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

Annotation of mandoc/read.c, Revision 1.35

1.35    ! schwarze    1: /*     $Id: read.c,v 1.34 2012/11/19 22:30:58 schwarze Exp $ */
1.1       kristaps    2: /*
                      3:  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.35    ! schwarze    4:  * Copyright (c) 2010, 2011, 2012, 2013 Ingo Schwarze <schwarze@openbsd.org>
1.1       kristaps    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
                      7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
                      9:  *
                     10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
                     11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
                     12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
                     13:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     14:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     15:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     16:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
                     17:  */
1.11      kristaps   18: #ifdef HAVE_CONFIG_H
                     19: #include "config.h"
                     20: #endif
                     21:
1.15      kristaps   22: #ifdef HAVE_MMAP
                     23: # include <sys/stat.h>
                     24: # include <sys/mman.h>
                     25: #endif
1.1       kristaps   26:
                     27: #include <assert.h>
                     28: #include <ctype.h>
                     29: #include <fcntl.h>
1.3       kristaps   30: #include <stdarg.h>
1.28      joerg      31: #include <stdint.h>
1.1       kristaps   32: #include <stdio.h>
                     33: #include <stdlib.h>
                     34: #include <string.h>
                     35: #include <unistd.h>
                     36:
                     37: #include "mandoc.h"
1.3       kristaps   38: #include "libmandoc.h"
1.1       kristaps   39: #include "mdoc.h"
                     40: #include "man.h"
1.28      joerg      41: #include "main.h"
1.1       kristaps   42:
                     43: #ifndef MAP_FILE
                     44: #define        MAP_FILE        0
                     45: #endif
                     46:
                     47: #define        REPARSE_LIMIT   1000
                     48:
                     49: struct buf {
                     50:        char             *buf; /* binary input buffer */
                     51:        size_t            sz; /* size of binary buffer */
                     52: };
                     53:
                     54: struct mparse {
                     55:        enum mandoclevel  file_status; /* status of current parse */
1.3       kristaps   56:        enum mandoclevel  wlevel; /* ignore messages below this */
1.1       kristaps   57:        int               line; /* line number in the file */
                     58:        enum mparset      inttype; /* which parser to use */
                     59:        struct man       *pman; /* persistent man parser */
                     60:        struct mdoc      *pmdoc; /* persistent mdoc parser */
                     61:        struct man       *man; /* man parser */
                     62:        struct mdoc      *mdoc; /* mdoc parser */
                     63:        struct roff      *roff; /* roff parser (!NULL) */
                     64:        int               reparse_count; /* finite interp. stack */
                     65:        mandocmsg         mmsg; /* warning/error message handler */
                     66:        void             *arg; /* argument to mmsg */
1.3       kristaps   67:        const char       *file;
1.24      kristaps   68:        struct buf       *secondary;
1.29      schwarze   69:        char             *defos; /* default operating system */
1.1       kristaps   70: };
                     71:
                     72: static void      resize_buf(struct buf *, size_t);
                     73: static void      mparse_buf_r(struct mparse *, struct buf, int);
                     74: static void      mparse_readfd_r(struct mparse *, int, const char *, int);
                     75: static void      pset(const char *, int, struct mparse *);
                     76: static int       read_whole_file(const char *, int, struct buf *, int *);
                     77: static void      mparse_end(struct mparse *);
                     78:
1.3       kristaps   79: static const enum mandocerr    mandoclimits[MANDOCLEVEL_MAX] = {
                     80:        MANDOCERR_OK,
                     81:        MANDOCERR_WARNING,
                     82:        MANDOCERR_WARNING,
                     83:        MANDOCERR_ERROR,
                     84:        MANDOCERR_FATAL,
                     85:        MANDOCERR_MAX,
                     86:        MANDOCERR_MAX
                     87: };
                     88:
1.7       kristaps   89: static const char * const      mandocerrs[MANDOCERR_MAX] = {
                     90:        "ok",
                     91:
                     92:        "generic warning",
                     93:
                     94:        /* related to the prologue */
                     95:        "no title in document",
                     96:        "document title should be all caps",
                     97:        "unknown manual section",
1.32      schwarze   98:        "unknown manual volume or arch",
1.7       kristaps   99:        "date missing, using today's date",
                    100:        "cannot parse date, using it verbatim",
                    101:        "prologue macros out of order",
                    102:        "duplicate prologue macro",
                    103:        "macro not allowed in prologue",
                    104:        "macro not allowed in body",
                    105:
                    106:        /* related to document structure */
                    107:        ".so is fragile, better use ln(1)",
                    108:        "NAME section must come first",
                    109:        "bad NAME section contents",
                    110:        "sections out of conventional order",
                    111:        "duplicate section name",
                    112:        "section not in conventional manual section",
                    113:
                    114:        /* related to macros and nesting */
                    115:        "skipping obsolete macro",
                    116:        "skipping paragraph macro",
1.31      schwarze  117:        "moving paragraph macro out of list",
