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

Annotation of mandoc/read.c, Revision 1.165

1.165   ! schwarze    1: /*     $Id: read.c,v 1.164 2017/04/29 12:45:42 schwarze Exp $ */
1.1       kristaps    2: /*
                      3:  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.156     schwarze    4:  * Copyright (c) 2010-2017 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:  *
1.133     schwarze   11:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1       kristaps   12:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.133     schwarze   13:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1       kristaps   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: #include "config.h"
                     20:
1.80      schwarze   21: #include <sys/types.h>
1.82      schwarze   22: #include <sys/mman.h>
1.80      schwarze   23: #include <sys/stat.h>
1.1       kristaps   24:
                     25: #include <assert.h>
                     26: #include <ctype.h>
1.146     schwarze   27: #if HAVE_ERR
1.143     schwarze   28: #include <err.h>
1.146     schwarze   29: #endif
1.40      schwarze   30: #include <errno.h>
1.1       kristaps   31: #include <fcntl.h>
1.3       kristaps   32: #include <stdarg.h>
1.1       kristaps   33: #include <stdio.h>
                     34: #include <stdlib.h>
                     35: #include <string.h>
                     36: #include <unistd.h>
1.140     schwarze   37: #include <zlib.h>
1.1       kristaps   38:
1.133     schwarze   39: #include "mandoc_aux.h"
1.1       kristaps   40: #include "mandoc.h"
1.133     schwarze   41: #include "roff.h"
1.1       kristaps   42: #include "mdoc.h"
                     43: #include "man.h"
1.133     schwarze   44: #include "libmandoc.h"
1.139     schwarze   45: #include "roff_int.h"
1.1       kristaps   46:
                     47: #define        REPARSE_LIMIT   1000
                     48:
                     49: struct mparse {
1.160     schwarze   50:        struct roff      *roff; /* roff parser (!NULL) */
1.134     schwarze   51:        struct roff_man  *man; /* man parser */
1.45      schwarze   52:        char             *sodest; /* filename pointed to by .so */
1.83      schwarze   53:        const char       *file; /* filename of current input file */
                     54:        struct buf       *primary; /* buffer currently being parsed */
                     55:        struct buf       *secondary; /* preprocessed copy of input */
                     56:        const char       *defos; /* default operating system */
                     57:        mandocmsg         mmsg; /* warning/error message handler */
                     58:        enum mandoclevel  file_status; /* status of current parse */
                     59:        enum mandoclevel  wlevel; /* ignore messages below this */
                     60:        int               options; /* parser options */
1.140     schwarze   61:        int               gzip; /* current input file is gzipped */
1.93      schwarze   62:        int               filenc; /* encoding of the current file */
1.1       kristaps   63:        int               reparse_count; /* finite interp. stack */
1.83      schwarze   64:        int               line; /* line number in the file */
1.1       kristaps   65: };
                     66:
1.84      schwarze   67: static void      choose_parser(struct mparse *);
1.1       kristaps   68: static void      resize_buf(struct buf *, size_t);
1.163     schwarze   69: static int       mparse_buf_r(struct mparse *, struct buf, size_t, int);
1.40      schwarze   70: static int       read_whole_file(struct mparse *, const char *, int,
                     71:                                struct buf *, int *);
1.1       kristaps   72: static void      mparse_end(struct mparse *);
1.37      schwarze   73: static void      mparse_parse_buffer(struct mparse *, struct buf,
                     74:                        const char *);
1.1       kristaps   75:
1.3       kristaps   76: static const enum mandocerr    mandoclimits[MANDOCLEVEL_MAX] = {
                     77:        MANDOCERR_OK,
1.165   ! schwarze   78:        MANDOCERR_STYLE,
1.3       kristaps   79:        MANDOCERR_WARNING,
                     80:        MANDOCERR_ERROR,
1.112     schwarze   81:        MANDOCERR_UNSUPP,
1.3       kristaps   82:        MANDOCERR_MAX,
                     83:        MANDOCERR_MAX
                     84: };
                     85:
1.7       kristaps   86: static const char * const      mandocerrs[MANDOCERR_MAX] = {
                     87:        "ok",
                     88:
1.165   ! schwarze   89:        "generic style suggestion",
1.7       kristaps   90:        "generic warning",
                     91:
                     92:        /* related to the prologue */
1.79      schwarze   93:        "missing manual title, using UNTITLED",
                     94:        "missing manual title, using \"\"",
1.54      schwarze   95:        "lower case character in document title",
1.79      schwarze   96:        "missing manual section, using \"\"",
1.7       kristaps   97:        "unknown manual section",
1.54      schwarze   98:        "missing date, using today's date",
1.7       kristaps   99:        "cannot parse date, using it verbatim",
1.79      schwarze  100:        "missing Os macro, using \"\"",
                    101:        "duplicate prologue macro",
                    102:        "late prologue macro",
                    103:        "skipping late title macro",
1.7       kristaps  104:        "prologue macros out of order",
                    105:
                    106:        /* related to document structure */
                    107:        ".so is fragile, better use ln(1)",
1.50      schwarze  108:        "no document body",
1.54      schwarze  109:        "content before first section header",
                    110:        "first section is not \"NAME\"",
1.156     schwarze  111:        "NAME section without Nm before Nd",
1.128     schwarze  112:        "NAME section without description",
                    113:        "description not at the end of NAME",
                    114:        "bad NAME section content",
1.156     schwarze  115:        "missing comma before name",
1.121     schwarze  116:        "missing description line, using \"\"",
1.162     schwarze  117:        "description line outside NAME section",
1.7       kristaps  118:        "sections out of conventional order",
1.54      schwarze  119:        "duplicate section title",
                    120:        "unexpected section",
1.87      schwarze  121:        "unusual Xr order",
                    122:        "unusual Xr punctuation",
