733127341273512736127371273812739127401274112742127431274412745127461274712748127491275012751127521275312754127551275612757127581275912760127611276212763127641276512766127671276812769127701277112772127731277412775127761277712778127791278012781127821278312784127851278612787127881278912790127911279212793127941279512796127971279812799128001280112802128031280412805128061280712808128091281012811128121281312814128151281612817128181281912820128211282212823128241282512826128271282812829128301283112832128331283412835128361283712838128391284012841128421284312844128451284612847128481284912850128511285212853128541285512856128571285812859128601286112862128631286412865128661286712868128691287012871128721287312874128751287612877128781287912880128811288212883128841288512886128871288812889128901289112892128931289412895128961289712898128991290012901129021290312904129051290612907129081290912910129111291212913129141291512916129171291812919129201292112922129231292412925129261292712928129291293012931129321293312934129351293612937129381293912940129411294212943129441294512946129471294812949129501295112952129531295412955129561295712958129591296012961129621296312964129651296612967129681296912970129711297212973129741297512976129771297812979129801298112982129831298412985129861298712988129891299012991129921299312994129951299612997129981299913000130011300213003130041300513006130071300813009130101301113012130131301413015130161301713018130191302013021130221302313024130251302613027130281302913030130311303213033130341303513036130371303813039130401304113042130431304413045130461304713048130491305013051130521305313054130551305613057130581305913060130611306213063130641306513066130671306813069130701307113072130731307413075130761307713078130791308013081130821308313084130851308613087130881308913090130911309213093130941309513096130971309813099131001310113102131031310413105131061310713108131091311013111131121311313114131151311613117131181311913120131211312213123131241312513126131271312813129131301313113132131331313413135131361313713138131391314013141131421314313144131451314613147131481314913150131511315213153131541315513156131571315813159131601316113162131631316413165131661316713168131691317013171131721317313174131751317613177131781317913180131811318213183131841318513186131871318813189131901319113192131931319413195131961319713198131991320013201132021320313204132051320613207132081320913210132111321213213132141321513216132171321813219132201322113222132231322413225132261322713228132291323013231132321323313234132351323613237132381323913240132411324213243132441324513246132471324813249132501325113252132531325413255132561325713258132591326013261132621326313264132651326613267132681326913270132711327213273132741327513276132771327813279132801328113282132831328413285132861328713288132891329013291132921329313294132951329613297132981329913300133011330213303133041330513306133071330813309133101331113312133131331413315133161331713318133191332013321133221332313324133251332613327133281332913330133311333213333133341333513336133371333813339133401334113342133431334413345133461334713348133491335013351133521335313354133551335613357133581335913360133611336213363133641336513366133671336813369133701337113372133731337413375133761337713378133791338013381133821338313384133851338613387133881338913390133911339213393133941339513396133971339813399134001340113402134031340413405134061340713408134091341013411134121341313414134151341613417134181341913420134211342213423134241342513426134271342813429134301343113432134331343413435134361343713438134391344013441134421344313444134451344613447134481344913450134511345213453134541345513456134571345813459134601346113462134631346413465134661346713468134691347013471134721347313474134751347613477134781347913480134811348213483134841348513486134871348813489134901349113492134931349413495134961349713498134991350013501135021350313504135051350613507135081350913510135111351213513135141351513516135171351813519135201352113522135231352413525135261352713528135291353013531135321353313534135351353613537135381353913540135411354213543135441354513546135471354813549135501355113552135531355413555135561355713558135591356013561135621356313564135651356613567135681356913570135711357213573135741357513576135771357813579135801358113582135831358413585135861358713588135891359013591135921359313594135951359613597135981359913600136011360213603136041360513606136071360813609136101361113612136131361413615136161361713618136191362013621136221362313624136251362613627136281362913630136311363213633136341363513636136371363813639136401364113642136431364413645136461364713648136491365013651136521365313654136551365613657136581365913660136611366213663136641366513666136671366813669136701367113672136731367413675136761367713678136791368013681136821368313684136851368613687136881368913690136911369213693136941369513696136971369813699137001370113702137031370413705137061370713708137091371013711137121371313714137151371613717137181371913720137211372213723137241372513726137271372813729137301373113732137331373413735137361373713738137391374013741137421374313744137451374613747137481374913750137511375213753137541375513756137571375813759137601376113762137631376413765137661376713768137691377013771137721377313774137751377613777137781377913780137811378213783137841378513786137871378813789137901379113792137931379413795137961379713798137991380013801138021380313804138051380613807138081380913810138111381213813138141381513816138171381813819138201382113822138231382413825138261382713828138291383013831138321383313834138351383613837138381383913840138411384213843138441384513846138471384813849138501385113852138531385413855138561385713858138591386013861138621386313864138651386613867138681386913870138711387213873138741387513876138771387813879138801388113882138831388413885138861388713888138891389013891138921389313894138951389613897138981389913900139011390213903139041390513906139071390813909139101391113912139131391413915139161391713918139191392013921139221392313924139251392613927139281392913930139311393213933139341393513936139371393813939139401394113942139431394413945139461394713948139491395013951139521395313954139551395613957139581395913960139611396213963139641396513966139671396813969139701397113972139731397413975139761397713978139791398013981139821398313984139851398613987139881398913990139911399213993139941399513996139971399813999140001400114002140031400414005140061400714008140091401014011140121401314014140151401614017140181401914020140211402214023140241402514026140271402814029140301403114032140331403414035140361403714038140391404014041140421404314044140451404614047140481404914050140511405214053140541405514056140571405814059140601406114062140631406414065140661406714068140691407014071140721407314074140751407614077140781407914080140811408214083140841408514086140871408814089140901409114092140931409414095140961409714098140991410014101141021410314104141051410614107141081410914110141111411214113141141411514116141171411814119141201412114122141231412414125141261412714128141291413014131141321413314134141351413614137141381413914140141411414214143141441414514146141471414814149141501415114152141531415414155141561415714158141591416014161141621416314164141651416614167141681416914170141711417214173141741417514176141771417814179141801418114182141831418414185141861418714188141891419014191141921419314194141951419614197141981419914200142011420214203142041420514206142071420814209142101421114212142131421414215142161421714218142191422014221142221422314224142251422614227142281422914230142311423214233142341423514236142371423814239142401424114242142431424414245142461424714248142491425014251142521425314254142551425614257142581425914260142611426214263142641426514266142671426814269142701427114272142731427414275142761427714278142791428014281142821428314284142851428614287142881428914290142911429214293142941429514296142971429814299143001430114302143031430414305143061430714308143091431014311143121431314314143151431614317143181431914320143211432214323143241432514326143271432814329143301433114332143331433414335143361433714338143391434014341143421434314344143451434614347143481434914350143511435214353143541435514356143571435814359143601436114362143631436414365143661436714368143691437014371143721437314374143751437614377143781437914380143811438214383143841438514386143871438814389143901439114392143931439414395143961439714398143991440014401144021440314404144051440614407144081440914410144111441214413144141441514416144171441814419144201442114422144231442414425144261442714428144291443014431144321443314434144351443614437144381443914440144411444214443144441444514446144471444814449144501445114452144531445414455144561445714458144591446014461144621446314464144651446614467144681446914470144711447214473144741447514476144771447814479144801448114482144831448414485144861448714488144891449014491144921449314494144951449614497144981449914500145011450214503145041450514506145071450814509145101451114512145131451414515145161451714518145191452014521145221452314524145251452614527145281452914530145311453214533145341453514536145371453814539145401454114542145431454414545145461454714548145491455014551145521455314554145551455614557145581455914560145611456214563145641456514566145671456814569145701457114572145731457414575145761457714578145791458014581145821458314584145851458614587145881458914590145911459214593145941459514596145971459814599146001460114602146031460414605146061460714608146091461014611146121461314614146151461614617146181461914620146211462214623146241462514626146271462814629146301463114632146331463414635146361463714638146391464014641146421464314644146451464614647146481464914650146511465214653146541465514656146571465814659146601466114662146631466414665146661466714668146691467014671146721467314674146751467614677146781467914680146811468214683146841468514686146871468814689146901469114692146931469414695146961469714698146991470014701147021470314704147051470614707147081470914710147111471214713147141471514716147171471814719147201472114722147231472414725147261472714728147291473014731147321473314734147351473614737147381473914740147411474214743147441474514746147471474814749147501475114752147531475414755147561475714758147591476014761147621476314764147651476614767147681476914770147711477214773147741477514776147771477814779147801478114782147831478414785147861478714788147891479014791147921479314794147951479614797147981479914800148011480214803148041480514806148071480814809148101481114812148131481414815148161481714818148191482014821148221482314824148251482614827148281482914830148311483214833148341483514836148371483814839148401484114842148431484414845148461484714848148491485014851148521485314854148551485614857148581485914860148611486214863148641486514866148671486814869148701487114872148731487414875148761487714878148791488014881148821488314884148851488614887148881488914890148911489214893148941489514896148971489814899149001490114902149031490414905149061490714908149091491014911149121491314914149151491614917149181491914920149211492214923149241492514926149271492814929149301493114932149331493414935149361493714938149391494014941149421494314944149451494614947149481494914950149511495214953149541495514956149571495814959149601496114962149631496414965149661496714968149691497014971149721497314974149751497614977149781497914980149811498214983149841498514986149871498814989149901499114992149931499414995149961499714998149991500015001150021500315004150051500615007150081500915010150111501215013150141501515016150171501815019150201502115022150231502415025150261502715028150291503015031150321503315034150351503615037150381503915040150411504215043150441504515046150471504815049150501505115052150531505415055150561505715058150591506015061150621506315064150651506615067150681506915070150711507215073150741507515076150771507815079150801508115082150831508415085150861508715088150891509015091150921509315094150951509615097150981509915100151011510215103151041510515106151071510815109151101511115112151131511415115151161511715118151191512015121151221512315124151251512615127151281512915130151311513215133151341513515136151371513815139151401514115142151431514415145151461514715148151491515015151151521515315154151551515615157151581515915160151611516215163151641516515166151671516815169151701517115172151731517415175151761517715178151791518015181151821518315184151851518615187151881518915190151911519215193151941519515196151971519815199152001520115202152031520415205152061520715208152091521015211152121521315214152151521615217152181521915220152211522215223152241522515226152271522815229152301523115232152331523415235152361523715238152391524015241152421524315244152451524615247152481524915250152511525215253152541525515256152571525815259152601526115262152631526415265152661526715268152691527015271152721527315274152751527615277152781527915280152811528215283152841528515286152871528815289152901529115292152931529415295152961529715298152991530015301153021530315304153051530615307153081530915310153111531215313153141531515316153171531815319153201532115322153231532415325153261532715328153291533015331153321533315334153351533615337153381533915340153411534215343153441534515346153471534815349153501535115352153531535415355153561535715358153591536015361153621536315364153651536615367153681536915370153711537215373153741537515376153771537815379153801538115382153831538415385153861538715388153891539015391153921539315394153951539615397153981539915400154011540215403154041540515406154071540815409154101541115412154131541415415154161541715418154191542015421154221542315424154251542615427154281542915430154311543215433154341543515436154371543815439154401544115442154431544415445154461544715448154491545015451154521545315454154551545615457154581545915460154611546215463154641546515466154671546815469154701547115472154731547415475154761547715478154791548015481154821548315484154851548615487154881548915490154911549215493154941549515496154971549815499155001550115502155031550415505155061550715508155091551015511155121551315514155151551615517155181551915520155211552215523155241552515526155271552815529155301553115532155331553415535155361553715538155391554015541155421554315544155451554615547155481554915550155511555215553155541555515556 |
- /*
- * parser.c : an XML 1.0 parser, namespaces and validity support are mostly
- * implemented on top of the SAX interfaces
- *
- * References:
- * The XML specification:
- * http://www.w3.org/TR/REC-xml
- * Original 1.0 version:
- * http://www.w3.org/TR/1998/REC-xml-19980210
- * XML second edition working draft
- * http://www.w3.org/TR/2000/WD-xml-2e-20000814
- *
- * Okay this is a big file, the parser core is around 7000 lines, then it
- * is followed by the progressive parser top routines, then the various
- * high level APIs to call the parser and a few miscellaneous functions.
- * A number of helper functions and deprecated ones have been moved to
- * parserInternals.c to reduce this file size.
- * As much as possible the functions are associated with their relative
- * production in the XML specification. A few productions defining the
- * different ranges of character are actually implanted either in
- * parserInternals.h or parserInternals.c
- * The DOM tree build is realized from the default SAX callbacks in
- * the module SAX.c.
- * The routines doing the validation checks are in valid.c and called either
- * from the SAX callbacks or as standalone functions using a preparsed
- * document.
- *
- * See Copyright for the status of this software.
- *
- * daniel@veillard.com
- */
- /* To avoid EBCDIC trouble when parsing on zOS */
- #if defined(__MVS__)
- #pragma convert("ISO8859-1")
- #endif
- #define IN_LIBXML
- #include "libxml.h"
- #if defined(_WIN32) && !defined (__CYGWIN__)
- #define XML_DIR_SEP '\\'
- #else
- #define XML_DIR_SEP '/'
- #endif
- #include <stdlib.h>
- #include <limits.h>
- #include <string.h>
- #include <stdarg.h>
- #include <stddef.h>
- #include <libxml/xmlmemory.h>
- #include <libxml/threads.h>
- #include <libxml/globals.h>
- #include <libxml/tree.h>
- #include <libxml/parser.h>
- #include <libxml/parserInternals.h>
- #include <libxml/valid.h>
- #include <libxml/entities.h>
- #include <libxml/xmlerror.h>
- #include <libxml/encoding.h>
- #include <libxml/xmlIO.h>
- #include <libxml/uri.h>
- #ifdef LIBXML_CATALOG_ENABLED
- #include <libxml/catalog.h>
- #endif
- #ifdef LIBXML_SCHEMAS_ENABLED
- #include <libxml/xmlschemastypes.h>
- #include <libxml/relaxng.h>
- #endif
- #ifdef HAVE_CTYPE_H
- #include <ctype.h>
- #endif
- #ifdef HAVE_STDLIB_H
- #include <stdlib.h>
- #endif
- #ifdef HAVE_SYS_STAT_H
- #include <sys/stat.h>
- #endif
- #ifdef HAVE_FCNTL_H
- #include <fcntl.h>
- #endif
- #ifdef HAVE_UNISTD_H
- #include <unistd.h>
- #endif
- #include "buf.h"
- #include "enc.h"
- struct _xmlStartTag {
- const xmlChar *prefix;
- const xmlChar *URI;
- int line;
- int nsNr;
- };
- static void
- xmlFatalErr(xmlParserCtxtPtr ctxt, xmlParserErrors error, const char *info);
- static xmlParserCtxtPtr
- xmlCreateEntityParserCtxtInternal(const xmlChar *URL, const xmlChar *ID,
- const xmlChar *base, xmlParserCtxtPtr pctx);
- static void xmlHaltParser(xmlParserCtxtPtr ctxt);
- static int
- xmlParseElementStart(xmlParserCtxtPtr ctxt);
- static void
- xmlParseElementEnd(xmlParserCtxtPtr ctxt);
- /************************************************************************
- * *
- * Arbitrary limits set in the parser. See XML_PARSE_HUGE *
- * *
- ************************************************************************/
- #define XML_PARSER_BIG_ENTITY 1000
- #define XML_PARSER_LOT_ENTITY 5000
- /*
- * XML_PARSER_NON_LINEAR is the threshold where the ratio of parsed entity
- * replacement over the size in byte of the input indicates that you have
- * and exponential behaviour. A value of 10 correspond to at least 3 entity
- * replacement per byte of input.
- */
- #define XML_PARSER_NON_LINEAR 10
- /*
- * xmlParserEntityCheck
- *
- * Function to check non-linear entity expansion behaviour
- * This is here to detect and stop exponential linear entity expansion
- * This is not a limitation of the parser but a safety
- * boundary feature. It can be disabled with the XML_PARSE_HUGE
- * parser option.
- */
- static int
- xmlParserEntityCheck(xmlParserCtxtPtr ctxt, size_t size,
- xmlEntityPtr ent, size_t replacement)
- {
- size_t consumed = 0;
- int i;
- if ((ctxt == NULL) || (ctxt->options & XML_PARSE_HUGE))
- return (0);
- if (ctxt->lastError.code == XML_ERR_ENTITY_LOOP)
- return (1);
- /*
- * This may look absurd but is needed to detect
- * entities problems
- */
- if ((ent != NULL) && (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
- (ent->content != NULL) && (ent->checked == 0) &&
- (ctxt->errNo != XML_ERR_ENTITY_LOOP)) {
- unsigned long oldnbent = ctxt->nbentities, diff;
- xmlChar *rep;
- ent->checked = 1;
- ++ctxt->depth;
- rep = xmlStringDecodeEntities(ctxt, ent->content,
- XML_SUBSTITUTE_REF, 0, 0, 0);
- --ctxt->depth;
- if ((rep == NULL) || (ctxt->errNo == XML_ERR_ENTITY_LOOP)) {
- ent->content[0] = 0;
- }
- diff = ctxt->nbentities - oldnbent + 1;
- if (diff > INT_MAX / 2)
- diff = INT_MAX / 2;
- ent->checked = diff * 2;
- if (rep != NULL) {
- if (xmlStrchr(rep, '<'))
- ent->checked |= 1;
- xmlFree(rep);
- rep = NULL;
- }
- }
- /*
- * Prevent entity exponential check, not just replacement while
- * parsing the DTD
- * The check is potentially costly so do that only once in a thousand
- */
- if ((ctxt->instate == XML_PARSER_DTD) && (ctxt->nbentities > 10000) &&
- (ctxt->nbentities % 1024 == 0)) {
- for (i = 0;i < ctxt->inputNr;i++) {
- consumed += ctxt->inputTab[i]->consumed +
- (ctxt->inputTab[i]->cur - ctxt->inputTab[i]->base);
- }
- if (ctxt->nbentities > consumed * XML_PARSER_NON_LINEAR) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- ctxt->instate = XML_PARSER_EOF;
- return (1);
- }
- consumed = 0;
- }
- if (replacement != 0) {
- if (replacement < XML_MAX_TEXT_LENGTH)
- return(0);
- /*
- * If the volume of entity copy reaches 10 times the
- * amount of parsed data and over the large text threshold
- * then that's very likely to be an abuse.
- */
- if (ctxt->input != NULL) {
- consumed = ctxt->input->consumed +
- (ctxt->input->cur - ctxt->input->base);
- }
- consumed += ctxt->sizeentities;
- if (replacement < XML_PARSER_NON_LINEAR * consumed)
- return(0);
- } else if (size != 0) {
- /*
- * Do the check based on the replacement size of the entity
- */
- if (size < XML_PARSER_BIG_ENTITY)
- return(0);
- /*
- * A limit on the amount of text data reasonably used
- */
- if (ctxt->input != NULL) {
- consumed = ctxt->input->consumed +
- (ctxt->input->cur - ctxt->input->base);
- }
- consumed += ctxt->sizeentities;
- if ((size < XML_PARSER_NON_LINEAR * consumed) &&
- (ctxt->nbentities * 3 < XML_PARSER_NON_LINEAR * consumed))
- return (0);
- } else if (ent != NULL) {
- /*
- * use the number of parsed entities in the replacement
- */
- size = ent->checked / 2;
- /*
- * The amount of data parsed counting entities size only once
- */
- if (ctxt->input != NULL) {
- consumed = ctxt->input->consumed +
- (ctxt->input->cur - ctxt->input->base);
- }
- consumed += ctxt->sizeentities;
- /*
- * Check the density of entities for the amount of data
- * knowing an entity reference will take at least 3 bytes
- */
- if (size * 3 < consumed * XML_PARSER_NON_LINEAR)
- return (0);
- } else {
- /*
- * strange we got no data for checking
- */
- if (((ctxt->lastError.code != XML_ERR_UNDECLARED_ENTITY) &&
- (ctxt->lastError.code != XML_WAR_UNDECLARED_ENTITY)) ||
- (ctxt->nbentities <= 10000))
- return (0);
- }
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- return (1);
- }
- /**
- * xmlParserMaxDepth:
- *
- * arbitrary depth limit for the XML documents that we allow to
- * process. This is not a limitation of the parser but a safety
- * boundary feature. It can be disabled with the XML_PARSE_HUGE
- * parser option.
- */
- unsigned int xmlParserMaxDepth = 256;
- #define SAX2 1
- #define XML_PARSER_BIG_BUFFER_SIZE 300
- #define XML_PARSER_BUFFER_SIZE 100
- #define SAX_COMPAT_MODE BAD_CAST "SAX compatibility mode document"
- /**
- * XML_PARSER_CHUNK_SIZE
- *
- * When calling GROW that's the minimal amount of data
- * the parser expected to have received. It is not a hard
- * limit but an optimization when reading strings like Names
- * It is not strictly needed as long as inputs available characters
- * are followed by 0, which should be provided by the I/O level
- */
- #define XML_PARSER_CHUNK_SIZE 100
- /*
- * List of XML prefixed PI allowed by W3C specs
- */
- static const char *xmlW3CPIs[] = {
- "xml-stylesheet",
- "xml-model",
- NULL
- };
- /* DEPR void xmlParserHandleReference(xmlParserCtxtPtr ctxt); */
- static xmlEntityPtr xmlParseStringPEReference(xmlParserCtxtPtr ctxt,
- const xmlChar **str);
- static xmlParserErrors
- xmlParseExternalEntityPrivate(xmlDocPtr doc, xmlParserCtxtPtr oldctxt,
- xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *URL,
- const xmlChar *ID, xmlNodePtr *list);
- static int
- xmlCtxtUseOptionsInternal(xmlParserCtxtPtr ctxt, int options,
- const char *encoding);
- #ifdef LIBXML_LEGACY_ENABLED
- static void
- xmlAddEntityReference(xmlEntityPtr ent, xmlNodePtr firstNode,
- xmlNodePtr lastNode);
- #endif /* LIBXML_LEGACY_ENABLED */
- static xmlParserErrors
- xmlParseBalancedChunkMemoryInternal(xmlParserCtxtPtr oldctxt,
- const xmlChar *string, void *user_data, xmlNodePtr *lst);
- static int
- xmlLoadEntityContent(xmlParserCtxtPtr ctxt, xmlEntityPtr entity);
- /************************************************************************
- * *
- * Some factorized error routines *
- * *
- ************************************************************************/
- /**
- * xmlErrAttributeDup:
- * @ctxt: an XML parser context
- * @prefix: the attribute prefix
- * @localname: the attribute localname
- *
- * Handle a redefinition of attribute error
- */
- static void
- xmlErrAttributeDup(xmlParserCtxtPtr ctxt, const xmlChar * prefix,
- const xmlChar * localname)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = XML_ERR_ATTRIBUTE_REDEFINED;
- if (prefix == NULL)
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER,
- XML_ERR_ATTRIBUTE_REDEFINED, XML_ERR_FATAL, NULL, 0,
- (const char *) localname, NULL, NULL, 0, 0,
- "Attribute %s redefined\n", localname);
- else
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER,
- XML_ERR_ATTRIBUTE_REDEFINED, XML_ERR_FATAL, NULL, 0,
- (const char *) prefix, (const char *) localname,
- NULL, 0, 0, "Attribute %s:%s redefined\n", prefix,
- localname);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @extra: extra information string
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void
- xmlFatalErr(xmlParserCtxtPtr ctxt, xmlParserErrors error, const char *info)
- {
- const char *errmsg;
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- switch (error) {
- case XML_ERR_INVALID_HEX_CHARREF:
- errmsg = "CharRef: invalid hexadecimal value";
- break;
- case XML_ERR_INVALID_DEC_CHARREF:
- errmsg = "CharRef: invalid decimal value";
- break;
- case XML_ERR_INVALID_CHARREF:
- errmsg = "CharRef: invalid value";
- break;
- case XML_ERR_INTERNAL_ERROR:
- errmsg = "internal error";
- break;
- case XML_ERR_PEREF_AT_EOF:
- errmsg = "PEReference at end of document";
- break;
- case XML_ERR_PEREF_IN_PROLOG:
- errmsg = "PEReference in prolog";
- break;
- case XML_ERR_PEREF_IN_EPILOG:
- errmsg = "PEReference in epilog";
- break;
- case XML_ERR_PEREF_NO_NAME:
- errmsg = "PEReference: no name";
- break;
- case XML_ERR_PEREF_SEMICOL_MISSING:
- errmsg = "PEReference: expecting ';'";
- break;
- case XML_ERR_ENTITY_LOOP:
- errmsg = "Detected an entity reference loop";
- break;
- case XML_ERR_ENTITY_NOT_STARTED:
- errmsg = "EntityValue: \" or ' expected";
- break;
- case XML_ERR_ENTITY_PE_INTERNAL:
- errmsg = "PEReferences forbidden in internal subset";
- break;
- case XML_ERR_ENTITY_NOT_FINISHED:
- errmsg = "EntityValue: \" or ' expected";
- break;
- case XML_ERR_ATTRIBUTE_NOT_STARTED:
- errmsg = "AttValue: \" or ' expected";
- break;
- case XML_ERR_LT_IN_ATTRIBUTE:
- errmsg = "Unescaped '<' not allowed in attributes values";
- break;
- case XML_ERR_LITERAL_NOT_STARTED:
- errmsg = "SystemLiteral \" or ' expected";
- break;
- case XML_ERR_LITERAL_NOT_FINISHED:
- errmsg = "Unfinished System or Public ID \" or ' expected";
- break;
- case XML_ERR_MISPLACED_CDATA_END:
- errmsg = "Sequence ']]>' not allowed in content";
- break;
- case XML_ERR_URI_REQUIRED:
- errmsg = "SYSTEM or PUBLIC, the URI is missing";
- break;
- case XML_ERR_PUBID_REQUIRED:
- errmsg = "PUBLIC, the Public Identifier is missing";
- break;
- case XML_ERR_HYPHEN_IN_COMMENT:
- errmsg = "Comment must not contain '--' (double-hyphen)";
- break;
- case XML_ERR_PI_NOT_STARTED:
- errmsg = "xmlParsePI : no target name";
- break;
- case XML_ERR_RESERVED_XML_NAME:
- errmsg = "Invalid PI name";
- break;
- case XML_ERR_NOTATION_NOT_STARTED:
- errmsg = "NOTATION: Name expected here";
- break;
- case XML_ERR_NOTATION_NOT_FINISHED:
- errmsg = "'>' required to close NOTATION declaration";
- break;
- case XML_ERR_VALUE_REQUIRED:
- errmsg = "Entity value required";
- break;
- case XML_ERR_URI_FRAGMENT:
- errmsg = "Fragment not allowed";
- break;
- case XML_ERR_ATTLIST_NOT_STARTED:
- errmsg = "'(' required to start ATTLIST enumeration";
- break;
- case XML_ERR_NMTOKEN_REQUIRED:
- errmsg = "NmToken expected in ATTLIST enumeration";
- break;
- case XML_ERR_ATTLIST_NOT_FINISHED:
- errmsg = "')' required to finish ATTLIST enumeration";
- break;
- case XML_ERR_MIXED_NOT_STARTED:
- errmsg = "MixedContentDecl : '|' or ')*' expected";
- break;
- case XML_ERR_PCDATA_REQUIRED:
- errmsg = "MixedContentDecl : '#PCDATA' expected";
- break;
- case XML_ERR_ELEMCONTENT_NOT_STARTED:
- errmsg = "ContentDecl : Name or '(' expected";
- break;
- case XML_ERR_ELEMCONTENT_NOT_FINISHED:
- errmsg = "ContentDecl : ',' '|' or ')' expected";
- break;
- case XML_ERR_PEREF_IN_INT_SUBSET:
- errmsg =
- "PEReference: forbidden within markup decl in internal subset";
- break;
- case XML_ERR_GT_REQUIRED:
- errmsg = "expected '>'";
- break;
- case XML_ERR_CONDSEC_INVALID:
- errmsg = "XML conditional section '[' expected";
- break;
- case XML_ERR_EXT_SUBSET_NOT_FINISHED:
- errmsg = "Content error in the external subset";
- break;
- case XML_ERR_CONDSEC_INVALID_KEYWORD:
- errmsg =
- "conditional section INCLUDE or IGNORE keyword expected";
- break;
- case XML_ERR_CONDSEC_NOT_FINISHED:
- errmsg = "XML conditional section not closed";
- break;
- case XML_ERR_XMLDECL_NOT_STARTED:
- errmsg = "Text declaration '<?xml' required";
- break;
- case XML_ERR_XMLDECL_NOT_FINISHED:
- errmsg = "parsing XML declaration: '?>' expected";
- break;
- case XML_ERR_EXT_ENTITY_STANDALONE:
- errmsg = "external parsed entities cannot be standalone";
- break;
- case XML_ERR_ENTITYREF_SEMICOL_MISSING:
- errmsg = "EntityRef: expecting ';'";
- break;
- case XML_ERR_DOCTYPE_NOT_FINISHED:
- errmsg = "DOCTYPE improperly terminated";
- break;
- case XML_ERR_LTSLASH_REQUIRED:
- errmsg = "EndTag: '</' not found";
- break;
- case XML_ERR_EQUAL_REQUIRED:
- errmsg = "expected '='";
- break;
- case XML_ERR_STRING_NOT_CLOSED:
- errmsg = "String not closed expecting \" or '";
- break;
- case XML_ERR_STRING_NOT_STARTED:
- errmsg = "String not started expecting ' or \"";
- break;
- case XML_ERR_ENCODING_NAME:
- errmsg = "Invalid XML encoding name";
- break;
- case XML_ERR_STANDALONE_VALUE:
- errmsg = "standalone accepts only 'yes' or 'no'";
- break;
- case XML_ERR_DOCUMENT_EMPTY:
- errmsg = "Document is empty";
- break;
- case XML_ERR_DOCUMENT_END:
- errmsg = "Extra content at the end of the document";
- break;
- case XML_ERR_NOT_WELL_BALANCED:
- errmsg = "chunk is not well balanced";
- break;
- case XML_ERR_EXTRA_CONTENT:
- errmsg = "extra content at the end of well balanced chunk";
- break;
- case XML_ERR_VERSION_MISSING:
- errmsg = "Malformed declaration expecting version";
- break;
- case XML_ERR_NAME_TOO_LONG:
- errmsg = "Name too long use XML_PARSE_HUGE option";
- break;
- #if 0
- case:
- errmsg = "";
- break;
- #endif
- default:
- errmsg = "Unregistered error message";
- }
- if (ctxt != NULL)
- ctxt->errNo = error;
- if (info == NULL) {
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_FATAL, NULL, 0, info, NULL, NULL, 0, 0, "%s\n",
- errmsg);
- } else {
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_FATAL, NULL, 0, info, NULL, NULL, 0, 0, "%s: %s\n",
- errmsg, info);
- }
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErrMsg:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsg(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_FATAL, NULL, 0, NULL, NULL, NULL, 0, 0, "%s", msg);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlWarningMsg:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @str1: extra data
- * @str2: extra data
- *
- * Handle a warning.
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlWarningMsg(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar *str1, const xmlChar *str2)
- {
- xmlStructuredErrorFunc schannel = NULL;
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if ((ctxt != NULL) && (ctxt->sax != NULL) &&
- (ctxt->sax->initialized == XML_SAX2_MAGIC))
- schannel = ctxt->sax->serror;
- if (ctxt != NULL) {
- __xmlRaiseError(schannel,
- (ctxt->sax) ? ctxt->sax->warning : NULL,
- ctxt->userData,
- ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_WARNING, NULL, 0,
- (const char *) str1, (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- } else {
- __xmlRaiseError(schannel, NULL, NULL,
- ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_WARNING, NULL, 0,
- (const char *) str1, (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- }
- }
- /**
- * xmlValidityError:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @str1: extra data
- *
- * Handle a validity error.
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlValidityError(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar *str1, const xmlChar *str2)
- {
- xmlStructuredErrorFunc schannel = NULL;
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL) {
- ctxt->errNo = error;
- if ((ctxt->sax != NULL) && (ctxt->sax->initialized == XML_SAX2_MAGIC))
- schannel = ctxt->sax->serror;
- }
- if (ctxt != NULL) {
- __xmlRaiseError(schannel,
- ctxt->vctxt.error, ctxt->vctxt.userData,
- ctxt, NULL, XML_FROM_DTD, error,
- XML_ERR_ERROR, NULL, 0, (const char *) str1,
- (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- ctxt->valid = 0;
- } else {
- __xmlRaiseError(schannel, NULL, NULL,
- ctxt, NULL, XML_FROM_DTD, error,
- XML_ERR_ERROR, NULL, 0, (const char *) str1,
- (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- }
- }
- /**
- * xmlFatalErrMsgInt:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @val: an integer value
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsgInt(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, int val)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL,
- ctxt, NULL, XML_FROM_PARSER, error, XML_ERR_FATAL,
- NULL, 0, NULL, NULL, NULL, val, 0, msg, val);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErrMsgStrIntStr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @str1: an string info
- * @val: an integer value
- * @str2: an string info
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsgStrIntStr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar *str1, int val,
- const xmlChar *str2)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL,
- ctxt, NULL, XML_FROM_PARSER, error, XML_ERR_FATAL,
- NULL, 0, (const char *) str1, (const char *) str2,
- NULL, val, 0, msg, str1, val, str2);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErrMsgStr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @val: a string value
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsgStr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar * val)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL,
- XML_FROM_PARSER, error, XML_ERR_FATAL,
- NULL, 0, (const char *) val, NULL, NULL, 0, 0, msg,
- val);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlErrMsgStr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @val: a string value
- *
- * Handle a non fatal parser error
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlErrMsgStr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar * val)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL,
- XML_FROM_PARSER, error, XML_ERR_ERROR,
- NULL, 0, (const char *) val, NULL, NULL, 0, 0, msg,
- val);
- }
- /**
- * xmlNsErr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the message
- * @info1: extra information string
- * @info2: extra information string
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlNsErr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg,
- const xmlChar * info1, const xmlChar * info2,
- const xmlChar * info3)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_NAMESPACE, error,
- XML_ERR_ERROR, NULL, 0, (const char *) info1,
- (const char *) info2, (const char *) info3, 0, 0, msg,
- info1, info2, info3);
- if (ctxt != NULL)
- ctxt->nsWellFormed = 0;
- }
- /**
- * xmlNsWarn
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the message
- * @info1: extra information string
- * @info2: extra information string
- *
- * Handle a namespace warning error
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlNsWarn(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg,
- const xmlChar * info1, const xmlChar * info2,
- const xmlChar * info3)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_NAMESPACE, error,
- XML_ERR_WARNING, NULL, 0, (const char *) info1,
- (const char *) info2, (const char *) info3, 0, 0, msg,
- info1, info2, info3);
- }
- /************************************************************************
- * *
- * Library wide options *
- * *
- ************************************************************************/
- /**
- * xmlHasFeature:
- * @feature: the feature to be examined
- *
- * Examines if the library has been compiled with a given feature.
- *
- * Returns a non-zero value if the feature exist, otherwise zero.
- * Returns zero (0) if the feature does not exist or an unknown
- * unknown feature is requested, non-zero otherwise.
- */
- int
- xmlHasFeature(xmlFeature feature)
- {
- switch (feature) {
- case XML_WITH_THREAD:
- #ifdef LIBXML_THREAD_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_TREE:
- #ifdef LIBXML_TREE_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_OUTPUT:
- #ifdef LIBXML_OUTPUT_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_PUSH:
- #ifdef LIBXML_PUSH_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_READER:
- #ifdef LIBXML_READER_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_PATTERN:
- #ifdef LIBXML_PATTERN_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_WRITER:
- #ifdef LIBXML_WRITER_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_SAX1:
- #ifdef LIBXML_SAX1_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_FTP:
- #ifdef LIBXML_FTP_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_HTTP:
- #ifdef LIBXML_HTTP_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_VALID:
- #ifdef LIBXML_VALID_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_HTML:
- #ifdef LIBXML_HTML_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_LEGACY:
- #ifdef LIBXML_LEGACY_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_C14N:
- #ifdef LIBXML_C14N_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_CATALOG:
- #ifdef LIBXML_CATALOG_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_XPATH:
- #ifdef LIBXML_XPATH_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_XPTR:
- #ifdef LIBXML_XPTR_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_XINCLUDE:
- #ifdef LIBXML_XINCLUDE_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ICONV:
- #ifdef LIBXML_ICONV_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ISO8859X:
- #ifdef LIBXML_ISO8859X_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_UNICODE:
- #ifdef LIBXML_UNICODE_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_REGEXP:
- #ifdef LIBXML_REGEXP_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_AUTOMATA:
- #ifdef LIBXML_AUTOMATA_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_EXPR:
- #ifdef LIBXML_EXPR_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_SCHEMAS:
- #ifdef LIBXML_SCHEMAS_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_SCHEMATRON:
- #ifdef LIBXML_SCHEMATRON_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_MODULES:
- #ifdef LIBXML_MODULES_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_DEBUG:
- #ifdef LIBXML_DEBUG_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_DEBUG_MEM:
- #ifdef DEBUG_MEMORY_LOCATION
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_DEBUG_RUN:
- #ifdef LIBXML_DEBUG_RUNTIME
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ZLIB:
- #ifdef LIBXML_ZLIB_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_LZMA:
- #ifdef LIBXML_LZMA_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ICU:
- #ifdef LIBXML_ICU_ENABLED
- return(1);
- #else
- return(0);
- #endif
- default:
- break;
- }
- return(0);
- }
- /************************************************************************
- * *
- * SAX2 defaulted attributes handling *
- * *
- ************************************************************************/
- /**
- * xmlDetectSAX2:
- * @ctxt: an XML parser context
- *
- * Do the SAX2 detection and specific initialization
- */
- static void
- xmlDetectSAX2(xmlParserCtxtPtr ctxt) {
- xmlSAXHandlerPtr sax;
- if (ctxt == NULL) return;
- sax = ctxt->sax;
- #ifdef LIBXML_SAX1_ENABLED
- if ((sax) && (sax->initialized == XML_SAX2_MAGIC) &&
- ((sax->startElementNs != NULL) ||
- (sax->endElementNs != NULL) ||
- ((sax->startElement == NULL) && (sax->endElement == NULL))))
- ctxt->sax2 = 1;
- #else
- ctxt->sax2 = 1;
- #endif /* LIBXML_SAX1_ENABLED */
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- if ((ctxt->str_xml==NULL) || (ctxt->str_xmlns==NULL) ||
- (ctxt->str_xml_ns == NULL)) {
- xmlErrMemory(ctxt, NULL);
- }
- }
- typedef struct _xmlDefAttrs xmlDefAttrs;
- typedef xmlDefAttrs *xmlDefAttrsPtr;
- struct _xmlDefAttrs {
- int nbAttrs; /* number of defaulted attributes on that element */
- int maxAttrs; /* the size of the array */
- #if __STDC_VERSION__ >= 199901L
- /* Using a C99 flexible array member avoids UBSan errors. */
- const xmlChar *values[]; /* array of localname/prefix/values/external */
- #else
- const xmlChar *values[5];
- #endif
- };
- /**
- * xmlAttrNormalizeSpace:
- * @src: the source string
- * @dst: the target string
- *
- * Normalize the space in non CDATA attribute values:
- * If the attribute type is not CDATA, then the XML processor MUST further
- * process the normalized attribute value by discarding any leading and
- * trailing space (#x20) characters, and by replacing sequences of space
- * (#x20) characters by a single space (#x20) character.
- * Note that the size of dst need to be at least src, and if one doesn't need
- * to preserve dst (and it doesn't come from a dictionary or read-only) then
- * passing src as dst is just fine.
- *
- * Returns a pointer to the normalized value (dst) or NULL if no conversion
- * is needed.
- */
- static xmlChar *
- xmlAttrNormalizeSpace(const xmlChar *src, xmlChar *dst)
- {
- if ((src == NULL) || (dst == NULL))
- return(NULL);
- while (*src == 0x20) src++;
- while (*src != 0) {
- if (*src == 0x20) {
- while (*src == 0x20) src++;
- if (*src != 0)
- *dst++ = 0x20;
- } else {
- *dst++ = *src++;
- }
- }
- *dst = 0;
- if (dst == src)
- return(NULL);
- return(dst);
- }
- /**
- * xmlAttrNormalizeSpace2:
- * @src: the source string
- *
- * Normalize the space in non CDATA attribute values, a slightly more complex
- * front end to avoid allocation problems when running on attribute values
- * coming from the input.
- *
- * Returns a pointer to the normalized value (dst) or NULL if no conversion
- * is needed.
- */
- static const xmlChar *
- xmlAttrNormalizeSpace2(xmlParserCtxtPtr ctxt, xmlChar *src, int *len)
- {
- int i;
- int remove_head = 0;
- int need_realloc = 0;
- const xmlChar *cur;
- if ((ctxt == NULL) || (src == NULL) || (len == NULL))
- return(NULL);
- i = *len;
- if (i <= 0)
- return(NULL);
- cur = src;
- while (*cur == 0x20) {
- cur++;
- remove_head++;
- }
- while (*cur != 0) {
- if (*cur == 0x20) {
- cur++;
- if ((*cur == 0x20) || (*cur == 0)) {
- need_realloc = 1;
- break;
- }
- } else
- cur++;
- }
- if (need_realloc) {
- xmlChar *ret;
- ret = xmlStrndup(src + remove_head, i - remove_head + 1);
- if (ret == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- xmlAttrNormalizeSpace(ret, ret);
- *len = (int) strlen((const char *)ret);
- return(ret);
- } else if (remove_head) {
- *len -= remove_head;
- memmove(src, src + remove_head, 1 + *len);
- return(src);
- }
- return(NULL);
- }
- /**
- * xmlAddDefAttrs:
- * @ctxt: an XML parser context
- * @fullname: the element fullname
- * @fullattr: the attribute fullname
- * @value: the attribute value
- *
- * Add a defaulted attribute for an element
- */
- static void
- xmlAddDefAttrs(xmlParserCtxtPtr ctxt,
- const xmlChar *fullname,
- const xmlChar *fullattr,
- const xmlChar *value) {
- xmlDefAttrsPtr defaults;
- int len;
- const xmlChar *name;
- const xmlChar *prefix;
- /*
- * Allows to detect attribute redefinitions
- */
- if (ctxt->attsSpecial != NULL) {
- if (xmlHashLookup2(ctxt->attsSpecial, fullname, fullattr) != NULL)
- return;
- }
- if (ctxt->attsDefault == NULL) {
- ctxt->attsDefault = xmlHashCreateDict(10, ctxt->dict);
- if (ctxt->attsDefault == NULL)
- goto mem_error;
- }
- /*
- * split the element name into prefix:localname , the string found
- * are within the DTD and then not associated to namespace names.
- */
- name = xmlSplitQName3(fullname, &len);
- if (name == NULL) {
- name = xmlDictLookup(ctxt->dict, fullname, -1);
- prefix = NULL;
- } else {
- name = xmlDictLookup(ctxt->dict, name, -1);
- prefix = xmlDictLookup(ctxt->dict, fullname, len);
- }
- /*
- * make sure there is some storage
- */
- defaults = xmlHashLookup2(ctxt->attsDefault, name, prefix);
- if (defaults == NULL) {
- defaults = (xmlDefAttrsPtr) xmlMalloc(sizeof(xmlDefAttrs) +
- (4 * 5) * sizeof(const xmlChar *));
- if (defaults == NULL)
- goto mem_error;
- defaults->nbAttrs = 0;
- defaults->maxAttrs = 4;
- if (xmlHashUpdateEntry2(ctxt->attsDefault, name, prefix,
- defaults, NULL) < 0) {
- xmlFree(defaults);
- goto mem_error;
- }
- } else if (defaults->nbAttrs >= defaults->maxAttrs) {
- xmlDefAttrsPtr temp;
- temp = (xmlDefAttrsPtr) xmlRealloc(defaults, sizeof(xmlDefAttrs) +
- (2 * defaults->maxAttrs * 5) * sizeof(const xmlChar *));
- if (temp == NULL)
- goto mem_error;
- defaults = temp;
- defaults->maxAttrs *= 2;
- if (xmlHashUpdateEntry2(ctxt->attsDefault, name, prefix,
- defaults, NULL) < 0) {
- xmlFree(defaults);
- goto mem_error;
- }
- }
- /*
- * Split the element name into prefix:localname , the string found
- * are within the DTD and hen not associated to namespace names.
- */
- name = xmlSplitQName3(fullattr, &len);
- if (name == NULL) {
- name = xmlDictLookup(ctxt->dict, fullattr, -1);
- prefix = NULL;
- } else {
- name = xmlDictLookup(ctxt->dict, name, -1);
- prefix = xmlDictLookup(ctxt->dict, fullattr, len);
- }
- defaults->values[5 * defaults->nbAttrs] = name;
- defaults->values[5 * defaults->nbAttrs + 1] = prefix;
- /* intern the string and precompute the end */
- len = xmlStrlen(value);
- value = xmlDictLookup(ctxt->dict, value, len);
- defaults->values[5 * defaults->nbAttrs + 2] = value;
- defaults->values[5 * defaults->nbAttrs + 3] = value + len;
- if (ctxt->external)
- defaults->values[5 * defaults->nbAttrs + 4] = BAD_CAST "external";
- else
- defaults->values[5 * defaults->nbAttrs + 4] = NULL;
- defaults->nbAttrs++;
- return;
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return;
- }
- /**
- * xmlAddSpecialAttr:
- * @ctxt: an XML parser context
- * @fullname: the element fullname
- * @fullattr: the attribute fullname
- * @type: the attribute type
- *
- * Register this attribute type
- */
- static void
- xmlAddSpecialAttr(xmlParserCtxtPtr ctxt,
- const xmlChar *fullname,
- const xmlChar *fullattr,
- int type)
- {
- if (ctxt->attsSpecial == NULL) {
- ctxt->attsSpecial = xmlHashCreateDict(10, ctxt->dict);
- if (ctxt->attsSpecial == NULL)
- goto mem_error;
- }
- if (xmlHashLookup2(ctxt->attsSpecial, fullname, fullattr) != NULL)
- return;
- xmlHashAddEntry2(ctxt->attsSpecial, fullname, fullattr,
- (void *) (ptrdiff_t) type);
- return;
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return;
- }
- /**
- * xmlCleanSpecialAttrCallback:
- *
- * Removes CDATA attributes from the special attribute table
- */
- static void
- xmlCleanSpecialAttrCallback(void *payload, void *data,
- const xmlChar *fullname, const xmlChar *fullattr,
- const xmlChar *unused ATTRIBUTE_UNUSED) {
- xmlParserCtxtPtr ctxt = (xmlParserCtxtPtr) data;
- if (((ptrdiff_t) payload) == XML_ATTRIBUTE_CDATA) {
- xmlHashRemoveEntry2(ctxt->attsSpecial, fullname, fullattr, NULL);
- }
- }
- /**
- * xmlCleanSpecialAttr:
- * @ctxt: an XML parser context
- *
- * Trim the list of attributes defined to remove all those of type
- * CDATA as they are not special. This call should be done when finishing
- * to parse the DTD and before starting to parse the document root.
- */
- static void
- xmlCleanSpecialAttr(xmlParserCtxtPtr ctxt)
- {
- if (ctxt->attsSpecial == NULL)
- return;
- xmlHashScanFull(ctxt->attsSpecial, xmlCleanSpecialAttrCallback, ctxt);
- if (xmlHashSize(ctxt->attsSpecial) == 0) {
- xmlHashFree(ctxt->attsSpecial, NULL);
- ctxt->attsSpecial = NULL;
- }
- return;
- }
- /**
- * xmlCheckLanguageID:
- * @lang: pointer to the string value
- *
- * Checks that the value conforms to the LanguageID production:
- *
- * NOTE: this is somewhat deprecated, those productions were removed from
- * the XML Second edition.
- *
- * [33] LanguageID ::= Langcode ('-' Subcode)*
- * [34] Langcode ::= ISO639Code | IanaCode | UserCode
- * [35] ISO639Code ::= ([a-z] | [A-Z]) ([a-z] | [A-Z])
- * [36] IanaCode ::= ('i' | 'I') '-' ([a-z] | [A-Z])+
- * [37] UserCode ::= ('x' | 'X') '-' ([a-z] | [A-Z])+
- * [38] Subcode ::= ([a-z] | [A-Z])+
- *
- * The current REC reference the successors of RFC 1766, currently 5646
- *
- * http://www.rfc-editor.org/rfc/rfc5646.txt
- * langtag = language
- * ["-" script]
- * ["-" region]
- * *("-" variant)
- * *("-" extension)
- * ["-" privateuse]
- * language = 2*3ALPHA ; shortest ISO 639 code
- * ["-" extlang] ; sometimes followed by
- * ; extended language subtags
- * / 4ALPHA ; or reserved for future use
- * / 5*8ALPHA ; or registered language subtag
- *
- * extlang = 3ALPHA ; selected ISO 639 codes
- * *2("-" 3ALPHA) ; permanently reserved
- *
- * script = 4ALPHA ; ISO 15924 code
- *
- * region = 2ALPHA ; ISO 3166-1 code
- * / 3DIGIT ; UN M.49 code
- *
- * variant = 5*8alphanum ; registered variants
- * / (DIGIT 3alphanum)
- *
- * extension = singleton 1*("-" (2*8alphanum))
- *
- * ; Single alphanumerics
- * ; "x" reserved for private use
- * singleton = DIGIT ; 0 - 9
- * / %x41-57 ; A - W
- * / %x59-5A ; Y - Z
- * / %x61-77 ; a - w
- * / %x79-7A ; y - z
- *
- * it sounds right to still allow Irregular i-xxx IANA and user codes too
- * The parser below doesn't try to cope with extension or privateuse
- * that could be added but that's not interoperable anyway
- *
- * Returns 1 if correct 0 otherwise
- **/
- int
- xmlCheckLanguageID(const xmlChar * lang)
- {
- const xmlChar *cur = lang, *nxt;
- if (cur == NULL)
- return (0);
- if (((cur[0] == 'i') && (cur[1] == '-')) ||
- ((cur[0] == 'I') && (cur[1] == '-')) ||
- ((cur[0] == 'x') && (cur[1] == '-')) ||
- ((cur[0] == 'X') && (cur[1] == '-'))) {
- /*
- * Still allow IANA code and user code which were coming
- * from the previous version of the XML-1.0 specification
- * it's deprecated but we should not fail
- */
- cur += 2;
- while (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
- ((cur[0] >= 'a') && (cur[0] <= 'z')))
- cur++;
- return(cur[0] == 0);
- }
- nxt = cur;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if (nxt - cur >= 4) {
- /*
- * Reserved
- */
- if ((nxt - cur > 8) || (nxt[0] != 0))
- return(0);
- return(1);
- }
- if (nxt - cur < 2)
- return(0);
- /* we got an ISO 639 code */
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can have extlang or script or region or variant */
- if ((nxt[0] >= '0') && (nxt[0] <= '9'))
- goto region_m49;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if (nxt - cur == 4)
- goto script;
- if (nxt - cur == 2)
- goto region;
- if ((nxt - cur >= 5) && (nxt - cur <= 8))
- goto variant;
- if (nxt - cur != 3)
- return(0);
- /* we parsed an extlang */
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can have script or region or variant */
- if ((nxt[0] >= '0') && (nxt[0] <= '9'))
- goto region_m49;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if (nxt - cur == 2)
- goto region;
- if ((nxt - cur >= 5) && (nxt - cur <= 8))
- goto variant;
- if (nxt - cur != 4)
- return(0);
- /* we parsed a script */
- script:
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can have region or variant */
- if ((nxt[0] >= '0') && (nxt[0] <= '9'))
- goto region_m49;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if ((nxt - cur >= 5) && (nxt - cur <= 8))
- goto variant;
- if (nxt - cur != 2)
- return(0);
- /* we parsed a region */
- region:
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can just have a variant */
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if ((nxt - cur < 5) || (nxt - cur > 8))
- return(0);
- /* we parsed a variant */
- variant:
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- /* extensions and private use subtags not checked */
- return (1);
- region_m49:
- if (((nxt[1] >= '0') && (nxt[1] <= '9')) &&
- ((nxt[2] >= '0') && (nxt[2] <= '9'))) {
- nxt += 3;
- goto region;
- }
- return(0);
- }
- /************************************************************************
- * *
- * Parser stacks related functions and macros *
- * *
- ************************************************************************/
- static xmlEntityPtr xmlParseStringEntityRef(xmlParserCtxtPtr ctxt,
- const xmlChar ** str);
- #ifdef SAX2
- /**
- * nsPush:
- * @ctxt: an XML parser context
- * @prefix: the namespace prefix or NULL
- * @URL: the namespace name
- *
- * Pushes a new parser namespace on top of the ns stack
- *
- * Returns -1 in case of error, -2 if the namespace should be discarded
- * and the index in the stack otherwise.