1.7       kristaps  118:        "skipping no-space macro",
                    119:        "blocks badly nested",
                    120:        "child violates parent syntax",
                    121:        "nested displays are not portable",
                    122:        "already in literal mode",
                    123:        "line scope broken",
                    124:
                    125:        /* related to missing macro arguments */
                    126:        "skipping empty macro",
                    127:        "argument count wrong",
                    128:        "missing display type",
                    129:        "list type must come first",
                    130:        "tag lists require a width argument",
                    131:        "missing font type",
                    132:        "skipping end of block that is not open",
                    133:
                    134:        /* related to bad macro arguments */
                    135:        "skipping argument",
                    136:        "duplicate argument",
                    137:        "duplicate display type",
                    138:        "duplicate list type",
                    139:        "unknown AT&T UNIX version",
                    140:        "bad Boolean value",
                    141:        "unknown font",
                    142:        "unknown standard specifier",
                    143:        "bad width argument",
                    144:
                    145:        /* related to plain text */
                    146:        "blank line in non-literal context",
                    147:        "tab in non-literal context",
                    148:        "end of line whitespace",
                    149:        "bad comment style",
1.12      kristaps  150:        "bad escape sequence",
1.7       kristaps  151:        "unterminated quoted string",
1.16      kristaps  152:
                    153:        /* related to equations */
                    154:        "unexpected literal in equation",
1.7       kristaps  155:
                    156:        "generic error",
1.17      kristaps  157:
                    158:        /* related to equations */
1.20      kristaps  159:        "unexpected equation scope closure",
                    160:        "equation scope open on exit",
1.21      kristaps  161:        "overlapping equation scopes",
                    162:        "unexpected end of equation",
                    163:        "equation syntax error",
1.7       kristaps  164:
                    165:        /* related to tables */
                    166:        "bad table syntax",
                    167:        "bad table option",
                    168:        "bad table layout",
                    169:        "no table layout cells specified",
                    170:        "no table data cells specified",
                    171:        "ignore data in cell",
                    172:        "data block still open",
                    173:        "ignoring extra data cells",
                    174:
                    175:        "input stack limit exceeded, infinite loop?",
                    176:        "skipping bad character",
                    177:        "escaped character not allowed in a name",
1.30      schwarze  178:        "manual name not yet set",
1.7       kristaps  179:        "skipping text before the first section header",
                    180:        "skipping unknown macro",
                    181:        "NOT IMPLEMENTED, please use groff: skipping request",
                    182:        "argument count wrong",
1.34      schwarze  183:        "skipping column outside column list",
1.7       kristaps  184:        "skipping end of block that is not open",
                    185:        "missing end of block",
                    186:        "scope open on exit",
                    187:        "uname(3) system call failed",
                    188:        "macro requires line argument(s)",
                    189:        "macro requires body argument(s)",
                    190:        "macro requires argument(s)",
                    191:        "missing list type",
                    192:        "line argument(s) will be lost",
                    193:        "body argument(s) will be lost",
                    194:
                    195:        "generic fatal error",
                    196:
                    197:        "not a manual",
                    198:        "column syntax is inconsistent",
                    199:        "NOT IMPLEMENTED: .Bd -file",
                    200:        "argument count wrong, violates syntax",
                    201:        "child violates parent syntax",
                    202:        "argument count wrong, violates syntax",
                    203:        "NOT IMPLEMENTED: .so with absolute path or \"..\"",
                    204:        "no document body",
                    205:        "no document prologue",
                    206:        "static buffer exhausted",
                    207: };
                    208:
                    209: static const char * const      mandoclevels[MANDOCLEVEL_MAX] = {
                    210:        "SUCCESS",
                    211:        "RESERVED",
                    212:        "WARNING",
                    213:        "ERROR",
                    214:        "FATAL",
                    215:        "BADARG",
                    216:        "SYSERR"
                    217: };
                    218:
1.1       kristaps  219: static void
                    220: resize_buf(struct buf *buf, size_t initial)
                    221: {
                    222:
                    223:        buf->sz = buf->sz > initial/2 ? 2 * buf->sz : initial;
                    224:        buf->buf = mandoc_realloc(buf->buf, buf->sz);
                    225: }
                    226:
                    227: static void
                    228: pset(const char *buf, int pos, struct mparse *curp)
                    229: {
                    230:        int              i;
                    231:
                    232:        /*
                    233:         * Try to intuit which kind of manual parser should be used.  If
                    234:         * passed in by command-line (-man, -mdoc), then use that
                    235:         * explicitly.  If passed as -mandoc, then try to guess from the
                    236:         * line: either skip dot-lines, use -mdoc when finding `.Dt', or
                    237:         * default to -man, which is more lenient.
