Annotation of src/usr.bin/m4/gnum4.c, Revision 1.29
1.29 ! espie 1: /* $OpenBSD: gnum4.c,v 1.28 2004/11/07 10:26:18 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:
28: /*
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>
1.8 espie 36: #include <paths.h>
1.5 espie 37: #include <regex.h>
1.1 espie 38: #include <stddef.h>
39: #include <stdlib.h>
40: #include <stdio.h>
41: #include <string.h>
42: #include <err.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: }
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");
111: if (!envpath)
112: return;
113: /* for portability: getenv result is read-only */
114: envpath = strdup(envpath);
115: if (!envpath)
116: errx(1, "out of memory");
117: for (sweep = envpath;
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:
1.5 espie 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)
1.5 espie 167: errx(1, "undefined macro %s", argv[2]);
168: argv[1] = p->defn;
1.23 espie 169:
1.25 espie 170: eval(argv+1, argc-1, p->type, is_traced(n));
1.5 espie 171: }
172:
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
183: errx(1, "unknown builtin %s", argv[2]);
184: }
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:
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:
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:
243: static void
1.18 espie 244: exit_regerror(int er, regex_t *re)
1.5 espie 245: {
246: size_t errlen;
247: char *errbuf;
248:
249: errlen = regerror(er, re, NULL, 0);
1.26 espie 250: errbuf = xalloc(errlen,
251: "malloc in regerror: %lu", (unsigned long)errlen);
1.5 espie 252: regerror(er, re, errbuf, errlen);
1.19 espie 253: errx(1, "%s at line %lu: regular expression error: %s",
254: CURRENT_NAME, CURRENT_LINE, errbuf);
1.5 espie 255: }
256:
257: static void
1.18 espie 258: add_sub(int n, const char *string, regex_t *re, regmatch_t *pm)
1.5 espie 259: {
260: if (n > re->re_nsub)
261: warnx("No subexpression %d", n);
262: /* Subexpressions that did not match are
263: * not an error. */
264: else if (pm[n].rm_so != -1 &&
265: pm[n].rm_eo != -1) {
266: addchars(string + pm[n].rm_so,
267: pm[n].rm_eo - pm[n].rm_so);
268: }
269: }
270:
271: /* Add replacement string to the output buffer, recognizing special
272: * constructs and replacing them with substrings of the original string.
273: */
274: static void
1.18 espie 275: add_replace(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
1.5 espie 276: {
277: const char *p;
278:
279: for (p = replace; *p != '\0'; p++) {
280: if (*p == '&' && !mimic_gnu) {
281: add_sub(0, string, re, pm);
282: continue;
283: }
284: if (*p == '\\') {
285: if (p[1] == '\\') {
286: addchar(p[1]);
1.10 espie 287: p++;
1.5 espie 288: continue;
289: }
290: if (p[1] == '&') {
291: if (mimic_gnu)
292: add_sub(0, string, re, pm);
293: else
294: addchar(p[1]);
295: p++;
296: continue;
297: }
298: if (isdigit(p[1])) {
299: add_sub(*(++p) - '0', string, re, pm);
300: continue;
301: }
302: }
303: addchar(*p);
304: }
305: }
306:
307: static void
1.18 espie 308: do_subst(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
1.5 espie 309: {
310: int error;
1.11 espie 311: int flags = 0;
312: const char *last_match = NULL;
1.5 espie 313:
1.11 espie 314: while ((error = regexec(re, string, re->re_nsub+1, pm, flags)) == 0) {
1.13 espie 315: if (pm[0].rm_eo != 0) {
316: if (string[pm[0].rm_eo-1] == '\n')
317: flags = 0;
318: else
319: flags = REG_NOTBOL;
320: }
1.5 espie 321:
322: /* NULL length matches are special... We use the `vi-mode'
323: * rule: don't allow a NULL-match at the last match
324: * position.
325: */
1.11 espie 326: if (pm[0].rm_so == pm[0].rm_eo &&
327: string + pm[0].rm_so == last_match) {
1.5 espie 328: if (*string == '\0')
329: return;
330: addchar(*string);
1.13 espie 331: if (*string++ == '\n')
332: flags = 0;
333: else
334: flags = REG_NOTBOL;
1.5 espie 335: continue;
336: }
1.11 espie 337: last_match = string + pm[0].rm_so;
338: addchars(string, pm[0].rm_so);
1.5 espie 339: add_replace(string, re, replace, pm);
340: string += pm[0].rm_eo;
341: }
342: if (error != REG_NOMATCH)
343: exit_regerror(error, re);
344: pbstr(string);
345: }
346:
347: static void
1.18 espie 348: do_regexp(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
1.5 espie 349: {
350: int error;
351:
352: switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
353: case 0:
354: add_replace(string, re, replace, pm);
355: pbstr(getstring());
356: break;
357: case REG_NOMATCH:
358: break;
359: default:
360: exit_regerror(error, re);
361: }
362: }
363:
364: static void
1.18 espie 365: do_regexpindex(const char *string, regex_t *re, regmatch_t *pm)
1.5 espie 366: {
367: int error;
368:
369: switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
370: case 0:
371: pbunsigned(pm[0].rm_so);
372: break;
373: case REG_NOMATCH:
374: pbnum(-1);
375: break;
376: default:
377: exit_regerror(error, re);
378: }
379: }
380:
381: /* In Gnu m4 mode, parentheses for backmatch don't work like POSIX 1003.2
382: * says. So we twiddle with the regexp before passing it to regcomp.
