[BACK]Return to util.c CVS log [TXT][DIR] Up to [cvsweb.bsd.lv] / texi2mdoc

Annotation of texi2mdoc/util.c, Revision 1.19

1.19    ! kristaps    1: /*     $Id: util.c,v 1.18 2015/02/28 08:41:59 kristaps Exp $ */
1.1       kristaps    2: /*
                      3:  * Copyright (c) 2015 Kristaps Dzonsons <kristaps@bsd.lv>
                      4:  *
                      5:  * Permission to use, copy, modify, and distribute this software for any
                      6:  * purpose with or without fee is hereby granted, provided that the above
                      7:  * copyright notice and this permission notice appear in all copies.
                      8:  *
                      9:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
                     10:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
                     11:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
                     12:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     13:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     14:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     15:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
                     16:  */
                     17: #include <sys/mman.h>
                     18: #include <sys/stat.h>
                     19:
                     20: #include <assert.h>
                     21: #include <ctype.h>
                     22: #include <fcntl.h>
                     23: #include <getopt.h>
                     24: #include <libgen.h>
                     25: #include <limits.h>
                     26: #include <stdarg.h>
                     27: #include <stdio.h>
                     28: #include <stdlib.h>
                     29: #include <string.h>
                     30: #include <time.h>
                     31: #include <unistd.h>
                     32:
                     33: #include "extern.h"
                     34:
                     35: /*
                     36:  * Unmap the top-most file in the stack of files currently opened (that
                     37:  * is, nested calls to parsefile()).
                     38:  */
                     39: void
                     40: texifilepop(struct texi *p)
                     41: {
                     42:        struct texifile *f;
                     43:
                     44:        assert(p->filepos > 0);
                     45:        f = &p->files[--p->filepos];
1.14      kristaps   46:        free(f->map);
1.1       kristaps   47: }
                     48:
1.7       kristaps   49: static void
                     50: teximacrofree(struct teximacro *p)
                     51: {
                     52:        size_t   i;
                     53:
                     54:        for (i = 0; i < p->argsz; i++)
                     55:                free(p->args[i]);
                     56:
                     57:        free(p->args);
                     58:        free(p->key);
                     59:        free(p->value);
                     60: }
                     61:
                     62: static void
                     63: texivaluefree(struct texivalue *p)
                     64: {
                     65:
                     66:        free(p->key);
                     67:        free(p->value);
                     68: }
                     69:
1.1       kristaps   70: /*
                     71:  * Unmap all files that we're currently using and free all resources
                     72:  * that we've allocated during the parse.
                     73:  * The utility should exit(...) after this is called.
                     74:  */
                     75: void
                     76: texiexit(struct texi *p)
                     77: {
                     78:        size_t   i;
                     79:
                     80:        /* Make sure we're newline-terminated. */
                     81:        if (p->outcol)
                     82:                putchar('\n');
                     83:
                     84:        /* Unmap all files. */
                     85:        while (p->filepos > 0)
                     86:                texifilepop(p);
                     87:
1.7       kristaps   88:        for (i = 0; i < p->macrosz; i++)
                     89:                teximacrofree(&p->macros[i]);
1.1       kristaps   90:        for (i = 0; i < p->dirsz; i++)
                     91:                free(p->dirs[i]);
1.4       kristaps   92:        for (i = 0; i < p->indexsz; i++)
                     93:                free(p->indexs[i]);
1.7       kristaps   94:        for (i = 0; i < p->valsz; i++)
                     95:                texivaluefree(&p->vals[i]);
1.4       kristaps   96:
1.7       kristaps   97:        free(p->macros);
1.1       kristaps   98:        free(p->vals);
1.4       kristaps   99:        free(p->indexs);
1.1       kristaps  100:        free(p->dirs);
                    101:        free(p->subtitle);
                    102:        free(p->title);
                    103: }
                    104:
                    105: /*
                    106:  * Fatal error: unmap all files and exit.
                    107:  * The "errstring" is passed to perror(3).
                    108:  */
                    109: void
                    110: texiabort(struct texi *p, const char *errstring)
                    111: {
                    112:
                    113:        perror(errstring);
                    114:        texiexit(p);
                    115:        exit(EXIT_FAILURE);
                    116: }
                    117:
                    118: /*
                    119:  * Print a generic warning message (to stderr) tied to our current
                    120:  * location in the parse sequence.
                    121:  */
                    122: void
                    123: texiwarn(const struct texi *p, const char *fmt, ...)
                    124: {
1.15      kristaps  125:        va_list                  ap;
                    126:        const struct texifile   *f;
                    127:
                    128:        f = &p->files[p->filepos - 1];
                    129:
                    130:        if (f->insplice)
                    131:                fprintf(stderr, "%s:%zu:%zu (%zuB left in splice): "
                    132:                        "warning: ", f->name, f->line + 1,
                    133:                        f->col + 1, f->insplice);
                    134:        else
                    135:                fprintf(stderr, "%s:%zu:%zu: warning: ",
                    136:                        f->name, f->line + 1, f->col + 1);
1.1       kristaps  137:
                    138:        va_start(ap, fmt);
                    139:        vfprintf(stderr, fmt, ap);
                    140:        va_end(ap);
                    141:        fputc('\n', stderr);
                    142: }
                    143:
                    144: /*
                    145:  * Print an error message (to stderr) tied to our current location in
                    146:  * the parse sequence, invoke texiexit(), then die.
                    147:  */
                    148: void
                    149: texierr(struct texi *p, const char *fmt, ...)
                    150: {
1.15      kristaps  151:        va_list          ap;
                    152:        struct texifile *f;
                    153:
                    154:        f = &p->files[p->filepos - 1];
                    155:
                    156:        if (f->insplice)
                    157:                fprintf(stderr, "%s:%zu:%zu: (%zuB left in splice): "
                    158:                        "error: ", f->name, f->line + 1,
                    159:                        f->col + 1, f->insplice);
                    160:        else
                    161:                fprintf(stderr, "%s:%zu:%zu: error: ",
                    162:                        f->name, f->line + 1, f->col + 1);
1.1       kristaps  163:
                    164:        va_start(ap, fmt);
                    165:        vfprintf(stderr, fmt, ap);
                    166:        va_end(ap);
                    167:        fputc('\n', stderr);
                    168:        texiexit(p);
                    169:        exit(EXIT_FAILURE);
                    170: }
                    171:
                    172: /*
                    173:  * Put a single data character to the output if we're not ignoring.
1.13      kristaps  174:  * Escape starting a line with a control character and slashes.
1.1       kristaps  175:  */
                    176: void
                    177: texiputchar(struct texi *p, char c)
                    178: {
                    179:
                    180:        if (p->ign)
                    181:                return;
                    182:        if ('.' == c && 0 == p->outcol)
                    183:                fputs("\\&", stdout);
1.10      kristaps  184:        if ('\'' == c && 0 == p->outcol)
                    185:                fputs("\\&", stdout);
1.1       kristaps  186:
                    187:        putchar(c);
1.13      kristaps  188:        if ('\\' == c)
                    189:                putchar('e');
1.1       kristaps  190:        p->seenvs = 0;
                    191:        if ('\n' == c) {
                    192:                p->outcol = 0;
                    193:                p->seenws = 0;
                    194:        } else
                    195:                p->outcol++;
                    196: }
                    197:
                    198: /*
1.13      kristaps  199:  * Put an opaque series of characters.