                    238:         *
                    239:         * Separate out pmdoc/pman from mdoc/man: the first persists
                    240:         * through all parsers, while the latter is used per-parse.
                    241:         */
                    242:
                    243:        if ('.' == buf[0] || '\'' == buf[0]) {
                    244:                for (i = 1; buf[i]; i++)
                    245:                        if (' ' != buf[i] && '\t' != buf[i])
                    246:                                break;
                    247:                if ('\0' == buf[i])
                    248:                        return;
                    249:        }
                    250:
                    251:        switch (curp->inttype) {
                    252:        case (MPARSE_MDOC):
                    253:                if (NULL == curp->pmdoc)
1.29      schwarze  254:                        curp->pmdoc = mdoc_alloc(curp->roff, curp,
                    255:                                        curp->defos);
1.1       kristaps  256:                assert(curp->pmdoc);
                    257:                curp->mdoc = curp->pmdoc;
                    258:                return;
                    259:        case (MPARSE_MAN):
                    260:                if (NULL == curp->pman)
1.18      kristaps  261:                        curp->pman = man_alloc(curp->roff, curp);
1.1       kristaps  262:                assert(curp->pman);
                    263:                curp->man = curp->pman;
                    264:                return;
                    265:        default:
                    266:                break;
                    267:        }
                    268:
                    269:        if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3))  {
                    270:                if (NULL == curp->pmdoc)
1.29      schwarze  271:                        curp->pmdoc = mdoc_alloc(curp->roff, curp,
                    272:                                        curp->defos);
1.1       kristaps  273:                assert(curp->pmdoc);
                    274:                curp->mdoc = curp->pmdoc;
                    275:                return;
                    276:        }
                    277:
                    278:        if (NULL == curp->pman)
1.18      kristaps  279:                curp->pman = man_alloc(curp->roff, curp);
1.1       kristaps  280:        assert(curp->pman);
                    281:        curp->man = curp->pman;
                    282: }
                    283:
                    284: /*
                    285:  * Main parse routine for an opened file.  This is called for each
                    286:  * opened file and simply loops around the full input file, possibly
                    287:  * nesting (i.e., with `so').
                    288:  */
                    289: static void
                    290: mparse_buf_r(struct mparse *curp, struct buf blk, int start)
                    291: {
                    292:        const struct tbl_span   *span;
                    293:        struct buf       ln;
                    294:        enum rofferr     rr;
                    295:        int              i, of, rc;
                    296:        int              pos; /* byte number in the ln buffer */
                    297:        int              lnn; /* line number in the real file */
                    298:        unsigned char    c;
                    299:
                    300:        memset(&ln, 0, sizeof(struct buf));
                    301:
                    302:        lnn = curp->line;
                    303:        pos = 0;
                    304:
                    305:        for (i = 0; i < (int)blk.sz; ) {
                    306:                if (0 == pos && '\0' == blk.buf[i])
                    307:                        break;
                    308:
                    309:                if (start) {
                    310:                        curp->line = lnn;
                    311:                        curp->reparse_count = 0;
                    312:                }
                    313:
                    314:                while (i < (int)blk.sz && (start || '\0' != blk.buf[i])) {
                    315:
                    316:                        /*
                    317:                         * When finding an unescaped newline character,
                    318:                         * leave the character loop to process the line.