1.86      schwarze  123:        "AUTHORS section without An macro",
1.7       kristaps  124:
                    125:        /* related to macros and nesting */
1.55      schwarze  126:        "obsolete macro",
1.102     schwarze  127:        "macro neither callable nor escaped",
1.7       kristaps  128:        "skipping paragraph macro",
1.31      schwarze  129:        "moving paragraph macro out of list",
1.7       kristaps  130:        "skipping no-space macro",
                    131:        "blocks badly nested",
                    132:        "nested displays are not portable",
1.57      schwarze  133:        "moving content out of list",
1.78      schwarze  134:        "fill mode already enabled, skipping",
                    135:        "fill mode already disabled, skipping",
1.7       kristaps  136:        "line scope broken",
                    137:
                    138:        /* related to missing macro arguments */
1.58      schwarze  139:        "skipping empty request",
                    140:        "conditional request controls empty scope",
1.7       kristaps  141:        "skipping empty macro",
1.124     schwarze  142:        "empty block",
1.62      schwarze  143:        "empty argument, using 0n",
1.60      schwarze  144:        "missing display type, using -ragged",
                    145:        "list type is not the first argument",
1.152     schwarze  146:        "missing -width in -tag list, using 6n",
1.78      schwarze  147:        "missing utility name, using \"\"",
1.123     schwarze  148:        "missing function name, using \"\"",
1.60      schwarze  149:        "empty head in list item",
                    150:        "empty list item",
1.61      schwarze  151:        "missing font type, using \\fR",
                    152:        "unknown font type, using \\fR",
1.103     schwarze  153:        "nothing follows prefix",
1.122     schwarze  154:        "empty reference block",
1.155     schwarze  155:        "missing section argument",
1.60      schwarze  156:        "missing -std argument, adding it",
1.125     schwarze  157:        "missing option string, using \"\"",
                    158:        "missing resource identifier, using \"\"",
1.90      schwarze  159:        "missing eqn box, using \"\"",
1.7       kristaps  160:
                    161:        /* related to bad macro arguments */
1.64      schwarze  162:        "unterminated quoted argument",
1.7       kristaps  163:        "duplicate argument",
1.76      schwarze  164:        "skipping duplicate argument",
1.63      schwarze  165:        "skipping duplicate display type",
                    166:        "skipping duplicate list type",
1.76      schwarze  167:        "skipping -width argument",
1.126     schwarze  168:        "wrong number of cells",
1.7       kristaps  169:        "unknown AT&T UNIX version",
1.88      schwarze  170:        "comma in function argument",
1.89      schwarze  171:        "parenthesis in function name",
1.67      schwarze  172:        "invalid content in Rs block",
1.63      schwarze  173:        "invalid Boolean argument",
                    174:        "unknown font, skipping request",
1.126     schwarze  175:        "odd number of characters in request",
1.7       kristaps  176:
                    177:        /* related to plain text */
1.64      schwarze  178:        "blank line in fill mode, using .sp",
                    179:        "tab in filled text",
                    180:        "whitespace at end of input line",
1.158     schwarze  181:        "new sentence, new line",
1.7       kristaps  182:        "bad comment style",
1.64      schwarze  183:        "invalid escape sequence",
                    184:        "undefined string, using \"\"",
1.16      kristaps  185:
1.118     schwarze  186:        /* related to tables */
                    187:        "tbl line starts with span",
                    188:        "tbl column starts with span",
                    189:        "skipping vertical bar in tbl layout",
                    190:
1.7       kristaps  191:        "generic error",
                    192:
                    193:        /* related to tables */
1.116     schwarze  194:        "non-alphabetic character in tbl options",
                    195:        "skipping unknown tbl option",
                    196:        "missing tbl option argument",
                    197:        "wrong tbl option argument size",
1.118     schwarze  198:        "empty tbl layout",
                    199:        "invalid character in tbl layout",
                    200:        "unmatched parenthesis in tbl layout",
1.119     schwarze  201:        "tbl without any data cells",
                    202:        "ignoring data in spanned tbl cell",
                    203:        "ignoring extra tbl data cells",
                    204:        "data block open at end of tbl",
1.7       kristaps  205:
1.68      schwarze  206:        /* related to document structure and macros */
1.108     schwarze  207:        NULL,
1.7       kristaps  208:        "input stack limit exceeded, infinite loop?",
                    209:        "skipping bad character",
1.68      schwarze  210:        "skipping unknown macro",
1.112     schwarze  211:        "skipping insecure request",
1.73      schwarze  212:        "skipping item outside list",
1.68      schwarze  213:        "skipping column outside column list",
                    214:        "skipping end of block that is not open",
1.115     schwarze  215:        "fewer RS blocks open, skipping",
1.68      schwarze  216:        "inserting missing end of block",
                    217:        "appending missing end of block",
                    218:
                    219:        /* related to request and macro arguments */
1.7       kristaps  220:        "escaped character not allowed in a name",
1.99      schwarze  221:        "NOT IMPLEMENTED: Bd -file",
1.145     schwarze  222:        "skipping display without arguments",
1.71      schwarze  223:        "missing list type, using -item",
1.70      schwarze  224:        "missing manual name, using \"\"",
1.71      schwarze  225:        "uname(3) system call failed, using UNKNOWN",
1.63      schwarze  226:        "unknown standard specifier",
1.71      schwarze  227:        "skipping request without numeric argument",
1.109     schwarze  228:        "NOT IMPLEMENTED: .so with absolute path or \"..\"",
1.110     schwarze  229:        ".so request failed",
1.61      schwarze  230:        "skipping all arguments",
                    231:        "skipping excess arguments",
1.92      kristaps  232:        "divide by zero",
1.112     schwarze  233:
                    234:        "unsupported feature",
