Annotation of src/usr.bin/sed/process.c, Revision 1.32
1.32 ! tom 1: /* $OpenBSD: process.c,v 1.31 2017/01/21 10:05:23 krw Exp $ */
1.2 deraadt 2:
1.1 deraadt 3: /*-
4: * Copyright (c) 1992 Diomidis Spinellis.
5: * Copyright (c) 1992, 1993
6: * The Regents of the University of California. All rights reserved.
7: *
8: * This code is derived from software contributed to Berkeley by
9: * Diomidis Spinellis of Imperial College, University of London.
10: *
11: * Redistribution and use in source and binary forms, with or without
12: * modification, are permitted provided that the following conditions
13: * are met:
14: * 1. Redistributions of source code must retain the above copyright
15: * notice, this list of conditions and the following disclaimer.
16: * 2. Redistributions in binary form must reproduce the above copyright
17: * notice, this list of conditions and the following disclaimer in the
18: * documentation and/or other materials provided with the distribution.
1.10 millert 19: * 3. Neither the name of the University nor the names of its contributors
1.1 deraadt 20: * may be used to endorse or promote products derived from this software
21: * without specific prior written permission.
22: *
23: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33: * SUCH DAMAGE.
34: */
35:
36: #include <sys/types.h>
37: #include <sys/stat.h>
38: #include <sys/uio.h>
39:
40: #include <ctype.h>
41: #include <errno.h>
42: #include <fcntl.h>
43: #include <limits.h>
44: #include <regex.h>
45: #include <stdio.h>
46: #include <stdlib.h>
47: #include <string.h>
48: #include <unistd.h>
49:
50: #include "defs.h"
51: #include "extern.h"
52:
53: static SPACE HS, PS, SS;
54: #define pd PS.deleted
55: #define ps PS.space
56: #define psl PS.len
1.24 jasper 57: #define psanl PS.append_newline
1.1 deraadt 58: #define hs HS.space
59: #define hsl HS.len
60:
1.7 millert 61: static inline int applies(struct s_command *);
62: static void flush_appends(void);
63: static void lputs(char *);
1.28 martijn 64: static inline int regexec_e(regex_t *, const char *, int, int, size_t,
65: size_t);
1.7 millert 66: static void regsub(SPACE *, char *, char *);
67: static int substitute(struct s_command *);
1.1 deraadt 68:
69: struct s_appends *appends; /* Array of pointers to strings to append. */
70: static int appendx; /* Index into appends array. */
1.22 deraadt 71: size_t appendnum; /* Size of appends array. */
1.1 deraadt 72:
73: static int lastaddr; /* Set by applies if last address of a range. */
74: static int sdone; /* If any substitutes since last line input. */
75: /* Iov structure for 'w' commands. */
76: static regex_t *defpreg;
77: size_t maxnsub;
78: regmatch_t *match;
79:
1.24 jasper 80: #define OUT() do {\
81: fwrite(ps, 1, psl, outfile);\
82: if (psanl) fputc('\n', outfile);\
83: } while (0)
1.1 deraadt 84:
85: void
1.11 deraadt 86: process(void)
1.1 deraadt 87: {
88: struct s_command *cp;
89: SPACE tspace;
90: size_t len, oldpsl;
91: char *p;
92:
93: for (linenum = 0; mf_fgets(&PS, REPLACE);) {
94: pd = 0;
95: top:
96: cp = prog;
97: redirect:
98: while (cp != NULL) {
99: if (!applies(cp)) {
100: cp = cp->next;
101: continue;
102: }
103: switch (cp->code) {
104: case '{':
105: cp = cp->u.c;
106: goto redirect;
107: case 'a':
1.12 tedu 108: if (appendx >= appendnum) {
1.20 deraadt 109: appends = xreallocarray(appends,
1.22 deraadt 110: appendnum,
111: 2 * sizeof(struct s_appends));
1.12 tedu 112: appendnum *= 2;
113: }
1.1 deraadt 114: appends[appendx].type = AP_STRING;
115: appends[appendx].s = cp->t;
