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

Annotation of mandoc/read.c, Revision 1.76

1.76    ! schwarze    1: /*     $Id: read.c,v 1.75 2014/07/31 00:41:10 schwarze Exp $ */
1.1       kristaps    2: /*
                      3:  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.40      schwarze    4:  * Copyright (c) 2010-2014 Ingo Schwarze <schwarze@openbsd.org>
1.42      schwarze    5:  * Copyright (c) 2010, 2012 Joerg Sonnenberger <joerg@netbsd.org>
1.1       kristaps    6:  *
                      7:  * Permission to use, copy, modify, and distribute this software for any
                      8:  * purpose with or without fee is hereby granted, provided that the above
                      9:  * copyright notice and this permission notice appear in all copies.
                     10:  *
                     11:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
                     12:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
                     13:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
                     14:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     15:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     16:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     17:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
                     18:  */
1.11      kristaps   19: #ifdef HAVE_CONFIG_H
                     20: #include "config.h"
                     21: #endif
                     22:
1.15      kristaps   23: #ifdef HAVE_MMAP
                     24: # include <sys/stat.h>
                     25: # include <sys/mman.h>
                     26: #endif
1.1       kristaps   27:
                     28: #include <assert.h>
                     29: #include <ctype.h>
1.40      schwarze   30: #include <errno.h>
1.1       kristaps   31: #include <fcntl.h>
1.3       kristaps   32: #include <stdarg.h>
1.28      joerg      33: #include <stdint.h>
1.1       kristaps   34: #include <stdio.h>
                     35: #include <stdlib.h>
                     36: #include <string.h>
                     37: #include <unistd.h>
                     38:
                     39: #include "mandoc.h"
1.46      schwarze   40: #include "mandoc_aux.h"
1.3       kristaps   41: #include "libmandoc.h"
1.1       kristaps   42: #include "mdoc.h"
                     43: #include "man.h"
1.28      joerg      44: #include "main.h"
1.1       kristaps   45:
                     46: #define        REPARSE_LIMIT   1000
                     47:
                     48: struct buf {
1.47      schwarze   49:        char             *buf; /* binary input buffer */
1.1       kristaps   50:        size_t            sz; /* size of binary buffer */
                     51: };
                     52:
                     53: struct mparse {
                     54:        enum mandoclevel  file_status; /* status of current parse */
1.3       kristaps   55:        enum mandoclevel  wlevel; /* ignore messages below this */
1.1       kristaps   56:        int               line; /* line number in the file */
1.44      schwarze   57:        int               options; /* parser options */
1.1       kristaps   58:        struct man       *pman; /* persistent man parser */
                     59:        struct mdoc      *pmdoc; /* persistent mdoc parser */
                     60:        struct man       *man; /* man parser */
                     61:        struct mdoc      *mdoc; /* mdoc parser */
                     62:        struct roff      *roff; /* roff parser (!NULL) */
1.45      schwarze   63:        char             *sodest; /* filename pointed to by .so */
1.1       kristaps   64:        int               reparse_count; /* finite interp. stack */
                     65:        mandocmsg         mmsg; /* warning/error message handler */
1.47      schwarze   66:        const char       *file;
1.24      kristaps   67:        struct buf       *secondary;
1.69      schwarze   68:        const char       *defos; /* default operating system */
1.1       kristaps   69: };
                     70:
                     71: static void      resize_buf(struct buf *, size_t);
                     72: static void      mparse_buf_r(struct mparse *, struct buf, int);
                     73: static void      pset(const char *, int, struct mparse *);
1.40      schwarze   74: static int       read_whole_file(struct mparse *, const char *, int,
                     75:                                struct buf *, int *);
1.1       kristaps   76: static void      mparse_end(struct mparse *);
1.37      schwarze   77: static void      mparse_parse_buffer(struct mparse *, struct buf,
                     78:                        const char *);
1.1       kristaps   79:
1.3       kristaps   80: static const enum mandocerr    mandoclimits[MANDOCLEVEL_MAX] = {
                     81:        MANDOCERR_OK,
                     82:        MANDOCERR_WARNING,
                     83:        MANDOCERR_WARNING,
                     84:        MANDOCERR_ERROR,
                     85:        MANDOCERR_FATAL,
                     86:        MANDOCERR_MAX,
                     87:        MANDOCERR_MAX
                     88: };
                     89:
1.7       kristaps   90: static const char * const      mandocerrs[MANDOCERR_MAX] = {
                     91:        "ok",
                     92:
                     93:        "generic warning",
                     94:
                     95:        /* related to the prologue */
1.54      schwarze   96:        "missing .TH macro, using \"unknown 1\"",
                     97:        "lower case character in document title",
1.7       kristaps   98:        "unknown manual section",
1.32      schwarze   99:        "unknown manual volume or arch",
1.54      schwarze  100:        "missing date, using today's date",
1.7       kristaps  101:        "cannot parse date, using it verbatim",
                    102:        "prologue macros out of order",
                    103:        "duplicate prologue macro",
1.54      schwarze  104:        "incomplete prologue, terminated by",
                    105:        "skipping prologue macro in body",
