Annotation of src/usr.bin/m4/gnum4.c, Revision 1.41
1.41 ! marco 1: /* $OpenBSD: gnum4.c,v 1.40 2010/03/22 20:40:43 espie Exp $ */
1.1 espie 2:
3: /*
4: * Copyright (c) 1999 Marc Espie
5: *
6: * Redistribution and use in source and binary forms, with or without
7: * modification, are permitted provided that the following conditions
8: * are met:
9: * 1. Redistributions of source code must retain the above copyright
10: * notice, this list of conditions and the following disclaimer.
11: * 2. Redistributions in binary form must reproduce the above copyright
12: * notice, this list of conditions and the following disclaimer in the
13: * documentation and/or other materials provided with the distribution.
14: *
15: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
16: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
19: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25: * SUCH DAMAGE.
26: */
27:
1.41 ! marco 28: /*
1.1 espie 29: * functions needed to support gnu-m4 extensions, including a fake freezing
30: */
31:
32: #include <sys/param.h>
1.5 espie 33: #include <sys/types.h>
1.8 espie 34: #include <sys/wait.h>
1.5 espie 35: #include <ctype.h>
1.36 espie 36: #include <err.h>
1.8 espie 37: #include <paths.h>
1.5 espie 38: #include <regex.h>
1.1 espie 39: #include <stddef.h>
40: #include <stdlib.h>
41: #include <stdio.h>
42: #include <string.h>
1.8 espie 43: #include <errno.h>
44: #include <unistd.h>
1.1 espie 45: #include "mdef.h"
46: #include "stdd.h"
47: #include "extern.h"
48:
1.5 espie 49:
50: int mimic_gnu = 0;
51:
1.1 espie 52: /*
53: * Support for include path search
1.27 espie 54: * First search in the current directory.
1.1 espie 55: * If not found, and the path is not absolute, include path kicks in.
56: * First, -I options, in the order found on the command line.
57: * Then M4PATH env variable
58: */
59:
60: struct path_entry {
61: char *name;
62: struct path_entry *next;
63: } *first, *last;
64:
1.16 millert 65: static struct path_entry *new_path_entry(const char *);
66: static void ensure_m4path(void);
67: static struct input_file *dopath(struct input_file *, const char *);
1.3 espie 68:
1.1 espie 69: static struct path_entry *
1.18 espie 70: new_path_entry(const char *dirname)
1.1 espie 71: {
72: struct path_entry *n;
73:
74: n = malloc(sizeof(struct path_entry));
75: if (!n)
76: errx(1, "out of memory");
77: n->name = strdup(dirname);
78: if (!n->name)
79: errx(1, "out of memory");
80: n->next = 0;
81: return n;
82: }
1.41 ! marco 83:
! 84: void
1.18 espie 85: addtoincludepath(const char *dirname)
1.1 espie 86: {
87: struct path_entry *n;
88:
89: n = new_path_entry(dirname);
90:
91: if (last) {
92: last->next = n;
93: last = n;
94: }
95: else
96: last = first = n;
97: }
98:
99: static void
100: ensure_m4path()
101: {
102: static int envpathdone = 0;
103: char *envpath;
104: char *sweep;
105: char *path;
106:
107: if (envpathdone)
108: return;
109: envpathdone = TRUE;
110: envpath = getenv("M4PATH");
1.41 ! marco 111: if (!envpath)
1.1 espie 112: return;
113: /* for portability: getenv result is read-only */
114: envpath = strdup(envpath);
115: if (!envpath)
116: errx(1, "out of memory");
1.41 ! marco 117: for (sweep = envpath;
1.1 espie 118: (path = strsep(&sweep, ":")) != NULL;)
119: addtoincludepath(path);
120: free(envpath);
121: }
122:
123: static
1.4 espie 124: struct input_file *
1.18 espie 125: dopath(struct input_file *i, const char *filename)
1.1 espie 126: {
127: char path[MAXPATHLEN];
128: struct path_entry *pe;
1.4 espie 129: FILE *f;
1.1 espie 130:
131: for (pe = first; pe; pe = pe->next) {
132: snprintf(path, sizeof(path), "%s/%s", pe->name, filename);
1.4 espie 133: if ((f = fopen(path, "r")) != 0) {
134: set_input(i, f, path);
135: return i;
136: }
1.1 espie 137: }
138: return NULL;
139: }
140:
1.4 espie 141: struct input_file *
1.18 espie 142: fopen_trypath(struct input_file *i, const char *filename)
1.1 espie 143: {