383: */
384: static char *
1.18 espie 385: twiddle(const char *p)
1.5 espie 386: {
1.28 espie 387: /* + at start of regexp is a normal character for Gnu m4 */
388: if (*p == '^') {
389: addchar(*p);
390: p++;
391: }
392: if (*p == '+') {
393: addchar('\\');
394: }
1.5 espie 395: /* This could use strcspn for speed... */
396: while (*p != '\0') {
1.14 espie 397: if (*p == '\\') {
398: switch(p[1]) {
399: case '(':
400: case ')':
401: case '|':
402: addchar(p[1]);
403: break;
404: case 'w':
405: addconstantstring("[_a-zA-Z0-9]");
406: break;
407: case 'W':
408: addconstantstring("[^_a-zA-Z0-9]");
409: break;
410: case '<':
411: addconstantstring("[[:<:]]");
412: break;
413: case '>':
414: addconstantstring("[[:>:]]");
415: break;
416: default:
417: addchars(p, 2);
418: break;
419: }
1.5 espie 420: p+=2;
421: continue;
422: }
1.14 espie 423: if (*p == '(' || *p == ')' || *p == '|')
1.5 espie 424: addchar('\\');
425:
426: addchar(*p);
427: p++;
428: }
429: return getstring();
430: }
431:
432: /* patsubst(string, regexp, opt replacement) */
433: /* argv[2]: string
434: * argv[3]: regexp
435: * argv[4]: opt rep
436: */
437: void
1.18 espie 438: dopatsubst(const char *argv[], int argc)
1.5 espie 439: {
440: if (argc <= 3) {
441: warnx("Too few arguments to patsubst");
442: return;
443: }
1.20 espie 444: /* special case: empty regexp */
445: if (argv[3][0] == '\0') {
446: const char *s;
1.21 espie 447: size_t len;
448: if (argv[4] && argc > 4)
449: len = strlen(argv[4]);
450: else
451: len = 0;
1.20 espie 452: for (s = argv[2]; *s != '\0'; s++) {
453: addchars(argv[4], len);
454: addchar(*s);
455: }
456: } else {
457: int error;
458: regex_t re;
459: regmatch_t *pmatch;
460:
461: error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
1.29 ! espie 462: mimic_gnu ? REG_EXTENDED : REG_NEWLINE | REG_EXTENDED);
1.20 espie 463: if (error != 0)
464: exit_regerror(error, &re);
465:
1.26 espie 466: pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1), NULL);
1.20 espie 467: do_subst(argv[2], &re,
1.21 espie 468: argc > 4 && argv[4] != NULL ? argv[4] : "", pmatch);
1.20 espie 469: free(pmatch);
470: regfree(&re);
471: }
1.5 espie 472: pbstr(getstring());
473: }
474:
475: void
1.18 espie 476: doregexp(const char *argv[], int argc)
1.5 espie 477: {
478: int error;
479: regex_t re;
480: regmatch_t *pmatch;
481:
482: if (argc <= 3) {
1.7 espie 483: warnx("Too few arguments to regexp");
1.5 espie 484: return;
485: }
486: error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
487: REG_EXTENDED);
488: if (error != 0)
489: exit_regerror(error, &re);
490:
1.26 espie 491: pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1), NULL);
1.7 espie 492: if (argv[4] == NULL || argc == 4)
1.5 espie 493: do_regexpindex(argv[2], &re, pmatch);
494: else
495: do_regexp(argv[2], &re, argv[4], pmatch);
496: free(pmatch);
497: regfree(&re);
1.8 espie 498: }
499:
500: void
1.18 espie 501: doesyscmd(const char *cmd)
1.8 espie 502: {
503: int p[2];
504: pid_t pid, cpid;
505: char *argv[4];
506: int cc;
507: int status;
508:
509: /* Follow gnu m4 documentation: first flush buffers. */
510: fflush(NULL);
511:
512: argv[0] = "sh";
513: argv[1] = "-c";
514: argv[2] = (char *)cmd;
515: argv[3] = NULL;
516:
517: /* Just set up standard output, share stderr and stdin with m4 */
518: if (pipe(p) == -1)
519: err(1, "bad pipe");
520: switch(cpid = fork()) {
521: case -1:
522: err(1, "bad fork");
523: /* NOTREACHED */
524: case 0:
525: (void) close(p[0]);
526: (void) dup2(p[1], 1);
527: (void) close(p[1]);
528: execv(_PATH_BSHELL, argv);
529: exit(1);
530: default:
531: /* Read result in two stages, since m4's buffer is
532: * pushback-only. */
533: (void) close(p[1]);
534: do {
535: char result[BUFSIZE];
536: cc = read(p[0], result, sizeof result);
537: if (cc > 0)
538: addchars(result, cc);
539: } while (cc > 0 || (cc == -1 && errno == EINTR));
540:
541: (void) close(p[0]);
542: while ((pid = wait(&status)) != cpid && pid >= 0)
543: continue;
544: pbstr(getstring());
545: }
1.5 espie 546: }
![[BACK]](/icons/back.gif){kind=icon}
Return to gnum4.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / m4