Annotation of src/usr.bin/sed/process.c, Revision 1.1.1.1
1.1 deraadt 1: /*-
2: * Copyright (c) 1992 Diomidis Spinellis.
3: * Copyright (c) 1992, 1993
4: * The Regents of the University of California. All rights reserved.
5: *
6: * This code is derived from software contributed to Berkeley by
7: * Diomidis Spinellis of Imperial College, University of London.
8: *
9: * Redistribution and use in source and binary forms, with or without
10: * modification, are permitted provided that the following conditions
11: * are met:
12: * 1. Redistributions of source code must retain the above copyright
13: * notice, this list of conditions and the following disclaimer.
14: * 2. Redistributions in binary form must reproduce the above copyright
15: * notice, this list of conditions and the following disclaimer in the
16: * documentation and/or other materials provided with the distribution.
17: * 3. All advertising materials mentioning features or use of this software
18: * must display the following acknowledgement:
19: * This product includes software developed by the University of
20: * California, Berkeley and its contributors.
21: * 4. Neither the name of the University nor the names of its contributors
22: * may be used to endorse or promote products derived from this software
23: * without specific prior written permission.
24: *
25: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35: * SUCH DAMAGE.
36: */
37:
38: #ifndef lint
39: /* from: static char sccsid[] = "@(#)process.c 8.1 (Berkeley) 6/6/93"; */
40: static char *rcsid = "$Id: process.c,v 1.17 1995/07/11 04:09:50 cgd Exp $";
41: #endif /* not lint */
42:
43: #include <sys/types.h>
44: #include <sys/stat.h>
45: #include <sys/ioctl.h>
46: #include <sys/uio.h>
47:
48: #include <ctype.h>
49: #include <errno.h>
50: #include <fcntl.h>
51: #include <limits.h>
52: #include <regex.h>
53: #include <stdio.h>
54: #include <stdlib.h>
55: #include <string.h>
56: #include <unistd.h>
57:
58: #include "defs.h"
59: #include "extern.h"
60:
61: static SPACE HS, PS, SS;
62: #define pd PS.deleted
63: #define ps PS.space
64: #define psl PS.len
65: #define hs HS.space
66: #define hsl HS.len
67:
68: static inline int applies __P((struct s_command *));
69: static void flush_appends __P((void));
70: static void lputs __P((char *));
71: static inline int regexec_e __P((regex_t *, const char *, int, int, size_t));
72: static void regsub __P((SPACE *, char *, char *));
73: static int substitute __P((struct s_command *));
74:
75: struct s_appends *appends; /* Array of pointers to strings to append. */
76: static int appendx; /* Index into appends array. */
77: int appendnum; /* Size of appends array. */
78:
79: static int lastaddr; /* Set by applies if last address of a range. */
80: static int sdone; /* If any substitutes since last line input. */
81: /* Iov structure for 'w' commands. */
82: static regex_t *defpreg;
83: size_t maxnsub;
84: regmatch_t *match;
85:
86: #define OUT(s) { fwrite(s, sizeof(u_char), psl, stdout); }
87:
88: void
89: process()
90: {
91: struct s_command *cp;
92: SPACE tspace;
93: size_t len, oldpsl;
94: char *p;
95:
96: for (linenum = 0; mf_fgets(&PS, REPLACE);) {
97: pd = 0;
98: top:
99: cp = prog;
100: redirect:
101: while (cp != NULL) {