144: FILE *f;
145:
146: f = fopen(filename, "r");
1.4 espie 147: if (f != NULL) {
148: set_input(i, f, filename);
149: return i;
150: }
1.1 espie 151: if (filename[0] == '/')
152: return NULL;
153:
154: ensure_m4path();
155:
1.4 espie 156: return dopath(i, filename);
1.1 espie 157: }
158:
1.41 ! marco 159: void
1.18 espie 160: doindir(const char *argv[], int argc)
1.5 espie 161: {
1.25 espie 162: ndptr n;
1.22 espie 163: struct macro_definition *p;
1.5 espie 164:
1.25 espie 165: n = lookup(argv[2]);
166: if (n == NULL || (p = macro_getdef(n)) == NULL)
1.32 espie 167: m4errx(1, "indir: undefined macro %s.", argv[2]);
1.5 espie 168: argv[1] = p->defn;
1.41 ! marco 169:
1.25 espie 170: eval(argv+1, argc-1, p->type, is_traced(n));
1.5 espie 171: }
172:
1.41 ! marco 173: void
1.18 espie 174: dobuiltin(const char *argv[], int argc)
1.5 espie 175: {
1.24 espie 176: ndptr p;
1.25 espie 177:
1.5 espie 178: argv[1] = NULL;
1.24 espie 179: p = macro_getbuiltin(argv[2]);
180: if (p != NULL)
1.25 espie 181: eval(argv+1, argc-1, macro_builtin_type(p), is_traced(p));
1.5 espie 182: else
1.32 espie 183: m4errx(1, "unknown builtin %s.", argv[2]);
1.41 ! marco 184: }
1.5 espie 185:
186:
187: /* We need some temporary buffer space, as pb pushes BACK and substitution
188: * proceeds forward... */
189: static char *buffer;
190: static size_t bufsize = 0;
191: static size_t current = 0;
192:
1.16 millert 193: static void addchars(const char *, size_t);
1.17 espie 194: static void addchar(int);
1.16 millert 195: static char *twiddle(const char *);
196: static char *getstring(void);
197: static void exit_regerror(int, regex_t *);
198: static void do_subst(const char *, regex_t *, const char *, regmatch_t *);
199: static void do_regexpindex(const char *, regex_t *, regmatch_t *);
200: static void do_regexp(const char *, regex_t *, const char *, regmatch_t *);
201: static void add_sub(int, const char *, regex_t *, regmatch_t *);
202: static void add_replace(const char *, regex_t *, const char *, regmatch_t *);
1.14 espie 203: #define addconstantstring(s) addchars((s), sizeof(s)-1)
1.5 espie 204:
1.41 ! marco 205: static void
1.18 espie 206: addchars(const char *c, size_t n)
1.5 espie 207: {
208: if (n == 0)
209: return;
1.15 espie 210: while (current + n > bufsize) {
1.5 espie 211: if (bufsize == 0)
212: bufsize = 1024;
213: else
214: bufsize *= 2;
1.26 espie 215: buffer = xrealloc(buffer, bufsize, NULL);
1.5 espie 216: }
217: memcpy(buffer+current, c, n);
218: current += n;
219: }
220:
1.41 ! marco 221: static void
1.18 espie 222: addchar(int c)
1.5 espie 223: {
224: if (current +1 > bufsize) {
225: if (bufsize == 0)
226: bufsize = 1024;
227: else
228: bufsize *= 2;
1.26 espie 229: buffer = xrealloc(buffer, bufsize, NULL);
1.5 espie 230: }
231: buffer[current++] = c;
232: }
233:
234: static char *
235: getstring()
236: {
237: addchar('\0');
238: current = 0;
239: return buffer;
240: }
241:
242:
1.41 ! marco 243: static void
1.18 espie 244: exit_regerror(int er, regex_t *re)
1.5 espie 245: {
1.41 ! marco 246: size_t errlen;
! 247: char *errbuf;
1.5 espie 248:
249: errlen = regerror(er, re, NULL, 0);
1.41 ! marco 250: errbuf = xalloc(errlen,
1.26 espie 251: "malloc in regerror: %lu", (unsigned long)errlen);
1.5 espie 252: regerror(er, re, errbuf, errlen);
1.32 espie 253: m4errx(1, "regular expression error: %s.", errbuf);
1.5 espie 254: }
255:
256: static void
1.18 espie 257: add_sub(int n, const char *string, regex_t *re, regmatch_t *pm)
1.5 espie 258: {
259: if (n > re->re_nsub)
260: warnx("No subexpression %d", n);
261: /* Subexpressions that did not match are
262: * not an error. */
263: else if (pm[n].rm_so != -1 &&
264: pm[n].rm_eo != -1) {
265: addchars(string + pm[n].rm_so,
266: pm[n].rm_eo - pm[n].rm_so);
267: }
268: }
269:
270: /* Add replacement string to the output buffer, recognizing special
271: * constructs and replacing them with substrings of the original string.