116: appends[appendx].len = strlen(cp->t);
117: appendx++;
118: break;
119: case 'b':
120: cp = cp->u.c;
121: goto redirect;
122: case 'c':
123: pd = 1;
124: psl = 0;
1.24 jasper 125: if (cp->a2 == NULL || lastaddr || lastline())
126: (void)fprintf(outfile, "%s", cp->t);
1.1 deraadt 127: break;
128: case 'd':
129: pd = 1;
130: goto new;
131: case 'D':
132: if (pd)
133: goto new;
1.8 millert 134: if (psl == 0 ||
1.24 jasper 135: (p = memchr(ps, '\n', psl)) == NULL) {
1.1 deraadt 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':
1.24 jasper 147: cspace(&PS, "\n", 1, 0);
1.1 deraadt 148: cspace(&PS, hs, hsl, 0);
149: break;
150: case 'h':
151: cspace(&HS, ps, psl, REPLACE);
152: break;
153: case 'H':
1.24 jasper 154: cspace(&HS, "\n", 1, 0);
1.1 deraadt 155: cspace(&HS, ps, psl, 0);
156: break;
157: case 'i':
1.24 jasper 158: (void)fprintf(outfile, "%s", cp->t);
1.1 deraadt 159: break;
160: case 'l':
161: lputs(ps);
162: break;
163: case 'n':
164: if (!nflag && !pd)
1.24 jasper 165: OUT();
1.1 deraadt 166: flush_appends();
167: if (!mf_fgets(&PS, REPLACE))
168: exit(0);
169: pd = 0;
170: break;
171: case 'N':
172: flush_appends();
1.24 jasper 173: cspace(&PS, "\n", 1, 0);
174: if (!mf_fgets(&PS, 0))
1.1 deraadt 175: exit(0);
176: break;
177: case 'p':
178: if (pd)
179: break;
1.24 jasper 180: OUT();
1.1 deraadt 181: break;
182: case 'P':
183: if (pd)
184: break;
1.24 jasper 185: if ((p = memchr(ps, '\n', psl)) != NULL) {
1.1 deraadt 186: oldpsl = psl;
1.24 jasper 187: psl = p - ps;
188: psanl = 1;
189: OUT();
1.21 jsg 190: psl = oldpsl;
191: } else {
1.24 jasper 192: OUT();
1.1 deraadt 193: }
194: break;
195: case 'q':
196: if (!nflag && !pd)
1.24 jasper 197: OUT();
1.1 deraadt 198: flush_appends();
199: exit(0);
200: case 'r':
1.22 deraadt 201: if (appendx >= appendnum) {
1.20 deraadt 202: appends = xreallocarray(appends,
1.22 deraadt 203: appendnum,
204: 2 * sizeof(struct s_appends));
205: appendnum *= 2;
206: }
1.1 deraadt 207: appends[appendx].type = AP_FILE;
208: appends[appendx].s = cp->t;
209: appends[appendx].len = strlen(cp->t);
210: appendx++;
211: break;
212: case 's':
213: sdone |= substitute(cp);
214: break;
215: case 't':
216: if (sdone) {
217: sdone = 0;
218: cp = cp->u.c;
219: goto redirect;
220: }
221: break;
222: case 'w':
223: if (pd)
224: break;
225: if (cp->u.fd == -1 && (cp->u.fd = open(cp->t,
226: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
227: DEFFILEMODE)) == -1)
1.27 mmcc 228: error(FATAL, "%s: %s",
1.1 deraadt 229: cp->t, strerror(errno));
1.24 jasper 230: if (write(cp->u.fd, ps, psl) != psl ||
231: write(cp->u.fd, "\n", 1) != 1)
1.27 mmcc 232: error(FATAL, "%s: %s",
1.1 deraadt 233: cp->t, strerror(errno));
234: break;
235: case 'x':
236: if (hs == NULL)
1.24 jasper 237: cspace(&HS, "", 0, REPLACE);
1.1 deraadt 238: tspace = PS;
239: PS = HS;
1.24 jasper 240: psanl = tspace.append_newline;
1.1 deraadt 241: HS = tspace;
242: break;
243: case 'y':
1.8 millert 244: if (pd || psl == 0)
1.1 deraadt 245: break;
1.24 jasper 246: for (p = ps, len = psl; len--; ++p)
1.4 deraadt 247: *p = cp->u.y[(unsigned char)*p];
1.1 deraadt 248: break;
249: case ':':
250: case '}':
251: break;
252: case '=':
1.24 jasper 253: (void)fprintf(outfile, "%lu\n", linenum);
1.1 deraadt 254: }
255: cp = cp->next;
256: } /* for all cp */
257:
258: new: if (!nflag && !pd)
1.24 jasper 259: OUT();
1.1 deraadt 260: flush_appends();
261: } /* for all lines */
262: }
263:
264: /*
265: * TRUE if the address passed matches the current program state
266: * (lastline, linenumber, ps).