                    200:  * Characters starting a line with a control character are escaped, but
                    201:  * that's it, so don't use this for non-controlled sequences of text.
1.1       kristaps  202:  */
                    203: void
                    204: texiputchars(struct texi *p, const char *s)
                    205: {
                    206:
1.13      kristaps  207:        if (p->ign)
                    208:                return;
                    209:        if ('.' == *s && 0 == p->outcol)
                    210:                fputs("\\&", stdout);
                    211:        if ('\'' == *s && 0 == p->outcol)
                    212:                fputs("\\&", stdout);
                    213:        p->outcol += fputs(s, stdout);
                    214:        p->seenvs = 0;
1.9       kristaps  215: }
                    216:
                    217: /*
                    218:  * This puts all characters onto the output stream but makes sure to
                    219:  * escape mdoc(7) slashes.
1.14      kristaps  220:  * FIXME: useless.
1.9       kristaps  221:  */
                    222: void
1.14      kristaps  223: texiputbuf(struct texi *p, size_t start, size_t end)
1.9       kristaps  224: {
                    225:
1.14      kristaps  226:        for ( ; start < end; start++)
                    227:                texiputchar(p, BUF(p)[start]);
1.1       kristaps  228: }
                    229:
                    230: /*
                    231:  * Close an mdoc(7) macro opened with teximacroopen().
                    232:  * If there are no more macros on the line, prints a newline.
                    233:  */
                    234: void
                    235: teximacroclose(struct texi *p)
                    236: {
                    237:
                    238:        if (p->ign)
                    239:                return;
                    240:
                    241:        if (0 == --p->outmacro) {
                    242:                putchar('\n');
                    243:                p->outcol = p->seenws = 0;
                    244:        }
                    245: }
                    246:
                    247: /*
                    248:  * Open a mdoc(7) macro.
                    249:  * This is used for line macros, e.g., Qq [foo bar baz].
                    250:  * It can be invoked for nested macros, e.g., Qq Li foo .
                    251:  * TODO: flush-right punctuation (e.g., parenthesis).
                    252:  */
                    253: void
                    254: teximacroopen(struct texi *p, const char *s)
                    255: {
                    256:        int      rc;
                    257:
                    258:        if (p->ign)
                    259:                return;
                    260:
                    261:        if (p->outcol && 0 == p->outmacro) {
                    262:                putchar('\n');
                    263:                p->outcol = 0;
                    264:        }
                    265:
                    266:        if (0 == p->outmacro)
                    267:                putchar('.');
                    268:        else
                    269:                putchar(' ');
                    270:
                    271:        if (EOF != (rc = fputs(s, stdout)))
                    272:                p->outcol += rc;
                    273:
                    274:        putchar(' ');
                    275:        p->outcol++;
                    276:        p->outmacro++;
                    277:        p->seenws = 0;
                    278: }
                    279:
                    280: /*
                    281:  * Put a stadnalone mdoc(7) command with the trailing newline.
                    282:  */
                    283: void
                    284: teximacro(struct texi *p, const char *s)
                    285: {
                    286:
                    287:        if (p->ign)
                    288:                return;
                    289:
                    290:        if (p->outmacro)
                    291:                texierr(p, "\"%s\" in open line scope!?", s);
                    292:        if (p->literal)
                    293:                texierr(p, "\"%s\" in a literal scope!?", s);
                    294:
                    295:        if (p->outcol)
                    296:                putchar('\n');
                    297:
                    298:        putchar('.');
                    299:        puts(s);
                    300:        p->outcol = p->seenws = 0;
                    301: }
                    302:
                    303: /*
                    304:  * Introduce vertical space during normal (non-macro) input.
                    305:  */
                    306: void
                    307: texivspace(struct texi *p)
                    308: {
                    309:
1.5       kristaps  310:        if (p->seenvs || TEXILIST_TABLE == p->list)
1.1       kristaps  311:                return;
                    312:        teximacro(p, "Pp");
                    313:        p->seenvs = 1;
                    314: }
                    315:
                    316: /*
                    317:  * Advance by a single byte in the input stream, adjusting our location
                    318:  * in the current input file.
                    319:  */
                    320: void
1.14      kristaps  321: advance(struct texi *p, size_t *pos)
1.1       kristaps  322: {
1.15      kristaps  323:        struct texifile *f;
1.1       kristaps  324:
1.15      kristaps  325:        f = &p->files[p->filepos - 1];
                    326:
                    327:        if (0 == f->insplice) {
                    328:                if ('\n' == BUF(p)[*pos]) {
                    329:                        f->line++;
                    330:                        f->col = 0;
                    331:                } else
                    332:                        f->col++;
1.17      kristaps  333:        } else {
1.15      kristaps  334:                --f->insplice;
1.17      kristaps  335:                if (0 == f->insplice)
                    336:                        f->depth = 0;
                    337:        }
1.1       kristaps  338:
                    339:        (*pos)++;
                    340: }
                    341:
                    342: /*
                    343:  * It's common to wait punctuation to float on the right side of macro
                    344:  * lines in mdoc(7), e.g., ".Em hello ) ."
                    345:  * This function does so, and should be called before teximacroclose().
                    346:  * It will detect that it's the last in the nested macros and
                    347:  * appropriately flush-left punctuation alongside the macro.
                    348:  */
                    349: void
1.14      kristaps  350: texipunctuate(struct texi *p, size_t *pos)
1.1       kristaps  351: {
                    352:        size_t   start, end;
                    353:
                    354:        if (1 != p->outmacro)
                    355:                return;
                    356:
1.14      kristaps  357:        for (start = end = *pos; end < BUFSZ(p); end++) {
                    358:                switch (BUF(p)[end]) {
1.1       kristaps  359:                case (','):
                    360:                case (')'):
                    361:                case ('.'):
                    362:                case ('"'):
                    363:                case (':'):
                    364:                case ('!'):
                    365:                case ('?'):
                    366:                        continue;
                    367:                default:
                    368:                        break;
                    369:                }
                    370:                break;
                    371:        }
                    372:        if (end == *pos)
                    373:                return;
1.14      kristaps  374:        if (end + 1 == BUFSZ(p) || ' ' == BUF(p)[end] ||
                    375:                '\n' == BUF(p)[end]) {
1.1       kristaps  376:                for ( ; start < end; start++) {
                    377:                        texiputchar(p, ' ');
1.14      kristaps  378:                        texiputchar(p, BUF(p)[start]);
                    379:                        advance(p, pos);
1.1       kristaps  380:                }
                    381:        }
                    382: }
                    383:
                    384: /*
                    385:  * Advance to the next non-whitespace word in the input stream.
                    386:  * If we're in literal mode, then print all of the whitespace as we're
                    387:  * doing so.
                    388:  */
                    389: static size_t
1.14      kristaps  390: advancenext(struct texi *p, size_t *pos)
1.1       kristaps  391: {
                    392:
                    393:        if (p->literal) {
1.14      kristaps  394:                while (*pos < BUFSZ(p) && ismspace(BUF(p)[*pos])) {
                    395:                        texiputchar(p, BUF(p)[*pos]);
                    396:                        advance(p, pos);
1.1       kristaps  397:                }
                    398:                return(*pos);
                    399:        }
                    400:
1.14      kristaps  401:        while (*pos < BUFSZ(p) && ismspace(BUF(p)[*pos])) {
1.1       kristaps  402:                p->seenws = 1;
                    403:                /*
                    404:                 * If it looks like we've printed a double-line, then
                    405:                 * output a paragraph.