272: */
1.41 ! marco 273: static void
1.18 espie 274: add_replace(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
1.5 espie 275: {
276: const char *p;
277:
278: for (p = replace; *p != '\0'; p++) {
279: if (*p == '&' && !mimic_gnu) {
280: add_sub(0, string, re, pm);
281: continue;
282: }
283: if (*p == '\\') {
284: if (p[1] == '\\') {
285: addchar(p[1]);
1.10 espie 286: p++;
1.5 espie 287: continue;
288: }
289: if (p[1] == '&') {
290: if (mimic_gnu)
291: add_sub(0, string, re, pm);
292: else
293: addchar(p[1]);
294: p++;
295: continue;
296: }
297: if (isdigit(p[1])) {
298: add_sub(*(++p) - '0', string, re, pm);
299: continue;
300: }
301: }
1.41 ! marco 302: addchar(*p);
1.5 espie 303: }
304: }
305:
1.41 ! marco 306: static void
1.18 espie 307: do_subst(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
1.5 espie 308: {
309: int error;
1.11 espie 310: int flags = 0;
311: const char *last_match = NULL;
1.5 espie 312:
1.11 espie 313: while ((error = regexec(re, string, re->re_nsub+1, pm, flags)) == 0) {
1.13 espie 314: if (pm[0].rm_eo != 0) {
315: if (string[pm[0].rm_eo-1] == '\n')
316: flags = 0;
317: else
318: flags = REG_NOTBOL;
319: }
1.5 espie 320:
1.41 ! marco 321: /* NULL length matches are special... We use the `vi-mode'
1.5 espie 322: * rule: don't allow a NULL-match at the last match
1.41 ! marco 323: * position.
1.5 espie 324: */
1.41 ! marco 325: if (pm[0].rm_so == pm[0].rm_eo &&
1.11 espie 326: string + pm[0].rm_so == last_match) {
1.5 espie 327: if (*string == '\0')
328: return;
329: addchar(*string);
1.13 espie 330: if (*string++ == '\n')
331: flags = 0;
332: else
333: flags = REG_NOTBOL;
1.5 espie 334: continue;
335: }
1.11 espie 336: last_match = string + pm[0].rm_so;
337: addchars(string, pm[0].rm_so);
1.5 espie 338: add_replace(string, re, replace, pm);
339: string += pm[0].rm_eo;
340: }
341: if (error != REG_NOMATCH)
342: exit_regerror(error, re);
343: pbstr(string);
344: }
345:
1.41 ! marco 346: static void
1.18 espie 347: do_regexp(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
1.5 espie 348: {
349: int error;
350:
351: switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
1.41 ! marco 352: case 0:
1.5 espie 353: add_replace(string, re, replace, pm);
354: pbstr(getstring());
355: break;
356: case REG_NOMATCH:
357: break;
358: default:
359: exit_regerror(error, re);
360: }
361: }
362:
1.41 ! marco 363: static void
1.18 espie 364: do_regexpindex(const char *string, regex_t *re, regmatch_t *pm)
1.5 espie 365: {
366: int error;
367:
368: switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
369: case 0:
370: pbunsigned(pm[0].rm_so);
371: break;
372: case REG_NOMATCH:
373: pbnum(-1);
374: break;
375: default:
376: exit_regerror(error, re);
377: }
378: }
379:
380: /* In Gnu m4 mode, parentheses for backmatch don't work like POSIX 1003.2
381: * says. So we twiddle with the regexp before passing it to regcomp.