- */
- static int
- nsPush(xmlParserCtxtPtr ctxt, const xmlChar *prefix, const xmlChar *URL)
- {
- if (ctxt->options & XML_PARSE_NSCLEAN) {
- int i;
- for (i = ctxt->nsNr - 2;i >= 0;i -= 2) {
- if (ctxt->nsTab[i] == prefix) {
- /* in scope */
- if (ctxt->nsTab[i + 1] == URL)
- return(-2);
- /* out of scope keep it */
- break;
- }
- }
- }
- if ((ctxt->nsMax == 0) || (ctxt->nsTab == NULL)) {
- ctxt->nsMax = 10;
- ctxt->nsNr = 0;
- ctxt->nsTab = (const xmlChar **)
- xmlMalloc(ctxt->nsMax * sizeof(xmlChar *));
- if (ctxt->nsTab == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->nsMax = 0;
- return (-1);
- }
- } else if (ctxt->nsNr >= ctxt->nsMax) {
- const xmlChar ** tmp;
- ctxt->nsMax *= 2;
- tmp = (const xmlChar **) xmlRealloc((char *) ctxt->nsTab,
- ctxt->nsMax * sizeof(ctxt->nsTab[0]));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->nsMax /= 2;
- return (-1);
- }
- ctxt->nsTab = tmp;
- }
- ctxt->nsTab[ctxt->nsNr++] = prefix;
- ctxt->nsTab[ctxt->nsNr++] = URL;
- return (ctxt->nsNr);
- }
- /**
- * nsPop:
- * @ctxt: an XML parser context
- * @nr: the number to pop
- *
- * Pops the top @nr parser prefix/namespace from the ns stack
- *
- * Returns the number of namespaces removed
- */
- static int
- nsPop(xmlParserCtxtPtr ctxt, int nr)
- {
- int i;
- if (ctxt->nsTab == NULL) return(0);
- if (ctxt->nsNr < nr) {
- xmlGenericError(xmlGenericErrorContext, "Pbm popping %d NS\n", nr);
- nr = ctxt->nsNr;
- }
- if (ctxt->nsNr <= 0)
- return (0);
- for (i = 0;i < nr;i++) {
- ctxt->nsNr--;
- ctxt->nsTab[ctxt->nsNr] = NULL;
- }
- return(nr);
- }
- #endif
- static int
- xmlCtxtGrowAttrs(xmlParserCtxtPtr ctxt, int nr) {
- const xmlChar **atts;
- int *attallocs;
- int maxatts;
- if (ctxt->atts == NULL) {
- maxatts = 55; /* allow for 10 attrs by default */
- atts = (const xmlChar **)
- xmlMalloc(maxatts * sizeof(xmlChar *));
- if (atts == NULL) goto mem_error;
- ctxt->atts = atts;
- attallocs = (int *) xmlMalloc((maxatts / 5) * sizeof(int));
- if (attallocs == NULL) goto mem_error;
- ctxt->attallocs = attallocs;
- ctxt->maxatts = maxatts;
- } else if (nr + 5 > ctxt->maxatts) {
- maxatts = (nr + 5) * 2;
- atts = (const xmlChar **) xmlRealloc((void *) ctxt->atts,
- maxatts * sizeof(const xmlChar *));
- if (atts == NULL) goto mem_error;
- ctxt->atts = atts;
- attallocs = (int *) xmlRealloc((void *) ctxt->attallocs,
- (maxatts / 5) * sizeof(int));
- if (attallocs == NULL) goto mem_error;
- ctxt->attallocs = attallocs;
- ctxt->maxatts = maxatts;
- }
- return(ctxt->maxatts);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return(-1);
- }
- /**
- * inputPush:
- * @ctxt: an XML parser context
- * @value: the parser input
- *
- * Pushes a new parser input on top of the input stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- int
- inputPush(xmlParserCtxtPtr ctxt, xmlParserInputPtr value)
- {
- if ((ctxt == NULL) || (value == NULL))
- return(-1);
- if (ctxt->inputNr >= ctxt->inputMax) {
- ctxt->inputMax *= 2;
- ctxt->inputTab =
- (xmlParserInputPtr *) xmlRealloc(ctxt->inputTab,
- ctxt->inputMax *
- sizeof(ctxt->inputTab[0]));
- if (ctxt->inputTab == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFreeInputStream(value);
- ctxt->inputMax /= 2;
- value = NULL;
- return (-1);
- }
- }
- ctxt->inputTab[ctxt->inputNr] = value;
- ctxt->input = value;
- return (ctxt->inputNr++);
- }
- /**
- * inputPop:
- * @ctxt: an XML parser context
- *
- * Pops the top parser input from the input stack
- *
- * Returns the input just removed
- */
- xmlParserInputPtr
- inputPop(xmlParserCtxtPtr ctxt)
- {
- xmlParserInputPtr ret;
- if (ctxt == NULL)
- return(NULL);
- if (ctxt->inputNr <= 0)
- return (NULL);
- ctxt->inputNr--;
- if (ctxt->inputNr > 0)
- ctxt->input = ctxt->inputTab[ctxt->inputNr - 1];
- else
- ctxt->input = NULL;
- ret = ctxt->inputTab[ctxt->inputNr];
- ctxt->inputTab[ctxt->inputNr] = NULL;
- return (ret);
- }
- /**
- * nodePush:
- * @ctxt: an XML parser context
- * @value: the element node
- *
- * Pushes a new element node on top of the node stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- int
- nodePush(xmlParserCtxtPtr ctxt, xmlNodePtr value)
- {
- if (ctxt == NULL) return(0);
- if (ctxt->nodeNr >= ctxt->nodeMax) {
- xmlNodePtr *tmp;
- tmp = (xmlNodePtr *) xmlRealloc(ctxt->nodeTab,
- ctxt->nodeMax * 2 *
- sizeof(ctxt->nodeTab[0]));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- return (-1);
- }
- ctxt->nodeTab = tmp;
- ctxt->nodeMax *= 2;
- }
- if ((((unsigned int) ctxt->nodeNr) > xmlParserMaxDepth) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INTERNAL_ERROR,
- "Excessive depth in document: %d use XML_PARSE_HUGE option\n",
- xmlParserMaxDepth);
- xmlHaltParser(ctxt);
- return(-1);
- }
- ctxt->nodeTab[ctxt->nodeNr] = value;
- ctxt->node = value;
- return (ctxt->nodeNr++);
- }
- /**
- * nodePop:
- * @ctxt: an XML parser context
- *
- * Pops the top element node from the node stack
- *
- * Returns the node just removed
- */
- xmlNodePtr
- nodePop(xmlParserCtxtPtr ctxt)
- {
- xmlNodePtr ret;
- if (ctxt == NULL) return(NULL);
- if (ctxt->nodeNr <= 0)
- return (NULL);
- ctxt->nodeNr--;
- if (ctxt->nodeNr > 0)
- ctxt->node = ctxt->nodeTab[ctxt->nodeNr - 1];
- else
- ctxt->node = NULL;
- ret = ctxt->nodeTab[ctxt->nodeNr];
- ctxt->nodeTab[ctxt->nodeNr] = NULL;
- return (ret);
- }
- /**
- * nameNsPush:
- * @ctxt: an XML parser context
- * @value: the element name
- * @prefix: the element prefix
- * @URI: the element namespace name
- * @line: the current line number for error messages
- * @nsNr: the number of namespaces pushed on the namespace table
- *
- * Pushes a new element name/prefix/URL on top of the name stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- static int
- nameNsPush(xmlParserCtxtPtr ctxt, const xmlChar * value,
- const xmlChar *prefix, const xmlChar *URI, int line, int nsNr)
- {
- xmlStartTag *tag;
- if (ctxt->nameNr >= ctxt->nameMax) {
- const xmlChar * *tmp;
- xmlStartTag *tmp2;
- ctxt->nameMax *= 2;
- tmp = (const xmlChar * *) xmlRealloc((xmlChar * *)ctxt->nameTab,
- ctxt->nameMax *
- sizeof(ctxt->nameTab[0]));
- if (tmp == NULL) {
- ctxt->nameMax /= 2;
- goto mem_error;
- }
- ctxt->nameTab = tmp;
- tmp2 = (xmlStartTag *) xmlRealloc((void * *)ctxt->pushTab,
- ctxt->nameMax *
- sizeof(ctxt->pushTab[0]));
- if (tmp2 == NULL) {
- ctxt->nameMax /= 2;
- goto mem_error;
- }
- ctxt->pushTab = tmp2;
- } else if (ctxt->pushTab == NULL) {
- ctxt->pushTab = (xmlStartTag *) xmlMalloc(ctxt->nameMax *
- sizeof(ctxt->pushTab[0]));
- if (ctxt->pushTab == NULL)
- goto mem_error;
- }
- ctxt->nameTab[ctxt->nameNr] = value;
- ctxt->name = value;
- tag = &ctxt->pushTab[ctxt->nameNr];
- tag->prefix = prefix;
- tag->URI = URI;
- tag->line = line;
- tag->nsNr = nsNr;
- return (ctxt->nameNr++);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return (-1);
- }
- #ifdef LIBXML_PUSH_ENABLED
- /**
- * nameNsPop:
- * @ctxt: an XML parser context
- *
- * Pops the top element/prefix/URI name from the name stack
- *
- * Returns the name just removed
- */
- static const xmlChar *
- nameNsPop(xmlParserCtxtPtr ctxt)
- {
- const xmlChar *ret;
- if (ctxt->nameNr <= 0)
- return (NULL);
- ctxt->nameNr--;
- if (ctxt->nameNr > 0)
- ctxt->name = ctxt->nameTab[ctxt->nameNr - 1];
- else
- ctxt->name = NULL;
- ret = ctxt->nameTab[ctxt->nameNr];
- ctxt->nameTab[ctxt->nameNr] = NULL;
- return (ret);
- }
- #endif /* LIBXML_PUSH_ENABLED */
- /**
- * namePush:
- * @ctxt: an XML parser context
- * @value: the element name
- *
- * Pushes a new element name on top of the name stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- int
- namePush(xmlParserCtxtPtr ctxt, const xmlChar * value)
- {
- if (ctxt == NULL) return (-1);
- if (ctxt->nameNr >= ctxt->nameMax) {
- const xmlChar * *tmp;
- tmp = (const xmlChar * *) xmlRealloc((xmlChar * *)ctxt->nameTab,
- ctxt->nameMax * 2 *
- sizeof(ctxt->nameTab[0]));
- if (tmp == NULL) {
- goto mem_error;
- }
- ctxt->nameTab = tmp;
- ctxt->nameMax *= 2;
- }
- ctxt->nameTab[ctxt->nameNr] = value;
- ctxt->name = value;
- return (ctxt->nameNr++);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return (-1);
- }
- /**
- * namePop:
- * @ctxt: an XML parser context
- *
- * Pops the top element name from the name stack
- *
- * Returns the name just removed
- */
- const xmlChar *
- namePop(xmlParserCtxtPtr ctxt)
- {
- const xmlChar *ret;
- if ((ctxt == NULL) || (ctxt->nameNr <= 0))
- return (NULL);
- ctxt->nameNr--;
- if (ctxt->nameNr > 0)
- ctxt->name = ctxt->nameTab[ctxt->nameNr - 1];
- else
- ctxt->name = NULL;
- ret = ctxt->nameTab[ctxt->nameNr];
- ctxt->nameTab[ctxt->nameNr] = NULL;
- return (ret);
- }
- static int spacePush(xmlParserCtxtPtr ctxt, int val) {
- if (ctxt->spaceNr >= ctxt->spaceMax) {
- int *tmp;
- ctxt->spaceMax *= 2;
- tmp = (int *) xmlRealloc(ctxt->spaceTab,
- ctxt->spaceMax * sizeof(ctxt->spaceTab[0]));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->spaceMax /=2;
- return(-1);
- }
- ctxt->spaceTab = tmp;
- }
- ctxt->spaceTab[ctxt->spaceNr] = val;
- ctxt->space = &ctxt->spaceTab[ctxt->spaceNr];
- return(ctxt->spaceNr++);
- }
- static int spacePop(xmlParserCtxtPtr ctxt) {
- int ret;
- if (ctxt->spaceNr <= 0) return(0);
- ctxt->spaceNr--;
- if (ctxt->spaceNr > 0)
- ctxt->space = &ctxt->spaceTab[ctxt->spaceNr - 1];
- else
- ctxt->space = &ctxt->spaceTab[0];
- ret = ctxt->spaceTab[ctxt->spaceNr];
- ctxt->spaceTab[ctxt->spaceNr] = -1;
- return(ret);
- }
- /*
- * Macros for accessing the content. Those should be used only by the parser,
- * and not exported.
- *
- * Dirty macros, i.e. one often need to make assumption on the context to
- * use them
- *
- * CUR_PTR return the current pointer to the xmlChar to be parsed.
- * To be used with extreme caution since operations consuming
- * characters may move the input buffer to a different location !
- * CUR returns the current xmlChar value, i.e. a 8 bit value if compiled
- * This should be used internally by the parser
- * only to compare to ASCII values otherwise it would break when
- * running with UTF-8 encoding.
- * RAW same as CUR but in the input buffer, bypass any token
- * extraction that may have been done
- * NXT(n) returns the n'th next xmlChar. Same as CUR is should be used only
- * to compare on ASCII based substring.
- * SKIP(n) Skip n xmlChar, and must also be used only to skip ASCII defined
- * strings without newlines within the parser.
- * NEXT1(l) Skip 1 xmlChar, and must also be used only to skip 1 non-newline ASCII
- * defined char within the parser.
- * Clean macros, not dependent of an ASCII context, expect UTF-8 encoding
- *
- * NEXT Skip to the next character, this does the proper decoding
- * in UTF-8 mode. It also pop-up unfinished entities on the fly.
- * NEXTL(l) Skip the current unicode character of l xmlChars long.
- * CUR_CHAR(l) returns the current unicode character (int), set l
- * to the number of xmlChars used for the encoding [0-5].
- * CUR_SCHAR same but operate on a string instead of the context
- * COPY_BUF copy the current unicode char to the target buffer, increment
- * the index
- * GROW, SHRINK handling of input buffers
- */
- #define RAW (*ctxt->input->cur)
- #define CUR (*ctxt->input->cur)
- #define NXT(val) ctxt->input->cur[(val)]
- #define CUR_PTR ctxt->input->cur
- #define BASE_PTR ctxt->input->base
- #define CMP4( s, c1, c2, c3, c4 ) \
- ( ((unsigned char *) s)[ 0 ] == c1 && ((unsigned char *) s)[ 1 ] == c2 && \
- ((unsigned char *) s)[ 2 ] == c3 && ((unsigned char *) s)[ 3 ] == c4 )
- #define CMP5( s, c1, c2, c3, c4, c5 ) \
- ( CMP4( s, c1, c2, c3, c4 ) && ((unsigned char *) s)[ 4 ] == c5 )
- #define CMP6( s, c1, c2, c3, c4, c5, c6 ) \
- ( CMP5( s, c1, c2, c3, c4, c5 ) && ((unsigned char *) s)[ 5 ] == c6 )
- #define CMP7( s, c1, c2, c3, c4, c5, c6, c7 ) \
- ( CMP6( s, c1, c2, c3, c4, c5, c6 ) && ((unsigned char *) s)[ 6 ] == c7 )
- #define CMP8( s, c1, c2, c3, c4, c5, c6, c7, c8 ) \
- ( CMP7( s, c1, c2, c3, c4, c5, c6, c7 ) && ((unsigned char *) s)[ 7 ] == c8 )
- #define CMP9( s, c1, c2, c3, c4, c5, c6, c7, c8, c9 ) \
- ( CMP8( s, c1, c2, c3, c4, c5, c6, c7, c8 ) && \
- ((unsigned char *) s)[ 8 ] == c9 )
- #define CMP10( s, c1, c2, c3, c4, c5, c6, c7, c8, c9, c10 ) \
- ( CMP9( s, c1, c2, c3, c4, c5, c6, c7, c8, c9 ) && \
- ((unsigned char *) s)[ 9 ] == c10 )
- #define SKIP(val) do { \
- ctxt->input->cur += (val),ctxt->input->col+=(val); \
- if (*ctxt->input->cur == 0) \
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK); \
- } while (0)
- #define SKIPL(val) do { \
- int skipl; \
- for(skipl=0; skipl<val; skipl++) { \
- if (*(ctxt->input->cur) == '\n') { \
- ctxt->input->line++; ctxt->input->col = 1; \
- } else ctxt->input->col++; \
- ctxt->input->cur++; \
- } \
- if (*ctxt->input->cur == 0) \
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK); \
- } while (0)
- #define SHRINK if ((ctxt->progressive == 0) && \
- (ctxt->input->cur - ctxt->input->base > 2 * INPUT_CHUNK) && \
- (ctxt->input->end - ctxt->input->cur < 2 * INPUT_CHUNK)) \
- xmlSHRINK (ctxt);
- static void xmlSHRINK (xmlParserCtxtPtr ctxt) {
- xmlParserInputShrink(ctxt->input);
- if (*ctxt->input->cur == 0)
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- }
- #define GROW if ((ctxt->progressive == 0) && \
- (ctxt->input->end - ctxt->input->cur < INPUT_CHUNK)) \
- xmlGROW (ctxt);
- static void xmlGROW (xmlParserCtxtPtr ctxt) {
- ptrdiff_t curEnd = ctxt->input->end - ctxt->input->cur;
- ptrdiff_t curBase = ctxt->input->cur - ctxt->input->base;
- if (((curEnd > XML_MAX_LOOKUP_LIMIT) ||
- (curBase > XML_MAX_LOOKUP_LIMIT)) &&
- ((ctxt->input->buf) &&
- (ctxt->input->buf->readcallback != xmlInputReadCallbackNop)) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR, "Huge input lookup");
- xmlHaltParser(ctxt);
- return;
- }
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- if ((ctxt->input->cur > ctxt->input->end) ||
- (ctxt->input->cur < ctxt->input->base)) {
- xmlHaltParser(ctxt);
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR, "cur index out of bound");
- return;
- }
- if ((ctxt->input->cur != NULL) && (*ctxt->input->cur == 0))
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- }
- #define SKIP_BLANKS xmlSkipBlankChars(ctxt)
- #define NEXT xmlNextChar(ctxt)
- #define NEXT1 { \
- ctxt->input->col++; \
- ctxt->input->cur++; \
- if (*ctxt->input->cur == 0) \
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK); \
- }
- #define NEXTL(l) do { \
- if (*(ctxt->input->cur) == '\n') { \
- ctxt->input->line++; ctxt->input->col = 1; \
- } else ctxt->input->col++; \
- ctxt->input->cur += l; \
- } while (0)
- #define CUR_CHAR(l) xmlCurrentChar(ctxt, &l)
- #define CUR_SCHAR(s, l) xmlStringCurrentChar(ctxt, s, &l)
- #define COPY_BUF(l,b,i,v) \
- if (l == 1) b[i++] = (xmlChar) v; \
- else i += xmlCopyCharMultiByte(&b[i],v)
- /**
- * xmlSkipBlankChars:
- * @ctxt: the XML parser context
- *
- * skip all blanks character found at that point in the input streams.
- * It pops up finished entities in the process if allowable at that point.
- *
- * Returns the number of space chars skipped
- */
- int
- xmlSkipBlankChars(xmlParserCtxtPtr ctxt) {
- int res = 0;
- /*
- * It's Okay to use CUR/NEXT here since all the blanks are on
- * the ASCII range.
- */
- if (((ctxt->inputNr == 1) && (ctxt->instate != XML_PARSER_DTD)) ||
- (ctxt->instate == XML_PARSER_START)) {
- const xmlChar *cur;
- /*
- * if we are in the document content, go really fast
- */
- cur = ctxt->input->cur;
- while (IS_BLANK_CH(*cur)) {
- if (*cur == '\n') {
- ctxt->input->line++; ctxt->input->col = 1;
- } else {
- ctxt->input->col++;
- }
- cur++;
- res++;
- if (*cur == 0) {
- ctxt->input->cur = cur;
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- cur = ctxt->input->cur;
- }
- }
- ctxt->input->cur = cur;
- } else {
- int expandPE = ((ctxt->external != 0) || (ctxt->inputNr != 1));
- while (1) {
- if (IS_BLANK_CH(CUR)) { /* CHECKED tstblanks.xml */
- NEXT;
- } else if (CUR == '%') {
- /*
- * Need to handle support of entities branching here
- */
- if ((expandPE == 0) || (IS_BLANK_CH(NXT(1))) || (NXT(1) == 0))
- break;
- xmlParsePEReference(ctxt);
- } else if (CUR == 0) {
- if (ctxt->inputNr <= 1)
- break;
- xmlPopInput(ctxt);
- } else {
- break;
- }
- /*
- * Also increase the counter when entering or exiting a PERef.
- * The spec says: "When a parameter-entity reference is recognized
- * in the DTD and included, its replacement text MUST be enlarged
- * by the attachment of one leading and one following space (#x20)
- * character."
- */
- res++;
- }
- }
- return(res);
- }
- /************************************************************************
- * *
- * Commodity functions to handle entities *
- * *
- ************************************************************************/
- /**
- * xmlPopInput:
- * @ctxt: an XML parser context
- *
- * xmlPopInput: the current input pointed by ctxt->input came to an end
- * pop it and return the next char.
- *
- * Returns the current xmlChar in the parser context
- */
- xmlChar
- xmlPopInput(xmlParserCtxtPtr ctxt) {
- if ((ctxt == NULL) || (ctxt->inputNr <= 1)) return(0);
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "Popping input %d\n", ctxt->inputNr);
- if ((ctxt->inputNr > 1) && (ctxt->inSubset == 0) &&
- (ctxt->instate != XML_PARSER_EOF))
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "Unfinished entity outside the DTD");
- xmlFreeInputStream(inputPop(ctxt));
- if (*ctxt->input->cur == 0)
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- return(CUR);
- }
- /**
- * xmlPushInput:
- * @ctxt: an XML parser context
- * @input: an XML parser input fragment (entity, XML fragment ...).
- *
- * xmlPushInput: switch to a new input stream which is stacked on top
- * of the previous one(s).
- * Returns -1 in case of error or the index in the input stack
- */
- int
- xmlPushInput(xmlParserCtxtPtr ctxt, xmlParserInputPtr input) {
- int ret;
- if (input == NULL) return(-1);
- if (xmlParserDebugEntities) {
- if ((ctxt->input != NULL) && (ctxt->input->filename))
- xmlGenericError(xmlGenericErrorContext,
- "%s(%d): ", ctxt->input->filename,
- ctxt->input->line);
- xmlGenericError(xmlGenericErrorContext,
- "Pushing input %d : %.30s\n", ctxt->inputNr+1, input->cur);
- }
- if (((ctxt->inputNr > 40) && ((ctxt->options & XML_PARSE_HUGE) == 0)) ||
- (ctxt->inputNr > 1024)) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- while (ctxt->inputNr > 1)
- xmlFreeInputStream(inputPop(ctxt));
- return(-1);
- }
- ret = inputPush(ctxt, input);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- GROW;
- return(ret);
- }
- /**
- * xmlParseCharRef:
- * @ctxt: an XML parser context
- *
- * parse Reference declarations
- *
- * [66] CharRef ::= '&#' [0-9]+ ';' |
- * '&#x' [0-9a-fA-F]+ ';'
- *
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- *
- * Returns the value parsed (as an int), 0 in case of error
- */
- int
- xmlParseCharRef(xmlParserCtxtPtr ctxt) {
- int val = 0;
- int count = 0;
- /*
- * Using RAW/CUR/NEXT is okay since we are working on ASCII range here
- */
- if ((RAW == '&') && (NXT(1) == '#') &&
- (NXT(2) == 'x')) {
- SKIP(3);
- GROW;
- while (RAW != ';') { /* loop blocked by count */
- if (count++ > 20) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(0);
- }
- if ((RAW >= '0') && (RAW <= '9'))
- val = val * 16 + (CUR - '0');
- else if ((RAW >= 'a') && (RAW <= 'f') && (count < 20))
- val = val * 16 + (CUR - 'a') + 10;
- else if ((RAW >= 'A') && (RAW <= 'F') && (count < 20))
- val = val * 16 + (CUR - 'A') + 10;
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_HEX_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- NEXT;
- count++;
- }
- if (RAW == ';') {
- /* on purpose to avoid reentrancy problems with NEXT and SKIP */
- ctxt->input->col++;
- ctxt->input->cur++;
- }
- } else if ((RAW == '&') && (NXT(1) == '#')) {
- SKIP(2);
- GROW;
- while (RAW != ';') { /* loop blocked by count */
- if (count++ > 20) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(0);
- }
- if ((RAW >= '0') && (RAW <= '9'))
- val = val * 10 + (CUR - '0');
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_DEC_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- NEXT;
- count++;
- }
- if (RAW == ';') {
- /* on purpose to avoid reentrancy problems with NEXT and SKIP */
- ctxt->input->col++;
- ctxt->input->cur++;
- }
- } else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_CHARREF, NULL);
- }
- /*
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- */
- if (val >= 0x110000) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseCharRef: character reference out of bounds\n",
- val);
- } else if (IS_CHAR(val)) {
- return(val);
- } else {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseCharRef: invalid xmlChar value %d\n",
- val);
- }
- return(0);
- }
- /**
- * xmlParseStringCharRef:
- * @ctxt: an XML parser context
- * @str: a pointer to an index in the string
- *
- * parse Reference declarations, variant parsing from a string rather
- * than an an input flow.
- *
- * [66] CharRef ::= '&#' [0-9]+ ';' |
- * '&#x' [0-9a-fA-F]+ ';'
- *
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- *
- * Returns the value parsed (as an int), 0 in case of error, str will be
- * updated to the current value of the index
- */
- static int
- xmlParseStringCharRef(xmlParserCtxtPtr ctxt, const xmlChar **str) {
- const xmlChar *ptr;
- xmlChar cur;
- int val = 0;
- if ((str == NULL) || (*str == NULL)) return(0);
- ptr = *str;
- cur = *ptr;
- if ((cur == '&') && (ptr[1] == '#') && (ptr[2] == 'x')) {
- ptr += 3;
- cur = *ptr;
- while (cur != ';') { /* Non input consuming loop */
- if ((cur >= '0') && (cur <= '9'))
- val = val * 16 + (cur - '0');
- else if ((cur >= 'a') && (cur <= 'f'))
- val = val * 16 + (cur - 'a') + 10;
- else if ((cur >= 'A') && (cur <= 'F'))
- val = val * 16 + (cur - 'A') + 10;
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_HEX_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- ptr++;
- cur = *ptr;
- }
- if (cur == ';')
- ptr++;
- } else if ((cur == '&') && (ptr[1] == '#')){
- ptr += 2;
- cur = *ptr;
- while (cur != ';') { /* Non input consuming loops */
- if ((cur >= '0') && (cur <= '9'))
- val = val * 10 + (cur - '0');
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_DEC_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- ptr++;
- cur = *ptr;
- }
- if (cur == ';')
- ptr++;
- } else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_CHARREF, NULL);
- return(0);
- }
- *str = ptr;
- /*
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- */
- if (val >= 0x110000) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseStringCharRef: character reference out of bounds\n",
- val);
- } else if (IS_CHAR(val)) {
- return(val);
- } else {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseStringCharRef: invalid xmlChar value %d\n",
- val);
- }
- return(0);
- }
- /**
- * xmlParserHandlePEReference:
- * @ctxt: the parser context
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive
- * reference to itself, either directly or indirectly.
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", ... ... The declaration of a parameter
- * entity must precede any reference to it...
- *
- * [ VC: Entity Declared ]
- * In a document with an external subset or external parameter entities
- * with "standalone='no'", ... ... The declaration of a parameter entity
- * must precede any reference to it...
- *
- * [ WFC: In DTD ]
- * Parameter-entity references may only appear in the DTD.
- * NOTE: misleading but this is handled.
- *
- * A PEReference may have been detected in the current input stream
- * the handling is done accordingly to
- * http://www.w3.org/TR/REC-xml#entproc
- * i.e.
- * - Included in literal in entity values
- * - Included as Parameter Entity reference within DTDs
- */
- void
- xmlParserHandlePEReference(xmlParserCtxtPtr ctxt) {
- switch(ctxt->instate) {
- case XML_PARSER_CDATA_SECTION:
- return;
- case XML_PARSER_COMMENT:
- return;
- case XML_PARSER_START_TAG:
- return;
- case XML_PARSER_END_TAG:
- return;
- case XML_PARSER_EOF:
- xmlFatalErr(ctxt, XML_ERR_PEREF_AT_EOF, NULL);
- return;
- case XML_PARSER_PROLOG:
- case XML_PARSER_START:
- case XML_PARSER_MISC:
- xmlFatalErr(ctxt, XML_ERR_PEREF_IN_PROLOG, NULL);
- return;
- case XML_PARSER_ENTITY_DECL:
- case XML_PARSER_CONTENT:
- case XML_PARSER_ATTRIBUTE_VALUE:
- case XML_PARSER_PI:
- case XML_PARSER_SYSTEM_LITERAL:
- case XML_PARSER_PUBLIC_LITERAL:
- /* we just ignore it there */
- return;
- case XML_PARSER_EPILOG:
- xmlFatalErr(ctxt, XML_ERR_PEREF_IN_EPILOG, NULL);
- return;
- case XML_PARSER_ENTITY_VALUE:
- /*
- * NOTE: in the case of entity values, we don't do the
- * substitution here since we need the literal
- * entity value to be able to save the internal
- * subset of the document.
- * This will be handled by xmlStringDecodeEntities
- */
- return;
- case XML_PARSER_DTD:
- /*
- * [WFC: Well-Formedness Constraint: PEs in Internal Subset]
- * In the internal DTD subset, parameter-entity references
- * can occur only where markup declarations can occur, not
- * within markup declarations.
- * In that case this is handled in xmlParseMarkupDecl
- */
- if ((ctxt->external == 0) && (ctxt->inputNr == 1))
- return;
- if (IS_BLANK_CH(NXT(1)) || NXT(1) == 0)
- return;
- break;
- case XML_PARSER_IGNORE:
- return;
- }
- xmlParsePEReference(ctxt);
- }
- /*
- * Macro used to grow the current buffer.
- * buffer##_size is expected to be a size_t
- * mem_error: is expected to handle memory allocation failures
- */
- #define growBuffer(buffer, n) { \
- xmlChar *tmp; \
- size_t new_size = buffer##_size * 2 + n; \
- if (new_size < buffer##_size) goto mem_error; \
- tmp = (xmlChar *) xmlRealloc(buffer, new_size); \
- if (tmp == NULL) goto mem_error; \
- buffer = tmp; \
- buffer##_size = new_size; \
- }
- /**
- * xmlStringLenDecodeEntities:
- * @ctxt: the parser context
- * @str: the input string
- * @len: the string length
- * @what: combination of XML_SUBSTITUTE_REF and XML_SUBSTITUTE_PEREF
- * @end: an end marker xmlChar, 0 if none
- * @end2: an end marker xmlChar, 0 if none
- * @end3: an end marker xmlChar, 0 if none
- *
- * Takes a entity string content and process to do the adequate substitutions.
- *
- * [67] Reference ::= EntityRef | CharRef
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * Returns A newly allocated string with the substitution done. The caller
- * must deallocate it !
- */
- xmlChar *
- xmlStringLenDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int len,
- int what, xmlChar end, xmlChar end2, xmlChar end3) {
- xmlChar *buffer = NULL;
- size_t buffer_size = 0;
- size_t nbchars = 0;
- xmlChar *current = NULL;
- xmlChar *rep = NULL;
- const xmlChar *last;
- xmlEntityPtr ent;
- int c,l;
- if ((ctxt == NULL) || (str == NULL) || (len < 0))
- return(NULL);
- last = str + len;
- if (((ctxt->depth > 40) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) ||
- (ctxt->depth > 1024)) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- return(NULL);
- }
- /*
- * allocate a translation buffer.
- */
- buffer_size = XML_PARSER_BIG_BUFFER_SIZE;
- buffer = (xmlChar *) xmlMallocAtomic(buffer_size);
- if (buffer == NULL) goto mem_error;
- /*
- * OK loop until we reach one of the ending char or a size limit.
- * we are operating on already parsed values.
- */
- if (str < last)
- c = CUR_SCHAR(str, l);
- else
- c = 0;
- while ((c != 0) && (c != end) && /* non input consuming loop */
- (c != end2) && (c != end3) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- if (c == 0) break;
- if ((c == '&') && (str[1] == '#')) {
- int val = xmlParseStringCharRef(ctxt, &str);
- if (val == 0)
- goto int_error;
- COPY_BUF(0,buffer,nbchars,val);
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- } else if ((c == '&') && (what & XML_SUBSTITUTE_REF)) {
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "String decoding Entity Reference: %.30s\n",
- str);
- ent = xmlParseStringEntityRef(ctxt, &str);
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- if (ent != NULL)
- ctxt->nbentities += ent->checked / 2;
- if ((ent != NULL) &&
- (ent->etype == XML_INTERNAL_PREDEFINED_ENTITY)) {
- if (ent->content != NULL) {
- COPY_BUF(0,buffer,nbchars,ent->content[0]);
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- } else {
- xmlFatalErrMsg(ctxt, XML_ERR_INTERNAL_ERROR,
- "predefined entity has no content\n");
- goto int_error;
- }
- } else if ((ent != NULL) && (ent->content != NULL)) {
- ctxt->depth++;
- rep = xmlStringDecodeEntities(ctxt, ent->content, what,
- 0, 0, 0);
- ctxt->depth--;
- if (rep == NULL) {
- ent->content[0] = 0;
- goto int_error;
- }
- current = rep;
- while (*current != 0) { /* non input consuming loop */
- buffer[nbchars++] = *current++;
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- if (xmlParserEntityCheck(ctxt, nbchars, ent, 0))
- goto int_error;
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- }
- xmlFree(rep);
- rep = NULL;
- } else if (ent != NULL) {
- int i = xmlStrlen(ent->name);
- const xmlChar *cur = ent->name;
- buffer[nbchars++] = '&';
- if (nbchars + i + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, i + XML_PARSER_BUFFER_SIZE);
- }
- for (;i > 0;i--)
- buffer[nbchars++] = *cur++;
- buffer[nbchars++] = ';';
- }
- } else if (c == '%' && (what & XML_SUBSTITUTE_PEREF)) {
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "String decoding PE Reference: %.30s\n", str);
- ent = xmlParseStringPEReference(ctxt, &str);
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- if (ent != NULL)
- ctxt->nbentities += ent->checked / 2;
- if (ent != NULL) {
- if (ent->content == NULL) {
- /*
- * Note: external parsed entities will not be loaded,
- * it is not required for a non-validating parser to
- * complete external PEReferences coming from the
- * internal subset
- */
- if (((ctxt->options & XML_PARSE_NOENT) != 0) ||
- ((ctxt->options & XML_PARSE_DTDVALID) != 0) ||
- (ctxt->validate != 0)) {
- xmlLoadEntityContent(ctxt, ent);
- } else {
- xmlWarningMsg(ctxt, XML_ERR_ENTITY_PROCESSING,
- "not validating will not read content for PE entity %s\n",
- ent->name, NULL);
- }
- }
- ctxt->depth++;
- rep = xmlStringDecodeEntities(ctxt, ent->content, what,
- 0, 0, 0);
- ctxt->depth--;
- if (rep == NULL) {
- if (ent->content != NULL)
- ent->content[0] = 0;
- goto int_error;
- }
- current = rep;
- while (*current != 0) { /* non input consuming loop */
- buffer[nbchars++] = *current++;
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- if (xmlParserEntityCheck(ctxt, nbchars, ent, 0))
- goto int_error;
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- }
- xmlFree(rep);
- rep = NULL;
- }
- } else {
- COPY_BUF(l,buffer,nbchars,c);
- str += l;
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- }
- if (str < last)
- c = CUR_SCHAR(str, l);
- else
- c = 0;
- }
- buffer[nbchars] = 0;
- return(buffer);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- int_error:
- if (rep != NULL)
- xmlFree(rep);
- if (buffer != NULL)
- xmlFree(buffer);
- return(NULL);
- }
- /**
- * xmlStringDecodeEntities:
- * @ctxt: the parser context
- * @str: the input string
- * @what: combination of XML_SUBSTITUTE_REF and XML_SUBSTITUTE_PEREF
- * @end: an end marker xmlChar, 0 if none
- * @end2: an end marker xmlChar, 0 if none
- * @end3: an end marker xmlChar, 0 if none
- *
- * Takes a entity string content and process to do the adequate substitutions.
- *
- * [67] Reference ::= EntityRef | CharRef
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * Returns A newly allocated string with the substitution done. The caller
- * must deallocate it !
- */
- xmlChar *
- xmlStringDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int what,
- xmlChar end, xmlChar end2, xmlChar end3) {
- if ((ctxt == NULL) || (str == NULL)) return(NULL);
- return(xmlStringLenDecodeEntities(ctxt, str, xmlStrlen(str), what,
- end, end2, end3));
- }
- /************************************************************************
- * *
- * Commodity functions, cleanup needed ? *
- * *
- ************************************************************************/
- /**
- * areBlanks:
- * @ctxt: an XML parser context
- * @str: a xmlChar *
- * @len: the size of @str
- * @blank_chars: we know the chars are blanks
- *
- * Is this a sequence of blank chars that one can ignore ?
- *
- * Returns 1 if ignorable 0 otherwise.
- */
- static int areBlanks(xmlParserCtxtPtr ctxt, const xmlChar *str, int len,
- int blank_chars) {
- int i, ret;
- xmlNodePtr lastChild;
- /*
- * Don't spend time trying to differentiate them, the same callback is
- * used !
- */
- if (ctxt->sax->ignorableWhitespace == ctxt->sax->characters)
- return(0);
- /*
- * Check for xml:space value.
- */
- if ((ctxt->space == NULL) || (*(ctxt->space) == 1) ||
- (*(ctxt->space) == -2))
- return(0);
- /*
- * Check that the string is made of blanks
- */
- if (blank_chars == 0) {
- for (i = 0;i < len;i++)
- if (!(IS_BLANK_CH(str[i]))) return(0);
- }
- /*
- * Look if the element is mixed content in the DTD if available
- */
- if (ctxt->node == NULL) return(0);
- if (ctxt->myDoc != NULL) {
- ret = xmlIsMixedElement(ctxt->myDoc, ctxt->node->name);
- if (ret == 0) return(1);
- if (ret == 1) return(0);
- }
- /*
- * Otherwise, heuristic :-\
- */
- if ((RAW != '<') && (RAW != 0xD)) return(0);
- if ((ctxt->node->children == NULL) &&
- (RAW == '<') && (NXT(1) == '/')) return(0);
- lastChild = xmlGetLastChild(ctxt->node);
- if (lastChild == NULL) {
- if ((ctxt->node->type != XML_ELEMENT_NODE) &&
- (ctxt->node->content != NULL)) return(0);
- } else if (xmlNodeIsText(lastChild))
- return(0);
- else if ((ctxt->node->children != NULL) &&
- (xmlNodeIsText(ctxt->node->children)))
- return(0);
- return(1);
- }
- /************************************************************************
- * *
- * Extra stuff for namespace support *
- * Relates to http://www.w3.org/TR/WD-xml-names *
- * *
- ************************************************************************/
- /**
- * xmlSplitQName:
- * @ctxt: an XML parser context
- * @name: an XML parser context
- * @prefix: a xmlChar **
- *
- * parse an UTF8 encoded XML qualified name string
- *
- * [NS 5] QName ::= (Prefix ':')? LocalPart
- *
- * [NS 6] Prefix ::= NCName
- *
- * [NS 7] LocalPart ::= NCName
- *
- * Returns the local part, and prefix is updated
- * to get the Prefix if any.
- */
- xmlChar *
- xmlSplitQName(xmlParserCtxtPtr ctxt, const xmlChar *name, xmlChar **prefix) {
- xmlChar buf[XML_MAX_NAMELEN + 5];
- xmlChar *buffer = NULL;
- int len = 0;
- int max = XML_MAX_NAMELEN;
- xmlChar *ret = NULL;
- const xmlChar *cur = name;
- int c;
- if (prefix == NULL) return(NULL);
- *prefix = NULL;
- if (cur == NULL) return(NULL);
- #ifndef XML_XML_NAMESPACE
- /* xml: prefix is not really a namespace */
- if ((cur[0] == 'x') && (cur[1] == 'm') &&
- (cur[2] == 'l') && (cur[3] == ':'))
- return(xmlStrdup(name));
- #endif
- /* nasty but well=formed */
- if (cur[0] == ':')
- return(xmlStrdup(name));
- c = *cur++;
- while ((c != 0) && (c != ':') && (len < max)) { /* tested bigname.xml */
- buf[len++] = c;
- c = *cur++;
- }
- if (len >= max) {
- /*
- * Okay someone managed to make a huge name, so he's ready to pay
- * for the processing speed.
- */
- max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while ((c != 0) && (c != ':')) { /* tested bigname.xml */
- if (len + 10 > max) {
- xmlChar *tmp;
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buffer);
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- buffer = tmp;
- }
- buffer[len++] = c;
- c = *cur++;
- }
- buffer[len] = 0;
- }
- if ((c == ':') && (*cur == 0)) {
- if (buffer != NULL)
- xmlFree(buffer);
- *prefix = NULL;
- return(xmlStrdup(name));
- }
- if (buffer == NULL)
- ret = xmlStrndup(buf, len);
- else {
- ret = buffer;
- buffer = NULL;
- max = XML_MAX_NAMELEN;
- }
- if (c == ':') {
- c = *cur;
- *prefix = ret;
- if (c == 0) {
- return(xmlStrndup(BAD_CAST "", 0));
- }
- len = 0;
- /*
- * Check that the first character is proper to start
- * a new name
- */
- if (!(((c >= 0x61) && (c <= 0x7A)) ||
- ((c >= 0x41) && (c <= 0x5A)) ||
- (c == '_') || (c == ':'))) {
- int l;
- int first = CUR_SCHAR(cur, l);
- if (!IS_LETTER(first) && (first != '_')) {
- xmlFatalErrMsgStr(ctxt, XML_NS_ERR_QNAME,
- "Name %s is not XML Namespace compliant\n",
- name);
- }
- }
- cur++;
- while ((c != 0) && (len < max)) { /* tested bigname2.xml */
- buf[len++] = c;
- c = *cur++;
- }
- if (len >= max) {
- /*
- * Okay someone managed to make a huge name, so he's ready to pay
- * for the processing speed.
- */
- max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while (c != 0) { /* tested bigname2.xml */
- if (len + 10 > max) {
- xmlChar *tmp;
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buffer);
- return(NULL);
- }
- buffer = tmp;
- }
- buffer[len++] = c;
- c = *cur++;
- }
- buffer[len] = 0;
- }
- if (buffer == NULL)
- ret = xmlStrndup(buf, len);
- else {
- ret = buffer;
- }
- }
- return(ret);
- }
- /************************************************************************
- * *
- * The parser itself *
- * Relates to http://www.w3.org/TR/REC-xml *
- * *
- ************************************************************************/
- /************************************************************************
- * *
- * Routines to parse Name, NCName and NmToken *
- * *
- ************************************************************************/
- #ifdef DEBUG
- static unsigned long nbParseName = 0;
- static unsigned long nbParseNmToken = 0;
- static unsigned long nbParseNCName = 0;
- static unsigned long nbParseNCNameComplex = 0;
- static unsigned long nbParseNameComplex = 0;
- static unsigned long nbParseStringName = 0;
- #endif
- /*
- * The two following functions are related to the change of accepted
- * characters for Name and NmToken in the Revision 5 of XML-1.0
- * They correspond to the modified production [4] and the new production [4a]
- * changes in that revision. Also note that the macros used for the
- * productions Letter, Digit, CombiningChar and Extender are not needed
- * anymore.
- * We still keep compatibility to pre-revision5 parsing semantic if the
- * new XML_PARSE_OLD10 option is given to the parser.
- */
- static int
- xmlIsNameStartChar(xmlParserCtxtPtr ctxt, int c) {
- if ((ctxt->options & XML_PARSE_OLD10) == 0) {
- /*
- * Use the new checks of production [4] [4a] amd [5] of the
- * Update 5 of XML-1.0
- */
- if ((c != ' ') && (c != '>') && (c != '/') && /* accelerators */
- (((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- (c == '_') || (c == ':') ||
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))))
- return(1);
- } else {
- if (IS_LETTER(c) || (c == '_') || (c == ':'))
- return(1);
- }
- return(0);
- }
- static int
- xmlIsNameChar(xmlParserCtxtPtr ctxt, int c) {
- if ((ctxt->options & XML_PARSE_OLD10) == 0) {
- /*
- * Use the new checks of production [4] [4a] amd [5] of the
- * Update 5 of XML-1.0
- */
- if ((c != ' ') && (c != '>') && (c != '/') && /* accelerators */
- (((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- ((c >= '0') && (c <= '9')) || /* !start */
- (c == '_') || (c == ':') ||
- (c == '-') || (c == '.') || (c == 0xB7) || /* !start */
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x300) && (c <= 0x36F)) || /* !start */
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x203F) && (c <= 0x2040)) || /* !start */
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))))
- return(1);
- } else {
- if ((IS_LETTER(c)) || (IS_DIGIT(c)) ||
- (c == '.') || (c == '-') ||
- (c == '_') || (c == ':') ||
- (IS_COMBINING(c)) ||
- (IS_EXTENDER(c)))
- return(1);
- }
- return(0);
- }
- static xmlChar * xmlParseAttValueInternal(xmlParserCtxtPtr ctxt,
- int *len, int *alloc, int normalize);
- static const xmlChar *
- xmlParseNameComplex(xmlParserCtxtPtr ctxt) {
- int len = 0, l;
- int c;
- int count = 0;
- #ifdef DEBUG
- nbParseNameComplex++;
- #endif
- /*
- * Handler for more complex cases
- */
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- c = CUR_CHAR(l);
- if ((ctxt->options & XML_PARSE_OLD10) == 0) {
- /*
- * Use the new checks of production [4] [4a] amd [5] of the
- * Update 5 of XML-1.0
- */
- if ((c == ' ') || (c == '>') || (c == '/') || /* accelerators */
- (!(((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- (c == '_') || (c == ':') ||
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))))) {
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- while ((c != ' ') && (c != '>') && (c != '/') && /* accelerators */
- (((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- ((c >= '0') && (c <= '9')) || /* !start */
- (c == '_') || (c == ':') ||
- (c == '-') || (c == '.') || (c == 0xB7) || /* !start */
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x300) && (c <= 0x36F)) || /* !start */
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x203F) && (c <= 0x2040)) || /* !start */
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))
- )) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- }
- } else {
- if ((c == ' ') || (c == '>') || (c == '/') || /* accelerators */
- (!IS_LETTER(c) && (c != '_') &&
- (c != ':'))) {
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- while ((c != ' ') && (c != '>') && (c != '/') && /* test bigname.xml */
- ((IS_LETTER(c)) || (IS_DIGIT(c)) ||
- (c == '.') || (c == '-') ||
- (c == '_') || (c == ':') ||
- (IS_COMBINING(c)) ||
- (IS_EXTENDER(c)))) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- }
- }
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "Name");
- return(NULL);
- }
- if (ctxt->input->cur - ctxt->input->base < len) {
- /*
- * There were a couple of bugs where PERefs lead to to a change
- * of the buffer. Check the buffer size to avoid passing an invalid
- * pointer to xmlDictLookup.
- */
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "unexpected change of input buffer");
- return (NULL);
- }
- if ((*ctxt->input->cur == '\n') && (ctxt->input->cur[-1] == '\r'))
- return(xmlDictLookup(ctxt->dict, ctxt->input->cur - (len + 1), len));
- return(xmlDictLookup(ctxt->dict, ctxt->input->cur - len, len));
- }
- /**
- * xmlParseName:
- * @ctxt: an XML parser context
- *
- * parse an XML name.
- *
- * [4] NameChar ::= Letter | Digit | '.' | '-' | '_' | ':' |
- * CombiningChar | Extender
- *
- * [5] Name ::= (Letter | '_' | ':') (NameChar)*
- *
- * [6] Names ::= Name (#x20 Name)*
- *
- * Returns the Name parsed or NULL
- */
- const xmlChar *
- xmlParseName(xmlParserCtxtPtr ctxt) {
- const xmlChar *in;
- const xmlChar *ret;
- int count = 0;
- GROW;
- #ifdef DEBUG
- nbParseName++;
- #endif
- /*
- * Accelerator for simple ASCII names
- */
- in = ctxt->input->cur;
- if (((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- (*in == '_') || (*in == ':')) {
- in++;
- while (((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- ((*in >= 0x30) && (*in <= 0x39)) ||
- (*in == '_') || (*in == '-') ||
- (*in == ':') || (*in == '.'))
