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

Annotation of mandoc/read.c, Revision 1.169

1.169   ! schwarze    1: /*     $Id: read.c,v 1.168 2017/06/01 15:25:39 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",
1.166     schwarze   90:
                     91:        "useless macro",
1.167     schwarze   92:        "consider using OS macro",
1.168     schwarze   93:        "description line ends with a full stop",
1.166     schwarze   94:
1.7       kristaps   95:        "generic warning",
                     96:
                     97:        /* related to the prologue */
1.79      schwarze   98:        "missing manual title, using UNTITLED",
                     99:        "missing manual title, using \"\"",
1.54      schwarze  100:        "lower case character in document title",
1.79      schwarze  101:        "missing manual section, using \"\"",
1.7       kristaps  102:        "unknown manual section",
1.54      schwarze  103:        "missing date, using today's date",
1.7       kristaps  104:        "cannot parse date, using it verbatim",
1.79      schwarze  105:        "missing Os macro, using \"\"",
                    106:        "duplicate prologue macro",
                    107:        "late prologue macro",
                    108:        "skipping late title macro",
1.7       kristaps  109:        "prologue macros out of order",
                    110:
                    111:        /* related to document structure */
                    112:        ".so is fragile, better use ln(1)",
1.50      schwarze  113:        "no document body",
1.54      schwarze  114:        "content before first section header",
                    115:        "first section is not \"NAME\"",
1.156     schwarze  116:        "NAME section without Nm before Nd",
1.128     schwarze  117:        "NAME section without description",
                    118:        "description not at the end of NAME",
                    119:        "bad NAME section content",
1.156     schwarze  120:        "missing comma before name",
1.121     schwarze  121:        "missing description line, using \"\"",
1.162     schwarze  122:        "description line outside NAME section",
1.7       kristaps  123:        "sections out of conventional order",
1.54      schwarze  124:        "duplicate section title",
                    125:        "unexpected section",
1.87      schwarze  126:        "unusual Xr order",
                    127:        "unusual Xr punctuation",
1.86      schwarze  128:        "AUTHORS section without An macro",
1.7       kristaps  129:
                    130:        /* related to macros and nesting */
1.55      schwarze  131:        "obsolete macro",
1.102     schwarze  132:        "macro neither callable nor escaped",
1.7       kristaps  133:        "skipping paragraph macro",
1.31      schwarze  134:        "moving paragraph macro out of list",
1.7       kristaps  135:        "skipping no-space macro",
                    136:        "blocks badly nested",
                    137:        "nested displays are not portable",
1.57      schwarze  138:        "moving content out of list",
1.78      schwarze  139:        "fill mode already enabled, skipping",
                    140:        "fill mode already disabled, skipping",
1.7       kristaps  141:        "line scope broken",
1.169   ! schwarze  142:        "skipping blank line in line scope",
1.7       kristaps  143:
                    144:        /* related to missing macro arguments */
1.58      schwarze  145:        "skipping empty request",
                    146:        "conditional request controls empty scope",
1.7       kristaps  147:        "skipping empty macro",
1.124     schwarze  148:        "empty block",
1.62      schwarze  149:        "empty argument, using 0n",
1.60      schwarze  150:        "missing display type, using -ragged",
                    151:        "list type is not the first argument",
1.152     schwarze  152:        "missing -width in -tag list, using 6n",
1.78      schwarze  153:        "missing utility name, using \"\"",
1.123     schwarze  154:        "missing function name, using \"\"",
1.60      schwarze  155:        "empty head in list item",
                    156:        "empty list item",
1.61      schwarze  157:        "missing font type, using \\fR",
                    158:        "unknown font type, using \\fR",
1.103     schwarze  159:        "nothing follows prefix",
1.122     schwarze  160:        "empty reference block",
1.155     schwarze  161:        "missing section argument",
1.60      schwarze  162:        "missing -std argument, adding it",
1.125     schwarze  163:        "missing option string, using \"\"",
                    164:        "missing resource identifier, using \"\"",
1.90      schwarze  165:        "missing eqn box, using \"\"",
1.7       kristaps  166:
                    167:        /* related to bad macro arguments */
1.64      schwarze  168:        "unterminated quoted argument",
1.7       kristaps  169:        "duplicate argument",
1.76      schwarze  170:        "skipping duplicate argument",
1.63      schwarze  171:        "skipping duplicate display type",
                    172:        "skipping duplicate list type",
1.76      schwarze  173:        "skipping -width argument",
1.126     schwarze  174:        "wrong number of cells",
1.7       kristaps  175:        "unknown AT&T UNIX version",
1.88      schwarze  176:        "comma in function argument",
1.89      schwarze  177:        "parenthesis in function name",
