gnu1987/gnucmds/sed/sedexec.c

/* sedexec.c -- execute compiled form of stream editor commands

   The single entry point of this module is the function execute(). It
may take a string argument (the name of a file to be used as text)  or
the argument NULL which tells it to filter standard input. It executes
the compiled commands in cmds[] on each line in turn.
   The function command() does most of the work. Match() and advance()
are used for matching text against precompiled regular expressions and
dosub() does right-hand-side substitution.  Getline() does text input;
readout() and memcmp() are output and string-comparison utilities.
   If your environment includes a  memcmp() in the standard libraries,
define the symbol MEMCMP to suppress the version given here,  as yours
is probably hacked to use whatever special string-compare instructions
are available on your hardware and accordingly faster.

==== Written for the GNU programming environment by Eric S. Raymond ==== */

#include <stdio.h>	/* {f}puts, {f}printf, getc/putc, f{re}open, fclose */
#include <ctype.h>	/* for isprint(), isdigit(), toascii() macros */
#include "sed.h"	/* compiled-command structure and various constants */

/* shared variables imported from sedcomp.c */
extern char	linebuf[];	/* current-line buffer */
extern sedcmd	cmds[];		/* hold compiled commands */
extern long	linenum[];	/* numeric-addresses table */
extern int	dflag;		/* -d option flag */
extern int	nflag;		/* -n option flag */
extern int	eargc;		/* scratch copy of argument count */
extern char	bits[];		/* the bits table */

/***** end of imported stuff *****/

#define	MAXHOLD		MAXBUF	/* size of the hold space */
#define GENSIZ		71	/* maximum genbuf size */

#define TRUE		1
#define FALSE		0

/* error messages */
static char TLITL[]	= "sed: line too long\n";
static char COSTF[]	= "sed: can't open text file %s\n";
static char REBAD[]	= "sed: regular expression (internal) error, %o\n";
static char TMAPP[]	= "sed: too many appends after line %ld\n";
static char TMRDS[]	= "sed: too many reads after line %ld\n";

/* pointers and data areas used in RE interpretation */
static char	*tagend[MAXTAGS];	/* tagged pattern start pointers */
static char	*tagstart[MAXTAGS];	/* tagged pattern end pointers */
static char	genbuf[GENSIZ];		/* right-hand-sides expanded here */
static char	*loc1, *loc2, *locs;	/* pointers to pattern-match locs */

/* miscellaneous internals */
static sedcmd	*appends[MAXAPPENDS];	/* array of ptrs to a,i,c commands */
static sedcmd	**aptr = appends;	/* ptr to current append */
static long	lnum = 0L;		/* current source line number */
static char	*spend = linebuf;	/* current end-of-line pointer */

/* command-logic flags */
static int	lastline;	/* do-line flag */
static int	jump;		/* jump to cmd's link address if set */
static int	delete;		/* delete command flag */


void execute(file)
/* execute the compiled commands in cmds[] on a file */
char *file;		/* name of text source file to be filtered */
{
	register sedcmd	*ipc;			/* ptr to current command */
	static sedcmd	*pending = NULL;	/* next cmd to be executed */
	sedcmd		*command();		/* main command executive */
	char		*newspend;		/* ptr to source, linebuf */
	char		*getline();		/* input-getting functions */

	if (file != NULL)		/* filter text from a named file */ 
		if (freopen(file, "r", stdin) == NULL)
			fprintf(stderr, COSTF, file);

	if (pending)			/* there's a command waiting */
	{
		ipc = pending;		/* it will be first executed */
		pending = NULL;		/*   clear the waiting ptr */
		goto doit;		/*   and go to execute it immediately */
	}

	/* here's the main command-execution loop */
	while ((newspend = getline(linebuf)) != BAD)
	{
		spend = newspend;	/*   update buffer-end ptr */

		/* loop through compiled commands, executing them */
		for(ipc = cmds; ipc->command; ipc++)
		{
			if (!selected(ipc))
				continue;

			/* execute the command pointed at */
	doit:		pending = command(ipc);

			if (delete)	/* if delete flag is set */
				break;	/* don't exec rest of compiled cmds */

			if (jump)	/* if jump set, follow cmd's link */
			{
				jump = FALSE;
				if ((ipc = ipc->u.link) == NULL)
				{
					ipc = cmds;	/* restart commands */
					break;
				}
				else
					ipc--;	/* so ipc++ won't screw us */
			}