102: if (!applies(cp)) {
103: cp = cp->next;
104: continue;
105: }
106: switch (cp->code) {
107: case '{':
108: cp = cp->u.c;
109: goto redirect;
110: case 'a':
111: if (appendx >= appendnum)
112: appends = xrealloc(appends,
113: sizeof(struct s_appends) *
114: (appendnum *= 2));
115: appends[appendx].type = AP_STRING;
116: appends[appendx].s = cp->t;
117: appends[appendx].len = strlen(cp->t);
118: appendx++;
119: break;
120: case 'b':
121: cp = cp->u.c;
122: goto redirect;
123: case 'c':
124: pd = 1;
125: psl = 0;
126: if (cp->a2 == NULL || lastaddr)
127: (void)printf("%s", cp->t);
128: break;
129: case 'd':
130: pd = 1;
131: goto new;
132: case 'D':
133: if (pd)
134: goto new;
135: if ((p = memchr(ps, '\n', psl - 1)) == NULL) {
136: pd = 1;
137: goto new;
138: } else {
139: psl -= (p + 1) - ps;
140: memmove(ps, p + 1, psl);
141: goto top;
142: }
143: case 'g':
144: cspace(&PS, hs, hsl, REPLACE);
145: break;
146: case 'G':
147: cspace(&PS, hs, hsl, 0);
148: break;
149: case 'h':
150: cspace(&HS, ps, psl, REPLACE);
151: break;
152: case 'H':
153: cspace(&HS, ps, psl, 0);
154: break;
155: case 'i':
156: (void)printf("%s", cp->t);
157: break;
158: case 'l':
159: lputs(ps);
160: break;
161: case 'n':
162: if (!nflag && !pd)
163: OUT(ps)
164: flush_appends();
165: if (!mf_fgets(&PS, REPLACE))
166: exit(0);
167: pd = 0;
168: break;
169: case 'N':
170: flush_appends();
171: if (!mf_fgets(&PS, 0)) {
172: if (!nflag && !pd)
173: OUT(ps)
174: exit(0);
175: }
176: break;
177: case 'p':
178: if (pd)
179: break;
180: OUT(ps)
181: break;
182: case 'P':
183: if (pd)
184: break;
185: if ((p = memchr(ps, '\n', psl - 1)) != NULL) {
186: oldpsl = psl;
187: psl = (p + 1) - ps;
188: }
189: OUT(ps)
190: if (p != NULL)
191: psl = oldpsl;
192: break;
193: case 'q':
194: if (!nflag && !pd)
195: OUT(ps)
196: flush_appends();
197: exit(0);
198: case 'r':
199: if (appendx >= appendnum)
200: appends = xrealloc(appends,
201: sizeof(struct s_appends) *
202: (appendnum *= 2));
203: appends[appendx].type = AP_FILE;
204: appends[appendx].s = cp->t;
205: appends[appendx].len = strlen(cp->t);
206: appendx++;
207: break;
208: case 's':
209: sdone |= substitute(cp);
210: break;
211: case 't':
212: if (sdone) {
213: sdone = 0;
214: cp = cp->u.c;
215: goto redirect;
216: }
217: break;
218: case 'w':
219: if (pd)
220: break;
221: if (cp->u.fd == -1 && (cp->u.fd = open(cp->t,
222: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
223: DEFFILEMODE)) == -1)
224: err(FATAL, "%s: %s\n",
225: cp->t, strerror(errno));
226: if (write(cp->u.fd, ps, psl) != psl)
227: err(FATAL, "%s: %s\n",
228: cp->t, strerror(errno));
229: break;
230: case 'x':
231: if (hs == NULL)
232: cspace(&HS, "\n", 1, REPLACE);
233: tspace = PS;
234: PS = HS;
235: HS = tspace;
236: break;
237: case 'y':
238: if (pd)
239: break;
240: for (p = ps, len = psl; --len; ++p)
241: *p = cp->u.y[*p];
242: break;
243: case ':':
244: case '}':
245: break;
246: case '=':
247: (void)printf("%lu\n", linenum);
248: }
249: cp = cp->next;
250: } /* for all cp */
251:
252: new: if (!nflag && !pd)
253: OUT(ps)
254: flush_appends();
255: } /* for all lines */
256: }
257:
258: /*
259: * TRUE if the address passed matches the current program state
260: * (lastline, linenumber, ps).