1.7       kristaps  106:
                    107:        /* related to document structure */
                    108:        ".so is fragile, better use ln(1)",
1.50      schwarze  109:        "no document body",
1.54      schwarze  110:        "content before first section header",
                    111:        "first section is not \"NAME\"",
1.7       kristaps  112:        "bad NAME section contents",
                    113:        "sections out of conventional order",
1.54      schwarze  114:        "duplicate section title",
                    115:        "unexpected section",
1.7       kristaps  116:
                    117:        /* related to macros and nesting */
1.55      schwarze  118:        "obsolete macro",
1.7       kristaps  119:        "skipping paragraph macro",
1.31      schwarze  120:        "moving paragraph macro out of list",
1.7       kristaps  121:        "skipping no-space macro",
                    122:        "blocks badly nested",
                    123:        "nested displays are not portable",
1.57      schwarze  124:        "moving content out of list",
                    125:        ".Vt block has child macro",
1.56      schwarze  126:        "fill mode already enabled, skipping .fi",
                    127:        "fill mode already disabled, skipping .nf",
1.7       kristaps  128:        "line scope broken",
                    129:
                    130:        /* related to missing macro arguments */
1.58      schwarze  131:        "skipping empty request",
                    132:        "conditional request controls empty scope",
1.7       kristaps  133:        "skipping empty macro",
1.62      schwarze  134:        "empty argument, using 0n",
1.7       kristaps  135:        "argument count wrong",
1.60      schwarze  136:        "missing display type, using -ragged",
                    137:        "list type is not the first argument",
                    138:        "missing -width in -tag list, using 8n",
1.70      schwarze  139:        "missing name for .Ex, using \"\"",
1.60      schwarze  140:        "empty head in list item",
                    141:        "empty list item",
1.61      schwarze  142:        "missing font type, using \\fR",
                    143:        "unknown font type, using \\fR",
1.60      schwarze  144:        "missing -std argument, adding it",
1.7       kristaps  145:
                    146:        /* related to bad macro arguments */
1.64      schwarze  147:        "unterminated quoted argument",
1.7       kristaps  148:        "duplicate argument",
1.76    ! schwarze  149:        "skipping duplicate argument",
1.63      schwarze  150:        "skipping duplicate display type",
                    151:        "skipping duplicate list type",
1.76    ! schwarze  152:        "skipping -width argument",
1.7       kristaps  153:        "unknown AT&T UNIX version",
1.67      schwarze  154:        "invalid content in Rs block",
1.63      schwarze  155:        "invalid Boolean argument",
                    156:        "unknown font, skipping request",
1.7       kristaps  157:
                    158:        /* related to plain text */
1.64      schwarze  159:        "blank line in fill mode, using .sp",
                    160:        "tab in filled text",
                    161:        "whitespace at end of input line",
1.7       kristaps  162:        "bad comment style",
1.64      schwarze  163:        "invalid escape sequence",
                    164:        "undefined string, using \"\"",
1.16      kristaps  165:
1.7       kristaps  166:        "generic error",
1.17      kristaps  167:
                    168:        /* related to equations */
1.20      kristaps  169:        "unexpected equation scope closure",
                    170:        "equation scope open on exit",
1.21      kristaps  171:        "overlapping equation scopes",
                    172:        "unexpected end of equation",
                    173:        "equation syntax error",
1.7       kristaps  174:
                    175:        /* related to tables */
                    176:        "bad table syntax",
                    177:        "bad table option",
                    178:        "bad table layout",
                    179:        "no table layout cells specified",
                    180:        "no table data cells specified",
                    181:        "ignore data in cell",
                    182:        "data block still open",
                    183:        "ignoring extra data cells",
                    184:
1.68      schwarze  185:        /* related to document structure and macros */
1.7       kristaps  186:        "input stack limit exceeded, infinite loop?",
                    187:        "skipping bad character",
1.68      schwarze  188:        "skipping unknown macro",
1.73      schwarze  189:        "skipping item outside list",
1.68      schwarze  190:        "skipping column outside column list",
                    191:        "skipping end of block that is not open",
                    192:        "inserting missing end of block",
                    193:        "appending missing end of block",
                    194:
                    195:        /* related to request and macro arguments */
1.7       kristaps  196:        "escaped character not allowed in a name",
                    197:        "argument count wrong",
1.71      schwarze  198:        "missing list type, using -item",
1.70      schwarze  199:        "missing manual name, using \"\"",
1.71      schwarze  200:        "uname(3) system call failed, using UNKNOWN",
1.63      schwarze  201:        "unknown standard specifier",
1.71      schwarze  202:        "skipping request without numeric argument",
1.61      schwarze  203:        "skipping all arguments",
                    204:        "skipping excess arguments",
1.7       kristaps  205:
                    206:        "generic fatal error",
                    207:
1.40      schwarze  208:        "input too large",
1.7       kristaps  209:        "NOT IMPLEMENTED: .Bd -file",
                    210:        "NOT IMPLEMENTED: .so with absolute path or \"..\"",