		}
		/* we've now done all modification commands on the line */

		/* here's where the transformed line is output */
		if (!nflag && !delete)
		{
			for(newspend = linebuf; newspend < spend; newspend++)
				putc(*newspend, stdout);
			putc('\n', stdout);
		}

		/* if we've been set up for append, emit the text from it */
		if (aptr > appends)
			readout();

		delete = FALSE;	/* clear delete flag; about to get next cmd */
	}
}


static int selected(ipc)	/* uses lnum, linenum */
/* return TRUE if the command is currently selected, FALSE otherwise */
sedcmd	*ipc;
{
	register char	*start_at = ipc->addr1;	/* point p1 at first address */
	register char	*end_at   = ipc->addr2;	/*   and p2 at second */
	int		c;			/* scratch character holder */

	if (start_at)
	{
		if (ipc->flags.active)		/* command is selected */
		{
			if (*end_at == CEND)
				start_at = NULL;
			else if (*end_at == CLNUM)
			{
				if (lnum > linenum[end_at[1]])
				{
					ipc->flags.active = FALSE;
					return(ipc->flags.allbut);
				}
				if (lnum == linenum[c])
					ipc->flags.active = FALSE;
			}
			else if (match(linebuf, end_at, FALSE))
				ipc->flags.active = FALSE;
		}
		else if (*start_at == CEND)
		{
			if (!lastline)
				return(ipc->flags.allbut);
		}
		else if (*start_at == CLNUM)
		{
			if (lnum != linenum[start_at[1]])
				return(ipc->flags.allbut);
			ipc->flags.active = (end_at != NULL);
		}
		else if (match(linebuf, start_at, FALSE))
			ipc->flags.active = (end_at != NULL);
		else
			return(ipc->flags.allbut);
	}

	/* skip selected command if flags.allbut is on */
	return (!ipc->flags.allbut);
}

static int match(lp, ep, gf)		/* uses genbuf */
/* match RE at ep... against lp...; if gf set, copy lp... from genbuf first */
register char	*lp, *ep;
int		gf;
{
	char	*scp;
	char	c;

	if (gf)
	{
		if (*ep)
			return(FALSE);
		strcpy(lp, genbuf);
		locs = lp = loc2;
	}
	else
		locs = NULL;

	if (*ep++)
	{
		loc1 = lp;
		if(*ep == CCHR && ep[1] != *lp)	/* 1st char is wrong */
			return(FALSE);		/*   so fail */
		return(advance(lp, ep));	/* else try to match rest */
	}

	/* quick check for 1st character if it's literal */
	if (*ep == CCHR)
	{
		c = ep[1];		/* pull out character to search for */
		do {
			if (*lp != c)
				continue;	/* scan the source string */
			if (advance(lp, ep))	/* found it, match the rest */
				return(loc1 = lp, 1);
		} while
			(*lp++);
		return(FALSE);		/* didn't find that first char */
	}

	/* else try for unanchored match of the pattern */
	do {
		if (advance(lp, ep))
			return(loc1 = lp, TRUE);
	} while
		(*lp++);

	/* if got here, didn't match either way */
	return(FALSE);
}

static int advance(lp, ep)
/* attempt to advance match pointer by one pattern element */
register char	*lp;		/* source (linebuf) ptr */
register char	*ep;		/* regular expression element ptr */
{
	register char	*curlp;		/* save ptr for closures */ 
	char		c;		/* scratch character holder */
	char		*bbeg;
	int		ct;

	for (;;)
		switch (*ep++)
		{
		case CCHR:		/* literal character */
			if (*ep++ == *lp++)	/* if chars are equal */
				continue;	/* matched */
			return(FALSE);		/* else return false */

		case CDOT:		/* anything but newline */
			if (*lp++)		/* first NUL is at EOL */
				continue;	/* keep going if didn't find */
			return(FALSE);		/* else return false */

		case CNL:		/* start-of-line */
		case CDOL:		/* end-of-line */
			if (*lp == 0)		/* found that first NUL? */
				continue;	/* yes, keep going */
			return(FALSE);		/* else return false */

		case CEOF:		/* end-of-address mark */
			loc2 = lp;		/* set second loc */
			return(TRUE);		/* return true */

		case CCL:		/* a character class */
			c = *lp++ & 0177;
			if (ep[c>>3] & bits[c & 07])	/* is char in set? */
			{
				ep += 16;	/* then skip rest of bitmask */
				continue;	/*   and keep going */
			}
			return(FALSE);		/* else return false */