                    406:                 * FIXME: this is stupid.
                    407:                 */
1.14      kristaps  408:                if (*pos && '\n' == BUF(p)[*pos] && '\n' == BUF(p)[*pos - 1])
1.1       kristaps  409:                        texivspace(p);
1.14      kristaps  410:                advance(p, pos);
1.1       kristaps  411:        }
                    412:        return(*pos);
                    413: }
                    414:
                    415: /*
                    416:  * Advance to the EOLN in the input stream.
                    417:  * NOTE: THIS SHOULD NOT BE CALLED ON BLANK TEXT, as it will read up to
                    418:  * the @\n.
                    419:  */
                    420: size_t
1.14      kristaps  421: advanceeoln(struct texi *p, size_t *pos, int consumenl)
1.1       kristaps  422: {
                    423:
1.14      kristaps  424:        while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos])
                    425:                advance(p, pos);
                    426:        if (*pos < BUFSZ(p) && consumenl)
                    427:                advance(p, pos);
1.1       kristaps  428:        return(*pos);
                    429: }
                    430:
                    431: /*
                    432:  * Advance to position "end", which is an absolute position in the
                    433:  * current buffer greater than or equal to the current position.
                    434:  */
                    435: void
1.14      kristaps  436: advanceto(struct texi *p, size_t *pos, size_t end)
1.1       kristaps  437: {
                    438:
                    439:        assert(*pos <= end);
                    440:        while (*pos < end)
1.14      kristaps  441:                advance(p, pos);
1.1       kristaps  442: }
                    443:
1.7       kristaps  444: static void
1.17      kristaps  445: texiexecmacro(struct texi *p, struct teximacro *m, size_t sv, size_t *pos)
1.7       kristaps  446: {
1.11      kristaps  447:        size_t            valsz, realsz, aasz, asz,
                    448:                           ssz, i, j, k, start, end;
                    449:        char             *val;
                    450:        char            **args;
                    451:        const char       *cp;
1.7       kristaps  452:
1.17      kristaps  453:        /* Disregard empty macros. */
                    454:        if (0 == (valsz = realsz = strlen(m->value)))
                    455:                return;
                    456:
                    457:        /*
                    458:         * This is important: it protect us from macros that invoke more
                    459:         * macros, possibly going on infinitely.
                    460:         * We use "sv" instead of the current position because we might
                    461:         * be invoked at the end of the macro (i.e., insplice == 0).
                    462:         * The "sv" value was initialised at the start of the macro.
                    463:         */
                    464:        if (sv > 0)
                    465:                if (++p->files[p->filepos].depth > 64)
                    466:                        texierr(p, "maximium recursive depth");
                    467:
1.14      kristaps  468:        args = argparse(p, pos, &asz, m->argsz);
1.7       kristaps  469:        if (asz != m->argsz)
                    470:                texiwarn(p, "invalid macro argument length");
                    471:        aasz = asz < m->argsz ? asz : m->argsz;
                    472:
                    473:        if (0 == aasz) {
1.17      kristaps  474:                texisplice(p, m->value, valsz, pos);
1.7       kristaps  475:                return;
                    476:        }
                    477:
                    478:        val = strdup(m->value);
                    479:
                    480:        for (i = j = 0; i < realsz; i++) {
                    481:                /* Parse blindly til the backslash delimiter. */
                    482:                if ('\\' != m->value[i]) {
                    483:                        val[j++] = m->value[i];
                    484:                        val[j] = '\0';
                    485:                        continue;
                    486:                } else if (i == realsz - 1)
                    487:                        texierr(p, "trailing argument name delimiter");
                    488:
                    489:                /* Double-backslash is escaped. */
                    490:                if ('\\' == m->value[i + 1]) {
                    491:                        val[j++] = m->value[i++];
                    492:                        val[j] = '\0';
                    493:                        continue;
                    494:                }
                    495:
                    496:                assert('\\' == m->value[i] && i < realsz - 1);
                    497:
                    498:                /* Parse to terminating delimiter. */
                    499:                /* FIXME: embedded, escaped delimiters? */
                    500:                for (start = end = i + 1; end < realsz; end++)
                    501:                        if ('\\' == m->value[end])
                    502:                                break;
                    503:                if (end == realsz)
                    504:                        texierr(p, "unterminated argument name");
                    505:
                    506:                for (k = 0; k < aasz; k++) {
                    507:                        if ((ssz = strlen(m->args[k])) != (end - start))
                    508:                                continue;
                    509:                        if (strncmp(&m->value[start], m->args[k], ssz))
                    510:                                continue;
                    511:                        break;
                    512:                }
                    513:
                    514:                /*
                    515:                 * Argument didn't exist in argument table.
1.14      kristaps  516:                 * Just ignore it.
1.7       kristaps  517:                 */
                    518:                if (k == aasz) {
1.14      kristaps  519:                        i = end;
1.7       kristaps  520:                        continue;
                    521:                }
                    522:
                    523:                if (strlen(args[k]) > ssz) {
                    524:                        valsz += strlen(args[k]);
                    525:                        val = realloc(val, valsz + 1);
                    526:                        if (NULL == val)
                    527:                                texiabort(p, NULL);
                    528:                }
                    529:
1.11      kristaps  530:                for (cp = args[k]; '\0' != *cp; cp++)
                    531:                        val[j++] = *cp;
                    532:
                    533:                val[j] = '\0';
1.7       kristaps  534:                i = end;
                    535:        }
                    536:
1.14      kristaps  537:        texisplice(p, val, strlen(val), pos);
1.7       kristaps  538:
                    539:        for (i = 0; i < asz; i++)
                    540:                free(args[i]);
                    541:        free(args);
                    542:        free(val);
                    543: }
                    544:
1.1       kristaps  545: /*
                    546:  * Output a free-form word in the input stream, progressing to the next
                    547:  * command or white-space.
                    548:  * This also will advance the input stream.
                    549:  */
                    550: static void
1.14      kristaps  551: parseword(struct texi *p, size_t *pos, char extra)
1.1       kristaps  552: {
                    553:
                    554:        if (p->seenws && 0 == p->outmacro &&
                    555:                 p->outcol > 72 && 0 == p->literal)
                    556:                texiputchar(p, '\n');
                    557:        /* FIXME: abstract this: we use it elsewhere. */
                    558:        if (p->seenws && p->outcol && 0 == p->literal)
                    559:                texiputchar(p, ' ');
                    560:
                    561:        p->seenws = 0;
                    562:
1.14      kristaps  563:        while (*pos < BUFSZ(p) && ! ismspace(BUF(p)[*pos])) {
                    564:                switch (BUF(p)[*pos]) {
1.1       kristaps  565:                case ('@'):
                    566:                case ('}'):
                    567:                case ('{'):
                    568:                        return;
                    569:                }
1.14      kristaps  570:                if ('\0' != extra && BUF(p)[*pos] == extra)
1.1       kristaps  571:                        return;
1.14      kristaps  572:                if (*pos < BUFSZ(p) - 1 &&
                    573:                         '`' == BUF(p)[*pos] &&
                    574:                         '`' == BUF(p)[*pos + 1]) {
1.1       kristaps  575:                        texiputchars(p, "\\(lq");
1.14      kristaps  576:                        advance(p, pos);
                    577:                } else if (*pos < BUFSZ(p) - 1 &&
                    578:                         '\'' == BUF(p)[*pos] &&
                    579:                         '\'' == BUF(p)[*pos + 1]) {
1.1       kristaps  580:                        texiputchars(p, "\\(rq");
1.14      kristaps  581:                        advance(p, pos);
1.1       kristaps  582:                } else
1.14      kristaps  583:                        texiputchar(p, BUF(p)[*pos]);
                    584:                advance(p, pos);
1.1       kristaps  585:        }
                    586: }
                    587:
                    588: /*
                    589:  * Look up the command at position "pos" in the buffer, returning it (or
                    590:  * TEXICMD__MAX if none found) and setting "end" to be the absolute
                    591:  * index after the command name.