                    235:        "input too large",
1.114     schwarze  236:        "unsupported control character",
1.112     schwarze  237:        "unsupported roff request",
1.119     schwarze  238:        "eqn delim option in tbl",
1.118     schwarze  239:        "unsupported tbl layout modifier",
1.112     schwarze  240:        "ignoring macro in table",
1.7       kristaps  241: };
                    242:
                    243: static const char * const      mandoclevels[MANDOCLEVEL_MAX] = {
                    244:        "SUCCESS",
1.165   ! schwarze  245:        "STYLE",
1.7       kristaps  246:        "WARNING",
                    247:        "ERROR",
1.112     schwarze  248:        "UNSUPP",
1.7       kristaps  249:        "BADARG",
                    250:        "SYSERR"
                    251: };
                    252:
1.47      schwarze  253:
1.1       kristaps  254: static void
                    255: resize_buf(struct buf *buf, size_t initial)
                    256: {
                    257:
                    258:        buf->sz = buf->sz > initial/2 ? 2 * buf->sz : initial;
                    259:        buf->buf = mandoc_realloc(buf->buf, buf->sz);
                    260: }
                    261:
                    262: static void
1.84      schwarze  263: choose_parser(struct mparse *curp)
1.1       kristaps  264: {
1.83      schwarze  265:        char            *cp, *ep;
                    266:        int              format;
1.1       kristaps  267:
1.83      schwarze  268:        /*
                    269:         * If neither command line arguments -mdoc or -man select
                    270:         * a parser nor the roff parser found a .Dd or .TH macro
                    271:         * yet, look ahead in the main input buffer.
                    272:         */
                    273:
                    274:        if ((format = roff_getformat(curp->roff)) == 0) {
                    275:                cp = curp->primary->buf;
                    276:                ep = cp + curp->primary->sz;
                    277:                while (cp < ep) {
1.85      schwarze  278:                        if (*cp == '.' || *cp == '\'') {
1.83      schwarze  279:                                cp++;
                    280:                                if (cp[0] == 'D' && cp[1] == 'd') {
                    281:                                        format = MPARSE_MDOC;
                    282:                                        break;
                    283:                                }
                    284:                                if (cp[0] == 'T' && cp[1] == 'H') {
                    285:                                        format = MPARSE_MAN;
                    286:                                        break;
                    287:                                }
                    288:                        }
                    289:                        cp = memchr(cp, '\n', ep - cp);
                    290:                        if (cp == NULL)
                    291:                                break;
                    292:                        cp++;
                    293:                }
1.1       kristaps  294:        }
                    295:
1.83      schwarze  296:        if (format == MPARSE_MDOC) {
1.137     schwarze  297:                curp->man->macroset = MACROSET_MDOC;
1.164     schwarze  298:                if (curp->man->mdocmac == NULL)
                    299:                        curp->man->mdocmac = roffhash_alloc(MDOC_Dd, MDOC_MAX);
1.137     schwarze  300:        } else {
                    301:                curp->man->macroset = MACROSET_MAN;
1.164     schwarze  302:                if (curp->man->manmac == NULL)
                    303:                        curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX);
1.47      schwarze  304:        }
1.164     schwarze  305:        curp->man->first->tok = TOKEN_NONE;
1.1       kristaps  306: }
                    307:
                    308: /*
1.95      schwarze  309:  * Main parse routine for a buffer.
                    310:  * It assumes encoding and line numbering are already set up.
                    311:  * It can recurse directly (for invocations of user-defined
                    312:  * macros, inline equations, and input line traps)
                    313:  * and indirectly (for .so file inclusion).
1.1       kristaps  314:  */
1.163     schwarze  315: static int
1.95      schwarze  316: mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
1.1       kristaps  317: {
                    318:        const struct tbl_span   *span;
                    319:        struct buf       ln;
1.113     schwarze  320:        const char      *save_file;
1.110     schwarze  321:        char            *cp;
1.95      schwarze  322:        size_t           pos; /* byte number in the ln buffer */
1.154     schwarze  323:        size_t           j;  /* auxiliary byte number in the blk buffer */
1.1       kristaps  324:        enum rofferr     rr;
1.100     schwarze  325:        int              of;
1.1       kristaps  326:        int              lnn; /* line number in the real file */
1.110     schwarze  327:        int              fd;
1.1       kristaps  328:        unsigned char    c;
                    329:
1.95      schwarze  330:        memset(&ln, 0, sizeof(ln));
1.1       kristaps  331:
1.47      schwarze  332:        lnn = curp->line;
                    333:        pos = 0;
1.1       kristaps  334:
1.95      schwarze  335:        while (i < blk.sz) {
1.1       kristaps  336:                if (0 == pos && '\0' == blk.buf[i])
                    337:                        break;
                    338:
                    339:                if (start) {
                    340:                        curp->line = lnn;
                    341:                        curp->reparse_count = 0;
1.93      schwarze  342:
                    343:                        if (lnn < 3 &&
                    344:                            curp->filenc & MPARSE_UTF8 &&
1.95      schwarze  345:                            curp->filenc & MPARSE_LATIN1)
                    346:                                curp->filenc = preconv_cue(&blk, i);
1.1       kristaps  347:                }
                    348:
1.95      schwarze  349:                while (i < blk.sz && (start || blk.buf[i] != '\0')) {
1.1       kristaps  350:
                    351:                        /*
                    352:                         * When finding an unescaped newline character,
                    353:                         * leave the character loop to process the line.