		case CBRA:		/* start of tagged pattern */
			tagstart[*ep++] = lp;	/* mark it */
			continue;		/* and go */

		case CKET:		/* end of tagged pattern */
			tagend[*ep++] = lp;	/* mark it */
			continue;		/* and go */

		case CBACK: 		/* a pattern tag reference */
			bbeg = tagstart[*ep];		/* find the start */
			ct = tagend[*ep++] - bbeg;	/* and the length */
			if (memcmp(bbeg, lp, ct))	/* matching text */
			{
				lp += ct;	/* yes, look past end */
				continue;	/* and keep going */
			}
			return(FALSE);			/* else match failed */

		case CBACK|STAR:	/* pattern tag with Kleene star */
			bbeg = tagstart[*ep];		/* get start */
			ct = tagend[*ep++] - bbeg;	/* and length */
			curlp = lp;			/* save the location */
			while(memcmp(bbeg, lp, ct))	/* while we match */
				lp += ct;		/*  go to next one */

			while(lp >= curlp)	/* while 1 match left */
			{
				if (advance(lp, ep))	/* try rest of RE */
					return(TRUE);	/* if matched, O.K. */
				lp -= ct;		/* else backtrack */
			}
			return(FALSE);		/* no matches left, fail */


		case CDOT|STAR:		/* match .* */
			curlp = lp;		/* save closure start loc */
			while (*lp++);		/* match anything */ 
			goto star;		/* now look for followers */

		case CCHR|STAR:		/* match <literal char>* */
			curlp = lp;		/* save closure start loc */
			while (*lp++ == *ep);	/* match many of that char */
			ep++;			/* to start of next element */
			goto star;		/* match it and followers */

		case CCL|STAR:		/* match [...]* */
			curlp = lp;		/* save closure start loc */
			do {
				c = *lp++ & 0x7F;	/* match any in set */
			} while
				(ep[c>>3] & bits[c & 07]);
			ep += 16;		/* skip past the set */
			goto star;		/* match followers */

		star:		/* the recursion part of a * or + match */
			if (--lp == curlp)	/* 0 matches */
				continue;

			if (*ep == CCHR)
			{
				c = ep[1];
				do {
					if (*lp != c)
						continue;
					if (advance(lp, ep))
						return(TRUE);
				} while
					(lp-- > curlp);
				return(FALSE);
			}

			if (*ep == CBACK)
			{
				c = *(tagstart[ep[1]]);
				do {
					if (*lp != c)
						continue;
					if (advance(lp, ep))
						return(TRUE);
				} while
					(lp-- > curlp);
				return(FALSE);
			}
	
			do {
				if (lp == locs)
					break;
				if (advance(lp, ep))
					return(TRUE);
			} while
				(lp-- > curlp);
			return(FALSE);

		default:
			fprintf(stderr, REBAD, *--ep);
		}
}

static int substitute(lhs, rhs, gf)	/* uses linebuf, spend */
/* perform s command */
char	*lhs, *rhs;			/* left and right sides of command */
int	gf;				/* the global-substitute flag */
{
	char *dosub();			/* for if we find a match */

	if (match(linebuf, lhs, FALSE))		/* if 1 match */
		spend = dosub(linebuf, rhs);	/* perform substitution once */
	else
		return(FALSE);		/* command fails */

	while(gf && match(linebuf, lhs, TRUE))	/* cycle through possibles */
		spend = dosub(linebuf, rhs);	/* substitute each one */

	return(TRUE);			/* we succeeded */
}

static char *dosub(linep, rhsbuf)	/* uses genbuf, loc1, loc2 */
/* generate substituted right-hand side (of s command) */
char	*linep;		/* line buffer to substitute in */
char	*rhsbuf;	/* where to put the result */
{
	register char	*lp, *sp, *rp;
	int		c;
	char		*place();

	/* copy linebuf to genbuf up to location  1 */
	sp = genbuf; lp = linep; while (lp < loc1) *sp++ = *lp++;