382: */
383: static char *
1.18 espie 384: twiddle(const char *p)
1.5 espie 385: {
1.28 espie 386: /* + at start of regexp is a normal character for Gnu m4 */
387: if (*p == '^') {
388: addchar(*p);
389: p++;
390: }
391: if (*p == '+') {
392: addchar('\\');
393: }
1.5 espie 394: /* This could use strcspn for speed... */
395: while (*p != '\0') {
1.14 espie 396: if (*p == '\\') {
397: switch(p[1]) {
398: case '(':
399: case ')':
400: case '|':
401: addchar(p[1]);
402: break;
403: case 'w':
404: addconstantstring("[_a-zA-Z0-9]");
405: break;
406: case 'W':
407: addconstantstring("[^_a-zA-Z0-9]");
408: break;
409: case '<':
410: addconstantstring("[[:<:]]");
411: break;
412: case '>':
413: addconstantstring("[[:>:]]");
414: break;
415: default:
416: addchars(p, 2);
417: break;
418: }
1.5 espie 419: p+=2;
420: continue;
421: }
1.14 espie 422: if (*p == '(' || *p == ')' || *p == '|')
1.5 espie 423: addchar('\\');
424:
425: addchar(*p);
426: p++;
427: }
428: return getstring();
429: }
430:
431: /* patsubst(string, regexp, opt replacement) */
432: /* argv[2]: string
433: * argv[3]: regexp
434: * argv[4]: opt rep
435: */
436: void
1.18 espie 437: dopatsubst(const char *argv[], int argc)
1.5 espie 438: {
439: if (argc <= 3) {
440: warnx("Too few arguments to patsubst");
441: return;
442: }
1.20 espie 443: /* special case: empty regexp */
444: if (argv[3][0] == '\0') {
445: const char *s;
1.21 espie 446: size_t len;
1.41 ! marco 447: if (argc > 4 && argv[4])
1.21 espie 448: len = strlen(argv[4]);
449: else
450: len = 0;
1.20 espie 451: for (s = argv[2]; *s != '\0'; s++) {
452: addchars(argv[4], len);
453: addchar(*s);
454: }
455: } else {
456: int error;
457: regex_t re;
458: regmatch_t *pmatch;
1.30 espie 459: int mode = REG_EXTENDED;
460: size_t l = strlen(argv[3]);
461:
462: if (!mimic_gnu ||
1.41 ! marco 463: (argv[3][0] == '^') ||
1.30 espie 464: (l > 0 && argv[3][l-1] == '$'))
465: mode |= REG_NEWLINE;
1.20 espie 466:
1.41 ! marco 467: error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
1.30 espie 468: mode);
1.20 espie 469: if (error != 0)
470: exit_regerror(error, &re);
1.41 ! marco 471:
1.26 espie 472: pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1), NULL);
1.41 ! marco 473: do_subst(argv[2], &re,
1.21 espie 474: argc > 4 && argv[4] != NULL ? argv[4] : "", pmatch);
1.20 espie 475: free(pmatch);
476: regfree(&re);
477: }
1.5 espie 478: pbstr(getstring());
479: }
480:
481: void
1.18 espie 482: doregexp(const char *argv[], int argc)
1.5 espie 483: {
484: int error;
485: regex_t re;
486: regmatch_t *pmatch;
487:
488: if (argc <= 3) {
1.7 espie 489: warnx("Too few arguments to regexp");
1.5 espie 490: return;
491: }
1.40 espie 492: /* special gnu case */
493: if (argv[3][0] == '\0' && mimic_gnu) {
494: if (argc == 4 || argv[4] == NULL)
495: return;
1.41 ! marco 496: else
1.40 espie 497: pbstr(argv[4]);
498: }
1.41 ! marco 499: error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
1.5 espie 500: REG_EXTENDED);
501: if (error != 0)
502: exit_regerror(error, &re);
1.41 ! marco 503:
1.26 espie 504: pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1), NULL);
1.40 espie 505: if (argc == 4 || argv[4] == NULL)
1.5 espie 506: do_regexpindex(argv[2], &re, pmatch);
507: else
508: do_regexp(argv[2], &re, argv[4], pmatch);
509: free(pmatch);
510: regfree(&re);
1.33 espie 511: }
512:
513: void
514: doformat(const char *argv[], int argc)