261: */
262: #define MATCH(a) \
263: (a)->type == AT_RE ? regexec_e((a)->u.r, ps, 0, 1, psl) : \
264: (a)->type == AT_LINE ? linenum == (a)->u.l : lastline
265:
266: /*
267: * Return TRUE if the command applies to the current line. Sets the inrange
268: * flag to process ranges. Interprets the non-select (``!'') flag.
269: */
270: static inline int
271: applies(cp)
272: struct s_command *cp;
273: {
274: int r;
275:
276: lastaddr = 0;
277: if (cp->a1 == NULL && cp->a2 == NULL)
278: r = 1;
279: else if (cp->a2)
280: if (cp->inrange) {
281: if (MATCH(cp->a2)) {
282: cp->inrange = 0;
283: lastaddr = 1;
284: }
285: r = 1;
286: } else if (MATCH(cp->a1)) {
287: /*
288: * If the second address is a number less than or
289: * equal to the line number first selected, only
290: * one line shall be selected.
291: * -- POSIX 1003.2
292: */
293: if (cp->a2->type == AT_LINE &&
294: linenum >= cp->a2->u.l)
295: lastaddr = 1;
296: else
297: cp->inrange = 1;
298: r = 1;
299: } else
300: r = 0;
301: else
302: r = MATCH(cp->a1);
303: return (cp->nonsel ? ! r : r);
304: }
305:
306: /*
307: * substitute --
308: * Do substitutions in the pattern space. Currently, we build a
309: * copy of the new pattern space in the substitute space structure
310: * and then swap them.
311: */
312: static int
313: substitute(cp)
314: struct s_command *cp;
315: {
316: SPACE tspace;
317: regex_t *re;
318: size_t re_off, slen;
319: int n, lastempty;
320: char *s;
321:
322: s = ps;
323: re = cp->u.s->re;
324: if (re == NULL) {
325: if (defpreg != NULL && cp->u.s->maxbref > defpreg->re_nsub) {
326: linenum = cp->u.s->linenum;
327: err(COMPILE, "\\%d not defined in the RE",
328: cp->u.s->maxbref);
329: }
330: }
331: if (!regexec_e(re, s, 0, 0, psl))
332: return (0);
333:
334: SS.len = 0; /* Clean substitute space. */
335: slen = psl;
336: n = cp->u.s->n;
337: lastempty = 1;
338:
339: switch (n) {
340: case 0: /* Global */
341: do {
342: if (lastempty || match[0].rm_so != match[0].rm_eo) {
343: /* Locate start of replaced string. */
344: re_off = match[0].rm_so;
345: /* Copy leading retained string. */
346: cspace(&SS, s, re_off, APPEND);
347: /* Add in regular expression. */
348: regsub(&SS, s, cp->u.s->new);
349: }
350:
351: /* Move past this match. */
352: if (match[0].rm_so != match[0].rm_eo) {
353: s += match[0].rm_eo;
354: slen -= match[0].rm_eo;
355: lastempty = 0;
356: } else {
357: if (match[0].rm_so == 0)
358: cspace(&SS, s, match[0].rm_so + 1,
359: APPEND);
360: else
361: cspace(&SS, s + match[0].rm_so, 1,
362: APPEND);
363: s += match[0].rm_so + 1;
364: slen -= match[0].rm_so + 1;
365: lastempty = 1;
366: }
367: } while (slen > 0 && regexec_e(re, s, REG_NOTBOL, 0, slen));
368: /* Copy trailing retained string. */
369: if (slen > 0)
370: cspace(&SS, s, slen, APPEND);
371: break;
372: default: /* Nth occurrence */
373: while (--n) {
374: s += match[0].rm_eo;
375: slen -= match[0].rm_eo;
376: if (!regexec_e(re, s, REG_NOTBOL, 0, slen))
377: return (0);
378: }
379: /* FALLTHROUGH */
380: case 1: /* 1st occurrence */
381: /* Locate start of replaced string. */
382: re_off = match[0].rm_so + (s - ps);
383: /* Copy leading retained string. */
384: cspace(&SS, ps, re_off, APPEND);
385: /* Add in regular expression. */
386: regsub(&SS, s, cp->u.s->new);
387: /* Copy trailing retained string. */
388: s += match[0].rm_eo;
389: slen -= match[0].rm_eo;
390: cspace(&SS, s, slen, APPEND);
391: break;
392: }
393:
394: /*
395: * Swap the substitute space and the pattern space, and make sure
396: * that any leftover pointers into stdio memory get lost.
