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

Annotation of mandoc/read.c, Revision 1.164

1.164   ! schwarze    1: /*     $Id: read.c,v 1.163 2017/03/07 20:00:08 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,
                     78:        MANDOCERR_WARNING,
                     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:
                     89:        "generic warning",
                     90:
                     91:        /* related to the prologue */
1.79      schwarze   92:        "missing manual title, using UNTITLED",
                     93:        "missing manual title, using \"\"",
1.54      schwarze   94:        "lower case character in document title",
1.79      schwarze   95:        "missing manual section, using \"\"",
1.7       kristaps   96:        "unknown manual section",
1.54      schwarze   97:        "missing date, using today's date",
1.7       kristaps   98:        "cannot parse date, using it verbatim",
1.79      schwarze   99:        "missing Os macro, using \"\"",
                    100:        "duplicate prologue macro",
                    101:        "late prologue macro",
                    102:        "skipping late title macro",
1.7       kristaps  103:        "prologue macros out of order",
                    104:
                    105:        /* related to document structure */
                    106:        ".so is fragile, better use ln(1)",
1.50      schwarze  107:        "no document body",
1.54      schwarze  108:        "content before first section header",
                    109:        "first section is not \"NAME\"",
1.156     schwarze  110:        "NAME section without Nm before Nd",
1.128     schwarze  111:        "NAME section without description",
                    112:        "description not at the end of NAME",
                    113:        "bad NAME section content",
1.156     schwarze  114:        "missing comma before name",
1.121     schwarze  115:        "missing description line, using \"\"",
1.162     schwarze  116:        "description line outside NAME section",
1.7       kristaps  117:        "sections out of conventional order",
1.54      schwarze  118:        "duplicate section title",
                    119:        "unexpected section",
1.87      schwarze  120:        "unusual Xr order",
                    121:        "unusual Xr punctuation",
1.86      schwarze  122:        "AUTHORS section without An macro",
1.7       kristaps  123:
                    124:        /* related to macros and nesting */
1.55      schwarze  125:        "obsolete macro",
1.102     schwarze  126:        "macro neither callable nor escaped",
1.7       kristaps  127:        "skipping paragraph macro",
1.31      schwarze  128:        "moving paragraph macro out of list",
1.7       kristaps  129:        "skipping no-space macro",
                    130:        "blocks badly nested",
                    131:        "nested displays are not portable",
1.57      schwarze  132:        "moving content out of list",
1.78      schwarze  133:        "fill mode already enabled, skipping",
                    134:        "fill mode already disabled, skipping",
1.7       kristaps  135:        "line scope broken",
                    136:
                    137:        /* related to missing macro arguments */
1.58      schwarze  138:        "skipping empty request",
                    139:        "conditional request controls empty scope",
1.7       kristaps  140:        "skipping empty macro",
1.124     schwarze  141:        "empty block",
1.62      schwarze  142:        "empty argument, using 0n",
1.60      schwarze  143:        "missing display type, using -ragged",
                    144:        "list type is not the first argument",
1.152     schwarze  145:        "missing -width in -tag list, using 6n",
1.78      schwarze  146:        "missing utility name, using \"\"",
1.123     schwarze  147:        "missing function name, using \"\"",
1.60      schwarze  148:        "empty head in list item",
                    149:        "empty list item",
1.61      schwarze  150:        "missing font type, using \\fR",
                    151:        "unknown font type, using \\fR",
1.103     schwarze  152:        "nothing follows prefix",
1.122     schwarze  153:        "empty reference block",
1.155     schwarze  154:        "missing section argument",
1.60      schwarze  155:        "missing -std argument, adding it",
1.125     schwarze  156:        "missing option string, using \"\"",
                    157:        "missing resource identifier, using \"\"",
1.90      schwarze  158:        "missing eqn box, using \"\"",
1.7       kristaps  159:
                    160:        /* related to bad macro arguments */
1.64      schwarze  161:        "unterminated quoted argument",
1.7       kristaps  162:        "duplicate argument",
1.76      schwarze  163:        "skipping duplicate argument",
1.63      schwarze  164:        "skipping duplicate display type",
                    165:        "skipping duplicate list type",
1.76      schwarze  166:        "skipping -width argument",
1.126     schwarze  167:        "wrong number of cells",
1.7       kristaps  168:        "unknown AT&T UNIX version",
1.88      schwarze  169:        "comma in function argument",
1.89      schwarze  170:        "parenthesis in function name",
1.67      schwarze  171:        "invalid content in Rs block",
1.63      schwarze  172:        "invalid Boolean argument",
                    173:        "unknown font, skipping request",
1.126     schwarze  174:        "odd number of characters in request",
1.7       kristaps  175:
                    176:        /* related to plain text */
1.64      schwarze  177:        "blank line in fill mode, using .sp",
                    178:        "tab in filled text",
                    179:        "whitespace at end of input line",
1.158     schwarze  180:        "new sentence, new line",
1.7       kristaps  181:        "bad comment style",
1.64      schwarze  182:        "invalid escape sequence",
                    183:        "undefined string, using \"\"",
1.16      kristaps  184:
1.118     schwarze  185:        /* related to tables */
                    186:        "tbl line starts with span",
                    187:        "tbl column starts with span",
                    188:        "skipping vertical bar in tbl layout",
                    189:
1.7       kristaps  190:        "generic error",
                    191:
                    192:        /* related to tables */
1.116     schwarze  193:        "non-alphabetic character in tbl options",