1.67      schwarze  178:        "invalid content in Rs block",
1.63      schwarze  179:        "invalid Boolean argument",
                    180:        "unknown font, skipping request",
1.126     schwarze  181:        "odd number of characters in request",
1.7       kristaps  182:
                    183:        /* related to plain text */
1.64      schwarze  184:        "blank line in fill mode, using .sp",
                    185:        "tab in filled text",
                    186:        "whitespace at end of input line",
1.158     schwarze  187:        "new sentence, new line",
1.7       kristaps  188:        "bad comment style",
1.64      schwarze  189:        "invalid escape sequence",
                    190:        "undefined string, using \"\"",
1.16      kristaps  191:
1.118     schwarze  192:        /* related to tables */
                    193:        "tbl line starts with span",
                    194:        "tbl column starts with span",
                    195:        "skipping vertical bar in tbl layout",
                    196:
1.7       kristaps  197:        "generic error",
                    198:
                    199:        /* related to tables */
1.116     schwarze  200:        "non-alphabetic character in tbl options",
                    201:        "skipping unknown tbl option",
                    202:        "missing tbl option argument",
                    203:        "wrong tbl option argument size",
1.118     schwarze  204:        "empty tbl layout",
                    205:        "invalid character in tbl layout",
                    206:        "unmatched parenthesis in tbl layout",
1.119     schwarze  207:        "tbl without any data cells",
                    208:        "ignoring data in spanned tbl cell",
                    209:        "ignoring extra tbl data cells",
                    210:        "data block open at end of tbl",
1.7       kristaps  211:
1.68      schwarze  212:        /* related to document structure and macros */
1.108     schwarze  213:        NULL,
1.7       kristaps  214:        "input stack limit exceeded, infinite loop?",
                    215:        "skipping bad character",
1.68      schwarze  216:        "skipping unknown macro",
1.112     schwarze  217:        "skipping insecure request",
1.73      schwarze  218:        "skipping item outside list",
1.68      schwarze  219:        "skipping column outside column list",
                    220:        "skipping end of block that is not open",
1.115     schwarze  221:        "fewer RS blocks open, skipping",
1.68      schwarze  222:        "inserting missing end of block",
                    223:        "appending missing end of block",
                    224:
                    225:        /* related to request and macro arguments */
1.7       kristaps  226:        "escaped character not allowed in a name",
1.99      schwarze  227:        "NOT IMPLEMENTED: Bd -file",
1.145     schwarze  228:        "skipping display without arguments",
1.71      schwarze  229:        "missing list type, using -item",
1.70      schwarze  230:        "missing manual name, using \"\"",
1.71      schwarze  231:        "uname(3) system call failed, using UNKNOWN",
1.63      schwarze  232:        "unknown standard specifier",
1.71      schwarze  233:        "skipping request without numeric argument",
1.109     schwarze  234:        "NOT IMPLEMENTED: .so with absolute path or \"..\"",
1.110     schwarze  235:        ".so request failed",
1.61      schwarze  236:        "skipping all arguments",
                    237:        "skipping excess arguments",
1.92      kristaps  238:        "divide by zero",
1.112     schwarze  239:
                    240:        "unsupported feature",
                    241:        "input too large",
1.114     schwarze  242:        "unsupported control character",
1.112     schwarze  243:        "unsupported roff request",
1.119     schwarze  244:        "eqn delim option in tbl",
1.118     schwarze  245:        "unsupported tbl layout modifier",
1.112     schwarze  246:        "ignoring macro in table",
1.7       kristaps  247: };
                    248:
                    249: static const char * const      mandoclevels[MANDOCLEVEL_MAX] = {
                    250:        "SUCCESS",
1.165     schwarze  251:        "STYLE",
1.7       kristaps  252:        "WARNING",
                    253:        "ERROR",
1.112     schwarze  254:        "UNSUPP",
1.7       kristaps  255:        "BADARG",
                    256:        "SYSERR"
                    257: };
                    258:
1.47      schwarze  259:
1.1       kristaps  260: static void
                    261: resize_buf(struct buf *buf, size_t initial)
                    262: {
                    263:
                    264:        buf->sz = buf->sz > initial/2 ? 2 * buf->sz : initial;
                    265:        buf->buf = mandoc_realloc(buf->buf, buf->sz);
                    266: }
                    267:
                    268: static void
1.84      schwarze  269: choose_parser(struct mparse *curp)
1.1       kristaps  270: {
1.83      schwarze  271:        char            *cp, *ep;
                    272:        int              format;
1.1       kristaps  273:
1.83      schwarze  274:        /*
                    275:         * If neither command line arguments -mdoc or -man select
                    276:         * a parser nor the roff parser found a .Dd or .TH macro
                    277:         * yet, look ahead in the main input buffer.