	/* insert substitute right-hand-side with tags expanded in genbuf */
	for (rp = rhsbuf; c = *rp++; )
	{
		if (c == '&')
			sp = place(sp, loc1, loc2);
		else if (c & TAGMARK && (c &= 0x7F) >= '1' && c < '1'+MAXTAGS)
			sp = place(sp, tagstart[c - '1'], tagend[c - '1']);
		else if (c & TAGMARK)
			fprintf(stderr, "sed: bad tag value %x\n", c);
		else
			*sp++ = toascii(c);

		if (sp >= genbuf + GENSIZ)
			fprintf(stderr, TLITL);
	}

	/* copy the part of the line after the substituted area */
	lp = loc2;
	loc2 = sp - genbuf + linep;
	while (*sp++ = *lp++)
		if (sp >= genbuf + GENSIZ)
			fprintf(stderr, TLITL);

	/* copy the substituted pattern from genbuf to linebuf */
	lp = linep; sp = genbuf; while (*lp++ = *sp++) continue;
	return(lp - 1);
}

static char *place(asp, al1, al2)		/* uses genbuf */
/* place chars at *al1...*(al1 - 1) at asp... in genbuf[] */
register char	*asp, *al1, *al2;
{
	while (al1 < al2 && asp < genbuf + GENSIZ)
		*asp++ = *al1++;
	return(asp);
}

static void listto(p1, fp)
/* write a hex dump expansion of *p1... to fp */
register char	*p1;		/* the source */
FILE		*fp;		/* output stream to write to */
{
	p1--;
	while(*p1++)
		if (isprint(*p1))
			putc(*p1, fp);		/* pass it through */
		else
		{
			putc('\134', fp);		/* emit a backslash */
			switch(*p1)
			{
			case '\10':	putc('b', fp); break;	/* BS */
			case '\09':	putc('t', fp); break;	/* TAB */
			case '\12':	putc('n', fp); break;	/* NL */
			case '\15':	putc('r', fp); break;	/* CR */
			case '\33':	putc('e', fp); break;	/* ESC */
			default:	fprintf(fp, "%02x", *p1 & 0xFF);
			}
		}
	putc('\n', fp);
}


static void dumpline(ipc)
/* execute p, P, w, W commands and options */
sedcmd	*ipc;
{
	register char	*p;

	if (ipc->flags.firstl)
		for(p = linebuf; *p != '\n' && *p != '\0'; )
			putc(*p++, ipc->fout);
	else
		fputs(linebuf, ipc->fout);
	putc('\n', ipc->fout);
}

static sedcmd *command(ipc)
/* execute compiled command pointed at by ipc */
sedcmd	*ipc;
{
	register char	*p1, *p2;		/* scratch pointers */
	static char	holdsp[MAXHOLD];	/* the hold space */
	static char	*hspend = holdsp;	/* hold space end pointer */
	static int	sflag;			/* true if last s succeeded */
	char		c, *newspend;

	if (dflag)
		fprintf(stderr,
			"sed: executing %x on arguments \"%s\" and \"%s\"\n",
			ipc->command, ipc->addr1, ipc->addr2);

	switch(ipc->command)
	{
	case ACMD:		/* append */
		*aptr++ = ipc;
		if (aptr >= appends + MAXAPPENDS)
			fprintf(stderr, TMAPP, lnum);
		*aptr = 0;
		break;

	case CCMD:		/* change pattern space */
		if (!ipc->flags.active || lastline)
			puts(ipc->u.lhs);		
	case DCMD:		/* delete pattern space */
		delete = TRUE;
		break;

	case CDCMD:		/* delete a line in pattern space */
		p1 = p2 = linebuf;
		while(*p1 != '\n')
			if (delete = (*p1++ == 0))
				return;
		p1++;
		while(*p2++ = *p1++);
		spend = p2-1;
		break;

	case EQCMD:		/* show current line number */
		fprintf(stdout, "%ld\n", lnum);
		break;

	case GCMD:		/* copy hold space to pattern space */
		p1 = linebuf; p2 = holdsp; while (*p1++ = *p2++);
		spend = p1-1;
		break;

	case CGCMD:		/* append hold space to pattern space */
		*spend++ = '\n';
		p1 = spend;	p2 = holdsp;
		while(*p1++ = *p2++)
			if (p1 >= linebuf + MAXBUF)
				break;
		spend = p1-1;
		break;

	case HCMD:		/* copy pattern space to hold space */
		p1 = holdsp;	p2 = linebuf;	while(*p1++ = *p2++);
		hspend = p1-1;
		break;

	case CHCMD:		/* append pattern space to hold space */
		*hspend++ = '\n';
		p1 = hspend;	p2 = linebuf;
		while(*p1++ = *p2++)
			if (p1 >= holdsp + MAXHOLD)
				break;
		hspend = p1-1;
		break;

