Annotation of src/usr.bin/m4/gnum4.c, Revision 1.5
1.5 ! espie 1: /* $OpenBSD: gnum4.c,v 1.4 2000/01/12 17:49:53 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>
! 34: #include <ctype.h>
! 35: #include <regex.h>
1.1 espie 36: #include <stddef.h>
37: #include <stdlib.h>
38: #include <stdio.h>
39: #include <string.h>
40: #include <err.h>
41: #include "mdef.h"
42: #include "stdd.h"
43: #include "extern.h"
44:
1.5 ! espie 45:
! 46: int mimic_gnu = 0;
! 47:
1.1 espie 48: /*
49: * Support for include path search
50: * First search in the the current directory.
51: * If not found, and the path is not absolute, include path kicks in.
52: * First, -I options, in the order found on the command line.
53: * Then M4PATH env variable
54: */
55:
56: struct path_entry {
57: char *name;
58: struct path_entry *next;
59: } *first, *last;
60:
1.3 espie 61: static struct path_entry *new_path_entry __P((const char *));
62: static void ensure_m4path __P((void));
1.4 espie 63: static struct input_file *dopath __P((struct input_file *, const char *));
1.3 espie 64:
1.1 espie 65: static struct path_entry *
66: new_path_entry(dirname)
1.3 espie 67: const char *dirname;
1.1 espie 68: {
69: struct path_entry *n;
70:
71: n = malloc(sizeof(struct path_entry));
72: if (!n)
73: errx(1, "out of memory");
74: n->name = strdup(dirname);
75: if (!n->name)
76: errx(1, "out of memory");
77: n->next = 0;
78: return n;
79: }
80:
81: void
82: addtoincludepath(dirname)
83: const char *dirname;
84: {
85: struct path_entry *n;
86:
87: n = new_path_entry(dirname);
88:
89: if (last) {
90: last->next = n;
91: last = n;
92: }
93: else
94: last = first = n;
95: }
96:
97: static void
98: ensure_m4path()
99: {
100: static int envpathdone = 0;
101: char *envpath;
102: char *sweep;
103: char *path;
104:
105: if (envpathdone)
106: return;
107: envpathdone = TRUE;
108: envpath = getenv("M4PATH");
109: if (!envpath)
110: return;
111: /* for portability: getenv result is read-only */
112: envpath = strdup(envpath);
113: if (!envpath)
114: errx(1, "out of memory");
115: for (sweep = envpath;
116: (path = strsep(&sweep, ":")) != NULL;)
117: addtoincludepath(path);
118: free(envpath);
119: }
120:
121: static
1.4 espie 122: struct input_file *
123: dopath(i, filename)
124: struct input_file *i;
1.3 espie 125: 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 *
142: fopen_trypath(i, filename)
143: struct input_file *i;
1.1 espie 144: const char *filename;
145: {
146: FILE *f;
147:
148: f = fopen(filename, "r");
1.4 espie 149: if (f != NULL) {
150: set_input(i, f, filename);
151: return i;
152: }
1.1 espie 153: if (filename[0] == '/')
154: return NULL;
155:
156: ensure_m4path();
157:
1.4 espie 158: return dopath(i, filename);
1.1 espie 159: }
160:
1.5 ! espie 161: void
! 162: doindir(argv, argc)
! 163: const char *argv[];
! 164: int argc;
! 165: {
! 166: ndptr p;
! 167:
! 168: p = lookup(argv[2]);
! 169: if (p == NULL)
! 170: errx(1, "undefined macro %s", argv[2]);
! 171: argv[1] = p->defn;
! 172: if (p->type == MACRTYPE)
! 173: expand(argv+1, argc-1);
! 174: else
! 175: eval(argv+1, argc-1, p->type);
! 176: }
! 177:
! 178: void
! 179: dobuiltin(argv, argc)
! 180: const char *argv[];
! 181: int argc;
! 182: {
! 183: int n;
! 184: argv[1] = NULL;
! 185: n = builtin_type(argv[2]);
! 186: if (n != -1)
! 187: eval(argv+1, argc, n);
! 188: else
! 189: errx(1, "unknown builtin %s", argv[2]);
! 190: }
! 191:
! 192:
! 193: /* We need some temporary buffer space, as pb pushes BACK and substitution
! 194: * proceeds forward... */
! 195: static char *buffer;
! 196: static size_t bufsize = 0;
! 197: static size_t current = 0;
! 198:
! 199: static void addchars __P((const char *, size_t));