                    319:                         * Skip a preceding carriage return, if any.
                    320:                         */
                    321:
                    322:                        if ('\r' == blk.buf[i] && i + 1 < (int)blk.sz &&
                    323:                            '\n' == blk.buf[i + 1])
                    324:                                ++i;
                    325:                        if ('\n' == blk.buf[i]) {
                    326:                                ++i;
                    327:                                ++lnn;
                    328:                                break;
                    329:                        }
                    330:
1.35    ! schwarze  331:                        /*
        !           332:                         * Make sure we have space for at least
        !           333:                         * one backslash and one other character
        !           334:                         * and the trailing NUL byte.
        !           335:                         */
        !           336:
        !           337:                        if (pos + 2 >= (int)ln.sz)
        !           338:                                resize_buf(&ln, 256);
        !           339:
1.1       kristaps  340:                        /*
                    341:                         * Warn about bogus characters.  If you're using
                    342:                         * non-ASCII encoding, you're screwing your
                    343:                         * readers.  Since I'd rather this not happen,
1.27      joerg     344:                         * I'll be helpful and replace these characters
                    345:                         * with "?", so we don't display gibberish.
                    346:                         * Note to manual writers: use special characters.
1.1       kristaps  347:                         */
                    348:
                    349:                        c = (unsigned char) blk.buf[i];
                    350:
                    351:                        if ( ! (isascii(c) &&
                    352:                                        (isgraph(c) || isblank(c)))) {
1.3       kristaps  353:                                mandoc_msg(MANDOCERR_BADCHAR, curp,
1.27      joerg     354:                                                curp->line, pos, NULL);
1.1       kristaps  355:                                i++;
1.27      joerg     356:                                ln.buf[pos++] = '?';
1.1       kristaps  357:                                continue;
                    358:                        }
                    359:
                    360:                        /* Trailing backslash = a plain char. */
                    361:
                    362:                        if ('\\' != blk.buf[i] || i + 1 == (int)blk.sz) {
                    363:                                ln.buf[pos++] = blk.buf[i++];
                    364:                                continue;
                    365:                        }
                    366:
                    367:                        /*
                    368:                         * Found escape and at least one other character.
                    369:                         * When it's a newline character, skip it.
                    370:                         * When there is a carriage return in between,
                    371:                         * skip that one as well.
                    372:                         */
                    373:
                    374:                        if ('\r' == blk.buf[i + 1] && i + 2 < (int)blk.sz &&
                    375:                            '\n' == blk.buf[i + 2])
                    376:                                ++i;
                    377:                        if ('\n' == blk.buf[i + 1]) {
                    378:                                i += 2;
                    379:                                ++lnn;
                    380:                                continue;
                    381:                        }
                    382:
1.13      kristaps  383:                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {
1.1       kristaps  384:                                i += 2;
                    385:                                /* Comment, skip to end of line */
                    386:                                for (; i < (int)blk.sz; ++i) {
                    387:                                        if ('\n' == blk.buf[i]) {
                    388:                                                ++i;
                    389:                                                ++lnn;
                    390:                                                break;
                    391:                                        }
                    392:                                }
                    393:
                    394:                                /* Backout trailing whitespaces */
                    395:                                for (; pos > 0; --pos) {
                    396:                                        if (ln.buf[pos - 1] != ' ')
                    397:                                                break;
                    398:                                        if (pos > 2 && ln.buf[pos - 2] == '\\')
                    399:                                                break;
                    400:                                }
                    401:                                break;
                    402:                        }
                    403:
1.35    ! schwarze  404:                        /* Catch escaped bogus characters. */
        !           405:
        !           406:                        c = (unsigned char) blk.buf[i+1];
        !           407:
        !           408:                        if ( ! (isascii(c) &&
        !           409:                                        (isgraph(c) || isblank(c)))) {
        !           410:                                mandoc_msg(MANDOCERR_BADCHAR, curp,
        !           411:                                                curp->line, pos, NULL);
        !           412:                                i += 2;
        !           413:                                ln.buf[pos++] = '?';
        !           414:                                continue;
        !           415:                        }
        !           416:
1.1       kristaps  417:                        /* Some other escape sequence, copy & cont. */
                    418:
                    419:                        ln.buf[pos++] = blk.buf[i++];
                    420:                        ln.buf[pos++] = blk.buf[i++];
                    421:                }
                    422:
                    423:                if (pos >= (int)ln.sz)
                    424:                        resize_buf(&ln, 256);
                    425:
                    426:                ln.buf[pos] = '\0';
                    427:
                    428:                /*
                    429:                 * A significant amount of complexity is contained by
                    430:                 * the roff preprocessor.  It's line-oriented but can be
                    431:                 * expressed on one line, so we need at times to
                    432:                 * readjust our starting point and re-run it.  The roff
                    433:                 * preprocessor can also readjust the buffers with new
                    434:                 * data, so we pass them in wholesale.