267: */
268: #define MATCH(a) \
1.28 martijn 269: (a)->type == AT_RE ? regexec_e((a)->u.r, ps, 0, 1, 0, psl) : \
1.24 jasper 270: (a)->type == AT_LINE ? linenum == (a)->u.l : lastline()
1.1 deraadt 271:
272: /*
273: * Return TRUE if the command applies to the current line. Sets the inrange
274: * flag to process ranges. Interprets the non-select (``!'') flag.
275: */
276: static inline int
1.11 deraadt 277: applies(struct s_command *cp)
1.1 deraadt 278: {
279: int r;
280:
281: lastaddr = 0;
282: if (cp->a1 == NULL && cp->a2 == NULL)
283: r = 1;
284: else if (cp->a2)
285: if (cp->inrange) {
286: if (MATCH(cp->a2)) {
287: cp->inrange = 0;
288: lastaddr = 1;
289: }
290: r = 1;
291: } else if (MATCH(cp->a1)) {
292: /*
293: * If the second address is a number less than or
294: * equal to the line number first selected, only
295: * one line shall be selected.
296: * -- POSIX 1003.2
297: */
298: if (cp->a2->type == AT_LINE &&
299: linenum >= cp->a2->u.l)
300: lastaddr = 1;
301: else
302: cp->inrange = 1;
303: r = 1;
304: } else
305: r = 0;
306: else
307: r = MATCH(cp->a1);
1.13 tedu 308: return (cp->nonsel ? !r : r);
1.1 deraadt 309: }
310:
311: /*
1.24 jasper 312: * Reset all inrange markers.
313: */
314: void
315: resetranges(void)
316: {
317: struct s_command *cp;
318:
319: for (cp = prog; cp; cp = cp->code == '{' ? cp->u.c : cp->next)
320: if (cp->a2)
321: cp->inrange = 0;
322: }
323:
324: /*
1.1 deraadt 325: * substitute --
326: * Do substitutions in the pattern space. Currently, we build a
327: * copy of the new pattern space in the substitute space structure
328: * and then swap them.
