Home Explore Help
Register Sign In
gearsix
Repositories Public Activity

gearsix pushed to buf at gearsix/ec

  • aec8f5bb13 added (*Buffer).find,index;started (*Buffer).split,insert At this point these are pretty much directly ported from the C implementation I made. Split is going to need to consider how to handle the file caching outlined in the *Table Data Format*. Also going to need to give undo/redo's a think.
  • 842ba1ce7e Added file caching for buffers The read file is going to be cached by creating a hard link, then when the file is re-opened this hard link will be checked against the next read file to check to make sure they're the same. Hopefully this will avoid issues with the input file being different (even if it has the same name). Files that have the same name as the read file with "app" and "tbl" appended will also be created containing the corresponding data. - loadCache loads a cached read file and checks if it matches the input file given (by calling SameFile on a Stat of both). - loadCache requires *all* cache files be loaded for a cache to be valid.
  • 24c616241a tarted working on logic for read file caching The read file is going to be cached by creating a hard link, then when the file is re-opened this hard link will be checked against the next read file to check if it's the same (despite having the same name). Hopefully this will avoid issues with the input file being different (even if it has the same name). The loadcache function needs to be tidied up next, with the read file cache check it's gotten a little messy. - added skeleton for initcache - loadcache loads a cached read file and checks if it matches the input file given (by calling SameFile on a Stat of both). - loadcache requires *all* cache files be loaded for a cache to be valid. - some minor compilation bugfixes (mostly typos, etc).
  • View comparison for these 3 commits »

11 hours ago

gearsix pushed to buf at gearsix/ec

  • 24c616241a tarted working on logic for read file caching The read file is going to be cached by creating a hard link, then when the file is re-opened this hard link will be checked against the next read file to check if it's the same (despite having the same name). Hopefully this will avoid issues with the input file being different (even if it has the same name). The loadcache function needs to be tidied up next, with the read file cache check it's gotten a little messy. - added skeleton for initcache - loadcache loads a cached read file and checks if it matches the input file given (by calling SameFile on a Stat of both). - loadcache requires *all* cache files be loaded for a cache to be valid. - some minor compilation bugfixes (mostly typos, etc).
  • ed5db690c2 Define the *Table Data Format* and implemented Buffer.Reload() The *Table Data Format* could just be JSON but this is much more efficient and easier to write a parser for (although everything has a library for parsing JSON anyway I guess). The format contains a line for each piece, where each line contains (in regex: `[ >][ra] \d+ \d+`. The first character indicates if it's the Buffer.pos, the next indicates if the piece `f` value is read/append, the next is a number of the `off` value and then the number for the `siz` value. - renamed `piece.len` -> `piece.siz` (avoid nameclash with len()) - added comment doc at the top of *buf.go* defining the Table Data Format. - added a comment noting that the read file should be cached with a link - renamed `(*Buffer).recover()` -> `(*Buffer).readcache` to avoid nameclash with recover()
  • 02d1ddbde5 started work on buf.go buf.go implements the text buffering system that this program will use. It implements a Piece Table (see file header comment for details), which is originally the reason I started prototyping this tool. Using Golang has made it easy enough for me to use files for a nice caching system of the piece table runtime data, so this commit has added cfg.go as well. This file provides a basic config struct and initialisation for now (data files are cached in the user's CacheDir). The buffer will attempt to load any files from the cache that exist for a given filename. If these files do not exist, it will create them, if they do it loads them and reloads the buffer data from them. **Reminder:** The runtime should clear any cached files upon a clean exit.
  • cf5ea34f5a Started Search/Substitute implementation Squashed commit of the following: commit ca61676229e12a02969c1b4181b1f60aabba4b39 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:45:07 2024 +0000 began implementing srchSubstitute commit 377b80e568e788432f8de7eace11261141ff0233 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:43:22 2024 +0000 draft impl of flagTokCur commit c099ff43e63c9c01dc3360c3ab38eb74228652b3 Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:34:52 2024 +0000 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time. commit 0dcf32dc02ac266e8a48905bc5296c8fed81f2fe Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:31:59 2024 +0000 All File and Addr references use int64 references now (not uint64). This was just to make it easier to work with the golang stdlib. I would have preferred to keep uint64, since a file index can't be less than 0. if I change it again, I'm making a dedicated type for the Idx.i value, so that I can flip flop on it easier. - Added NewAddrIndex - (a *Addr).Stringify returns "<nil>" if a is nil commit 262aa1c5154ed043a813f9e18aa37bd8dc3df1a5 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 15 13:21:44 2024 +0000 started working on TestSrchSearch Also refactored TestSrchParseArg to use "type input", "type output" style of testing (easier when working with structs). commit 0dd93caf1a2fe6b2f70c957b024bbaae191521b3 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:41:36 2024 +0000 Further developments in srch.go Started adding tests in srch_test.go as well - Added errSrchNoFile - Added SrchCmd returns - Added argument parsing for Srch (srchParseArg) - Added srchSearch commit fbec2c9a59c8ae77824a38f39ea9ac499fdf4c3a Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:38:52 2024 +0000 addr updates - (*Idx).GetFileIndex -> (*Idx).FileIndex - (*Idx).FileIndex, AddrCur is now default index init value - (*Idx).GetIndex -> (*Idx).line - (*Idx).GetCol -> (*Idx).col - Added (*Addr).Size commit 3af2edcfe7e2b734d94d645bfe8b3d32a82783ff Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:55 2024 +0000 scaffolding for SrchCmd commit d5bc43e6de5b26012a1cdb3e39aabedd009926f8 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:42 2024 +0000 spec: reviewed and amended ;s (search)
  • 02f6edd685 Implemented (*Addr).GetFileIndex; added GetTok