- in++;
- if ((*in > 0) && (*in < 0x80)) {
- count = in - ctxt->input->cur;
- if ((count > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "Name");
- return(NULL);
- }
- ret = xmlDictLookup(ctxt->dict, ctxt->input->cur, count);
- ctxt->input->cur = in;
- ctxt->input->col += count;
- if (ret == NULL)
- xmlErrMemory(ctxt, NULL);
- return(ret);
- }
- }
- /* accelerator for special cases */
- return(xmlParseNameComplex(ctxt));
- }
- static const xmlChar *
- xmlParseNCNameComplex(xmlParserCtxtPtr ctxt) {
- int len = 0, l;
- int c;
- int count = 0;
- size_t startPosition = 0;
- #ifdef DEBUG
- nbParseNCNameComplex++;
- #endif
- /*
- * Handler for more complex cases
- */
- GROW;
- startPosition = CUR_PTR - BASE_PTR;
- c = CUR_CHAR(l);
- if ((c == ' ') || (c == '>') || (c == '/') || /* accelerators */
- (!xmlIsNameStartChar(ctxt, c) || (c == ':'))) {
- return(NULL);
- }
- while ((c != ' ') && (c != '>') && (c != '/') && /* test bigname.xml */
- (xmlIsNameChar(ctxt, c) && (c != ':'))) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- if (c == 0) {
- count = 0;
- /*
- * when shrinking to extend the buffer we really need to preserve
- * the part of the name we already parsed. Hence rolling back
- * by current length.
- */
- ctxt->input->cur -= l;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- ctxt->input->cur += l;
- c = CUR_CHAR(l);
- }
- }
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- return(xmlDictLookup(ctxt->dict, (BASE_PTR + startPosition), len));
- }
- /**
- * xmlParseNCName:
- * @ctxt: an XML parser context
- * @len: length of the string parsed
- *
- * parse an XML name.
- *
- * [4NS] NCNameChar ::= Letter | Digit | '.' | '-' | '_' |
- * CombiningChar | Extender
- *
- * [5NS] NCName ::= (Letter | '_') (NCNameChar)*
- *
- * Returns the Name parsed or NULL
- */
- static const xmlChar *
- xmlParseNCName(xmlParserCtxtPtr ctxt) {
- const xmlChar *in, *e;
- const xmlChar *ret;
- int count = 0;
- #ifdef DEBUG
- nbParseNCName++;
- #endif
- /*
- * Accelerator for simple ASCII names
- */
- in = ctxt->input->cur;
- e = ctxt->input->end;
- if ((((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- (*in == '_')) && (in < e)) {
- in++;
- while ((((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- ((*in >= 0x30) && (*in <= 0x39)) ||
- (*in == '_') || (*in == '-') ||
- (*in == '.')) && (in < e))
- in++;
- if (in >= e)
- goto complex;
- if ((*in > 0) && (*in < 0x80)) {
- count = in - ctxt->input->cur;
- if ((count > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- ret = xmlDictLookup(ctxt->dict, ctxt->input->cur, count);
- ctxt->input->cur = in;
- ctxt->input->col += count;
- if (ret == NULL) {
- xmlErrMemory(ctxt, NULL);
- }
- return(ret);
- }
- }
- complex:
- return(xmlParseNCNameComplex(ctxt));
- }
- /**
- * xmlParseNameAndCompare:
- * @ctxt: an XML parser context
- *
- * parse an XML name and compares for match
- * (specialized for endtag parsing)
- *
- * Returns NULL for an illegal name, (xmlChar*) 1 for success
- * and the name for mismatch
- */
- static const xmlChar *
- xmlParseNameAndCompare(xmlParserCtxtPtr ctxt, xmlChar const *other) {
- register const xmlChar *cmp = other;
- register const xmlChar *in;
- const xmlChar *ret;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- in = ctxt->input->cur;
- while (*in != 0 && *in == *cmp) {
- ++in;
- ++cmp;
- }
- if (*cmp == 0 && (*in == '>' || IS_BLANK_CH (*in))) {
- /* success */
- ctxt->input->col += in - ctxt->input->cur;
- ctxt->input->cur = in;
- return (const xmlChar*) 1;
- }
- /* failure (or end of input buffer), check with full function */
- ret = xmlParseName (ctxt);
- /* strings coming from the dictionary direct compare possible */
- if (ret == other) {
- return (const xmlChar*) 1;
- }
- return ret;
- }
- /**
- * xmlParseStringName:
- * @ctxt: an XML parser context
- * @str: a pointer to the string pointer (IN/OUT)
- *
- * parse an XML name.
- *
- * [4] NameChar ::= Letter | Digit | '.' | '-' | '_' | ':' |
- * CombiningChar | Extender
- *
- * [5] Name ::= (Letter | '_' | ':') (NameChar)*
- *
- * [6] Names ::= Name (#x20 Name)*
- *
- * Returns the Name parsed or NULL. The @str pointer
- * is updated to the current location in the string.
- */
- static xmlChar *
- xmlParseStringName(xmlParserCtxtPtr ctxt, const xmlChar** str) {
- xmlChar buf[XML_MAX_NAMELEN + 5];
- const xmlChar *cur = *str;
- int len = 0, l;
- int c;
- #ifdef DEBUG
- nbParseStringName++;
- #endif
- c = CUR_SCHAR(cur, l);
- if (!xmlIsNameStartChar(ctxt, c)) {
- return(NULL);
- }
- COPY_BUF(l,buf,len,c);
- cur += l;
- c = CUR_SCHAR(cur, l);
- while (xmlIsNameChar(ctxt, c)) {
- COPY_BUF(l,buf,len,c);
- cur += l;
- c = CUR_SCHAR(cur, l);
- if (len >= XML_MAX_NAMELEN) { /* test bigentname.xml */
- /*
- * Okay someone managed to make a huge name, so he's ready to pay
- * for the processing speed.
- */
- xmlChar *buffer;
- int max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while (xmlIsNameChar(ctxt, c)) {
- if (len + 10 > max) {
- xmlChar *tmp;
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- xmlFree(buffer);
- return(NULL);
- }
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buffer);
- return(NULL);
- }
- buffer = tmp;
- }
- COPY_BUF(l,buffer,len,c);
- cur += l;
- c = CUR_SCHAR(cur, l);
- }
- buffer[len] = 0;
- *str = cur;
- return(buffer);
- }
- }
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- *str = cur;
- return(xmlStrndup(buf, len));
- }
- /**
- * xmlParseNmtoken:
- * @ctxt: an XML parser context
- *
- * parse an XML Nmtoken.
- *
- * [7] Nmtoken ::= (NameChar)+
- *
- * [8] Nmtokens ::= Nmtoken (#x20 Nmtoken)*
- *
- * Returns the Nmtoken parsed or NULL
- */
- xmlChar *
- xmlParseNmtoken(xmlParserCtxtPtr ctxt) {
- xmlChar buf[XML_MAX_NAMELEN + 5];
- int len = 0, l;
- int c;
- int count = 0;
- #ifdef DEBUG
- nbParseNmToken++;
- #endif
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- c = CUR_CHAR(l);
- while (xmlIsNameChar(ctxt, c)) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- }
- COPY_BUF(l,buf,len,c);
- NEXTL(l);
- c = CUR_CHAR(l);
- if (c == 0) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- c = CUR_CHAR(l);
- }
- if (len >= XML_MAX_NAMELEN) {
- /*
- * Okay someone managed to make a huge token, so he's ready to pay
- * for the processing speed.
- */
- xmlChar *buffer;
- int max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while (xmlIsNameChar(ctxt, c)) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buffer);
- return(NULL);
- }
- }
- if (len + 10 > max) {
- xmlChar *tmp;
- if ((max > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NmToken");
- xmlFree(buffer);
- return(NULL);
- }
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buffer);
- return(NULL);
- }
- buffer = tmp;
- }
- COPY_BUF(l,buffer,len,c);
- NEXTL(l);
- c = CUR_CHAR(l);
- }
- buffer[len] = 0;
- return(buffer);
- }
- }
- if (len == 0)
- return(NULL);
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NmToken");
- return(NULL);
- }
- return(xmlStrndup(buf, len));
- }
- /**
- * xmlParseEntityValue:
- * @ctxt: an XML parser context
- * @orig: if non-NULL store a copy of the original entity value
- *
- * parse a value for ENTITY declarations
- *
- * [9] EntityValue ::= '"' ([^%&"] | PEReference | Reference)* '"' |
- * "'" ([^%&'] | PEReference | Reference)* "'"
- *
- * Returns the EntityValue parsed with reference substituted or NULL
- */
- xmlChar *
- xmlParseEntityValue(xmlParserCtxtPtr ctxt, xmlChar **orig) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- int c, l;
- xmlChar stop;
- xmlChar *ret = NULL;
- const xmlChar *cur = NULL;
- xmlParserInputPtr input;
- if (RAW == '"') stop = '"';
- else if (RAW == '\'') stop = '\'';
- else {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_NOT_STARTED, NULL);
- return(NULL);
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- /*
- * The content of the entity definition is copied in a buffer.
- */
- ctxt->instate = XML_PARSER_ENTITY_VALUE;
- input = ctxt->input;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- goto error;
- NEXT;
- c = CUR_CHAR(l);
- /*
- * NOTE: 4.4.5 Included in Literal
- * When a parameter entity reference appears in a literal entity
- * value, ... a single or double quote character in the replacement
- * text is always treated as a normal data character and will not
- * terminate the literal.
- * In practice it means we stop the loop only when back at parsing
- * the initial entity and the quote is found
- */
- while (((IS_CHAR(c)) && ((c != stop) || /* checked */
- (ctxt->input != input))) && (ctxt->instate != XML_PARSER_EOF)) {
- if (len + 5 >= size) {
- xmlChar *tmp;
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- goto error;
- }
- buf = tmp;
- }
- COPY_BUF(l,buf,len,c);
- NEXTL(l);
- GROW;
- c = CUR_CHAR(l);
- if (c == 0) {
- GROW;
- c = CUR_CHAR(l);
- }
- }
- buf[len] = 0;
- if (ctxt->instate == XML_PARSER_EOF)
- goto error;
- if (c != stop) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_NOT_FINISHED, NULL);
- goto error;
- }
- NEXT;
- /*
- * Raise problem w.r.t. '&' and '%' being used in non-entities
- * reference constructs. Note Charref will be handled in
- * xmlStringDecodeEntities()
- */
- cur = buf;
- while (*cur != 0) { /* non input consuming */
- if ((*cur == '%') || ((*cur == '&') && (cur[1] != '#'))) {
- xmlChar *name;
- xmlChar tmp = *cur;
- int nameOk = 0;
- cur++;
- name = xmlParseStringName(ctxt, &cur);
- if (name != NULL) {
- nameOk = 1;
- xmlFree(name);
- }
- if ((nameOk == 0) || (*cur != ';')) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_ENTITY_CHAR_ERROR,
- "EntityValue: '%c' forbidden except for entities references\n",
- tmp);
- goto error;
- }
- if ((tmp == '%') && (ctxt->inSubset == 1) &&
- (ctxt->inputNr == 1)) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_PE_INTERNAL, NULL);
- goto error;
- }
- if (*cur == 0)
- break;
- }
- cur++;
- }
- /*
- * Then PEReference entities are substituted.
- *
- * NOTE: 4.4.7 Bypassed
- * When a general entity reference appears in the EntityValue in
- * an entity declaration, it is bypassed and left as is.
- * so XML_SUBSTITUTE_REF is not set here.
- */
- ++ctxt->depth;
- ret = xmlStringDecodeEntities(ctxt, buf, XML_SUBSTITUTE_PEREF,
- 0, 0, 0);
- --ctxt->depth;
- if (orig != NULL) {
- *orig = buf;
- buf = NULL;
- }
- error:
- if (buf != NULL)
- xmlFree(buf);
- return(ret);
- }
- /**
- * xmlParseAttValueComplex:
- * @ctxt: an XML parser context
- * @len: the resulting attribute len
- * @normalize: whether to apply the inner normalization
- *
- * parse a value for an attribute, this is the fallback function
- * of xmlParseAttValue() when the attribute parsing requires handling
- * of non-ASCII characters, or normalization compaction.
- *
- * Returns the AttValue parsed or NULL. The value has to be freed by the caller.
- */
- static xmlChar *
- xmlParseAttValueComplex(xmlParserCtxtPtr ctxt, int *attlen, int normalize) {
- xmlChar limit = 0;
- xmlChar *buf = NULL;
- xmlChar *rep = NULL;
- size_t len = 0;
- size_t buf_size = 0;
- int c, l, in_space = 0;
- xmlChar *current = NULL;
- xmlEntityPtr ent;
- if (NXT(0) == '"') {
- ctxt->instate = XML_PARSER_ATTRIBUTE_VALUE;
- limit = '"';
- NEXT;
- } else if (NXT(0) == '\'') {
- limit = '\'';
- ctxt->instate = XML_PARSER_ATTRIBUTE_VALUE;
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_ATTRIBUTE_NOT_STARTED, NULL);
- return(NULL);
- }
- /*
- * allocate a translation buffer.
- */
- buf_size = XML_PARSER_BUFFER_SIZE;
- buf = (xmlChar *) xmlMallocAtomic(buf_size);
- if (buf == NULL) goto mem_error;
- /*
- * OK loop until we reach one of the ending char or a size limit.
- */
- c = CUR_CHAR(l);
- while (((NXT(0) != limit) && /* checked */
- (IS_CHAR(c)) && (c != '<')) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- /*
- * Impose a reasonable limit on attribute size, unless XML_PARSE_HUGE
- * special option is given
- */
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- goto mem_error;
- }
- if (c == '&') {
- in_space = 0;
- if (NXT(1) == '#') {
- int val = xmlParseCharRef(ctxt);
- if (val == '&') {
- if (ctxt->replaceEntities) {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- buf[len++] = '&';
- } else {
- /*
- * The reparsing will be done in xmlStringGetNodeList()
- * called by the attribute() function in SAX.c
- */
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- buf[len++] = '&';
- buf[len++] = '#';
- buf[len++] = '3';
- buf[len++] = '8';
- buf[len++] = ';';
- }
- } else if (val != 0) {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- len += xmlCopyChar(0, &buf[len], val);
- }
- } else {
- ent = xmlParseEntityRef(ctxt);
- ctxt->nbentities++;
- if (ent != NULL)
- ctxt->nbentities += ent->owner;
- if ((ent != NULL) &&
- (ent->etype == XML_INTERNAL_PREDEFINED_ENTITY)) {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- if ((ctxt->replaceEntities == 0) &&
- (ent->content[0] == '&')) {
- buf[len++] = '&';
- buf[len++] = '#';
- buf[len++] = '3';
- buf[len++] = '8';
- buf[len++] = ';';
- } else {
- buf[len++] = ent->content[0];
- }
- } else if ((ent != NULL) &&
- (ctxt->replaceEntities != 0)) {
- if (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) {
- ++ctxt->depth;
- rep = xmlStringDecodeEntities(ctxt, ent->content,
- XML_SUBSTITUTE_REF,
- 0, 0, 0);
- --ctxt->depth;
- if (rep != NULL) {
- current = rep;
- while (*current != 0) { /* non input consuming */
- if ((*current == 0xD) || (*current == 0xA) ||
- (*current == 0x9)) {
- buf[len++] = 0x20;
- current++;
- } else
- buf[len++] = *current++;
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- }
- xmlFree(rep);
- rep = NULL;
- }
- } else {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- if (ent->content != NULL)
- buf[len++] = ent->content[0];
- }
- } else if (ent != NULL) {
- int i = xmlStrlen(ent->name);
- const xmlChar *cur = ent->name;
- /*
- * This may look absurd but is needed to detect
- * entities problems
- */
- if ((ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
- (ent->content != NULL) && (ent->checked == 0)) {
- unsigned long oldnbent = ctxt->nbentities, diff;
- ++ctxt->depth;
- rep = xmlStringDecodeEntities(ctxt, ent->content,
- XML_SUBSTITUTE_REF, 0, 0, 0);
- --ctxt->depth;
- diff = ctxt->nbentities - oldnbent + 1;
- if (diff > INT_MAX / 2)
- diff = INT_MAX / 2;
- ent->checked = diff * 2;
- if (rep != NULL) {
- if (xmlStrchr(rep, '<'))
- ent->checked |= 1;
- xmlFree(rep);
- rep = NULL;
- } else {
- ent->content[0] = 0;
- }
- }
- /*
- * Just output the reference
- */
- buf[len++] = '&';
- while (len + i + 10 > buf_size) {
- growBuffer(buf, i + 10);
- }
- for (;i > 0;i--)
- buf[len++] = *cur++;
- buf[len++] = ';';
- }
- }
- } else {
- if ((c == 0x20) || (c == 0xD) || (c == 0xA) || (c == 0x9)) {
- if ((len != 0) || (!normalize)) {
- if ((!normalize) || (!in_space)) {
- COPY_BUF(l,buf,len,0x20);
- while (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- }
- in_space = 1;
- }
- } else {
- in_space = 0;
- COPY_BUF(l,buf,len,c);
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- }
- NEXTL(l);
- }
- GROW;
- c = CUR_CHAR(l);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto error;
- if ((in_space) && (normalize)) {
- while ((len > 0) && (buf[len - 1] == 0x20)) len--;
- }
- buf[len] = 0;
- if (RAW == '<') {
- xmlFatalErr(ctxt, XML_ERR_LT_IN_ATTRIBUTE, NULL);
- } else if (RAW != limit) {
- if ((c != 0) && (!IS_CHAR(c))) {
- xmlFatalErrMsg(ctxt, XML_ERR_INVALID_CHAR,
- "invalid character in attribute value\n");
- } else {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue: ' expected\n");
- }
- } else
- NEXT;
- /*
- * There we potentially risk an overflow, don't allow attribute value of
- * length more than INT_MAX it is a very reasonable assumption !
- */
- if (len >= INT_MAX) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- goto mem_error;
- }
- if (attlen != NULL) *attlen = (int) len;
- return(buf);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- error:
- if (buf != NULL)
- xmlFree(buf);
- if (rep != NULL)
- xmlFree(rep);
- return(NULL);
- }
- /**
- * xmlParseAttValue:
- * @ctxt: an XML parser context
- *
- * parse a value for an attribute
- * Note: the parser won't do substitution of entities here, this
- * will be handled later in xmlStringGetNodeList
- *
- * [10] AttValue ::= '"' ([^<&"] | Reference)* '"' |
- * "'" ([^<&'] | Reference)* "'"
- *
- * 3.3.3 Attribute-Value Normalization:
- * Before the value of an attribute is passed to the application or
- * checked for validity, the XML processor must normalize it as follows:
- * - a character reference is processed by appending the referenced
- * character to the attribute value
- * - an entity reference is processed by recursively processing the
- * replacement text of the entity
- * - a whitespace character (#x20, #xD, #xA, #x9) is processed by
- * appending #x20 to the normalized value, except that only a single
- * #x20 is appended for a "#xD#xA" sequence that is part of an external
- * parsed entity or the literal entity value of an internal parsed entity
- * - other characters are processed by appending them to the normalized value
- * If the declared value is not CDATA, then the XML processor must further
- * process the normalized attribute value by discarding any leading and
- * trailing space (#x20) characters, and by replacing sequences of space
- * (#x20) characters by a single space (#x20) character.
- * All attributes for which no declaration has been read should be treated
- * by a non-validating parser as if declared CDATA.
- *
- * Returns the AttValue parsed or NULL. The value has to be freed by the caller.
- */
- xmlChar *
- xmlParseAttValue(xmlParserCtxtPtr ctxt) {
- if ((ctxt == NULL) || (ctxt->input == NULL)) return(NULL);
- return(xmlParseAttValueInternal(ctxt, NULL, NULL, 0));
- }
- /**
- * xmlParseSystemLiteral:
- * @ctxt: an XML parser context
- *
- * parse an XML Literal
- *
- * [11] SystemLiteral ::= ('"' [^"]* '"') | ("'" [^']* "'")
- *
- * Returns the SystemLiteral parsed or NULL
- */
- xmlChar *
- xmlParseSystemLiteral(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- int cur, l;
- xmlChar stop;
- int state = ctxt->instate;
- int count = 0;
- SHRINK;
- if (RAW == '"') {
- NEXT;
- stop = '"';
- } else if (RAW == '\'') {
- NEXT;
- stop = '\'';
- } else {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_STARTED, NULL);
- return(NULL);
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- ctxt->instate = XML_PARSER_SYSTEM_LITERAL;
- cur = CUR_CHAR(l);
- while ((IS_CHAR(cur)) && (cur != stop)) { /* checked */
- if (len + 5 >= size) {
- xmlChar *tmp;
- if ((size > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "SystemLiteral");
- xmlFree(buf);
- ctxt->instate = (xmlParserInputState) state;
- return(NULL);
- }
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buf);
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = (xmlParserInputState) state;
- return(NULL);
- }
- buf = tmp;
- }
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return(NULL);
- }
- }
- COPY_BUF(l,buf,len,cur);
- NEXTL(l);
- cur = CUR_CHAR(l);
- if (cur == 0) {
- GROW;
- SHRINK;
- cur = CUR_CHAR(l);
- }
- }
- buf[len] = 0;
- ctxt->instate = (xmlParserInputState) state;
- if (!IS_CHAR(cur)) {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_FINISHED, NULL);
- } else {
- NEXT;
- }
- return(buf);
- }
- /**
- * xmlParsePubidLiteral:
- * @ctxt: an XML parser context
- *
- * parse an XML public literal
- *
- * [12] PubidLiteral ::= '"' PubidChar* '"' | "'" (PubidChar - "'")* "'"
- *
- * Returns the PubidLiteral parsed or NULL.
- */
- xmlChar *
- xmlParsePubidLiteral(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- xmlChar cur;
- xmlChar stop;
- int count = 0;
- xmlParserInputState oldstate = ctxt->instate;
- SHRINK;
- if (RAW == '"') {
- NEXT;
- stop = '"';
- } else if (RAW == '\'') {
- NEXT;
- stop = '\'';
- } else {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_STARTED, NULL);
- return(NULL);
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- ctxt->instate = XML_PARSER_PUBLIC_LITERAL;
- cur = CUR;
- while ((IS_PUBIDCHAR_CH(cur)) && (cur != stop)) { /* checked */
- if (len + 1 >= size) {
- xmlChar *tmp;
- if ((size > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "Public ID");
- xmlFree(buf);
- return(NULL);
- }
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buf);
- return(NULL);
- }
- buf = tmp;
- }
- buf[len++] = cur;
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return(NULL);
- }
- }
- NEXT;
- cur = CUR;
- if (cur == 0) {
- GROW;
- SHRINK;
- cur = CUR;
- }
- }
- buf[len] = 0;
- if (cur != stop) {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_FINISHED, NULL);
- } else {
- NEXT;
- }
- ctxt->instate = oldstate;
- return(buf);
- }
- static void xmlParseCharDataComplex(xmlParserCtxtPtr ctxt, int cdata);
- /*
- * used for the test in the inner loop of the char data testing
- */
- static const unsigned char test_char_data[256] = {
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x9, CR/LF separated */
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x00, 0x27, /* & */
- 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E, 0x2F,
- 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
- 0x38, 0x39, 0x3A, 0x3B, 0x00, 0x3D, 0x3E, 0x3F, /* < */
- 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47,
- 0x48, 0x49, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F,
- 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57,
- 0x58, 0x59, 0x5A, 0x5B, 0x5C, 0x00, 0x5E, 0x5F, /* ] */
- 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
- 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F,
- 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77,
- 0x78, 0x79, 0x7A, 0x7B, 0x7C, 0x7D, 0x7E, 0x7F,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* non-ascii */
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
- };
- /**
- * xmlParseCharData:
- * @ctxt: an XML parser context
- * @cdata: int indicating whether we are within a CDATA section
- *
- * parse a CharData section.
- * if we are within a CDATA section ']]>' marks an end of section.
- *
- * The right angle bracket (>) may be represented using the string ">",
- * and must, for compatibility, be escaped using ">" or a character
- * reference when it appears in the string "]]>" in content, when that
- * string is not marking the end of a CDATA section.
- *
- * [14] CharData ::= [^<&]* - ([^<&]* ']]>' [^<&]*)
- */
- void
- xmlParseCharData(xmlParserCtxtPtr ctxt, int cdata) {
- const xmlChar *in;
- int nbchar = 0;
- int line = ctxt->input->line;
- int col = ctxt->input->col;
- int ccol;
- SHRINK;
- GROW;
- /*
- * Accelerated common case where input don't need to be
- * modified before passing it to the handler.
- */
- if (!cdata) {
- in = ctxt->input->cur;
- do {
- get_more_space:
- while (*in == 0x20) { in++; ctxt->input->col++; }
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- goto get_more_space;
- }
- if (*in == '<') {
- nbchar = in - ctxt->input->cur;
- if (nbchar > 0) {
- const xmlChar *tmp = ctxt->input->cur;
- ctxt->input->cur = in;
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->ignorableWhitespace !=
- ctxt->sax->characters)) {
- if (areBlanks(ctxt, tmp, nbchar, 1)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData,
- tmp, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- tmp, nbchar);
- if (*ctxt->space == -1)
- *ctxt->space = -2;
- }
- } else if ((ctxt->sax != NULL) &&
- (ctxt->sax->characters != NULL)) {
- ctxt->sax->characters(ctxt->userData,
- tmp, nbchar);
- }
- }
- return;
- }
- get_more:
- ccol = ctxt->input->col;
- while (test_char_data[*in]) {
- in++;
- ccol++;
- }
- ctxt->input->col = ccol;
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- goto get_more;
- }
- if (*in == ']') {
- if ((in[1] == ']') && (in[2] == '>')) {
- xmlFatalErr(ctxt, XML_ERR_MISPLACED_CDATA_END, NULL);
- ctxt->input->cur = in + 1;
- return;
- }
- in++;
- ctxt->input->col++;
- goto get_more;
- }
- nbchar = in - ctxt->input->cur;
- if (nbchar > 0) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->ignorableWhitespace !=
- ctxt->sax->characters) &&
- (IS_BLANK_CH(*ctxt->input->cur))) {
- const xmlChar *tmp = ctxt->input->cur;
- ctxt->input->cur = in;
- if (areBlanks(ctxt, tmp, nbchar, 0)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData,
- tmp, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- tmp, nbchar);
- if (*ctxt->space == -1)
- *ctxt->space = -2;
- }
- line = ctxt->input->line;
- col = ctxt->input->col;
- } else if (ctxt->sax != NULL) {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- ctxt->input->cur, nbchar);
- line = ctxt->input->line;
- col = ctxt->input->col;
- }
- /* something really bad happened in the SAX callback */
- if (ctxt->instate != XML_PARSER_CONTENT)
- return;
- }
- ctxt->input->cur = in;
- if (*in == 0xD) {
- in++;
- if (*in == 0xA) {
- ctxt->input->cur = in;
- in++;
- ctxt->input->line++; ctxt->input->col = 1;
- continue; /* while */
- }
- in--;
- }
- if (*in == '<') {
- return;
- }
- if (*in == '&') {
- return;
- }
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- in = ctxt->input->cur;
- } while (((*in >= 0x20) && (*in <= 0x7F)) || (*in == 0x09) || (*in == 0x0a));
- nbchar = 0;
- }
- ctxt->input->line = line;
- ctxt->input->col = col;
- xmlParseCharDataComplex(ctxt, cdata);
- }
- /**
- * xmlParseCharDataComplex:
- * @ctxt: an XML parser context
- * @cdata: int indicating whether we are within a CDATA section
- *
- * parse a CharData section.this is the fallback function
- * of xmlParseCharData() when the parsing requires handling
- * of non-ASCII characters.
- */
- static void
- xmlParseCharDataComplex(xmlParserCtxtPtr ctxt, int cdata) {
- xmlChar buf[XML_PARSER_BIG_BUFFER_SIZE + 5];
- int nbchar = 0;
- int cur, l;
- int count = 0;
- SHRINK;
- GROW;
- cur = CUR_CHAR(l);
- while ((cur != '<') && /* checked */
- (cur != '&') &&
- (IS_CHAR(cur))) /* test also done in xmlCurrentChar() */ {
- if ((cur == ']') && (NXT(1) == ']') &&
- (NXT(2) == '>')) {
- if (cdata) break;
- else {
- xmlFatalErr(ctxt, XML_ERR_MISPLACED_CDATA_END, NULL);
- }
- }
- COPY_BUF(l,buf,nbchar,cur);
- /* move current position before possible calling of ctxt->sax->characters */
- NEXTL(l);
- cur = CUR_CHAR(l);
- if (nbchar >= XML_PARSER_BIG_BUFFER_SIZE) {
- buf[nbchar] = 0;
- /*
- * OK the segment is to be consumed as chars.
- */
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (areBlanks(ctxt, buf, nbchar, 0)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData,
- buf, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData, buf, nbchar);
- if ((ctxt->sax->characters !=
- ctxt->sax->ignorableWhitespace) &&
- (*ctxt->space == -1))
- *ctxt->space = -2;
- }
- }
- nbchar = 0;
- /* something really bad happened in the SAX callback */
- if (ctxt->instate != XML_PARSER_CONTENT)
- return;
- }
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- }
- }
- if (nbchar != 0) {
- buf[nbchar] = 0;
- /*
- * OK the segment is to be consumed as chars.
- */
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (areBlanks(ctxt, buf, nbchar, 0)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData, buf, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData, buf, nbchar);
- if ((ctxt->sax->characters != ctxt->sax->ignorableWhitespace) &&
- (*ctxt->space == -1))
- *ctxt->space = -2;
- }
- }
- }
- if ((cur != 0) && (!IS_CHAR(cur))) {
- /* Generate the error and skip the offending character */
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "PCDATA invalid Char value %d\n",
- cur);
- NEXTL(l);
- }
- }
- /**
- * xmlParseExternalID:
- * @ctxt: an XML parser context
- * @publicID: a xmlChar** receiving PubidLiteral
- * @strict: indicate whether we should restrict parsing to only
- * production [75], see NOTE below
- *
- * Parse an External ID or a Public ID
- *
- * NOTE: Productions [75] and [83] interact badly since [75] can generate
- * 'PUBLIC' S PubidLiteral S SystemLiteral
- *
- * [75] ExternalID ::= 'SYSTEM' S SystemLiteral
- * | 'PUBLIC' S PubidLiteral S SystemLiteral
- *
- * [83] PublicID ::= 'PUBLIC' S PubidLiteral
- *
- * Returns the function returns SystemLiteral and in the second
- * case publicID receives PubidLiteral, is strict is off
- * it is possible to return NULL and have publicID set.
- */
- xmlChar *
- xmlParseExternalID(xmlParserCtxtPtr ctxt, xmlChar **publicID, int strict) {
- xmlChar *URI = NULL;
- SHRINK;
- *publicID = NULL;
- if (CMP6(CUR_PTR, 'S', 'Y', 'S', 'T', 'E', 'M')) {
- SKIP(6);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'SYSTEM'\n");
- }
- URI = xmlParseSystemLiteral(ctxt);
- if (URI == NULL) {
- xmlFatalErr(ctxt, XML_ERR_URI_REQUIRED, NULL);
- }
- } else if (CMP6(CUR_PTR, 'P', 'U', 'B', 'L', 'I', 'C')) {
- SKIP(6);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'PUBLIC'\n");
- }
- *publicID = xmlParsePubidLiteral(ctxt);
- if (*publicID == NULL) {
- xmlFatalErr(ctxt, XML_ERR_PUBID_REQUIRED, NULL);
- }
- if (strict) {
- /*
- * We don't handle [83] so "S SystemLiteral" is required.
- */
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the Public Identifier\n");
- }
- } else {
- /*
- * We handle [83] so we return immediately, if
- * "S SystemLiteral" is not detected. We skip blanks if no
- * system literal was found, but this is harmless since we must
- * be at the end of a NotationDecl.
- */
- if (SKIP_BLANKS == 0) return(NULL);
- if ((CUR != '\'') && (CUR != '"')) return(NULL);
- }
- URI = xmlParseSystemLiteral(ctxt);
- if (URI == NULL) {
- xmlFatalErr(ctxt, XML_ERR_URI_REQUIRED, NULL);
- }
- }
- return(URI);
- }
- /**
- * xmlParseCommentComplex:
- * @ctxt: an XML parser context
- * @buf: the already parsed part of the buffer
- * @len: number of bytes in the buffer
- * @size: allocated size of the buffer
- *
- * Skip an XML (SGML) comment <!-- .... -->
- * The spec says that "For compatibility, the string "--" (double-hyphen)
- * must not occur within comments. "
- * This is the slow routine in case the accelerator for ascii didn't work
- *
- * [15] Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
- */
- static void
- xmlParseCommentComplex(xmlParserCtxtPtr ctxt, xmlChar *buf,
- size_t len, size_t size) {
- int q, ql;
- int r, rl;
- int cur, l;
- size_t count = 0;
- int inputid;
- inputid = ctxt->input->id;
- if (buf == NULL) {
- len = 0;
- size = XML_PARSER_BUFFER_SIZE;
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return;
- }
- }
- GROW; /* Assure there's enough input data */
- q = CUR_CHAR(ql);
- if (q == 0)
- goto not_terminated;
- if (!IS_CHAR(q)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseComment: invalid xmlChar value %d\n",
- q);
- xmlFree (buf);
- return;
- }
- NEXTL(ql);
- r = CUR_CHAR(rl);
- if (r == 0)
- goto not_terminated;
- if (!IS_CHAR(r)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseComment: invalid xmlChar value %d\n",
- q);
- xmlFree (buf);
- return;
- }
- NEXTL(rl);
- cur = CUR_CHAR(l);
- if (cur == 0)
- goto not_terminated;
- while (IS_CHAR(cur) && /* checked */
- ((cur != '>') ||
- (r != '-') || (q != '-'))) {
- if ((r == '-') && (q == '-')) {
- xmlFatalErr(ctxt, XML_ERR_HYPHEN_IN_COMMENT, NULL);
- }
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment too big found", NULL);
- xmlFree (buf);
- return;
- }
- if (len + 5 >= size) {
- xmlChar *new_buf;
- size_t new_size;
- new_size = size * 2;
- new_buf = (xmlChar *) xmlRealloc(buf, new_size);
- if (new_buf == NULL) {
- xmlFree (buf);
- xmlErrMemory(ctxt, NULL);
- return;
- }
- buf = new_buf;
- size = new_size;
- }
- COPY_BUF(ql,buf,len,q);
- q = r;
- ql = rl;
- r = cur;
- rl = l;
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- }
- NEXTL(l);
- cur = CUR_CHAR(l);
- if (cur == 0) {
- SHRINK;
- GROW;
- cur = CUR_CHAR(l);
- }
- }
- buf[len] = 0;
- if (cur == 0) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment not terminated \n<!--%.50s\n", buf);
- } else if (!IS_CHAR(cur)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseComment: invalid xmlChar value %d\n",
- cur);
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Comment doesn't start and stop in the same"
- " entity\n");
- }
- NEXT;
- if ((ctxt->sax != NULL) && (ctxt->sax->comment != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->comment(ctxt->userData, buf);
- }
- xmlFree(buf);
- return;
- not_terminated:
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment not terminated\n", NULL);
- xmlFree(buf);
- return;
- }
- /**
- * xmlParseComment:
- * @ctxt: an XML parser context
- *
- * Skip an XML (SGML) comment <!-- .... -->
- * The spec says that "For compatibility, the string "--" (double-hyphen)
- * must not occur within comments. "
- *
- * [15] Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
- */
- void
- xmlParseComment(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- size_t size = XML_PARSER_BUFFER_SIZE;
- size_t len = 0;
- xmlParserInputState state;
- const xmlChar *in;
- size_t nbchar = 0;
- int ccol;
- int inputid;
- /*
- * Check that there is a comment right here.
- */
- if ((RAW != '<') || (NXT(1) != '!') ||
- (NXT(2) != '-') || (NXT(3) != '-')) return;
- state = ctxt->instate;
- ctxt->instate = XML_PARSER_COMMENT;
- inputid = ctxt->input->id;
- SKIP(4);
- SHRINK;
- GROW;
- /*
- * Accelerated common case where input don't need to be
- * modified before passing it to the handler.
- */
- in = ctxt->input->cur;
- do {
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- }
- get_more:
- ccol = ctxt->input->col;
- while (((*in > '-') && (*in <= 0x7F)) ||
- ((*in >= 0x20) && (*in < '-')) ||
- (*in == 0x09)) {
- in++;
- ccol++;
- }
- ctxt->input->col = ccol;
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- goto get_more;
- }
- nbchar = in - ctxt->input->cur;
- /*
- * save current set of data
- */
- if (nbchar > 0) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->comment != NULL)) {
- if (buf == NULL) {
- if ((*in == '-') && (in[1] == '-'))
- size = nbchar + 1;
- else
- size = XML_PARSER_BUFFER_SIZE + nbchar;
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = state;
- return;
- }
- len = 0;
- } else if (len + nbchar + 1 >= size) {
- xmlChar *new_buf;
- size += len + nbchar + XML_PARSER_BUFFER_SIZE;
- new_buf = (xmlChar *) xmlRealloc(buf,
- size * sizeof(xmlChar));
- if (new_buf == NULL) {
- xmlFree (buf);
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = state;
- return;
- }
- buf = new_buf;
- }
- memcpy(&buf[len], ctxt->input->cur, nbchar);
- len += nbchar;
- buf[len] = 0;
- }
- }
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment too big found", NULL);
- xmlFree (buf);
- return;
- }
- ctxt->input->cur = in;
- if (*in == 0xA) {
- in++;
- ctxt->input->line++; ctxt->input->col = 1;
- }
- if (*in == 0xD) {
- in++;
- if (*in == 0xA) {
- ctxt->input->cur = in;
- in++;
- ctxt->input->line++; ctxt->input->col = 1;
- goto get_more;
- }
- in--;
- }
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- in = ctxt->input->cur;
- if (*in == '-') {
- if (in[1] == '-') {
- if (in[2] == '>') {
- if (ctxt->input->id != inputid) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "comment doesn't start and stop in the"
- " same entity\n");
- }
- SKIP(3);
- if ((ctxt->sax != NULL) && (ctxt->sax->comment != NULL) &&
- (!ctxt->disableSAX)) {
- if (buf != NULL)
- ctxt->sax->comment(ctxt->userData, buf);
- else
- ctxt->sax->comment(ctxt->userData, BAD_CAST "");
- }
- if (buf != NULL)
- xmlFree(buf);
- if (ctxt->instate != XML_PARSER_EOF)
- ctxt->instate = state;
- return;
- }
- if (buf != NULL) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_HYPHEN_IN_COMMENT,
- "Double hyphen within comment: "
- "<!--%.50s\n",
- buf);
- } else
- xmlFatalErrMsgStr(ctxt, XML_ERR_HYPHEN_IN_COMMENT,
- "Double hyphen within comment\n", NULL);
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- in++;
- ctxt->input->col++;
- }
- in++;
- ctxt->input->col++;
- goto get_more;
- }
- } while (((*in >= 0x20) && (*in <= 0x7F)) || (*in == 0x09) || (*in == 0x0a));
- xmlParseCommentComplex(ctxt, buf, len, size);
- ctxt->instate = state;
- return;
- }
- /**
- * xmlParsePITarget:
- * @ctxt: an XML parser context
- *
- * parse the name of a PI
- *
- * [17] PITarget ::= Name - (('X' | 'x') ('M' | 'm') ('L' | 'l'))
- *
- * Returns the PITarget name or NULL
- */
- const xmlChar *
- xmlParsePITarget(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- name = xmlParseName(ctxt);
- if ((name != NULL) &&
- ((name[0] == 'x') || (name[0] == 'X')) &&
- ((name[1] == 'm') || (name[1] == 'M')) &&
- ((name[2] == 'l') || (name[2] == 'L'))) {
- int i;
- if ((name[0] == 'x') && (name[1] == 'm') &&
- (name[2] == 'l') && (name[3] == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_RESERVED_XML_NAME,
- "XML declaration allowed only at the start of the document\n");
- return(name);
- } else if (name[3] == 0) {
- xmlFatalErr(ctxt, XML_ERR_RESERVED_XML_NAME, NULL);
- return(name);
- }
- for (i = 0;;i++) {
- if (xmlW3CPIs[i] == NULL) break;
- if (xmlStrEqual(name, (const xmlChar *)xmlW3CPIs[i]))
- return(name);
- }
- xmlWarningMsg(ctxt, XML_ERR_RESERVED_XML_NAME,
- "xmlParsePITarget: invalid name prefix 'xml'\n",
- NULL, NULL);
- }
- if ((name != NULL) && (xmlStrchr(name, ':') != NULL)) {
- xmlNsErr(ctxt, XML_NS_ERR_COLON,
- "colons are forbidden from PI names '%s'\n", name, NULL, NULL);
- }
- return(name);
- }
- #ifdef LIBXML_CATALOG_ENABLED
- /**
- * xmlParseCatalogPI:
- * @ctxt: an XML parser context
- * @catalog: the PI value string
- *
- * parse an XML Catalog Processing Instruction.
- *
- * <?oasis-xml-catalog catalog="http://example.com/catalog.xml"?>
- *
- * Occurs only if allowed by the user and if happening in the Misc
- * part of the document before any doctype information
- * This will add the given catalog to the parsing context in order
- * to be used if there is a resolution need further down in the document
- */
- static void
- xmlParseCatalogPI(xmlParserCtxtPtr ctxt, const xmlChar *catalog) {
- xmlChar *URL = NULL;
- const xmlChar *tmp, *base;
- xmlChar marker;
- tmp = catalog;
- while (IS_BLANK_CH(*tmp)) tmp++;
- if (xmlStrncmp(tmp, BAD_CAST"catalog", 7))
- goto error;
- tmp += 7;
- while (IS_BLANK_CH(*tmp)) tmp++;
- if (*tmp != '=') {
- return;
- }
- tmp++;
- while (IS_BLANK_CH(*tmp)) tmp++;
- marker = *tmp;
- if ((marker != '\'') && (marker != '"'))
- goto error;
- tmp++;
- base = tmp;
- while ((*tmp != 0) && (*tmp != marker)) tmp++;
- if (*tmp == 0)
- goto error;
- URL = xmlStrndup(base, tmp - base);
- tmp++;
- while (IS_BLANK_CH(*tmp)) tmp++;
- if (*tmp != 0)
- goto error;
- if (URL != NULL) {
- ctxt->catalogs = xmlCatalogAddLocal(ctxt->catalogs, URL);
- xmlFree(URL);
- }
- return;
- error:
- xmlWarningMsg(ctxt, XML_WAR_CATALOG_PI,
- "Catalog PI syntax error: %s\n",
- catalog, NULL);
- if (URL != NULL)
- xmlFree(URL);
- }
- #endif
- /**
- * xmlParsePI:
- * @ctxt: an XML parser context
- *
- * parse an XML Processing Instruction.
- *
- * [16] PI ::= '<?' PITarget (S (Char* - (Char* '?>' Char*)))? '?>'
- *
- * The processing is transferred to SAX once parsed.
- */
- void
- xmlParsePI(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- size_t len = 0;
- size_t size = XML_PARSER_BUFFER_SIZE;
- int cur, l;
- const xmlChar *target;
- xmlParserInputState state;
- int count = 0;
- if ((RAW == '<') && (NXT(1) == '?')) {
- int inputid = ctxt->input->id;
- state = ctxt->instate;
- ctxt->instate = XML_PARSER_PI;
- /*
- * this is a Processing Instruction.
- */
- SKIP(2);
- SHRINK;
- /*
- * Parse the target name and check for special support like
- * namespace.
- */
- target = xmlParsePITarget(ctxt);
- if (target != NULL) {
- if ((RAW == '?') && (NXT(1) == '>')) {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "PI declaration doesn't start and stop in"
- " the same entity\n");
- }
- SKIP(2);
- /*
- * SAX: PI detected.
- */
- if ((ctxt->sax) && (!ctxt->disableSAX) &&
- (ctxt->sax->processingInstruction != NULL))
- ctxt->sax->processingInstruction(ctxt->userData,
- target, NULL);
- if (ctxt->instate != XML_PARSER_EOF)
- ctxt->instate = state;
- return;
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = state;
- return;
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_SPACE_REQUIRED,
- "ParsePI: PI %s space expected\n", target);
- }
- cur = CUR_CHAR(l);
- while (IS_CHAR(cur) && /* checked */
- ((cur != '?') || (NXT(1) != '>'))) {
- if (len + 5 >= size) {
- xmlChar *tmp;
- size_t new_size = size * 2;
- tmp = (xmlChar *) xmlRealloc(buf, new_size);
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buf);
- ctxt->instate = state;
- return;
- }
- buf = tmp;
- size = new_size;
- }
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- count = 0;
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_PI_NOT_FINISHED,
- "PI %s too big found", target);
- xmlFree(buf);
- ctxt->instate = state;
- return;
- }
- }
- COPY_BUF(l,buf,len,cur);
- NEXTL(l);
- cur = CUR_CHAR(l);
- if (cur == 0) {
- SHRINK;
- GROW;
- cur = CUR_CHAR(l);
- }
- }
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_PI_NOT_FINISHED,
- "PI %s too big found", target);
- xmlFree(buf);
- ctxt->instate = state;
- return;
- }
- buf[len] = 0;
- if (cur != '?') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_PI_NOT_FINISHED,
- "ParsePI: PI %s never end ...\n", target);
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "PI declaration doesn't start and stop in"
- " the same entity\n");
- }
- SKIP(2);
- #ifdef LIBXML_CATALOG_ENABLED
- if (((state == XML_PARSER_MISC) ||
- (state == XML_PARSER_START)) &&
- (xmlStrEqual(target, XML_CATALOG_PI))) {
- xmlCatalogAllow allow = xmlCatalogGetDefaults();
- if ((allow == XML_CATA_ALLOW_DOCUMENT) ||
- (allow == XML_CATA_ALLOW_ALL))
- xmlParseCatalogPI(ctxt, buf);
- }
- #endif
- /*
- * SAX: PI detected.
- */
- if ((ctxt->sax) && (!ctxt->disableSAX) &&
- (ctxt->sax->processingInstruction != NULL))
- ctxt->sax->processingInstruction(ctxt->userData,
- target, buf);
- }
- xmlFree(buf);
- } else {
- xmlFatalErr(ctxt, XML_ERR_PI_NOT_STARTED, NULL);
- }
- if (ctxt->instate != XML_PARSER_EOF)
- ctxt->instate = state;
- }
- }
- /**
- * xmlParseNotationDecl:
- * @ctxt: an XML parser context
- *
- * parse a notation declaration
- *
- * [82] NotationDecl ::= '<!NOTATION' S Name S (ExternalID | PublicID) S? '>'
- *
- * Hence there is actually 3 choices:
- * 'PUBLIC' S PubidLiteral
- * 'PUBLIC' S PubidLiteral S SystemLiteral
- * and 'SYSTEM' S SystemLiteral
- *
- * See the NOTE on xmlParseExternalID().
- */
- void
- xmlParseNotationDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- xmlChar *Pubid;
- xmlChar *Systemid;
- if (CMP10(CUR_PTR, '<', '!', 'N', 'O', 'T', 'A', 'T', 'I', 'O', 'N')) {
- int inputid = ctxt->input->id;
- SHRINK;
- SKIP(10);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '<!NOTATION'\n");
- return;
- }
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_STARTED, NULL);
- return;
- }
- if (xmlStrchr(name, ':') != NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_COLON,
- "colons are forbidden from notation names '%s'\n",
- name, NULL, NULL);
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the NOTATION name'\n");
- return;
- }
- /*
- * Parse the IDs.
- */
- Systemid = xmlParseExternalID(ctxt, &Pubid, 0);
- SKIP_BLANKS;
- if (RAW == '>') {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Notation declaration doesn't start and stop"
- " in the same entity\n");
- }
- NEXT;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->notationDecl != NULL))
- ctxt->sax->notationDecl(ctxt->userData, name, Pubid, Systemid);
- } else {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_FINISHED, NULL);
- }
- if (Systemid != NULL) xmlFree(Systemid);
- if (Pubid != NULL) xmlFree(Pubid);
- }
- }
- /**
- * xmlParseEntityDecl:
- * @ctxt: an XML parser context
- *
- * parse <!ENTITY declarations
- *
- * [70] EntityDecl ::= GEDecl | PEDecl
- *
- * [71] GEDecl ::= '<!ENTITY' S Name S EntityDef S? '>'
- *
- * [72] PEDecl ::= '<!ENTITY' S '%' S Name S PEDef S? '>'
- *
- * [73] EntityDef ::= EntityValue | (ExternalID NDataDecl?)
- *
- * [74] PEDef ::= EntityValue | ExternalID
- *
- * [76] NDataDecl ::= S 'NDATA' S Name
- *
- * [ VC: Notation Declared ]
- * The Name must match the declared name of a notation.
- */
- void
- xmlParseEntityDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name = NULL;
- xmlChar *value = NULL;
- xmlChar *URI = NULL, *literal = NULL;
- const xmlChar *ndata = NULL;
- int isParameter = 0;
- xmlChar *orig = NULL;
- /* GROW; done in the caller */
- if (CMP8(CUR_PTR, '<', '!', 'E', 'N', 'T', 'I', 'T', 'Y')) {
- int inputid = ctxt->input->id;
- SHRINK;
- SKIP(8);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '<!ENTITY'\n");
- }
- if (RAW == '%') {
- NEXT;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '%%'\n");
- }
- isParameter = 1;
- }
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseEntityDecl: no name\n");
- return;
- }
- if (xmlStrchr(name, ':') != NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_COLON,
- "colons are forbidden from entities names '%s'\n",
- name, NULL, NULL);
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the entity name\n");
- }
- ctxt->instate = XML_PARSER_ENTITY_DECL;
- /*
- * handle the various case of definitions...
- */
- if (isParameter) {
- if ((RAW == '"') || (RAW == '\'')) {
- value = xmlParseEntityValue(ctxt, &orig);
- if (value) {
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) && (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_INTERNAL_PARAMETER_ENTITY,
- NULL, NULL, value);
- }
- } else {
- URI = xmlParseExternalID(ctxt, &literal, 1);
- if ((URI == NULL) && (literal == NULL)) {
- xmlFatalErr(ctxt, XML_ERR_VALUE_REQUIRED, NULL);
- }
- if (URI) {
- xmlURIPtr uri;
- uri = xmlParseURI((const char *) URI);
- if (uri == NULL) {
- xmlErrMsgStr(ctxt, XML_ERR_INVALID_URI,
- "Invalid URI: %s\n", URI);
- /*
- * This really ought to be a well formedness error
- * but the XML Core WG decided otherwise c.f. issue
- * E26 of the XML erratas.
- */
- } else {
- if (uri->fragment != NULL) {
- /*
- * Okay this is foolish to block those but not
- * invalid URIs.
- */
- xmlFatalErr(ctxt, XML_ERR_URI_FRAGMENT, NULL);
- } else {
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) &&
- (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_EXTERNAL_PARAMETER_ENTITY,
- literal, URI, NULL);
- }
- xmlFreeURI(uri);
- }
- }
- }
- } else {
- if ((RAW == '"') || (RAW == '\'')) {
- value = xmlParseEntityValue(ctxt, &orig);
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) && (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_INTERNAL_GENERAL_ENTITY,
- NULL, NULL, value);
- /*
- * For expat compatibility in SAX mode.