                    194:        "skipping unknown tbl option",
                    195:        "missing tbl option argument",
                    196:        "wrong tbl option argument size",
1.118     schwarze  197:        "empty tbl layout",
                    198:        "invalid character in tbl layout",
                    199:        "unmatched parenthesis in tbl layout",
1.119     schwarze  200:        "tbl without any data cells",
                    201:        "ignoring data in spanned tbl cell",
                    202:        "ignoring extra tbl data cells",
                    203:        "data block open at end of tbl",
1.7       kristaps  204:
1.68      schwarze  205:        /* related to document structure and macros */
1.108     schwarze  206:        NULL,
1.7       kristaps  207:        "input stack limit exceeded, infinite loop?",
                    208:        "skipping bad character",
1.68      schwarze  209:        "skipping unknown macro",
1.112     schwarze  210:        "skipping insecure request",
1.73      schwarze  211:        "skipping item outside list",
1.68      schwarze  212:        "skipping column outside column list",
                    213:        "skipping end of block that is not open",
1.115     schwarze  214:        "fewer RS blocks open, skipping",
1.68      schwarze  215:        "inserting missing end of block",
                    216:        "appending missing end of block",
                    217:
                    218:        /* related to request and macro arguments */
1.7       kristaps  219:        "escaped character not allowed in a name",
1.99      schwarze  220:        "NOT IMPLEMENTED: Bd -file",
1.145     schwarze  221:        "skipping display without arguments",
1.71      schwarze  222:        "missing list type, using -item",
1.70      schwarze  223:        "missing manual name, using \"\"",
1.71      schwarze  224:        "uname(3) system call failed, using UNKNOWN",
1.63      schwarze  225:        "unknown standard specifier",
1.71      schwarze  226:        "skipping request without numeric argument",
1.109     schwarze  227:        "NOT IMPLEMENTED: .so with absolute path or \"..\"",
1.110     schwarze  228:        ".so request failed",
1.61      schwarze  229:        "skipping all arguments",
                    230:        "skipping excess arguments",
1.92      kristaps  231:        "divide by zero",
1.112     schwarze  232:
                    233:        "unsupported feature",
                    234:        "input too large",
1.114     schwarze  235:        "unsupported control character",
1.112     schwarze  236:        "unsupported roff request",
1.119     schwarze  237:        "eqn delim option in tbl",
1.118     schwarze  238:        "unsupported tbl layout modifier",
1.112     schwarze  239:        "ignoring macro in table",
1.7       kristaps  240: };
                    241:
                    242: static const char * const      mandoclevels[MANDOCLEVEL_MAX] = {
                    243:        "SUCCESS",
                    244:        "RESERVED",
                    245:        "WARNING",
                    246:        "ERROR",
1.112     schwarze  247:        "UNSUPP",
1.7       kristaps  248:        "BADARG",
                    249:        "SYSERR"
                    250: };
                    251:
1.47      schwarze  252:
1.1       kristaps  253: static void
                    254: resize_buf(struct buf *buf, size_t initial)
                    255: {
                    256:
                    257:        buf->sz = buf->sz > initial/2 ? 2 * buf->sz : initial;
                    258:        buf->buf = mandoc_realloc(buf->buf, buf->sz);
                    259: }
                    260:
                    261: static void
1.84      schwarze  262: choose_parser(struct mparse *curp)
1.1       kristaps  263: {
1.83      schwarze  264:        char            *cp, *ep;
                    265:        int              format;
1.1       kristaps  266:
1.83      schwarze  267:        /*
                    268:         * If neither command line arguments -mdoc or -man select
                    269:         * a parser nor the roff parser found a .Dd or .TH macro
                    270:         * yet, look ahead in the main input buffer.
                    271:         */
                    272:
                    273:        if ((format = roff_getformat(curp->roff)) == 0) {
                    274:                cp = curp->primary->buf;
                    275:                ep = cp + curp->primary->sz;
                    276:                while (cp < ep) {
1.85      schwarze  277:                        if (*cp == '.' || *cp == '\'') {
1.83      schwarze  278:                                cp++;
                    279:                                if (cp[0] == 'D' && cp[1] == 'd') {
                    280:                                        format = MPARSE_MDOC;
                    281:                                        break;
                    282:                                }
                    283:                                if (cp[0] == 'T' && cp[1] == 'H') {
                    284:                                        format = MPARSE_MAN;
                    285:                                        break;
                    286:                                }
                    287:                        }
                    288:                        cp = memchr(cp, '\n', ep - cp);
                    289:                        if (cp == NULL)
                    290:                                break;
                    291:                        cp++;
                    292:                }
1.1       kristaps  293:        }
                    294:
1.83      schwarze  295:        if (format == MPARSE_MDOC) {
1.137     schwarze  296:                curp->man->macroset = MACROSET_MDOC;
1.164   ! schwarze  297:                if (curp->man->mdocmac == NULL)
        !           298:                        curp->man->mdocmac = roffhash_alloc(MDOC_Dd, MDOC_MAX);
1.137     schwarze  299:        } else {
                    300:                curp->man->macroset = MACROSET_MAN;
1.164   ! schwarze  301:                if (curp->man->manmac == NULL)
        !           302:                        curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX);
1.47      schwarze  303:        }
1.164   ! schwarze  304:        curp->man->first->tok = TOKEN_NONE;
1.1       kristaps  305: }
                    306:
                    307: /*
1.95      schwarze  308:  * Main parse routine for a buffer.