1 week ago

gearsix created new branch buf at gearsix/ec

1 week ago

gearsix pushed to main at gearsix/ec

  • ed5db690c2 Define the *Table Data Format* and implemented Buffer.Reload() The *Table Data Format* could just be JSON but this is much more efficient and easier to write a parser for (although everything has a library for parsing JSON anyway I guess). The format contains a line for each piece, where each line contains (in regex: `[ >][ra] \d+ \d+`. The first character indicates if it's the Buffer.pos, the next indicates if the piece `f` value is read/append, the next is a number of the `off` value and then the number for the `siz` value. - renamed `piece.len` -> `piece.siz` (avoid nameclash with len()) - added comment doc at the top of *buf.go* defining the Table Data Format. - added a comment noting that the read file should be cached with a link - renamed `(*Buffer).recover()` -> `(*Buffer).readcache` to avoid nameclash with recover()

1 week ago

gearsix pushed to develop at gearsix/sya

2 weeks ago

gearsix pushed to develop at gearsix/sya

2 weeks ago

gearsix pushed to master at gearsix/scripts

  • 63c236c1db preview-md: added cmark-gfm as supported default renderer

2 weeks ago

gearsix pushed to main at gearsix/xpm

2 weeks ago

gearsix pushed to main at gearsix/xpm

2 weeks ago

gearsix pushed to develop at gearsix/dotfm

  • 771b8a6019 Merge quick patch into dev
  • 21746ca0d7 added install script for users without root access - Split 'install.sh' into 'install-root.sh' and 'install-user.sh' - Updated README to reflect the above - Added TODO item to merge both install scripts
  • View comparison for these 2 commits »