                    592:  */
                    593: enum texicmd
1.19    ! kristaps  594: texicmd(const struct texi *p, size_t pos, size_t *end, struct teximacro **macro)
1.1       kristaps  595: {
1.4       kristaps  596:        size_t   i, len, toksz;
1.1       kristaps  597:
1.14      kristaps  598:        assert('@' == BUF(p)[pos]);
1.1       kristaps  599:
1.7       kristaps  600:        if (NULL != macro)
                    601:                *macro = NULL;
                    602:
1.14      kristaps  603:        if ((*end = pos) == BUFSZ(p))
1.1       kristaps  604:                return(TEXICMD__MAX);
1.14      kristaps  605:        else if ((*end = ++pos) == BUFSZ(p))
1.1       kristaps  606:                return(TEXICMD__MAX);
                    607:
                    608:        /* Alphabetic commands are special. */
1.16      kristaps  609:        if ( ! isalpha((unsigned char)BUF(p)[pos])) {
1.14      kristaps  610:                if ((*end = pos + 1) == BUFSZ(p))
1.1       kristaps  611:                        return(TEXICMD__MAX);
                    612:                for (i = 0; i < TEXICMD__MAX; i++) {
                    613:                        if (1 != texitoks[i].len)
                    614:                                continue;
1.14      kristaps  615:                        if (0 == strncmp(texitoks[i].tok, &BUF(p)[pos], 1))
1.1       kristaps  616:                                return(i);
                    617:                }
1.14      kristaps  618:                texiwarn(p, "bad command: @%c", BUF(p)[pos]);
1.1       kristaps  619:                return(TEXICMD__MAX);
                    620:        }
                    621:
1.4       kristaps  622:        /* Scan to the end of the possible command name. */
1.14      kristaps  623:        for (*end = pos; *end < BUFSZ(p) && ! ismspace(BUF(p)[*end]); (*end)++)
                    624:                if ((*end > pos && ('@' == BUF(p)[*end] ||
                    625:                          '{' == BUF(p)[*end] || '}' == BUF(p)[*end])))
1.1       kristaps  626:                        break;
                    627:
1.4       kristaps  628:        /* Look for the command. */
1.1       kristaps  629:        len = *end - pos;
                    630:        for (i = 0; i < TEXICMD__MAX; i++) {
                    631:                if (len != texitoks[i].len)
                    632:                        continue;
1.14      kristaps  633:                if (0 == strncmp(texitoks[i].tok, &BUF(p)[pos], len))
1.1       kristaps  634:                        return(i);
                    635:        }
                    636:
1.4       kristaps  637:        /* Look for it in our indices. */
                    638:        for (i = 0; i < p->indexsz; i++) {
                    639:                toksz = strlen(p->indexs[i]);
                    640:                if (len != 5 + toksz)
                    641:                        continue;
1.14      kristaps  642:                if (strncmp(&BUF(p)[pos], p->indexs[i], toksz))
1.4       kristaps  643:                        continue;
1.14      kristaps  644:                if (0 == strncmp(&BUF(p)[pos + toksz], "index", 5))
1.7       kristaps  645:                        return(TEXICMD_USER_INDEX);
                    646:        }
                    647:
                    648:        for (i = 0; i < p->macrosz; i++) {
                    649:                if (len != strlen(p->macros[i].key))
                    650:                        continue;
1.14      kristaps  651:                if (strncmp(&BUF(p)[pos], p->macros[i].key, len))
1.7       kristaps  652:                        continue;
                    653:                if (NULL != macro)
                    654:                        *macro = &p->macros[i];
                    655:                return(TEXICMD__MAX);
1.4       kristaps  656:        }
                    657:
1.14      kristaps  658:        texiwarn(p, "bad command: @%.*s", (int)len, &BUF(p)[pos]);
1.1       kristaps  659:        return(TEXICMD__MAX);
                    660: }
                    661:
                    662: /*
                    663:  * Parse an argument from a bracketed command, e.g., @url{foo, baz}.
                    664:  * Num should be set to the argument we're currently parsing, although
                    665:  * it suffixes for it to be zero or non-zero.
                    666:  * This will return 1 if there are more arguments, 0 otherwise.
                    667:  * This will stop (returning 0) in the event of EOF or if we're not at a
                    668:  * bracket for the zeroth parse.
                    669:  */
                    670: int
1.14      kristaps  671: parsearg(struct texi *p, size_t *pos, size_t num)
1.1       kristaps  672: {
1.17      kristaps  673:        size_t            end, sv;
1.7       kristaps  674:        enum texicmd      cmd;
                    675:        struct teximacro *macro;
1.1       kristaps  676:
1.14      kristaps  677:        while (*pos < BUFSZ(p) && ismspace(BUF(p)[*pos]))
                    678:                advance(p, pos);
                    679:        if (*pos == BUFSZ(p) || (0 == num && '{' != BUF(p)[*pos]))
1.1       kristaps  680:                return(0);
                    681:        if (0 == num)
1.14      kristaps  682:                advance(p, pos);
1.1       kristaps  683:
1.14      kristaps  684:        while ((*pos = advancenext(p, pos)) < BUFSZ(p)) {
                    685:                switch (BUF(p)[*pos]) {
1.1       kristaps  686:                case (','):
1.14      kristaps  687:                        advance(p, pos);
1.1       kristaps  688:                        return(1);
                    689:                case ('}'):
1.14      kristaps  690:                        advance(p, pos);
1.1       kristaps  691:                        return(0);
                    692:                case ('{'):
                    693:                        if (0 == p->ign)
                    694:                                texiwarn(p, "unexpected \"{\"");
1.14      kristaps  695:                        advance(p, pos);
1.1       kristaps  696:                        continue;
                    697:                case ('@'):
                    698:                        break;
                    699:                default:
1.14      kristaps  700:                        parseword(p, pos, ',');
1.1       kristaps  701:                        continue;
                    702:                }
                    703:
1.17      kristaps  704:                sv = p->files[p->filepos - 1].insplice;
1.14      kristaps  705:                cmd = texicmd(p, *pos, &end, &macro);
                    706:                advanceto(p, pos, end);
1.7       kristaps  707:                if (NULL != macro)
1.17      kristaps  708:                        texiexecmacro(p, macro, sv, pos);
1.1       kristaps  709:                if (TEXICMD__MAX == cmd)
                    710:                        continue;
                    711:                if (NULL != texitoks[cmd].fp)
1.14      kristaps  712:                        (*texitoks[cmd].fp)(p, cmd, pos);
1.1       kristaps  713:        }
                    714:        return(0);
                    715: }
                    716:
                    717: /*
                    718:  * Parse until the end of a bracketed statement, e.g., @foo{bar baz}.
                    719:  * This will stop in the event of EOF or if we're not at a bracket.