- */
- if ((ctxt->myDoc == NULL) ||
- (xmlStrEqual(ctxt->myDoc->version, SAX_COMPAT_MODE))) {
- if (ctxt->myDoc == NULL) {
- ctxt->myDoc = xmlNewDoc(SAX_COMPAT_MODE);
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return;
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- }
- if (ctxt->myDoc->intSubset == NULL)
- ctxt->myDoc->intSubset = xmlNewDtd(ctxt->myDoc,
- BAD_CAST "fake", NULL, NULL);
- xmlSAX2EntityDecl(ctxt, name, XML_INTERNAL_GENERAL_ENTITY,
- NULL, NULL, value);
- }
- } else {
- URI = xmlParseExternalID(ctxt, &literal, 1);
- if ((URI == NULL) && (literal == NULL)) {
- xmlFatalErr(ctxt, XML_ERR_VALUE_REQUIRED, NULL);
- }
- if (URI) {
- xmlURIPtr uri;
- uri = xmlParseURI((const char *)URI);
- if (uri == NULL) {
- xmlErrMsgStr(ctxt, XML_ERR_INVALID_URI,
- "Invalid URI: %s\n", URI);
- /*
- * This really ought to be a well formedness error
- * but the XML Core WG decided otherwise c.f. issue
- * E26 of the XML erratas.
- */
- } else {
- if (uri->fragment != NULL) {
- /*
- * Okay this is foolish to block those but not
- * invalid URIs.
- */
- xmlFatalErr(ctxt, XML_ERR_URI_FRAGMENT, NULL);
- }
- xmlFreeURI(uri);
- }
- }
- if ((RAW != '>') && (SKIP_BLANKS == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required before 'NDATA'\n");
- }
- if (CMP5(CUR_PTR, 'N', 'D', 'A', 'T', 'A')) {
- SKIP(5);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'NDATA'\n");
- }
- ndata = xmlParseName(ctxt);
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->unparsedEntityDecl != NULL))
- ctxt->sax->unparsedEntityDecl(ctxt->userData, name,
- literal, URI, ndata);
- } else {
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) && (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_EXTERNAL_GENERAL_PARSED_ENTITY,
- literal, URI, NULL);
- /*
- * For expat compatibility in SAX mode.
- * assuming the entity replacement was asked for
- */
- if ((ctxt->replaceEntities != 0) &&
- ((ctxt->myDoc == NULL) ||
- (xmlStrEqual(ctxt->myDoc->version, SAX_COMPAT_MODE)))) {
- if (ctxt->myDoc == NULL) {
- ctxt->myDoc = xmlNewDoc(SAX_COMPAT_MODE);
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return;
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- }
- if (ctxt->myDoc->intSubset == NULL)
- ctxt->myDoc->intSubset = xmlNewDtd(ctxt->myDoc,
- BAD_CAST "fake", NULL, NULL);
- xmlSAX2EntityDecl(ctxt, name,
- XML_EXTERNAL_GENERAL_PARSED_ENTITY,
- literal, URI, NULL);
- }
- }
- }
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- SKIP_BLANKS;
- if (RAW != '>') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_NOT_FINISHED,
- "xmlParseEntityDecl: entity %s not terminated\n", name);
- xmlHaltParser(ctxt);
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Entity declaration doesn't start and stop in"
- " the same entity\n");
- }
- NEXT;
- }
- if (orig != NULL) {
- /*
- * Ugly mechanism to save the raw entity value.
- */
- xmlEntityPtr cur = NULL;
- if (isParameter) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getParameterEntity != NULL))
- cur = ctxt->sax->getParameterEntity(ctxt->userData, name);
- } else {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getEntity != NULL))
- cur = ctxt->sax->getEntity(ctxt->userData, name);
- if ((cur == NULL) && (ctxt->userData==ctxt)) {
- cur = xmlSAX2GetEntity(ctxt, name);
- }
- }
- if ((cur != NULL) && (cur->orig == NULL)) {
- cur->orig = orig;
- orig = NULL;
- }
- }
- done:
- if (value != NULL) xmlFree(value);
- if (URI != NULL) xmlFree(URI);
- if (literal != NULL) xmlFree(literal);
- if (orig != NULL) xmlFree(orig);
- }
- }
- /**
- * xmlParseDefaultDecl:
- * @ctxt: an XML parser context
- * @value: Receive a possible fixed default value for the attribute
- *
- * Parse an attribute default declaration
- *
- * [60] DefaultDecl ::= '#REQUIRED' | '#IMPLIED' | (('#FIXED' S)? AttValue)
- *
- * [ VC: Required Attribute ]
- * if the default declaration is the keyword #REQUIRED, then the
- * attribute must be specified for all elements of the type in the
- * attribute-list declaration.
- *
- * [ VC: Attribute Default Legal ]
- * The declared default value must meet the lexical constraints of
- * the declared attribute type c.f. xmlValidateAttributeDecl()
- *
- * [ VC: Fixed Attribute Default ]
- * if an attribute has a default value declared with the #FIXED
- * keyword, instances of that attribute must match the default value.
- *
- * [ WFC: No < in Attribute Values ]
- * handled in xmlParseAttValue()
- *
- * returns: XML_ATTRIBUTE_NONE, XML_ATTRIBUTE_REQUIRED, XML_ATTRIBUTE_IMPLIED
- * or XML_ATTRIBUTE_FIXED.
- */
- int
- xmlParseDefaultDecl(xmlParserCtxtPtr ctxt, xmlChar **value) {
- int val;
- xmlChar *ret;
- *value = NULL;
- if (CMP9(CUR_PTR, '#', 'R', 'E', 'Q', 'U', 'I', 'R', 'E', 'D')) {
- SKIP(9);
- return(XML_ATTRIBUTE_REQUIRED);
- }
- if (CMP8(CUR_PTR, '#', 'I', 'M', 'P', 'L', 'I', 'E', 'D')) {
- SKIP(8);
- return(XML_ATTRIBUTE_IMPLIED);
- }
- val = XML_ATTRIBUTE_NONE;
- if (CMP6(CUR_PTR, '#', 'F', 'I', 'X', 'E', 'D')) {
- SKIP(6);
- val = XML_ATTRIBUTE_FIXED;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '#FIXED'\n");
- }
- }
- ret = xmlParseAttValue(ctxt);
- ctxt->instate = XML_PARSER_DTD;
- if (ret == NULL) {
- xmlFatalErrMsg(ctxt, (xmlParserErrors)ctxt->errNo,
- "Attribute default value declaration error\n");
- } else
- *value = ret;
- return(val);
- }
- /**
- * xmlParseNotationType:
- * @ctxt: an XML parser context
- *
- * parse an Notation attribute type.
- *
- * Note: the leading 'NOTATION' S part has already being parsed...
- *
- * [58] NotationType ::= 'NOTATION' S '(' S? Name (S? '|' S? Name)* S? ')'
- *
- * [ VC: Notation Attributes ]
- * Values of this type must match one of the notation names included
- * in the declaration; all notation names in the declaration must be declared.
- *
- * Returns: the notation attribute tree built while parsing
- */
- xmlEnumerationPtr
- xmlParseNotationType(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- xmlEnumerationPtr ret = NULL, last = NULL, cur, tmp;
- if (RAW != '(') {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_STARTED, NULL);
- return(NULL);
- }
- SHRINK;
- do {
- NEXT;
- SKIP_BLANKS;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "Name expected in NOTATION declaration\n");
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- tmp = ret;
- while (tmp != NULL) {
- if (xmlStrEqual(name, tmp->name)) {
- xmlValidityError(ctxt, XML_DTD_DUP_TOKEN,
- "standalone: attribute notation value token %s duplicated\n",
- name, NULL);
- if (!xmlDictOwns(ctxt->dict, name))
- xmlFree((xmlChar *) name);
- break;
- }
- tmp = tmp->next;
- }
- if (tmp == NULL) {
- cur = xmlCreateEnumeration(name);
- if (cur == NULL) {
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- if (last == NULL) ret = last = cur;
- else {
- last->next = cur;
- last = cur;
- }
- }
- SKIP_BLANKS;
- } while (RAW == '|');
- if (RAW != ')') {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_FINISHED, NULL);
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- NEXT;
- return(ret);
- }
- /**
- * xmlParseEnumerationType:
- * @ctxt: an XML parser context
- *
- * parse an Enumeration attribute type.
- *
- * [59] Enumeration ::= '(' S? Nmtoken (S? '|' S? Nmtoken)* S? ')'
- *
- * [ VC: Enumeration ]
- * Values of this type must match one of the Nmtoken tokens in
- * the declaration
- *
- * Returns: the enumeration attribute tree built while parsing
- */
- xmlEnumerationPtr
- xmlParseEnumerationType(xmlParserCtxtPtr ctxt) {
- xmlChar *name;
- xmlEnumerationPtr ret = NULL, last = NULL, cur, tmp;
- if (RAW != '(') {
- xmlFatalErr(ctxt, XML_ERR_ATTLIST_NOT_STARTED, NULL);
- return(NULL);
- }
- SHRINK;
- do {
- NEXT;
- SKIP_BLANKS;
- name = xmlParseNmtoken(ctxt);
- if (name == NULL) {
- xmlFatalErr(ctxt, XML_ERR_NMTOKEN_REQUIRED, NULL);
- return(ret);
- }
- tmp = ret;
- while (tmp != NULL) {
- if (xmlStrEqual(name, tmp->name)) {
- xmlValidityError(ctxt, XML_DTD_DUP_TOKEN,
- "standalone: attribute enumeration value token %s duplicated\n",
- name, NULL);
- if (!xmlDictOwns(ctxt->dict, name))
- xmlFree(name);
- break;
- }
- tmp = tmp->next;
- }
- if (tmp == NULL) {
- cur = xmlCreateEnumeration(name);
- if (!xmlDictOwns(ctxt->dict, name))
- xmlFree(name);
- if (cur == NULL) {
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- if (last == NULL) ret = last = cur;
- else {
- last->next = cur;
- last = cur;
- }
- }
- SKIP_BLANKS;
- } while (RAW == '|');
- if (RAW != ')') {
- xmlFatalErr(ctxt, XML_ERR_ATTLIST_NOT_FINISHED, NULL);
- return(ret);
- }
- NEXT;
- return(ret);
- }
- /**
- * xmlParseEnumeratedType:
- * @ctxt: an XML parser context
- * @tree: the enumeration tree built while parsing
- *
- * parse an Enumerated attribute type.
- *
- * [57] EnumeratedType ::= NotationType | Enumeration
- *
- * [58] NotationType ::= 'NOTATION' S '(' S? Name (S? '|' S? Name)* S? ')'
- *
- *
- * Returns: XML_ATTRIBUTE_ENUMERATION or XML_ATTRIBUTE_NOTATION
- */
- int
- xmlParseEnumeratedType(xmlParserCtxtPtr ctxt, xmlEnumerationPtr *tree) {
- if (CMP8(CUR_PTR, 'N', 'O', 'T', 'A', 'T', 'I', 'O', 'N')) {
- SKIP(8);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'NOTATION'\n");
- return(0);
- }
- *tree = xmlParseNotationType(ctxt);
- if (*tree == NULL) return(0);
- return(XML_ATTRIBUTE_NOTATION);
- }
- *tree = xmlParseEnumerationType(ctxt);
- if (*tree == NULL) return(0);
- return(XML_ATTRIBUTE_ENUMERATION);
- }
- /**
- * xmlParseAttributeType:
- * @ctxt: an XML parser context
- * @tree: the enumeration tree built while parsing
- *
- * parse the Attribute list def for an element
- *
- * [54] AttType ::= StringType | TokenizedType | EnumeratedType
- *
- * [55] StringType ::= 'CDATA'
- *
- * [56] TokenizedType ::= 'ID' | 'IDREF' | 'IDREFS' | 'ENTITY' |
- * 'ENTITIES' | 'NMTOKEN' | 'NMTOKENS'
- *
- * Validity constraints for attribute values syntax are checked in
- * xmlValidateAttributeValue()
- *
- * [ VC: ID ]
- * Values of type ID must match the Name production. A name must not
- * appear more than once in an XML document as a value of this type;
- * i.e., ID values must uniquely identify the elements which bear them.
- *
- * [ VC: One ID per Element Type ]
- * No element type may have more than one ID attribute specified.
- *
- * [ VC: ID Attribute Default ]
- * An ID attribute must have a declared default of #IMPLIED or #REQUIRED.
- *
- * [ VC: IDREF ]
- * Values of type IDREF must match the Name production, and values
- * of type IDREFS must match Names; each IDREF Name must match the value
- * of an ID attribute on some element in the XML document; i.e. IDREF
- * values must match the value of some ID attribute.
- *
- * [ VC: Entity Name ]
- * Values of type ENTITY must match the Name production, values
- * of type ENTITIES must match Names; each Entity Name must match the
- * name of an unparsed entity declared in the DTD.
- *
- * [ VC: Name Token ]
- * Values of type NMTOKEN must match the Nmtoken production; values
- * of type NMTOKENS must match Nmtokens.
- *
- * Returns the attribute type
- */
- int
- xmlParseAttributeType(xmlParserCtxtPtr ctxt, xmlEnumerationPtr *tree) {
- SHRINK;
- if (CMP5(CUR_PTR, 'C', 'D', 'A', 'T', 'A')) {
- SKIP(5);
- return(XML_ATTRIBUTE_CDATA);
- } else if (CMP6(CUR_PTR, 'I', 'D', 'R', 'E', 'F', 'S')) {
- SKIP(6);
- return(XML_ATTRIBUTE_IDREFS);
- } else if (CMP5(CUR_PTR, 'I', 'D', 'R', 'E', 'F')) {
- SKIP(5);
- return(XML_ATTRIBUTE_IDREF);
- } else if ((RAW == 'I') && (NXT(1) == 'D')) {
- SKIP(2);
- return(XML_ATTRIBUTE_ID);
- } else if (CMP6(CUR_PTR, 'E', 'N', 'T', 'I', 'T', 'Y')) {
- SKIP(6);
- return(XML_ATTRIBUTE_ENTITY);
- } else if (CMP8(CUR_PTR, 'E', 'N', 'T', 'I', 'T', 'I', 'E', 'S')) {
- SKIP(8);
- return(XML_ATTRIBUTE_ENTITIES);
- } else if (CMP8(CUR_PTR, 'N', 'M', 'T', 'O', 'K', 'E', 'N', 'S')) {
- SKIP(8);
- return(XML_ATTRIBUTE_NMTOKENS);
- } else if (CMP7(CUR_PTR, 'N', 'M', 'T', 'O', 'K', 'E', 'N')) {
- SKIP(7);
- return(XML_ATTRIBUTE_NMTOKEN);
- }
- return(xmlParseEnumeratedType(ctxt, tree));
- }
- /**
- * xmlParseAttributeListDecl:
- * @ctxt: an XML parser context
- *
- * : parse the Attribute list def for an element
- *
- * [52] AttlistDecl ::= '<!ATTLIST' S Name AttDef* S? '>'
- *
- * [53] AttDef ::= S Name S AttType S DefaultDecl
- *
- */
- void
- xmlParseAttributeListDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *elemName;
- const xmlChar *attrName;
- xmlEnumerationPtr tree;
- if (CMP9(CUR_PTR, '<', '!', 'A', 'T', 'T', 'L', 'I', 'S', 'T')) {
- int inputid = ctxt->input->id;
- SKIP(9);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '<!ATTLIST'\n");
- }
- elemName = xmlParseName(ctxt);
- if (elemName == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "ATTLIST: no name for Element\n");
- return;
- }
- SKIP_BLANKS;
- GROW;
- while ((RAW != '>') && (ctxt->instate != XML_PARSER_EOF)) {
- int type;
- int def;
- xmlChar *defaultValue = NULL;
- GROW;
- tree = NULL;
- attrName = xmlParseName(ctxt);
- if (attrName == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "ATTLIST: no name for Attribute\n");
- break;
- }
- GROW;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the attribute name\n");
- break;
- }
- type = xmlParseAttributeType(ctxt, &tree);
- if (type <= 0) {
- break;
- }
- GROW;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the attribute type\n");
- if (tree != NULL)
- xmlFreeEnumeration(tree);
- break;
- }
- def = xmlParseDefaultDecl(ctxt, &defaultValue);
- if (def <= 0) {
- if (defaultValue != NULL)
- xmlFree(defaultValue);
- if (tree != NULL)
- xmlFreeEnumeration(tree);
- break;
- }
- if ((type != XML_ATTRIBUTE_CDATA) && (defaultValue != NULL))
- xmlAttrNormalizeSpace(defaultValue, defaultValue);
- GROW;
- if (RAW != '>') {
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the attribute default value\n");
- if (defaultValue != NULL)
- xmlFree(defaultValue);
- if (tree != NULL)
- xmlFreeEnumeration(tree);
- break;
- }
- }
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->attributeDecl != NULL))
- ctxt->sax->attributeDecl(ctxt->userData, elemName, attrName,
- type, def, defaultValue, tree);
- else if (tree != NULL)
- xmlFreeEnumeration(tree);
- if ((ctxt->sax2) && (defaultValue != NULL) &&
- (def != XML_ATTRIBUTE_IMPLIED) &&
- (def != XML_ATTRIBUTE_REQUIRED)) {
- xmlAddDefAttrs(ctxt, elemName, attrName, defaultValue);
- }
- if (ctxt->sax2) {
- xmlAddSpecialAttr(ctxt, elemName, attrName, type);
- }
- if (defaultValue != NULL)
- xmlFree(defaultValue);
- GROW;
- }
- if (RAW == '>') {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Attribute list declaration doesn't start and"
- " stop in the same entity\n");
- }
- NEXT;
- }
- }
- }
- /**
- * xmlParseElementMixedContentDecl:
- * @ctxt: an XML parser context
- * @inputchk: the input used for the current entity, needed for boundary checks
- *
- * parse the declaration for a Mixed Element content
- * The leading '(' and spaces have been skipped in xmlParseElementContentDecl
- *
- * [51] Mixed ::= '(' S? '#PCDATA' (S? '|' S? Name)* S? ')*' |
- * '(' S? '#PCDATA' S? ')'
- *
- * [ VC: Proper Group/PE Nesting ] applies to [51] too (see [49])
- *
- * [ VC: No Duplicate Types ]
- * The same name must not appear more than once in a single
- * mixed-content declaration.
- *
- * returns: the list of the xmlElementContentPtr describing the element choices
- */
- xmlElementContentPtr
- xmlParseElementMixedContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
- xmlElementContentPtr ret = NULL, cur = NULL, n;
- const xmlChar *elem = NULL;
- GROW;
- if (CMP7(CUR_PTR, '#', 'P', 'C', 'D', 'A', 'T', 'A')) {
- SKIP(7);
- SKIP_BLANKS;
- SHRINK;
- if (RAW == ')') {
- if (ctxt->input->id != inputchk) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element content declaration doesn't start and"
- " stop in the same entity\n");
- }
- NEXT;
- ret = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_PCDATA);
- if (ret == NULL)
- return(NULL);
- if (RAW == '*') {
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- NEXT;
- }
- return(ret);
- }
- if ((RAW == '(') || (RAW == '|')) {
- ret = cur = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_PCDATA);
- if (ret == NULL) return(NULL);
- }
- while ((RAW == '|') && (ctxt->instate != XML_PARSER_EOF)) {
- NEXT;
- if (elem == NULL) {
- ret = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
- if (ret == NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, cur);
- return(NULL);
- }
- ret->c1 = cur;
- if (cur != NULL)
- cur->parent = ret;
- cur = ret;
- } else {
- n = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
- if (n == NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- n->c1 = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
- if (n->c1 != NULL)
- n->c1->parent = n;
- cur->c2 = n;
- if (n != NULL)
- n->parent = cur;
- cur = n;
- }
- SKIP_BLANKS;
- elem = xmlParseName(ctxt);
- if (elem == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseElementMixedContentDecl : Name expected\n");
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- SKIP_BLANKS;
- GROW;
- }
- if ((RAW == ')') && (NXT(1) == '*')) {
- if (elem != NULL) {
- cur->c2 = xmlNewDocElementContent(ctxt->myDoc, elem,
- XML_ELEMENT_CONTENT_ELEMENT);
- if (cur->c2 != NULL)
- cur->c2->parent = cur;
- }
- if (ret != NULL)
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- if (ctxt->input->id != inputchk) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element content declaration doesn't start and"
- " stop in the same entity\n");
- }
- SKIP(2);
- } else {
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- xmlFatalErr(ctxt, XML_ERR_MIXED_NOT_STARTED, NULL);
- return(NULL);
- }
- } else {
- xmlFatalErr(ctxt, XML_ERR_PCDATA_REQUIRED, NULL);
- }
- return(ret);
- }
- /**
- * xmlParseElementChildrenContentDeclPriv:
- * @ctxt: an XML parser context
- * @inputchk: the input used for the current entity, needed for boundary checks
- * @depth: the level of recursion
- *
- * parse the declaration for a Mixed Element content
- * The leading '(' and spaces have been skipped in xmlParseElementContentDecl
- *
- *
- * [47] children ::= (choice | seq) ('?' | '*' | '+')?
- *
- * [48] cp ::= (Name | choice | seq) ('?' | '*' | '+')?
- *
- * [49] choice ::= '(' S? cp ( S? '|' S? cp )* S? ')'
- *
- * [50] seq ::= '(' S? cp ( S? ',' S? cp )* S? ')'
- *
- * [ VC: Proper Group/PE Nesting ] applies to [49] and [50]
- * TODO Parameter-entity replacement text must be properly nested
- * with parenthesized groups. That is to say, if either of the
- * opening or closing parentheses in a choice, seq, or Mixed
- * construct is contained in the replacement text for a parameter
- * entity, both must be contained in the same replacement text. For
- * interoperability, if a parameter-entity reference appears in a
- * choice, seq, or Mixed construct, its replacement text should not
- * be empty, and neither the first nor last non-blank character of
- * the replacement text should be a connector (| or ,).
- *
- * Returns the tree of xmlElementContentPtr describing the element
- * hierarchy.
- */
- static xmlElementContentPtr
- xmlParseElementChildrenContentDeclPriv(xmlParserCtxtPtr ctxt, int inputchk,
- int depth) {
- xmlElementContentPtr ret = NULL, cur = NULL, last = NULL, op = NULL;
- const xmlChar *elem;
- xmlChar type = 0;
- if (((depth > 128) && ((ctxt->options & XML_PARSE_HUGE) == 0)) ||
- (depth > 2048)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_ELEMCONTENT_NOT_FINISHED,
- "xmlParseElementChildrenContentDecl : depth %d too deep, use XML_PARSE_HUGE\n",
- depth);
- return(NULL);
- }
- SKIP_BLANKS;
- GROW;
- if (RAW == '(') {
- int inputid = ctxt->input->id;
- /* Recurse on first child */
- NEXT;
- SKIP_BLANKS;
- cur = ret = xmlParseElementChildrenContentDeclPriv(ctxt, inputid,
- depth + 1);
- if (cur == NULL)
- return(NULL);
- SKIP_BLANKS;
- GROW;
- } else {
- elem = xmlParseName(ctxt);
- if (elem == NULL) {
- xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED, NULL);
- return(NULL);
- }
- cur = ret = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
- if (cur == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- GROW;
- if (RAW == '?') {
- cur->ocur = XML_ELEMENT_CONTENT_OPT;
- NEXT;
- } else if (RAW == '*') {
- cur->ocur = XML_ELEMENT_CONTENT_MULT;
- NEXT;
- } else if (RAW == '+') {
- cur->ocur = XML_ELEMENT_CONTENT_PLUS;
- NEXT;
- } else {
- cur->ocur = XML_ELEMENT_CONTENT_ONCE;
- }
- GROW;
- }
- SKIP_BLANKS;
- SHRINK;
- while ((RAW != ')') && (ctxt->instate != XML_PARSER_EOF)) {
- /*
- * Each loop we parse one separator and one element.
- */
- if (RAW == ',') {
- if (type == 0) type = CUR;
- /*
- * Detect "Name | Name , Name" error
- */
- else if (type != CUR) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_SEPARATOR_REQUIRED,
- "xmlParseElementChildrenContentDecl : '%c' expected\n",
- type);
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- NEXT;
- op = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_SEQ);
- if (op == NULL) {
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- if (last == NULL) {
- op->c1 = ret;
- if (ret != NULL)
- ret->parent = op;
- ret = cur = op;
- } else {
- cur->c2 = op;
- if (op != NULL)
- op->parent = cur;
- op->c1 = last;
- if (last != NULL)
- last->parent = op;
- cur =op;
- last = NULL;
- }
- } else if (RAW == '|') {
- if (type == 0) type = CUR;
- /*
- * Detect "Name , Name | Name" error
- */
- else if (type != CUR) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_SEPARATOR_REQUIRED,
- "xmlParseElementChildrenContentDecl : '%c' expected\n",
- type);
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- NEXT;
- op = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
- if (op == NULL) {
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- if (last == NULL) {
- op->c1 = ret;
- if (ret != NULL)
- ret->parent = op;
- ret = cur = op;
- } else {
- cur->c2 = op;
- if (op != NULL)
- op->parent = cur;
- op->c1 = last;
- if (last != NULL)
- last->parent = op;
- cur =op;
- last = NULL;
- }
- } else {
- xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_FINISHED, NULL);
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- GROW;
- SKIP_BLANKS;
- GROW;
- if (RAW == '(') {
- int inputid = ctxt->input->id;
- /* Recurse on second child */
- NEXT;
- SKIP_BLANKS;
- last = xmlParseElementChildrenContentDeclPriv(ctxt, inputid,
- depth + 1);
- if (last == NULL) {
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- SKIP_BLANKS;
- } else {
- elem = xmlParseName(ctxt);
- if (elem == NULL) {
- xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED, NULL);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- last = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
- if (last == NULL) {
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- if (RAW == '?') {
- last->ocur = XML_ELEMENT_CONTENT_OPT;
- NEXT;
- } else if (RAW == '*') {
- last->ocur = XML_ELEMENT_CONTENT_MULT;
- NEXT;
- } else if (RAW == '+') {
- last->ocur = XML_ELEMENT_CONTENT_PLUS;
- NEXT;
- } else {
- last->ocur = XML_ELEMENT_CONTENT_ONCE;
- }
- }
- SKIP_BLANKS;
- GROW;
- }
- if ((cur != NULL) && (last != NULL)) {
- cur->c2 = last;
- if (last != NULL)
- last->parent = cur;
- }
- if (ctxt->input->id != inputchk) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element content declaration doesn't start and stop in"
- " the same entity\n");
- }
- NEXT;
- if (RAW == '?') {
- if (ret != NULL) {
- if ((ret->ocur == XML_ELEMENT_CONTENT_PLUS) ||
- (ret->ocur == XML_ELEMENT_CONTENT_MULT))
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- else
- ret->ocur = XML_ELEMENT_CONTENT_OPT;
- }
- NEXT;
- } else if (RAW == '*') {
- if (ret != NULL) {
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- cur = ret;
- /*
- * Some normalization:
- * (a | b* | c?)* == (a | b | c)*
- */
- while ((cur != NULL) && (cur->type == XML_ELEMENT_CONTENT_OR)) {
- if ((cur->c1 != NULL) &&
- ((cur->c1->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c1->ocur == XML_ELEMENT_CONTENT_MULT)))
- cur->c1->ocur = XML_ELEMENT_CONTENT_ONCE;
- if ((cur->c2 != NULL) &&
- ((cur->c2->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c2->ocur == XML_ELEMENT_CONTENT_MULT)))
- cur->c2->ocur = XML_ELEMENT_CONTENT_ONCE;
- cur = cur->c2;
- }
- }
- NEXT;
- } else if (RAW == '+') {
- if (ret != NULL) {
- int found = 0;
- if ((ret->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (ret->ocur == XML_ELEMENT_CONTENT_MULT))
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- else
- ret->ocur = XML_ELEMENT_CONTENT_PLUS;
- /*
- * Some normalization:
- * (a | b*)+ == (a | b)*
- * (a | b?)+ == (a | b)*
- */
- while ((cur != NULL) && (cur->type == XML_ELEMENT_CONTENT_OR)) {
- if ((cur->c1 != NULL) &&
- ((cur->c1->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c1->ocur == XML_ELEMENT_CONTENT_MULT))) {
- cur->c1->ocur = XML_ELEMENT_CONTENT_ONCE;
- found = 1;
- }
- if ((cur->c2 != NULL) &&
- ((cur->c2->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c2->ocur == XML_ELEMENT_CONTENT_MULT))) {
- cur->c2->ocur = XML_ELEMENT_CONTENT_ONCE;
- found = 1;
- }
- cur = cur->c2;
- }
- if (found)
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- }
- NEXT;
- }
- return(ret);
- }
- /**
- * xmlParseElementChildrenContentDecl:
- * @ctxt: an XML parser context
- * @inputchk: the input used for the current entity, needed for boundary checks
- *
- * parse the declaration for a Mixed Element content
- * The leading '(' and spaces have been skipped in xmlParseElementContentDecl
- *
- * [47] children ::= (choice | seq) ('?' | '*' | '+')?
- *
- * [48] cp ::= (Name | choice | seq) ('?' | '*' | '+')?
- *
- * [49] choice ::= '(' S? cp ( S? '|' S? cp )* S? ')'
- *
- * [50] seq ::= '(' S? cp ( S? ',' S? cp )* S? ')'
- *
- * [ VC: Proper Group/PE Nesting ] applies to [49] and [50]
- * TODO Parameter-entity replacement text must be properly nested
- * with parenthesized groups. That is to say, if either of the
- * opening or closing parentheses in a choice, seq, or Mixed
- * construct is contained in the replacement text for a parameter
- * entity, both must be contained in the same replacement text. For
- * interoperability, if a parameter-entity reference appears in a
- * choice, seq, or Mixed construct, its replacement text should not
- * be empty, and neither the first nor last non-blank character of
- * the replacement text should be a connector (| or ,).
- *
- * Returns the tree of xmlElementContentPtr describing the element
- * hierarchy.
- */
- xmlElementContentPtr
- xmlParseElementChildrenContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
- /* stub left for API/ABI compat */
- return(xmlParseElementChildrenContentDeclPriv(ctxt, inputchk, 1));
- }
- /**
- * xmlParseElementContentDecl:
- * @ctxt: an XML parser context
- * @name: the name of the element being defined.
- * @result: the Element Content pointer will be stored here if any
- *
- * parse the declaration for an Element content either Mixed or Children,
- * the cases EMPTY and ANY are handled directly in xmlParseElementDecl
- *
- * [46] contentspec ::= 'EMPTY' | 'ANY' | Mixed | children
- *
- * returns: the type of element content XML_ELEMENT_TYPE_xxx
- */
- int
- xmlParseElementContentDecl(xmlParserCtxtPtr ctxt, const xmlChar *name,
- xmlElementContentPtr *result) {
- xmlElementContentPtr tree = NULL;
- int inputid = ctxt->input->id;
- int res;
- *result = NULL;
- if (RAW != '(') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED,
- "xmlParseElementContentDecl : %s '(' expected\n", name);
- return(-1);
- }
- NEXT;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- SKIP_BLANKS;
- if (CMP7(CUR_PTR, '#', 'P', 'C', 'D', 'A', 'T', 'A')) {
- tree = xmlParseElementMixedContentDecl(ctxt, inputid);
- res = XML_ELEMENT_TYPE_MIXED;
- } else {
- tree = xmlParseElementChildrenContentDeclPriv(ctxt, inputid, 1);
- res = XML_ELEMENT_TYPE_ELEMENT;
- }
- SKIP_BLANKS;
- *result = tree;
- return(res);
- }
- /**
- * xmlParseElementDecl:
- * @ctxt: an XML parser context
- *
- * parse an Element declaration.
- *
- * [45] elementdecl ::= '<!ELEMENT' S Name S contentspec S? '>'
- *
- * [ VC: Unique Element Type Declaration ]
- * No element type may be declared more than once
- *
- * Returns the type of the element, or -1 in case of error
- */
- int
- xmlParseElementDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- int ret = -1;
- xmlElementContentPtr content = NULL;
- /* GROW; done in the caller */
- if (CMP9(CUR_PTR, '<', '!', 'E', 'L', 'E', 'M', 'E', 'N', 'T')) {
- int inputid = ctxt->input->id;
- SKIP(9);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'ELEMENT'\n");
- return(-1);
- }
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseElementDecl: no name for Element\n");
- return(-1);
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the element name\n");
- }
- if (CMP5(CUR_PTR, 'E', 'M', 'P', 'T', 'Y')) {
- SKIP(5);
- /*
- * Element must always be empty.
- */
- ret = XML_ELEMENT_TYPE_EMPTY;
- } else if ((RAW == 'A') && (NXT(1) == 'N') &&
- (NXT(2) == 'Y')) {
- SKIP(3);
- /*
- * Element is a generic container.
- */
- ret = XML_ELEMENT_TYPE_ANY;
- } else if (RAW == '(') {
- ret = xmlParseElementContentDecl(ctxt, name, &content);
- } else {
- /*
- * [ WFC: PEs in Internal Subset ] error handling.
- */
- if ((RAW == '%') && (ctxt->external == 0) &&
- (ctxt->inputNr == 1)) {
- xmlFatalErrMsg(ctxt, XML_ERR_PEREF_IN_INT_SUBSET,
- "PEReference: forbidden within markup decl in internal subset\n");
- } else {
- xmlFatalErrMsg(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED,
- "xmlParseElementDecl: 'EMPTY', 'ANY' or '(' expected\n");
- }
- return(-1);
- }
- SKIP_BLANKS;
- if (RAW != '>') {
- xmlFatalErr(ctxt, XML_ERR_GT_REQUIRED, NULL);
- if (content != NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, content);
- }
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element declaration doesn't start and stop in"
- " the same entity\n");
- }
- NEXT;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->elementDecl != NULL)) {
- if (content != NULL)
- content->parent = NULL;
- ctxt->sax->elementDecl(ctxt->userData, name, ret,
- content);
- if ((content != NULL) && (content->parent == NULL)) {
- /*
- * this is a trick: if xmlAddElementDecl is called,
- * instead of copying the full tree it is plugged directly
- * if called from the parser. Avoid duplicating the
- * interfaces or change the API/ABI
- */
- xmlFreeDocElementContent(ctxt->myDoc, content);
- }
- } else if (content != NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, content);
- }
- }
- }
- return(ret);
- }
- /**
- * xmlParseConditionalSections
- * @ctxt: an XML parser context
- *
- * [61] conditionalSect ::= includeSect | ignoreSect
- * [62] includeSect ::= '<![' S? 'INCLUDE' S? '[' extSubsetDecl ']]>'
- * [63] ignoreSect ::= '<![' S? 'IGNORE' S? '[' ignoreSectContents* ']]>'
- * [64] ignoreSectContents ::= Ignore ('<![' ignoreSectContents ']]>' Ignore)*
- * [65] Ignore ::= Char* - (Char* ('<![' | ']]>') Char*)
- */
- static void
- xmlParseConditionalSections(xmlParserCtxtPtr ctxt) {
- int *inputIds = NULL;
- size_t inputIdsSize = 0;
- size_t depth = 0;
- while (ctxt->instate != XML_PARSER_EOF) {
- if ((RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- int id = ctxt->input->id;
- SKIP(3);
- SKIP_BLANKS;
- if (CMP7(CUR_PTR, 'I', 'N', 'C', 'L', 'U', 'D', 'E')) {
- SKIP(7);
- SKIP_BLANKS;
- if (RAW != '[') {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_INVALID, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- if (ctxt->input->id != id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is"
- " not in the same entity\n");
- }
- NEXT;
- if (inputIdsSize <= depth) {
- int *tmp;
- inputIdsSize = (inputIdsSize == 0 ? 4 : inputIdsSize * 2);
- tmp = (int *) xmlRealloc(inputIds,
- inputIdsSize * sizeof(int));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- goto error;
- }
- inputIds = tmp;
- }
- inputIds[depth] = id;
- depth++;
- } else if (CMP6(CUR_PTR, 'I', 'G', 'N', 'O', 'R', 'E')) {
- int state;
- xmlParserInputState instate;
- size_t ignoreDepth = 0;
- SKIP(6);
- SKIP_BLANKS;
- if (RAW != '[') {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_INVALID, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- if (ctxt->input->id != id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is"
- " not in the same entity\n");
- }
- NEXT;
- /*
- * Parse up to the end of the conditional section but disable
- * SAX event generating DTD building in the meantime
- */
- state = ctxt->disableSAX;
- instate = ctxt->instate;
- if (ctxt->recovery == 0) ctxt->disableSAX = 1;
- ctxt->instate = XML_PARSER_IGNORE;
- while (RAW != 0) {
- if ((RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- SKIP(3);
- ignoreDepth++;
- /* Check for integer overflow */
- if (ignoreDepth == 0) {
- xmlErrMemory(ctxt, NULL);
- goto error;
- }
- } else if ((RAW == ']') && (NXT(1) == ']') &&
- (NXT(2) == '>')) {
- if (ignoreDepth == 0)
- break;
- SKIP(3);
- ignoreDepth--;
- } else {
- NEXT;
- }
- }
- ctxt->disableSAX = state;
- ctxt->instate = instate;
- if (RAW == 0) {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_NOT_FINISHED, NULL);
- goto error;
- }
- if (ctxt->input->id != id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is"
- " not in the same entity\n");
- }
- SKIP(3);
- } else {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_INVALID_KEYWORD, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- } else if ((depth > 0) &&
- (RAW == ']') && (NXT(1) == ']') && (NXT(2) == '>')) {
- depth--;
- if (ctxt->input->id != inputIds[depth]) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is not"
- " in the same entity\n");
- }
- SKIP(3);
- } else {
- const xmlChar *check = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- xmlParseMarkupDecl(ctxt);
- if ((CUR_PTR == check) && (cons == ctxt->input->consumed)) {
- xmlFatalErr(ctxt, XML_ERR_EXT_SUBSET_NOT_FINISHED, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- }
- if (depth == 0)
- break;
- SKIP_BLANKS;
- GROW;
- }
- error:
- xmlFree(inputIds);
- }
- /**
- * xmlParseMarkupDecl:
- * @ctxt: an XML parser context
- *
- * parse Markup declarations
- *
- * [29] markupdecl ::= elementdecl | AttlistDecl | EntityDecl |
- * NotationDecl | PI | Comment
- *
- * [ VC: Proper Declaration/PE Nesting ]
- * Parameter-entity replacement text must be properly nested with
- * markup declarations. That is to say, if either the first character
- * or the last character of a markup declaration (markupdecl above) is
- * contained in the replacement text for a parameter-entity reference,
- * both must be contained in the same replacement text.
- *
- * [ WFC: PEs in Internal Subset ]
- * In the internal DTD subset, parameter-entity references can occur
- * only where markup declarations can occur, not within markup declarations.
- * (This does not apply to references that occur in external parameter
- * entities or to the external subset.)
- */
- void
- xmlParseMarkupDecl(xmlParserCtxtPtr ctxt) {
- GROW;
- if (CUR == '<') {
- if (NXT(1) == '!') {
- switch (NXT(2)) {
- case 'E':
- if (NXT(3) == 'L')
- xmlParseElementDecl(ctxt);
- else if (NXT(3) == 'N')
- xmlParseEntityDecl(ctxt);
- break;
- case 'A':
- xmlParseAttributeListDecl(ctxt);
- break;
- case 'N':
- xmlParseNotationDecl(ctxt);
- break;
- case '-':
- xmlParseComment(ctxt);
- break;
- default:
- /* there is an error but it will be detected later */
- break;
- }
- } else if (NXT(1) == '?') {
- xmlParsePI(ctxt);
- }
- }
- /*
- * detect requirement to exit there and act accordingly
- * and avoid having instate overridden later on
- */
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- ctxt->instate = XML_PARSER_DTD;
- }
- /**
- * xmlParseTextDecl:
- * @ctxt: an XML parser context
- *
- * parse an XML declaration header for external entities
- *
- * [77] TextDecl ::= '<?xml' VersionInfo? EncodingDecl S? '?>'
- */
- void
- xmlParseTextDecl(xmlParserCtxtPtr ctxt) {
- xmlChar *version;
- const xmlChar *encoding;
- int oldstate;
- /*
- * We know that '<?xml' is here.
- */
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- SKIP(5);
- } else {
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_STARTED, NULL);
- return;
- }
- /* Avoid expansion of parameter entities when skipping blanks. */
- oldstate = ctxt->instate;
- ctxt->instate = XML_PARSER_START;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space needed after '<?xml'\n");
- }
- /*
- * We may have the VersionInfo here.
- */
- version = xmlParseVersionInfo(ctxt);
- if (version == NULL)
- version = xmlCharStrdup(XML_DEFAULT_VERSION);
- else {
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space needed here\n");
- }
- }
- ctxt->input->version = version;
- /*
- * We must have the encoding declaration
- */
- encoding = xmlParseEncodingDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- ctxt->instate = oldstate;
- return;
- }
- if ((encoding == NULL) && (ctxt->errNo == XML_ERR_OK)) {
- xmlFatalErrMsg(ctxt, XML_ERR_MISSING_ENCODING,
- "Missing encoding in text declaration\n");
- }
- SKIP_BLANKS;
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- } else if (RAW == '>') {
- /* Deprecated old WD ... */
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- MOVETO_ENDTAG(CUR_PTR);
- NEXT;
- }
- ctxt->instate = oldstate;
- }
- /**
- * xmlParseExternalSubset:
- * @ctxt: an XML parser context
- * @ExternalID: the external identifier
- * @SystemID: the system identifier (or URL)
- *
- * parse Markup declarations from an external subset
- *
- * [30] extSubset ::= textDecl? extSubsetDecl
- *
- * [31] extSubsetDecl ::= (markupdecl | conditionalSect | PEReference | S) *
- */
- void
- xmlParseExternalSubset(xmlParserCtxtPtr ctxt, const xmlChar *ExternalID,
- const xmlChar *SystemID) {
- xmlDetectSAX2(ctxt);
- GROW;
- if ((ctxt->encoding == NULL) &&
- (ctxt->input->end - ctxt->input->cur >= 4)) {
- xmlChar start[4];
- xmlCharEncoding enc;
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE)
- xmlSwitchEncoding(ctxt, enc);
- }
- if (CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) {
- xmlParseTextDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- xmlHaltParser(ctxt);
- return;
- }
- }
- if (ctxt->myDoc == NULL) {
- ctxt->myDoc = xmlNewDoc(BAD_CAST "1.0");
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return;
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- }
- if ((ctxt->myDoc != NULL) && (ctxt->myDoc->intSubset == NULL))
- xmlCreateIntSubset(ctxt->myDoc, NULL, ExternalID, SystemID);
- ctxt->instate = XML_PARSER_DTD;
- ctxt->external = 1;
- SKIP_BLANKS;
- while (((RAW == '<') && (NXT(1) == '?')) ||
- ((RAW == '<') && (NXT(1) == '!')) ||
- (RAW == '%')) {
- const xmlChar *check = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- GROW;
- if ((RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- xmlParseConditionalSections(ctxt);
- } else
- xmlParseMarkupDecl(ctxt);
- SKIP_BLANKS;
- if ((CUR_PTR == check) && (cons == ctxt->input->consumed)) {
- xmlFatalErr(ctxt, XML_ERR_EXT_SUBSET_NOT_FINISHED, NULL);
- break;
- }
- }
- if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXT_SUBSET_NOT_FINISHED, NULL);
- }
- }
- /**
- * xmlParseReference:
- * @ctxt: an XML parser context
- *
- * parse and handle entity references in content, depending on the SAX
- * interface, this may end-up in a call to character() if this is a
- * CharRef, a predefined entity, if there is no reference() callback.
- * or if the parser was asked to switch to that mode.
- *
- * [67] Reference ::= EntityRef | CharRef
- */
- void
- xmlParseReference(xmlParserCtxtPtr ctxt) {
- xmlEntityPtr ent;
- xmlChar *val;
- int was_checked;
- xmlNodePtr list = NULL;
- xmlParserErrors ret = XML_ERR_OK;
- if (RAW != '&')
- return;
- /*
- * Simple case of a CharRef
- */
- if (NXT(1) == '#') {
- int i = 0;
- xmlChar out[16];
- int hex = NXT(2);
- int value = xmlParseCharRef(ctxt);
- if (value == 0)
- return;
- if (ctxt->charset != XML_CHAR_ENCODING_UTF8) {
- /*
- * So we are using non-UTF-8 buffers
- * Check that the char fit on 8bits, if not
- * generate a CharRef.
- */
- if (value <= 0xFF) {
- out[0] = value;
- out[1] = 0;
- if ((ctxt->sax != NULL) && (ctxt->sax->characters != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->characters(ctxt->userData, out, 1);
- } else {
- if ((hex == 'x') || (hex == 'X'))
- snprintf((char *)out, sizeof(out), "#x%X", value);
- else
- snprintf((char *)out, sizeof(out), "#%d", value);
- if ((ctxt->sax != NULL) && (ctxt->sax->reference != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->reference(ctxt->userData, out);
- }
- } else {
- /*
- * Just encode the value in UTF-8
- */
- COPY_BUF(0 ,out, i, value);
- out[i] = 0;
- if ((ctxt->sax != NULL) && (ctxt->sax->characters != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->characters(ctxt->userData, out, i);
- }
- return;
- }
- /*
- * We are seeing an entity reference
- */
- ent = xmlParseEntityRef(ctxt);
- if (ent == NULL) return;
- if (!ctxt->wellFormed)
- return;
- was_checked = ent->checked;
- /* special case of predefined entities */
- if ((ent->name == NULL) ||
- (ent->etype == XML_INTERNAL_PREDEFINED_ENTITY)) {
- val = ent->content;
- if (val == NULL) return;
- /*
- * inline the entity.
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->characters != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->characters(ctxt->userData, val, xmlStrlen(val));
- return;
- }
- /*
- * The first reference to the entity trigger a parsing phase
- * where the ent->children is filled with the result from
- * the parsing.
- * Note: external parsed entities will not be loaded, it is not
- * required for a non-validating parser, unless the parsing option
- * of validating, or substituting entities were given. Doing so is
- * far more secure as the parser will only process data coming from
- * the document entity by default.
- */
- if (((ent->checked == 0) ||
- ((ent->children == NULL) && (ctxt->options & XML_PARSE_NOENT))) &&
- ((ent->etype != XML_EXTERNAL_GENERAL_PARSED_ENTITY) ||
- (ctxt->options & (XML_PARSE_NOENT | XML_PARSE_DTDVALID)))) {
- unsigned long oldnbent = ctxt->nbentities, diff;
- /*
- * This is a bit hackish but this seems the best
- * way to make sure both SAX and DOM entity support
- * behaves okay.
- */
- void *user_data;
- if (ctxt->userData == ctxt)
- user_data = NULL;
- else
- user_data = ctxt->userData;
- /*
- * Check that this entity is well formed
- * 4.3.2: An internal general parsed entity is well-formed
- * if its replacement text matches the production labeled
- * content.
- */
- if (ent->etype == XML_INTERNAL_GENERAL_ENTITY) {
- ctxt->depth++;
- ret = xmlParseBalancedChunkMemoryInternal(ctxt, ent->content,
- user_data, &list);
- ctxt->depth--;
- } else if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY) {
- ctxt->depth++;
- ret = xmlParseExternalEntityPrivate(ctxt->myDoc, ctxt, ctxt->sax,
- user_data, ctxt->depth, ent->URI,
- ent->ExternalID, &list);
- ctxt->depth--;
- } else {
- ret = XML_ERR_ENTITY_PE_INTERNAL;
- xmlErrMsgStr(ctxt, XML_ERR_INTERNAL_ERROR,
- "invalid entity type found\n", NULL);
- }
- /*
- * Store the number of entities needing parsing for this entity
- * content and do checkings
- */
- diff = ctxt->nbentities - oldnbent + 1;
- if (diff > INT_MAX / 2)
- diff = INT_MAX / 2;
- ent->checked = diff * 2;
- if ((ent->content != NULL) && (xmlStrchr(ent->content, '<')))
- ent->checked |= 1;
- if (ret == XML_ERR_ENTITY_LOOP) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- xmlHaltParser(ctxt);
- xmlFreeNodeList(list);
- return;
- }
- if (xmlParserEntityCheck(ctxt, 0, ent, 0)) {
- xmlFreeNodeList(list);
- return;
- }
- if ((ret == XML_ERR_OK) && (list != NULL)) {
- if (((ent->etype == XML_INTERNAL_GENERAL_ENTITY) ||
- (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY))&&
- (ent->children == NULL)) {
- ent->children = list;
- /*
- * Prune it directly in the generated document
- * except for single text nodes.
- */
- if ((ctxt->replaceEntities == 0) ||
- (ctxt->parseMode == XML_PARSE_READER) ||
- ((list->type == XML_TEXT_NODE) &&
- (list->next == NULL))) {
- ent->owner = 1;
- while (list != NULL) {
- list->parent = (xmlNodePtr) ent;
- xmlSetTreeDoc(list, ent->doc);
- if (list->next == NULL)
- ent->last = list;
- list = list->next;
- }
- list = NULL;
- } else {
- ent->owner = 0;
- while (list != NULL) {
- list->parent = (xmlNodePtr) ctxt->node;
- list->doc = ctxt->myDoc;
- if (list->next == NULL)
- ent->last = list;
- list = list->next;
- }
- list = ent->children;
- #ifdef LIBXML_LEGACY_ENABLED
- if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)
- xmlAddEntityReference(ent, list, NULL);
- #endif /* LIBXML_LEGACY_ENABLED */
- }
- } else {
- xmlFreeNodeList(list);
- list = NULL;
- }
- } else if ((ret != XML_ERR_OK) &&
- (ret != XML_WAR_UNDECLARED_ENTITY)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "Entity '%s' failed to parse\n", ent->name);
- if (ent->content != NULL)
- ent->content[0] = 0;
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- } else if (list != NULL) {
- xmlFreeNodeList(list);
- list = NULL;
- }
- if (ent->checked == 0)
- ent->checked = 2;
- /* Prevent entity from being parsed and expanded twice (Bug 760367). */
- was_checked = 0;
- } else if (ent->checked != 1) {
- ctxt->nbentities += ent->checked / 2;
- }
- /*
- * Now that the entity content has been gathered
- * provide it to the application, this can take different forms based
- * on the parsing modes.
- */
- if (ent->children == NULL) {
- /*
- * Probably running in SAX mode and the callbacks don't
- * build the entity content. So unless we already went
- * though parsing for first checking go though the entity
- * content to generate callbacks associated to the entity
- */
- if (was_checked != 0) {
- void *user_data;
- /*
- * This is a bit hackish but this seems the best
- * way to make sure both SAX and DOM entity support
- * behaves okay.