                    354:                         * Skip a preceding carriage return, if any.
                    355:                         */
                    356:
1.95      schwarze  357:                        if ('\r' == blk.buf[i] && i + 1 < blk.sz &&
1.1       kristaps  358:                            '\n' == blk.buf[i + 1])
                    359:                                ++i;
                    360:                        if ('\n' == blk.buf[i]) {
                    361:                                ++i;
                    362:                                ++lnn;
                    363:                                break;
                    364:                        }
                    365:
1.35      schwarze  366:                        /*
1.93      schwarze  367:                         * Make sure we have space for the worst
                    368:                         * case of 11 bytes: "\\[u10ffff]\0"
1.35      schwarze  369:                         */
                    370:
1.95      schwarze  371:                        if (pos + 11 > ln.sz)
1.35      schwarze  372:                                resize_buf(&ln, 256);
                    373:
1.47      schwarze  374:                        /*
1.93      schwarze  375:                         * Encode 8-bit input.
1.1       kristaps  376:                         */
                    377:
1.93      schwarze  378:                        c = blk.buf[i];
                    379:                        if (c & 0x80) {
1.95      schwarze  380:                                if ( ! (curp->filenc && preconv_encode(
                    381:                                    &blk, &i, &ln, &pos, &curp->filenc))) {
1.114     schwarze  382:                                        mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
                    383:                                            curp->line, pos, "0x%x", c);
1.93      schwarze  384:                                        ln.buf[pos++] = '?';
                    385:                                        i++;
                    386:                                }
                    387:                                continue;
                    388:                        }
                    389:
                    390:                        /*
                    391:                         * Exclude control characters.
                    392:                         */
1.1       kristaps  393:
1.93      schwarze  394:                        if (c == 0x7f || (c < 0x20 && c != 0x09)) {
1.114     schwarze  395:                                mandoc_vmsg(c == 0x00 || c == 0x04 ||
                    396:                                    c > 0x0a ? MANDOCERR_CHAR_BAD :
                    397:                                    MANDOCERR_CHAR_UNSUPP,
                    398:                                    curp, curp->line, pos, "0x%x", c);
1.1       kristaps  399:                                i++;
1.127     schwarze  400:                                if (c != '\r')
                    401:                                        ln.buf[pos++] = '?';
1.1       kristaps  402:                                continue;
                    403:                        }
                    404:
                    405:                        /* Trailing backslash = a plain char. */
                    406:
1.95      schwarze  407:                        if (blk.buf[i] != '\\' || i + 1 == blk.sz) {
1.1       kristaps  408:                                ln.buf[pos++] = blk.buf[i++];
                    409:                                continue;
                    410:                        }
                    411:
                    412:                        /*
                    413:                         * Found escape and at least one other character.
                    414:                         * When it's a newline character, skip it.
                    415:                         * When there is a carriage return in between,
                    416:                         * skip that one as well.
                    417:                         */
                    418:
1.95      schwarze  419:                        if ('\r' == blk.buf[i + 1] && i + 2 < blk.sz &&
1.1       kristaps  420:                            '\n' == blk.buf[i + 2])
                    421:                                ++i;
                    422:                        if ('\n' == blk.buf[i + 1]) {
                    423:                                i += 2;
                    424:                                ++lnn;
                    425:                                continue;
                    426:                        }
                    427:
1.13      kristaps  428:                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {
1.154     schwarze  429:                                j = i;
1.1       kristaps  430:                                i += 2;
                    431:                                /* Comment, skip to end of line */
1.95      schwarze  432:                                for (; i < blk.sz; ++i) {
1.153     schwarze  433:                                        if (blk.buf[i] != '\n')
                    434:                                                continue;
                    435:                                        if (blk.buf[i - 1] == ' ' ||
                    436:                                            blk.buf[i - 1] == '\t')
                    437:                                                mandoc_msg(
                    438:                                                    MANDOCERR_SPACE_EOL,
                    439:                                                    curp, curp->line,
1.154     schwarze  440:                                                    pos + i-1 - j, NULL);
1.153     schwarze  441:                                        ++i;
                    442:                                        ++lnn;
                    443:                                        break;
1.1       kristaps  444:                                }
                    445:
                    446:                                /* Backout trailing whitespaces */
                    447:                                for (; pos > 0; --pos) {
                    448:                                        if (ln.buf[pos - 1] != ' ')
                    449:                                                break;
                    450:                                        if (pos > 2 && ln.buf[pos - 2] == '\\')
                    451:                                                break;
                    452:                                }
                    453:                                break;
                    454:                        }
                    455:
1.35      schwarze  456:                        /* Catch escaped bogus characters. */
                    457:
                    458:                        c = (unsigned char) blk.buf[i+1];
                    459:
1.47      schwarze  460:                        if ( ! (isascii(c) &&
                    461:                            (isgraph(c) || isblank(c)))) {
1.114     schwarze  462:                                mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
1.78      schwarze  463:                                    curp->line, pos, "0x%x", c);
1.35      schwarze  464:                                i += 2;
                    465:                                ln.buf[pos++] = '?';
                    466:                                continue;
                    467:                        }
                    468:
1.1       kristaps  469:                        /* Some other escape sequence, copy & cont. */
                    470:
                    471:                        ln.buf[pos++] = blk.buf[i++];
                    472:                        ln.buf[pos++] = blk.buf[i++];
                    473:                }
                    474:
1.95      schwarze  475:                if (pos >= ln.sz)
1.1       kristaps  476:                        resize_buf(&ln, 256);
                    477:
                    478:                ln.buf[pos] = '\0';
                    479:
                    480:                /*
                    481:                 * A significant amount of complexity is contained by
                    482:                 * the roff preprocessor.  It's line-oriented but can be
                    483:                 * expressed on one line, so we need at times to
                    484:                 * readjust our starting point and re-run it.  The roff
                    485:                 * preprocessor can also readjust the buffers with new
                    486:                 * data, so we pass them in wholesale.