                    720:  */
                    721: void
1.18      kristaps  722: parsebracket(struct texi *p, size_t *pos, int dostack)
1.1       kristaps  723: {
1.18      kristaps  724:        size_t            end, sv, stack;
1.7       kristaps  725:        enum texicmd      cmd;
                    726:        struct teximacro *macro;
1.1       kristaps  727:
1.14      kristaps  728:        while (*pos < BUFSZ(p) && ismspace(BUF(p)[*pos]))
                    729:                advance(p, pos);
1.1       kristaps  730:
1.14      kristaps  731:        if (*pos == BUFSZ(p) || '{' != BUF(p)[*pos])
1.1       kristaps  732:                return;
1.14      kristaps  733:        advance(p, pos);
1.1       kristaps  734:
1.18      kristaps  735:        stack = 0;
1.14      kristaps  736:        while ((*pos = advancenext(p, pos)) < BUFSZ(p)) {
                    737:                switch (BUF(p)[*pos]) {
1.1       kristaps  738:                case ('}'):
1.18      kristaps  739:                        if (stack > 0) {
                    740:                                stack--;
                    741:                                advance(p, pos);
                    742:                                texiputchar(p, '}');
                    743:                                continue;
                    744:                        }
1.14      kristaps  745:                        advance(p, pos);
1.1       kristaps  746:                        return;
                    747:                case ('{'):
1.18      kristaps  748:                        if (dostack) {
                    749:                                stack++;
                    750:                                advance(p, pos);
                    751:                                texiputchar(p, '{');
                    752:                                continue;
                    753:                        }
1.1       kristaps  754:                        if (0 == p->ign)
                    755:                                texiwarn(p, "unexpected \"{\"");
1.14      kristaps  756:                        advance(p, pos);
1.1       kristaps  757:                        continue;
                    758:                case ('@'):
                    759:                        break;
                    760:                default:
1.14      kristaps  761:                        parseword(p, pos, '\0');
1.1       kristaps  762:                        continue;
                    763:                }
                    764:
1.17      kristaps  765:                sv = p->files[p->filepos - 1].insplice;
1.14      kristaps  766:                cmd = texicmd(p, *pos, &end, &macro);
                    767:                advanceto(p, pos, end);
1.7       kristaps  768:                if (NULL != macro)
1.17      kristaps  769:                        texiexecmacro(p, macro, sv, pos);
1.1       kristaps  770:                if (TEXICMD__MAX == cmd)
                    771:                        continue;
                    772:                if (NULL != texitoks[cmd].fp)
1.14      kristaps  773:                        (*texitoks[cmd].fp)(p, cmd, pos);
1.1       kristaps  774:        }
                    775: }
                    776:
                    777: /*
                    778:  * This should be invoked when we're on a macro line and want to process
                    779:  * to the end of the current input line, doing all of our macros along
                    780:  * the way.
                    781:  */
                    782: void
1.14      kristaps  783: parseeoln(struct texi *p, size_t *pos)
1.1       kristaps  784: {
1.17      kristaps  785:        size_t            end, sv;
1.7       kristaps  786:        enum texicmd      cmd;
                    787:        struct teximacro *macro;
1.1       kristaps  788:
1.14      kristaps  789:        while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) {
                    790:                while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) {
1.1       kristaps  791:                        p->seenws = 1;
                    792:                        if (p->literal)
1.14      kristaps  793:                                texiputchar(p, BUF(p)[*pos]);
                    794:                        advance(p, pos);
1.1       kristaps  795:                }
1.14      kristaps  796:                switch (BUF(p)[*pos]) {
1.1       kristaps  797:                case ('}'):
                    798:                        if (0 == p->ign)
                    799:                                texiwarn(p, "unexpected \"}\"");
1.14      kristaps  800:                        advance(p, pos);
1.1       kristaps  801:                        continue;
                    802:                case ('{'):
                    803:                        if (0 == p->ign)
                    804:                                texiwarn(p, "unexpected \"{\"");
1.14      kristaps  805:                        advance(p, pos);
1.1       kristaps  806:                        continue;
                    807:                case ('@'):
                    808:                        break;
                    809:                default:
1.14      kristaps  810:                        parseword(p, pos, '\0');
1.1       kristaps  811:                        continue;
                    812:                }
                    813:
1.17      kristaps  814:                sv = p->files[p->filepos - 1].insplice;
1.14      kristaps  815:                cmd = texicmd(p, *pos, &end, &macro);
                    816:                advanceto(p, pos, end);
1.7       kristaps  817:                if (NULL != macro)
1.17      kristaps  818:                        texiexecmacro(p, macro, sv, pos);
1.1       kristaps  819:                if (TEXICMD__MAX == cmd)
                    820:                        continue;
                    821:                if (NULL != texitoks[cmd].fp)
1.14      kristaps  822:                        (*texitoks[cmd].fp)(p, cmd, pos);
1.1       kristaps  823:        }
1.14      kristaps  824:
                    825:        if (*pos < BUFSZ(p) && '\n' == BUF(p)[*pos])
                    826:                advance(p, pos);
1.19    ! kristaps  827: }
        !           828:
        !           829: /*
        !           830:  * Peek to see if there's a command after subsequent whitespace.
        !           831:  * If so, return the macro identifier.
        !           832:  * This DOES NOT work with user-defined macros.
        !           833:  */
        !           834: enum texicmd
        !           835: peekcmd(const struct texi *p, size_t pos)
        !           836: {
        !           837:        size_t          end;
        !           838:
        !           839:        while (pos < BUFSZ(p) && ismspace(BUF(p)[pos]))
        !           840:                pos++;
        !           841:        if (pos == BUFSZ(p) || '@' != BUF(p)[pos])
        !           842:                return(TEXICMD__MAX);
        !           843:        return(texicmd(p, pos, &end, NULL));
1.1       kristaps  844: }
                    845:
                    846: /*
                    847:  * Parse a single word or command.
                    848:  * This will return immediately at the EOF.
                    849:  */
1.14      kristaps  850: static void
                    851: parsesingle(struct texi *p, size_t *pos)
1.1       kristaps  852: {
1.17      kristaps  853:        size_t            end, sv;
1.7       kristaps  854:        enum texicmd      cmd;
                    855:        struct teximacro *macro;
1.1       kristaps  856:
1.14      kristaps  857:        if ((*pos = advancenext(p, pos)) >= BUFSZ(p))
1.1       kristaps  858:                return;
                    859:
1.14      kristaps  860:        switch (BUF(p)[*pos]) {
1.1       kristaps  861:        case ('}'):
                    862:                if (0 == p->ign)
                    863:                        texiwarn(p, "unexpected \"}\"");
1.14      kristaps  864:                advance(p, pos);
1.1       kristaps  865:                return;
                    866:        case ('{'):
                    867:                if (0 == p->ign)
                    868:                        texiwarn(p, "unexpected \"{\"");
1.14      kristaps  869:                advance(p, pos);
1.1       kristaps  870:                return;
                    871:        case ('@'):
                    872:                break;
                    873:        default:
1.14      kristaps  874:                parseword(p, pos, '\0');
1.1       kristaps  875:                return;
                    876:        }
                    877:
1.17      kristaps  878:        sv = p->files[p->filepos - 1].insplice;
1.14      kristaps  879:        cmd = texicmd(p, *pos, &end, &macro);
                    880:        advanceto(p, pos, end);
1.7       kristaps  881:        if (NULL != macro)
1.17      kristaps  882:                texiexecmacro(p, macro, sv, pos);
1.1       kristaps  883:        if (TEXICMD__MAX == cmd)
                    884:                return;
                    885:        if (NULL != texitoks[cmd].fp)
1.14      kristaps  886:                (*texitoks[cmd].fp)(p, cmd, pos);
1.1       kristaps  887: }
                    888:
                    889: /*
                    890:  * This is used in the @deffn type of command.