1.52      schwarze  211:        ".so request failed",
1.7       kristaps  212:        "static buffer exhausted",
1.40      schwarze  213:
                    214:        /* system errors */
1.51      schwarze  215:        NULL,
1.40      schwarze  216:        "cannot stat file",
                    217:        "cannot read file",
1.7       kristaps  218: };
                    219:
                    220: static const char * const      mandoclevels[MANDOCLEVEL_MAX] = {
                    221:        "SUCCESS",
                    222:        "RESERVED",
                    223:        "WARNING",
                    224:        "ERROR",
                    225:        "FATAL",
                    226:        "BADARG",
                    227:        "SYSERR"
                    228: };
                    229:
1.47      schwarze  230:
1.1       kristaps  231: static void
                    232: resize_buf(struct buf *buf, size_t initial)
                    233: {
                    234:
                    235:        buf->sz = buf->sz > initial/2 ? 2 * buf->sz : initial;
                    236:        buf->buf = mandoc_realloc(buf->buf, buf->sz);
                    237: }
                    238:
                    239: static void
                    240: pset(const char *buf, int pos, struct mparse *curp)
                    241: {
                    242:        int              i;
                    243:
                    244:        /*
                    245:         * Try to intuit which kind of manual parser should be used.  If
                    246:         * passed in by command-line (-man, -mdoc), then use that
                    247:         * explicitly.  If passed as -mandoc, then try to guess from the
                    248:         * line: either skip dot-lines, use -mdoc when finding `.Dt', or
                    249:         * default to -man, which is more lenient.
                    250:         *
                    251:         * Separate out pmdoc/pman from mdoc/man: the first persists
                    252:         * through all parsers, while the latter is used per-parse.
                    253:         */
                    254:
                    255:        if ('.' == buf[0] || '\'' == buf[0]) {
                    256:                for (i = 1; buf[i]; i++)
                    257:                        if (' ' != buf[i] && '\t' != buf[i])
                    258:                                break;
                    259:                if ('\0' == buf[i])
                    260:                        return;
                    261:        }
                    262:
1.44      schwarze  263:        if (MPARSE_MDOC & curp->options) {
1.1       kristaps  264:                curp->mdoc = curp->pmdoc;
                    265:                return;
1.44      schwarze  266:        } else if (MPARSE_MAN & curp->options) {
1.1       kristaps  267:                curp->man = curp->pman;
                    268:                return;
                    269:        }
                    270:
                    271:        if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3))  {
1.47      schwarze  272:                if (NULL == curp->pmdoc)
1.44      schwarze  273:                        curp->pmdoc = mdoc_alloc(
                    274:                            curp->roff, curp, curp->defos,
                    275:                            MPARSE_QUICK & curp->options ? 1 : 0);
1.1       kristaps  276:                assert(curp->pmdoc);
                    277:                curp->mdoc = curp->pmdoc;
                    278:                return;
1.47      schwarze  279:        }
1.1       kristaps  280:
1.47      schwarze  281:        if (NULL == curp->pman)
1.44      schwarze  282:                curp->pman = man_alloc(curp->roff, curp,
                    283:                    MPARSE_QUICK & curp->options ? 1 : 0);
1.1       kristaps  284:        assert(curp->pman);
                    285:        curp->man = curp->pman;
                    286: }
                    287:
                    288: /*
                    289:  * Main parse routine for an opened file.  This is called for each
                    290:  * opened file and simply loops around the full input file, possibly
                    291:  * nesting (i.e., with `so').
                    292:  */
                    293: static void
                    294: mparse_buf_r(struct mparse *curp, struct buf blk, int start)
                    295: {
                    296:        const struct tbl_span   *span;
                    297:        struct buf       ln;
                    298:        enum rofferr     rr;
                    299:        int              i, of, rc;
                    300:        int              pos; /* byte number in the ln buffer */
                    301:        int              lnn; /* line number in the real file */
                    302:        unsigned char    c;
                    303:
                    304:        memset(&ln, 0, sizeof(struct buf));
                    305:
1.47      schwarze  306:        lnn = curp->line;
                    307:        pos = 0;
1.1       kristaps  308:
                    309:        for (i = 0; i < (int)blk.sz; ) {
                    310:                if (0 == pos && '\0' == blk.buf[i])
                    311:                        break;
                    312:
                    313:                if (start) {
                    314:                        curp->line = lnn;
                    315:                        curp->reparse_count = 0;
                    316:                }
                    317:
                    318:                while (i < (int)blk.sz && (start || '\0' != blk.buf[i])) {
                    319:
                    320:                        /*
                    321:                         * When finding an unescaped newline character,
                    322:                         * leave the character loop to process the line.
                    323:                         * Skip a preceding carriage return, if any.