                    487:                 */
                    488:
                    489:                of = 0;
                    490:
1.24      kristaps  491:                /*
                    492:                 * Maintain a lookaside buffer of all parsed lines.  We
                    493:                 * only do this if mparse_keep() has been invoked (the
                    494:                 * buffer may be accessed with mparse_getkeep()).
                    495:                 */
                    496:
                    497:                if (curp->secondary) {
1.47      schwarze  498:                        curp->secondary->buf = mandoc_realloc(
                    499:                            curp->secondary->buf,
                    500:                            curp->secondary->sz + pos + 2);
                    501:                        memcpy(curp->secondary->buf +
                    502:                            curp->secondary->sz,
                    503:                            ln.buf, pos);
1.24      kristaps  504:                        curp->secondary->sz += pos;
                    505:                        curp->secondary->buf
                    506:                                [curp->secondary->sz] = '\n';
                    507:                        curp->secondary->sz++;
                    508:                        curp->secondary->buf
                    509:                                [curp->secondary->sz] = '\0';
                    510:                }
1.1       kristaps  511: rerun:
1.96      schwarze  512:                rr = roff_parseln(curp->roff, curp->line, &ln, &of);
1.1       kristaps  513:
                    514:                switch (rr) {
1.47      schwarze  515:                case ROFF_REPARSE:
1.163     schwarze  516:                        if (++curp->reparse_count > REPARSE_LIMIT)
1.3       kristaps  517:                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
1.47      schwarze  518:                                    curp->line, pos, NULL);
1.163     schwarze  519:                        else if (mparse_buf_r(curp, ln, of, 0) == 1 ||
                    520:                            start == 1) {
                    521:                                pos = 0;
                    522:                                continue;
                    523:                        }
                    524:                        free(ln.buf);
                    525:                        return 0;
1.47      schwarze  526:                case ROFF_APPEND:
1.95      schwarze  527:                        pos = strlen(ln.buf);
1.1       kristaps  528:                        continue;
1.47      schwarze  529:                case ROFF_RERUN:
1.1       kristaps  530:                        goto rerun;
1.47      schwarze  531:                case ROFF_IGN:
1.1       kristaps  532:                        pos = 0;
                    533:                        continue;
1.47      schwarze  534:                case ROFF_SO:
1.95      schwarze  535:                        if ( ! (curp->options & MPARSE_SO) &&
                    536:                            (i >= blk.sz || blk.buf[i] == '\0')) {
1.45      schwarze  537:                                curp->sodest = mandoc_strdup(ln.buf + of);
                    538:                                free(ln.buf);
1.163     schwarze  539:                                return 1;
1.45      schwarze  540:                        }
1.24      kristaps  541:                        /*
                    542:                         * We remove `so' clauses from our lookaside
                    543:                         * buffer because we're going to descend into
                    544:                         * the file recursively.
                    545:                         */
1.47      schwarze  546:                        if (curp->secondary)
1.25      kristaps  547:                                curp->secondary->sz -= pos + 1;
1.113     schwarze  548:                        save_file = curp->file;
1.148     schwarze  549:                        if ((fd = mparse_open(curp, ln.buf + of)) != -1) {
1.110     schwarze  550:                                mparse_readfd(curp, fd, ln.buf + of);
1.147     schwarze  551:                                close(fd);
1.113     schwarze  552:                                curp->file = save_file;
                    553:                        } else {
                    554:                                curp->file = save_file;
1.52      schwarze  555:                                mandoc_vmsg(MANDOCERR_SO_FAIL,
                    556:                                    curp, curp->line, pos,
                    557:                                    ".so %s", ln.buf + of);
1.110     schwarze  558:                                ln.sz = mandoc_asprintf(&cp,
                    559:                                    ".sp\nSee the file %s.\n.sp",
                    560:                                    ln.buf + of);
                    561:                                free(ln.buf);
                    562:                                ln.buf = cp;
                    563:                                of = 0;
                    564:                                mparse_buf_r(curp, ln, of, 0);
1.52      schwarze  565:                        }
1.1       kristaps  566:                        pos = 0;
                    567:                        continue;
                    568:                default:
                    569:                        break;
                    570:                }
                    571:
1.150     schwarze  572:                if (curp->man->macroset == MACROSET_NONE)
1.84      schwarze  573:                        choose_parser(curp);
1.1       kristaps  574:
1.47      schwarze  575:                /*
1.84      schwarze  576:                 * Lastly, push down into the parsers themselves.
1.1       kristaps  577:                 * If libroff returns ROFF_TBL, then add it to the
                    578:                 * currently open parse.  Since we only get here if
                    579:                 * there does exist data (see tbl_data.c), we're
                    580:                 * guaranteed that something's been allocated.