2 weeks ago

gearsix pushed to main at gearsix/ec

  • 02d1ddbde5 started work on buf.go buf.go implements the text buffering system that this program will use. It implements a Piece Table (see file header comment for details), which is originally the reason I started prototyping this tool. Using Golang has made it easy enough for me to use files for a nice caching system of the piece table runtime data, so this commit has added cfg.go as well. This file provides a basic config struct and initialisation for now (data files are cached in the user's CacheDir). The buffer will attempt to load any files from the cache that exist for a given filename. If these files do not exist, it will create them, if they do it loads them and reloads the buffer data from them. **Reminder:** The runtime should clear any cached files upon a clean exit.
  • cf5ea34f5a Started Search/Substitute implementation Squashed commit of the following: commit ca61676229e12a02969c1b4181b1f60aabba4b39 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:45:07 2024 +0000 began implementing srchSubstitute commit 377b80e568e788432f8de7eace11261141ff0233 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:43:22 2024 +0000 draft impl of flagTokCur commit c099ff43e63c9c01dc3360c3ab38eb74228652b3 Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:34:52 2024 +0000 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time. commit 0dcf32dc02ac266e8a48905bc5296c8fed81f2fe Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:31:59 2024 +0000 All File and Addr references use int64 references now (not uint64). This was just to make it easier to work with the golang stdlib. I would have preferred to keep uint64, since a file index can't be less than 0. if I change it again, I'm making a dedicated type for the Idx.i value, so that I can flip flop on it easier. - Added NewAddrIndex - (a *Addr).Stringify returns "<nil>" if a is nil commit 262aa1c5154ed043a813f9e18aa37bd8dc3df1a5 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 15 13:21:44 2024 +0000 started working on TestSrchSearch Also refactored TestSrchParseArg to use "type input", "type output" style of testing (easier when working with structs). commit 0dd93caf1a2fe6b2f70c957b024bbaae191521b3 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:41:36 2024 +0000 Further developments in srch.go Started adding tests in srch_test.go as well - Added errSrchNoFile - Added SrchCmd returns - Added argument parsing for Srch (srchParseArg) - Added srchSearch commit fbec2c9a59c8ae77824a38f39ea9ac499fdf4c3a Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:38:52 2024 +0000 addr updates - (*Idx).GetFileIndex -> (*Idx).FileIndex - (*Idx).FileIndex, AddrCur is now default index init value - (*Idx).GetIndex -> (*Idx).line - (*Idx).GetCol -> (*Idx).col - Added (*Addr).Size commit 3af2edcfe7e2b734d94d645bfe8b3d32a82783ff Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:55 2024 +0000 scaffolding for SrchCmd commit d5bc43e6de5b26012a1cdb3e39aabedd009926f8 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:42 2024 +0000 spec: reviewed and amended ;s (search)
  • 6a55d1892b Started Search/Substitute implementation Squashed commit of the following: commit ca61676229e12a02969c1b4181b1f60aabba4b39 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:45:07 2024 +0000 began implementing srchSubstitute commit 377b80e568e788432f8de7eace11261141ff0233 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:43:22 2024 +0000 draft impl of flagTokCur commit c099ff43e63c9c01dc3360c3ab38eb74228652b3 Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:34:52 2024 +0000 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time. commit 0dcf32dc02ac266e8a48905bc5296c8fed81f2fe Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:31:59 2024 +0000 All File and Addr references use int64 references now (not uint64). This was just to make it easier to work with the golang stdlib. I would have preferred to keep uint64, since a file index can't be less than 0. if I change it again, I'm making a dedicated type for the Idx.i value, so that I can flip flop on it easier. - Added NewAddrIndex - (a *Addr).Stringify returns "<nil>" if a is nil commit 262aa1c5154ed043a813f9e18aa37bd8dc3df1a5 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 15 13:21:44 2024 +0000 started working on TestSrchSearch Also refactored TestSrchParseArg to use "type input", "type output" style of testing (easier when working with structs). commit 0dd93caf1a2fe6b2f70c957b024bbaae191521b3 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:41:36 2024 +0000 Further developments in srch.go Started adding tests in srch_test.go as well - Added errSrchNoFile - Added SrchCmd returns - Added argument parsing for Srch (srchParseArg) - Added srchSearch commit fbec2c9a59c8ae77824a38f39ea9ac499fdf4c3a Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:38:52 2024 +0000 addr updates - (*Idx).GetFileIndex -> (*Idx).FileIndex - (*Idx).FileIndex, AddrCur is now default index init value - (*Idx).GetIndex -> (*Idx).line - (*Idx).GetCol -> (*Idx).col - Added (*Addr).Size commit 3af2edcfe7e2b734d94d645bfe8b3d32a82783ff Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:55 2024 +0000 scaffolding for SrchCmd commit d5bc43e6de5b26012a1cdb3e39aabedd009926f8 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:42 2024 +0000 spec: reviewed and amended ;s (search)
  • View comparison for these 3 commits »