                    309:  * It assumes encoding and line numbering are already set up.
                    310:  * It can recurse directly (for invocations of user-defined
                    311:  * macros, inline equations, and input line traps)
                    312:  * and indirectly (for .so file inclusion).
1.1       kristaps  313:  */
1.163     schwarze  314: static int
1.95      schwarze  315: mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
1.1       kristaps  316: {
                    317:        const struct tbl_span   *span;
                    318:        struct buf       ln;
1.113     schwarze  319:        const char      *save_file;
1.110     schwarze  320:        char            *cp;
1.95      schwarze  321:        size_t           pos; /* byte number in the ln buffer */
1.154     schwarze  322:        size_t           j;  /* auxiliary byte number in the blk buffer */
1.1       kristaps  323:        enum rofferr     rr;
1.100     schwarze  324:        int              of;
1.1       kristaps  325:        int              lnn; /* line number in the real file */
1.110     schwarze  326:        int              fd;
1.1       kristaps  327:        unsigned char    c;
                    328:
1.95      schwarze  329:        memset(&ln, 0, sizeof(ln));
1.1       kristaps  330:
1.47      schwarze  331:        lnn = curp->line;
                    332:        pos = 0;
1.1       kristaps  333:
1.95      schwarze  334:        while (i < blk.sz) {
1.1       kristaps  335:                if (0 == pos && '\0' == blk.buf[i])
                    336:                        break;
                    337:
                    338:                if (start) {
                    339:                        curp->line = lnn;
                    340:                        curp->reparse_count = 0;
1.93      schwarze  341:
                    342:                        if (lnn < 3 &&
                    343:                            curp->filenc & MPARSE_UTF8 &&
1.95      schwarze  344:                            curp->filenc & MPARSE_LATIN1)
                    345:                                curp->filenc = preconv_cue(&blk, i);
1.1       kristaps  346:                }
                    347:
1.95      schwarze  348:                while (i < blk.sz && (start || blk.buf[i] != '\0')) {
1.1       kristaps  349:
                    350:                        /*
                    351:                         * When finding an unescaped newline character,
                    352:                         * leave the character loop to process the line.
                    353:                         * Skip a preceding carriage return, if any.
                    354:                         */
                    355:
1.95      schwarze  356:                        if ('\r' == blk.buf[i] && i + 1 < blk.sz &&
1.1       kristaps  357:                            '\n' == blk.buf[i + 1])
                    358:                                ++i;
                    359:                        if ('\n' == blk.buf[i]) {
                    360:                                ++i;
                    361:                                ++lnn;
                    362:                                break;
                    363:                        }
                    364:
1.35      schwarze  365:                        /*
1.93      schwarze  366:                         * Make sure we have space for the worst
                    367:                         * case of 11 bytes: "\\[u10ffff]\0"
1.35      schwarze  368:                         */
                    369:
1.95      schwarze  370:                        if (pos + 11 > ln.sz)
1.35      schwarze  371:                                resize_buf(&ln, 256);
                    372:
1.47      schwarze  373:                        /*
1.93      schwarze  374:                         * Encode 8-bit input.
1.1       kristaps  375:                         */
                    376:
1.93      schwarze  377:                        c = blk.buf[i];
                    378:                        if (c & 0x80) {
1.95      schwarze  379:                                if ( ! (curp->filenc && preconv_encode(
                    380:                                    &blk, &i, &ln, &pos, &curp->filenc))) {
1.114     schwarze  381:                                        mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
                    382:                                            curp->line, pos, "0x%x", c);
1.93      schwarze  383:                                        ln.buf[pos++] = '?';
                    384:                                        i++;
                    385:                                }
                    386:                                continue;
                    387:                        }
                    388:
                    389:                        /*
                    390:                         * Exclude control characters.