                    435:                 */
                    436:
                    437:                of = 0;
                    438:
1.24      kristaps  439:                /*
                    440:                 * Maintain a lookaside buffer of all parsed lines.  We
                    441:                 * only do this if mparse_keep() has been invoked (the
                    442:                 * buffer may be accessed with mparse_getkeep()).
                    443:                 */
                    444:
                    445:                if (curp->secondary) {
                    446:                        curp->secondary->buf =
                    447:                                mandoc_realloc
                    448:                                (curp->secondary->buf,
                    449:                                 curp->secondary->sz + pos + 2);
                    450:                        memcpy(curp->secondary->buf +
                    451:                                        curp->secondary->sz,
                    452:                                        ln.buf, pos);
                    453:                        curp->secondary->sz += pos;
                    454:                        curp->secondary->buf
                    455:                                [curp->secondary->sz] = '\n';
                    456:                        curp->secondary->sz++;
                    457:                        curp->secondary->buf
                    458:                                [curp->secondary->sz] = '\0';
                    459:                }
1.1       kristaps  460: rerun:
                    461:                rr = roff_parseln
                    462:                        (curp->roff, curp->line,
                    463:                         &ln.buf, &ln.sz, of, &of);
                    464:
                    465:                switch (rr) {
                    466:                case (ROFF_REPARSE):
                    467:                        if (REPARSE_LIMIT >= ++curp->reparse_count)
                    468:                                mparse_buf_r(curp, ln, 0);
                    469:                        else
1.3       kristaps  470:                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
1.1       kristaps  471:                                        curp->line, pos, NULL);
                    472:                        pos = 0;
                    473:                        continue;
                    474:                case (ROFF_APPEND):
                    475:                        pos = (int)strlen(ln.buf);
                    476:                        continue;
                    477:                case (ROFF_RERUN):
                    478:                        goto rerun;
                    479:                case (ROFF_IGN):
                    480:                        pos = 0;
                    481:                        continue;
                    482:                case (ROFF_ERR):
                    483:                        assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    484:                        break;
                    485:                case (ROFF_SO):
1.24      kristaps  486:                        /*
                    487:                         * We remove `so' clauses from our lookaside
                    488:                         * buffer because we're going to descend into
                    489:                         * the file recursively.
                    490:                         */
1.25      kristaps  491:                        if (curp->secondary)
                    492:                                curp->secondary->sz -= pos + 1;
1.1       kristaps  493:                        mparse_readfd_r(curp, -1, ln.buf + of, 1);
                    494:                        if (MANDOCLEVEL_FATAL <= curp->file_status)
                    495:                                break;
                    496:                        pos = 0;
                    497:                        continue;
                    498:                default:
                    499:                        break;
                    500:                }
                    501:
                    502:                /*
                    503:                 * If we encounter errors in the recursive parse, make
                    504:                 * sure we don't continue parsing.
                    505:                 */
                    506:
                    507:                if (MANDOCLEVEL_FATAL <= curp->file_status)
                    508:                        break;
                    509:
                    510:                /*
                    511:                 * If input parsers have not been allocated, do so now.
1.14      kristaps  512:                 * We keep these instanced between parsers, but set them
1.1       kristaps  513:                 * locally per parse routine since we can use different
                    514:                 * parsers with each one.
                    515:                 */
                    516:
                    517:                if ( ! (curp->man || curp->mdoc))
                    518:                        pset(ln.buf + of, pos - of, curp);
                    519:
                    520:                /*
                    521:                 * Lastly, push down into the parsers themselves.  One
                    522:                 * of these will have already been set in the pset()
                    523:                 * routine.