- */
- if (ctxt->userData == ctxt)
- user_data = NULL;
- else
- user_data = ctxt->userData;
- if (ent->etype == XML_INTERNAL_GENERAL_ENTITY) {
- ctxt->depth++;
- ret = xmlParseBalancedChunkMemoryInternal(ctxt,
- ent->content, user_data, NULL);
- ctxt->depth--;
- } else if (ent->etype ==
- XML_EXTERNAL_GENERAL_PARSED_ENTITY) {
- ctxt->depth++;
- ret = xmlParseExternalEntityPrivate(ctxt->myDoc, ctxt,
- ctxt->sax, user_data, ctxt->depth,
- ent->URI, ent->ExternalID, NULL);
- ctxt->depth--;
- } else {
- ret = XML_ERR_ENTITY_PE_INTERNAL;
- xmlErrMsgStr(ctxt, XML_ERR_INTERNAL_ERROR,
- "invalid entity type found\n", NULL);
- }
- if (ret == XML_ERR_ENTITY_LOOP) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- return;
- }
- }
- if ((ctxt->sax != NULL) && (ctxt->sax->reference != NULL) &&
- (ctxt->replaceEntities == 0) && (!ctxt->disableSAX)) {
- /*
- * Entity reference callback comes second, it's somewhat
- * superfluous but a compatibility to historical behaviour
- */
- ctxt->sax->reference(ctxt->userData, ent->name);
- }
- return;
- }
- /*
- * If we didn't get any children for the entity being built
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->reference != NULL) &&
- (ctxt->replaceEntities == 0) && (!ctxt->disableSAX)) {
- /*
- * Create a node.
- */
- ctxt->sax->reference(ctxt->userData, ent->name);
- return;
- }
- if ((ctxt->replaceEntities) || (ent->children == NULL)) {
- /*
- * There is a problem on the handling of _private for entities
- * (bug 155816): Should we copy the content of the field from
- * the entity (possibly overwriting some value set by the user
- * when a copy is created), should we leave it alone, or should
- * we try to take care of different situations? The problem
- * is exacerbated by the usage of this field by the xmlReader.
- * To fix this bug, we look at _private on the created node
- * and, if it's NULL, we copy in whatever was in the entity.
- * If it's not NULL we leave it alone. This is somewhat of a
- * hack - maybe we should have further tests to determine
- * what to do.
- */
- if ((ctxt->node != NULL) && (ent->children != NULL)) {
- /*
- * Seems we are generating the DOM content, do
- * a simple tree copy for all references except the first
- * In the first occurrence list contains the replacement.
- */
- if (((list == NULL) && (ent->owner == 0)) ||
- (ctxt->parseMode == XML_PARSE_READER)) {
- xmlNodePtr nw = NULL, cur, firstChild = NULL;
- /*
- * We are copying here, make sure there is no abuse
- */
- ctxt->sizeentcopy += ent->length + 5;
- if (xmlParserEntityCheck(ctxt, 0, ent, ctxt->sizeentcopy))
- return;
- /*
- * when operating on a reader, the entities definitions
- * are always owning the entities subtree.
- if (ctxt->parseMode == XML_PARSE_READER)
- ent->owner = 1;
- */
- cur = ent->children;
- while (cur != NULL) {
- nw = xmlDocCopyNode(cur, ctxt->myDoc, 1);
- if (nw != NULL) {
- if (nw->_private == NULL)
- nw->_private = cur->_private;
- if (firstChild == NULL){
- firstChild = nw;
- }
- nw = xmlAddChild(ctxt->node, nw);
- }
- if (cur == ent->last) {
- /*
- * needed to detect some strange empty
- * node cases in the reader tests
- */
- if ((ctxt->parseMode == XML_PARSE_READER) &&
- (nw != NULL) &&
- (nw->type == XML_ELEMENT_NODE) &&
- (nw->children == NULL))
- nw->extra = 1;
- break;
- }
- cur = cur->next;
- }
- #ifdef LIBXML_LEGACY_ENABLED
- if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)
- xmlAddEntityReference(ent, firstChild, nw);
- #endif /* LIBXML_LEGACY_ENABLED */
- } else if ((list == NULL) || (ctxt->inputNr > 0)) {
- xmlNodePtr nw = NULL, cur, next, last,
- firstChild = NULL;
- /*
- * We are copying here, make sure there is no abuse
- */
- ctxt->sizeentcopy += ent->length + 5;
- if (xmlParserEntityCheck(ctxt, 0, ent, ctxt->sizeentcopy))
- return;
- /*
- * Copy the entity child list and make it the new
- * entity child list. The goal is to make sure any
- * ID or REF referenced will be the one from the
- * document content and not the entity copy.
- */
- cur = ent->children;
- ent->children = NULL;
- last = ent->last;
- ent->last = NULL;
- while (cur != NULL) {
- next = cur->next;
- cur->next = NULL;
- cur->parent = NULL;
- nw = xmlDocCopyNode(cur, ctxt->myDoc, 1);
- if (nw != NULL) {
- if (nw->_private == NULL)
- nw->_private = cur->_private;
- if (firstChild == NULL){
- firstChild = cur;
- }
- xmlAddChild((xmlNodePtr) ent, nw);
- xmlAddChild(ctxt->node, cur);
- }
- if (cur == last)
- break;
- cur = next;
- }
- if (ent->owner == 0)
- ent->owner = 1;
- #ifdef LIBXML_LEGACY_ENABLED
- if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)
- xmlAddEntityReference(ent, firstChild, nw);
- #endif /* LIBXML_LEGACY_ENABLED */
- } else {
- const xmlChar *nbktext;
- /*
- * the name change is to avoid coalescing of the
- * node with a possible previous text one which
- * would make ent->children a dangling pointer
- */
- nbktext = xmlDictLookup(ctxt->dict, BAD_CAST "nbktext",
- -1);
- if (ent->children->type == XML_TEXT_NODE)
- ent->children->name = nbktext;
- if ((ent->last != ent->children) &&
- (ent->last->type == XML_TEXT_NODE))
- ent->last->name = nbktext;
- xmlAddChildList(ctxt->node, ent->children);
- }
- /*
- * This is to avoid a nasty side effect, see
- * characters() in SAX.c
- */
- ctxt->nodemem = 0;
- ctxt->nodelen = 0;
- return;
- }
- }
- }
- /**
- * xmlParseEntityRef:
- * @ctxt: an XML parser context
- *
- * parse ENTITY references declarations
- *
- * [68] EntityRef ::= '&' Name ';'
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", the Name given in the entity reference
- * must match that in an entity declaration, except that well-formed
- * documents need not declare any of the following entities: amp, lt,
- * gt, apos, quot. The declaration of a parameter entity must precede
- * any reference to it. Similarly, the declaration of a general entity
- * must precede any reference to it which appears in a default value in an
- * attribute-list declaration. Note that if entities are declared in the
- * external subset or in external parameter entities, a non-validating
- * processor is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be declared is a
- * well-formedness constraint only if standalone='yes'.
- *
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an unparsed entity
- *
- * Returns the xmlEntityPtr if found, or NULL otherwise.
- */
- xmlEntityPtr
- xmlParseEntityRef(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- xmlEntityPtr ent = NULL;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- if (RAW != '&')
- return(NULL);
- NEXT;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseEntityRef: no name\n");
- return(NULL);
- }
- if (RAW != ';') {
- xmlFatalErr(ctxt, XML_ERR_ENTITYREF_SEMICOL_MISSING, NULL);
- return(NULL);
- }
- NEXT;
- /*
- * Predefined entities override any extra definition
- */
- if ((ctxt->options & XML_PARSE_OLDSAX) == 0) {
- ent = xmlGetPredefinedEntity(name);
- if (ent != NULL)
- return(ent);
- }
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Ask first SAX for entity resolution, otherwise try the
- * entities which may have stored in the parser context.
- */
- if (ctxt->sax != NULL) {
- if (ctxt->sax->getEntity != NULL)
- ent = ctxt->sax->getEntity(ctxt->userData, name);
- if ((ctxt->wellFormed == 1 ) && (ent == NULL) &&
- (ctxt->options & XML_PARSE_OLDSAX))
- ent = xmlGetPredefinedEntity(name);
- if ((ctxt->wellFormed == 1 ) && (ent == NULL) &&
- (ctxt->userData==ctxt)) {
- ent = xmlSAX2GetEntity(ctxt, name);
- }
- }
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", the
- * Name given in the entity reference must match that in an
- * entity declaration, except that well-formed documents
- * need not declare any of the following entities: amp, lt,
- * gt, apos, quot.
- * The declaration of a parameter entity must precede any
- * reference to it.
- * Similarly, the declaration of a general entity must
- * precede any reference to it which appears in a default
- * value in an attribute-list declaration. Note that if
- * entities are declared in the external subset or in
- * external parameter entities, a non-validating processor
- * is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be
- * declared is a well-formedness constraint only if
- * standalone='yes'.
- */
- if (ent == NULL) {
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) &&
- (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n", name);
- } else {
- xmlErrMsgStr(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n", name);
- if ((ctxt->inSubset == 0) &&
- (ctxt->sax != NULL) &&
- (ctxt->sax->reference != NULL)) {
- ctxt->sax->reference(ctxt->userData, name);
- }
- }
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- ctxt->valid = 0;
- }
- /*
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an
- * unparsed entity
- */
- else if (ent->etype == XML_EXTERNAL_GENERAL_UNPARSED_ENTITY) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNPARSED_ENTITY,
- "Entity reference to unparsed entity %s\n", name);
- }
- /*
- * [ WFC: No External Entity References ]
- * Attribute values cannot contain direct or indirect
- * entity references to external entities.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_EXTERNAL,
- "Attribute references external entity '%s'\n", name);
- }
- /*
- * [ WFC: No < in Attribute Values ]
- * The replacement text of any entity referred to directly or
- * indirectly in an attribute value (other than "<") must
- * not contain a <.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent != NULL) &&
- (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY)) {
- if (((ent->checked & 1) || (ent->checked == 0)) &&
- (ent->content != NULL) && (xmlStrchr(ent->content, '<'))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_LT_IN_ATTRIBUTE,
- "'<' in entity '%s' is not allowed in attributes values\n", name);
- }
- }
- /*
- * Internal check, no parameter entities here ...
- */
- else {
- switch (ent->etype) {
- case XML_INTERNAL_PARAMETER_ENTITY:
- case XML_EXTERNAL_PARAMETER_ENTITY:
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_PARAMETER,
- "Attempt to reference the parameter entity '%s'\n",
- name);
- break;
- default:
- break;
- }
- }
- /*
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive reference
- * to itself, either directly or indirectly.
- * Done somewhere else
- */
- return(ent);
- }
- /**
- * xmlParseStringEntityRef:
- * @ctxt: an XML parser context
- * @str: a pointer to an index in the string
- *
- * parse ENTITY references declarations, but this version parses it from
- * a string value.
- *
- * [68] EntityRef ::= '&' Name ';'
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", the Name given in the entity reference
- * must match that in an entity declaration, except that well-formed
- * documents need not declare any of the following entities: amp, lt,
- * gt, apos, quot. The declaration of a parameter entity must precede
- * any reference to it. Similarly, the declaration of a general entity
- * must precede any reference to it which appears in a default value in an
- * attribute-list declaration. Note that if entities are declared in the
- * external subset or in external parameter entities, a non-validating
- * processor is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be declared is a
- * well-formedness constraint only if standalone='yes'.
- *
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an unparsed entity
- *
- * Returns the xmlEntityPtr if found, or NULL otherwise. The str pointer
- * is updated to the current location in the string.
- */
- static xmlEntityPtr
- xmlParseStringEntityRef(xmlParserCtxtPtr ctxt, const xmlChar ** str) {
- xmlChar *name;
- const xmlChar *ptr;
- xmlChar cur;
- xmlEntityPtr ent = NULL;
- if ((str == NULL) || (*str == NULL))
- return(NULL);
- ptr = *str;
- cur = *ptr;
- if (cur != '&')
- return(NULL);
- ptr++;
- name = xmlParseStringName(ctxt, &ptr);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseStringEntityRef: no name\n");
- *str = ptr;
- return(NULL);
- }
- if (*ptr != ';') {
- xmlFatalErr(ctxt, XML_ERR_ENTITYREF_SEMICOL_MISSING, NULL);
- xmlFree(name);
- *str = ptr;
- return(NULL);
- }
- ptr++;
- /*
- * Predefined entities override any extra definition
- */
- if ((ctxt->options & XML_PARSE_OLDSAX) == 0) {
- ent = xmlGetPredefinedEntity(name);
- if (ent != NULL) {
- xmlFree(name);
- *str = ptr;
- return(ent);
- }
- }
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Ask first SAX for entity resolution, otherwise try the
- * entities which may have stored in the parser context.
- */
- if (ctxt->sax != NULL) {
- if (ctxt->sax->getEntity != NULL)
- ent = ctxt->sax->getEntity(ctxt->userData, name);
- if ((ent == NULL) && (ctxt->options & XML_PARSE_OLDSAX))
- ent = xmlGetPredefinedEntity(name);
- if ((ent == NULL) && (ctxt->userData==ctxt)) {
- ent = xmlSAX2GetEntity(ctxt, name);
- }
- }
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(name);
- return(NULL);
- }
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", the
- * Name given in the entity reference must match that in an
- * entity declaration, except that well-formed documents
- * need not declare any of the following entities: amp, lt,
- * gt, apos, quot.
- * The declaration of a parameter entity must precede any
- * reference to it.
- * Similarly, the declaration of a general entity must
- * precede any reference to it which appears in a default
- * value in an attribute-list declaration. Note that if
- * entities are declared in the external subset or in
- * external parameter entities, a non-validating processor
- * is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be
- * declared is a well-formedness constraint only if
- * standalone='yes'.
- */
- if (ent == NULL) {
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) &&
- (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n", name);
- } else {
- xmlErrMsgStr(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n",
- name);
- }
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- /* TODO ? check regressions ctxt->valid = 0; */
- }
- /*
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an
- * unparsed entity
- */
- else if (ent->etype == XML_EXTERNAL_GENERAL_UNPARSED_ENTITY) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNPARSED_ENTITY,
- "Entity reference to unparsed entity %s\n", name);
- }
- /*
- * [ WFC: No External Entity References ]
- * Attribute values cannot contain direct or indirect
- * entity references to external entities.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_EXTERNAL,
- "Attribute references external entity '%s'\n", name);
- }
- /*
- * [ WFC: No < in Attribute Values ]
- * The replacement text of any entity referred to directly or
- * indirectly in an attribute value (other than "<") must
- * not contain a <.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent != NULL) && (ent->content != NULL) &&
- (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
- (xmlStrchr(ent->content, '<'))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_LT_IN_ATTRIBUTE,
- "'<' in entity '%s' is not allowed in attributes values\n",
- name);
- }
- /*
- * Internal check, no parameter entities here ...
- */
- else {
- switch (ent->etype) {
- case XML_INTERNAL_PARAMETER_ENTITY:
- case XML_EXTERNAL_PARAMETER_ENTITY:
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_PARAMETER,
- "Attempt to reference the parameter entity '%s'\n",
- name);
- break;
- default:
- break;
- }
- }
- /*
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive reference
- * to itself, either directly or indirectly.
- * Done somewhere else
- */
- xmlFree(name);
- *str = ptr;
- return(ent);
- }
- /**
- * xmlParsePEReference:
- * @ctxt: an XML parser context
- *
- * parse PEReference declarations
- * The entity content is handled directly by pushing it's content as
- * a new input stream.
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive
- * reference to itself, either directly or indirectly.
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", ... ... The declaration of a parameter
- * entity must precede any reference to it...
- *
- * [ VC: Entity Declared ]
- * In a document with an external subset or external parameter entities
- * with "standalone='no'", ... ... The declaration of a parameter entity
- * must precede any reference to it...
- *
- * [ WFC: In DTD ]
- * Parameter-entity references may only appear in the DTD.
- * NOTE: misleading but this is handled.
- */
- void
- xmlParsePEReference(xmlParserCtxtPtr ctxt)
- {
- const xmlChar *name;
- xmlEntityPtr entity = NULL;
- xmlParserInputPtr input;
- if (RAW != '%')
- return;
- NEXT;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_PEREF_NO_NAME, "PEReference: no name\n");
- return;
- }
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "PEReference: %s\n", name);
- if (RAW != ';') {
- xmlFatalErr(ctxt, XML_ERR_PEREF_SEMICOL_MISSING, NULL);
- return;
- }
- NEXT;
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Request the entity from SAX
- */
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getParameterEntity != NULL))
- entity = ctxt->sax->getParameterEntity(ctxt->userData, name);
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- if (entity == NULL) {
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", ...
- * ... The declaration of a parameter entity must precede
- * any reference to it...
- */
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) &&
- (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name);
- } else {
- /*
- * [ VC: Entity Declared ]
- * In a document with an external subset or external
- * parameter entities with "standalone='no'", ...
- * ... The declaration of a parameter entity must
- * precede any reference to it...
- */
- if ((ctxt->validate) && (ctxt->vctxt.error != NULL)) {
- xmlValidityError(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name, NULL);
- } else
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name, NULL);
- ctxt->valid = 0;
- }
- xmlParserEntityCheck(ctxt, 0, NULL, 0);
- } else {
- /*
- * Internal checking in case the entity quest barfed
- */
- if ((entity->etype != XML_INTERNAL_PARAMETER_ENTITY) &&
- (entity->etype != XML_EXTERNAL_PARAMETER_ENTITY)) {
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "Internal: %%%s; is not a parameter entity\n",
- name, NULL);
- } else {
- xmlChar start[4];
- xmlCharEncoding enc;
- if (xmlParserEntityCheck(ctxt, 0, entity, 0))
- return;
- if ((entity->etype == XML_EXTERNAL_PARAMETER_ENTITY) &&
- ((ctxt->options & XML_PARSE_NOENT) == 0) &&
- ((ctxt->options & XML_PARSE_DTDVALID) == 0) &&
- ((ctxt->options & XML_PARSE_DTDLOAD) == 0) &&
- ((ctxt->options & XML_PARSE_DTDATTR) == 0) &&
- (ctxt->replaceEntities == 0) &&
- (ctxt->validate == 0))
- return;
- input = xmlNewEntityInputStream(ctxt, entity);
- if (xmlPushInput(ctxt, input) < 0) {
- xmlFreeInputStream(input);
- return;
- }
- if (entity->etype == XML_EXTERNAL_PARAMETER_ENTITY) {
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- * Note that, since we may have some non-UTF8
- * encoding (like UTF16, bug 135229), the 'length'
- * is not known, but we can calculate based upon
- * the amount of data in the buffer.
- */
- GROW
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- if ((ctxt->input->end - ctxt->input->cur)>=4) {
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) &&
- (IS_BLANK_CH(NXT(5)))) {
- xmlParseTextDecl(ctxt);
- }
- }
- }
- }
- ctxt->hasPErefs = 1;
- }
- /**
- * xmlLoadEntityContent:
- * @ctxt: an XML parser context
- * @entity: an unloaded system entity
- *
- * Load the original content of the given system entity from the
- * ExternalID/SystemID given. This is to be used for Included in Literal
- * http://www.w3.org/TR/REC-xml/#inliteral processing of entities references
- *
- * Returns 0 in case of success and -1 in case of failure
- */
- static int
- xmlLoadEntityContent(xmlParserCtxtPtr ctxt, xmlEntityPtr entity) {
- xmlParserInputPtr input;
- xmlBufferPtr buf;
- int l, c;
- int count = 0;
- if ((ctxt == NULL) || (entity == NULL) ||
- ((entity->etype != XML_EXTERNAL_PARAMETER_ENTITY) &&
- (entity->etype != XML_EXTERNAL_GENERAL_PARSED_ENTITY)) ||
- (entity->content != NULL)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlLoadEntityContent parameter error");
- return(-1);
- }
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "Reading %s entity content input\n", entity->name);
- buf = xmlBufferCreate();
- if (buf == NULL) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlLoadEntityContent parameter error");
- return(-1);
- }
- input = xmlNewEntityInputStream(ctxt, entity);
- if (input == NULL) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlLoadEntityContent input error");
- xmlBufferFree(buf);
- return(-1);
- }
- /*
- * Push the entity as the current input, read char by char
- * saving to the buffer until the end of the entity or an error
- */
- if (xmlPushInput(ctxt, input) < 0) {
- xmlBufferFree(buf);
- return(-1);
- }
- GROW;
- c = CUR_CHAR(l);
- while ((ctxt->input == input) && (ctxt->input->cur < ctxt->input->end) &&
- (IS_CHAR(c))) {
- xmlBufferAdd(buf, ctxt->input->cur, l);
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlBufferFree(buf);
- return(-1);
- }
- }
- NEXTL(l);
- c = CUR_CHAR(l);
- if (c == 0) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlBufferFree(buf);
- return(-1);
- }
- c = CUR_CHAR(l);
- }
- }
- if ((ctxt->input == input) && (ctxt->input->cur >= ctxt->input->end)) {
- xmlPopInput(ctxt);
- } else if (!IS_CHAR(c)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlLoadEntityContent: invalid char value %d\n",
- c);
- xmlBufferFree(buf);
- return(-1);
- }
- entity->content = buf->content;
- buf->content = NULL;
- xmlBufferFree(buf);
- return(0);
- }
- /**
- * xmlParseStringPEReference:
- * @ctxt: an XML parser context
- * @str: a pointer to an index in the string
- *
- * parse PEReference declarations
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive
- * reference to itself, either directly or indirectly.
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", ... ... The declaration of a parameter
- * entity must precede any reference to it...
- *
- * [ VC: Entity Declared ]
- * In a document with an external subset or external parameter entities
- * with "standalone='no'", ... ... The declaration of a parameter entity
- * must precede any reference to it...
- *
- * [ WFC: In DTD ]
- * Parameter-entity references may only appear in the DTD.
- * NOTE: misleading but this is handled.
- *
- * Returns the string of the entity content.
- * str is updated to the current value of the index
- */
- static xmlEntityPtr
- xmlParseStringPEReference(xmlParserCtxtPtr ctxt, const xmlChar **str) {
- const xmlChar *ptr;
- xmlChar cur;
- xmlChar *name;
- xmlEntityPtr entity = NULL;
- if ((str == NULL) || (*str == NULL)) return(NULL);
- ptr = *str;
- cur = *ptr;
- if (cur != '%')
- return(NULL);
- ptr++;
- name = xmlParseStringName(ctxt, &ptr);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseStringPEReference: no name\n");
- *str = ptr;
- return(NULL);
- }
- cur = *ptr;
- if (cur != ';') {
- xmlFatalErr(ctxt, XML_ERR_ENTITYREF_SEMICOL_MISSING, NULL);
- xmlFree(name);
- *str = ptr;
- return(NULL);
- }
- ptr++;
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Request the entity from SAX
- */
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getParameterEntity != NULL))
- entity = ctxt->sax->getParameterEntity(ctxt->userData, name);
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(name);
- *str = ptr;
- return(NULL);
- }
- if (entity == NULL) {
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", ...
- * ... The declaration of a parameter entity must precede
- * any reference to it...
- */
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) && (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n", name);
- } else {
- /*
- * [ VC: Entity Declared ]
- * In a document with an external subset or external
- * parameter entities with "standalone='no'", ...
- * ... The declaration of a parameter entity must
- * precede any reference to it...
- */
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name, NULL);
- ctxt->valid = 0;
- }
- xmlParserEntityCheck(ctxt, 0, NULL, 0);
- } else {
- /*
- * Internal checking in case the entity quest barfed
- */
- if ((entity->etype != XML_INTERNAL_PARAMETER_ENTITY) &&
- (entity->etype != XML_EXTERNAL_PARAMETER_ENTITY)) {
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "%%%s; is not a parameter entity\n",
- name, NULL);
- }
- }
- ctxt->hasPErefs = 1;
- xmlFree(name);
- *str = ptr;
- return(entity);
- }
- /**
- * xmlParseDocTypeDecl:
- * @ctxt: an XML parser context
- *
- * parse a DOCTYPE declaration
- *
- * [28] doctypedecl ::= '<!DOCTYPE' S Name (S ExternalID)? S?
- * ('[' (markupdecl | PEReference | S)* ']' S?)? '>'
- *
- * [ VC: Root Element Type ]
- * The Name in the document type declaration must match the element
- * type of the root element.
- */
- void
- xmlParseDocTypeDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name = NULL;
- xmlChar *ExternalID = NULL;
- xmlChar *URI = NULL;
- /*
- * We know that '<!DOCTYPE' has been detected.
- */
- SKIP(9);
- SKIP_BLANKS;
- /*
- * Parse the DOCTYPE name.
- */
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseDocTypeDecl : no DOCTYPE name !\n");
- }
- ctxt->intSubName = name;
- SKIP_BLANKS;
- /*
- * Check for SystemID and ExternalID
- */
- URI = xmlParseExternalID(ctxt, &ExternalID, 1);
- if ((URI != NULL) || (ExternalID != NULL)) {
- ctxt->hasExternalSubset = 1;
- }
- ctxt->extSubURI = URI;
- ctxt->extSubSystem = ExternalID;
- SKIP_BLANKS;
- /*
- * Create and update the internal subset.
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->internalSubset != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->internalSubset(ctxt->userData, name, ExternalID, URI);
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- /*
- * Is there any internal subset declarations ?
- * they are handled separately in xmlParseInternalSubset()
- */
- if (RAW == '[')
- return;
- /*
- * We should be at the end of the DOCTYPE declaration.
- */
- if (RAW != '>') {
- xmlFatalErr(ctxt, XML_ERR_DOCTYPE_NOT_FINISHED, NULL);
- }
- NEXT;
- }
- /**
- * xmlParseInternalSubset:
- * @ctxt: an XML parser context
- *
- * parse the internal subset declaration
- *
- * [28 end] ('[' (markupdecl | PEReference | S)* ']' S?)? '>'
- */
- static void
- xmlParseInternalSubset(xmlParserCtxtPtr ctxt) {
- /*
- * Is there any DTD definition ?
- */
- if (RAW == '[') {
- int baseInputNr = ctxt->inputNr;
- ctxt->instate = XML_PARSER_DTD;
- NEXT;
- /*
- * Parse the succession of Markup declarations and
- * PEReferences.
- * Subsequence (markupdecl | PEReference | S)*
- */
- while (((RAW != ']') || (ctxt->inputNr > baseInputNr)) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *check = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- SKIP_BLANKS;
- xmlParseMarkupDecl(ctxt);
- xmlParsePEReference(ctxt);
- /*
- * Conditional sections are allowed from external entities included
- * by PE References in the internal subset.
- */
- if ((ctxt->inputNr > 1) && (ctxt->input->filename != NULL) &&
- (RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- xmlParseConditionalSections(ctxt);
- }
- if ((CUR_PTR == check) && (cons == ctxt->input->consumed)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlParseInternalSubset: error detected in Markup declaration\n");
- if (ctxt->inputNr > baseInputNr)
- xmlPopInput(ctxt);
- else
- break;
- }
- }
- if (RAW == ']') {
- NEXT;
- SKIP_BLANKS;
- }
- }
- /*
- * We should be at the end of the DOCTYPE declaration.
- */
- if (RAW != '>') {
- xmlFatalErr(ctxt, XML_ERR_DOCTYPE_NOT_FINISHED, NULL);
- return;
- }
- NEXT;
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlParseAttribute:
- * @ctxt: an XML parser context
- * @value: a xmlChar ** used to store the value of the attribute
- *
- * parse an attribute
- *
- * [41] Attribute ::= Name Eq AttValue
- *
- * [ WFC: No External Entity References ]
- * Attribute values cannot contain direct or indirect entity references
- * to external entities.
- *
- * [ WFC: No < in Attribute Values ]
- * The replacement text of any entity referred to directly or indirectly in
- * an attribute value (other than "<") must not contain a <.
- *
- * [ VC: Attribute Value Type ]
- * The attribute must have been declared; the value must be of the type
- * declared for it.
- *
- * [25] Eq ::= S? '=' S?
- *
- * With namespace:
- *
- * [NS 11] Attribute ::= QName Eq AttValue
- *
- * Also the case QName == xmlns:??? is handled independently as a namespace
- * definition.
- *
- * Returns the attribute name, and the value in *value.
- */
- const xmlChar *
- xmlParseAttribute(xmlParserCtxtPtr ctxt, xmlChar **value) {
- const xmlChar *name;
- xmlChar *val;
- *value = NULL;
- GROW;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "error parsing attribute name\n");
- return(NULL);
- }
- /*
- * read the value
- */
- SKIP_BLANKS;
- if (RAW == '=') {
- NEXT;
- SKIP_BLANKS;
- val = xmlParseAttValue(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ATTRIBUTE_WITHOUT_VALUE,
- "Specification mandates value for attribute %s\n", name);
- return(NULL);
- }
- /*
- * Check that xml:lang conforms to the specification
- * No more registered as an error, just generate a warning now
- * since this was deprecated in XML second edition
- */
- if ((ctxt->pedantic) && (xmlStrEqual(name, BAD_CAST "xml:lang"))) {
- if (!xmlCheckLanguageID(val)) {
- xmlWarningMsg(ctxt, XML_WAR_LANG_VALUE,
- "Malformed value for xml:lang : %s\n",
- val, NULL);
- }
- }
- /*
- * Check that xml:space conforms to the specification
- */
- if (xmlStrEqual(name, BAD_CAST "xml:space")) {
- if (xmlStrEqual(val, BAD_CAST "default"))
- *(ctxt->space) = 0;
- else if (xmlStrEqual(val, BAD_CAST "preserve"))
- *(ctxt->space) = 1;
- else {
- xmlWarningMsg(ctxt, XML_WAR_SPACE_VALUE,
- "Invalid value \"%s\" for xml:space : \"default\" or \"preserve\" expected\n",
- val, NULL);
- }
- }
- *value = val;
- return(name);
- }
- /**
- * xmlParseStartTag:
- * @ctxt: an XML parser context
- *
- * parse a start of tag either for rule element or
- * EmptyElement. In both case we don't parse the tag closing chars.
- *
- * [40] STag ::= '<' Name (S Attribute)* S? '>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * [44] EmptyElemTag ::= '<' Name (S Attribute)* S? '/>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * With namespace:
- *
- * [NS 8] STag ::= '<' QName (S Attribute)* S? '>'
- *
- * [NS 10] EmptyElement ::= '<' QName (S Attribute)* S? '/>'
- *
- * Returns the element name parsed
- */
- const xmlChar *
- xmlParseStartTag(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- const xmlChar *attname;
- xmlChar *attvalue;
- const xmlChar **atts = ctxt->atts;
- int nbatts = 0;
- int maxatts = ctxt->maxatts;
- int i;
- if (RAW != '<') return(NULL);
- NEXT1;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseStartTag: invalid element name\n");
- return(NULL);
- }
- /*
- * Now parse the attributes, it ends up with the ending
- *
- * (S Attribute)* S?
- */
- SKIP_BLANKS;
- GROW;
- while (((RAW != '>') &&
- ((RAW != '/') || (NXT(1) != '>')) &&
- (IS_BYTE_CHAR(RAW))) && (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *q = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- attname = xmlParseAttribute(ctxt, &attvalue);
- if ((attname != NULL) && (attvalue != NULL)) {
- /*
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same
- * start-tag or empty-element tag.
- */
- for (i = 0; i < nbatts;i += 2) {
- if (xmlStrEqual(atts[i], attname)) {
- xmlErrAttributeDup(ctxt, NULL, attname);
- xmlFree(attvalue);
- goto failed;
- }
- }
- /*
- * Add the pair to atts
- */
- if (atts == NULL) {
- maxatts = 22; /* allow for 10 attrs by default */
- atts = (const xmlChar **)
- xmlMalloc(maxatts * sizeof(xmlChar *));
- if (atts == NULL) {
- xmlErrMemory(ctxt, NULL);
- if (attvalue != NULL)
- xmlFree(attvalue);
- goto failed;
- }
- ctxt->atts = atts;
- ctxt->maxatts = maxatts;
- } else if (nbatts + 4 > maxatts) {
- const xmlChar **n;
- maxatts *= 2;
- n = (const xmlChar **) xmlRealloc((void *) atts,
- maxatts * sizeof(const xmlChar *));
- if (n == NULL) {
- xmlErrMemory(ctxt, NULL);
- if (attvalue != NULL)
- xmlFree(attvalue);
- goto failed;
- }
- atts = n;
- ctxt->atts = atts;
- ctxt->maxatts = maxatts;
- }
- atts[nbatts++] = attname;
- atts[nbatts++] = attvalue;
- atts[nbatts] = NULL;
- atts[nbatts + 1] = NULL;
- } else {
- if (attvalue != NULL)
- xmlFree(attvalue);
- }
- failed:
- GROW
- if ((RAW == '>') || (((RAW == '/') && (NXT(1) == '>'))))
- break;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "attributes construct error\n");
- }
- if ((cons == ctxt->input->consumed) && (q == CUR_PTR) &&
- (attname == NULL) && (attvalue == NULL)) {
- xmlFatalErrMsg(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlParseStartTag: problem parsing attributes\n");
- break;
- }
- SHRINK;
- GROW;
- }
- /*
- * SAX: Start of Element !
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->startElement != NULL) &&
- (!ctxt->disableSAX)) {
- if (nbatts > 0)
- ctxt->sax->startElement(ctxt->userData, name, atts);
- else
- ctxt->sax->startElement(ctxt->userData, name, NULL);
- }
- if (atts != NULL) {
- /* Free only the content strings */
- for (i = 1;i < nbatts;i+=2)
- if (atts[i] != NULL)
- xmlFree((xmlChar *) atts[i]);
- }
- return(name);
- }
- /**
- * xmlParseEndTag1:
- * @ctxt: an XML parser context
- * @line: line of the start tag
- * @nsNr: number of namespaces on the start tag
- *
- * parse an end of tag
- *
- * [42] ETag ::= '</' Name S? '>'
- *
- * With namespace
- *
- * [NS 9] ETag ::= '</' QName S? '>'
- */
- static void
- xmlParseEndTag1(xmlParserCtxtPtr ctxt, int line) {
- const xmlChar *name;
- GROW;
- if ((RAW != '<') || (NXT(1) != '/')) {
- xmlFatalErrMsg(ctxt, XML_ERR_LTSLASH_REQUIRED,
- "xmlParseEndTag: '</' not found\n");
- return;
- }
- SKIP(2);
- name = xmlParseNameAndCompare(ctxt,ctxt->name);
- /*
- * We should definitely be at the ending "S? '>'" part
- */
- GROW;
- SKIP_BLANKS;
- if ((!IS_BYTE_CHAR(RAW)) || (RAW != '>')) {
- xmlFatalErr(ctxt, XML_ERR_GT_REQUIRED, NULL);
- } else
- NEXT1;
- /*
- * [ WFC: Element Type Match ]
- * The Name in an element's end-tag must match the element type in the
- * start-tag.
- *
- */
- if (name != (xmlChar*)1) {
- if (name == NULL) name = BAD_CAST "unparsable";
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NAME_MISMATCH,
- "Opening and ending tag mismatch: %s line %d and %s\n",
- ctxt->name, line, name);
- }
- /*
- * SAX: End of Tag
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->endElement != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElement(ctxt->userData, ctxt->name);
- namePop(ctxt);
- spacePop(ctxt);
- return;
- }
- /**
- * xmlParseEndTag:
- * @ctxt: an XML parser context
- *
- * parse an end of tag
- *
- * [42] ETag ::= '</' Name S? '>'
- *
- * With namespace
- *
- * [NS 9] ETag ::= '</' QName S? '>'
- */
- void
- xmlParseEndTag(xmlParserCtxtPtr ctxt) {
- xmlParseEndTag1(ctxt, 0);
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /************************************************************************
- * *
- * SAX 2 specific operations *
- * *
- ************************************************************************/
- /*
- * xmlGetNamespace:
- * @ctxt: an XML parser context
- * @prefix: the prefix to lookup
- *
- * Lookup the namespace name for the @prefix (which ca be NULL)
- * The prefix must come from the @ctxt->dict dictionary
- *
- * Returns the namespace name or NULL if not bound
- */
- static const xmlChar *
- xmlGetNamespace(xmlParserCtxtPtr ctxt, const xmlChar *prefix) {
- int i;
- if (prefix == ctxt->str_xml) return(ctxt->str_xml_ns);
- for (i = ctxt->nsNr - 2;i >= 0;i-=2)
- if (ctxt->nsTab[i] == prefix) {
- if ((prefix == NULL) && (*ctxt->nsTab[i + 1] == 0))
- return(NULL);
- return(ctxt->nsTab[i + 1]);
- }
- return(NULL);
- }
- /**
- * xmlParseQName:
- * @ctxt: an XML parser context
- * @prefix: pointer to store the prefix part
- *
- * parse an XML Namespace QName
- *
- * [6] QName ::= (Prefix ':')? LocalPart
- * [7] Prefix ::= NCName
- * [8] LocalPart ::= NCName
- *
- * Returns the Name parsed or NULL
- */
- static const xmlChar *
- xmlParseQName(xmlParserCtxtPtr ctxt, const xmlChar **prefix) {
- const xmlChar *l, *p;
- GROW;
- l = xmlParseNCName(ctxt);
- if (l == NULL) {
- if (CUR == ':') {
- l = xmlParseName(ctxt);
- if (l != NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_QNAME,
- "Failed to parse QName '%s'\n", l, NULL, NULL);
- *prefix = NULL;
- return(l);
- }
- }
- return(NULL);
- }
- if (CUR == ':') {
- NEXT;
- p = l;
- l = xmlParseNCName(ctxt);
- if (l == NULL) {
- xmlChar *tmp;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- xmlNsErr(ctxt, XML_NS_ERR_QNAME,
- "Failed to parse QName '%s:'\n", p, NULL, NULL);
- l = xmlParseNmtoken(ctxt);
- if (l == NULL) {
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- tmp = xmlBuildQName(BAD_CAST "", p, NULL, 0);
- } else {
- tmp = xmlBuildQName(l, p, NULL, 0);
- xmlFree((char *)l);
- }
- p = xmlDictLookup(ctxt->dict, tmp, -1);
- if (tmp != NULL) xmlFree(tmp);
- *prefix = NULL;
- return(p);
- }
- if (CUR == ':') {
- xmlChar *tmp;
- xmlNsErr(ctxt, XML_NS_ERR_QNAME,
- "Failed to parse QName '%s:%s:'\n", p, l, NULL);
- NEXT;
- tmp = (xmlChar *) xmlParseName(ctxt);
- if (tmp != NULL) {
- tmp = xmlBuildQName(tmp, l, NULL, 0);
- l = xmlDictLookup(ctxt->dict, tmp, -1);
- if (tmp != NULL) xmlFree(tmp);
- *prefix = p;
- return(l);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- tmp = xmlBuildQName(BAD_CAST "", l, NULL, 0);
- l = xmlDictLookup(ctxt->dict, tmp, -1);
- if (tmp != NULL) xmlFree(tmp);
- *prefix = p;
- return(l);
- }
- *prefix = p;
- } else
- *prefix = NULL;
- return(l);
- }
- /**
- * xmlParseQNameAndCompare:
- * @ctxt: an XML parser context
- * @name: the localname
- * @prefix: the prefix, if any.
- *
- * parse an XML name and compares for match
- * (specialized for endtag parsing)
- *
- * Returns NULL for an illegal name, (xmlChar*) 1 for success
- * and the name for mismatch
- */
- static const xmlChar *
- xmlParseQNameAndCompare(xmlParserCtxtPtr ctxt, xmlChar const *name,
- xmlChar const *prefix) {
- const xmlChar *cmp;
- const xmlChar *in;
- const xmlChar *ret;
- const xmlChar *prefix2;
- if (prefix == NULL) return(xmlParseNameAndCompare(ctxt, name));
- GROW;
- in = ctxt->input->cur;
- cmp = prefix;
- while (*in != 0 && *in == *cmp) {
- ++in;
- ++cmp;
- }
- if ((*cmp == 0) && (*in == ':')) {
- in++;
- cmp = name;
- while (*in != 0 && *in == *cmp) {
- ++in;
- ++cmp;
- }
- if (*cmp == 0 && (*in == '>' || IS_BLANK_CH (*in))) {
- /* success */
- ctxt->input->col += in - ctxt->input->cur;
- ctxt->input->cur = in;
- return((const xmlChar*) 1);
- }
- }
- /*
- * all strings coms from the dictionary, equality can be done directly
- */
- ret = xmlParseQName (ctxt, &prefix2);
- if ((ret == name) && (prefix == prefix2))
- return((const xmlChar*) 1);
- return ret;
- }
- /**
- * xmlParseAttValueInternal:
- * @ctxt: an XML parser context
- * @len: attribute len result
- * @alloc: whether the attribute was reallocated as a new string
- * @normalize: if 1 then further non-CDATA normalization must be done
- *
- * parse a value for an attribute.
- * NOTE: if no normalization is needed, the routine will return pointers
- * directly from the data buffer.
- *
- * 3.3.3 Attribute-Value Normalization:
- * Before the value of an attribute is passed to the application or
- * checked for validity, the XML processor must normalize it as follows:
- * - a character reference is processed by appending the referenced
- * character to the attribute value
- * - an entity reference is processed by recursively processing the
- * replacement text of the entity
- * - a whitespace character (#x20, #xD, #xA, #x9) is processed by
- * appending #x20 to the normalized value, except that only a single
- * #x20 is appended for a "#xD#xA" sequence that is part of an external
- * parsed entity or the literal entity value of an internal parsed entity
- * - other characters are processed by appending them to the normalized value
- * If the declared value is not CDATA, then the XML processor must further
- * process the normalized attribute value by discarding any leading and
- * trailing space (#x20) characters, and by replacing sequences of space
- * (#x20) characters by a single space (#x20) character.
- * All attributes for which no declaration has been read should be treated
- * by a non-validating parser as if declared CDATA.
- *
- * Returns the AttValue parsed or NULL. The value has to be freed by the
- * caller if it was copied, this can be detected by val[*len] == 0.
- */
- #define GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end) \
- const xmlChar *oldbase = ctxt->input->base;\
- GROW;\
- if (ctxt->instate == XML_PARSER_EOF)\
- return(NULL);\
- if (oldbase != ctxt->input->base) {\
- ptrdiff_t delta = ctxt->input->base - oldbase;\
- start = start + delta;\
- in = in + delta;\
- }\
- end = ctxt->input->end;
- static xmlChar *
- xmlParseAttValueInternal(xmlParserCtxtPtr ctxt, int *len, int *alloc,
- int normalize)
- {
- xmlChar limit = 0;
- const xmlChar *in = NULL, *start, *end, *last;
- xmlChar *ret = NULL;
- int line, col;
- GROW;
- in = (xmlChar *) CUR_PTR;
- line = ctxt->input->line;
- col = ctxt->input->col;
- if (*in != '"' && *in != '\'') {
- xmlFatalErr(ctxt, XML_ERR_ATTRIBUTE_NOT_STARTED, NULL);
- return (NULL);
- }
- ctxt->instate = XML_PARSER_ATTRIBUTE_VALUE;
- /*
- * try to handle in this routine the most common case where no
- * allocation of a new string is required and where content is
- * pure ASCII.
- */
- limit = *in++;
- col++;
- end = ctxt->input->end;
- start = in;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- }
- if (normalize) {
- /*
- * Skip any leading spaces
- */
- while ((in < end) && (*in != limit) &&
- ((*in == 0x20) || (*in == 0x9) ||
- (*in == 0xA) || (*in == 0xD))) {
- if (*in == 0xA) {
- line++; col = 1;
- } else {
- col++;
- }
- in++;
- start = in;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- while ((in < end) && (*in != limit) && (*in >= 0x20) &&
- (*in <= 0x7f) && (*in != '&') && (*in != '<')) {
- col++;
- if ((*in++ == 0x20) && (*in == 0x20)) break;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- last = in;
- /*
- * skip the trailing blanks
- */
- while ((last[-1] == 0x20) && (last > start)) last--;
- while ((in < end) && (*in != limit) &&
- ((*in == 0x20) || (*in == 0x9) ||
- (*in == 0xA) || (*in == 0xD))) {
- if (*in == 0xA) {
- line++, col = 1;
- } else {
- col++;
- }
- in++;
- if (in >= end) {
- const xmlChar *oldbase = ctxt->input->base;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- if (oldbase != ctxt->input->base) {
- ptrdiff_t delta = ctxt->input->base - oldbase;
- start = start + delta;
- in = in + delta;
- last = last + delta;
- }
- end = ctxt->input->end;
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- if (*in != limit) goto need_complex;
- } else {
- while ((in < end) && (*in != limit) && (*in >= 0x20) &&
- (*in <= 0x7f) && (*in != '&') && (*in != '<')) {
- in++;
- col++;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- last = in;
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- if (*in != limit) goto need_complex;
- }
- in++;
- col++;
- if (len != NULL) {
- *len = last - start;
- ret = (xmlChar *) start;
- } else {
- if (alloc) *alloc = 1;
- ret = xmlStrndup(start, last - start);
- }
- CUR_PTR = in;
- ctxt->input->line = line;
- ctxt->input->col = col;
- if (alloc) *alloc = 0;
- return ret;
- need_complex:
- if (alloc) *alloc = 1;
- return xmlParseAttValueComplex(ctxt, len, normalize);
- }
- /**
- * xmlParseAttribute2:
- * @ctxt: an XML parser context
- * @pref: the element prefix
- * @elem: the element name
- * @prefix: a xmlChar ** used to store the value of the attribute prefix
- * @value: a xmlChar ** used to store the value of the attribute
- * @len: an int * to save the length of the attribute
- * @alloc: an int * to indicate if the attribute was allocated
- *
- * parse an attribute in the new SAX2 framework.
- *
- * Returns the attribute name, and the value in *value, .
- */
- static const xmlChar *
- xmlParseAttribute2(xmlParserCtxtPtr ctxt,
- const xmlChar * pref, const xmlChar * elem,
- const xmlChar ** prefix, xmlChar ** value,
- int *len, int *alloc)
- {
- const xmlChar *name;
- xmlChar *val, *internal_val = NULL;
- int normalize = 0;
- *value = NULL;
- GROW;
- name = xmlParseQName(ctxt, prefix);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "error parsing attribute name\n");
- return (NULL);
- }
- /*
- * get the type if needed
- */
- if (ctxt->attsSpecial != NULL) {
- int type;
- type = (int) (ptrdiff_t) xmlHashQLookup2(ctxt->attsSpecial,
- pref, elem, *prefix, name);
- if (type != 0)
- normalize = 1;
- }
- /*
- * read the value
- */
- SKIP_BLANKS;
- if (RAW == '=') {
- NEXT;
- SKIP_BLANKS;
- val = xmlParseAttValueInternal(ctxt, len, alloc, normalize);
- if (normalize) {
- /*
- * Sometimes a second normalisation pass for spaces is needed
- * but that only happens if charrefs or entities references
- * have been used in the attribute value, i.e. the attribute
- * value have been extracted in an allocated string already.
- */
- if (*alloc) {
- const xmlChar *val2;
- val2 = xmlAttrNormalizeSpace2(ctxt, val, len);
- if ((val2 != NULL) && (val2 != val)) {
- xmlFree(val);
- val = (xmlChar *) val2;
- }
- }
- }
- ctxt->instate = XML_PARSER_CONTENT;
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ATTRIBUTE_WITHOUT_VALUE,
- "Specification mandates value for attribute %s\n",
- name);
- return (NULL);
- }
- if (*prefix == ctxt->str_xml) {
- /*
- * Check that xml:lang conforms to the specification
- * No more registered as an error, just generate a warning now
- * since this was deprecated in XML second edition
- */
- if ((ctxt->pedantic) && (xmlStrEqual(name, BAD_CAST "lang"))) {
- internal_val = xmlStrndup(val, *len);
- if (!xmlCheckLanguageID(internal_val)) {
- xmlWarningMsg(ctxt, XML_WAR_LANG_VALUE,
- "Malformed value for xml:lang : %s\n",
- internal_val, NULL);
- }
- }
- /*
- * Check that xml:space conforms to the specification
- */
- if (xmlStrEqual(name, BAD_CAST "space")) {
- internal_val = xmlStrndup(val, *len);
- if (xmlStrEqual(internal_val, BAD_CAST "default"))
- *(ctxt->space) = 0;
- else if (xmlStrEqual(internal_val, BAD_CAST "preserve"))
- *(ctxt->space) = 1;
- else {
- xmlWarningMsg(ctxt, XML_WAR_SPACE_VALUE,
- "Invalid value \"%s\" for xml:space : \"default\" or \"preserve\" expected\n",
- internal_val, NULL);
- }
- }
- if (internal_val) {
- xmlFree(internal_val);
- }
- }
- *value = val;
- return (name);
- }
- /**
- * xmlParseStartTag2:
- * @ctxt: an XML parser context
- *
- * parse a start of tag either for rule element or
- * EmptyElement. In both case we don't parse the tag closing chars.
- * This routine is called when running SAX2 parsing
- *
- * [40] STag ::= '<' Name (S Attribute)* S? '>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * [44] EmptyElemTag ::= '<' Name (S Attribute)* S? '/>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * With namespace:
- *
- * [NS 8] STag ::= '<' QName (S Attribute)* S? '>'
- *
- * [NS 10] EmptyElement ::= '<' QName (S Attribute)* S? '/>'
- *
- * Returns the element name parsed
- */
- static const xmlChar *
- xmlParseStartTag2(xmlParserCtxtPtr ctxt, const xmlChar **pref,
- const xmlChar **URI, int *tlen) {
- const xmlChar *localname;
- const xmlChar *prefix;
- const xmlChar *attname;
- const xmlChar *aprefix;
- const xmlChar *nsname;
- xmlChar *attvalue;
- const xmlChar **atts = ctxt->atts;
- int maxatts = ctxt->maxatts;
- int nratts, nbatts, nbdef, inputid;
- int i, j, nbNs, attval;
- unsigned long cur;
- int nsNr = ctxt->nsNr;
- if (RAW != '<') return(NULL);
- NEXT1;
- /*
- * NOTE: it is crucial with the SAX2 API to never call SHRINK beyond that
- * point since the attribute values may be stored as pointers to
- * the buffer and calling SHRINK would destroy them !
- * The Shrinking is only possible once the full set of attribute
- * callbacks have been done.
- */
- SHRINK;
- cur = ctxt->input->cur - ctxt->input->base;
- inputid = ctxt->input->id;
- nbatts = 0;
- nratts = 0;
- nbdef = 0;
- nbNs = 0;
- attval = 0;
- /* Forget any namespaces added during an earlier parse of this element. */
- ctxt->nsNr = nsNr;
- localname = xmlParseQName(ctxt, &prefix);
- if (localname == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "StartTag: invalid element name\n");
- return(NULL);
- }
- *tlen = ctxt->input->cur - ctxt->input->base - cur;
- /*
- * Now parse the attributes, it ends up with the ending
- *
- * (S Attribute)* S?