                    324:                         */
                    325:
                    326:                        if ('\r' == blk.buf[i] && i + 1 < (int)blk.sz &&
                    327:                            '\n' == blk.buf[i + 1])
                    328:                                ++i;
                    329:                        if ('\n' == blk.buf[i]) {
                    330:                                ++i;
                    331:                                ++lnn;
                    332:                                break;
                    333:                        }
                    334:
1.35      schwarze  335:                        /*
                    336:                         * Make sure we have space for at least
                    337:                         * one backslash and one other character
                    338:                         * and the trailing NUL byte.
                    339:                         */
                    340:
                    341:                        if (pos + 2 >= (int)ln.sz)
                    342:                                resize_buf(&ln, 256);
                    343:
1.47      schwarze  344:                        /*
1.1       kristaps  345:                         * Warn about bogus characters.  If you're using
                    346:                         * non-ASCII encoding, you're screwing your
                    347:                         * readers.  Since I'd rather this not happen,
1.27      joerg     348:                         * I'll be helpful and replace these characters
                    349:                         * with "?", so we don't display gibberish.
                    350:                         * Note to manual writers: use special characters.
1.1       kristaps  351:                         */
                    352:
                    353:                        c = (unsigned char) blk.buf[i];
                    354:
1.47      schwarze  355:                        if ( ! (isascii(c) &&
                    356:                            (isgraph(c) || isblank(c)))) {
1.3       kristaps  357:                                mandoc_msg(MANDOCERR_BADCHAR, curp,
1.47      schwarze  358:                                    curp->line, pos, NULL);
1.1       kristaps  359:                                i++;
1.27      joerg     360:                                ln.buf[pos++] = '?';
1.1       kristaps  361:                                continue;
                    362:                        }
                    363:
                    364:                        /* Trailing backslash = a plain char. */
                    365:
                    366:                        if ('\\' != blk.buf[i] || i + 1 == (int)blk.sz) {
                    367:                                ln.buf[pos++] = blk.buf[i++];
                    368:                                continue;
                    369:                        }
                    370:
                    371:                        /*
                    372:                         * Found escape and at least one other character.
                    373:                         * When it's a newline character, skip it.
                    374:                         * When there is a carriage return in between,
                    375:                         * skip that one as well.
                    376:                         */
                    377:
                    378:                        if ('\r' == blk.buf[i + 1] && i + 2 < (int)blk.sz &&
                    379:                            '\n' == blk.buf[i + 2])
                    380:                                ++i;
                    381:                        if ('\n' == blk.buf[i + 1]) {
                    382:                                i += 2;
                    383:                                ++lnn;
                    384:                                continue;
                    385:                        }
                    386:
1.13      kristaps  387:                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {
1.1       kristaps  388:                                i += 2;
                    389:                                /* Comment, skip to end of line */
                    390:                                for (; i < (int)blk.sz; ++i) {
                    391:                                        if ('\n' == blk.buf[i]) {
                    392:                                                ++i;
                    393:                                                ++lnn;
                    394:                                                break;
                    395:                                        }
                    396:                                }
                    397:
                    398:                                /* Backout trailing whitespaces */
                    399:                                for (; pos > 0; --pos) {
                    400:                                        if (ln.buf[pos - 1] != ' ')
                    401:                                                break;
                    402:                                        if (pos > 2 && ln.buf[pos - 2] == '\\')
                    403:                                                break;
                    404:                                }
                    405:                                break;
                    406:                        }
                    407:
1.35      schwarze  408:                        /* Catch escaped bogus characters. */
                    409:
                    410:                        c = (unsigned char) blk.buf[i+1];
                    411:
1.47      schwarze  412:                        if ( ! (isascii(c) &&
                    413:                            (isgraph(c) || isblank(c)))) {
1.35      schwarze  414:                                mandoc_msg(MANDOCERR_BADCHAR, curp,
1.47      schwarze  415:                                    curp->line, pos, NULL);
1.35      schwarze  416:                                i += 2;
                    417:                                ln.buf[pos++] = '?';
                    418:                                continue;
                    419:                        }
                    420:
1.1       kristaps  421:                        /* Some other escape sequence, copy & cont. */
                    422:
                    423:                        ln.buf[pos++] = blk.buf[i++];
                    424:                        ln.buf[pos++] = blk.buf[i++];
                    425:                }
                    426:
1.47      schwarze  427:                if (pos >= (int)ln.sz)
1.1       kristaps  428:                        resize_buf(&ln, 256);
                    429:
                    430:                ln.buf[pos] = '\0';
                    431:
                    432:                /*
                    433:                 * A significant amount of complexity is contained by
                    434:                 * the roff preprocessor.  It's line-oriented but can be
                    435:                 * expressed on one line, so we need at times to
                    436:                 * readjust our starting point and re-run it.  The roff
                    437:                 * preprocessor can also readjust the buffers with new
                    438:                 * data, so we pass them in wholesale.
                    439:                 */
                    440:
                    441:                of = 0;
                    442:
1.24      kristaps  443:                /*
                    444:                 * Maintain a lookaside buffer of all parsed lines.  We
                    445:                 * only do this if mparse_keep() has been invoked (the
                    446:                 * buffer may be accessed with mparse_getkeep()).