                    391:                         */
1.1       kristaps  392:
1.93      schwarze  393:                        if (c == 0x7f || (c < 0x20 && c != 0x09)) {
1.114     schwarze  394:                                mandoc_vmsg(c == 0x00 || c == 0x04 ||
                    395:                                    c > 0x0a ? MANDOCERR_CHAR_BAD :
                    396:                                    MANDOCERR_CHAR_UNSUPP,
                    397:                                    curp, curp->line, pos, "0x%x", c);
1.1       kristaps  398:                                i++;
1.127     schwarze  399:                                if (c != '\r')
                    400:                                        ln.buf[pos++] = '?';
1.1       kristaps  401:                                continue;
                    402:                        }
                    403:
                    404:                        /* Trailing backslash = a plain char. */
                    405:
1.95      schwarze  406:                        if (blk.buf[i] != '\\' || i + 1 == blk.sz) {
1.1       kristaps  407:                                ln.buf[pos++] = blk.buf[i++];
                    408:                                continue;
                    409:                        }
                    410:
                    411:                        /*
                    412:                         * Found escape and at least one other character.
                    413:                         * When it's a newline character, skip it.
                    414:                         * When there is a carriage return in between,
                    415:                         * skip that one as well.
                    416:                         */
                    417:
1.95      schwarze  418:                        if ('\r' == blk.buf[i + 1] && i + 2 < blk.sz &&
1.1       kristaps  419:                            '\n' == blk.buf[i + 2])
                    420:                                ++i;
                    421:                        if ('\n' == blk.buf[i + 1]) {
                    422:                                i += 2;
                    423:                                ++lnn;
                    424:                                continue;
                    425:                        }
                    426:
1.13      kristaps  427:                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {
1.154     schwarze  428:                                j = i;
1.1       kristaps  429:                                i += 2;
                    430:                                /* Comment, skip to end of line */
1.95      schwarze  431:                                for (; i < blk.sz; ++i) {
1.153     schwarze  432:                                        if (blk.buf[i] != '\n')
                    433:                                                continue;
                    434:                                        if (blk.buf[i - 1] == ' ' ||
                    435:                                            blk.buf[i - 1] == '\t')
                    436:                                                mandoc_msg(
                    437:                                                    MANDOCERR_SPACE_EOL,
                    438:                                                    curp, curp->line,
1.154     schwarze  439:                                                    pos + i-1 - j, NULL);
1.153     schwarze  440:                                        ++i;
                    441:                                        ++lnn;
                    442:                                        break;
1.1       kristaps  443:                                }
                    444:
                    445:                                /* Backout trailing whitespaces */
                    446:                                for (; pos > 0; --pos) {
                    447:                                        if (ln.buf[pos - 1] != ' ')
                    448:                                                break;
                    449:                                        if (pos > 2 && ln.buf[pos - 2] == '\\')
                    450:                                                break;
                    451:                                }
                    452:                                break;
                    453:                        }
                    454:
1.35      schwarze  455:                        /* Catch escaped bogus characters. */
                    456:
                    457:                        c = (unsigned char) blk.buf[i+1];
                    458:
1.47      schwarze  459:                        if ( ! (isascii(c) &&
                    460:                            (isgraph(c) || isblank(c)))) {
1.114     schwarze  461:                                mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
1.78      schwarze  462:                                    curp->line, pos, "0x%x", c);
1.35      schwarze  463:                                i += 2;
                    464:                                ln.buf[pos++] = '?';
                    465:                                continue;
                    466:                        }
                    467:
1.1       kristaps  468:                        /* Some other escape sequence, copy & cont. */
                    469:
                    470:                        ln.buf[pos++] = blk.buf[i++];
                    471:                        ln.buf[pos++] = blk.buf[i++];
                    472:                }
                    473:
1.95      schwarze  474:                if (pos >= ln.sz)
1.1       kristaps  475:                        resize_buf(&ln, 256);
                    476:
                    477:                ln.buf[pos] = '\0';
                    478:
                    479:                /*
                    480:                 * A significant amount of complexity is contained by
                    481:                 * the roff preprocessor.  It's line-oriented but can be
                    482:                 * expressed on one line, so we need at times to
                    483:                 * readjust our starting point and re-run it.  The roff
                    484:                 * preprocessor can also readjust the buffers with new
                    485:                 * data, so we pass them in wholesale.
                    486:                 */
                    487:
                    488:                of = 0;
                    489:
1.24      kristaps  490:                /*
                    491:                 * Maintain a lookaside buffer of all parsed lines.  We
                    492:                 * only do this if mparse_keep() has been invoked (the
                    493:                 * buffer may be accessed with mparse_getkeep()).