                    524:                 * If libroff returns ROFF_TBL, then add it to the
                    525:                 * currently open parse.  Since we only get here if
                    526:                 * there does exist data (see tbl_data.c), we're
                    527:                 * guaranteed that something's been allocated.
                    528:                 * Do the same for ROFF_EQN.
                    529:                 */
                    530:
                    531:                rc = -1;
                    532:
                    533:                if (ROFF_TBL == rr)
                    534:                        while (NULL != (span = roff_span(curp->roff))) {
                    535:                                rc = curp->man ?
                    536:                                        man_addspan(curp->man, span) :
                    537:                                        mdoc_addspan(curp->mdoc, span);
                    538:                                if (0 == rc)
                    539:                                        break;
                    540:                        }
                    541:                else if (ROFF_EQN == rr)
                    542:                        rc = curp->mdoc ?
                    543:                                mdoc_addeqn(curp->mdoc,
                    544:                                        roff_eqn(curp->roff)) :
                    545:                                man_addeqn(curp->man,
                    546:                                        roff_eqn(curp->roff));
                    547:                else if (curp->man || curp->mdoc)
                    548:                        rc = curp->man ?
                    549:                                man_parseln(curp->man,
                    550:                                        curp->line, ln.buf, of) :
                    551:                                mdoc_parseln(curp->mdoc,
                    552:                                        curp->line, ln.buf, of);
                    553:
                    554:                if (0 == rc) {
                    555:                        assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    556:                        break;
                    557:                }
                    558:
                    559:                /* Temporary buffers typically are not full. */
                    560:
                    561:                if (0 == start && '\0' == blk.buf[i])
                    562:                        break;
                    563:
                    564:                /* Start the next input line. */
                    565:
                    566:                pos = 0;
                    567:        }
                    568:
                    569:        free(ln.buf);
                    570: }
                    571:
                    572: static int
                    573: read_whole_file(const char *file, int fd, struct buf *fb, int *with_mmap)
                    574: {
                    575:        size_t           off;
                    576:        ssize_t          ssz;
                    577:
1.15      kristaps  578: #ifdef HAVE_MMAP
                    579:        struct stat      st;
1.1       kristaps  580:        if (-1 == fstat(fd, &st)) {
                    581:                perror(file);
                    582:                return(0);
                    583:        }
                    584:
                    585:        /*
                    586:         * If we're a regular file, try just reading in the whole entry
                    587:         * via mmap().  This is faster than reading it into blocks, and
                    588:         * since each file is only a few bytes to begin with, I'm not
                    589:         * concerned that this is going to tank any machines.
                    590:         */
                    591:
                    592:        if (S_ISREG(st.st_mode)) {
                    593:                if (st.st_size >= (1U << 31)) {
                    594:                        fprintf(stderr, "%s: input too large\n", file);
                    595:                        return(0);
                    596:                }
                    597:                *with_mmap = 1;
                    598:                fb->sz = (size_t)st.st_size;
                    599:                fb->buf = mmap(NULL, fb->sz, PROT_READ,
                    600:                                MAP_FILE|MAP_SHARED, fd, 0);
                    601:                if (fb->buf != MAP_FAILED)
                    602:                        return(1);
                    603:        }
1.15      kristaps  604: #endif
1.1       kristaps  605:
                    606:        /*
                    607:         * If this isn't a regular file (like, say, stdin), then we must
                    608:         * go the old way and just read things in bit by bit.