                    581:                 * Do the same for ROFF_EQN.
                    582:                 */
                    583:
1.139     schwarze  584:                if (rr == ROFF_TBL)
1.100     schwarze  585:                        while ((span = roff_span(curp->roff)) != NULL)
1.139     schwarze  586:                                roff_addtbl(curp->man, span);
                    587:                else if (rr == ROFF_EQN)
                    588:                        roff_addeqn(curp->man, roff_eqn(curp->roff));
                    589:                else if ((curp->man->macroset == MACROSET_MDOC ?
1.135     schwarze  590:                    mdoc_parseln(curp->man, curp->line, ln.buf, of) :
1.100     schwarze  591:                    man_parseln(curp->man, curp->line, ln.buf, of)) == 2)
                    592:                                break;
1.1       kristaps  593:
                    594:                /* Temporary buffers typically are not full. */
                    595:
                    596:                if (0 == start && '\0' == blk.buf[i])
                    597:                        break;
                    598:
                    599:                /* Start the next input line. */
                    600:
                    601:                pos = 0;
                    602:        }
                    603:
                    604:        free(ln.buf);
1.163     schwarze  605:        return 1;
1.1       kristaps  606: }
                    607:
                    608: static int
1.40      schwarze  609: read_whole_file(struct mparse *curp, const char *file, int fd,
                    610:                struct buf *fb, int *with_mmap)
1.1       kristaps  611: {
1.161     schwarze  612:        struct stat      st;
1.140     schwarze  613:        gzFile           gz;
1.1       kristaps  614:        size_t           off;
                    615:        ssize_t          ssz;
1.143     schwarze  616:
                    617:        if (fstat(fd, &st) == -1)
                    618:                err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  619:
                    620:        /*
                    621:         * If we're a regular file, try just reading in the whole entry
                    622:         * via mmap().  This is faster than reading it into blocks, and
                    623:         * since each file is only a few bytes to begin with, I'm not
                    624:         * concerned that this is going to tank any machines.
                    625:         */
                    626:
1.140     schwarze  627:        if (curp->gzip == 0 && S_ISREG(st.st_mode)) {
1.131     schwarze  628:                if (st.st_size > 0x7fffffff) {
1.111     schwarze  629:                        mandoc_msg(MANDOCERR_TOOLARGE, curp, 0, 0, NULL);
1.142     schwarze  630:                        return 0;
1.1       kristaps  631:                }
                    632:                *with_mmap = 1;
                    633:                fb->sz = (size_t)st.st_size;
1.37      schwarze  634:                fb->buf = mmap(NULL, fb->sz, PROT_READ, MAP_SHARED, fd, 0);
1.1       kristaps  635:                if (fb->buf != MAP_FAILED)
1.142     schwarze  636:                        return 1;
1.1       kristaps  637:        }
                    638:
1.140     schwarze  639:        if (curp->gzip) {
1.143     schwarze  640:                if ((gz = gzdopen(fd, "rb")) == NULL)
                    641:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.140     schwarze  642:        } else
                    643:                gz = NULL;
                    644:
1.1       kristaps  645:        /*
                    646:         * If this isn't a regular file (like, say, stdin), then we must
                    647:         * go the old way and just read things in bit by bit.
                    648:         */
                    649:
                    650:        *with_mmap = 0;
                    651:        off = 0;
                    652:        fb->sz = 0;
                    653:        fb->buf = NULL;
                    654:        for (;;) {
                    655:                if (off == fb->sz) {
                    656:                        if (fb->sz == (1U << 31)) {
1.111     schwarze  657:                                mandoc_msg(MANDOCERR_TOOLARGE, curp,
                    658:                                    0, 0, NULL);
1.1       kristaps  659:                                break;
                    660:                        }
                    661:                        resize_buf(fb, 65536);
                    662:                }
1.140     schwarze  663:                ssz = curp->gzip ?