- */
- SKIP_BLANKS;
- GROW;
- while (((RAW != '>') &&
- ((RAW != '/') || (NXT(1) != '>')) &&
- (IS_BYTE_CHAR(RAW))) && (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *q = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- int len = -1, alloc = 0;
- attname = xmlParseAttribute2(ctxt, prefix, localname,
- &aprefix, &attvalue, &len, &alloc);
- if ((attname == NULL) || (attvalue == NULL))
- goto next_attr;
- if (len < 0) len = xmlStrlen(attvalue);
- if ((attname == ctxt->str_xmlns) && (aprefix == NULL)) {
- const xmlChar *URL = xmlDictLookup(ctxt->dict, attvalue, len);
- xmlURIPtr uri;
- if (URL == NULL) {
- xmlErrMemory(ctxt, "dictionary allocation failure");
- if ((attvalue != NULL) && (alloc != 0))
- xmlFree(attvalue);
- localname = NULL;
- goto done;
- }
- if (*URL != 0) {
- uri = xmlParseURI((const char *) URL);
- if (uri == NULL) {
- xmlNsErr(ctxt, XML_WAR_NS_URI,
- "xmlns: '%s' is not a valid URI\n",
- URL, NULL, NULL);
- } else {
- if (uri->scheme == NULL) {
- xmlNsWarn(ctxt, XML_WAR_NS_URI_RELATIVE,
- "xmlns: URI %s is not absolute\n",
- URL, NULL, NULL);
- }
- xmlFreeURI(uri);
- }
- if (URL == ctxt->str_xml_ns) {
- if (attname != ctxt->str_xml) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xml namespace URI cannot be the default namespace\n",
- NULL, NULL, NULL);
- }
- goto next_attr;
- }
- if ((len == 29) &&
- (xmlStrEqual(URL,
- BAD_CAST "http://www.w3.org/2000/xmlns/"))) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "reuse of the xmlns namespace name is forbidden\n",
- NULL, NULL, NULL);
- goto next_attr;
- }
- }
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == NULL)
- break;
- if (j <= nbNs)
- xmlErrAttributeDup(ctxt, NULL, attname);
- else
- if (nsPush(ctxt, NULL, URL) > 0) nbNs++;
- } else if (aprefix == ctxt->str_xmlns) {
- const xmlChar *URL = xmlDictLookup(ctxt->dict, attvalue, len);
- xmlURIPtr uri;
- if (attname == ctxt->str_xml) {
- if (URL != ctxt->str_xml_ns) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xml namespace prefix mapped to wrong URI\n",
- NULL, NULL, NULL);
- }
- /*
- * Do not keep a namespace definition node
- */
- goto next_attr;
- }
- if (URL == ctxt->str_xml_ns) {
- if (attname != ctxt->str_xml) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xml namespace URI mapped to wrong prefix\n",
- NULL, NULL, NULL);
- }
- goto next_attr;
- }
- if (attname == ctxt->str_xmlns) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "redefinition of the xmlns prefix is forbidden\n",
- NULL, NULL, NULL);
- goto next_attr;
- }
- if ((len == 29) &&
- (xmlStrEqual(URL,
- BAD_CAST "http://www.w3.org/2000/xmlns/"))) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "reuse of the xmlns namespace name is forbidden\n",
- NULL, NULL, NULL);
- goto next_attr;
- }
- if ((URL == NULL) || (URL[0] == 0)) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xmlns:%s: Empty XML namespace is not allowed\n",
- attname, NULL, NULL);
- goto next_attr;
- } else {
- uri = xmlParseURI((const char *) URL);
- if (uri == NULL) {
- xmlNsErr(ctxt, XML_WAR_NS_URI,
- "xmlns:%s: '%s' is not a valid URI\n",
- attname, URL, NULL);
- } else {
- if ((ctxt->pedantic) && (uri->scheme == NULL)) {
- xmlNsWarn(ctxt, XML_WAR_NS_URI_RELATIVE,
- "xmlns:%s: URI %s is not absolute\n",
- attname, URL, NULL);
- }
- xmlFreeURI(uri);
- }
- }
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == attname)
- break;
- if (j <= nbNs)
- xmlErrAttributeDup(ctxt, aprefix, attname);
- else
- if (nsPush(ctxt, attname, URL) > 0) nbNs++;
- } else {
- /*
- * Add the pair to atts
- */
- if ((atts == NULL) || (nbatts + 5 > maxatts)) {
- if (xmlCtxtGrowAttrs(ctxt, nbatts + 5) < 0) {
- goto next_attr;
- }
- maxatts = ctxt->maxatts;
- atts = ctxt->atts;
- }
- ctxt->attallocs[nratts++] = alloc;
- atts[nbatts++] = attname;
- atts[nbatts++] = aprefix;
- /*
- * The namespace URI field is used temporarily to point at the
- * base of the current input buffer for non-alloced attributes.
- * When the input buffer is reallocated, all the pointers become
- * invalid, but they can be reconstructed later.
- */
- if (alloc)
- atts[nbatts++] = NULL;
- else
- atts[nbatts++] = ctxt->input->base;
- atts[nbatts++] = attvalue;
- attvalue += len;
- atts[nbatts++] = attvalue;
- /*
- * tag if some deallocation is needed
- */
- if (alloc != 0) attval = 1;
- attvalue = NULL; /* moved into atts */
- }
- next_attr:
- if ((attvalue != NULL) && (alloc != 0)) {
- xmlFree(attvalue);
- attvalue = NULL;
- }
- GROW
- if (ctxt->instate == XML_PARSER_EOF)
- break;
- if ((RAW == '>') || (((RAW == '/') && (NXT(1) == '>'))))
- break;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "attributes construct error\n");
- break;
- }
- if ((cons == ctxt->input->consumed) && (q == CUR_PTR) &&
- (attname == NULL) && (attvalue == NULL)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlParseStartTag: problem parsing attributes\n");
- break;
- }
- GROW;
- }
- if (ctxt->input->id != inputid) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "Unexpected change of input\n");
- localname = NULL;
- goto done;
- }
- /* Reconstruct attribute value pointers. */
- for (i = 0, j = 0; j < nratts; i += 5, j++) {
- if (atts[i+2] != NULL) {
- /*
- * Arithmetic on dangling pointers is technically undefined
- * behavior, but well...
- */
- ptrdiff_t offset = ctxt->input->base - atts[i+2];
- atts[i+2] = NULL; /* Reset repurposed namespace URI */
- atts[i+3] += offset; /* value */
- atts[i+4] += offset; /* valuend */
- }
- }
- /*
- * The attributes defaulting
- */
- if (ctxt->attsDefault != NULL) {
- xmlDefAttrsPtr defaults;
- defaults = xmlHashLookup2(ctxt->attsDefault, localname, prefix);
- if (defaults != NULL) {
- for (i = 0;i < defaults->nbAttrs;i++) {
- attname = defaults->values[5 * i];
- aprefix = defaults->values[5 * i + 1];
- /*
- * special work for namespaces defaulted defs
- */
- if ((attname == ctxt->str_xmlns) && (aprefix == NULL)) {
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == NULL)
- break;
- if (j <= nbNs) continue;
- nsname = xmlGetNamespace(ctxt, NULL);
- if (nsname != defaults->values[5 * i + 2]) {
- if (nsPush(ctxt, NULL,
- defaults->values[5 * i + 2]) > 0)
- nbNs++;
- }
- } else if (aprefix == ctxt->str_xmlns) {
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == attname)
- break;
- if (j <= nbNs) continue;
- nsname = xmlGetNamespace(ctxt, attname);
- if (nsname != defaults->values[2]) {
- if (nsPush(ctxt, attname,
- defaults->values[5 * i + 2]) > 0)
- nbNs++;
- }
- } else {
- /*
- * check that it's not a defined attribute
- */
- for (j = 0;j < nbatts;j+=5) {
- if ((attname == atts[j]) && (aprefix == atts[j+1]))
- break;
- }
- if (j < nbatts) continue;
- if ((atts == NULL) || (nbatts + 5 > maxatts)) {
- if (xmlCtxtGrowAttrs(ctxt, nbatts + 5) < 0) {
- localname = NULL;
- goto done;
- }
- maxatts = ctxt->maxatts;
- atts = ctxt->atts;
- }
- atts[nbatts++] = attname;
- atts[nbatts++] = aprefix;
- if (aprefix == NULL)
- atts[nbatts++] = NULL;
- else
- atts[nbatts++] = xmlGetNamespace(ctxt, aprefix);
- atts[nbatts++] = defaults->values[5 * i + 2];
- atts[nbatts++] = defaults->values[5 * i + 3];
- if ((ctxt->standalone == 1) &&
- (defaults->values[5 * i + 4] != NULL)) {
- xmlValidityError(ctxt, XML_DTD_STANDALONE_DEFAULTED,
- "standalone: attribute %s on %s defaulted from external subset\n",
- attname, localname);
- }
- nbdef++;
- }
- }
- }
- }
- /*
- * The attributes checkings
- */
- for (i = 0; i < nbatts;i += 5) {
- /*
- * The default namespace does not apply to attribute names.
- */
- if (atts[i + 1] != NULL) {
- nsname = xmlGetNamespace(ctxt, atts[i + 1]);
- if (nsname == NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_UNDEFINED_NAMESPACE,
- "Namespace prefix %s for %s on %s is not defined\n",
- atts[i + 1], atts[i], localname);
- }
- atts[i + 2] = nsname;
- } else
- nsname = NULL;
- /*
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same
- * start-tag or empty-element tag.
- * As extended by the Namespace in XML REC.
- */
- for (j = 0; j < i;j += 5) {
- if (atts[i] == atts[j]) {
- if (atts[i+1] == atts[j+1]) {
- xmlErrAttributeDup(ctxt, atts[i+1], atts[i]);
- break;
- }
- if ((nsname != NULL) && (atts[j + 2] == nsname)) {
- xmlNsErr(ctxt, XML_NS_ERR_ATTRIBUTE_REDEFINED,
- "Namespaced Attribute %s in '%s' redefined\n",
- atts[i], nsname, NULL);
- break;
- }
- }
- }
- }
- nsname = xmlGetNamespace(ctxt, prefix);
- if ((prefix != NULL) && (nsname == NULL)) {
- xmlNsErr(ctxt, XML_NS_ERR_UNDEFINED_NAMESPACE,
- "Namespace prefix %s on %s is not defined\n",
- prefix, localname, NULL);
- }
- *pref = prefix;
- *URI = nsname;
- /*
- * SAX: Start of Element !
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->startElementNs != NULL) &&
- (!ctxt->disableSAX)) {
- if (nbNs > 0)
- ctxt->sax->startElementNs(ctxt->userData, localname, prefix,
- nsname, nbNs, &ctxt->nsTab[ctxt->nsNr - 2 * nbNs],
- nbatts / 5, nbdef, atts);
- else
- ctxt->sax->startElementNs(ctxt->userData, localname, prefix,
- nsname, 0, NULL, nbatts / 5, nbdef, atts);
- }
- done:
- /*
- * Free up attribute allocated strings if needed
- */
- if (attval != 0) {
- for (i = 3,j = 0; j < nratts;i += 5,j++)
- if ((ctxt->attallocs[j] != 0) && (atts[i] != NULL))
- xmlFree((xmlChar *) atts[i]);
- }
- return(localname);
- }
- /**
- * xmlParseEndTag2:
- * @ctxt: an XML parser context
- * @line: line of the start tag
- * @nsNr: number of namespaces on the start tag
- *
- * parse an end of tag
- *
- * [42] ETag ::= '</' Name S? '>'
- *
- * With namespace
- *
- * [NS 9] ETag ::= '</' QName S? '>'
- */
- static void
- xmlParseEndTag2(xmlParserCtxtPtr ctxt, const xmlStartTag *tag) {
- const xmlChar *name;
- GROW;
- if ((RAW != '<') || (NXT(1) != '/')) {
- xmlFatalErr(ctxt, XML_ERR_LTSLASH_REQUIRED, NULL);
- return;
- }
- SKIP(2);
- if (tag->prefix == NULL)
- name = xmlParseNameAndCompare(ctxt, ctxt->name);
- else
- name = xmlParseQNameAndCompare(ctxt, ctxt->name, tag->prefix);
- /*
- * We should definitely be at the ending "S? '>'" part
- */
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- SKIP_BLANKS;
- if ((!IS_BYTE_CHAR(RAW)) || (RAW != '>')) {
- xmlFatalErr(ctxt, XML_ERR_GT_REQUIRED, NULL);
- } else
- NEXT1;
- /*
- * [ WFC: Element Type Match ]
- * The Name in an element's end-tag must match the element type in the
- * start-tag.
- *
- */
- if (name != (xmlChar*)1) {
- if (name == NULL) name = BAD_CAST "unparsable";
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NAME_MISMATCH,
- "Opening and ending tag mismatch: %s line %d and %s\n",
- ctxt->name, tag->line, name);
- }
- /*
- * SAX: End of Tag
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->endElementNs != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElementNs(ctxt->userData, ctxt->name, tag->prefix,
- tag->URI);
- spacePop(ctxt);
- if (tag->nsNr != 0)
- nsPop(ctxt, tag->nsNr);
- }
- /**
- * xmlParseCDSect:
- * @ctxt: an XML parser context
- *
- * Parse escaped pure raw content.
- *
- * [18] CDSect ::= CDStart CData CDEnd
- *
- * [19] CDStart ::= '<![CDATA['
- *
- * [20] Data ::= (Char* - (Char* ']]>' Char*))
- *
- * [21] CDEnd ::= ']]>'
- */
- void
- xmlParseCDSect(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- int r, rl;
- int s, sl;
- int cur, l;
- int count = 0;
- /* Check 2.6.0 was NXT(0) not RAW */
- if (CMP9(CUR_PTR, '<', '!', '[', 'C', 'D', 'A', 'T', 'A', '[')) {
- SKIP(9);
- } else
- return;
- ctxt->instate = XML_PARSER_CDATA_SECTION;
- r = CUR_CHAR(rl);
- if (!IS_CHAR(r)) {
- xmlFatalErr(ctxt, XML_ERR_CDATA_NOT_FINISHED, NULL);
- ctxt->instate = XML_PARSER_CONTENT;
- return;
- }
- NEXTL(rl);
- s = CUR_CHAR(sl);
- if (!IS_CHAR(s)) {
- xmlFatalErr(ctxt, XML_ERR_CDATA_NOT_FINISHED, NULL);
- ctxt->instate = XML_PARSER_CONTENT;
- return;
- }
- NEXTL(sl);
- cur = CUR_CHAR(l);
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return;
- }
- while (IS_CHAR(cur) &&
- ((r != ']') || (s != ']') || (cur != '>'))) {
- if (len + 5 >= size) {
- xmlChar *tmp;
- if ((size > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_CDATA_NOT_FINISHED,
- "CData section too big found", NULL);
- xmlFree (buf);
- return;
- }
- tmp = (xmlChar *) xmlRealloc(buf, size * 2 * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buf);
- xmlErrMemory(ctxt, NULL);
- return;
- }
- buf = tmp;
- size *= 2;
- }
- COPY_BUF(rl,buf,len,r);
- r = s;
- rl = sl;
- s = cur;
- sl = l;
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- count = 0;
- }
- NEXTL(l);
- cur = CUR_CHAR(l);
- }
- buf[len] = 0;
- ctxt->instate = XML_PARSER_CONTENT;
- if (cur != '>') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_CDATA_NOT_FINISHED,
- "CData section not finished\n%.50s\n", buf);
- xmlFree(buf);
- return;
- }
- NEXTL(l);
- /*
- * OK the buffer is to be consumed as cdata.
- */
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (ctxt->sax->cdataBlock != NULL)
- ctxt->sax->cdataBlock(ctxt->userData, buf, len);
- else if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData, buf, len);
- }
- xmlFree(buf);
- }
- /**
- * xmlParseContentInternal:
- * @ctxt: an XML parser context
- *
- * Parse a content sequence. Stops at EOF or '</'. Leaves checking of
- * unexpected EOF to the caller.
- */
- static void
- xmlParseContentInternal(xmlParserCtxtPtr ctxt) {
- int nameNr = ctxt->nameNr;
- GROW;
- while ((RAW != 0) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *test = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- const xmlChar *cur = ctxt->input->cur;
- /*
- * First case : a Processing Instruction.
- */
- if ((*cur == '<') && (cur[1] == '?')) {
- xmlParsePI(ctxt);
- }
- /*
- * Second case : a CDSection
- */
- /* 2.6.0 test was *cur not RAW */
- else if (CMP9(CUR_PTR, '<', '!', '[', 'C', 'D', 'A', 'T', 'A', '[')) {
- xmlParseCDSect(ctxt);
- }
- /*
- * Third case : a comment
- */
- else if ((*cur == '<') && (NXT(1) == '!') &&
- (NXT(2) == '-') && (NXT(3) == '-')) {
- xmlParseComment(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- }
- /*
- * Fourth case : a sub-element.
- */
- else if (*cur == '<') {
- if (NXT(1) == '/') {
- if (ctxt->nameNr <= nameNr)
- break;
- xmlParseElementEnd(ctxt);
- } else {
- xmlParseElementStart(ctxt);
- }
- }
- /*
- * Fifth case : a reference. If if has not been resolved,
- * parsing returns it's Name, create the node
- */
- else if (*cur == '&') {
- xmlParseReference(ctxt);
- }
- /*
- * Last case, text. Note that References are handled directly.
- */
- else {
- xmlParseCharData(ctxt, 0);
- }
- GROW;
- SHRINK;
- if ((cons == ctxt->input->consumed) && (test == CUR_PTR)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "detected an error in element content\n");
- xmlHaltParser(ctxt);
- break;
- }
- }
- }
- /**
- * xmlParseContent:
- * @ctxt: an XML parser context
- *
- * Parse a content sequence. Stops at EOF or '</'.
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- */
- void
- xmlParseContent(xmlParserCtxtPtr ctxt) {
- int nameNr = ctxt->nameNr;
- xmlParseContentInternal(ctxt);
- if ((ctxt->instate != XML_PARSER_EOF) && (ctxt->nameNr > nameNr)) {
- const xmlChar *name = ctxt->nameTab[ctxt->nameNr - 1];
- int line = ctxt->pushTab[ctxt->nameNr - 1].line;
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NOT_FINISHED,
- "Premature end of data in tag %s line %d\n",
- name, line, NULL);
- }
- }
- /**
- * xmlParseElement:
- * @ctxt: an XML parser context
- *
- * parse an XML element
- *
- * [39] element ::= EmptyElemTag | STag content ETag
- *
- * [ WFC: Element Type Match ]
- * The Name in an element's end-tag must match the element type in the
- * start-tag.
- *
- */
- void
- xmlParseElement(xmlParserCtxtPtr ctxt) {
- if (xmlParseElementStart(ctxt) != 0)
- return;
- xmlParseContentInternal(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- if (CUR == 0) {
- const xmlChar *name = ctxt->nameTab[ctxt->nameNr - 1];
- int line = ctxt->pushTab[ctxt->nameNr - 1].line;
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NOT_FINISHED,
- "Premature end of data in tag %s line %d\n",
- name, line, NULL);
- return;
- }
- xmlParseElementEnd(ctxt);
- }
- /**
- * xmlParseElementStart:
- * @ctxt: an XML parser context
- *
- * Parse the start of an XML element. Returns -1 in case of error, 0 if an
- * opening tag was parsed, 1 if an empty element was parsed.
- */
- static int
- xmlParseElementStart(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- const xmlChar *prefix = NULL;
- const xmlChar *URI = NULL;
- xmlParserNodeInfo node_info;
- int line, tlen = 0;
- xmlNodePtr ret;
- int nsNr = ctxt->nsNr;
- if (((unsigned int) ctxt->nameNr > xmlParserMaxDepth) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INTERNAL_ERROR,
- "Excessive depth in document: %d use XML_PARSE_HUGE option\n",
- xmlParserMaxDepth);
- xmlHaltParser(ctxt);
- return(-1);
- }
- /* Capture start position */
- if (ctxt->record_info) {
- node_info.begin_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.begin_line = ctxt->input->line;
- }
- if (ctxt->spaceNr == 0)
- spacePush(ctxt, -1);
- else if (*ctxt->space == -2)
- spacePush(ctxt, -1);
- else
- spacePush(ctxt, *ctxt->space);
- line = ctxt->input->line;
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax2)
- #endif /* LIBXML_SAX1_ENABLED */
- name = xmlParseStartTag2(ctxt, &prefix, &URI, &tlen);
- #ifdef LIBXML_SAX1_ENABLED
- else
- name = xmlParseStartTag(ctxt);
- #endif /* LIBXML_SAX1_ENABLED */
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if (name == NULL) {
- spacePop(ctxt);
- return(-1);
- }
- nameNsPush(ctxt, name, prefix, URI, line, ctxt->nsNr - nsNr);
- ret = ctxt->node;
- #ifdef LIBXML_VALID_ENABLED
- /*
- * [ VC: Root Element Type ]
- * The Name in the document type declaration must match the element
- * type of the root element.
- */
- if (ctxt->validate && ctxt->wellFormed && ctxt->myDoc &&
- ctxt->node && (ctxt->node == ctxt->myDoc->children))
- ctxt->valid &= xmlValidateRoot(&ctxt->vctxt, ctxt->myDoc);
- #endif /* LIBXML_VALID_ENABLED */
- /*
- * Check for an Empty Element.
- */
- if ((RAW == '/') && (NXT(1) == '>')) {
- SKIP(2);
- if (ctxt->sax2) {
- if ((ctxt->sax != NULL) && (ctxt->sax->endElementNs != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElementNs(ctxt->userData, name, prefix, URI);
- #ifdef LIBXML_SAX1_ENABLED
- } else {
- if ((ctxt->sax != NULL) && (ctxt->sax->endElement != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElement(ctxt->userData, name);
- #endif /* LIBXML_SAX1_ENABLED */
- }
- namePop(ctxt);
- spacePop(ctxt);
- if (nsNr != ctxt->nsNr)
- nsPop(ctxt, ctxt->nsNr - nsNr);
- if ( ret != NULL && ctxt->record_info ) {
- node_info.end_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.end_line = ctxt->input->line;
- node_info.node = ret;
- xmlParserAddNodeInfo(ctxt, &node_info);
- }
- return(1);
- }
- if (RAW == '>') {
- NEXT1;
- } else {
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_GT_REQUIRED,
- "Couldn't find end of Start Tag %s line %d\n",
- name, line, NULL);
- /*
- * end of parsing of this node.
- */
- nodePop(ctxt);
- namePop(ctxt);
- spacePop(ctxt);
- if (nsNr != ctxt->nsNr)
- nsPop(ctxt, ctxt->nsNr - nsNr);
- /*
- * Capture end position and add node
- */
- if ( ret != NULL && ctxt->record_info ) {
- node_info.end_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.end_line = ctxt->input->line;
- node_info.node = ret;
- xmlParserAddNodeInfo(ctxt, &node_info);
- }
- return(-1);
- }
- return(0);
- }
- /**
- * xmlParseElementEnd:
- * @ctxt: an XML parser context
- *
- * Parse the end of an XML element.
- */
- static void
- xmlParseElementEnd(xmlParserCtxtPtr ctxt) {
- xmlParserNodeInfo node_info;
- xmlNodePtr ret = ctxt->node;
- if (ctxt->nameNr <= 0)
- return;
- /*
- * parse the end of tag: '</' should be here.
- */
- if (ctxt->sax2) {
- xmlParseEndTag2(ctxt, &ctxt->pushTab[ctxt->nameNr - 1]);
- namePop(ctxt);
- }
- #ifdef LIBXML_SAX1_ENABLED
- else
- xmlParseEndTag1(ctxt, 0);
- #endif /* LIBXML_SAX1_ENABLED */
- /*
- * Capture end position and add node
- */
- if ( ret != NULL && ctxt->record_info ) {
- node_info.end_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.end_line = ctxt->input->line;
- node_info.node = ret;
- xmlParserAddNodeInfo(ctxt, &node_info);
- }
- }
- /**
- * xmlParseVersionNum:
- * @ctxt: an XML parser context
- *
- * parse the XML version value.
- *
- * [26] VersionNum ::= '1.' [0-9]+
- *
- * In practice allow [0-9].[0-9]+ at that level
- *
- * Returns the string giving the XML version number, or NULL
- */
- xmlChar *
- xmlParseVersionNum(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = 10;
- xmlChar cur;
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- cur = CUR;
- if (!((cur >= '0') && (cur <= '9'))) {
- xmlFree(buf);
- return(NULL);
- }
- buf[len++] = cur;
- NEXT;
- cur=CUR;
- if (cur != '.') {
- xmlFree(buf);
- return(NULL);
- }
- buf[len++] = cur;
- NEXT;
- cur=CUR;
- while ((cur >= '0') && (cur <= '9')) {
- if (len + 1 >= size) {
- xmlChar *tmp;
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buf);
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- buf = tmp;
- }
- buf[len++] = cur;
- NEXT;
- cur=CUR;
- }
- buf[len] = 0;
- return(buf);
- }
- /**
- * xmlParseVersionInfo:
- * @ctxt: an XML parser context
- *
- * parse the XML version.
- *
- * [24] VersionInfo ::= S 'version' Eq (' VersionNum ' | " VersionNum ")
- *
- * [25] Eq ::= S? '=' S?
- *
- * Returns the version string, e.g. "1.0"
- */
- xmlChar *
- xmlParseVersionInfo(xmlParserCtxtPtr ctxt) {
- xmlChar *version = NULL;
- if (CMP7(CUR_PTR, 'v', 'e', 'r', 's', 'i', 'o', 'n')) {
- SKIP(7);
- SKIP_BLANKS;
- if (RAW != '=') {
- xmlFatalErr(ctxt, XML_ERR_EQUAL_REQUIRED, NULL);
- return(NULL);
- }
- NEXT;
- SKIP_BLANKS;
- if (RAW == '"') {
- NEXT;
- version = xmlParseVersionNum(ctxt);
- if (RAW != '"') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else if (RAW == '\''){
- NEXT;
- version = xmlParseVersionNum(ctxt);
- if (RAW != '\'') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_STARTED, NULL);
- }
- }
- return(version);
- }
- /**
- * xmlParseEncName:
- * @ctxt: an XML parser context
- *
- * parse the XML encoding name
- *
- * [81] EncName ::= [A-Za-z] ([A-Za-z0-9._] | '-')*
- *
- * Returns the encoding name value or NULL
- */
- xmlChar *
- xmlParseEncName(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = 10;
- xmlChar cur;
- cur = CUR;
- if (((cur >= 'a') && (cur <= 'z')) ||
- ((cur >= 'A') && (cur <= 'Z'))) {
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- buf[len++] = cur;
- NEXT;
- cur = CUR;
- while (((cur >= 'a') && (cur <= 'z')) ||
- ((cur >= 'A') && (cur <= 'Z')) ||
- ((cur >= '0') && (cur <= '9')) ||
- (cur == '.') || (cur == '_') ||
- (cur == '-')) {
- if (len + 1 >= size) {
- xmlChar *tmp;
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buf);
- return(NULL);
- }
- buf = tmp;
- }
- buf[len++] = cur;
- NEXT;
- cur = CUR;
- if (cur == 0) {
- SHRINK;
- GROW;
- cur = CUR;
- }
- }
- buf[len] = 0;
- } else {
- xmlFatalErr(ctxt, XML_ERR_ENCODING_NAME, NULL);
- }
- return(buf);
- }
- /**
- * xmlParseEncodingDecl:
- * @ctxt: an XML parser context
- *
- * parse the XML encoding declaration
- *
- * [80] EncodingDecl ::= S 'encoding' Eq ('"' EncName '"' | "'" EncName "'")
- *
- * this setups the conversion filters.
- *
- * Returns the encoding value or NULL
- */
- const xmlChar *
- xmlParseEncodingDecl(xmlParserCtxtPtr ctxt) {
- xmlChar *encoding = NULL;
- SKIP_BLANKS;
- if (CMP8(CUR_PTR, 'e', 'n', 'c', 'o', 'd', 'i', 'n', 'g')) {
- SKIP(8);
- SKIP_BLANKS;
- if (RAW != '=') {
- xmlFatalErr(ctxt, XML_ERR_EQUAL_REQUIRED, NULL);
- return(NULL);
- }
- NEXT;
- SKIP_BLANKS;
- if (RAW == '"') {
- NEXT;
- encoding = xmlParseEncName(ctxt);
- if (RAW != '"') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- xmlFree((xmlChar *) encoding);
- return(NULL);
- } else
- NEXT;
- } else if (RAW == '\''){
- NEXT;
- encoding = xmlParseEncName(ctxt);
- if (RAW != '\'') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- xmlFree((xmlChar *) encoding);
- return(NULL);
- } else
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_STARTED, NULL);
- }
- /*
- * Non standard parsing, allowing the user to ignore encoding
- */
- if (ctxt->options & XML_PARSE_IGNORE_ENC) {
- xmlFree((xmlChar *) encoding);
- return(NULL);
- }
- /*
- * UTF-16 encoding switch has already taken place at this stage,
- * more over the little-endian/big-endian selection is already done
- */
- if ((encoding != NULL) &&
- ((!xmlStrcasecmp(encoding, BAD_CAST "UTF-16")) ||
- (!xmlStrcasecmp(encoding, BAD_CAST "UTF16")))) {
- /*
- * If no encoding was passed to the parser, that we are
- * using UTF-16 and no decoder is present i.e. the
- * document is apparently UTF-8 compatible, then raise an
- * encoding mismatch fatal error
- */
- if ((ctxt->encoding == NULL) &&
- (ctxt->input->buf != NULL) &&
- (ctxt->input->buf->encoder == NULL)) {
- xmlFatalErrMsg(ctxt, XML_ERR_INVALID_ENCODING,
- "Document labelled UTF-16 but has UTF-8 content\n");
- }
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = encoding;
- }
- /*
- * UTF-8 encoding is handled natively
- */
- else if ((encoding != NULL) &&
- ((!xmlStrcasecmp(encoding, BAD_CAST "UTF-8")) ||
- (!xmlStrcasecmp(encoding, BAD_CAST "UTF8")))) {
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = encoding;
- }
- else if (encoding != NULL) {
- xmlCharEncodingHandlerPtr handler;
- if (ctxt->input->encoding != NULL)
- xmlFree((xmlChar *) ctxt->input->encoding);
- ctxt->input->encoding = encoding;
- handler = xmlFindCharEncodingHandler((const char *) encoding);
- if (handler != NULL) {
- if (xmlSwitchToEncoding(ctxt, handler) < 0) {
- /* failed to convert */
- ctxt->errNo = XML_ERR_UNSUPPORTED_ENCODING;
- return(NULL);
- }
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNSUPPORTED_ENCODING,
- "Unsupported encoding %s\n", encoding);
- return(NULL);
- }
- }
- }
- return(encoding);
- }
- /**
- * xmlParseSDDecl:
- * @ctxt: an XML parser context
- *
- * parse the XML standalone declaration
- *
- * [32] SDDecl ::= S 'standalone' Eq
- * (("'" ('yes' | 'no') "'") | ('"' ('yes' | 'no')'"'))
- *
- * [ VC: Standalone Document Declaration ]
- * TODO The standalone document declaration must have the value "no"
- * if any external markup declarations contain declarations of:
- * - attributes with default values, if elements to which these
- * attributes apply appear in the document without specifications
- * of values for these attributes, or
- * - entities (other than amp, lt, gt, apos, quot), if references
- * to those entities appear in the document, or
- * - attributes with values subject to normalization, where the
- * attribute appears in the document with a value which will change
- * as a result of normalization, or
- * - element types with element content, if white space occurs directly
- * within any instance of those types.
- *
- * Returns:
- * 1 if standalone="yes"
- * 0 if standalone="no"
- * -2 if standalone attribute is missing or invalid
- * (A standalone value of -2 means that the XML declaration was found,
- * but no value was specified for the standalone attribute).
- */
- int
- xmlParseSDDecl(xmlParserCtxtPtr ctxt) {
- int standalone = -2;
- SKIP_BLANKS;
- if (CMP10(CUR_PTR, 's', 't', 'a', 'n', 'd', 'a', 'l', 'o', 'n', 'e')) {
- SKIP(10);
- SKIP_BLANKS;
- if (RAW != '=') {
- xmlFatalErr(ctxt, XML_ERR_EQUAL_REQUIRED, NULL);
- return(standalone);
- }
- NEXT;
- SKIP_BLANKS;
- if (RAW == '\''){
- NEXT;
- if ((RAW == 'n') && (NXT(1) == 'o')) {
- standalone = 0;
- SKIP(2);
- } else if ((RAW == 'y') && (NXT(1) == 'e') &&
- (NXT(2) == 's')) {
- standalone = 1;
- SKIP(3);
- } else {
- xmlFatalErr(ctxt, XML_ERR_STANDALONE_VALUE, NULL);
- }
- if (RAW != '\'') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else if (RAW == '"'){
- NEXT;
- if ((RAW == 'n') && (NXT(1) == 'o')) {
- standalone = 0;
- SKIP(2);
- } else if ((RAW == 'y') && (NXT(1) == 'e') &&
- (NXT(2) == 's')) {
- standalone = 1;
- SKIP(3);
- } else {
- xmlFatalErr(ctxt, XML_ERR_STANDALONE_VALUE, NULL);
- }
- if (RAW != '"') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_STARTED, NULL);
- }
- }
- return(standalone);
- }
- /**
- * xmlParseXMLDecl:
- * @ctxt: an XML parser context
- *
- * parse an XML declaration header
- *
- * [23] XMLDecl ::= '<?xml' VersionInfo EncodingDecl? SDDecl? S? '?>'
- */
- void
- xmlParseXMLDecl(xmlParserCtxtPtr ctxt) {
- xmlChar *version;
- /*
- * This value for standalone indicates that the document has an
- * XML declaration but it does not have a standalone attribute.
- * It will be overwritten later if a standalone attribute is found.
- */
- ctxt->input->standalone = -2;
- /*
- * We know that '<?xml' is here.
- */
- SKIP(5);
- if (!IS_BLANK_CH(RAW)) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Blank needed after '<?xml'\n");
- }
- SKIP_BLANKS;
- /*
- * We must have the VersionInfo here.
- */
- version = xmlParseVersionInfo(ctxt);
- if (version == NULL) {
- xmlFatalErr(ctxt, XML_ERR_VERSION_MISSING, NULL);
- } else {
- if (!xmlStrEqual(version, (const xmlChar *) XML_DEFAULT_VERSION)) {
- /*
- * Changed here for XML-1.0 5th edition
- */
- if (ctxt->options & XML_PARSE_OLD10) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNKNOWN_VERSION,
- "Unsupported version '%s'\n",
- version);
- } else {
- if ((version[0] == '1') && ((version[1] == '.'))) {
- xmlWarningMsg(ctxt, XML_WAR_UNKNOWN_VERSION,
- "Unsupported version '%s'\n",
- version, NULL);
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNKNOWN_VERSION,
- "Unsupported version '%s'\n",
- version);
- }
- }
- }
- if (ctxt->version != NULL)
- xmlFree((void *) ctxt->version);
- ctxt->version = version;
- }
- /*
- * We may have the encoding declaration
- */
- if (!IS_BLANK_CH(RAW)) {
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- return;
- }
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED, "Blank needed here\n");
- }
- xmlParseEncodingDecl(ctxt);
- if ((ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) ||
- (ctxt->instate == XML_PARSER_EOF)) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- return;
- }
- /*
- * We may have the standalone status.
- */
- if ((ctxt->input->encoding != NULL) && (!IS_BLANK_CH(RAW))) {
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- return;
- }
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED, "Blank needed here\n");
- }
- /*
- * We can grow the input buffer freely at that point
- */
- GROW;
- SKIP_BLANKS;
- ctxt->input->standalone = xmlParseSDDecl(ctxt);
- SKIP_BLANKS;
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- } else if (RAW == '>') {
- /* Deprecated old WD ... */
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- MOVETO_ENDTAG(CUR_PTR);
- NEXT;
- }
- }
- /**
- * xmlParseMisc:
- * @ctxt: an XML parser context
- *
- * parse an XML Misc* optional field.
- *
- * [27] Misc ::= Comment | PI | S
- */
- void
- xmlParseMisc(xmlParserCtxtPtr ctxt) {
- while (ctxt->instate != XML_PARSER_EOF) {
- SKIP_BLANKS;
- GROW;
- if ((RAW == '<') && (NXT(1) == '?')) {
- xmlParsePI(ctxt);
- } else if (CMP4(CUR_PTR, '<', '!', '-', '-')) {
- xmlParseComment(ctxt);
- } else {
- break;
- }
- }
- }
- /**
- * xmlParseDocument:
- * @ctxt: an XML parser context
- *
- * parse an XML document (and build a tree if using the standard SAX
- * interface).
- *
- * [1] document ::= prolog element Misc*
- *
- * [22] prolog ::= XMLDecl? Misc* (doctypedecl Misc*)?
- *
- * Returns 0, -1 in case of error. the parser context is augmented
- * as a result of the parsing.
- */
- int
- xmlParseDocument(xmlParserCtxtPtr ctxt) {
- xmlChar start[4];
- xmlCharEncoding enc;
- xmlInitParser();
- if ((ctxt == NULL) || (ctxt->input == NULL))
- return(-1);
- GROW;
- /*
- * SAX: detecting the level.
- */
- xmlDetectSAX2(ctxt);
- /*
- * SAX: beginning of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData, &xmlDefaultSAXLocator);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if ((ctxt->encoding == NULL) &&
- ((ctxt->input->end - ctxt->input->cur) >= 4)) {
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(&start[0], 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- if (CUR == 0) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- return(-1);
- }
- /*
- * Check for the XMLDecl in the Prolog.
- * do not GROW here to avoid the detected encoder to decode more
- * than just the first line, unless the amount of data is really
- * too small to hold "<?xml version="1.0" encoding="foo"
- */
- if ((ctxt->input->end - ctxt->input->cur) < 35) {
- GROW;
- }
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- /*
- * Note that we will switch encoding on the fly.
- */
- xmlParseXMLDecl(ctxt);
- if ((ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) ||
- (ctxt->instate == XML_PARSER_EOF)) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- return(-1);
- }
- ctxt->standalone = ctxt->input->standalone;
- SKIP_BLANKS;
- } else {
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- }
- if ((ctxt->sax) && (ctxt->sax->startDocument) && (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if ((ctxt->myDoc != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL) && (ctxt->input->buf->compressed >= 0)) {
- ctxt->myDoc->compression = ctxt->input->buf->compressed;
- }
- /*
- * The Misc part of the Prolog
- */
- xmlParseMisc(ctxt);
- /*
- * Then possibly doc type declaration(s) and more Misc
- * (doctypedecl Misc*)?
- */
- GROW;
- if (CMP9(CUR_PTR, '<', '!', 'D', 'O', 'C', 'T', 'Y', 'P', 'E')) {
- ctxt->inSubset = 1;
- xmlParseDocTypeDecl(ctxt);
- if (RAW == '[') {
- ctxt->instate = XML_PARSER_DTD;
- xmlParseInternalSubset(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- }
- /*
- * Create and update the external subset.
- */
- ctxt->inSubset = 2;
- if ((ctxt->sax != NULL) && (ctxt->sax->externalSubset != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->externalSubset(ctxt->userData, ctxt->intSubName,
- ctxt->extSubSystem, ctxt->extSubURI);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- ctxt->inSubset = 0;
- xmlCleanSpecialAttr(ctxt);
- ctxt->instate = XML_PARSER_PROLOG;
- xmlParseMisc(ctxt);
- }
- /*
- * Time to start parsing the tree itself
- */
- GROW;
- if (RAW != '<') {
- xmlFatalErrMsg(ctxt, XML_ERR_DOCUMENT_EMPTY,
- "Start tag expected, '<' not found\n");
- } else {
- ctxt->instate = XML_PARSER_CONTENT;
- xmlParseElement(ctxt);
- ctxt->instate = XML_PARSER_EPILOG;
- /*
- * The Misc part at the end
- */
- xmlParseMisc(ctxt);
- if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- }
- ctxt->instate = XML_PARSER_EOF;
- }
- /*
- * SAX: end of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- /*
- * Remove locally kept entity definitions if the tree was not built
- */
- if ((ctxt->myDoc != NULL) &&
- (xmlStrEqual(ctxt->myDoc->version, SAX_COMPAT_MODE))) {
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if ((ctxt->wellFormed) && (ctxt->myDoc != NULL)) {
- ctxt->myDoc->properties |= XML_DOC_WELLFORMED;
- if (ctxt->valid)
- ctxt->myDoc->properties |= XML_DOC_DTDVALID;
- if (ctxt->nsWellFormed)
- ctxt->myDoc->properties |= XML_DOC_NSVALID;
- if (ctxt->options & XML_PARSE_OLD10)
- ctxt->myDoc->properties |= XML_DOC_OLD10;
- }
- if (! ctxt->wellFormed) {
- ctxt->valid = 0;
- return(-1);
- }
- return(0);
- }
- /**
- * xmlParseExtParsedEnt:
- * @ctxt: an XML parser context
- *
- * parse a general parsed entity
- * An external general parsed entity is well-formed if it matches the
- * production labeled extParsedEnt.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * Returns 0, -1 in case of error. the parser context is augmented
- * as a result of the parsing.
- */
- int
- xmlParseExtParsedEnt(xmlParserCtxtPtr ctxt) {
- xmlChar start[4];
- xmlCharEncoding enc;
- if ((ctxt == NULL) || (ctxt->input == NULL))
- return(-1);
- xmlDefaultSAXHandlerInit();
- xmlDetectSAX2(ctxt);
- GROW;
- /*
- * SAX: beginning of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData, &xmlDefaultSAXLocator);
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- if ((ctxt->input->end - ctxt->input->cur) >= 4) {
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- if (CUR == 0) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- }
- /*
- * Check for the XMLDecl in the Prolog.
- */
- GROW;
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- /*
- * Note that we will switch encoding on the fly.
- */
- xmlParseXMLDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- return(-1);
- }
- SKIP_BLANKS;
- } else {
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- }
- if ((ctxt->sax) && (ctxt->sax->startDocument) && (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- /*
- * Doing validity checking on chunk doesn't make sense
- */
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->validate = 0;
- ctxt->loadsubset = 0;
- ctxt->depth = 0;
- xmlParseContent(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- /*
- * SAX: end of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- if (! ctxt->wellFormed) return(-1);
- return(0);
- }
- #ifdef LIBXML_PUSH_ENABLED
- /************************************************************************
- * *
- * Progressive parsing interfaces *
- * *
- ************************************************************************/
- /**
- * xmlParseLookupSequence:
- * @ctxt: an XML parser context
- * @first: the first char to lookup
- * @next: the next char to lookup or zero
- * @third: the next char to lookup or zero
- *
- * Try to find if a sequence (first, next, third) or just (first next) or
- * (first) is available in the input stream.
- * This function has a side effect of (possibly) incrementing ctxt->checkIndex
- * to avoid rescanning sequences of bytes, it DOES change the state of the
- * parser, do not use liberally.
- *
- * Returns the index to the current parsing point if the full sequence
- * is available, -1 otherwise.
- */
- static int
- xmlParseLookupSequence(xmlParserCtxtPtr ctxt, xmlChar first,
- xmlChar next, xmlChar third) {
- int base, len;
- xmlParserInputPtr in;
- const xmlChar *buf;
- in = ctxt->input;
- if (in == NULL) return(-1);
- base = in->cur - in->base;
- if (base < 0) return(-1);
- if (ctxt->checkIndex > base)
- base = ctxt->checkIndex;
- if (in->buf == NULL) {
- buf = in->base;
- len = in->length;
- } else {
- buf = xmlBufContent(in->buf->buffer);
- len = xmlBufUse(in->buf->buffer);
- }
- /* take into account the sequence length */
- if (third) len -= 2;
- else if (next) len --;
- for (;base < len;base++) {
- if (buf[base] == first) {
- if (third != 0) {
- if ((buf[base + 1] != next) ||
- (buf[base + 2] != third)) continue;
- } else if (next != 0) {
- if (buf[base + 1] != next) continue;
- }
- ctxt->checkIndex = 0;
- #ifdef DEBUG_PUSH
- if (next == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c' found at %d\n",
- first, base);
- else if (third == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c' found at %d\n",
- first, next, base);
- else
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c%c' found at %d\n",
- first, next, third, base);
- #endif
- return(base - (in->cur - in->base));
- }
- }
- ctxt->checkIndex = base;
- #ifdef DEBUG_PUSH
- if (next == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c' failed\n", first);
- else if (third == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c' failed\n", first, next);
- else
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c%c' failed\n", first, next, third);
- #endif
- return(-1);
- }
- /**
- * xmlParseGetLasts:
- * @ctxt: an XML parser context
- * @lastlt: pointer to store the last '<' from the input
- * @lastgt: pointer to store the last '>' from the input
- *
- * Lookup the last < and > in the current chunk
- */
- static void
- xmlParseGetLasts(xmlParserCtxtPtr ctxt, const xmlChar **lastlt,
- const xmlChar **lastgt) {
- const xmlChar *tmp;
- if ((ctxt == NULL) || (lastlt == NULL) || (lastgt == NULL)) {
- xmlGenericError(xmlGenericErrorContext,
- "Internal error: xmlParseGetLasts\n");
- return;
- }
- if ((ctxt->progressive != 0) && (ctxt->inputNr == 1)) {
- tmp = ctxt->input->end;
- tmp--;
- while ((tmp >= ctxt->input->base) && (*tmp != '<')) tmp--;
- if (tmp < ctxt->input->base) {
- *lastlt = NULL;
- *lastgt = NULL;
- } else {
- *lastlt = tmp;
- tmp++;
- while ((tmp < ctxt->input->end) && (*tmp != '>')) {
- if (*tmp == '\'') {
- tmp++;
- while ((tmp < ctxt->input->end) && (*tmp != '\'')) tmp++;
- if (tmp < ctxt->input->end) tmp++;
- } else if (*tmp == '"') {
- tmp++;
- while ((tmp < ctxt->input->end) && (*tmp != '"')) tmp++;
- if (tmp < ctxt->input->end) tmp++;
- } else
- tmp++;
- }
- if (tmp < ctxt->input->end)
- *lastgt = tmp;
- else {
- tmp = *lastlt;
- tmp--;
- while ((tmp >= ctxt->input->base) && (*tmp != '>')) tmp--;
- if (tmp >= ctxt->input->base)
- *lastgt = tmp;
- else
- *lastgt = NULL;
- }
- }
- } else {
- *lastlt = NULL;
- *lastgt = NULL;
- }
- }
- /**
- * xmlCheckCdataPush:
- * @cur: pointer to the block of characters
- * @len: length of the block in bytes
- * @complete: 1 if complete CDATA block is passed in, 0 if partial block
- *
- * Check that the block of characters is okay as SCdata content [20]
- *
- * Returns the number of bytes to pass if okay, a negative index where an
- * UTF-8 error occurred otherwise
- */
- static int
- xmlCheckCdataPush(const xmlChar *utf, int len, int complete) {
- int ix;
- unsigned char c;
- int codepoint;
- if ((utf == NULL) || (len <= 0))
- return(0);
- for (ix = 0; ix < len;) { /* string is 0-terminated */
- c = utf[ix];
- if ((c & 0x80) == 0x00) { /* 1-byte code, starts with 10 */
- if (c >= 0x20)
- ix++;
- else if ((c == 0xA) || (c == 0xD) || (c == 0x9))
- ix++;
- else
- return(-ix);
- } else if ((c & 0xe0) == 0xc0) {/* 2-byte code, starts with 110 */
- if (ix + 2 > len) return(complete ? -ix : ix);
- if ((utf[ix+1] & 0xc0 ) != 0x80)
- return(-ix);
- codepoint = (utf[ix] & 0x1f) << 6;
- codepoint |= utf[ix+1] & 0x3f;
- if (!xmlIsCharQ(codepoint))
- return(-ix);
- ix += 2;
- } else if ((c & 0xf0) == 0xe0) {/* 3-byte code, starts with 1110 */
- if (ix + 3 > len) return(complete ? -ix : ix);
- if (((utf[ix+1] & 0xc0) != 0x80) ||
- ((utf[ix+2] & 0xc0) != 0x80))
- return(-ix);
- codepoint = (utf[ix] & 0xf) << 12;
- codepoint |= (utf[ix+1] & 0x3f) << 6;
- codepoint |= utf[ix+2] & 0x3f;
- if (!xmlIsCharQ(codepoint))
- return(-ix);
- ix += 3;
- } else if ((c & 0xf8) == 0xf0) {/* 4-byte code, starts with 11110 */
- if (ix + 4 > len) return(complete ? -ix : ix);
- if (((utf[ix+1] & 0xc0) != 0x80) ||
- ((utf[ix+2] & 0xc0) != 0x80) ||
- ((utf[ix+3] & 0xc0) != 0x80))
- return(-ix);
- codepoint = (utf[ix] & 0x7) << 18;
- codepoint |= (utf[ix+1] & 0x3f) << 12;
- codepoint |= (utf[ix+2] & 0x3f) << 6;
- codepoint |= utf[ix+3] & 0x3f;
- if (!xmlIsCharQ(codepoint))
- return(-ix);
- ix += 4;
- } else /* unknown encoding */
- return(-ix);
- }
- return(ix);
- }
- /**
- * xmlParseTryOrFinish:
- * @ctxt: an XML parser context
- * @terminate: last chunk indicator
- *
- * Try to progress on parsing
- *
- * Returns zero if no parsing was possible
- */
- static int
- xmlParseTryOrFinish(xmlParserCtxtPtr ctxt, int terminate) {
- int ret = 0;
- int avail, tlen;
- xmlChar cur, next;
- const xmlChar *lastlt, *lastgt;
- if (ctxt->input == NULL)
- return(0);
- #ifdef DEBUG_PUSH
- switch (ctxt->instate) {
- case XML_PARSER_EOF:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try EOF\n"); break;
- case XML_PARSER_START:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try START\n"); break;
- case XML_PARSER_MISC:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try MISC\n");break;
- case XML_PARSER_COMMENT:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try COMMENT\n");break;
- case XML_PARSER_PROLOG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try PROLOG\n");break;
- case XML_PARSER_START_TAG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try START_TAG\n");break;
- case XML_PARSER_CONTENT:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try CONTENT\n");break;
- case XML_PARSER_CDATA_SECTION:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try CDATA_SECTION\n");break;
- case XML_PARSER_END_TAG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try END_TAG\n");break;
- case XML_PARSER_ENTITY_DECL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try ENTITY_DECL\n");break;
- case XML_PARSER_ENTITY_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try ENTITY_VALUE\n");break;
- case XML_PARSER_ATTRIBUTE_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try ATTRIBUTE_VALUE\n");break;
- case XML_PARSER_DTD:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try DTD\n");break;
- case XML_PARSER_EPILOG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try EPILOG\n");break;
- case XML_PARSER_PI:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try PI\n");break;
- case XML_PARSER_IGNORE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try IGNORE\n");break;
- }
- #endif
- if ((ctxt->input != NULL) &&
- (ctxt->input->cur - ctxt->input->base > 4096)) {
- xmlSHRINK(ctxt);
- ctxt->checkIndex = 0;
- }
- xmlParseGetLasts(ctxt, &lastlt, &lastgt);
- while (ctxt->instate != XML_PARSER_EOF) {
- if ((ctxt->errNo != XML_ERR_OK) && (ctxt->disableSAX == 1))
- return(0);
- if (ctxt->input == NULL) break;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length -
- (ctxt->input->cur - ctxt->input->base);
- else {
- /*
- * If we are operating on converted input, try to flush
- * remaining chars to avoid them stalling in the non-converted
- * buffer. But do not do this in document start where
- * encoding="..." may not have been read and we work on a
- * guessed encoding.