! 200: static void addchar __P((char));
! 201: static char *twiddle __P((const char *));
! 202: static char *getstring __P((void));
! 203: static void exit_regerror __P((int, regex_t *));
! 204: static void do_subst __P((const char *, regex_t *, const char *, regmatch_t *));
! 205: static void do_regexpindex __P((const char *, regex_t *, regmatch_t *));
! 206: static void do_regexp __P((const char *, regex_t *, const char *, regmatch_t *));
! 207: static void add_sub __P((int, const char *, regex_t *, regmatch_t *));
! 208: static void add_replace __P((const char *, regex_t *, const char *, regmatch_t *));
! 209:
! 210: static void
! 211: addchars(c, n)
! 212: const char *c;
! 213: size_t n;
! 214: {
! 215: if (n == 0)
! 216: return;
! 217: if (current + n > bufsize) {
! 218: if (bufsize == 0)
! 219: bufsize = 1024;
! 220: else
! 221: bufsize *= 2;
! 222: buffer = realloc(buffer, bufsize);
! 223: if (buffer == NULL)
! 224: errx(1, "out of memory");
! 225: }
! 226: memcpy(buffer+current, c, n);
! 227: current += n;
! 228: }
! 229:
! 230: static void
! 231: addchar(c)
! 232: char c;
! 233: {
! 234: if (current +1 > bufsize) {
! 235: if (bufsize == 0)
! 236: bufsize = 1024;
! 237: else
! 238: bufsize *= 2;
! 239: buffer = realloc(buffer, bufsize);
! 240: if (buffer == NULL)
! 241: errx(1, "out of memory");
! 242: }
! 243: buffer[current++] = c;
! 244: }
! 245:
! 246: static char *
! 247: getstring()
! 248: {
! 249: addchar('\0');
! 250: current = 0;
! 251: return buffer;
! 252: }
! 253:
! 254:
! 255: static void
! 256: exit_regerror(er, re)
! 257: int er;
! 258: regex_t *re;
! 259: {
! 260: size_t errlen;
! 261: char *errbuf;
! 262:
! 263: errlen = regerror(er, re, NULL, 0);
! 264: errbuf = xalloc(errlen);
! 265: regerror(er, re, errbuf, errlen);
! 266: errx(1, "regular expression error: %s", errbuf);
! 267: }
! 268:
! 269: static void
! 270: add_sub(n, string, re, pm)
! 271: int n;
! 272: const char *string;
! 273: regex_t *re;
! 274: regmatch_t *pm;
! 275: {
! 276: if (n > re->re_nsub)
! 277: warnx("No subexpression %d", n);
! 278: /* Subexpressions that did not match are
! 279: * not an error. */
! 280: else if (pm[n].rm_so != -1 &&
! 281: pm[n].rm_eo != -1) {
! 282: addchars(string + pm[n].rm_so,
! 283: pm[n].rm_eo - pm[n].rm_so);
! 284: }
! 285: }
! 286:
! 287: /* Add replacement string to the output buffer, recognizing special
! 288: * constructs and replacing them with substrings of the original string.
! 289: */
! 290: static void
! 291: add_replace(string, re, replace, pm)
! 292: const char *string;
! 293: regex_t *re;
! 294: const char *replace;
! 295: regmatch_t *pm;
! 296: {
! 297: const char *p;
! 298:
! 299: for (p = replace; *p != '\0'; p++) {
! 300: if (*p == '&' && !mimic_gnu) {
! 301: add_sub(0, string, re, pm);
! 302: continue;
! 303: }
! 304: if (*p == '\\') {
! 305: if (p[1] == '\\') {
! 306: addchar(p[1]);
! 307: continue;
! 308: }
! 309: if (p[1] == '&') {
! 310: if (mimic_gnu)
! 311: add_sub(0, string, re, pm);
! 312: else
! 313: addchar(p[1]);
! 314: p++;
! 315: continue;
! 316: }
! 317: if (isdigit(p[1])) {
! 318: add_sub(*(++p) - '0', string, re, pm);
! 319: continue;
! 320: }
! 321: }
! 322: addchar(*p);
! 323: }
! 324: }
! 325:
! 326: static void
! 327: do_subst(string, re, replace, pm)
! 328: const char *string;
! 329: regex_t *re;
! 330: const char *replace;
! 331: regmatch_t *pm;
! 332: {
! 333: int error;
! 334: regoff_t last_match = -1;
! 335:
! 336: while ((error = regexec(re, string, re->re_nsub+1, pm, 0)) == 0) {
! 337:
! 338: /* NULL length matches are special... We use the `vi-mode'
! 339: * rule: don't allow a NULL-match at the last match
! 340: * position.