2 weeks ago

gearsix pushed to master at gearsix/scripts

3 weeks ago

gearsix pushed to develop at gearsix/sya

3 weeks ago

gearsix pushed to main at gearsix/ec

  • 6a55d1892b Started Search/Substitute implementation Squashed commit of the following: commit ca61676229e12a02969c1b4181b1f60aabba4b39 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:45:07 2024 +0000 began implementing srchSubstitute commit 377b80e568e788432f8de7eace11261141ff0233 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 23 11:43:22 2024 +0000 draft impl of flagTokCur commit c099ff43e63c9c01dc3360c3ab38eb74228652b3 Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:34:52 2024 +0000 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time. commit 0dcf32dc02ac266e8a48905bc5296c8fed81f2fe Author: gearsix <gearsix@tuta.io> Date: Sun Mar 17 01:31:59 2024 +0000 All File and Addr references use int64 references now (not uint64). This was just to make it easier to work with the golang stdlib. I would have preferred to keep uint64, since a file index can't be less than 0. if I change it again, I'm making a dedicated type for the Idx.i value, so that I can flip flop on it easier. - Added NewAddrIndex - (a *Addr).Stringify returns "<nil>" if a is nil commit 262aa1c5154ed043a813f9e18aa37bd8dc3df1a5 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 15 13:21:44 2024 +0000 started working on TestSrchSearch Also refactored TestSrchParseArg to use "type input", "type output" style of testing (easier when working with structs). commit 0dd93caf1a2fe6b2f70c957b024bbaae191521b3 Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:41:36 2024 +0000 Further developments in srch.go Started adding tests in srch_test.go as well - Added errSrchNoFile - Added SrchCmd returns - Added argument parsing for Srch (srchParseArg) - Added srchSearch commit fbec2c9a59c8ae77824a38f39ea9ac499fdf4c3a Author: gearsix <gearsix@tuta.io> Date: Sat Mar 9 23:38:52 2024 +0000 addr updates - (*Idx).GetFileIndex -> (*Idx).FileIndex - (*Idx).FileIndex, AddrCur is now default index init value - (*Idx).GetIndex -> (*Idx).line - (*Idx).GetCol -> (*Idx).col - Added (*Addr).Size commit 3af2edcfe7e2b734d94d645bfe8b3d32a82783ff Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:55 2024 +0000 scaffolding for SrchCmd commit d5bc43e6de5b26012a1cdb3e39aabedd009926f8 Author: gearsix <gearsix@tuta.io> Date: Fri Mar 1 15:34:42 2024 +0000 spec: reviewed and amended ;s (search)

3 weeks ago

gearsix deleted branch srch at gearsix/ec

3 weeks ago

gearsix pushed to srch at gearsix/ec

  • c099ff43e6 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time.
  • 7ce67eb983 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time.
  • View comparison for these 2 commits »

3 weeks ago

gearsix pushed to srch at gearsix/ec

  • 7ce67eb983 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time.
  • 67304100a9 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time.
  • View comparison for these 2 commits »

1 month ago

gearsix pushed to srch at gearsix/ec

  • 67304100a9 finished SrchSearch tests Turns out the srchSearch didn't work because the regexp.ReaderFindIndex function eats bytes from the Reader after ReaderFindIndex is done, which messed up the absolute indexes given by the function (returned values were relative to the previous finding). The only way to deal with this was to create a new Reader from the new offset (the end of the previously found value) every time.
  • 0dcf32dc02 All File and Addr references use int64 references now (not uint64). This was just to make it easier to work with the golang stdlib. I would have preferred to keep uint64, since a file index can't be less than 0. if I change it again, I'm making a dedicated type for the Idx.i value, so that I can flip flop on it easier. - Added NewAddrIndex - (a *Addr).Stringify returns "<nil>" if a is nil
  • View comparison for these 2 commits »

1 month ago

gearsix pushed to srch at gearsix/ec

  • 262aa1c515 started working on TestSrchSearch Also refactored TestSrchParseArg to use "type input", "type output" style of testing (easier when working with structs).

1 month ago

gearsix pushed to srch at gearsix/ec

  • 0dd93caf1a Further developments in srch.go Started adding tests in srch_test.go as well - Added errSrchNoFile - Added SrchCmd returns - Added argument parsing for Srch (srchParseArg) - Added srchSearch
  • b6b0276d7a Further developments in srch.go Started adding tests in srch_test.go as well - Added errSrchNoFile - Added SrchCmd returns - Added argument parsing for Srch (srchParseArg) - Added srchSearch
  • View comparison for these 2 commits »

1 month ago