                    891:  * These have an arbitrary number of line arguments; however, these
                    892:  * arguments may or may not be surrounded by brackets.
                    893:  * In this function, we parse each one as either a bracketed or
                    894:  * non-bracketed argument, returning 0 when we've reached the end of
                    895:  * line or 1 otherwise.
                    896:  */
                    897: int
1.14      kristaps  898: parselinearg(struct texi *p, size_t *pos)
1.1       kristaps  899: {
                    900:
1.14      kristaps  901:        while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) {
1.1       kristaps  902:                p->seenws = 1;
1.14      kristaps  903:                advance(p, pos);
1.1       kristaps  904:        }
                    905:
1.14      kristaps  906:        if (*pos < BUFSZ(p) && '{' == BUF(p)[*pos])
1.18      kristaps  907:                parsebracket(p, pos, 0);
1.14      kristaps  908:        else if (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos])
                    909:                parsesingle(p, pos);
1.1       kristaps  910:        else
                    911:                return(0);
                    912:
                    913:        return(1);
                    914: }
                    915:
                    916: /*
                    917:  * Parse til the end of the buffer.
                    918:  */
1.14      kristaps  919: static void
                    920: parseeof(struct texi *p)
1.1       kristaps  921: {
                    922:        size_t   pos;
                    923:
1.14      kristaps  924:        for (pos = 0; pos < BUFSZ(p); )
                    925:                parsesingle(p, &pos);
1.1       kristaps  926: }
                    927:
1.8       kristaps  928: void
1.14      kristaps  929: texisplice(struct texi *p, const char *buf, size_t sz, size_t *pos)
1.8       kristaps  930: {
1.14      kristaps  931:        char            *cp;
                    932:        struct texifile *f;
1.8       kristaps  933:
1.14      kristaps  934:        assert(p->filepos > 0);
                    935:        f = &p->files[p->filepos - 1];
1.8       kristaps  936:
1.14      kristaps  937:        if (f->mapsz + sz > f->mapmaxsz) {
                    938:                f->mapmaxsz = f->mapsz + sz + 1024;
                    939:                cp = realloc(f->map, f->mapmaxsz);
                    940:                if (NULL == cp)
                    941:                        texiabort(p, NULL);
                    942:                f->map = cp;
                    943:        }
1.8       kristaps  944:
1.15      kristaps  945:        f->insplice += sz;
1.14      kristaps  946:        memmove(f->map + *pos + sz, f->map + *pos, f->mapsz - *pos);
                    947:        memcpy(f->map + *pos, buf, sz);
                    948:        f->mapsz += sz;
1.8       kristaps  949: }
                    950:
                    951: /*
1.1       kristaps  952:  * Parse a block sequence until we have the "@end endtoken" command
                    953:  * invocation.
                    954:  * This will return immediately at EOF.
                    955:  */
                    956: void
1.14      kristaps  957: parseto(struct texi *p, size_t *pos, const char *endtoken)
1.1       kristaps  958: {
1.17      kristaps  959:        size_t            end, sv;
1.7       kristaps  960:        enum texicmd      cmd;
                    961:        size_t            endtoksz;
                    962:        struct teximacro *macro;
1.1       kristaps  963:
                    964:        endtoksz = strlen(endtoken);
                    965:        assert(endtoksz > 0);
                    966:
1.14      kristaps  967:        while ((*pos = advancenext(p, pos)) < BUFSZ(p)) {
                    968:                switch (BUF(p)[*pos]) {
1.1       kristaps  969:                case ('}'):
                    970:                        if (0 == p->ign)
                    971:                                texiwarn(p, "unexpected \"}\"");
1.14      kristaps  972:                        advance(p, pos);
1.1       kristaps  973:                        continue;
                    974:                case ('{'):
                    975:                        if (0 == p->ign)
                    976:                                texiwarn(p, "unexpected \"{\"");
1.14      kristaps  977:                        advance(p, pos);
1.1       kristaps  978:                        continue;
                    979:                case ('@'):
                    980:                        break;
                    981:                default:
1.14      kristaps  982:                        parseword(p, pos, '\0');
1.1       kristaps  983:                        continue;
                    984:                }
                    985:
1.17      kristaps  986:                sv = p->files[p->filepos - 1].insplice;
1.14      kristaps  987:                cmd = texicmd(p, *pos, &end, &macro);
                    988:                advanceto(p, pos, end);
1.1       kristaps  989:                if (TEXICMD_END == cmd) {
1.14      kristaps  990:                        while (*pos < BUFSZ(p) && isws(BUF(p)[*pos]))
                    991:                                advance(p, pos);
1.1       kristaps  992:                        /*
                    993:                         * FIXME: check the full word, not just its
                    994:                         * initial substring!
                    995:                         */
1.14      kristaps  996:                        if (BUFSZ(p) - *pos >= endtoksz && 0 == strncmp
                    997:                                 (&BUF(p)[*pos], endtoken, endtoksz)) {
                    998:                                advanceeoln(p, pos, 0);
1.1       kristaps  999:                                break;
                   1000:                        }
                   1001:                        if (0 == p->ign)
                   1002:                                texiwarn(p, "unexpected \"end\"");
1.14      kristaps 1003:                        advanceeoln(p, pos, 0);
1.1       kristaps 1004:                        continue;
1.7       kristaps 1005:                }
                   1006:                if (NULL != macro)
1.17      kristaps 1007:                        texiexecmacro(p, macro, sv, pos);
1.7       kristaps 1008:                if (TEXICMD__MAX == cmd)
                   1009:                        continue;
                   1010:                if (NULL != texitoks[cmd].fp)
1.14      kristaps 1011:                        (*texitoks[cmd].fp)(p, cmd, pos);
1.1       kristaps 1012:        }
                   1013: }
                   1014:
                   1015: /*
1.12      kristaps 1016:  * Like parsefile() but used for reading from stdandard input.
                   1017:  * This can only be called for the first file!
                   1018:  */
                   1019: void
                   1020: parsestdin(struct texi *p)
                   1021: {
                   1022:        struct texifile *f;
                   1023:        ssize_t          ssz;
                   1024:
                   1025:        assert(0 == p->filepos);
                   1026:        f = &p->files[p->filepos];
                   1027:        memset(f, 0, sizeof(struct texifile));
                   1028:
                   1029:        f->type = TEXISRC_STDIN;
                   1030:        f->name = "<stdin>";
                   1031:
1.14      kristaps 1032:        for (f->mapsz = 0; ; f->mapsz += (size_t)ssz) {
                   1033:                if (f->mapsz == f->mapmaxsz) {
                   1034:                        if (f->mapmaxsz == (1U << 31))
1.12      kristaps 1035:                                texierr(p, "stdin buffer too long");
1.14      kristaps 1036:                        f->mapmaxsz = f->mapmaxsz > 65536 / 2 ?