- */
- if ((ctxt->instate != XML_PARSER_START) &&
- (ctxt->input->buf->raw != NULL) &&
- (xmlBufIsEmpty(ctxt->input->buf->raw) == 0)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer,
- ctxt->input);
- size_t current = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, 0, "");
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input,
- base, current);
- }
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- }
- if (avail < 1)
- goto done;
- switch (ctxt->instate) {
- case XML_PARSER_EOF:
- /*
- * Document parsing is done !
- */
- goto done;
- case XML_PARSER_START:
- if (ctxt->charset == XML_CHAR_ENCODING_NONE) {
- xmlChar start[4];
- xmlCharEncoding enc;
- /*
- * Very first chars read from the document flow.
- */
- if (avail < 4)
- goto done;
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines,
- * else xmlSwitchEncoding will set to (default)
- * UTF8.
- */
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- xmlSwitchEncoding(ctxt, enc);
- break;
- }
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if (cur == 0) {
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData,
- &xmlDefaultSAXLocator);
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- xmlHaltParser(ctxt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering EOF\n");
- #endif
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- if ((cur == '<') && (next == '?')) {
- /* PI or XML decl */
- if (avail < 5) return(ret);
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0))
- return(ret);
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData,
- &xmlDefaultSAXLocator);
- if ((ctxt->input->cur[2] == 'x') &&
- (ctxt->input->cur[3] == 'm') &&
- (ctxt->input->cur[4] == 'l') &&
- (IS_BLANK_CH(ctxt->input->cur[5]))) {
- ret += 5;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing XML Decl\n");
- #endif
- xmlParseXMLDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right
- * here
- */
- xmlHaltParser(ctxt);
- return(0);
- }
- ctxt->standalone = ctxt->input->standalone;
- if ((ctxt->encoding == NULL) &&
- (ctxt->input->encoding != NULL))
- ctxt->encoding = xmlStrdup(ctxt->input->encoding);
- if ((ctxt->sax) && (ctxt->sax->startDocument) &&
- (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- ctxt->instate = XML_PARSER_MISC;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering MISC\n");
- #endif
- } else {
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- if ((ctxt->sax) && (ctxt->sax->startDocument) &&
- (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- ctxt->instate = XML_PARSER_MISC;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering MISC\n");
- #endif
- }
- } else {
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData,
- &xmlDefaultSAXLocator);
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- if (ctxt->version == NULL) {
- xmlErrMemory(ctxt, NULL);
- break;
- }
- if ((ctxt->sax) && (ctxt->sax->startDocument) &&
- (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- ctxt->instate = XML_PARSER_MISC;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering MISC\n");
- #endif
- }
- break;
- case XML_PARSER_START_TAG: {
- const xmlChar *name;
- const xmlChar *prefix = NULL;
- const xmlChar *URI = NULL;
- int line = ctxt->input->line;
- int nsNr = ctxt->nsNr;
- if ((avail < 2) && (ctxt->inputNr == 1))
- goto done;
- cur = ctxt->input->cur[0];
- if (cur != '<') {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- xmlHaltParser(ctxt);
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- if (!terminate) {
- if (ctxt->progressive) {
- /* > can be found unescaped in attribute values */
- if ((lastgt == NULL) || (ctxt->input->cur >= lastgt))
- goto done;
- } else if (xmlParseLookupSequence(ctxt, '>', 0, 0) < 0) {
- goto done;
- }
- }
- if (ctxt->spaceNr == 0)
- spacePush(ctxt, -1);
- else if (*ctxt->space == -2)
- spacePush(ctxt, -1);
- else
- spacePush(ctxt, *ctxt->space);
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax2)
- #endif /* LIBXML_SAX1_ENABLED */
- name = xmlParseStartTag2(ctxt, &prefix, &URI, &tlen);
- #ifdef LIBXML_SAX1_ENABLED
- else
- name = xmlParseStartTag(ctxt);
- #endif /* LIBXML_SAX1_ENABLED */
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- if (name == NULL) {
- spacePop(ctxt);
- xmlHaltParser(ctxt);
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- #ifdef LIBXML_VALID_ENABLED
- /*
- * [ VC: Root Element Type ]
- * The Name in the document type declaration must match
- * the element type of the root element.
- */
- if (ctxt->validate && ctxt->wellFormed && ctxt->myDoc &&
- ctxt->node && (ctxt->node == ctxt->myDoc->children))
- ctxt->valid &= xmlValidateRoot(&ctxt->vctxt, ctxt->myDoc);
- #endif /* LIBXML_VALID_ENABLED */
- /*
- * Check for an Empty Element.
- */
- if ((RAW == '/') && (NXT(1) == '>')) {
- SKIP(2);
- if (ctxt->sax2) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->endElementNs != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElementNs(ctxt->userData, name,
- prefix, URI);
- if (ctxt->nsNr - nsNr > 0)
- nsPop(ctxt, ctxt->nsNr - nsNr);
- #ifdef LIBXML_SAX1_ENABLED
- } else {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->endElement != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElement(ctxt->userData, name);
- #endif /* LIBXML_SAX1_ENABLED */
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- spacePop(ctxt);
- if (ctxt->nameNr == 0) {
- ctxt->instate = XML_PARSER_EPILOG;
- } else {
- ctxt->instate = XML_PARSER_CONTENT;
- }
- ctxt->progressive = 1;
- break;
- }
- if (RAW == '>') {
- NEXT;
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_GT_REQUIRED,
- "Couldn't find end of Start Tag %s\n",
- name);
- nodePop(ctxt);
- spacePop(ctxt);
- }
- nameNsPush(ctxt, name, prefix, URI, line, ctxt->nsNr - nsNr);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->progressive = 1;
- break;
- }
- case XML_PARSER_CONTENT: {
- const xmlChar *test;
- unsigned int cons;
- if ((avail < 2) && (ctxt->inputNr == 1))
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- test = CUR_PTR;
- cons = ctxt->input->consumed;
- if ((cur == '<') && (next == '/')) {
- ctxt->instate = XML_PARSER_END_TAG;
- break;
- } else if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- xmlParsePI(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next != '!')) {
- ctxt->instate = XML_PARSER_START_TAG;
- break;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') &&
- (ctxt->input->cur[3] == '-')) {
- int term;
- if (avail < 4)
- goto done;
- ctxt->input->cur += 4;
- term = xmlParseLookupSequence(ctxt, '-', '-', '>');
- ctxt->input->cur -= 4;
- if ((!terminate) && (term < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- xmlParseComment(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (ctxt->input->cur[1] == '!') &&
- (ctxt->input->cur[2] == '[') &&
- (ctxt->input->cur[3] == 'C') &&
- (ctxt->input->cur[4] == 'D') &&
- (ctxt->input->cur[5] == 'A') &&
- (ctxt->input->cur[6] == 'T') &&
- (ctxt->input->cur[7] == 'A') &&
- (ctxt->input->cur[8] == '[')) {
- SKIP(9);
- ctxt->instate = XML_PARSER_CDATA_SECTION;
- break;
- } else if ((cur == '<') && (next == '!') &&
- (avail < 9)) {
- goto done;
- } else if (cur == '&') {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, ';', 0, 0) < 0))
- goto done;
- xmlParseReference(ctxt);
- } else {
- /* TODO Avoid the extra copy, handle directly !!! */
- /*
- * Goal of the following test is:
- * - minimize calls to the SAX 'character' callback
- * when they are mergeable
- * - handle an problem for isBlank when we only parse
- * a sequence of blank chars and the next one is
- * not available to check against '<' presence.
- * - tries to homogenize the differences in SAX
- * callbacks between the push and pull versions
- * of the parser.
- */
- if ((ctxt->inputNr == 1) &&
- (avail < XML_PARSER_BIG_BUFFER_SIZE)) {
- if (!terminate) {
- if (ctxt->progressive) {
- if ((lastlt == NULL) ||
- (ctxt->input->cur > lastlt))
- goto done;
- } else if (xmlParseLookupSequence(ctxt,
- '<', 0, 0) < 0) {
- goto done;
- }
- }
- }
- ctxt->checkIndex = 0;
- xmlParseCharData(ctxt, 0);
- }
- if ((cons == ctxt->input->consumed) && (test == CUR_PTR)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "detected an error in element content\n");
- xmlHaltParser(ctxt);
- break;
- }
- break;
- }
- case XML_PARSER_END_TAG:
- if (avail < 2)
- goto done;
- if (!terminate) {
- if (ctxt->progressive) {
- /* > can be found unescaped in attribute values */
- if ((lastgt == NULL) || (ctxt->input->cur >= lastgt))
- goto done;
- } else if (xmlParseLookupSequence(ctxt, '>', 0, 0) < 0) {
- goto done;
- }
- }
- if (ctxt->sax2) {
- xmlParseEndTag2(ctxt, &ctxt->pushTab[ctxt->nameNr - 1]);
- nameNsPop(ctxt);
- }
- #ifdef LIBXML_SAX1_ENABLED
- else
- xmlParseEndTag1(ctxt, 0);
- #endif /* LIBXML_SAX1_ENABLED */
- if (ctxt->instate == XML_PARSER_EOF) {
- /* Nothing */
- } else if (ctxt->nameNr == 0) {
- ctxt->instate = XML_PARSER_EPILOG;
- } else {
- ctxt->instate = XML_PARSER_CONTENT;
- }
- break;
- case XML_PARSER_CDATA_SECTION: {
- /*
- * The Push mode need to have the SAX callback for
- * cdataBlock merge back contiguous callbacks.
- */
- int base;
- base = xmlParseLookupSequence(ctxt, ']', ']', '>');
- if (base < 0) {
- if (avail >= XML_PARSER_BIG_BUFFER_SIZE + 2) {
- int tmp;
- tmp = xmlCheckCdataPush(ctxt->input->cur,
- XML_PARSER_BIG_BUFFER_SIZE, 0);
- if (tmp < 0) {
- tmp = -tmp;
- ctxt->input->cur += tmp;
- goto encoding_error;
- }
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (ctxt->sax->cdataBlock != NULL)
- ctxt->sax->cdataBlock(ctxt->userData,
- ctxt->input->cur, tmp);
- else if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- ctxt->input->cur, tmp);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- SKIPL(tmp);
- ctxt->checkIndex = 0;
- }
- goto done;
- } else {
- int tmp;
- tmp = xmlCheckCdataPush(ctxt->input->cur, base, 1);
- if ((tmp < 0) || (tmp != base)) {
- tmp = -tmp;
- ctxt->input->cur += tmp;
- goto encoding_error;
- }
- if ((ctxt->sax != NULL) && (base == 0) &&
- (ctxt->sax->cdataBlock != NULL) &&
- (!ctxt->disableSAX)) {
- /*
- * Special case to provide identical behaviour
- * between pull and push parsers on enpty CDATA
- * sections
- */
- if ((ctxt->input->cur - ctxt->input->base >= 9) &&
- (!strncmp((const char *)&ctxt->input->cur[-9],
- "<![CDATA[", 9)))
- ctxt->sax->cdataBlock(ctxt->userData,
- BAD_CAST "", 0);
- } else if ((ctxt->sax != NULL) && (base > 0) &&
- (!ctxt->disableSAX)) {
- if (ctxt->sax->cdataBlock != NULL)
- ctxt->sax->cdataBlock(ctxt->userData,
- ctxt->input->cur, base);
- else if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- ctxt->input->cur, base);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- SKIPL(base + 3);
- ctxt->checkIndex = 0;
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering CONTENT\n");
- #endif
- }
- break;
- }
- case XML_PARSER_MISC:
- SKIP_BLANKS;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length -
- (ctxt->input->cur - ctxt->input->base);
- else
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing PI\n");
- #endif
- xmlParsePI(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_MISC;
- ctxt->progressive = 1;
- ctxt->checkIndex = 0;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') &&
- (ctxt->input->cur[3] == '-')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '-', '-', '>') < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing Comment\n");
- #endif
- xmlParseComment(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_MISC;
- ctxt->progressive = 1;
- ctxt->checkIndex = 0;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == 'D') &&
- (ctxt->input->cur[3] == 'O') &&
- (ctxt->input->cur[4] == 'C') &&
- (ctxt->input->cur[5] == 'T') &&
- (ctxt->input->cur[6] == 'Y') &&
- (ctxt->input->cur[7] == 'P') &&
- (ctxt->input->cur[8] == 'E')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '>', 0, 0) < 0)) {
- ctxt->progressive = XML_PARSER_DTD;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing internal subset\n");
- #endif
- ctxt->inSubset = 1;
- ctxt->progressive = 0;
- ctxt->checkIndex = 0;
- xmlParseDocTypeDecl(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- if (RAW == '[') {
- ctxt->instate = XML_PARSER_DTD;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- } else {
- /*
- * Create and update the external subset.
- */
- ctxt->inSubset = 2;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->externalSubset != NULL))
- ctxt->sax->externalSubset(ctxt->userData,
- ctxt->intSubName, ctxt->extSubSystem,
- ctxt->extSubURI);
- ctxt->inSubset = 0;
- xmlCleanSpecialAttr(ctxt);
- ctxt->instate = XML_PARSER_PROLOG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering PROLOG\n");
- #endif
- }
- } else if ((cur == '<') && (next == '!') &&
- (avail < 9)) {
- goto done;
- } else {
- ctxt->instate = XML_PARSER_START_TAG;
- ctxt->progressive = XML_PARSER_START_TAG;
- xmlParseGetLasts(ctxt, &lastlt, &lastgt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- }
- break;
- case XML_PARSER_PROLOG:
- SKIP_BLANKS;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length - (ctxt->input->cur - ctxt->input->base);
- else
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing PI\n");
- #endif
- xmlParsePI(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_PROLOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') && (ctxt->input->cur[3] == '-')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '-', '-', '>') < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing Comment\n");
- #endif
- xmlParseComment(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_PROLOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (avail < 4)) {
- goto done;
- } else {
- ctxt->instate = XML_PARSER_START_TAG;
- if (ctxt->progressive == 0)
- ctxt->progressive = XML_PARSER_START_TAG;
- xmlParseGetLasts(ctxt, &lastlt, &lastgt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- }
- break;
- case XML_PARSER_EPILOG:
- SKIP_BLANKS;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length - (ctxt->input->cur - ctxt->input->base);
- else
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing PI\n");
- #endif
- xmlParsePI(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_EPILOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') && (ctxt->input->cur[3] == '-')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '-', '-', '>') < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing Comment\n");
- #endif
- xmlParseComment(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_EPILOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (avail < 4)) {
- goto done;
- } else {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- xmlHaltParser(ctxt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering EOF\n");
- #endif
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- break;
- case XML_PARSER_DTD: {
- /*
- * Sorry but progressive parsing of the internal subset
- * is not expected to be supported. We first check that
- * the full content of the internal subset is available and
- * the parsing is launched only at that point.
- * Internal subset ends up with "']' S? '>'" in an unescaped
- * section and not in a ']]>' sequence which are conditional
- * sections (whoever argued to keep that crap in XML deserve
- * a place in hell !).
- */
- int base, i;
- xmlChar *buf;
- xmlChar quote = 0;
- size_t use;
- base = ctxt->input->cur - ctxt->input->base;
- if (base < 0) return(0);
- if (ctxt->checkIndex > base)
- base = ctxt->checkIndex;
- buf = xmlBufContent(ctxt->input->buf->buffer);
- use = xmlBufUse(ctxt->input->buf->buffer);
- for (;(unsigned int) base < use; base++) {
- if (quote != 0) {
- if (buf[base] == quote)
- quote = 0;
- continue;
- }
- if ((quote == 0) && (buf[base] == '<')) {
- int found = 0;
- /* special handling of comments */
- if (((unsigned int) base + 4 < use) &&
- (buf[base + 1] == '!') &&
- (buf[base + 2] == '-') &&
- (buf[base + 3] == '-')) {
- for (;(unsigned int) base + 3 < use; base++) {
- if ((buf[base] == '-') &&
- (buf[base + 1] == '-') &&
- (buf[base + 2] == '>')) {
- found = 1;
- base += 2;
- break;
- }
- }
- if (!found) {
- #if 0
- fprintf(stderr, "unfinished comment\n");
- #endif
- break; /* for */
- }
- continue;
- }
- }
- if (buf[base] == '"') {
- quote = '"';
- continue;
- }
- if (buf[base] == '\'') {
- quote = '\'';
- continue;
- }
- if (buf[base] == ']') {
- #if 0
- fprintf(stderr, "%c%c%c%c: ", buf[base],
- buf[base + 1], buf[base + 2], buf[base + 3]);
- #endif
- if ((unsigned int) base +1 >= use)
- break;
- if (buf[base + 1] == ']') {
- /* conditional crap, skip both ']' ! */
- base++;
- continue;
- }
- for (i = 1; (unsigned int) base + i < use; i++) {
- if (buf[base + i] == '>') {
- #if 0
- fprintf(stderr, "found\n");
- #endif
- goto found_end_int_subset;
- }
- if (!IS_BLANK_CH(buf[base + i])) {
- #if 0
- fprintf(stderr, "not found\n");
- #endif
- goto not_end_of_int_subset;
- }
- }
- #if 0
- fprintf(stderr, "end of stream\n");
- #endif
- break;
- }
- not_end_of_int_subset:
- continue; /* for */
- }
- /*
- * We didn't found the end of the Internal subset
- */
- if (quote == 0)
- ctxt->checkIndex = base;
- else
- ctxt->checkIndex = 0;
- #ifdef DEBUG_PUSH
- if (next == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup of int subset end filed\n");
- #endif
- goto done;
- found_end_int_subset:
- ctxt->checkIndex = 0;
- xmlParseInternalSubset(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->inSubset = 2;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->externalSubset != NULL))
- ctxt->sax->externalSubset(ctxt->userData, ctxt->intSubName,
- ctxt->extSubSystem, ctxt->extSubURI);
- ctxt->inSubset = 0;
- xmlCleanSpecialAttr(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_PROLOG;
- ctxt->checkIndex = 0;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering PROLOG\n");
- #endif
- break;
- }
- case XML_PARSER_COMMENT:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == COMMENT\n");
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering CONTENT\n");
- #endif
- break;
- case XML_PARSER_IGNORE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == IGNORE");
- ctxt->instate = XML_PARSER_DTD;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- break;
- case XML_PARSER_PI:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == PI\n");
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering CONTENT\n");
- #endif
- break;
- case XML_PARSER_ENTITY_DECL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == ENTITY_DECL\n");
- ctxt->instate = XML_PARSER_DTD;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- break;
- case XML_PARSER_ENTITY_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == ENTITY_VALUE\n");
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- break;
- case XML_PARSER_ATTRIBUTE_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == ATTRIBUTE_VALUE\n");
- ctxt->instate = XML_PARSER_START_TAG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- break;
- case XML_PARSER_SYSTEM_LITERAL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == SYSTEM_LITERAL\n");
- ctxt->instate = XML_PARSER_START_TAG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- break;
- case XML_PARSER_PUBLIC_LITERAL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == PUBLIC_LITERAL\n");
- ctxt->instate = XML_PARSER_START_TAG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- break;
- }
- }
- done:
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: done %d\n", ret);
- #endif
- return(ret);
- encoding_error:
- {
- char buffer[150];
- snprintf(buffer, 149, "Bytes: 0x%02X 0x%02X 0x%02X 0x%02X\n",
- ctxt->input->cur[0], ctxt->input->cur[1],
- ctxt->input->cur[2], ctxt->input->cur[3]);
- __xmlErrEncoding(ctxt, XML_ERR_INVALID_CHAR,
- "Input is not proper UTF-8, indicate encoding !\n%s",
- BAD_CAST buffer, NULL);
- }
- return(0);
- }
- /**
- * xmlParseCheckTransition:
- * @ctxt: an XML parser context
- * @chunk: a char array
- * @size: the size in byte of the chunk
- *
- * Check depending on the current parser state if the chunk given must be
- * processed immediately or one need more data to advance on parsing.
- *
- * Returns -1 in case of error, 0 if the push is not needed and 1 if needed
- */
- static int
- xmlParseCheckTransition(xmlParserCtxtPtr ctxt, const char *chunk, int size) {
- if ((ctxt == NULL) || (chunk == NULL) || (size < 0))
- return(-1);
- if (ctxt->instate == XML_PARSER_START_TAG) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->progressive == XML_PARSER_COMMENT) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->instate == XML_PARSER_CDATA_SECTION) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->progressive == XML_PARSER_PI) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->instate == XML_PARSER_END_TAG) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if ((ctxt->progressive == XML_PARSER_DTD) ||
- (ctxt->instate == XML_PARSER_DTD)) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- return(1);
- }
- /**
- * xmlParseChunk:
- * @ctxt: an XML parser context
- * @chunk: an char array
- * @size: the size in byte of the chunk
- * @terminate: last chunk indicator
- *
- * Parse a Chunk of memory
- *
- * Returns zero if no error, the xmlParserErrors otherwise.
- */
- int
- xmlParseChunk(xmlParserCtxtPtr ctxt, const char *chunk, int size,
- int terminate) {
- int end_in_lf = 0;
- int remain = 0;
- size_t old_avail = 0;
- size_t avail = 0;
- if (ctxt == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- if ((ctxt->errNo != XML_ERR_OK) && (ctxt->disableSAX == 1))
- return(ctxt->errNo);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if (ctxt->instate == XML_PARSER_START)
- xmlDetectSAX2(ctxt);
- if ((size > 0) && (chunk != NULL) && (!terminate) &&
- (chunk[size - 1] == '\r')) {
- end_in_lf = 1;
- size--;
- }
- xmldecl_done:
- if ((size > 0) && (chunk != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL) && (ctxt->instate != XML_PARSER_EOF)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer, ctxt->input);
- size_t cur = ctxt->input->cur - ctxt->input->base;
- int res;
- old_avail = xmlBufUse(ctxt->input->buf->buffer);
- /*
- * Specific handling if we autodetected an encoding, we should not
- * push more than the first line ... which depend on the encoding
- * And only push the rest once the final encoding was detected
- */
- if ((ctxt->instate == XML_PARSER_START) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL) && (ctxt->input->buf->encoder != NULL)) {
- unsigned int len = 45;
- if ((xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UTF-16")) ||
- (xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UTF16")))
- len = 90;
- else if ((xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UCS-4")) ||
- (xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UCS4")))
- len = 180;
- if (ctxt->input->buf->rawconsumed < len)
- len -= ctxt->input->buf->rawconsumed;
- /*
- * Change size for reading the initial declaration only
- * if size is greater than len. Otherwise, memmove in xmlBufferAdd
- * will blindly copy extra bytes from memory.
- */
- if ((unsigned int) size > len) {
- remain = size - len;
- size = len;
- } else {
- remain = 0;
- }
- }
- res = xmlParserInputBufferPush(ctxt->input->buf, size, chunk);
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input, base, cur);
- if (res < 0) {
- ctxt->errNo = XML_PARSER_EOF;
- xmlHaltParser(ctxt);
- return (XML_PARSER_EOF);
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: pushed %d\n", size);
- #endif
- } else if (ctxt->instate != XML_PARSER_EOF) {
- if ((ctxt->input != NULL) && ctxt->input->buf != NULL) {
- xmlParserInputBufferPtr in = ctxt->input->buf;
- if ((in->encoder != NULL) && (in->buffer != NULL) &&
- (in->raw != NULL)) {
- int nbchars;
- size_t base = xmlBufGetInputBase(in->buffer, ctxt->input);
- size_t current = ctxt->input->cur - ctxt->input->base;
- nbchars = xmlCharEncInput(in, terminate);
- xmlBufSetInputBaseCur(in->buffer, ctxt->input, base, current);
- if (nbchars < 0) {
- /* TODO 2.6.0 */
- xmlGenericError(xmlGenericErrorContext,
- "xmlParseChunk: encoder error\n");
- xmlHaltParser(ctxt);
- return(XML_ERR_INVALID_ENCODING);
- }
- }
- }
- }
- if (remain != 0) {
- xmlParseTryOrFinish(ctxt, 0);
- } else {
- if ((ctxt->input != NULL) && (ctxt->input->buf != NULL))
- avail = xmlBufUse(ctxt->input->buf->buffer);
- /*
- * Depending on the current state it may not be such
- * a good idea to try parsing if there is nothing in the chunk
- * which would be worth doing a parser state transition and we
- * need to wait for more data
- */
- if ((terminate) || (avail > XML_MAX_TEXT_LENGTH) ||
- (old_avail == 0) || (avail == 0) ||
- (xmlParseCheckTransition(ctxt,
- (const char *)&ctxt->input->base[old_avail],
- avail - old_avail)))
- xmlParseTryOrFinish(ctxt, terminate);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- return(ctxt->errNo);
- if ((ctxt->input != NULL) &&
- (((ctxt->input->end - ctxt->input->cur) > XML_MAX_LOOKUP_LIMIT) ||
- ((ctxt->input->cur - ctxt->input->base) > XML_MAX_LOOKUP_LIMIT)) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR, "Huge input lookup");
- xmlHaltParser(ctxt);
- }
- if ((ctxt->errNo != XML_ERR_OK) && (ctxt->disableSAX == 1))
- return(ctxt->errNo);
- if (remain != 0) {
- chunk += size;
- size = remain;
- remain = 0;
- goto xmldecl_done;
- }
- if ((end_in_lf == 1) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer,
- ctxt->input);
- size_t current = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, 1, "\r");
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input,
- base, current);
- }
- if (terminate) {
- /*
- * Check for termination
- */
- int cur_avail = 0;
- if (ctxt->input != NULL) {
- if (ctxt->input->buf == NULL)
- cur_avail = ctxt->input->length -
- (ctxt->input->cur - ctxt->input->base);
- else
- cur_avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- }
- if ((ctxt->instate != XML_PARSER_EOF) &&
- (ctxt->instate != XML_PARSER_EPILOG)) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- }
- if ((ctxt->instate == XML_PARSER_EPILOG) && (cur_avail > 0)) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- }
- if (ctxt->instate != XML_PARSER_EOF) {
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- }
- ctxt->instate = XML_PARSER_EOF;
- }
- if (ctxt->wellFormed == 0)
- return((xmlParserErrors) ctxt->errNo);
- else
- return(0);
- }
- /************************************************************************
- * *
- * I/O front end functions to the parser *
- * *
- ************************************************************************/
- /**
- * xmlCreatePushParserCtxt:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @chunk: a pointer to an array of chars
- * @size: number of chars in the array
- * @filename: an optional file name or URI
- *
- * Create a parser context for using the XML parser in push mode.
- * If @buffer and @size are non-NULL, the data is used to detect
- * the encoding. The remaining characters will be parsed so they
- * don't need to be fed in again through xmlParseChunk.
- * To allow content encoding detection, @size should be >= 4
- * The value of @filename is used for fetching external entities
- * and error/warning reports.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreatePushParserCtxt(xmlSAXHandlerPtr sax, void *user_data,
- const char *chunk, int size, const char *filename) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- xmlParserInputBufferPtr buf;
- xmlCharEncoding enc = XML_CHAR_ENCODING_NONE;
- /*
- * plug some encoding conversion routines
- */
- if ((chunk != NULL) && (size >= 4))
- enc = xmlDetectCharEncoding((const xmlChar *) chunk, size);
- buf = xmlAllocParserInputBuffer(enc);
- if (buf == NULL) return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlErrMemory(NULL, "creating parser: out of memory\n");
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- ctxt->dictNames = 1;
- if (sax != NULL) {
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- #endif /* LIBXML_SAX1_ENABLED */
- xmlFree(ctxt->sax);
- ctxt->sax = (xmlSAXHandlerPtr) xmlMalloc(sizeof(xmlSAXHandler));
- if (ctxt->sax == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFreeParserInputBuffer(buf);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- memset(ctxt->sax, 0, sizeof(xmlSAXHandler));
- if (sax->initialized == XML_SAX2_MAGIC)
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandler));
- else
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandlerV1));
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- if (filename == NULL) {
- ctxt->directory = NULL;
- } else {
- ctxt->directory = xmlParserGetDirectory(filename);
- }
- inputStream = xmlNewInputStream(ctxt);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- if (filename == NULL)
- inputStream->filename = NULL;
- else {
- inputStream->filename = (char *)
- xmlCanonicPath((const xmlChar *) filename);
- if (inputStream->filename == NULL) {
- xmlFreeParserCtxt(ctxt);
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- }
- inputStream->buf = buf;
- xmlBufResetInput(inputStream->buf->buffer, inputStream);
- inputPush(ctxt, inputStream);
- /*
- * If the caller didn't provide an initial 'chunk' for determining
- * the encoding, we set the context to XML_CHAR_ENCODING_NONE so
- * that it can be automatically determined later
- */
- if ((size == 0) || (chunk == NULL)) {
- ctxt->charset = XML_CHAR_ENCODING_NONE;
- } else if ((ctxt->input != NULL) && (ctxt->input->buf != NULL)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer, ctxt->input);
- size_t cur = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, size, chunk);
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input, base, cur);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: pushed %d\n", size);
- #endif
- }
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- return(ctxt);
- }
- #endif /* LIBXML_PUSH_ENABLED */
- /**
- * xmlHaltParser:
- * @ctxt: an XML parser context
- *
- * Blocks further parser processing don't override error
- * for internal use
- */
- static void
- xmlHaltParser(xmlParserCtxtPtr ctxt) {
- if (ctxt == NULL)
- return;
- ctxt->instate = XML_PARSER_EOF;
- ctxt->disableSAX = 1;
- while (ctxt->inputNr > 1)
- xmlFreeInputStream(inputPop(ctxt));
- if (ctxt->input != NULL) {
- /*
- * in case there was a specific allocation deallocate before
- * overriding base
- */
- if (ctxt->input->free != NULL) {
- ctxt->input->free((xmlChar *) ctxt->input->base);
- ctxt->input->free = NULL;
- }
- if (ctxt->input->buf != NULL) {
- xmlFreeParserInputBuffer(ctxt->input->buf);
- ctxt->input->buf = NULL;
- }
- ctxt->input->cur = BAD_CAST"";
- ctxt->input->length = 0;
- ctxt->input->base = ctxt->input->cur;
- ctxt->input->end = ctxt->input->cur;
- }
- }
- /**
- * xmlStopParser:
- * @ctxt: an XML parser context
- *
- * Blocks further parser processing
- */
- void
- xmlStopParser(xmlParserCtxtPtr ctxt) {
- if (ctxt == NULL)
- return;
- xmlHaltParser(ctxt);
- ctxt->errNo = XML_ERR_USER_STOP;
- }
- /**
- * xmlCreateIOParserCtxt:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @ioread: an I/O read function
- * @ioclose: an I/O close function
- * @ioctx: an I/O handler
- * @enc: the charset encoding if known
- *
- * Create a parser context for using the XML parser with an existing
- * I/O stream
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateIOParserCtxt(xmlSAXHandlerPtr sax, void *user_data,
- xmlInputReadCallback ioread, xmlInputCloseCallback ioclose,
- void *ioctx, xmlCharEncoding enc) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- xmlParserInputBufferPtr buf;
- if (ioread == NULL) return(NULL);
- buf = xmlParserInputBufferCreateIO(ioread, ioclose, ioctx, enc);
- if (buf == NULL) {
- if (ioclose != NULL)
- ioclose(ioctx);
- return (NULL);
- }
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- if (sax != NULL) {
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- #endif /* LIBXML_SAX1_ENABLED */
- xmlFree(ctxt->sax);
- ctxt->sax = (xmlSAXHandlerPtr) xmlMalloc(sizeof(xmlSAXHandler));
- if (ctxt->sax == NULL) {
- xmlFreeParserInputBuffer(buf);
- xmlErrMemory(ctxt, NULL);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- memset(ctxt->sax, 0, sizeof(xmlSAXHandler));
- if (sax->initialized == XML_SAX2_MAGIC)
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandler));
- else
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandlerV1));
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- inputStream = xmlNewIOInputStream(ctxt, buf, enc);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- return(ctxt);
- }
- #ifdef LIBXML_VALID_ENABLED
- /************************************************************************
- * *
- * Front ends when parsing a DTD *
- * *
- ************************************************************************/
- /**
- * xmlIOParseDTD:
- * @sax: the SAX handler block or NULL
- * @input: an Input Buffer
- * @enc: the charset encoding if known
- *
- * Load and parse a DTD
- *
- * Returns the resulting xmlDtdPtr or NULL in case of error.
- * @input will be freed by the function in any case.
- */
- xmlDtdPtr
- xmlIOParseDTD(xmlSAXHandlerPtr sax, xmlParserInputBufferPtr input,
- xmlCharEncoding enc) {
- xmlDtdPtr ret = NULL;
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr pinput = NULL;
- xmlChar start[4];
- if (input == NULL)
- return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(input);
- return(NULL);
- }
- /* We are loading a DTD */
- ctxt->options |= XML_PARSE_DTDLOAD;
- /*
- * Set-up the SAX context
- */
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- ctxt->userData = ctxt;
- }
- xmlDetectSAX2(ctxt);
- /*
- * generate a parser input from the I/O handler
- */
- pinput = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (pinput == NULL) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserInputBuffer(input);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- /*
- * plug some encoding conversion routines here.
- */
- if (xmlPushInput(ctxt, pinput) < 0) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- pinput->filename = NULL;
- pinput->line = 1;
- pinput->col = 1;
- pinput->base = ctxt->input->cur;
- pinput->cur = ctxt->input->cur;
- pinput->free = NULL;
- /*
- * let's parse that entity knowing it's an external subset.
- */
- ctxt->inSubset = 2;
- ctxt->myDoc = xmlNewDoc(BAD_CAST "1.0");
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return(NULL);
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- ctxt->myDoc->extSubset = xmlNewDtd(ctxt->myDoc, BAD_CAST "none",
- BAD_CAST "none", BAD_CAST "none");
- if ((enc == XML_CHAR_ENCODING_NONE) &&
- ((ctxt->input->end - ctxt->input->cur) >= 4)) {
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- xmlParseExternalSubset(ctxt, BAD_CAST "none", BAD_CAST "none");
- if (ctxt->myDoc != NULL) {
- if (ctxt->wellFormed) {
- ret = ctxt->myDoc->extSubset;
- ctxt->myDoc->extSubset = NULL;
- if (ret != NULL) {
- xmlNodePtr tmp;
- ret->doc = NULL;
- tmp = ret->children;
- while (tmp != NULL) {
- tmp->doc = NULL;
- tmp = tmp->next;
- }
- }
- } else {
- ret = NULL;
- }
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlSAXParseDTD:
- * @sax: the SAX handler block
- * @ExternalID: a NAME* containing the External ID of the DTD
- * @SystemID: a NAME* containing the URL to the DTD
- *
- * Load and parse an external subset.
- *
- * Returns the resulting xmlDtdPtr or NULL in case of error.
- */
- xmlDtdPtr
- xmlSAXParseDTD(xmlSAXHandlerPtr sax, const xmlChar *ExternalID,
- const xmlChar *SystemID) {
- xmlDtdPtr ret = NULL;
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr input = NULL;
- xmlCharEncoding enc;
- xmlChar* systemIdCanonic;
- if ((ExternalID == NULL) && (SystemID == NULL)) return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- return(NULL);
- }
- /* We are loading a DTD */
- ctxt->options |= XML_PARSE_DTDLOAD;
- /*
- * Set-up the SAX context
- */
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- ctxt->userData = ctxt;
- }
- /*
- * Canonicalise the system ID
- */
- systemIdCanonic = xmlCanonicPath(SystemID);
- if ((SystemID != NULL) && (systemIdCanonic == NULL)) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- /*
- * Ask the Entity resolver to load the damn thing
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->resolveEntity != NULL))
- input = ctxt->sax->resolveEntity(ctxt->userData, ExternalID,
- systemIdCanonic);
- if (input == NULL) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- if (systemIdCanonic != NULL)
- xmlFree(systemIdCanonic);
- return(NULL);
- }
- /*
- * plug some encoding conversion routines here.
- */
- if (xmlPushInput(ctxt, input) < 0) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- if (systemIdCanonic != NULL)
- xmlFree(systemIdCanonic);
- return(NULL);
- }
- if ((ctxt->input->end - ctxt->input->cur) >= 4) {
- enc = xmlDetectCharEncoding(ctxt->input->cur, 4);
- xmlSwitchEncoding(ctxt, enc);
- }
- if (input->filename == NULL)
- input->filename = (char *) systemIdCanonic;
- else
- xmlFree(systemIdCanonic);
- input->line = 1;
- input->col = 1;
- input->base = ctxt->input->cur;
- input->cur = ctxt->input->cur;
- input->free = NULL;
- /*
- * let's parse that entity knowing it's an external subset.
- */
- ctxt->inSubset = 2;
- ctxt->myDoc = xmlNewDoc(BAD_CAST "1.0");
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- ctxt->myDoc->extSubset = xmlNewDtd(ctxt->myDoc, BAD_CAST "none",
- ExternalID, SystemID);
- xmlParseExternalSubset(ctxt, ExternalID, SystemID);
- if (ctxt->myDoc != NULL) {
- if (ctxt->wellFormed) {
- ret = ctxt->myDoc->extSubset;
- ctxt->myDoc->extSubset = NULL;
- if (ret != NULL) {
- xmlNodePtr tmp;
- ret->doc = NULL;
- tmp = ret->children;
- while (tmp != NULL) {
- tmp->doc = NULL;
- tmp = tmp->next;
- }
- }
- } else {
- ret = NULL;
- }
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlParseDTD:
- * @ExternalID: a NAME* containing the External ID of the DTD
- * @SystemID: a NAME* containing the URL to the DTD
- *
- * Load and parse an external subset.
- *
- * Returns the resulting xmlDtdPtr or NULL in case of error.
- */
- xmlDtdPtr
- xmlParseDTD(const xmlChar *ExternalID, const xmlChar *SystemID) {
- return(xmlSAXParseDTD(NULL, ExternalID, SystemID));
- }
- #endif /* LIBXML_VALID_ENABLED */
- /************************************************************************
- * *
- * Front ends when parsing an Entity *
- * *
- ************************************************************************/
- /**
- * xmlParseCtxtExternalEntity:
- * @ctx: the existing parsing context
- * @URL: the URL for the entity to load
- * @ID: the System ID for the entity to load
- * @lst: the return value for the set of parsed nodes
- *
- * Parse an external general entity within an existing parsing context
- * An external general parsed entity is well-formed if it matches the
- * production labeled extParsedEnt.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * Returns 0 if the entity is well formed, -1 in case of args problem and
- * the parser error code otherwise
- */
- int
- xmlParseCtxtExternalEntity(xmlParserCtxtPtr ctx, const xmlChar *URL,
- const xmlChar *ID, xmlNodePtr *lst) {
- void *userData;
- if (ctx == NULL) return(-1);
- /*
- * If the user provided their own SAX callbacks, then reuse the
- * userData callback field, otherwise the expected setup in a
- * DOM builder is to have userData == ctxt
- */
- if (ctx->userData == ctx)
- userData = NULL;
- else
- userData = ctx->userData;
- return xmlParseExternalEntityPrivate(ctx->myDoc, ctx, ctx->sax,
- userData, ctx->depth + 1,
- URL, ID, lst);
- }
- /**
- * xmlParseExternalEntityPrivate:
- * @doc: the document the chunk pertains to
- * @oldctxt: the previous parser context if available
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @URL: the URL for the entity to load
- * @ID: the System ID for the entity to load
- * @list: the return value for the set of parsed nodes
- *
- * Private version of xmlParseExternalEntity()
- *
- * Returns 0 if the entity is well formed, -1 in case of args problem and
- * the parser error code otherwise
- */
- static xmlParserErrors
- xmlParseExternalEntityPrivate(xmlDocPtr doc, xmlParserCtxtPtr oldctxt,
- xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *URL,
- const xmlChar *ID, xmlNodePtr *list) {
- xmlParserCtxtPtr ctxt;
- xmlDocPtr newDoc;
- xmlNodePtr newRoot;
- xmlSAXHandlerPtr oldsax = NULL;
- xmlParserErrors ret = XML_ERR_OK;
- xmlChar start[4];
- xmlCharEncoding enc;
- if (((depth > 40) &&
- ((oldctxt == NULL) || (oldctxt->options & XML_PARSE_HUGE) == 0)) ||
- (depth > 1024)) {
- return(XML_ERR_ENTITY_LOOP);
- }
- if (list != NULL)
- *list = NULL;
- if ((URL == NULL) && (ID == NULL))
- return(XML_ERR_INTERNAL_ERROR);
- if (doc == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- ctxt = xmlCreateEntityParserCtxtInternal(URL, ID, NULL, oldctxt);
- if (ctxt == NULL) return(XML_WAR_UNDECLARED_ENTITY);
- ctxt->userData = ctxt;
- if (sax != NULL) {
- oldsax = ctxt->sax;
- ctxt->sax = sax;
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- xmlDetectSAX2(ctxt);
- newDoc = xmlNewDoc(BAD_CAST "1.0");
- if (newDoc == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(XML_ERR_INTERNAL_ERROR);
- }
- newDoc->properties = XML_DOC_INTERNAL;
- if (doc) {
- newDoc->intSubset = doc->intSubset;
- newDoc->extSubset = doc->extSubset;
- if (doc->dict) {
- newDoc->dict = doc->dict;
- xmlDictReference(newDoc->dict);
- }
- if (doc->URL != NULL) {
- newDoc->URL = xmlStrdup(doc->URL);
- }
- }
- newRoot = xmlNewDocNode(newDoc, NULL, BAD_CAST "pseudoroot", NULL);
- if (newRoot == NULL) {
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- xmlFreeDoc(newDoc);
- return(XML_ERR_INTERNAL_ERROR);
- }
- xmlAddChild((xmlNodePtr) newDoc, newRoot);
- nodePush(ctxt, newDoc->children);
- if (doc == NULL) {
- ctxt->myDoc = newDoc;
- } else {
- ctxt->myDoc = doc;
- newRoot->doc = doc;
- }
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- GROW;
- if ((ctxt->input->end - ctxt->input->cur) >= 4) {
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- /*
- * Parse a possible text declaration first
- */
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- xmlParseTextDecl(ctxt);
- /*
- * An XML-1.0 document can't reference an entity not XML-1.0
- */
- if ((xmlStrEqual(oldctxt->version, BAD_CAST "1.0")) &&
- (!xmlStrEqual(ctxt->input->version, BAD_CAST "1.0"))) {
- xmlFatalErrMsg(ctxt, XML_ERR_VERSION_MISMATCH,
- "Version mismatch between document and entity\n");
- }
- }
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->depth = depth;
- if (oldctxt != NULL) {
- ctxt->_private = oldctxt->_private;
- ctxt->loadsubset = oldctxt->loadsubset;
- ctxt->validate = oldctxt->validate;
- ctxt->valid = oldctxt->valid;
- ctxt->replaceEntities = oldctxt->replaceEntities;
- if (oldctxt->validate) {
- ctxt->vctxt.error = oldctxt->vctxt.error;
- ctxt->vctxt.warning = oldctxt->vctxt.warning;
- ctxt->vctxt.userData = oldctxt->vctxt.userData;
- }
- ctxt->external = oldctxt->external;
- if (ctxt->dict) xmlDictFree(ctxt->dict);
- ctxt->dict = oldctxt->dict;
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- ctxt->dictNames = oldctxt->dictNames;
- ctxt->attsDefault = oldctxt->attsDefault;
- ctxt->attsSpecial = oldctxt->attsSpecial;
- ctxt->linenumbers = oldctxt->linenumbers;
- ctxt->record_info = oldctxt->record_info;
- ctxt->node_seq.maximum = oldctxt->node_seq.maximum;
- ctxt->node_seq.length = oldctxt->node_seq.length;
- ctxt->node_seq.buffer = oldctxt->node_seq.buffer;
- } else {
- /*
- * Doing validity checking on chunk without context
- * doesn't make sense
- */
- ctxt->_private = NULL;
- ctxt->validate = 0;
- ctxt->external = 2;
- ctxt->loadsubset = 0;
- }
- xmlParseContent(ctxt);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if (ctxt->node != newDoc->children) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = XML_ERR_INTERNAL_ERROR;
- else
- ret = (xmlParserErrors)ctxt->errNo;
- } else {
- if (list != NULL) {
- xmlNodePtr cur;
- /*
- * Return the newly created nodeset after unlinking it from
- * they pseudo parent.
- */
- cur = newDoc->children->children;
- *list = cur;
- while (cur != NULL) {
- cur->parent = NULL;
- cur = cur->next;
- }
- newDoc->children->children = NULL;
- }
- ret = XML_ERR_OK;
- }
- /*
- * Record in the parent context the number of entities replacement
- * done when parsing that reference.
- */
- if (oldctxt != NULL)
- oldctxt->nbentities += ctxt->nbentities;
- /*
- * Also record the size of the entity parsed
- */
- if (ctxt->input != NULL && oldctxt != NULL) {
- oldctxt->sizeentities += ctxt->input->consumed;
- oldctxt->sizeentities += (ctxt->input->cur - ctxt->input->base);
- }
- /*
- * And record the last error if any
- */
- if ((oldctxt != NULL) && (ctxt->lastError.code != XML_ERR_OK))
- xmlCopyError(&ctxt->lastError, &oldctxt->lastError);
- if (sax != NULL)
- ctxt->sax = oldsax;
- if (oldctxt != NULL) {
- ctxt->dict = NULL;
- ctxt->attsDefault = NULL;
- ctxt->attsSpecial = NULL;
- oldctxt->validate = ctxt->validate;
- oldctxt->valid = ctxt->valid;
- oldctxt->node_seq.maximum = ctxt->node_seq.maximum;
- oldctxt->node_seq.length = ctxt->node_seq.length;
- oldctxt->node_seq.buffer = ctxt->node_seq.buffer;
- }
- ctxt->node_seq.maximum = 0;
- ctxt->node_seq.length = 0;
- ctxt->node_seq.buffer = NULL;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- xmlFreeDoc(newDoc);
- return(ret);
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlParseExternalEntity:
- * @doc: the document the chunk pertains to
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @URL: the URL for the entity to load
- * @ID: the System ID for the entity to load
- * @lst: the return value for the set of parsed nodes
- *
- * Parse an external general entity
- * An external general parsed entity is well-formed if it matches the
- * production labeled extParsedEnt.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * Returns 0 if the entity is well formed, -1 in case of args problem and
- * the parser error code otherwise
- */
- int
- xmlParseExternalEntity(xmlDocPtr doc, xmlSAXHandlerPtr sax, void *user_data,
- int depth, const xmlChar *URL, const xmlChar *ID, xmlNodePtr *lst) {
- return(xmlParseExternalEntityPrivate(doc, NULL, sax, user_data, depth, URL,
- ID, lst));
- }
- /**
- * xmlParseBalancedChunkMemory:
- * @doc: the document the chunk pertains to (must not be NULL)
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @string: the input string in UTF8 or ISO-Latin (zero terminated)
- * @lst: the return value for the set of parsed nodes
- *
- * Parse a well-balanced chunk of an XML document
- * called by the parser
- * The allowed sequence for the Well Balanced Chunk is the one defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns 0 if the chunk is well balanced, -1 in case of args problem and
- * the parser error code otherwise
- */
- int
- xmlParseBalancedChunkMemory(xmlDocPtr doc, xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *string, xmlNodePtr *lst) {
- return xmlParseBalancedChunkMemoryRecover( doc, sax, user_data,
- depth, string, lst, 0 );
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /**
- * xmlParseBalancedChunkMemoryInternal:
- * @oldctxt: the existing parsing context
- * @string: the input string in UTF8 or ISO-Latin (zero terminated)
- * @user_data: the user data field for the parser context
- * @lst: the return value for the set of parsed nodes
- *
- *
- * Parse a well-balanced chunk of an XML document
- * called by the parser
- * The allowed sequence for the Well Balanced Chunk is the one defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns XML_ERR_OK if the chunk is well balanced, and the parser
- * error code otherwise
- *
- * In case recover is set to 1, the nodelist will not be empty even if
- * the parsed chunk is not well balanced.