	case ICMD:		/* insert text */
		printf("%s\n", ipc->u.lhs);
		break;

	case LCMD:		/* list text */
		listto(linebuf, (ipc->fout != NULL)?ipc->fout:stdout);
		break;

	case NCMD:	/* read next line into pattern space */
		if (!nflag)
			puts(linebuf);	/* flush out the current line */
		if (aptr > appends)
			readout();	/* do pending a, r commands */
		if ((newspend = getline(linebuf)) == BAD)
			return(delete = 1, ipc);
		spend = newspend;
		break;

	case CNCMD:	/* append next line to pattern space */
		if (aptr > appends)
			readout();
		*spend++ = '\n';
		if ((newspend = getline(spend)) == BAD)
			return(delete = 1, ipc);
		spend = newspend;
		break;

	case PCMD:		/* print pattern space */
	case CPCMD:		/* print one line from pattern space */
	case WCMD:		/* write pattern space to file */
	case CWCMD:		/* write one line from pattern space */
		dumpline(ipc);
		break;

	case QCMD:		/* quit the stream editor */
		if (!nflag)
			puts(linebuf);	/* flush out the current line */
		if (aptr > appends)
			readout();	/* do any pending a and r commands */
		exit(0);

	case RCMD:		/* read a file into the stream */
		*aptr++ = ipc;
		if (aptr >= appends + MAXAPPENDS)
			fprintf(stderr,	TMRDS, lnum);
		*aptr = 0;
		break;

	case SCMD:		/* substitute RE */
		sflag = substitute(ipc->u.lhs, ipc->rhs, ipc->flags.global);
		if (sflag)
			dumpline(ipc);
		break;

	case TCMD:		/* branch on last s successful */
	case CTCMD:		/* branch on last s failed */
		if (sflag == (ipc->command == CTCMD))
			break;		/* no branch if last s failed, else */
		sflag = FALSE;		/*   clear the s condition flag */
	case BCMD:		/* branch to label */
		jump = TRUE;		/*   set up to jump to assoc'd label */
		break;

	case XCMD:		/* exchange pattern and hold spaces */
		p1 = linebuf, p2 = holdsp;
		while (p1 <= spend || p2 <= hspend)
		{
			c = *p1; *p1++ = *p2; *p2++ = c;
		}
		newspend = hspend; hspend = spend; spend = newspend;
		break;

	case YCMD:		/* translate a line */
		p1 = linebuf;	p2 = ipc->u.lhs;
		while(*p1 = p2[*p1])
			p1++;
		break;
	}
	return(NULL);
}

static char *getline(buf)	/* uses lastline, eargc */
/* get next line of text to be filtered */
register char	*buf;		/* where to send the input */
{
	if (gets(buf) != NULL)
	{
		lnum++;			/* note that we got another line */
		while(*buf++);		/* find the end of the input */
		return(--buf);		/* return ptr to terminating null */ 
	}
	else
	{
		if (eargc == 0)			/* if there are no more args */
			lastline = TRUE;	/*    set a flag */
		return(BAD);
	}
}

#ifndef	MEMCMP
static int memcmp(a, b, count)
/* return 1 if *a... == *b... for count chars, 0 otherwise */
register char	*a, *b;
{
	while(count--)			/* look at count characters */
		if (*a++ != *b++)	/* if any are nonequal   */
			return(0);	/*    return 0 for false */
	return(TRUE);			/* compare succeeded */
}
#endif	MEMCMP

static void readout()		/* uses appends, aptr */
/* write file indicated by r command to output */
{
	register char	*p1;	/* character-fetching dummy */
	register int	t;	/* hold input char or EOF */
	FILE		*fi;	/* ptr to file to be read */

	aptr = appends - 1;	/* arrange for pre-increment to work right */
	while(*++aptr)
		if ((*aptr)->command == ACMD)		/* process "a" cmd */
			printf("%s\n", (*aptr)->u.lhs);
		else					/* process "r" cmd */
		{
			if ((fi = fopen((*aptr)->u.lhs, "r")) == NULL)
				continue;
			while((t = getc(fi)) != EOF)
				putc((char) t, stdout);
			fclose(fi);
		}
	aptr = appends;		/* reset the append ptr */
	*aptr = 0;
}

/* sedexec.c ends here */