329: */
330: static int
1.11 deraadt 331: substitute(struct s_command *cp)
1.1 deraadt 332: {
333: SPACE tspace;
334: regex_t *re;
1.18 schwarze 335: regoff_t slen;
1.1 deraadt 336: int n, lastempty;
1.28 martijn 337: size_t le = 0;
1.1 deraadt 338: char *s;
339:
340: s = ps;
341: re = cp->u.s->re;
342: if (re == NULL) {
343: if (defpreg != NULL && cp->u.s->maxbref > defpreg->re_nsub) {
344: linenum = cp->u.s->linenum;
1.27 mmcc 345: error(COMPILE, "\\%d not defined in the RE",
1.1 deraadt 346: cp->u.s->maxbref);
347: }
348: }
1.28 martijn 349: if (!regexec_e(re, ps, 0, 0, 0, psl))
1.1 deraadt 350: return (0);
351:
352: SS.len = 0; /* Clean substitute space. */
353: slen = psl;
354: n = cp->u.s->n;
355: lastempty = 1;
356:
1.18 schwarze 357: do {
358: /* Copy the leading retained string. */
1.28 martijn 359: if (n <= 1 && (match[0].rm_so > le))
360: cspace(&SS, s, match[0].rm_so - le, APPEND);
1.18 schwarze 361:
362: /* Skip zero-length matches right after other matches. */
1.28 martijn 363: if (lastempty || (match[0].rm_so - le) ||
1.18 schwarze 364: match[0].rm_so != match[0].rm_eo) {
365: if (n <= 1) {
366: /* Want this match: append replacement. */
1.28 martijn 367: regsub(&SS, ps, cp->u.s->new);
1.18 schwarze 368: if (n == 1)
369: n = -1;
1.1 deraadt 370: } else {
1.18 schwarze 371: /* Want a later match: append original. */
1.28 martijn 372: if (match[0].rm_eo - le)
373: cspace(&SS, s, match[0].rm_eo - le,
374: APPEND);
1.18 schwarze 375: n--;
1.1 deraadt 376: }
377: }
1.18 schwarze 378:
379: /* Move past this match. */
1.28 martijn 380: s = ps + match[0].rm_eo;
381: slen = psl - match[0].rm_eo;
382: le = match[0].rm_eo;
1.18 schwarze 383:
384: /*
385: * After a zero-length match, advance one byte,
386: * and at the end of the line, terminate.
387: */
388: if (match[0].rm_so == match[0].rm_eo) {
389: if (*s == '\0' || *s == '\n')
390: slen = -1;
391: else
392: slen--;
1.28 martijn 393: if (*s != '\0') {
1.31 krw 394: cspace(&SS, s++, 1, APPEND);
1.28 martijn 395: le++;
396: }
1.18 schwarze 397: lastempty = 1;
398: } else
399: lastempty = 0;
400:
1.28 martijn 401: } while (n >= 0 && slen >= 0 &&
402: regexec_e(re, ps, REG_NOTBOL, 0, le, psl));
1.18 schwarze 403:
404: /* Did not find the requested number of matches. */
1.29 martijn 405: if (n > 0)
1.18 schwarze 406: return (0);
407:
408: /* Copy the trailing retained string. */
409: if (slen > 0)
1.1 deraadt 410: cspace(&SS, s, slen, APPEND);
411:
412: /*
413: * Swap the substitute space and the pattern space, and make sure
414: * that any leftover pointers into stdio memory get lost.
415: */
416: tspace = PS;
417: PS = SS;
1.24 jasper 418: psanl = tspace.append_newline;
1.1 deraadt 419: SS = tspace;
420: SS.space = SS.back;
421:
422: /* Handle the 'p' flag. */
423: if (cp->u.s->p)
1.24 jasper 424: OUT();
1.1 deraadt 425:
426: /* Handle the 'w' flag. */
427: if (cp->u.s->wfile && !pd) {
428: if (cp->u.s->wfd == -1 && (cp->u.s->wfd = open(cp->u.s->wfile,
429: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, DEFFILEMODE)) == -1)
1.27 mmcc 430: error(FATAL, "%s: %s", cp->u.s->wfile, strerror(errno));
1.24 jasper 431: if (write(cp->u.s->wfd, ps, psl) != psl ||
432: write(cp->u.s->wfd, "\n", 1) != 1)
1.27 mmcc 433: error(FATAL, "%s: %s", cp->u.s->wfile, strerror(errno));
1.1 deraadt 434: }
435: return (1);
436: }
437:
438: /*
439: * Flush append requests. Always called before reading a line,
440: * therefore it also resets the substitution done (sdone) flag.