                    278:         */
                    279:
                    280:        if ((format = roff_getformat(curp->roff)) == 0) {
                    281:                cp = curp->primary->buf;
                    282:                ep = cp + curp->primary->sz;
                    283:                while (cp < ep) {
1.85      schwarze  284:                        if (*cp == '.' || *cp == '\'') {
1.83      schwarze  285:                                cp++;
                    286:                                if (cp[0] == 'D' && cp[1] == 'd') {
                    287:                                        format = MPARSE_MDOC;
                    288:                                        break;
                    289:                                }
                    290:                                if (cp[0] == 'T' && cp[1] == 'H') {
                    291:                                        format = MPARSE_MAN;
                    292:                                        break;
                    293:                                }
                    294:                        }
                    295:                        cp = memchr(cp, '\n', ep - cp);
                    296:                        if (cp == NULL)
                    297:                                break;
                    298:                        cp++;
                    299:                }
1.1       kristaps  300:        }
                    301:
1.83      schwarze  302:        if (format == MPARSE_MDOC) {
1.137     schwarze  303:                curp->man->macroset = MACROSET_MDOC;
1.164     schwarze  304:                if (curp->man->mdocmac == NULL)
                    305:                        curp->man->mdocmac = roffhash_alloc(MDOC_Dd, MDOC_MAX);
1.137     schwarze  306:        } else {
                    307:                curp->man->macroset = MACROSET_MAN;
1.164     schwarze  308:                if (curp->man->manmac == NULL)
                    309:                        curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX);
1.47      schwarze  310:        }
1.164     schwarze  311:        curp->man->first->tok = TOKEN_NONE;
1.1       kristaps  312: }
                    313:
                    314: /*
1.95      schwarze  315:  * Main parse routine for a buffer.
                    316:  * It assumes encoding and line numbering are already set up.
                    317:  * It can recurse directly (for invocations of user-defined
                    318:  * macros, inline equations, and input line traps)
                    319:  * and indirectly (for .so file inclusion).
1.1       kristaps  320:  */
1.163     schwarze  321: static int
1.95      schwarze  322: mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
1.1       kristaps  323: {
                    324:        const struct tbl_span   *span;
                    325:        struct buf       ln;
1.113     schwarze  326:        const char      *save_file;
1.110     schwarze  327:        char            *cp;
1.95      schwarze  328:        size_t           pos; /* byte number in the ln buffer */
1.154     schwarze  329:        size_t           j;  /* auxiliary byte number in the blk buffer */
1.1       kristaps  330:        enum rofferr     rr;
1.100     schwarze  331:        int              of;
1.1       kristaps  332:        int              lnn; /* line number in the real file */
1.110     schwarze  333:        int              fd;
1.1       kristaps  334:        unsigned char    c;
                    335:
1.95      schwarze  336:        memset(&ln, 0, sizeof(ln));
1.1       kristaps  337:
1.47      schwarze  338:        lnn = curp->line;
                    339:        pos = 0;
1.1       kristaps  340:
1.95      schwarze  341:        while (i < blk.sz) {
1.1       kristaps  342:                if (0 == pos && '\0' == blk.buf[i])
                    343:                        break;
                    344:
                    345:                if (start) {
                    346:                        curp->line = lnn;
                    347:                        curp->reparse_count = 0;
1.93      schwarze  348:
                    349:                        if (lnn < 3 &&
                    350:                            curp->filenc & MPARSE_UTF8 &&
1.95      schwarze  351:                            curp->filenc & MPARSE_LATIN1)
                    352:                                curp->filenc = preconv_cue(&blk, i);
1.1       kristaps  353:                }
                    354:
1.95      schwarze  355:                while (i < blk.sz && (start || blk.buf[i] != '\0')) {
1.1       kristaps  356:
                    357:                        /*
                    358:                         * When finding an unescaped newline character,
                    359:                         * leave the character loop to process the line.
                    360:                         * Skip a preceding carriage return, if any.