                    447:                 */
                    448:
                    449:                if (curp->secondary) {
1.47      schwarze  450:                        curp->secondary->buf = mandoc_realloc(
                    451:                            curp->secondary->buf,
                    452:                            curp->secondary->sz + pos + 2);
                    453:                        memcpy(curp->secondary->buf +
                    454:                            curp->secondary->sz,
                    455:                            ln.buf, pos);
1.24      kristaps  456:                        curp->secondary->sz += pos;
                    457:                        curp->secondary->buf
                    458:                                [curp->secondary->sz] = '\n';
                    459:                        curp->secondary->sz++;
                    460:                        curp->secondary->buf
                    461:                                [curp->secondary->sz] = '\0';
                    462:                }
1.1       kristaps  463: rerun:
1.47      schwarze  464:                rr = roff_parseln(curp->roff, curp->line,
                    465:                    &ln.buf, &ln.sz, of, &of);
1.1       kristaps  466:
                    467:                switch (rr) {
1.47      schwarze  468:                case ROFF_REPARSE:
1.1       kristaps  469:                        if (REPARSE_LIMIT >= ++curp->reparse_count)
                    470:                                mparse_buf_r(curp, ln, 0);
                    471:                        else
1.3       kristaps  472:                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
1.47      schwarze  473:                                    curp->line, pos, NULL);
1.1       kristaps  474:                        pos = 0;
                    475:                        continue;
1.47      schwarze  476:                case ROFF_APPEND:
1.1       kristaps  477:                        pos = (int)strlen(ln.buf);
                    478:                        continue;
1.47      schwarze  479:                case ROFF_RERUN:
1.1       kristaps  480:                        goto rerun;
1.47      schwarze  481:                case ROFF_IGN:
1.1       kristaps  482:                        pos = 0;
                    483:                        continue;
1.47      schwarze  484:                case ROFF_ERR:
1.1       kristaps  485:                        assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    486:                        break;
1.47      schwarze  487:                case ROFF_SO:
1.45      schwarze  488:                        if (0 == (MPARSE_SO & curp->options) &&
                    489:                            (i >= (int)blk.sz || '\0' == blk.buf[i])) {
                    490:                                curp->sodest = mandoc_strdup(ln.buf + of);
                    491:                                free(ln.buf);
                    492:                                return;
                    493:                        }
1.24      kristaps  494:                        /*
                    495:                         * We remove `so' clauses from our lookaside
                    496:                         * buffer because we're going to descend into
                    497:                         * the file recursively.
                    498:                         */
1.47      schwarze  499:                        if (curp->secondary)
1.25      kristaps  500:                                curp->secondary->sz -= pos + 1;
1.36      schwarze  501:                        mparse_readfd(curp, -1, ln.buf + of);
1.52      schwarze  502:                        if (MANDOCLEVEL_FATAL <= curp->file_status) {
                    503:                                mandoc_vmsg(MANDOCERR_SO_FAIL,
                    504:                                    curp, curp->line, pos,
                    505:                                    ".so %s", ln.buf + of);
1.1       kristaps  506:                                break;
1.52      schwarze  507:                        }
1.1       kristaps  508:                        pos = 0;
                    509:                        continue;
                    510:                default:
                    511:                        break;
                    512:                }
                    513:
                    514:                /*
                    515:                 * If we encounter errors in the recursive parse, make
                    516:                 * sure we don't continue parsing.
                    517:                 */
                    518:
                    519:                if (MANDOCLEVEL_FATAL <= curp->file_status)
                    520:                        break;
                    521:
                    522:                /*
                    523:                 * If input parsers have not been allocated, do so now.
1.14      kristaps  524:                 * We keep these instanced between parsers, but set them
1.1       kristaps  525:                 * locally per parse routine since we can use different
                    526:                 * parsers with each one.
                    527:                 */
                    528:
                    529:                if ( ! (curp->man || curp->mdoc))
                    530:                        pset(ln.buf + of, pos - of, curp);
                    531:
1.47      schwarze  532:                /*
1.1       kristaps  533:                 * Lastly, push down into the parsers themselves.  One
                    534:                 * of these will have already been set in the pset()
                    535:                 * routine.
                    536:                 * If libroff returns ROFF_TBL, then add it to the
                    537:                 * currently open parse.  Since we only get here if
                    538:                 * there does exist data (see tbl_data.c), we're
                    539:                 * guaranteed that something's been allocated.
                    540:                 * Do the same for ROFF_EQN.
                    541:                 */
                    542:
                    543:                rc = -1;
                    544:
                    545:                if (ROFF_TBL == rr)
                    546:                        while (NULL != (span = roff_span(curp->roff))) {
                    547:                                rc = curp->man ?
1.47      schwarze  548:                                    man_addspan(curp->man, span) :
                    549:                                    mdoc_addspan(curp->mdoc, span);
1.1       kristaps  550:                                if (0 == rc)
                    551:                                        break;
                    552:                        }
                    553:                else if (ROFF_EQN == rr)
1.47      schwarze  554:                        rc = curp->mdoc ?