                    494:                 */
                    495:
                    496:                if (curp->secondary) {
1.47      schwarze  497:                        curp->secondary->buf = mandoc_realloc(
                    498:                            curp->secondary->buf,
                    499:                            curp->secondary->sz + pos + 2);
                    500:                        memcpy(curp->secondary->buf +
                    501:                            curp->secondary->sz,
                    502:                            ln.buf, pos);
1.24      kristaps  503:                        curp->secondary->sz += pos;
                    504:                        curp->secondary->buf
                    505:                                [curp->secondary->sz] = '\n';
                    506:                        curp->secondary->sz++;
                    507:                        curp->secondary->buf
                    508:                                [curp->secondary->sz] = '\0';
                    509:                }
1.1       kristaps  510: rerun:
1.96      schwarze  511:                rr = roff_parseln(curp->roff, curp->line, &ln, &of);
1.1       kristaps  512:
                    513:                switch (rr) {
1.47      schwarze  514:                case ROFF_REPARSE:
1.163     schwarze  515:                        if (++curp->reparse_count > REPARSE_LIMIT)
1.3       kristaps  516:                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
1.47      schwarze  517:                                    curp->line, pos, NULL);
1.163     schwarze  518:                        else if (mparse_buf_r(curp, ln, of, 0) == 1 ||
                    519:                            start == 1) {
                    520:                                pos = 0;
                    521:                                continue;
                    522:                        }
                    523:                        free(ln.buf);
                    524:                        return 0;
1.47      schwarze  525:                case ROFF_APPEND:
1.95      schwarze  526:                        pos = strlen(ln.buf);
1.1       kristaps  527:                        continue;
1.47      schwarze  528:                case ROFF_RERUN:
1.1       kristaps  529:                        goto rerun;
1.47      schwarze  530:                case ROFF_IGN:
1.1       kristaps  531:                        pos = 0;
                    532:                        continue;
1.47      schwarze  533:                case ROFF_SO:
1.95      schwarze  534:                        if ( ! (curp->options & MPARSE_SO) &&
                    535:                            (i >= blk.sz || blk.buf[i] == '\0')) {
1.45      schwarze  536:                                curp->sodest = mandoc_strdup(ln.buf + of);
                    537:                                free(ln.buf);
1.163     schwarze  538:                                return 1;
1.45      schwarze  539:                        }
1.24      kristaps  540:                        /*
                    541:                         * We remove `so' clauses from our lookaside
                    542:                         * buffer because we're going to descend into
                    543:                         * the file recursively.
                    544:                         */
1.47      schwarze  545:                        if (curp->secondary)
1.25      kristaps  546:                                curp->secondary->sz -= pos + 1;
1.113     schwarze  547:                        save_file = curp->file;
1.148     schwarze  548:                        if ((fd = mparse_open(curp, ln.buf + of)) != -1) {
1.110     schwarze  549:                                mparse_readfd(curp, fd, ln.buf + of);
1.147     schwarze  550:                                close(fd);
1.113     schwarze  551:                                curp->file = save_file;
                    552:                        } else {
                    553:                                curp->file = save_file;
1.52      schwarze  554:                                mandoc_vmsg(MANDOCERR_SO_FAIL,
                    555:                                    curp, curp->line, pos,
                    556:                                    ".so %s", ln.buf + of);
1.110     schwarze  557:                                ln.sz = mandoc_asprintf(&cp,
                    558:                                    ".sp\nSee the file %s.\n.sp",
                    559:                                    ln.buf + of);
                    560:                                free(ln.buf);
                    561:                                ln.buf = cp;
                    562:                                of = 0;
                    563:                                mparse_buf_r(curp, ln, of, 0);
1.52      schwarze  564:                        }
1.1       kristaps  565:                        pos = 0;
                    566:                        continue;
                    567:                default:
                    568:                        break;
                    569:                }
                    570:
1.150     schwarze  571:                if (curp->man->macroset == MACROSET_NONE)
1.84      schwarze  572:                        choose_parser(curp);
1.1       kristaps  573:
1.47      schwarze  574:                /*
1.84      schwarze  575:                 * Lastly, push down into the parsers themselves.
1.1       kristaps  576:                 * If libroff returns ROFF_TBL, then add it to the
                    577:                 * currently open parse.  Since we only get here if
                    578:                 * there does exist data (see tbl_data.c), we're
                    579:                 * guaranteed that something's been allocated.
                    580:                 * Do the same for ROFF_EQN.
                    581:                 */
                    582:
1.139     schwarze  583:                if (rr == ROFF_TBL)
1.100     schwarze  584:                        while ((span = roff_span(curp->roff)) != NULL)
1.139     schwarze  585:                                roff_addtbl(curp->man, span);
                    586:                else if (rr == ROFF_EQN)
                    587:                        roff_addeqn(curp->man, roff_eqn(curp->roff));
                    588:                else if ((curp->man->macroset == MACROSET_MDOC ?
1.135     schwarze  589:                    mdoc_parseln(curp->man, curp->line, ln.buf, of) :
1.100     schwarze  590:                    man_parseln(curp->man, curp->line, ln.buf, of)) == 2)
                    591:                                break;
1.1       kristaps  592:
                    593:                /* Temporary buffers typically are not full. */
                    594:
                    595:                if (0 == start && '\0' == blk.buf[i])
                    596:                        break;
                    597:
                    598:                /* Start the next input line. */
                    599:
                    600:                pos = 0;
                    601:        }
                    602:
                    603:        free(ln.buf);
1.163     schwarze  604:        return 1;
1.1       kristaps  605: }
                    606:
                    607: static int
1.40      schwarze  608: read_whole_file(struct mparse *curp, const char *file, int fd,
                    609:                struct buf *fb, int *with_mmap)
1.1       kristaps  610: {
1.161     schwarze  611:        struct stat      st;
1.140     schwarze  612:        gzFile           gz;
1.1       kristaps  613:        size_t           off;
                    614:        ssize_t          ssz;
1.143     schwarze  615:
                    616:        if (fstat(fd, &st) == -1)
                    617:                err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  618:
                    619:        /*
                    620:         * If we're a regular file, try just reading in the whole entry
                    621:         * via mmap().  This is faster than reading it into blocks, and
                    622:         * since each file is only a few bytes to begin with, I'm not
                    623:         * concerned that this is going to tank any machines.