- */
- static xmlParserErrors
- xmlParseBalancedChunkMemoryInternal(xmlParserCtxtPtr oldctxt,
- const xmlChar *string, void *user_data, xmlNodePtr *lst) {
- xmlParserCtxtPtr ctxt;
- xmlDocPtr newDoc = NULL;
- xmlNodePtr newRoot;
- xmlSAXHandlerPtr oldsax = NULL;
- xmlNodePtr content = NULL;
- xmlNodePtr last = NULL;
- int size;
- xmlParserErrors ret = XML_ERR_OK;
- #ifdef SAX2
- int i;
- #endif
- if (((oldctxt->depth > 40) && ((oldctxt->options & XML_PARSE_HUGE) == 0)) ||
- (oldctxt->depth > 1024)) {
- return(XML_ERR_ENTITY_LOOP);
- }
- if (lst != NULL)
- *lst = NULL;
- if (string == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- size = xmlStrlen(string);
- ctxt = xmlCreateMemoryParserCtxt((char *) string, size);
- if (ctxt == NULL) return(XML_WAR_UNDECLARED_ENTITY);
- if (user_data != NULL)
- ctxt->userData = user_data;
- else
- ctxt->userData = ctxt;
- if (ctxt->dict != NULL) xmlDictFree(ctxt->dict);
- ctxt->dict = oldctxt->dict;
- ctxt->input_id = oldctxt->input_id + 1;
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- #ifdef SAX2
- /* propagate namespaces down the entity */
- for (i = 0;i < oldctxt->nsNr;i += 2) {
- nsPush(ctxt, oldctxt->nsTab[i], oldctxt->nsTab[i+1]);
- }
- #endif
- oldsax = ctxt->sax;
- ctxt->sax = oldctxt->sax;
- xmlDetectSAX2(ctxt);
- ctxt->replaceEntities = oldctxt->replaceEntities;
- ctxt->options = oldctxt->options;
- ctxt->_private = oldctxt->_private;
- if (oldctxt->myDoc == NULL) {
- newDoc = xmlNewDoc(BAD_CAST "1.0");
- if (newDoc == NULL) {
- ctxt->sax = oldsax;
- ctxt->dict = NULL;
- xmlFreeParserCtxt(ctxt);
- return(XML_ERR_INTERNAL_ERROR);
- }
- newDoc->properties = XML_DOC_INTERNAL;
- newDoc->dict = ctxt->dict;
- xmlDictReference(newDoc->dict);
- ctxt->myDoc = newDoc;
- } else {
- ctxt->myDoc = oldctxt->myDoc;
- content = ctxt->myDoc->children;
- last = ctxt->myDoc->last;
- }
- newRoot = xmlNewDocNode(ctxt->myDoc, NULL, BAD_CAST "pseudoroot", NULL);
- if (newRoot == NULL) {
- ctxt->sax = oldsax;
- ctxt->dict = NULL;
- xmlFreeParserCtxt(ctxt);
- if (newDoc != NULL) {
- xmlFreeDoc(newDoc);
- }
- return(XML_ERR_INTERNAL_ERROR);
- }
- ctxt->myDoc->children = NULL;
- ctxt->myDoc->last = NULL;
- xmlAddChild((xmlNodePtr) ctxt->myDoc, newRoot);
- nodePush(ctxt, ctxt->myDoc->children);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->depth = oldctxt->depth + 1;
- ctxt->validate = 0;
- ctxt->loadsubset = oldctxt->loadsubset;
- if ((oldctxt->validate) || (oldctxt->replaceEntities != 0)) {
- /*
- * ID/IDREF registration will be done in xmlValidateElement below
- */
- ctxt->loadsubset |= XML_SKIP_IDS;
- }
- ctxt->dictNames = oldctxt->dictNames;
- ctxt->attsDefault = oldctxt->attsDefault;
- ctxt->attsSpecial = oldctxt->attsSpecial;
- xmlParseContent(ctxt);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if (ctxt->node != ctxt->myDoc->children) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = XML_ERR_INTERNAL_ERROR;
- else
- ret = (xmlParserErrors)ctxt->errNo;
- } else {
- ret = XML_ERR_OK;
- }
- if ((lst != NULL) && (ret == XML_ERR_OK)) {
- xmlNodePtr cur;
- /*
- * Return the newly created nodeset after unlinking it from
- * they pseudo parent.
- */
- cur = ctxt->myDoc->children->children;
- *lst = cur;
- while (cur != NULL) {
- #ifdef LIBXML_VALID_ENABLED
- if ((oldctxt->validate) && (oldctxt->wellFormed) &&
- (oldctxt->myDoc) && (oldctxt->myDoc->intSubset) &&
- (cur->type == XML_ELEMENT_NODE)) {
- oldctxt->valid &= xmlValidateElement(&oldctxt->vctxt,
- oldctxt->myDoc, cur);
- }
- #endif /* LIBXML_VALID_ENABLED */
- cur->parent = NULL;
- cur = cur->next;
- }
- ctxt->myDoc->children->children = NULL;
- }
- if (ctxt->myDoc != NULL) {
- xmlFreeNode(ctxt->myDoc->children);
- ctxt->myDoc->children = content;
- ctxt->myDoc->last = last;
- }
- /*
- * Record in the parent context the number of entities replacement
- * done when parsing that reference.
- */
- if (oldctxt != NULL)
- oldctxt->nbentities += ctxt->nbentities;
- /*
- * Also record the last error if any
- */
- if (ctxt->lastError.code != XML_ERR_OK)
- xmlCopyError(&ctxt->lastError, &oldctxt->lastError);
- ctxt->sax = oldsax;
- ctxt->dict = NULL;
- ctxt->attsDefault = NULL;
- ctxt->attsSpecial = NULL;
- xmlFreeParserCtxt(ctxt);
- if (newDoc != NULL) {
- xmlFreeDoc(newDoc);
- }
- return(ret);
- }
- /**
- * xmlParseInNodeContext:
- * @node: the context node
- * @data: the input string
- * @datalen: the input string length in bytes
- * @options: a combination of xmlParserOption
- * @lst: the return value for the set of parsed nodes
- *
- * Parse a well-balanced chunk of an XML document
- * within the context (DTD, namespaces, etc ...) of the given node.
- *
- * The allowed sequence for the data is a Well Balanced Chunk defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns XML_ERR_OK if the chunk is well balanced, and the parser
- * error code otherwise
- */
- xmlParserErrors
- xmlParseInNodeContext(xmlNodePtr node, const char *data, int datalen,
- int options, xmlNodePtr *lst) {
- #ifdef SAX2
- xmlParserCtxtPtr ctxt;
- xmlDocPtr doc = NULL;
- xmlNodePtr fake, cur;
- int nsnr = 0;
- xmlParserErrors ret = XML_ERR_OK;
- /*
- * check all input parameters, grab the document
- */
- if ((lst == NULL) || (node == NULL) || (data == NULL) || (datalen < 0))
- return(XML_ERR_INTERNAL_ERROR);
- switch (node->type) {
- case XML_ELEMENT_NODE:
- case XML_ATTRIBUTE_NODE:
- case XML_TEXT_NODE:
- case XML_CDATA_SECTION_NODE:
- case XML_ENTITY_REF_NODE:
- case XML_PI_NODE:
- case XML_COMMENT_NODE:
- case XML_DOCUMENT_NODE:
- case XML_HTML_DOCUMENT_NODE:
- break;
- default:
- return(XML_ERR_INTERNAL_ERROR);
- }
- while ((node != NULL) && (node->type != XML_ELEMENT_NODE) &&
- (node->type != XML_DOCUMENT_NODE) &&
- (node->type != XML_HTML_DOCUMENT_NODE))
- node = node->parent;
- if (node == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- if (node->type == XML_ELEMENT_NODE)
- doc = node->doc;
- else
- doc = (xmlDocPtr) node;
- if (doc == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- /*
- * allocate a context and set-up everything not related to the
- * node position in the tree
- */
- if (doc->type == XML_DOCUMENT_NODE)
- ctxt = xmlCreateMemoryParserCtxt((char *) data, datalen);
- #ifdef LIBXML_HTML_ENABLED
- else if (doc->type == XML_HTML_DOCUMENT_NODE) {
- ctxt = htmlCreateMemoryParserCtxt((char *) data, datalen);
- /*
- * When parsing in context, it makes no sense to add implied
- * elements like html/body/etc...
- */
- options |= HTML_PARSE_NOIMPLIED;
- }
- #endif
- else
- return(XML_ERR_INTERNAL_ERROR);
- if (ctxt == NULL)
- return(XML_ERR_NO_MEMORY);
- /*
- * Use input doc's dict if present, else assure XML_PARSE_NODICT is set.
- * We need a dictionary for xmlDetectSAX2, so if there's no doc dict
- * we must wait until the last moment to free the original one.
- */
- if (doc->dict != NULL) {
- if (ctxt->dict != NULL)
- xmlDictFree(ctxt->dict);
- ctxt->dict = doc->dict;
- } else
- options |= XML_PARSE_NODICT;
- if (doc->encoding != NULL) {
- xmlCharEncodingHandlerPtr hdlr;
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = xmlStrdup((const xmlChar *) doc->encoding);
- hdlr = xmlFindCharEncodingHandler((const char *) doc->encoding);
- if (hdlr != NULL) {
- xmlSwitchToEncoding(ctxt, hdlr);
- } else {
- return(XML_ERR_UNSUPPORTED_ENCODING);
- }
- }
- xmlCtxtUseOptionsInternal(ctxt, options, NULL);
- xmlDetectSAX2(ctxt);
- ctxt->myDoc = doc;
- /* parsing in context, i.e. as within existing content */
- ctxt->input_id = 2;
- ctxt->instate = XML_PARSER_CONTENT;
- fake = xmlNewComment(NULL);
- if (fake == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(XML_ERR_NO_MEMORY);
- }
- xmlAddChild(node, fake);
- if (node->type == XML_ELEMENT_NODE) {
- nodePush(ctxt, node);
- /*
- * initialize the SAX2 namespaces stack
- */
- cur = node;
- while ((cur != NULL) && (cur->type == XML_ELEMENT_NODE)) {
- xmlNsPtr ns = cur->nsDef;
- const xmlChar *iprefix, *ihref;
- while (ns != NULL) {
- if (ctxt->dict) {
- iprefix = xmlDictLookup(ctxt->dict, ns->prefix, -1);
- ihref = xmlDictLookup(ctxt->dict, ns->href, -1);
- } else {
- iprefix = ns->prefix;
- ihref = ns->href;
- }
- if (xmlGetNamespace(ctxt, iprefix) == NULL) {
- nsPush(ctxt, iprefix, ihref);
- nsnr++;
- }
- ns = ns->next;
- }
- cur = cur->parent;
- }
- }
- if ((ctxt->validate) || (ctxt->replaceEntities != 0)) {
- /*
- * ID/IDREF registration will be done in xmlValidateElement below
- */
- ctxt->loadsubset |= XML_SKIP_IDS;
- }
- #ifdef LIBXML_HTML_ENABLED
- if (doc->type == XML_HTML_DOCUMENT_NODE)
- __htmlParseContent(ctxt);
- else
- #endif
- xmlParseContent(ctxt);
- nsPop(ctxt, nsnr);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if ((ctxt->node != NULL) && (ctxt->node != node)) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- ctxt->wellFormed = 0;
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = XML_ERR_INTERNAL_ERROR;
- else
- ret = (xmlParserErrors)ctxt->errNo;
- } else {
- ret = XML_ERR_OK;
- }
- /*
- * Return the newly created nodeset after unlinking it from
- * the pseudo sibling.
- */
- cur = fake->next;
- fake->next = NULL;
- node->last = fake;
- if (cur != NULL) {
- cur->prev = NULL;
- }
- *lst = cur;
- while (cur != NULL) {
- cur->parent = NULL;
- cur = cur->next;
- }
- xmlUnlinkNode(fake);
- xmlFreeNode(fake);
- if (ret != XML_ERR_OK) {
- xmlFreeNodeList(*lst);
- *lst = NULL;
- }
- if (doc->dict != NULL)
- ctxt->dict = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- #else /* !SAX2 */
- return(XML_ERR_INTERNAL_ERROR);
- #endif
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlParseBalancedChunkMemoryRecover:
- * @doc: the document the chunk pertains to (must not be NULL)
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @string: the input string in UTF8 or ISO-Latin (zero terminated)
- * @lst: the return value for the set of parsed nodes
- * @recover: return nodes even if the data is broken (use 0)
- *
- *
- * Parse a well-balanced chunk of an XML document
- * called by the parser
- * The allowed sequence for the Well Balanced Chunk is the one defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns 0 if the chunk is well balanced, -1 in case of args problem and
- * the parser error code otherwise
- *
- * In case recover is set to 1, the nodelist will not be empty even if
- * the parsed chunk is not well balanced, assuming the parsing succeeded to
- * some extent.
- */
- int
- xmlParseBalancedChunkMemoryRecover(xmlDocPtr doc, xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *string, xmlNodePtr *lst,
- int recover) {
- xmlParserCtxtPtr ctxt;
- xmlDocPtr newDoc;
- xmlSAXHandlerPtr oldsax = NULL;
- xmlNodePtr content, newRoot;
- int size;
- int ret = 0;
- if (depth > 40) {
- return(XML_ERR_ENTITY_LOOP);
- }
- if (lst != NULL)
- *lst = NULL;
- if (string == NULL)
- return(-1);
- size = xmlStrlen(string);
- ctxt = xmlCreateMemoryParserCtxt((char *) string, size);
- if (ctxt == NULL) return(-1);
- ctxt->userData = ctxt;
- if (sax != NULL) {
- oldsax = ctxt->sax;
- ctxt->sax = sax;
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- newDoc = xmlNewDoc(BAD_CAST "1.0");
- if (newDoc == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(-1);
- }
- newDoc->properties = XML_DOC_INTERNAL;
- if ((doc != NULL) && (doc->dict != NULL)) {
- xmlDictFree(ctxt->dict);
- ctxt->dict = doc->dict;
- xmlDictReference(ctxt->dict);
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- ctxt->dictNames = 1;
- } else {
- xmlCtxtUseOptionsInternal(ctxt, XML_PARSE_NODICT, NULL);
- }
- /* doc == NULL is only supported for historic reasons */
- if (doc != NULL) {
- newDoc->intSubset = doc->intSubset;
- newDoc->extSubset = doc->extSubset;
- }
- newRoot = xmlNewDocNode(newDoc, NULL, BAD_CAST "pseudoroot", NULL);
- if (newRoot == NULL) {
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- xmlFreeDoc(newDoc);
- return(-1);
- }
- xmlAddChild((xmlNodePtr) newDoc, newRoot);
- nodePush(ctxt, newRoot);
- /* doc == NULL is only supported for historic reasons */
- if (doc == NULL) {
- ctxt->myDoc = newDoc;
- } else {
- ctxt->myDoc = newDoc;
- newDoc->children->doc = doc;
- /* Ensure that doc has XML spec namespace */
- xmlSearchNsByHref(doc, (xmlNodePtr)doc, XML_XML_NAMESPACE);
- newDoc->oldNs = doc->oldNs;
- }
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->input_id = 2;
- ctxt->depth = depth;
- /*
- * Doing validity checking on chunk doesn't make sense
- */
- ctxt->validate = 0;
- ctxt->loadsubset = 0;
- xmlDetectSAX2(ctxt);
- if ( doc != NULL ){
- content = doc->children;
- doc->children = NULL;
- xmlParseContent(ctxt);
- doc->children = content;
- }
- else {
- xmlParseContent(ctxt);
- }
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if (ctxt->node != newDoc->children) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = 1;
- else
- ret = ctxt->errNo;
- } else {
- ret = 0;
- }
- if ((lst != NULL) && ((ret == 0) || (recover == 1))) {
- xmlNodePtr cur;
- /*
- * Return the newly created nodeset after unlinking it from
- * they pseudo parent.
- */
- cur = newDoc->children->children;
- *lst = cur;
- while (cur != NULL) {
- xmlSetTreeDoc(cur, doc);
- cur->parent = NULL;
- cur = cur->next;
- }
- newDoc->children->children = NULL;
- }
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- /* This leaks the namespace list if doc == NULL */
- newDoc->oldNs = NULL;
- xmlFreeDoc(newDoc);
- return(ret);
- }
- /**
- * xmlSAXParseEntity:
- * @sax: the SAX handler block
- * @filename: the filename
- *
- * parse an XML external entity out of context and build a tree.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * This correspond to a "Well Balanced" chunk
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseEntity(xmlSAXHandlerPtr sax, const char *filename) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- ctxt = xmlCreateFileParserCtxt(filename);
- if (ctxt == NULL) {
- return(NULL);
- }
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- ctxt->userData = NULL;
- }
- xmlParseExtParsedEnt(ctxt);
- if (ctxt->wellFormed)
- ret = ctxt->myDoc;
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlParseEntity:
- * @filename: the filename
- *
- * parse an XML external entity out of context and build a tree.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * This correspond to a "Well Balanced" chunk
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlParseEntity(const char *filename) {
- return(xmlSAXParseEntity(NULL, filename));
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /**
- * xmlCreateEntityParserCtxtInternal:
- * @URL: the entity URL
- * @ID: the entity PUBLIC ID
- * @base: a possible base for the target URI
- * @pctx: parser context used to set options on new context
- *
- * Create a parser context for an external entity
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time.
- *
- * Returns the new parser context or NULL
- */
- static xmlParserCtxtPtr
- xmlCreateEntityParserCtxtInternal(const xmlChar *URL, const xmlChar *ID,
- const xmlChar *base, xmlParserCtxtPtr pctx) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- char *directory = NULL;
- xmlChar *uri;
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- return(NULL);
- }
- if (pctx != NULL) {
- ctxt->options = pctx->options;
- ctxt->_private = pctx->_private;
- /*
- * this is a subparser of pctx, so the input_id should be
- * incremented to distinguish from main entity
- */
- ctxt->input_id = pctx->input_id + 1;
- }
- /* Don't read from stdin. */
- if (xmlStrcmp(URL, BAD_CAST "-") == 0)
- URL = BAD_CAST "./-";
- uri = xmlBuildURI(URL, base);
- if (uri == NULL) {
- inputStream = xmlLoadExternalEntity((char *)URL, (char *)ID, ctxt);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- if ((ctxt->directory == NULL) && (directory == NULL))
- directory = xmlParserGetDirectory((char *)URL);
- if ((ctxt->directory == NULL) && (directory != NULL))
- ctxt->directory = directory;
- } else {
- inputStream = xmlLoadExternalEntity((char *)uri, (char *)ID, ctxt);
- if (inputStream == NULL) {
- xmlFree(uri);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- if ((ctxt->directory == NULL) && (directory == NULL))
- directory = xmlParserGetDirectory((char *)uri);
- if ((ctxt->directory == NULL) && (directory != NULL))
- ctxt->directory = directory;
- xmlFree(uri);
- }
- return(ctxt);
- }
- /**
- * xmlCreateEntityParserCtxt:
- * @URL: the entity URL
- * @ID: the entity PUBLIC ID
- * @base: a possible base for the target URI
- *
- * Create a parser context for an external entity
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateEntityParserCtxt(const xmlChar *URL, const xmlChar *ID,
- const xmlChar *base) {
- return xmlCreateEntityParserCtxtInternal(URL, ID, base, NULL);
- }
- /************************************************************************
- * *
- * Front ends when parsing from a file *
- * *
- ************************************************************************/
- /**
- * xmlCreateURLParserCtxt:
- * @filename: the filename or URL
- * @options: a combination of xmlParserOption
- *
- * Create a parser context for a file or URL content.
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time and for file accesses
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateURLParserCtxt(const char *filename, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- char *directory = NULL;
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlErrMemory(NULL, "cannot allocate parser context");
- return(NULL);
- }
- if (options)
- xmlCtxtUseOptionsInternal(ctxt, options, NULL);
- ctxt->linenumbers = 1;
- inputStream = xmlLoadExternalEntity(filename, NULL, ctxt);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- if ((ctxt->directory == NULL) && (directory == NULL))
- directory = xmlParserGetDirectory(filename);
- if ((ctxt->directory == NULL) && (directory != NULL))
- ctxt->directory = directory;
- return(ctxt);
- }
- /**
- * xmlCreateFileParserCtxt:
- * @filename: the filename
- *
- * Create a parser context for a file content.
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateFileParserCtxt(const char *filename)
- {
- return(xmlCreateURLParserCtxt(filename, 0));
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlSAXParseFileWithData:
- * @sax: the SAX handler block
- * @filename: the filename
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- * @data: the userdata
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * User data (void *) is stored within the parser context in the
- * context's _private member, so it is available nearly everywhere in libxml
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseFileWithData(xmlSAXHandlerPtr sax, const char *filename,
- int recovery, void *data) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateFileParserCtxt(filename);
- if (ctxt == NULL) {
- return(NULL);
- }
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- }
- xmlDetectSAX2(ctxt);
- if (data!=NULL) {
- ctxt->_private = data;
- }
- if (ctxt->directory == NULL)
- ctxt->directory = xmlParserGetDirectory(filename);
- ctxt->recovery = recovery;
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || recovery) {
- ret = ctxt->myDoc;
- if ((ret != NULL) && (ctxt->input->buf != NULL)) {
- if (ctxt->input->buf->compressed > 0)
- ret->compression = 9;
- else
- ret->compression = ctxt->input->buf->compressed;
- }
- }
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlSAXParseFile:
- * @sax: the SAX handler block
- * @filename: the filename
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseFile(xmlSAXHandlerPtr sax, const char *filename,
- int recovery) {
- return(xmlSAXParseFileWithData(sax,filename,recovery,NULL));
- }
- /**
- * xmlRecoverDoc:
- * @cur: a pointer to an array of xmlChar
- *
- * parse an XML in-memory document and build a tree.
- * In the case the document is not Well Formed, a attempt to build a
- * tree is tried anyway
- *
- * Returns the resulting document tree or NULL in case of failure
- */
- xmlDocPtr
- xmlRecoverDoc(const xmlChar *cur) {
- return(xmlSAXParseDoc(NULL, cur, 1));
- }
- /**
- * xmlParseFile:
- * @filename: the filename
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- *
- * Returns the resulting document tree if the file was wellformed,
- * NULL otherwise.
- */
- xmlDocPtr
- xmlParseFile(const char *filename) {
- return(xmlSAXParseFile(NULL, filename, 0));
- }
- /**
- * xmlRecoverFile:
- * @filename: the filename
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- * In the case the document is not Well Formed, it attempts to build
- * a tree anyway
- *
- * Returns the resulting document tree or NULL in case of failure
- */
- xmlDocPtr
- xmlRecoverFile(const char *filename) {
- return(xmlSAXParseFile(NULL, filename, 1));
- }
- /**
- * xmlSetupParserForBuffer:
- * @ctxt: an XML parser context
- * @buffer: a xmlChar * buffer
- * @filename: a file name
- *
- * Setup the parser context to parse a new buffer; Clears any prior
- * contents from the parser context. The buffer parameter must not be
- * NULL, but the filename parameter can be
- */
- void
- xmlSetupParserForBuffer(xmlParserCtxtPtr ctxt, const xmlChar* buffer,
- const char* filename)
- {
- xmlParserInputPtr input;
- if ((ctxt == NULL) || (buffer == NULL))
- return;
- input = xmlNewInputStream(ctxt);
- if (input == NULL) {
- xmlErrMemory(NULL, "parsing new buffer: out of memory\n");
- xmlClearParserCtxt(ctxt);
- return;
- }
- xmlClearParserCtxt(ctxt);
- if (filename != NULL)
- input->filename = (char *) xmlCanonicPath((const xmlChar *)filename);
- input->base = buffer;
- input->cur = buffer;
- input->end = &buffer[xmlStrlen(buffer)];
- inputPush(ctxt, input);
- }
- /**
- * xmlSAXUserParseFile:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @filename: a file name
- *
- * parse an XML file and call the given SAX handler routines.
- * Automatic support for ZLIB/Compress compressed document is provided
- *
- * Returns 0 in case of success or a error number otherwise
- */
- int
- xmlSAXUserParseFile(xmlSAXHandlerPtr sax, void *user_data,
- const char *filename) {
- int ret = 0;
- xmlParserCtxtPtr ctxt;
- ctxt = xmlCreateFileParserCtxt(filename);
- if (ctxt == NULL) return -1;
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- xmlDetectSAX2(ctxt);
- if (user_data != NULL)
- ctxt->userData = user_data;
- xmlParseDocument(ctxt);
- if (ctxt->wellFormed)
- ret = 0;
- else {
- if (ctxt->errNo != 0)
- ret = ctxt->errNo;
- else
- ret = -1;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- if (ctxt->myDoc != NULL) {
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- xmlFreeParserCtxt(ctxt);
- return ret;
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /************************************************************************
- * *
- * Front ends when parsing from memory *
- * *
- ************************************************************************/
- /**
- * xmlCreateMemoryParserCtxt:
- * @buffer: a pointer to a char array
- * @size: the size of the array
- *
- * Create a parser context for an XML in-memory document.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateMemoryParserCtxt(const char *buffer, int size) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr input;
- xmlParserInputBufferPtr buf;
- if (buffer == NULL)
- return(NULL);
- if (size <= 0)
- return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL)
- return(NULL);
- /* TODO: xmlParserInputBufferCreateStatic, requires some serious changes */
- buf = xmlParserInputBufferCreateMem(buffer, size, XML_CHAR_ENCODING_NONE);
- if (buf == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- input = xmlNewInputStream(ctxt);
- if (input == NULL) {
- xmlFreeParserInputBuffer(buf);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- input->filename = NULL;
- input->buf = buf;
- xmlBufResetInput(input->buf->buffer, input);
- inputPush(ctxt, input);
- return(ctxt);
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlSAXParseMemoryWithData:
- * @sax: the SAX handler block
- * @buffer: an pointer to a char array
- * @size: the size of the array
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- * @data: the userdata
- *
- * parse an XML in-memory block and use the given SAX function block
- * to handle the parsing callback. If sax is NULL, fallback to the default
- * DOM tree building routines.
- *
- * User data (void *) is stored within the parser context in the
- * context's _private member, so it is available nearly everywhere in libxml
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseMemoryWithData(xmlSAXHandlerPtr sax, const char *buffer,
- int size, int recovery, void *data) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateMemoryParserCtxt(buffer, size);
- if (ctxt == NULL) return(NULL);
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- }
- xmlDetectSAX2(ctxt);
- if (data!=NULL) {
- ctxt->_private=data;
- }
- ctxt->recovery = recovery;
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || recovery) ret = ctxt->myDoc;
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlSAXParseMemory:
- * @sax: the SAX handler block
- * @buffer: an pointer to a char array
- * @size: the size of the array
- * @recovery: work in recovery mode, i.e. tries to read not Well Formed
- * documents
- *
- * parse an XML in-memory block and use the given SAX function block
- * to handle the parsing callback. If sax is NULL, fallback to the default
- * DOM tree building routines.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseMemory(xmlSAXHandlerPtr sax, const char *buffer,
- int size, int recovery) {
- return xmlSAXParseMemoryWithData(sax, buffer, size, recovery, NULL);
- }
- /**
- * xmlParseMemory:
- * @buffer: an pointer to a char array
- * @size: the size of the array
- *
- * parse an XML in-memory block and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr xmlParseMemory(const char *buffer, int size) {
- return(xmlSAXParseMemory(NULL, buffer, size, 0));
- }
- /**
- * xmlRecoverMemory:
- * @buffer: an pointer to a char array
- * @size: the size of the array
- *
- * parse an XML in-memory block and build a tree.
- * In the case the document is not Well Formed, an attempt to
- * build a tree is tried anyway
- *
- * Returns the resulting document tree or NULL in case of error
- */
- xmlDocPtr xmlRecoverMemory(const char *buffer, int size) {
- return(xmlSAXParseMemory(NULL, buffer, size, 1));
- }
- /**
- * xmlSAXUserParseMemory:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @buffer: an in-memory XML document input
- * @size: the length of the XML document in bytes
- *
- * A better SAX parsing routine.
- * parse an XML in-memory buffer and call the given SAX handler routines.
- *
- * Returns 0 in case of success or a error number otherwise
- */
- int xmlSAXUserParseMemory(xmlSAXHandlerPtr sax, void *user_data,
- const char *buffer, int size) {
- int ret = 0;
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateMemoryParserCtxt(buffer, size);
- if (ctxt == NULL) return -1;
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- xmlDetectSAX2(ctxt);
- if (user_data != NULL)
- ctxt->userData = user_data;
- xmlParseDocument(ctxt);
- if (ctxt->wellFormed)
- ret = 0;
- else {
- if (ctxt->errNo != 0)
- ret = ctxt->errNo;
- else
- ret = -1;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- if (ctxt->myDoc != NULL) {
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- xmlFreeParserCtxt(ctxt);
- return ret;
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /**
- * xmlCreateDocParserCtxt:
- * @cur: a pointer to an array of xmlChar
- *
- * Creates a parser context for an XML in-memory document.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateDocParserCtxt(const xmlChar *cur) {
- int len;
- if (cur == NULL)
- return(NULL);
- len = xmlStrlen(cur);
- return(xmlCreateMemoryParserCtxt((const char *)cur, len));
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlSAXParseDoc:
- * @sax: the SAX handler block
- * @cur: a pointer to an array of xmlChar
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- *
- * parse an XML in-memory document and build a tree.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseDoc(xmlSAXHandlerPtr sax, const xmlChar *cur, int recovery) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- xmlSAXHandlerPtr oldsax = NULL;
- if (cur == NULL) return(NULL);
- ctxt = xmlCreateDocParserCtxt(cur);
- if (ctxt == NULL) return(NULL);
- if (sax != NULL) {
- oldsax = ctxt->sax;
- ctxt->sax = sax;
- ctxt->userData = NULL;
- }
- xmlDetectSAX2(ctxt);
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || recovery) ret = ctxt->myDoc;
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlParseDoc:
- * @cur: a pointer to an array of xmlChar
- *
- * parse an XML in-memory document and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlParseDoc(const xmlChar *cur) {
- return(xmlSAXParseDoc(NULL, cur, 0));
- }
- #endif /* LIBXML_SAX1_ENABLED */
- #ifdef LIBXML_LEGACY_ENABLED
- /************************************************************************
- * *
- * Specific function to keep track of entities references *
- * and used by the XSLT debugger *
- * *
- ************************************************************************/
- static xmlEntityReferenceFunc xmlEntityRefFunc = NULL;
- /**
- * xmlAddEntityReference:
- * @ent : A valid entity
- * @firstNode : A valid first node for children of entity
- * @lastNode : A valid last node of children entity
- *
- * Notify of a reference to an entity of type XML_EXTERNAL_GENERAL_PARSED_ENTITY
- */
- static void
- xmlAddEntityReference(xmlEntityPtr ent, xmlNodePtr firstNode,
- xmlNodePtr lastNode)
- {
- if (xmlEntityRefFunc != NULL) {
- (*xmlEntityRefFunc) (ent, firstNode, lastNode);
- }
- }
- /**
- * xmlSetEntityReferenceFunc:
- * @func: A valid function
- *
- * Set the function to call call back when a xml reference has been made
- */
- void
- xmlSetEntityReferenceFunc(xmlEntityReferenceFunc func)
- {
- xmlEntityRefFunc = func;
- }
- #endif /* LIBXML_LEGACY_ENABLED */
- /************************************************************************
- * *
- * Miscellaneous *
- * *
- ************************************************************************/
- #ifdef LIBXML_XPATH_ENABLED
- #include <libxml/xpath.h>
- #endif
- extern void XMLCDECL xmlGenericErrorDefaultFunc(void *ctx, const char *msg, ...);
- static int xmlParserInitialized = 0;
- /**
- * xmlInitParser:
- *
- * Initialization function for the XML parser.
- * This is not reentrant. Call once before processing in case of
- * use in multithreaded programs.
- */
- void
- xmlInitParser(void) {
- if (xmlParserInitialized != 0)
- return;
- #if defined(_WIN32) && (!defined(LIBXML_STATIC) || defined(LIBXML_STATIC_FOR_DLL))
- if (xmlFree == free)
- atexit(xmlCleanupParser);
- #endif
- #ifdef LIBXML_THREAD_ENABLED
- __xmlGlobalInitMutexLock();
- if (xmlParserInitialized == 0) {
- #endif
- xmlInitThreads();
- xmlInitGlobals();
- if ((xmlGenericError == xmlGenericErrorDefaultFunc) ||
- (xmlGenericError == NULL))
- initGenericErrorDefaultFunc(NULL);
- xmlInitMemory();
- xmlInitializeDict();
- xmlInitCharEncodingHandlers();
- xmlDefaultSAXHandlerInit();
- xmlRegisterDefaultInputCallbacks();
- #ifdef LIBXML_OUTPUT_ENABLED
- xmlRegisterDefaultOutputCallbacks();
- #endif /* LIBXML_OUTPUT_ENABLED */
- #ifdef LIBXML_HTML_ENABLED
- htmlInitAutoClose();
- htmlDefaultSAXHandlerInit();
- #endif
- #ifdef LIBXML_XPATH_ENABLED
- xmlXPathInit();
- #endif
- xmlParserInitialized = 1;
- #ifdef LIBXML_THREAD_ENABLED
- }
- __xmlGlobalInitMutexUnlock();
- #endif
- }
- /**
- * xmlCleanupParser:
- *
- * This function name is somewhat misleading. It does not clean up
- * parser state, it cleans up memory allocated by the library itself.
- * It is a cleanup function for the XML library. It tries to reclaim all
- * related global memory allocated for the library processing.
- * It doesn't deallocate any document related memory. One should
- * call xmlCleanupParser() only when the process has finished using
- * the library and all XML/HTML documents built with it.
- * See also xmlInitParser() which has the opposite function of preparing
- * the library for operations.
- *
- * WARNING: if your application is multithreaded or has plugin support
- * calling this may crash the application if another thread or
- * a plugin is still using libxml2. It's sometimes very hard to
- * guess if libxml2 is in use in the application, some libraries
- * or plugins may use it without notice. In case of doubt abstain
- * from calling this function or do it just before calling exit()
- * to avoid leak reports from valgrind !
- */
- void
- xmlCleanupParser(void) {
- if (!xmlParserInitialized)
- return;
- xmlCleanupCharEncodingHandlers();
- #ifdef LIBXML_CATALOG_ENABLED
- xmlCatalogCleanup();
- #endif
- xmlDictCleanup();
- xmlCleanupInputCallbacks();
- #ifdef LIBXML_OUTPUT_ENABLED
- xmlCleanupOutputCallbacks();
- #endif
- #ifdef LIBXML_SCHEMAS_ENABLED
- xmlSchemaCleanupTypes();
- xmlRelaxNGCleanupTypes();
- #endif
- xmlResetLastError();
- xmlCleanupGlobals();
- xmlCleanupThreads(); /* must be last if called not from the main thread */
- xmlCleanupMemory();
- xmlParserInitialized = 0;
- }
- #if defined(HAVE_ATTRIBUTE_DESTRUCTOR) && !defined(LIBXML_STATIC) && \
- !defined(_WIN32)
- static void
- ATTRIBUTE_DESTRUCTOR
- xmlDestructor(void) {
- /*
- * Calling custom deallocation functions in a destructor can cause
- * problems, for example with Nokogiri.
- */
- if (xmlFree == free)
- xmlCleanupParser();
- }
- #endif
- /************************************************************************
- * *
- * New set (2.6.0) of simpler and more flexible APIs *
- * *
- ************************************************************************/
- /**
- * DICT_FREE:
- * @str: a string
- *
- * Free a string if it is not owned by the "dict" dictionary in the
- * current scope
- */
- #define DICT_FREE(str) \
- if ((str) && ((!dict) || \
- (xmlDictOwns(dict, (const xmlChar *)(str)) == 0))) \
- xmlFree((char *)(str));
- /**
- * xmlCtxtReset:
- * @ctxt: an XML parser context
- *
- * Reset a parser context
- */
- void
- xmlCtxtReset(xmlParserCtxtPtr ctxt)
- {
- xmlParserInputPtr input;
- xmlDictPtr dict;
- if (ctxt == NULL)
- return;
- dict = ctxt->dict;
- while ((input = inputPop(ctxt)) != NULL) { /* Non consuming */
- xmlFreeInputStream(input);
- }
- ctxt->inputNr = 0;
- ctxt->input = NULL;
- ctxt->spaceNr = 0;
- if (ctxt->spaceTab != NULL) {
- ctxt->spaceTab[0] = -1;
- ctxt->space = &ctxt->spaceTab[0];
- } else {
- ctxt->space = NULL;
- }
- ctxt->nodeNr = 0;
- ctxt->node = NULL;
- ctxt->nameNr = 0;
- ctxt->name = NULL;
- DICT_FREE(ctxt->version);
- ctxt->version = NULL;
- DICT_FREE(ctxt->encoding);
- ctxt->encoding = NULL;
- DICT_FREE(ctxt->directory);
- ctxt->directory = NULL;
- DICT_FREE(ctxt->extSubURI);
- ctxt->extSubURI = NULL;
- DICT_FREE(ctxt->extSubSystem);
- ctxt->extSubSystem = NULL;
- if (ctxt->myDoc != NULL)
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- ctxt->standalone = -1;
- ctxt->hasExternalSubset = 0;
- ctxt->hasPErefs = 0;
- ctxt->html = 0;
- ctxt->external = 0;
- ctxt->instate = XML_PARSER_START;
- ctxt->token = 0;
- ctxt->wellFormed = 1;
- ctxt->nsWellFormed = 1;
- ctxt->disableSAX = 0;
- ctxt->valid = 1;
- #if 0
- ctxt->vctxt.userData = ctxt;
- ctxt->vctxt.error = xmlParserValidityError;
- ctxt->vctxt.warning = xmlParserValidityWarning;
- #endif
- ctxt->record_info = 0;
- ctxt->checkIndex = 0;
- ctxt->inSubset = 0;
- ctxt->errNo = XML_ERR_OK;
- ctxt->depth = 0;
- ctxt->charset = XML_CHAR_ENCODING_UTF8;
- ctxt->catalogs = NULL;
- ctxt->nbentities = 0;
- ctxt->sizeentities = 0;
- ctxt->sizeentcopy = 0;
- xmlInitNodeInfoSeq(&ctxt->node_seq);
- if (ctxt->attsDefault != NULL) {
- xmlHashFree(ctxt->attsDefault, xmlHashDefaultDeallocator);
- ctxt->attsDefault = NULL;
- }
- if (ctxt->attsSpecial != NULL) {
- xmlHashFree(ctxt->attsSpecial, NULL);
- ctxt->attsSpecial = NULL;
- }
- #ifdef LIBXML_CATALOG_ENABLED
- if (ctxt->catalogs != NULL)
- xmlCatalogFreeLocal(ctxt->catalogs);
- #endif
- if (ctxt->lastError.code != XML_ERR_OK)
- xmlResetError(&ctxt->lastError);
- }
- /**
- * xmlCtxtResetPush:
- * @ctxt: an XML parser context
- * @chunk: a pointer to an array of chars
- * @size: number of chars in the array
- * @filename: an optional file name or URI
- * @encoding: the document encoding, or NULL
- *
- * Reset a push parser context
- *
- * Returns 0 in case of success and 1 in case of error
- */
- int
- xmlCtxtResetPush(xmlParserCtxtPtr ctxt, const char *chunk,
- int size, const char *filename, const char *encoding)
- {
- xmlParserInputPtr inputStream;
- xmlParserInputBufferPtr buf;
- xmlCharEncoding enc = XML_CHAR_ENCODING_NONE;
- if (ctxt == NULL)
- return(1);
- if ((encoding == NULL) && (chunk != NULL) && (size >= 4))
- enc = xmlDetectCharEncoding((const xmlChar *) chunk, size);
- buf = xmlAllocParserInputBuffer(enc);
- if (buf == NULL)
- return(1);
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(buf);
- return(1);
- }
- xmlCtxtReset(ctxt);
- if (filename == NULL) {
- ctxt->directory = NULL;
- } else {
- ctxt->directory = xmlParserGetDirectory(filename);
- }
- inputStream = xmlNewInputStream(ctxt);
- if (inputStream == NULL) {
- xmlFreeParserInputBuffer(buf);
- return(1);
- }
- if (filename == NULL)
- inputStream->filename = NULL;
- else
- inputStream->filename = (char *)
- xmlCanonicPath((const xmlChar *) filename);
- inputStream->buf = buf;
- xmlBufResetInput(buf->buffer, inputStream);
- inputPush(ctxt, inputStream);
- if ((size > 0) && (chunk != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer, ctxt->input);
- size_t cur = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, size, chunk);
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input, base, cur);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: pushed %d\n", size);
- #endif
- }
- if (encoding != NULL) {
- xmlCharEncodingHandlerPtr hdlr;
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = xmlStrdup((const xmlChar *) encoding);
- hdlr = xmlFindCharEncodingHandler(encoding);
- if (hdlr != NULL) {
- xmlSwitchToEncoding(ctxt, hdlr);
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNSUPPORTED_ENCODING,
- "Unsupported encoding %s\n", BAD_CAST encoding);
- }
- } else if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- return(0);
- }
- /**
- * xmlCtxtUseOptionsInternal:
- * @ctxt: an XML parser context
- * @options: a combination of xmlParserOption
- * @encoding: the user provided encoding to use
- *
- * Applies the options to the parser context
- *
- * Returns 0 in case of success, the set of unknown or unimplemented options
- * in case of error.
- */
- static int
- xmlCtxtUseOptionsInternal(xmlParserCtxtPtr ctxt, int options, const char *encoding)
- {
- if (ctxt == NULL)
- return(-1);
- if (encoding != NULL) {
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = xmlStrdup((const xmlChar *) encoding);
- }
- if (options & XML_PARSE_RECOVER) {
- ctxt->recovery = 1;
- options -= XML_PARSE_RECOVER;
- ctxt->options |= XML_PARSE_RECOVER;
- } else
- ctxt->recovery = 0;
- if (options & XML_PARSE_DTDLOAD) {
- ctxt->loadsubset = XML_DETECT_IDS;
- options -= XML_PARSE_DTDLOAD;
- ctxt->options |= XML_PARSE_DTDLOAD;
- } else
- ctxt->loadsubset = 0;
- if (options & XML_PARSE_DTDATTR) {
- ctxt->loadsubset |= XML_COMPLETE_ATTRS;
- options -= XML_PARSE_DTDATTR;
- ctxt->options |= XML_PARSE_DTDATTR;
- }
- if (options & XML_PARSE_NOENT) {
- ctxt->replaceEntities = 1;
- /* ctxt->loadsubset |= XML_DETECT_IDS; */
- options -= XML_PARSE_NOENT;
- ctxt->options |= XML_PARSE_NOENT;
- } else
- ctxt->replaceEntities = 0;
- if (options & XML_PARSE_PEDANTIC) {
- ctxt->pedantic = 1;
- options -= XML_PARSE_PEDANTIC;
- ctxt->options |= XML_PARSE_PEDANTIC;
- } else
- ctxt->pedantic = 0;
- if (options & XML_PARSE_NOBLANKS) {
- ctxt->keepBlanks = 0;
- ctxt->sax->ignorableWhitespace = xmlSAX2IgnorableWhitespace;
- options -= XML_PARSE_NOBLANKS;
- ctxt->options |= XML_PARSE_NOBLANKS;
- } else
- ctxt->keepBlanks = 1;
- if (options & XML_PARSE_DTDVALID) {
- ctxt->validate = 1;
- if (options & XML_PARSE_NOWARNING)
- ctxt->vctxt.warning = NULL;
- if (options & XML_PARSE_NOERROR)
- ctxt->vctxt.error = NULL;
- options -= XML_PARSE_DTDVALID;
- ctxt->options |= XML_PARSE_DTDVALID;
- } else
- ctxt->validate = 0;
- if (options & XML_PARSE_NOWARNING) {
- ctxt->sax->warning = NULL;
- options -= XML_PARSE_NOWARNING;
- }
- if (options & XML_PARSE_NOERROR) {
- ctxt->sax->error = NULL;
- ctxt->sax->fatalError = NULL;
- options -= XML_PARSE_NOERROR;
- }
- #ifdef LIBXML_SAX1_ENABLED
- if (options & XML_PARSE_SAX1) {
- ctxt->sax->startElement = xmlSAX2StartElement;
- ctxt->sax->endElement = xmlSAX2EndElement;
- ctxt->sax->startElementNs = NULL;
- ctxt->sax->endElementNs = NULL;
- ctxt->sax->initialized = 1;
- options -= XML_PARSE_SAX1;
- ctxt->options |= XML_PARSE_SAX1;
- }
- #endif /* LIBXML_SAX1_ENABLED */
- if (options & XML_PARSE_NODICT) {
- ctxt->dictNames = 0;
- options -= XML_PARSE_NODICT;
- ctxt->options |= XML_PARSE_NODICT;
- } else {
- ctxt->dictNames = 1;
- }
- if (options & XML_PARSE_NOCDATA) {
- ctxt->sax->cdataBlock = NULL;
- options -= XML_PARSE_NOCDATA;
- ctxt->options |= XML_PARSE_NOCDATA;
- }
- if (options & XML_PARSE_NSCLEAN) {
- ctxt->options |= XML_PARSE_NSCLEAN;
- options -= XML_PARSE_NSCLEAN;
- }
- if (options & XML_PARSE_NONET) {
- ctxt->options |= XML_PARSE_NONET;
- options -= XML_PARSE_NONET;
- }
- if (options & XML_PARSE_COMPACT) {
- ctxt->options |= XML_PARSE_COMPACT;
- options -= XML_PARSE_COMPACT;
- }
- if (options & XML_PARSE_OLD10) {
- ctxt->options |= XML_PARSE_OLD10;
- options -= XML_PARSE_OLD10;
- }
- if (options & XML_PARSE_NOBASEFIX) {
- ctxt->options |= XML_PARSE_NOBASEFIX;
- options -= XML_PARSE_NOBASEFIX;
- }
- if (options & XML_PARSE_HUGE) {
- ctxt->options |= XML_PARSE_HUGE;
- options -= XML_PARSE_HUGE;
- if (ctxt->dict != NULL)
- xmlDictSetLimit(ctxt->dict, 0);
- }
- if (options & XML_PARSE_OLDSAX) {
- ctxt->options |= XML_PARSE_OLDSAX;
- options -= XML_PARSE_OLDSAX;
- }
- if (options & XML_PARSE_IGNORE_ENC) {
- ctxt->options |= XML_PARSE_IGNORE_ENC;
- options -= XML_PARSE_IGNORE_ENC;
- }
- if (options & XML_PARSE_BIG_LINES) {
- ctxt->options |= XML_PARSE_BIG_LINES;
- options -= XML_PARSE_BIG_LINES;
- }
- ctxt->linenumbers = 1;
- return (options);
- }
- /**
- * xmlCtxtUseOptions:
- * @ctxt: an XML parser context
- * @options: a combination of xmlParserOption
- *
- * Applies the options to the parser context
- *
- * Returns 0 in case of success, the set of unknown or unimplemented options
- * in case of error.
- */
- int
- xmlCtxtUseOptions(xmlParserCtxtPtr ctxt, int options)
- {
- return(xmlCtxtUseOptionsInternal(ctxt, options, NULL));
- }
- /**
- * xmlDoRead:
- * @ctxt: an XML parser context
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- * @reuse: keep the context for reuse
- *
- * Common front-end for the xmlRead functions
- *
- * Returns the resulting document tree or NULL
- */
- static xmlDocPtr
- xmlDoRead(xmlParserCtxtPtr ctxt, const char *URL, const char *encoding,
- int options, int reuse)
- {
- xmlDocPtr ret;
- xmlCtxtUseOptionsInternal(ctxt, options, encoding);
- if (encoding != NULL) {
- xmlCharEncodingHandlerPtr hdlr;
- hdlr = xmlFindCharEncodingHandler(encoding);
- if (hdlr != NULL)
- xmlSwitchToEncoding(ctxt, hdlr);
- }
- if ((URL != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->filename == NULL))
- ctxt->input->filename = (char *) xmlStrdup((const xmlChar *) URL);
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || ctxt->recovery)
- ret = ctxt->myDoc;
- else {
- ret = NULL;
- if (ctxt->myDoc != NULL) {
- xmlFreeDoc(ctxt->myDoc);
- }
- }
- ctxt->myDoc = NULL;
- if (!reuse) {
- xmlFreeParserCtxt(ctxt);
- }
- return (ret);
- }
- /**
- * xmlReadDoc:
- * @cur: a pointer to a zero terminated string
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadDoc(const xmlChar * cur, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- if (cur == NULL)
- return (NULL);
- xmlInitParser();
- ctxt = xmlCreateDocParserCtxt(cur);
- if (ctxt == NULL)
- return (NULL);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlReadFile:
- * @filename: a file or URL
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML file from the filesystem or the network.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadFile(const char *filename, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateURLParserCtxt(filename, options);
- if (ctxt == NULL)
- return (NULL);
- return (xmlDoRead(ctxt, NULL, encoding, options, 0));
- }
- /**
- * xmlReadMemory:
- * @buffer: a pointer to a char array
- * @size: the size of the array
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadMemory(const char *buffer, int size, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateMemoryParserCtxt(buffer, size);
- if (ctxt == NULL)
- return (NULL);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlReadFd:
- * @fd: an open file descriptor
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML from a file descriptor and build a tree.
- * NOTE that the file descriptor will not be closed when the
- * reader is closed or reset.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadFd(int fd, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (fd < 0)
- return (NULL);
- xmlInitParser();
- input = xmlParserInputBufferCreateFd(fd, XML_CHAR_ENCODING_NONE);
- if (input == NULL)
- return (NULL);
- input->closecallback = NULL;
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- xmlFreeParserCtxt(ctxt);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlReadIO:
- * @ioread: an I/O read function
- * @ioclose: an I/O close function
- * @ioctx: an I/O handler
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML document from I/O functions and source and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadIO(xmlInputReadCallback ioread, xmlInputCloseCallback ioclose,
- void *ioctx, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (ioread == NULL)
- return (NULL);
- xmlInitParser();
- input = xmlParserInputBufferCreateIO(ioread, ioclose, ioctx,
- XML_CHAR_ENCODING_NONE);
- if (input == NULL) {
- if (ioclose != NULL)
- ioclose(ioctx);
- return (NULL);
- }
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- xmlFreeParserCtxt(ctxt);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlCtxtReadDoc:
- * @ctxt: an XML parser context
- * @cur: a pointer to a zero terminated string
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadDoc(xmlParserCtxtPtr ctxt, const xmlChar * cur,
- const char *URL, const char *encoding, int options)
- {
- xmlParserInputPtr stream;
- if (cur == NULL)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- stream = xmlNewStringInputStream(ctxt, cur);
- if (stream == NULL) {
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadFile:
- * @ctxt: an XML parser context
- * @filename: a file or URL
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML file from the filesystem or the network.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadFile(xmlParserCtxtPtr ctxt, const char *filename,
- const char *encoding, int options)
- {
- xmlParserInputPtr stream;
- if (filename == NULL)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- stream = xmlLoadExternalEntity(filename, NULL, ctxt);
- if (stream == NULL) {
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, NULL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadMemory:
- * @ctxt: an XML parser context
- * @buffer: a pointer to a char array
- * @size: the size of the array
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadMemory(xmlParserCtxtPtr ctxt, const char *buffer, int size,
- const char *URL, const char *encoding, int options)
- {
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (ctxt == NULL)
- return (NULL);
- if (buffer == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- input = xmlParserInputBufferCreateMem(buffer, size, XML_CHAR_ENCODING_NONE);
- if (input == NULL) {
- return(NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- return(NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadFd:
- * @ctxt: an XML parser context
- * @fd: an open file descriptor
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML from a file descriptor and build a tree.
- * This reuses the existing @ctxt parser context
- * NOTE that the file descriptor will not be closed when the
- * reader is closed or reset.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadFd(xmlParserCtxtPtr ctxt, int fd,
- const char *URL, const char *encoding, int options)
- {
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (fd < 0)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- input = xmlParserInputBufferCreateFd(fd, XML_CHAR_ENCODING_NONE);
- if (input == NULL)
- return (NULL);
- input->closecallback = NULL;
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadIO:
- * @ctxt: an XML parser context
- * @ioread: an I/O read function
- * @ioclose: an I/O close function
- * @ioctx: an I/O handler
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML document from I/O functions and source and build a tree.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadIO(xmlParserCtxtPtr ctxt, xmlInputReadCallback ioread,
- xmlInputCloseCallback ioclose, void *ioctx,
- const char *URL,
- const char *encoding, int options)
- {
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (ioread == NULL)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- input = xmlParserInputBufferCreateIO(ioread, ioclose, ioctx,
- XML_CHAR_ENCODING_NONE);
- if (input == NULL) {
- if (ioclose != NULL)
- ioclose(ioctx);
- return (NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- #define bottom_parser
- #include "elfgcchack.h"
|