                    609:         */
                    610:
                    611:        *with_mmap = 0;
                    612:        off = 0;
                    613:        fb->sz = 0;
                    614:        fb->buf = NULL;
                    615:        for (;;) {
                    616:                if (off == fb->sz) {
                    617:                        if (fb->sz == (1U << 31)) {
                    618:                                fprintf(stderr, "%s: input too large\n", file);
                    619:                                break;
                    620:                        }
                    621:                        resize_buf(fb, 65536);
                    622:                }
                    623:                ssz = read(fd, fb->buf + (int)off, fb->sz - off);
                    624:                if (ssz == 0) {
                    625:                        fb->sz = off;
                    626:                        return(1);
                    627:                }
                    628:                if (ssz == -1) {
                    629:                        perror(file);
                    630:                        break;
                    631:                }
                    632:                off += (size_t)ssz;
                    633:        }
                    634:
                    635:        free(fb->buf);
                    636:        fb->buf = NULL;
                    637:        return(0);
                    638: }
                    639:
                    640: static void
                    641: mparse_end(struct mparse *curp)
                    642: {
                    643:
                    644:        if (MANDOCLEVEL_FATAL <= curp->file_status)
                    645:                return;
                    646:
                    647:        if (curp->mdoc && ! mdoc_endparse(curp->mdoc)) {
                    648:                assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    649:                return;
                    650:        }
                    651:
                    652:        if (curp->man && ! man_endparse(curp->man)) {
                    653:                assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    654:                return;
                    655:        }
                    656:
                    657:        if ( ! (curp->man || curp->mdoc)) {
1.6       kristaps  658:                mandoc_msg(MANDOCERR_NOTMANUAL, curp, 1, 0, NULL);
1.1       kristaps  659:                curp->file_status = MANDOCLEVEL_FATAL;
1.6       kristaps  660:                return;
1.1       kristaps  661:        }
                    662:
                    663:        roff_endparse(curp->roff);
                    664: }
                    665:
                    666: static void
1.28      joerg     667: mparse_parse_buffer(struct mparse *curp, struct buf blk, const char *file,
                    668:                int re)
                    669: {
                    670:        const char      *svfile;
                    671:
                    672:        /* Line number is per-file. */
                    673:        svfile = curp->file;
                    674:        curp->file = file;
                    675:        curp->line = 1;
                    676:
                    677:        mparse_buf_r(curp, blk, 1);
                    678:
                    679:        if (0 == re && MANDOCLEVEL_FATAL > curp->file_status)
                    680:                mparse_end(curp);
                    681:
                    682:        curp->file = svfile;
                    683: }
                    684:
                    685: enum mandoclevel
                    686: mparse_readmem(struct mparse *curp, const void *buf, size_t len,
                    687:                const char *file)
                    688: {
                    689:        struct buf blk;
                    690:
                    691:        blk.buf = UNCONST(buf);
                    692:        blk.sz = len;
                    693:
                    694:        mparse_parse_buffer(curp, blk, file, 0);
                    695:        return(curp->file_status);
                    696: }
                    697:
                    698: static void
1.1       kristaps  699: mparse_readfd_r(struct mparse *curp, int fd, const char *file, int re)
                    700: {
1.28      joerg     701:        struct buf       blk;
                    702:        int              with_mmap;
1.1       kristaps  703:
                    704:        if (-1 == fd)
                    705:                if (-1 == (fd = open(file, O_RDONLY, 0))) {
                    706:                        perror(file);
                    707:                        curp->file_status = MANDOCLEVEL_SYSERR;
                    708:                        return;
                    709:                }
1.28      joerg     710:        /*
                    711:         * Run for each opened file; may be called more than once for
                    712:         * each full parse sequence if the opened file is nested (i.e.,
                    713:         * from `so').  Simply sucks in the whole file and moves into
                    714:         * the parse phase for the file.
                    715:         */
1.1       kristaps  716:
1.28      joerg     717:        if ( ! read_whole_file(file, fd, &blk, &with_mmap)) {
                    718:                curp->file_status = MANDOCLEVEL_SYSERR;
                    719:                return;
                    720:        }
1.1       kristaps  721:
1.28      joerg     722:        mparse_parse_buffer(curp, blk, file, re);
1.1       kristaps  723:
1.28      joerg     724: #ifdef HAVE_MMAP
                    725:        if (with_mmap)
                    726:                munmap(blk.buf, blk.sz);
                    727:        else
                    728: #endif
                    729:                free(blk.buf);
1.1       kristaps  730:
                    731:        if (STDIN_FILENO != fd && -1 == close(fd))