441: */
442: static void
1.11 deraadt 443: flush_appends(void)
1.1 deraadt 444: {
445: FILE *f;
446: int count, i;
447: char buf[8 * 1024];
448:
1.26 jasper 449: for (i = 0; i < appendx; i++)
1.1 deraadt 450: switch (appends[i].type) {
451: case AP_STRING:
1.26 jasper 452: fwrite(appends[i].s, sizeof(char), appends[i].len,
1.24 jasper 453: outfile);
1.1 deraadt 454: break;
455: case AP_FILE:
456: /*
457: * Read files probably shouldn't be cached. Since
458: * it's not an error to read a non-existent file,
459: * it's possible that another program is interacting
460: * with the sed script through the file system. It
461: * would be truly bizarre, but possible. It's probably
462: * not that big a performance win, anyhow.
463: */
464: if ((f = fopen(appends[i].s, "r")) == NULL)
465: break;
1.3 deraadt 466: while ((count = fread(buf, sizeof(char), sizeof(buf), f)))
1.24 jasper 467: (void)fwrite(buf, sizeof(char), count, outfile);
1.1 deraadt 468: (void)fclose(f);
469: break;
470: }
1.24 jasper 471: if (ferror(outfile))
1.27 mmcc 472: error(FATAL, "%s: %s", outfname, strerror(errno ? errno : EIO));
1.1 deraadt 473: appendx = sdone = 0;
474: }
475:
476: static void
1.11 deraadt 477: lputs(char *s)
1.1 deraadt 478: {
1.6 mpech 479: int count;
1.25 deraadt 480: extern int termwidth;
1.24 jasper 481: const char *escapes;
482: char *p;
1.1 deraadt 483:
1.26 jasper 484: for (count = 0; *s; ++s) {
1.1 deraadt 485: if (count >= termwidth) {
1.24 jasper 486: (void)fprintf(outfile, "\\\n");
1.1 deraadt 487: count = 0;
488: }
1.19 deraadt 489: if (isascii((unsigned char)*s) && isprint((unsigned char)*s)
490: && *s != '\\') {
1.24 jasper 491: (void)fputc(*s, outfile);
1.1 deraadt 492: count++;
1.24 jasper 493: } else if (*s == '\n') {
494: (void)fputc('$', outfile);
495: (void)fputc('\n', outfile);
496: count = 0;
497: } else {
1.14 millert 498: escapes = "\\\a\b\f\r\t\v";
1.24 jasper 499: (void)fputc('\\', outfile);
1.3 deraadt 500: if ((p = strchr(escapes, *s))) {
1.24 jasper 501: (void)fputc("\\abfrtv"[p - escapes], outfile);
1.1 deraadt 502: count += 2;
503: } else {
1.24 jasper 504: (void)fprintf(outfile, "%03o", *(u_char *)s);
1.1 deraadt 505: count += 4;
506: }
507: }
508: }
1.24 jasper 509: (void)fputc('$', outfile);
510: (void)fputc('\n', outfile);
511: if (ferror(outfile))
1.27 mmcc 512: error(FATAL, "%s: %s", outfname, strerror(errno ? errno : EIO));
1.1 deraadt 513: }
514:
515: static inline int
1.11 deraadt 516: regexec_e(regex_t *preg, const char *string, int eflags,
1.28 martijn 517: int nomatch, size_t start, size_t stop)
1.1 deraadt 518: {
519: int eval;
1.26 jasper 520:
1.1 deraadt 521: if (preg == NULL) {
522: if (defpreg == NULL)
1.27 mmcc 523: error(FATAL, "first RE may not be empty");
1.1 deraadt 524: } else
525: defpreg = preg;
526:
1.24 jasper 527: /* Set anchors */
1.28 martijn 528: match[0].rm_so = start;
529: match[0].rm_eo = stop;
1.26 jasper 530:
1.1 deraadt 531: eval = regexec(defpreg, string,