                    361:                         */
                    362:
1.95      schwarze  363:                        if ('\r' == blk.buf[i] && i + 1 < blk.sz &&
1.1       kristaps  364:                            '\n' == blk.buf[i + 1])
                    365:                                ++i;
                    366:                        if ('\n' == blk.buf[i]) {
                    367:                                ++i;
                    368:                                ++lnn;
                    369:                                break;
                    370:                        }
                    371:
1.35      schwarze  372:                        /*
1.93      schwarze  373:                         * Make sure we have space for the worst
                    374:                         * case of 11 bytes: "\\[u10ffff]\0"
1.35      schwarze  375:                         */
                    376:
1.95      schwarze  377:                        if (pos + 11 > ln.sz)
1.35      schwarze  378:                                resize_buf(&ln, 256);
                    379:
1.47      schwarze  380:                        /*
1.93      schwarze  381:                         * Encode 8-bit input.
1.1       kristaps  382:                         */
                    383:
1.93      schwarze  384:                        c = blk.buf[i];
                    385:                        if (c & 0x80) {
1.95      schwarze  386:                                if ( ! (curp->filenc && preconv_encode(
                    387:                                    &blk, &i, &ln, &pos, &curp->filenc))) {
1.114     schwarze  388:                                        mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
                    389:                                            curp->line, pos, "0x%x", c);
1.93      schwarze  390:                                        ln.buf[pos++] = '?';
                    391:                                        i++;
                    392:                                }
                    393:                                continue;
                    394:                        }
                    395:
                    396:                        /*
                    397:                         * Exclude control characters.
                    398:                         */
1.1       kristaps  399:
1.93      schwarze  400:                        if (c == 0x7f || (c < 0x20 && c != 0x09)) {
1.114     schwarze  401:                                mandoc_vmsg(c == 0x00 || c == 0x04 ||
                    402:                                    c > 0x0a ? MANDOCERR_CHAR_BAD :
                    403:                                    MANDOCERR_CHAR_UNSUPP,
                    404:                                    curp, curp->line, pos, "0x%x", c);
1.1       kristaps  405:                                i++;
1.127     schwarze  406:                                if (c != '\r')
                    407:                                        ln.buf[pos++] = '?';
1.1       kristaps  408:                                continue;
                    409:                        }
                    410:
                    411:                        /* Trailing backslash = a plain char. */
                    412:
1.95      schwarze  413:                        if (blk.buf[i] != '\\' || i + 1 == blk.sz) {
1.1       kristaps  414:                                ln.buf[pos++] = blk.buf[i++];
                    415:                                continue;
                    416:                        }
                    417:
                    418:                        /*
                    419:                         * Found escape and at least one other character.
                    420:                         * When it's a newline character, skip it.
                    421:                         * When there is a carriage return in between,
                    422:                         * skip that one as well.
                    423:                         */
                    424:
1.95      schwarze  425:                        if ('\r' == blk.buf[i + 1] && i + 2 < blk.sz &&
1.1       kristaps  426:                            '\n' == blk.buf[i + 2])
                    427:                                ++i;
                    428:                        if ('\n' == blk.buf[i + 1]) {
                    429:                                i += 2;
                    430:                                ++lnn;
                    431:                                continue;
                    432:                        }
                    433:
1.13      kristaps  434:                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {
1.154     schwarze  435:                                j = i;
1.1       kristaps  436:                                i += 2;
                    437:                                /* Comment, skip to end of line */
1.95      schwarze  438:                                for (; i < blk.sz; ++i) {
1.153     schwarze  439:                                        if (blk.buf[i] != '\n')
                    440:                                                continue;
                    441:                                        if (blk.buf[i - 1] == ' ' ||
                    442:                                            blk.buf[i - 1] == '\t')
                    443:                                                mandoc_msg(
                    444:                                                    MANDOCERR_SPACE_EOL,
                    445:                                                    curp, curp->line,
1.154     schwarze  446:                                                    pos + i-1 - j, NULL);
1.153     schwarze  447:                                        ++i;
                    448:                                        ++lnn;
                    449:                                        break;
1.1       kristaps  450:                                }
                    451:
                    452:                                /* Backout trailing whitespaces */
                    453:                                for (; pos > 0; --pos) {
                    454:                                        if (ln.buf[pos - 1] != ' ')
                    455:                                                break;
                    456:                                        if (pos > 2 && ln.buf[pos - 2] == '\\')
                    457:                                                break;
                    458:                                }
                    459:                                break;
                    460:                        }
                    461:
1.35      schwarze  462:                        /* Catch escaped bogus characters. */
                    463:
                    464:                        c = (unsigned char) blk.buf[i+1];
                    465:
1.47      schwarze  466:                        if ( ! (isascii(c) &&
                    467:                            (isgraph(c) || isblank(c)))) {
1.114     schwarze  468:                                mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
1.78      schwarze  469:                                    curp->line, pos, "0x%x", c);
1.35      schwarze  470:                                i += 2;
                    471:                                ln.buf[pos++] = '?';
                    472:                                continue;
                    473:                        }
                    474:
1.1       kristaps  475:                        /* Some other escape sequence, copy & cont. */
                    476:
                    477:                        ln.buf[pos++] = blk.buf[i++];
                    478:                        ln.buf[pos++] = blk.buf[i++];
                    479:                }
                    480:
1.95      schwarze  481:                if (pos >= ln.sz)
1.1       kristaps  482:                        resize_buf(&ln, 256);
                    483:
                    484:                ln.buf[pos] = '\0';
                    485:
                    486:                /*
                    487:                 * A significant amount of complexity is contained by
                    488:                 * the roff preprocessor.  It's line-oriented but can be
                    489:                 * expressed on one line, so we need at times to
                    490:                 * readjust our starting point and re-run it.  The roff
                    491:                 * preprocessor can also readjust the buffers with new
                    492:                 * data, so we pass them in wholesale.