                    624:         */
                    625:
1.140     schwarze  626:        if (curp->gzip == 0 && S_ISREG(st.st_mode)) {
1.131     schwarze  627:                if (st.st_size > 0x7fffffff) {
1.111     schwarze  628:                        mandoc_msg(MANDOCERR_TOOLARGE, curp, 0, 0, NULL);
1.142     schwarze  629:                        return 0;
1.1       kristaps  630:                }
                    631:                *with_mmap = 1;
                    632:                fb->sz = (size_t)st.st_size;
1.37      schwarze  633:                fb->buf = mmap(NULL, fb->sz, PROT_READ, MAP_SHARED, fd, 0);
1.1       kristaps  634:                if (fb->buf != MAP_FAILED)
1.142     schwarze  635:                        return 1;
1.1       kristaps  636:        }
                    637:
1.140     schwarze  638:        if (curp->gzip) {
1.143     schwarze  639:                if ((gz = gzdopen(fd, "rb")) == NULL)
                    640:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.140     schwarze  641:        } else
                    642:                gz = NULL;
                    643:
1.1       kristaps  644:        /*
                    645:         * If this isn't a regular file (like, say, stdin), then we must
                    646:         * go the old way and just read things in bit by bit.
                    647:         */
                    648:
                    649:        *with_mmap = 0;
                    650:        off = 0;
                    651:        fb->sz = 0;
                    652:        fb->buf = NULL;
                    653:        for (;;) {
                    654:                if (off == fb->sz) {
                    655:                        if (fb->sz == (1U << 31)) {
1.111     schwarze  656:                                mandoc_msg(MANDOCERR_TOOLARGE, curp,
                    657:                                    0, 0, NULL);
1.1       kristaps  658:                                break;
                    659:                        }
                    660:                        resize_buf(fb, 65536);
                    661:                }
1.140     schwarze  662:                ssz = curp->gzip ?
                    663:                    gzread(gz, fb->buf + (int)off, fb->sz - off) :
                    664:                    read(fd, fb->buf + (int)off, fb->sz - off);
1.1       kristaps  665:                if (ssz == 0) {
                    666:                        fb->sz = off;
1.142     schwarze  667:                        return 1;
1.1       kristaps  668:                }
1.143     schwarze  669:                if (ssz == -1)
                    670:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  671:                off += (size_t)ssz;
                    672:        }
                    673:
                    674:        free(fb->buf);
                    675:        fb->buf = NULL;
1.142     schwarze  676:        return 0;
1.1       kristaps  677: }
                    678:
                    679: static void
                    680: mparse_end(struct mparse *curp)
                    681: {
1.135     schwarze  682:        if (curp->man->macroset == MACROSET_NONE)
                    683:                curp->man->macroset = MACROSET_MAN;
                    684:        if (curp->man->macroset == MACROSET_MDOC)
                    685:                mdoc_endparse(curp->man);
                    686:        else
1.111     schwarze  687:                man_endparse(curp->man);
1.1       kristaps  688:        roff_endparse(curp->roff);
                    689: }
                    690:
                    691: static void
1.36      schwarze  692: mparse_parse_buffer(struct mparse *curp, struct buf blk, const char *file)
1.28      joerg     693: {
1.85      schwarze  694:        struct buf      *svprimary;
1.28      joerg     695:        const char      *svfile;
1.95      schwarze  696:        size_t           offset;
1.36      schwarze  697:        static int       recursion_depth;
                    698:
                    699:        if (64 < recursion_depth) {
                    700:                mandoc_msg(MANDOCERR_ROFFLOOP, curp, curp->line, 0, NULL);
                    701:                return;
                    702:        }
1.28      joerg     703:
                    704:        /* Line number is per-file. */
                    705:        svfile = curp->file;
                    706:        curp->file = file;
1.85      schwarze  707:        svprimary = curp->primary;
1.83      schwarze  708:        curp->primary = &blk;
1.28      joerg     709:        curp->line = 1;
1.36      schwarze  710:        recursion_depth++;
1.28      joerg     711:
1.93      schwarze  712:        /* Skip an UTF-8 byte order mark. */
                    713:        if (curp->filenc & MPARSE_UTF8 && blk.sz > 2 &&
                    714:            (unsigned char)blk.buf[0] == 0xef &&
                    715:            (unsigned char)blk.buf[1] == 0xbb &&
                    716:            (unsigned char)blk.buf[2] == 0xbf) {
1.95      schwarze  717:                offset = 3;
1.93      schwarze  718:                curp->filenc &= ~MPARSE_LATIN1;
1.95      schwarze  719:        } else
                    720:                offset = 0;
1.93      schwarze  721:
1.95      schwarze  722:        mparse_buf_r(curp, blk, offset, 1);
1.28      joerg     723:
1.111     schwarze  724:        if (--recursion_depth == 0)
1.28      joerg     725:                mparse_end(curp);
                    726:
1.85      schwarze  727:        curp->primary = svprimary;
1.28      joerg     728:        curp->file = svfile;
                    729: }
                    730:
                    731: enum mandoclevel
1.104     schwarze  732: mparse_readmem(struct mparse *curp, void *buf, size_t len,
1.28      joerg     733:                const char *file)
                    734: {
                    735:        struct buf blk;
                    736:
1.104     schwarze  737:        blk.buf = buf;
1.28      joerg     738:        blk.sz = len;
                    739:
1.36      schwarze  740:        mparse_parse_buffer(curp, blk, file);
1.142     schwarze  741:        return curp->file_status;
1.28      joerg     742: }
                    743:
1.98      schwarze  744: /*
                    745:  * Read the whole file into memory and call the parsers.