                    555:                            mdoc_addeqn(curp->mdoc,
                    556:                                roff_eqn(curp->roff)) :
                    557:                            man_addeqn(curp->man,
                    558:                                roff_eqn(curp->roff));
1.1       kristaps  559:                else if (curp->man || curp->mdoc)
                    560:                        rc = curp->man ?
1.47      schwarze  561:                            man_parseln(curp->man,
                    562:                                curp->line, ln.buf, of) :
                    563:                            mdoc_parseln(curp->mdoc,
                    564:                                curp->line, ln.buf, of);
1.1       kristaps  565:
                    566:                if (0 == rc) {
                    567:                        assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    568:                        break;
1.41      schwarze  569:                } else if (2 == rc)
                    570:                        break;
1.1       kristaps  571:
                    572:                /* Temporary buffers typically are not full. */
                    573:
                    574:                if (0 == start && '\0' == blk.buf[i])
                    575:                        break;
                    576:
                    577:                /* Start the next input line. */
                    578:
                    579:                pos = 0;
                    580:        }
                    581:
                    582:        free(ln.buf);
                    583: }
                    584:
                    585: static int
1.40      schwarze  586: read_whole_file(struct mparse *curp, const char *file, int fd,
                    587:                struct buf *fb, int *with_mmap)
1.1       kristaps  588: {
                    589:        size_t           off;
                    590:        ssize_t          ssz;
                    591:
1.15      kristaps  592: #ifdef HAVE_MMAP
                    593:        struct stat      st;
1.1       kristaps  594:        if (-1 == fstat(fd, &st)) {
1.40      schwarze  595:                curp->file_status = MANDOCLEVEL_SYSERR;
                    596:                if (curp->mmsg)
                    597:                        (*curp->mmsg)(MANDOCERR_SYSSTAT, curp->file_status,
                    598:                            file, 0, 0, strerror(errno));
1.1       kristaps  599:                return(0);
                    600:        }
                    601:
                    602:        /*
                    603:         * If we're a regular file, try just reading in the whole entry
                    604:         * via mmap().  This is faster than reading it into blocks, and
                    605:         * since each file is only a few bytes to begin with, I'm not
                    606:         * concerned that this is going to tank any machines.
                    607:         */
                    608:
                    609:        if (S_ISREG(st.st_mode)) {
                    610:                if (st.st_size >= (1U << 31)) {
1.40      schwarze  611:                        curp->file_status = MANDOCLEVEL_FATAL;
                    612:                        if (curp->mmsg)
                    613:                                (*curp->mmsg)(MANDOCERR_TOOLARGE,
                    614:                                    curp->file_status, file, 0, 0, NULL);
1.1       kristaps  615:                        return(0);
                    616:                }
                    617:                *with_mmap = 1;
                    618:                fb->sz = (size_t)st.st_size;
1.37      schwarze  619:                fb->buf = mmap(NULL, fb->sz, PROT_READ, MAP_SHARED, fd, 0);
1.1       kristaps  620:                if (fb->buf != MAP_FAILED)
                    621:                        return(1);
                    622:        }
1.15      kristaps  623: #endif
1.1       kristaps  624:
                    625:        /*
                    626:         * If this isn't a regular file (like, say, stdin), then we must
                    627:         * go the old way and just read things in bit by bit.
                    628:         */
                    629:
                    630:        *with_mmap = 0;
                    631:        off = 0;
                    632:        fb->sz = 0;
                    633:        fb->buf = NULL;
                    634:        for (;;) {
                    635:                if (off == fb->sz) {
                    636:                        if (fb->sz == (1U << 31)) {
1.40      schwarze  637:                                curp->file_status = MANDOCLEVEL_FATAL;
                    638:                                if (curp->mmsg)
                    639:                                        (*curp->mmsg)(MANDOCERR_TOOLARGE,
                    640:                                            curp->file_status,
                    641:                                            file, 0, 0, NULL);
1.1       kristaps  642:                                break;
                    643:                        }
                    644:                        resize_buf(fb, 65536);
                    645:                }
                    646:                ssz = read(fd, fb->buf + (int)off, fb->sz - off);
                    647:                if (ssz == 0) {
                    648:                        fb->sz = off;
                    649:                        return(1);
                    650:                }