! 341: */
! 342: if (pm[0].rm_so == pm[0].rm_eo && pm[0].rm_so == last_match) {
! 343: if (*string == '\0')
! 344: return;
! 345: addchar(*string);
! 346: string++;
! 347: continue;
! 348: }
! 349: last_match = pm[0].rm_so;
! 350: addchars(string, last_match);
! 351: add_replace(string, re, replace, pm);
! 352: string += pm[0].rm_eo;
! 353: }
! 354: if (error != REG_NOMATCH)
! 355: exit_regerror(error, re);
! 356: pbstr(string);
! 357: }
! 358:
! 359: static void
! 360: do_regexp(string, re, replace, pm)
! 361: const char *string;
! 362: regex_t *re;
! 363: const char *replace;
! 364: regmatch_t *pm;
! 365: {
! 366: int error;
! 367: const char *p;
! 368:
! 369: switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
! 370: case 0:
! 371: add_replace(string, re, replace, pm);
! 372: pbstr(getstring());
! 373: break;
! 374: case REG_NOMATCH:
! 375: break;
! 376: default:
! 377: exit_regerror(error, re);
! 378: }
! 379: }
! 380:
! 381: static void
! 382: do_regexpindex(string, re, pm)
! 383: const char *string;
! 384: regex_t *re;
! 385: regmatch_t *pm;
! 386: {
! 387: int error;
! 388:
! 389: switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
! 390: case 0:
! 391: pbunsigned(pm[0].rm_so);
! 392: break;
! 393: case REG_NOMATCH:
! 394: pbnum(-1);
! 395: break;
! 396: default:
! 397: exit_regerror(error, re);
! 398: }
! 399: }
! 400:
! 401: /* In Gnu m4 mode, parentheses for backmatch don't work like POSIX 1003.2
! 402: * says. So we twiddle with the regexp before passing it to regcomp.
! 403: */
! 404: static char *
! 405: twiddle(p)
! 406: const char *p;
! 407: {
! 408: /* This could use strcspn for speed... */
! 409: while (*p != '\0') {
! 410: if (*p == '\\' && (p[1] == '(' || p[1] == ')')) {
! 411: addchar(p[1]);
! 412: p+=2;
! 413: continue;
! 414: }
! 415: if (*p == '(' || *p == ')')
! 416: addchar('\\');
! 417:
! 418: addchar(*p);
! 419: p++;
! 420: }
! 421: return getstring();
! 422: }
! 423:
! 424: /* patsubst(string, regexp, opt replacement) */
! 425: /* argv[2]: string
! 426: * argv[3]: regexp
! 427: * argv[4]: opt rep
! 428: */
! 429: void
! 430: dopatsubst(argv, argc)
! 431: const char *argv[];
! 432: int argc;
! 433: {
! 434: int error;
! 435: regex_t re;
! 436: regmatch_t *pmatch;
! 437:
! 438: if (argc <= 3) {
! 439: warnx("Too few arguments to patsubst");
! 440: return;
! 441: }
! 442: error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
! 443: REG_EXTENDED);
! 444: if (error != 0)
! 445: exit_regerror(error, &re);
! 446:
! 447: pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
! 448: do_subst(argv[2], &re, argv[4] != NULL ? argv[4] : "", pmatch);
! 449: pbstr(getstring());
! 450: free(pmatch);
! 451: regfree(&re);
! 452: }
! 453:
! 454: void
! 455: doregexp(argv, argc)
! 456: const char *argv[];
! 457: int argc;
! 458: {
! 459: int error;
! 460: regex_t re;
! 461: regmatch_t *pmatch;
! 462:
! 463: if (argc <= 3) {
! 464: warnx("Too few arguments to patsubst");
! 465: return;
! 466: }
! 467: error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
! 468: REG_EXTENDED);
! 469: if (error != 0)
! 470: exit_regerror(error, &re);
! 471:
! 472: pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
! 473: if (argv[4] == NULL)
! 474: do_regexpindex(argv[2], &re, pmatch);
! 475: else
! 476: do_regexp(argv[2], &re, argv[4], pmatch);
! 477: free(pmatch);
! 478: regfree(&re);
! 479: }