                    732:                perror(file);
                    733: }
                    734:
                    735: enum mandoclevel
                    736: mparse_readfd(struct mparse *curp, int fd, const char *file)
                    737: {
                    738:
                    739:        mparse_readfd_r(curp, fd, file, 0);
                    740:        return(curp->file_status);
                    741: }
                    742:
                    743: struct mparse *
1.29      schwarze  744: mparse_alloc(enum mparset inttype, enum mandoclevel wlevel,
                    745:                mandocmsg mmsg, void *arg, char *defos)
1.1       kristaps  746: {
                    747:        struct mparse   *curp;
1.10      kristaps  748:
                    749:        assert(wlevel <= MANDOCLEVEL_FATAL);
1.1       kristaps  750:
                    751:        curp = mandoc_calloc(1, sizeof(struct mparse));
                    752:
1.3       kristaps  753:        curp->wlevel = wlevel;
1.1       kristaps  754:        curp->mmsg = mmsg;
                    755:        curp->arg = arg;
                    756:        curp->inttype = inttype;
1.29      schwarze  757:        curp->defos = defos;
1.1       kristaps  758:
1.33      schwarze  759:        curp->roff = roff_alloc(inttype, curp);
1.1       kristaps  760:        return(curp);
                    761: }
                    762:
                    763: void
                    764: mparse_reset(struct mparse *curp)
                    765: {
                    766:
                    767:        roff_reset(curp->roff);
                    768:
                    769:        if (curp->mdoc)
                    770:                mdoc_reset(curp->mdoc);
                    771:        if (curp->man)
                    772:                man_reset(curp->man);
1.24      kristaps  773:        if (curp->secondary)
                    774:                curp->secondary->sz = 0;
1.1       kristaps  775:
                    776:        curp->file_status = MANDOCLEVEL_OK;
                    777:        curp->mdoc = NULL;
                    778:        curp->man = NULL;
                    779: }
                    780:
                    781: void
                    782: mparse_free(struct mparse *curp)
                    783: {
                    784:
                    785:        if (curp->pmdoc)
                    786:                mdoc_free(curp->pmdoc);
                    787:        if (curp->pman)
                    788:                man_free(curp->pman);
                    789:        if (curp->roff)
                    790:                roff_free(curp->roff);
1.24      kristaps  791:        if (curp->secondary)
                    792:                free(curp->secondary->buf);
1.1       kristaps  793:
1.24      kristaps  794:        free(curp->secondary);
1.1       kristaps  795:        free(curp);
                    796: }
                    797:
                    798: void
                    799: mparse_result(struct mparse *curp, struct mdoc **mdoc, struct man **man)
                    800: {
                    801:
1.9       kristaps  802:        if (mdoc)
                    803:                *mdoc = curp->mdoc;
                    804:        if (man)
                    805:                *man = curp->man;
1.3       kristaps  806: }
                    807:
                    808: void
                    809: mandoc_vmsg(enum mandocerr t, struct mparse *m,
                    810:                int ln, int pos, const char *fmt, ...)
                    811: {
                    812:        char             buf[256];
                    813:        va_list          ap;
                    814:
                    815:        va_start(ap, fmt);
                    816:        vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
                    817:        va_end(ap);
                    818:
                    819:        mandoc_msg(t, m, ln, pos, buf);
                    820: }
                    821:
                    822: void
                    823: mandoc_msg(enum mandocerr er, struct mparse *m,
                    824:                int ln, int col, const char *msg)
                    825: {
                    826:        enum mandoclevel level;
                    827:
                    828:        level = MANDOCLEVEL_FATAL;
                    829:        while (er < mandoclimits[level])
                    830:                level--;
                    831:
                    832:        if (level < m->wlevel)
                    833:                return;
                    834:
1.8       kristaps  835:        if (m->mmsg)
                    836:                (*m->mmsg)(er, level, m->file, ln, col, msg);
1.3       kristaps  837:
                    838:        if (m->file_status < level)
                    839:                m->file_status = level;
1.7       kristaps  840: }
                    841:
                    842: const char *
                    843: mparse_strerror(enum mandocerr er)
                    844: {
                    845:
                    846:        return(mandocerrs[er]);
                    847: }
                    848:
                    849: const char *
                    850: mparse_strlevel(enum mandoclevel lvl)
                    851: {
                    852:        return(mandoclevels[lvl]);
1.24      kristaps  853: }
                    854:
                    855: void
                    856: mparse_keep(struct mparse *p)
                    857: {
                    858:
                    859:        assert(NULL == p->secondary);
                    860:        p->secondary = mandoc_calloc(1, sizeof(struct buf));
                    861: }
                    862:
                    863: const char *
                    864: mparse_getkeep(const struct mparse *p)
                    865: {
                    866:
                    867:        assert(p->secondary);
                    868:        return(p->secondary->sz ? p->secondary->buf : NULL);
1.1       kristaps  869: }

CVSweb