532: nomatch ? 0 : maxnsub + 1, match, eflags | REG_STARTEND);
1.13 tedu 533: switch (eval) {
1.1 deraadt 534: case 0:
535: return (1);
536: case REG_NOMATCH:
537: return (0);
538: }
1.27 mmcc 539: error(FATAL, "RE error: %s", strregerror(eval, defpreg));
1.1 deraadt 540: }
541:
542: /*
543: * regsub - perform substitutions after a regexp match
544: * Based on a routine by Henry Spencer
545: */
546: static void
1.11 deraadt 547: regsub(SPACE *sp, char *string, char *src)
1.1 deraadt 548: {
1.6 mpech 549: int len, no;
550: char c, *dst;
1.1 deraadt 551:
552: #define NEEDSP(reqlen) \
1.12 tedu 553: if (sp->len + (reqlen) + 1 >= sp->blen) { \
554: size_t newlen = sp->blen + (reqlen) + 1024; \
555: sp->space = sp->back = xrealloc(sp->back, newlen); \
556: sp->blen = newlen; \
1.1 deraadt 557: dst = sp->space + sp->len; \
558: }
559:
560: dst = sp->space + sp->len;
561: while ((c = *src++) != '\0') {
562: if (c == '&')
563: no = 0;
1.19 deraadt 564: else if (c == '\\' && isdigit((unsigned char)*src))
1.1 deraadt 565: no = *src++ - '0';
566: else
567: no = -1;
568: if (no < 0) { /* Ordinary character. */
1.31 krw 569: if (c == '\\' && (*src == '\\' || *src == '&'))
570: c = *src++;
1.1 deraadt 571: NEEDSP(1);
1.31 krw 572: *dst++ = c;
1.1 deraadt 573: ++sp->len;
1.31 krw 574: } else if (match[no].rm_so != -1 && match[no].rm_eo != -1) {
1.1 deraadt 575: len = match[no].rm_eo - match[no].rm_so;
576: NEEDSP(len);
577: memmove(dst, string + match[no].rm_so, len);
578: dst += len;
579: sp->len += len;
580: }
581: }
582: NEEDSP(1);
583: *dst = '\0';
584: }
585:
586: /*
587: * aspace --
588: * Append the source space to the destination space, allocating new
589: * space as necessary.
590: */
591: void
1.24 jasper 592: cspace(SPACE *sp, const char *p, size_t len, enum e_spflag spflag)
1.1 deraadt 593: {
594: size_t tlen;
595:
596: /* Make sure SPACE has enough memory and ramp up quickly. */
597: tlen = sp->len + len + 1;
598: if (tlen > sp->blen) {
1.12 tedu 599: size_t newlen = tlen + 1024;
600: sp->space = sp->back = xrealloc(sp->back, newlen);
601: sp->blen = newlen;
1.1 deraadt 602: }
603:
604: if (spflag == REPLACE)
605: sp->len = 0;
606:
607: memmove(sp->space + sp->len, p, len);
608:
609: sp->space[sp->len += len] = '\0';
610: }
611:
612: /*
613: * Close all cached opened files and report any errors
614: */
615: void
1.11 deraadt 616: cfclose(struct s_command *cp, struct s_command *end)
1.1 deraadt 617: {
618:
619: for (; cp != end; cp = cp->next)
1.13 tedu 620: switch (cp->code) {
1.1 deraadt 621: case 's':
622: if (cp->u.s->wfd != -1 && close(cp->u.s->wfd))
1.27 mmcc 623: error(FATAL,
1.1 deraadt 624: "%s: %s", cp->u.s->wfile, strerror(errno));
625: cp->u.s->wfd = -1;
626: break;
627: case 'w':
628: if (cp->u.fd != -1 && close(cp->u.fd))
1.27 mmcc 629: error(FATAL, "%s: %s", cp->t, strerror(errno));
1.1 deraadt 630: cp->u.fd = -1;
631: break;
632: case '{':
633: cfclose(cp->u.c, cp->next);
634: break;
635: }
636: }