                    746:  * Called recursively when an .so request is encountered.
                    747:  */
1.36      schwarze  748: enum mandoclevel
                    749: mparse_readfd(struct mparse *curp, int fd, const char *file)
1.1       kristaps  750: {
1.28      joerg     751:        struct buf       blk;
                    752:        int              with_mmap;
1.93      schwarze  753:        int              save_filenc;
1.1       kristaps  754:
1.91      schwarze  755:        if (read_whole_file(curp, file, fd, &blk, &with_mmap)) {
1.93      schwarze  756:                save_filenc = curp->filenc;
                    757:                curp->filenc = curp->options &
                    758:                    (MPARSE_UTF8 | MPARSE_LATIN1);
1.91      schwarze  759:                mparse_parse_buffer(curp, blk, file);
1.93      schwarze  760:                curp->filenc = save_filenc;
1.91      schwarze  761:                if (with_mmap)
                    762:                        munmap(blk.buf, blk.sz);
                    763:                else
                    764:                        free(blk.buf);
                    765:        }
1.142     schwarze  766:        return curp->file_status;
1.82      schwarze  767: }
                    768:
1.148     schwarze  769: int
                    770: mparse_open(struct mparse *curp, const char *file)
1.82      schwarze  771: {
                    772:        char             *cp;
1.148     schwarze  773:        int               fd;
1.82      schwarze  774:
                    775:        curp->file = file;
1.140     schwarze  776:        cp = strrchr(file, '.');
                    777:        curp->gzip = (cp != NULL && ! strcmp(cp + 1, "gz"));
1.98      schwarze  778:
1.140     schwarze  779:        /* First try to use the filename as it is. */
1.98      schwarze  780:
1.148     schwarze  781:        if ((fd = open(file, O_RDONLY)) != -1)
                    782:                return fd;
1.98      schwarze  783:
1.140     schwarze  784:        /*
                    785:         * If that doesn't work and the filename doesn't
                    786:         * already  end in .gz, try appending .gz.
                    787:         */
1.98      schwarze  788:
1.140     schwarze  789:        if ( ! curp->gzip) {
1.98      schwarze  790:                mandoc_asprintf(&cp, "%s.gz", file);
1.149     schwarze  791:                fd = open(cp, O_RDONLY);
1.108     schwarze  792:                free(cp);
1.148     schwarze  793:                if (fd != -1) {
1.140     schwarze  794:                        curp->gzip = 1;
1.148     schwarze  795:                        return fd;
1.82      schwarze  796:                }
                    797:        }
                    798:
1.140     schwarze  799:        /* Neither worked, give up. */
1.97      schwarze  800:
1.140     schwarze  801:        mandoc_msg(MANDOCERR_FILE, curp, 0, 0, strerror(errno));
1.148     schwarze  802:        return -1;
1.1       kristaps  803: }
                    804:
                    805: struct mparse *
1.94      schwarze  806: mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,
1.144     schwarze  807:     const char *defos)
1.1       kristaps  808: {
                    809:        struct mparse   *curp;
1.10      kristaps  810:
1.1       kristaps  811:        curp = mandoc_calloc(1, sizeof(struct mparse));
                    812:
1.44      schwarze  813:        curp->options = options;
1.3       kristaps  814:        curp->wlevel = wlevel;
1.1       kristaps  815:        curp->mmsg = mmsg;
1.29      schwarze  816:        curp->defos = defos;
1.1       kristaps  817:
1.144     schwarze  818:        curp->roff = roff_alloc(curp, options);
1.137     schwarze  819:        curp->man = roff_man_alloc( curp->roff, curp, curp->defos,
                    820:                curp->options & MPARSE_QUICK ? 1 : 0);
1.136     schwarze  821:        if (curp->options & MPARSE_MDOC) {
1.137     schwarze  822:                curp->man->macroset = MACROSET_MDOC;
1.164   ! schwarze  823:                if (curp->man->mdocmac == NULL)
        !           824:                        curp->man->mdocmac = roffhash_alloc(MDOC_Dd, MDOC_MAX);
1.137     schwarze  825:        } else if (curp->options & MPARSE_MAN) {
                    826:                curp->man->macroset = MACROSET_MAN;
1.164   ! schwarze  827:                if (curp->man->manmac == NULL)
        !           828:                        curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX);
1.136     schwarze  829:        }
1.138     schwarze  830:        curp->man->first->tok = TOKEN_NONE;
1.142     schwarze  831:        return curp;
1.1       kristaps  832: }
                    833:
                    834: void
                    835: mparse_reset(struct mparse *curp)
                    836: {
                    837:        roff_reset(curp->roff);