                   1037:                                2 * f->mapmaxsz : 65536;
                   1038:                        f->map = realloc(f->map, f->mapmaxsz);
1.12      kristaps 1039:                        if (NULL == f->map)
                   1040:                                texiabort(p, NULL);
                   1041:                }
1.14      kristaps 1042:                ssz = read(STDIN_FILENO, f->map +
                   1043:                        (int)f->mapsz, f->mapmaxsz - f->mapsz);
1.12      kristaps 1044:                if (0 == ssz)
                   1045:                        break;
                   1046:                else if (-1 == ssz)
                   1047:                        texiabort(p, NULL);
                   1048:        }
                   1049:
                   1050:        p->filepos++;
1.14      kristaps 1051:        parseeof(p);
1.12      kristaps 1052:        texifilepop(p);
                   1053: }
                   1054:
                   1055: /*
1.1       kristaps 1056:  * Memory-map the file "fname" and begin parsing it unless "parse" is
                   1057:  * zero, in which case we just dump the file to stdout (making sure it
                   1058:  * doesn't trip up mdoc(7) along the way).
                   1059:  * This can be called in a nested context.
                   1060:  */
                   1061: void
                   1062: parsefile(struct texi *p, const char *fname, int parse)
                   1063: {
                   1064:        struct texifile *f;
                   1065:        int              fd;
                   1066:        struct stat      st;
                   1067:        size_t           i;
1.14      kristaps 1068:        char            *map;
1.1       kristaps 1069:
1.5       kristaps 1070:        if (64 == p->filepos)
1.6       kristaps 1071:                texierr(p, "too many open files");
1.1       kristaps 1072:        f = &p->files[p->filepos];
                   1073:        memset(f, 0, sizeof(struct texifile));
                   1074:
1.12      kristaps 1075:        f->type = TEXISRC_FILE;
1.1       kristaps 1076:        f->name = fname;
                   1077:        if (-1 == (fd = open(fname, O_RDONLY, 0))) {
                   1078:                texiabort(p, fname);
                   1079:        } else if (-1 == fstat(fd, &st)) {
                   1080:                close(fd);
                   1081:                texiabort(p, fname);
                   1082:        }
                   1083:
1.14      kristaps 1084:        f->mapsz = f->mapmaxsz = st.st_size;
                   1085:        map = mmap(NULL, f->mapsz,
1.1       kristaps 1086:                PROT_READ, MAP_SHARED, fd, 0);
                   1087:        close(fd);
                   1088:
1.14      kristaps 1089:        if (MAP_FAILED == map)
1.1       kristaps 1090:                texiabort(p, fname);
                   1091:
                   1092:        if ( ! parse) {
1.13      kristaps 1093:                for (i = 0; i < f->mapsz; i++)
1.14      kristaps 1094:                        texiputchar(p, map[i]);
1.13      kristaps 1095:                if (p->outcol)
                   1096:                        texiputchar(p, '\n');
1.14      kristaps 1097:                munmap(map, f->mapsz);
                   1098:                return;
                   1099:        }
                   1100:
                   1101:        p->filepos++;
                   1102:        f->map = malloc(f->mapsz);
                   1103:        memcpy(f->map, map, f->mapsz);
                   1104:        munmap(map, f->mapsz);
                   1105:        parseeof(p);
1.1       kristaps 1106:        texifilepop(p);
                   1107: }
                   1108:
1.2       kristaps 1109: /*
                   1110:  * Look up the value to a stored pair's value starting in "buf" from
                   1111:  * start to end.
                   1112:  * Return the pointer to the value memory, which can be NULL if the
                   1113:  * pointer key does not exist.
                   1114:  * The pointer can point to NULL if the value has been unset.
                   1115:  */
                   1116: static char **
1.14      kristaps 1117: valuequery(const struct texi *p, size_t start, size_t end)
1.2       kristaps 1118: {
                   1119:        size_t   i, sz, len;
                   1120:
                   1121:        assert(end >= start);
                   1122:        /* Ignore zero-length. */
                   1123:        if (0 == (len = (end - start)))
                   1124:                return(NULL);
                   1125:        for (i = 0; i < p->valsz; i++) {
                   1126:                sz = strlen(p->vals[i].key);
                   1127:                if (sz != len)
                   1128:                        continue;
1.14      kristaps 1129:                if (0 == strncmp(p->vals[i].key, &BUF(p)[start], len))
1.2       kristaps 1130:                        return(&p->vals[i].value);
                   1131:        }
                   1132:        return(NULL);
                   1133: }
                   1134:
                   1135: /*
                   1136:  * Parse a key until the end of line, e.g., @clear foo\n, and return the
                   1137:  * pointer to its value via valuequery().
                   1138:  */
                   1139: static char **
1.14      kristaps 1140: valuelquery(struct texi *p, size_t *pos)
1.2       kristaps 1141: {
                   1142:        size_t    start, end;
                   1143:        char    **ret;
                   1144:
1.14      kristaps 1145:        while (*pos < BUFSZ(p) && isws(BUF(p)[*pos]))
                   1146:                advance(p, pos);
                   1147:        if (*pos == BUFSZ(p))
1.2       kristaps 1148:                return(NULL);
1.14      kristaps 1149:        for (start = end = *pos; end < BUFSZ(p); end++)
                   1150:                if ('\n' == BUF(p)[end])
1.2       kristaps 1151:                        break;
1.14      kristaps 1152:        advanceto(p, pos, end);
                   1153:        if (*pos < BUFSZ(p)) {
                   1154:                assert('\n' == BUF(p)[*pos]);
                   1155:                advance(p, pos);
1.2       kristaps 1156:        }
1.14      kristaps 1157:        if (NULL == (ret = valuequery(p, start, end)))
1.2       kristaps 1158:                return(NULL);
                   1159:        return(ret);
                   1160: }
                   1161:
                   1162: void
1.14      kristaps 1163: valuelclear(struct texi *p, size_t *pos)
1.2       kristaps 1164: {
                   1165:        char    **ret;
                   1166:
1.14      kristaps 1167:        if (NULL == (ret = valuelquery(p, pos)))
1.2       kristaps 1168:                return;
                   1169:        free(*ret);
                   1170:        *ret = NULL;
                   1171: }
                   1172:
                   1173: const char *
1.14      kristaps 1174: valuellookup(struct texi *p, size_t *pos)
1.2       kristaps 1175: {
                   1176:        char    **ret;
                   1177:
1.14      kristaps 1178:        if (NULL == (ret = valuelquery(p, pos)))
1.2       kristaps 1179:                return(NULL);
                   1180:        return(*ret);
                   1181: }
                   1182:
                   1183: /*
                   1184:  * Parse a key from a bracketed string, e.g., @value{foo}, and return
                   1185:  * the pointer to its value.
                   1186:  * If the returned pointer is NULL, either there was no string within
                   1187:  * the brackets (or no brackets), or the value was not found, or the
                   1188:  * value had previously been unset.