                    664:                    gzread(gz, fb->buf + (int)off, fb->sz - off) :
                    665:                    read(fd, fb->buf + (int)off, fb->sz - off);
1.1       kristaps  666:                if (ssz == 0) {
                    667:                        fb->sz = off;
1.142     schwarze  668:                        return 1;
1.1       kristaps  669:                }
1.143     schwarze  670:                if (ssz == -1)
                    671:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  672:                off += (size_t)ssz;
                    673:        }
                    674:
                    675:        free(fb->buf);
                    676:        fb->buf = NULL;
1.142     schwarze  677:        return 0;
1.1       kristaps  678: }
                    679:
                    680: static void
                    681: mparse_end(struct mparse *curp)
                    682: {
1.135     schwarze  683:        if (curp->man->macroset == MACROSET_NONE)
                    684:                curp->man->macroset = MACROSET_MAN;
                    685:        if (curp->man->macroset == MACROSET_MDOC)
                    686:                mdoc_endparse(curp->man);
                    687:        else
1.111     schwarze  688:                man_endparse(curp->man);
1.1       kristaps  689:        roff_endparse(curp->roff);
                    690: }
                    691:
                    692: static void
1.36      schwarze  693: mparse_parse_buffer(struct mparse *curp, struct buf blk, const char *file)
1.28      joerg     694: {
1.85      schwarze  695:        struct buf      *svprimary;
1.28      joerg     696:        const char      *svfile;
1.95      schwarze  697:        size_t           offset;
1.36      schwarze  698:        static int       recursion_depth;
                    699:
                    700:        if (64 < recursion_depth) {
                    701:                mandoc_msg(MANDOCERR_ROFFLOOP, curp, curp->line, 0, NULL);
                    702:                return;
                    703:        }
1.28      joerg     704:
                    705:        /* Line number is per-file. */
                    706:        svfile = curp->file;
                    707:        curp->file = file;
1.85      schwarze  708:        svprimary = curp->primary;
1.83      schwarze  709:        curp->primary = &blk;
1.28      joerg     710:        curp->line = 1;
1.36      schwarze  711:        recursion_depth++;
1.28      joerg     712:
1.93      schwarze  713:        /* Skip an UTF-8 byte order mark. */
                    714:        if (curp->filenc & MPARSE_UTF8 && blk.sz > 2 &&
                    715:            (unsigned char)blk.buf[0] == 0xef &&
                    716:            (unsigned char)blk.buf[1] == 0xbb &&
                    717:            (unsigned char)blk.buf[2] == 0xbf) {
1.95      schwarze  718:                offset = 3;
1.93      schwarze  719:                curp->filenc &= ~MPARSE_LATIN1;
1.95      schwarze  720:        } else
                    721:                offset = 0;
1.93      schwarze  722:
1.95      schwarze  723:        mparse_buf_r(curp, blk, offset, 1);
1.28      joerg     724:
1.111     schwarze  725:        if (--recursion_depth == 0)
1.28      joerg     726:                mparse_end(curp);
                    727:
1.85      schwarze  728:        curp->primary = svprimary;
1.28      joerg     729:        curp->file = svfile;
                    730: }
                    731:
                    732: enum mandoclevel
1.104     schwarze  733: mparse_readmem(struct mparse *curp, void *buf, size_t len,
1.28      joerg     734:                const char *file)
                    735: {
                    736:        struct buf blk;
                    737:
1.104     schwarze  738:        blk.buf = buf;
1.28      joerg     739:        blk.sz = len;
                    740:
1.36      schwarze  741:        mparse_parse_buffer(curp, blk, file);
1.142     schwarze  742:        return curp->file_status;
1.28      joerg     743: }
                    744:
1.98      schwarze  745: /*
                    746:  * Read the whole file into memory and call the parsers.
                    747:  * Called recursively when an .so request is encountered.
                    748:  */
1.36      schwarze  749: enum mandoclevel
                    750: mparse_readfd(struct mparse *curp, int fd, const char *file)
1.1       kristaps  751: {
1.28      joerg     752:        struct buf       blk;
                    753:        int              with_mmap;
1.93      schwarze  754:        int              save_filenc;
1.1       kristaps  755:
1.91      schwarze  756:        if (read_whole_file(curp, file, fd, &blk, &with_mmap)) {
1.93      schwarze  757:                save_filenc = curp->filenc;
                    758:                curp->filenc = curp->options &
                    759:                    (MPARSE_UTF8 | MPARSE_LATIN1);
1.91      schwarze  760:                mparse_parse_buffer(curp, blk, file);
1.93      schwarze  761:                curp->filenc = save_filenc;
1.91      schwarze  762:                if (with_mmap)
                    763:                        munmap(blk.buf, blk.sz);
                    764:                else
                    765:                        free(blk.buf);
                    766:        }
1.142     schwarze  767:        return curp->file_status;
1.82      schwarze  768: }
                    769:
1.148     schwarze  770: int
                    771: mparse_open(struct mparse *curp, const char *file)
1.82      schwarze  772: {
                    773:        char             *cp;
1.148     schwarze  774:        int               fd;
1.82      schwarze  775:
                    776:        curp->file = file;
1.140     schwarze  777:        cp = strrchr(file, '.');
                    778:        curp->gzip = (cp != NULL && ! strcmp(cp + 1, "gz"));
1.98      schwarze  779:
1.140     schwarze  780:        /* First try to use the filename as it is. */
1.98      schwarze  781:
1.148     schwarze  782:        if ((fd = open(file, O_RDONLY)) != -1)
                    783:                return fd;
1.98      schwarze  784:
1.140     schwarze  785:        /*
                    786:         * If that doesn't work and the filename doesn't
                    787:         * already  end in .gz, try appending .gz.