1.150     schwarze  838:        roff_man_reset(curp->man);
1.160     schwarze  839:
                    840:        free(curp->sodest);
                    841:        curp->sodest = NULL;
                    842:
1.24      kristaps  843:        if (curp->secondary)
                    844:                curp->secondary->sz = 0;
1.1       kristaps  845:
                    846:        curp->file_status = MANDOCLEVEL_OK;
1.159     schwarze  847:        curp->gzip = 0;
1.1       kristaps  848: }
                    849:
                    850: void
                    851: mparse_free(struct mparse *curp)
                    852: {
                    853:
1.164   ! schwarze  854:        roffhash_free(curp->man->mdocmac);
        !           855:        roffhash_free(curp->man->manmac);
1.137     schwarze  856:        roff_man_free(curp->man);
1.160     schwarze  857:        roff_free(curp->roff);
1.24      kristaps  858:        if (curp->secondary)
                    859:                free(curp->secondary->buf);
1.1       kristaps  860:
1.24      kristaps  861:        free(curp->secondary);
1.45      schwarze  862:        free(curp->sodest);
1.1       kristaps  863:        free(curp);
                    864: }
                    865:
                    866: void
1.135     schwarze  867: mparse_result(struct mparse *curp, struct roff_man **man,
                    868:        char **sodest)
1.1       kristaps  869: {
                    870:
1.45      schwarze  871:        if (sodest && NULL != (*sodest = curp->sodest)) {
                    872:                *man = NULL;
                    873:                return;
                    874:        }
1.9       kristaps  875:        if (man)
                    876:                *man = curp->man;
1.157     schwarze  877: }
                    878:
                    879: void
                    880: mparse_updaterc(struct mparse *curp, enum mandoclevel *rc)
                    881: {
                    882:        if (curp->file_status > *rc)
                    883:                *rc = curp->file_status;
1.3       kristaps  884: }
                    885:
                    886: void
                    887: mandoc_vmsg(enum mandocerr t, struct mparse *m,
                    888:                int ln, int pos, const char *fmt, ...)
                    889: {
                    890:        char             buf[256];
                    891:        va_list          ap;
                    892:
                    893:        va_start(ap, fmt);
1.48      schwarze  894:        (void)vsnprintf(buf, sizeof(buf), fmt, ap);
1.3       kristaps  895:        va_end(ap);
                    896:
                    897:        mandoc_msg(t, m, ln, pos, buf);
                    898: }
                    899:
                    900: void
1.47      schwarze  901: mandoc_msg(enum mandocerr er, struct mparse *m,
1.3       kristaps  902:                int ln, int col, const char *msg)
                    903: {
                    904:        enum mandoclevel level;
                    905:
1.112     schwarze  906:        level = MANDOCLEVEL_UNSUPP;
1.3       kristaps  907:        while (er < mandoclimits[level])
                    908:                level--;
                    909:
1.108     schwarze  910:        if (level < m->wlevel && er != MANDOCERR_FILE)
1.3       kristaps  911:                return;
                    912:
1.8       kristaps  913:        if (m->mmsg)
                    914:                (*m->mmsg)(er, level, m->file, ln, col, msg);
1.3       kristaps  915:
                    916:        if (m->file_status < level)
                    917:                m->file_status = level;
1.7       kristaps  918: }
                    919:
                    920: const char *
                    921: mparse_strerror(enum mandocerr er)
                    922: {
                    923:
1.142     schwarze  924:        return mandocerrs[er];
1.7       kristaps  925: }
                    926:
                    927: const char *
                    928: mparse_strlevel(enum mandoclevel lvl)
                    929: {
1.142     schwarze  930:        return mandoclevels[lvl];
1.24      kristaps  931: }
                    932:
                    933: void
                    934: mparse_keep(struct mparse *p)
                    935: {
                    936:
                    937:        assert(NULL == p->secondary);
                    938:        p->secondary = mandoc_calloc(1, sizeof(struct buf));
                    939: }
                    940:
                    941: const char *
                    942: mparse_getkeep(const struct mparse *p)
                    943: {
                    944:
                    945:        assert(p->secondary);
1.142     schwarze  946:        return p->secondary->sz ? p->secondary->buf : NULL;
1.1       kristaps  947: }

CVSweb