                    651:                if (ssz == -1) {
1.40      schwarze  652:                        curp->file_status = MANDOCLEVEL_SYSERR;
                    653:                        if (curp->mmsg)
                    654:                                (*curp->mmsg)(MANDOCERR_SYSREAD,
                    655:                                    curp->file_status, file, 0, 0,
                    656:                                    strerror(errno));
1.1       kristaps  657:                        break;
                    658:                }
                    659:                off += (size_t)ssz;
                    660:        }
                    661:
                    662:        free(fb->buf);
                    663:        fb->buf = NULL;
                    664:        return(0);
                    665: }
                    666:
                    667: static void
                    668: mparse_end(struct mparse *curp)
                    669: {
                    670:
                    671:        if (MANDOCLEVEL_FATAL <= curp->file_status)
                    672:                return;
                    673:
1.72      schwarze  674:        if (curp->mdoc == NULL &&
                    675:            curp->man == NULL &&
                    676:            curp->sodest == NULL) {
                    677:                if (curp->options & MPARSE_MDOC)
                    678:                        curp->mdoc = curp->pmdoc;
                    679:                else {
                    680:                        if (curp->pman == NULL)
                    681:                                curp->pman = man_alloc(curp->roff, curp,
                    682:                                    curp->options & MPARSE_QUICK ? 1 : 0);
                    683:                        curp->man = curp->pman;
                    684:                }
                    685:        }
                    686:
1.1       kristaps  687:        if (curp->mdoc && ! mdoc_endparse(curp->mdoc)) {
                    688:                assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    689:                return;
                    690:        }
                    691:
                    692:        if (curp->man && ! man_endparse(curp->man)) {
                    693:                assert(MANDOCLEVEL_FATAL <= curp->file_status);
                    694:                return;
                    695:        }
                    696:
                    697:        roff_endparse(curp->roff);
                    698: }
                    699:
                    700: static void
1.36      schwarze  701: mparse_parse_buffer(struct mparse *curp, struct buf blk, const char *file)
1.28      joerg     702: {
                    703:        const char      *svfile;
1.36      schwarze  704:        static int       recursion_depth;
                    705:
                    706:        if (64 < recursion_depth) {
                    707:                mandoc_msg(MANDOCERR_ROFFLOOP, curp, curp->line, 0, NULL);
                    708:                return;
                    709:        }
1.28      joerg     710:
                    711:        /* Line number is per-file. */
                    712:        svfile = curp->file;
                    713:        curp->file = file;
                    714:        curp->line = 1;
1.36      schwarze  715:        recursion_depth++;
1.28      joerg     716:
                    717:        mparse_buf_r(curp, blk, 1);
                    718:
1.36      schwarze  719:        if (0 == --recursion_depth && MANDOCLEVEL_FATAL > curp->file_status)
1.28      joerg     720:                mparse_end(curp);
                    721:
                    722:        curp->file = svfile;
                    723: }
                    724:
                    725: enum mandoclevel
                    726: mparse_readmem(struct mparse *curp, const void *buf, size_t len,
                    727:                const char *file)
                    728: {
                    729:        struct buf blk;
                    730:
                    731:        blk.buf = UNCONST(buf);
                    732:        blk.sz = len;
                    733:
1.36      schwarze  734:        mparse_parse_buffer(curp, blk, file);
1.28      joerg     735:        return(curp->file_status);
                    736: }
                    737:
1.36      schwarze  738: enum mandoclevel
                    739: mparse_readfd(struct mparse *curp, int fd, const char *file)
1.1       kristaps  740: {
1.28      joerg     741:        struct buf       blk;
                    742:        int              with_mmap;
1.1       kristaps  743:
1.40      schwarze  744:        if (-1 == fd && -1 == (fd = open(file, O_RDONLY, 0))) {
                    745:                curp->file_status = MANDOCLEVEL_SYSERR;
                    746:                if (curp->mmsg)
                    747:                        (*curp->mmsg)(MANDOCERR_SYSOPEN,
                    748:                            curp->file_status,
                    749:                            file, 0, 0, strerror(errno));
                    750:                goto out;
                    751:        }
                    752:
1.28      joerg     753:        /*
                    754:         * Run for each opened file; may be called more than once for
                    755:         * each full parse sequence if the opened file is nested (i.e.,
                    756:         * from `so').  Simply sucks in the whole file and moves into
                    757:         * the parse phase for the file.