397: */
398: tspace = PS;
399: PS = SS;
400: SS = tspace;
401: SS.space = SS.back;
402:
403: /* Handle the 'p' flag. */
404: if (cp->u.s->p)
405: OUT(ps)
406:
407: /* Handle the 'w' flag. */
408: if (cp->u.s->wfile && !pd) {
409: if (cp->u.s->wfd == -1 && (cp->u.s->wfd = open(cp->u.s->wfile,
410: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, DEFFILEMODE)) == -1)
411: err(FATAL, "%s: %s\n", cp->u.s->wfile, strerror(errno));
412: if (write(cp->u.s->wfd, ps, psl) != psl)
413: err(FATAL, "%s: %s\n", cp->u.s->wfile, strerror(errno));
414: }
415: return (1);
416: }
417:
418: /*
419: * Flush append requests. Always called before reading a line,
420: * therefore it also resets the substitution done (sdone) flag.
421: */
422: static void
423: flush_appends()
424: {
425: FILE *f;
426: int count, i;
427: char buf[8 * 1024];
428:
429: for (i = 0; i < appendx; i++)
430: switch (appends[i].type) {
431: case AP_STRING:
432: fwrite(appends[i].s, sizeof(char), appends[i].len,
433: stdout);
434: break;
435: case AP_FILE:
436: /*
437: * Read files probably shouldn't be cached. Since
438: * it's not an error to read a non-existent file,
439: * it's possible that another program is interacting
440: * with the sed script through the file system. It
441: * would be truly bizarre, but possible. It's probably
442: * not that big a performance win, anyhow.
443: */
444: if ((f = fopen(appends[i].s, "r")) == NULL)
445: break;
446: while (count = fread(buf, sizeof(char), sizeof(buf), f))
447: (void)fwrite(buf, sizeof(char), count, stdout);
448: (void)fclose(f);
449: break;
450: }
451: if (ferror(stdout))
452: err(FATAL, "stdout: %s", strerror(errno ? errno : EIO));
453: appendx = sdone = 0;
454: }
455:
456: static void
457: lputs(s)
458: register char *s;
459: {
460: register int count;
461: register char *escapes, *p;
462: struct winsize win;
463: static int termwidth = -1;
464:
465: if (termwidth == -1)
466: if (p = getenv("COLUMNS"))
467: termwidth = atoi(p);
468: else if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &win) == 0 &&
469: win.ws_col > 0)
470: termwidth = win.ws_col;
471: else
472: termwidth = 60;
473:
474: for (count = 0; *s; ++s) {
475: if (count >= termwidth) {
476: (void)printf("\\\n");
477: count = 0;
478: }
479: if (isascii(*s) && isprint(*s) && *s != '\\') {
480: (void)putchar(*s);
481: count++;
482: } else {
483: escapes = "\\\a\b\f\n\r\t\v";
484: (void)putchar('\\');
485: if (p = strchr(escapes, *s)) {
486: (void)putchar("\\abfnrtv"[p - escapes]);
487: count += 2;
488: } else {
489: (void)printf("%03o", *(u_char *)s);
490: count += 4;
491: }
492: }
493: }
494: (void)putchar('$');
495: (void)putchar('\n');
496: if (ferror(stdout))
497: err(FATAL, "stdout: %s", strerror(errno ? errno : EIO));
498: }
499:
500: static inline int
501: regexec_e(preg, string, eflags, nomatch, slen)
502: regex_t *preg;
503: const char *string;
504: int eflags, nomatch;
505: size_t slen;
506: {
507: int eval;
508:
509: if (preg == NULL) {