                   1189:  */
                   1190: const char *
1.14      kristaps 1191: valueblookup(struct texi *p, size_t *pos)
1.2       kristaps 1192: {
                   1193:        size_t    start, end;
                   1194:        char    **ret;
                   1195:
1.14      kristaps 1196:        while (*pos < BUFSZ(p) && isws(BUF(p)[*pos]))
                   1197:                advance(p, pos);
                   1198:        if (*pos == BUFSZ(p) || '{' != BUF(p)[*pos])
1.2       kristaps 1199:                return(NULL);
1.14      kristaps 1200:        advance(p, pos);
                   1201:        for (start = end = *pos; end < BUFSZ(p); end++)
                   1202:                if ('}' == BUF(p)[end])
1.2       kristaps 1203:                        break;
1.14      kristaps 1204:        advanceto(p, pos, end);
                   1205:        if (*pos < BUFSZ(p)) {
                   1206:                assert('}' == BUF(p)[*pos]);
                   1207:                advance(p, pos);
1.2       kristaps 1208:        }
1.14      kristaps 1209:        if (NULL == (ret = valuequery(p, start, end)))
1.2       kristaps 1210:                return(NULL);
                   1211:        return(*ret);
                   1212: }
                   1213:
                   1214: void
                   1215: valueadd(struct texi *p, char *key, char *val)
                   1216: {
                   1217:        size_t   i;
                   1218:
                   1219:        assert(NULL != key);
                   1220:        assert(NULL != val);
                   1221:
                   1222:        for (i = 0; i < p->valsz; i++)
                   1223:                if (0 == strcmp(p->vals[i].key, key))
                   1224:                        break;
                   1225:
                   1226:        if (i < p->valsz) {
                   1227:                free(key);
                   1228:                free(p->vals[i].value);
                   1229:                p->vals[i].value = val;
                   1230:        } else {
1.4       kristaps 1231:                /* FIXME: reallocarray() */
1.2       kristaps 1232:                p->vals = realloc(p->vals,
                   1233:                        (p->valsz + 1) *
                   1234:                         sizeof(struct texivalue));
1.4       kristaps 1235:                if (NULL == p->vals)
                   1236:                        texiabort(p, NULL);
1.2       kristaps 1237:                p->vals[p->valsz].key = key;
                   1238:                p->vals[p->valsz].value = val;
                   1239:                p->valsz++;
                   1240:        }
1.7       kristaps 1241: }
                   1242:
                   1243: /*
                   1244:  * Take the arguments to a macro, e.g., @foo{bar, baz, xyzzy} (or the
                   1245:  * declaration form, @macro foo {arg1, ...}) and textually convert it to
                   1246:  * an array of arguments of size "argsz".
                   1247:  * These need to be freed individually and as a whole.
                   1248:  * NOTE: this will puke on @, or @} macros, which can trick it into
                   1249:  * stopping argument parsing earlier.
                   1250:  * Ergo, textual: this doesn't interpret the arguments in any way.
                   1251:  */
                   1252: char **
1.14      kristaps 1253: argparse(struct texi *p, size_t *pos, size_t *argsz, size_t hint)
1.7       kristaps 1254: {
                   1255:        char    **args;
                   1256:        size_t    start, end, stack;
                   1257:
1.14      kristaps 1258:        while (*pos < BUFSZ(p) && isws(BUF(p)[*pos]))
                   1259:                advance(p, pos);
1.7       kristaps 1260:
                   1261:        args = NULL;
                   1262:        *argsz = 0;
                   1263:
1.17      kristaps 1264:        if (*pos == BUFSZ(p))
                   1265:                return(args);
                   1266:
1.14      kristaps 1267:        if ('{' != BUF(p)[*pos] && hint) {
1.10      kristaps 1268:                /*
                   1269:                 * Special case: if we encounter an unbracketed argument
                   1270:                 * and we're being invoked with non-zero arguments
                   1271:                 * (versus being set, i.e., hint>0), then parse until
                   1272:                 * the end of line.
                   1273:                 */
                   1274:                *argsz = 1;
                   1275:                args = calloc(1, sizeof(char *));
                   1276:                if (NULL == args)
                   1277:                        texiabort(p, NULL);
                   1278:                start = *pos;
1.14      kristaps 1279:                while (*pos < BUFSZ(p)) {
                   1280:                        if ('\n' == BUF(p)[*pos])
1.10      kristaps 1281:                                break;
1.14      kristaps 1282:                        advance(p, pos);
1.10      kristaps 1283:                }
                   1284:                args[0] = malloc(*pos - start + 1);
1.14      kristaps 1285:                memcpy(args[0], &BUF(p)[start], *pos - start);
1.10      kristaps 1286:                args[0][*pos - start] = '\0';
1.14      kristaps 1287:                if (*pos < BUFSZ(p) && '\n' == BUF(p)[*pos])
                   1288:                        advance(p, pos);
1.10      kristaps 1289:                return(args);
1.14      kristaps 1290:        } else if ('{' != BUF(p)[*pos])
1.7       kristaps 1291:                return(args);
1.17      kristaps 1292:
                   1293:        assert('{' == BUF(p)[*pos]);
1.7       kristaps 1294:
                   1295:        /* Parse til the closing '}', putting into the array. */
1.14      kristaps 1296:        advance(p, pos);
                   1297:        while (*pos < BUFSZ(p)) {
                   1298:                while (*pos < BUFSZ(p) && isws(BUF(p)[*pos]))
                   1299:                        advance(p, pos);
1.7       kristaps 1300:                start = *pos;
                   1301:                stack = 0;
1.14      kristaps 1302:                while (*pos < BUFSZ(p)) {
1.7       kristaps 1303:                        /*
                   1304:                         * According to the manual, commas within
                   1305:                         * embedded commands are escaped.
                   1306:                         * We keep track of embedded-ness in the "stack"
                   1307:                         * state anyway, so this is free.
                   1308:                         */
1.14      kristaps 1309:                        if (',' == BUF(p)[*pos] && 0 == stack && 1 != hint)
1.7       kristaps 1310:                                break;
1.14      kristaps 1311:                        else if (0 == stack && '}' == BUF(p)[*pos])
1.7       kristaps 1312:                                break;
1.14      kristaps 1313:                        else if (0 != stack && '}' == BUF(p)[*pos])
1.7       kristaps 1314:                                stack--;
1.14      kristaps 1315:                        else if ('{' == BUF(p)[*pos])
1.7       kristaps 1316:                                stack++;
1.14      kristaps 1317:                        advance(p, pos);
1.7       kristaps 1318:                }
                   1319:                if (stack)
                   1320:                        texiwarn(p, "unterminated macro "
                   1321:                                "in macro arguments");
1.14      kristaps 1322:                if ((end = *pos) == BUFSZ(p))
1.7       kristaps 1323:                        break;
                   1324:                /* Test for zero-length '{  }'. */
1.14      kristaps 1325:                if (start == end && '}' == BUF(p)[*pos] && 0 == *argsz)
1.7       kristaps 1326:                        break;
                   1327:                /* FIXME: use reallocarray. */
                   1328:                args = realloc
                   1329:                        (args, sizeof(char *) *
                   1330:                         (*argsz + 1));
                   1331:                if (NULL == args)
                   1332:                        texiabort(p, NULL);
                   1333:                args[*argsz] = malloc(end - start + 1);
                   1334:                if (NULL == args[*argsz])
                   1335:                        texiabort(p, NULL);
                   1336:                memcpy(args[*argsz],
1.14      kristaps 1337:                        &BUF(p)[start], end - start);
1.7       kristaps 1338:                args[*argsz][end - start] = '\0';
                   1339:                (*argsz)++;
1.14      kristaps 1340:                if ('}' == BUF(p)[*pos])
1.7       kristaps 1341:                        break;
1.14      kristaps 1342:                advance(p, pos);
1.7       kristaps 1343:        }
                   1344:
1.14      kristaps 1345:        if (*pos == BUFSZ(p))
1.7       kristaps 1346:                texierr(p, "unterminated arguments");
1.14      kristaps 1347:        assert('}' == BUF(p)[*pos]);
                   1348:        advance(p, pos);
1.7       kristaps 1349:        return(args);
1.2       kristaps 1350: }

CVSweb