                    788:         */
1.98      schwarze  789:
1.140     schwarze  790:        if ( ! curp->gzip) {
1.98      schwarze  791:                mandoc_asprintf(&cp, "%s.gz", file);
1.149     schwarze  792:                fd = open(cp, O_RDONLY);
1.108     schwarze  793:                free(cp);
1.148     schwarze  794:                if (fd != -1) {
1.140     schwarze  795:                        curp->gzip = 1;
1.148     schwarze  796:                        return fd;
1.82      schwarze  797:                }
                    798:        }
                    799:
1.140     schwarze  800:        /* Neither worked, give up. */
1.97      schwarze  801:
1.140     schwarze  802:        mandoc_msg(MANDOCERR_FILE, curp, 0, 0, strerror(errno));
1.148     schwarze  803:        return -1;
1.1       kristaps  804: }
                    805:
                    806: struct mparse *
1.94      schwarze  807: mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,
1.144     schwarze  808:     const char *defos)
1.1       kristaps  809: {
                    810:        struct mparse   *curp;
1.10      kristaps  811:
1.1       kristaps  812:        curp = mandoc_calloc(1, sizeof(struct mparse));
                    813:
1.44      schwarze  814:        curp->options = options;
1.3       kristaps  815:        curp->wlevel = wlevel;
1.1       kristaps  816:        curp->mmsg = mmsg;
1.29      schwarze  817:        curp->defos = defos;
1.1       kristaps  818:
1.144     schwarze  819:        curp->roff = roff_alloc(curp, options);
1.137     schwarze  820:        curp->man = roff_man_alloc( curp->roff, curp, curp->defos,
                    821:                curp->options & MPARSE_QUICK ? 1 : 0);
1.136     schwarze  822:        if (curp->options & MPARSE_MDOC) {
1.137     schwarze  823:                curp->man->macroset = MACROSET_MDOC;
1.164     schwarze  824:                if (curp->man->mdocmac == NULL)
                    825:                        curp->man->mdocmac = roffhash_alloc(MDOC_Dd, MDOC_MAX);
1.137     schwarze  826:        } else if (curp->options & MPARSE_MAN) {
                    827:                curp->man->macroset = MACROSET_MAN;
1.164     schwarze  828:                if (curp->man->manmac == NULL)
                    829:                        curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX);
1.136     schwarze  830:        }
1.138     schwarze  831:        curp->man->first->tok = TOKEN_NONE;
1.142     schwarze  832:        return curp;
1.1       kristaps  833: }
                    834:
                    835: void
                    836: mparse_reset(struct mparse *curp)
                    837: {
                    838:        roff_reset(curp->roff);
1.150     schwarze  839:        roff_man_reset(curp->man);
1.160     schwarze  840:
                    841:        free(curp->sodest);
                    842:        curp->sodest = NULL;
                    843:
1.24      kristaps  844:        if (curp->secondary)
                    845:                curp->secondary->sz = 0;
1.1       kristaps  846:
                    847:        curp->file_status = MANDOCLEVEL_OK;
1.159     schwarze  848:        curp->gzip = 0;
1.1       kristaps  849: }
                    850:
                    851: void
                    852: mparse_free(struct mparse *curp)
                    853: {
                    854:
1.164     schwarze  855:        roffhash_free(curp->man->mdocmac);
                    856:        roffhash_free(curp->man->manmac);
1.137     schwarze  857:        roff_man_free(curp->man);
1.160     schwarze  858:        roff_free(curp->roff);
1.24      kristaps  859:        if (curp->secondary)
                    860:                free(curp->secondary->buf);
1.1       kristaps  861:
1.24      kristaps  862:        free(curp->secondary);
1.45      schwarze  863:        free(curp->sodest);
1.1       kristaps  864:        free(curp);
                    865: }
                    866:
                    867: void
1.135     schwarze  868: mparse_result(struct mparse *curp, struct roff_man **man,
                    869:        char **sodest)
1.1       kristaps  870: {
                    871:
1.45      schwarze  872:        if (sodest && NULL != (*sodest = curp->sodest)) {
                    873:                *man = NULL;
                    874:                return;
                    875:        }
1.9       kristaps  876:        if (man)
                    877:                *man = curp->man;
1.157     schwarze  878: }
                    879:
                    880: void
                    881: mparse_updaterc(struct mparse *curp, enum mandoclevel *rc)
                    882: {
                    883:        if (curp->file_status > *rc)
                    884:                *rc = curp->file_status;
1.3       kristaps  885: }
                    886:
                    887: void
                    888: mandoc_vmsg(enum mandocerr t, struct mparse *m,
                    889:                int ln, int pos, const char *fmt, ...)
                    890: {
                    891:        char             buf[256];
                    892:        va_list          ap;
                    893:
                    894:        va_start(ap, fmt);
1.48      schwarze  895:        (void)vsnprintf(buf, sizeof(buf), fmt, ap);
1.3       kristaps  896:        va_end(ap);
                    897:
                    898:        mandoc_msg(t, m, ln, pos, buf);
                    899: }
                    900:
                    901: void
1.47      schwarze  902: mandoc_msg(enum mandocerr er, struct mparse *m,
1.3       kristaps  903:                int ln, int col, const char *msg)
                    904: {
                    905:        enum mandoclevel level;
                    906:
1.112     schwarze  907:        level = MANDOCLEVEL_UNSUPP;
1.3       kristaps  908:        while (er < mandoclimits[level])
                    909:                level--;
                    910:
1.108     schwarze  911:        if (level < m->wlevel && er != MANDOCERR_FILE)
1.3       kristaps  912:                return;
                    913:
1.8       kristaps  914:        if (m->mmsg)
                    915:                (*m->mmsg)(er, level, m->file, ln, col, msg);
1.3       kristaps  916:
                    917:        if (m->file_status < level)
                    918:                m->file_status = level;
1.7       kristaps  919: }
                    920:
                    921: const char *
                    922: mparse_strerror(enum mandocerr er)
                    923: {
                    924:
1.142     schwarze  925:        return mandocerrs[er];
1.7       kristaps  926: }
                    927:
                    928: const char *
                    929: mparse_strlevel(enum mandoclevel lvl)
                    930: {
1.142     schwarze  931:        return mandoclevels[lvl];
1.24      kristaps  932: }
                    933:
                    934: void
                    935: mparse_keep(struct mparse *p)
                    936: {
                    937:
                    938:        assert(NULL == p->secondary);
                    939:        p->secondary = mandoc_calloc(1, sizeof(struct buf));
                    940: }
                    941:
                    942: const char *
                    943: mparse_getkeep(const struct mparse *p)
                    944: {
                    945:
                    946:        assert(p->secondary);
1.142     schwarze  947:        return p->secondary->sz ? p->secondary->buf : NULL;
1.1       kristaps  948: }

CVSweb