510: if (defpreg == NULL)
511: err(FATAL, "first RE may not be empty");
512: } else
513: defpreg = preg;
514:
515: /* Set anchors, discounting trailing newline (if any). */
516: if (slen > 0 && string[slen - 1] == '\n')
517: slen--;
518: match[0].rm_so = 0;
519: match[0].rm_eo = slen;
520:
521: eval = regexec(defpreg, string,
522: nomatch ? 0 : maxnsub + 1, match, eflags | REG_STARTEND);
523: switch(eval) {
524: case 0:
525: return (1);
526: case REG_NOMATCH:
527: return (0);
528: }
529: err(FATAL, "RE error: %s", strregerror(eval, defpreg));
530: /* NOTREACHED */
531: }
532:
533: /*
534: * regsub - perform substitutions after a regexp match
535: * Based on a routine by Henry Spencer
536: */
537: static void
538: regsub(sp, string, src)
539: SPACE *sp;
540: char *string, *src;
541: {
542: register int len, no;
543: register char c, *dst;
544:
545: #define NEEDSP(reqlen) \
546: if (sp->len >= sp->blen - (reqlen) - 1) { \
547: sp->blen += (reqlen) + 1024; \
548: sp->space = sp->back = xrealloc(sp->back, sp->blen); \
549: dst = sp->space + sp->len; \
550: }
551:
552: dst = sp->space + sp->len;
553: while ((c = *src++) != '\0') {
554: if (c == '&')
555: no = 0;
556: else if (c == '\\' && isdigit(*src))
557: no = *src++ - '0';
558: else
559: no = -1;
560: if (no < 0) { /* Ordinary character. */
561: if (c == '\\' && (*src == '\\' || *src == '&'))
562: c = *src++;
563: NEEDSP(1);
564: *dst++ = c;
565: ++sp->len;
566: } else if (match[no].rm_so != -1 && match[no].rm_eo != -1) {
567: len = match[no].rm_eo - match[no].rm_so;
568: NEEDSP(len);
569: memmove(dst, string + match[no].rm_so, len);
570: dst += len;
571: sp->len += len;
572: }
573: }
574: NEEDSP(1);
575: *dst = '\0';
576: }
577:
578: /*
579: * aspace --
580: * Append the source space to the destination space, allocating new
581: * space as necessary.
582: */
583: void
584: cspace(sp, p, len, spflag)
585: SPACE *sp;
586: char *p;
587: size_t len;
588: enum e_spflag spflag;
589: {
590: size_t tlen;
591:
592: /* Make sure SPACE has enough memory and ramp up quickly. */
593: tlen = sp->len + len + 1;
594: if (tlen > sp->blen) {
595: sp->blen = tlen + 1024;
596: sp->space = sp->back = xrealloc(sp->back, sp->blen);
597: }
598:
599: if (spflag == REPLACE)
600: sp->len = 0;
601:
602: memmove(sp->space + sp->len, p, len);
603:
604: sp->space[sp->len += len] = '\0';
605: }
606:
607: /*
608: * Close all cached opened files and report any errors
609: */
610: void
611: cfclose(cp, end)
612: register struct s_command *cp, *end;
613: {
614:
615: for (; cp != end; cp = cp->next)
616: switch(cp->code) {
617: case 's':
618: if (cp->u.s->wfd != -1 && close(cp->u.s->wfd))
619: err(FATAL,
620: "%s: %s", cp->u.s->wfile, strerror(errno));
621: cp->u.s->wfd = -1;
622: break;
623: case 'w':
624: if (cp->u.fd != -1 && close(cp->u.fd))
625: err(FATAL, "%s: %s", cp->t, strerror(errno));
626: cp->u.fd = -1;
627: break;
628: case '{':
629: cfclose(cp->u.c, cp->next);
630: break;
631: }
632: }