515: {
516: const char *format = argv[2];
517: int pos = 3;
1.39 espie 518: int left_padded;
519: long width;
520: size_t l;
521: const char *thisarg;
522: char temp[2];
523: long extra;
1.35 espie 524:
1.33 espie 525: while (*format != 0) {
526: if (*format != '%') {
527: addchar(*format++);
1.39 espie 528: continue;
529: }
530:
531: format++;
532: if (*format == '%') {
533: addchar(*format++);
534: continue;
535: }
536: if (*format == 0) {
537: addchar('%');
538: break;
539: }
540:
541: if (*format == '*') {
542: format++;
543: if (pos >= argc)
1.41 ! marco 544: m4errx(1,
1.39 espie 545: "Format with too many format specifiers.");
546: width = strtol(argv[pos++], NULL, 10);
547: } else {
548: width = strtol(format, (char **)&format, 10);
549: }
550: if (width < 0) {
551: left_padded = 1;
552: width = -width;
1.37 espie 553: } else {
1.39 espie 554: left_padded = 0;
555: }
556: if (*format == '.') {
1.33 espie 557: format++;
1.39 espie 558: if (*format == '*') {
559: format++;
560: if (pos >= argc)
1.41 ! marco 561: m4errx(1,
1.39 espie 562: "Format with too many format specifiers.");
563: extra = strtol(argv[pos++], NULL, 10);
1.33 espie 564: } else {
1.39 espie 565: extra = strtol(format, (char **)&format, 10);
1.33 espie 566: }
1.39 espie 567: } else {
568: extra = LONG_MAX;
569: }
570: if (pos >= argc)
571: m4errx(1, "Format with too many format specifiers.");
572: switch(*format) {
573: case 's':
574: thisarg = argv[pos++];
575: break;
576: case 'c':
577: temp[0] = strtoul(argv[pos++], NULL, 10);
578: temp[1] = 0;
579: thisarg = temp;
580: break;
581: default:
1.41 ! marco 582: m4errx(1, "Unsupported format specification: %s.",
1.39 espie 583: argv[2]);
584: }
585: format++;
586: l = strlen(thisarg);
587: if (l > extra)
588: l = extra;
589: if (!left_padded) {
590: while (l < width--)
591: addchar(' ');
592: }
593: addchars(thisarg, l);
594: if (left_padded) {
595: while (l < width--)
596: addchar(' ');
1.33 espie 597: }
598: }
599: pbstr(getstring());
1.8 espie 600: }
601:
602: void
1.18 espie 603: doesyscmd(const char *cmd)
1.8 espie 604: {
605: int p[2];
606: pid_t pid, cpid;
607: char *argv[4];
608: int cc;
609: int status;
610:
611: /* Follow gnu m4 documentation: first flush buffers. */
612: fflush(NULL);
613:
614: argv[0] = "sh";
615: argv[1] = "-c";
616: argv[2] = (char *)cmd;
617: argv[3] = NULL;
618:
619: /* Just set up standard output, share stderr and stdin with m4 */
620: if (pipe(p) == -1)
621: err(1, "bad pipe");
622: switch(cpid = fork()) {
623: case -1:
624: err(1, "bad fork");
625: /* NOTREACHED */
626: case 0:
627: (void) close(p[0]);
628: (void) dup2(p[1], 1);
629: (void) close(p[1]);
630: execv(_PATH_BSHELL, argv);
631: exit(1);
632: default:
633: /* Read result in two stages, since m4's buffer is
634: * pushback-only. */
635: (void) close(p[1]);
636: do {
637: char result[BUFSIZE];
638: cc = read(p[0], result, sizeof result);
639: if (cc > 0)
640: addchars(result, cc);
641: } while (cc > 0 || (cc == -1 && errno == EINTR));
642:
643: (void) close(p[0]);
644: while ((pid = wait(&status)) != cpid && pid >= 0)
645: continue;
646: pbstr(getstring());
647: }
1.31 espie 648: }
649:
650: void
651: getdivfile(const char *name)
652: {
653: FILE *f;
654: int c;
655:
656: f = fopen(name, "r");
657: if (!f)
658: return;
659:
660: while ((c = getc(f))!= EOF)
661: putc(c, active);
662: (void) fclose(f);
1.5 espie 663: }