                    758:         */
1.1       kristaps  759:
1.40      schwarze  760:        if ( ! read_whole_file(curp, file, fd, &blk, &with_mmap))
1.36      schwarze  761:                goto out;
1.1       kristaps  762:
1.36      schwarze  763:        mparse_parse_buffer(curp, blk, file);
1.1       kristaps  764:
1.28      joerg     765: #ifdef HAVE_MMAP
                    766:        if (with_mmap)
                    767:                munmap(blk.buf, blk.sz);
                    768:        else
                    769: #endif
                    770:                free(blk.buf);
1.1       kristaps  771:
                    772:        if (STDIN_FILENO != fd && -1 == close(fd))
                    773:                perror(file);
1.36      schwarze  774: out:
1.1       kristaps  775:        return(curp->file_status);
                    776: }
                    777:
                    778: struct mparse *
1.44      schwarze  779: mparse_alloc(int options, enum mandoclevel wlevel,
1.69      schwarze  780:                mandocmsg mmsg, const char *defos)
1.1       kristaps  781: {
                    782:        struct mparse   *curp;
1.10      kristaps  783:
                    784:        assert(wlevel <= MANDOCLEVEL_FATAL);
1.1       kristaps  785:
                    786:        curp = mandoc_calloc(1, sizeof(struct mparse));
                    787:
1.44      schwarze  788:        curp->options = options;
1.3       kristaps  789:        curp->wlevel = wlevel;
1.1       kristaps  790:        curp->mmsg = mmsg;
1.29      schwarze  791:        curp->defos = defos;
1.1       kristaps  792:
1.44      schwarze  793:        curp->roff = roff_alloc(curp, options);
1.72      schwarze  794:        if (curp->options & MPARSE_MDOC)
                    795:                curp->pmdoc = mdoc_alloc(
                    796:                    curp->roff, curp, curp->defos,
                    797:                    curp->options & MPARSE_QUICK ? 1 : 0);
                    798:        if (curp->options & MPARSE_MAN)
                    799:                curp->pman = man_alloc(curp->roff, curp,
                    800:                    curp->options & MPARSE_QUICK ? 1 : 0);
                    801:
1.1       kristaps  802:        return(curp);
                    803: }
                    804:
                    805: void
                    806: mparse_reset(struct mparse *curp)
                    807: {
                    808:
                    809:        roff_reset(curp->roff);
                    810:
                    811:        if (curp->mdoc)
                    812:                mdoc_reset(curp->mdoc);
                    813:        if (curp->man)
                    814:                man_reset(curp->man);
1.24      kristaps  815:        if (curp->secondary)
                    816:                curp->secondary->sz = 0;
1.1       kristaps  817:
                    818:        curp->file_status = MANDOCLEVEL_OK;
                    819:        curp->mdoc = NULL;
                    820:        curp->man = NULL;
1.45      schwarze  821:
                    822:        free(curp->sodest);
                    823:        curp->sodest = NULL;
1.1       kristaps  824: }
                    825:
                    826: void
                    827: mparse_free(struct mparse *curp)
                    828: {
                    829:
                    830:        if (curp->pmdoc)
                    831:                mdoc_free(curp->pmdoc);
                    832:        if (curp->pman)
                    833:                man_free(curp->pman);
                    834:        if (curp->roff)
                    835:                roff_free(curp->roff);
1.24      kristaps  836:        if (curp->secondary)
                    837:                free(curp->secondary->buf);
1.1       kristaps  838:
1.24      kristaps  839:        free(curp->secondary);
1.45      schwarze  840:        free(curp->sodest);
1.1       kristaps  841:        free(curp);
                    842: }
                    843:
                    844: void
1.45      schwarze  845: mparse_result(struct mparse *curp,
                    846:        struct mdoc **mdoc, struct man **man, char **sodest)
1.1       kristaps  847: {
                    848:
1.45      schwarze  849:        if (sodest && NULL != (*sodest = curp->sodest)) {
                    850:                *mdoc = NULL;
                    851:                *man = NULL;
                    852:                return;
                    853:        }
1.9       kristaps  854:        if (mdoc)
                    855:                *mdoc = curp->mdoc;
                    856:        if (man)
                    857:                *man = curp->man;
1.3       kristaps  858: }
                    859:
                    860: void
                    861: mandoc_vmsg(enum mandocerr t, struct mparse *m,
                    862:                int ln, int pos, const char *fmt, ...)
                    863: {
                    864:        char             buf[256];
                    865:        va_list          ap;
                    866:
                    867:        va_start(ap, fmt);
1.48      schwarze  868:        (void)vsnprintf(buf, sizeof(buf), fmt, ap);
1.3       kristaps  869:        va_end(ap);
                    870:
                    871:        mandoc_msg(t, m, ln, pos, buf);
                    872: }
                    873:
                    874: void
1.47      schwarze  875: mandoc_msg(enum mandocerr er, struct mparse *m,
1.3       kristaps  876:                int ln, int col, const char *msg)
                    877: {
                    878:        enum mandoclevel level;
                    879:
                    880:        level = MANDOCLEVEL_FATAL;
                    881:        while (er < mandoclimits[level])
                    882:                level--;
                    883:
                    884:        if (level < m->wlevel)
                    885:                return;
                    886:
1.8       kristaps  887:        if (m->mmsg)
                    888:                (*m->mmsg)(er, level, m->file, ln, col, msg);
1.3       kristaps  889:
                    890:        if (m->file_status < level)
                    891:                m->file_status = level;
1.7       kristaps  892: }
                    893:
                    894: const char *
                    895: mparse_strerror(enum mandocerr er)
                    896: {
                    897:
                    898:        return(mandocerrs[er]);
                    899: }
                    900:
                    901: const char *
                    902: mparse_strlevel(enum mandoclevel lvl)
                    903: {
                    904:        return(mandoclevels[lvl]);
1.24      kristaps  905: }
                    906:
                    907: void
                    908: mparse_keep(struct mparse *p)
                    909: {
                    910:
                    911:        assert(NULL == p->secondary);
                    912:        p->secondary = mandoc_calloc(1, sizeof(struct buf));
                    913: }
                    914:
                    915: const char *
                    916: mparse_getkeep(const struct mparse *p)
                    917: {
                    918:
                    919:        assert(p->secondary);
                    920:        return(p->secondary->sz ? p->secondary->buf : NULL);
1.1       kristaps  921: }

CVSweb