                    493:                 */
                    494:
                    495:                of = 0;
                    496:
1.24      kristaps  497:                /*
                    498:                 * Maintain a lookaside buffer of all parsed lines.  We
                    499:                 * only do this if mparse_keep() has been invoked (the
                    500:                 * buffer may be accessed with mparse_getkeep()).
                    501:                 */
                    502:
                    503:                if (curp->secondary) {
1.47      schwarze  504:                        curp->secondary->buf = mandoc_realloc(
                    505:                            curp->secondary->buf,
                    506:                            curp->secondary->sz + pos + 2);
                    507:                        memcpy(curp->secondary->buf +
                    508:                            curp->secondary->sz,
                    509:                            ln.buf, pos);
1.24      kristaps  510:                        curp->secondary->sz += pos;
                    511:                        curp->secondary->buf
                    512:                                [curp->secondary->sz] = '\n';
                    513:                        curp->secondary->sz++;
                    514:                        curp->secondary->buf
                    515:                                [curp->secondary->sz] = '\0';
                    516:                }
1.1       kristaps  517: rerun:
1.96      schwarze  518:                rr = roff_parseln(curp->roff, curp->line, &ln, &of);
1.1       kristaps  519:
                    520:                switch (rr) {
1.47      schwarze  521:                case ROFF_REPARSE:
1.163     schwarze  522:                        if (++curp->reparse_count > REPARSE_LIMIT)
1.3       kristaps  523:                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
1.47      schwarze  524:                                    curp->line, pos, NULL);
1.163     schwarze  525:                        else if (mparse_buf_r(curp, ln, of, 0) == 1 ||
                    526:                            start == 1) {
                    527:                                pos = 0;
                    528:                                continue;
                    529:                        }
                    530:                        free(ln.buf);
                    531:                        return 0;
1.47      schwarze  532:                case ROFF_APPEND:
1.95      schwarze  533:                        pos = strlen(ln.buf);
1.1       kristaps  534:                        continue;
1.47      schwarze  535:                case ROFF_RERUN:
1.1       kristaps  536:                        goto rerun;
1.47      schwarze  537:                case ROFF_IGN:
1.1       kristaps  538:                        pos = 0;
                    539:                        continue;
1.47      schwarze  540:                case ROFF_SO:
1.95      schwarze  541:                        if ( ! (curp->options & MPARSE_SO) &&
                    542:                            (i >= blk.sz || blk.buf[i] == '\0')) {
1.45      schwarze  543:                                curp->sodest = mandoc_strdup(ln.buf + of);
                    544:                                free(ln.buf);
1.163     schwarze  545:                                return 1;
1.45      schwarze  546:                        }
1.24      kristaps  547:                        /*
                    548:                         * We remove `so' clauses from our lookaside
                    549:                         * buffer because we're going to descend into
                    550:                         * the file recursively.
                    551:                         */
1.47      schwarze  552:                        if (curp->secondary)
1.25      kristaps  553:                                curp->secondary->sz -= pos + 1;
1.113     schwarze  554:                        save_file = curp->file;
1.148     schwarze  555:                        if ((fd = mparse_open(curp, ln.buf + of)) != -1) {
1.110     schwarze  556:                                mparse_readfd(curp, fd, ln.buf + of);
1.147     schwarze  557:                                close(fd);
1.113     schwarze  558:                                curp->file = save_file;
                    559:                        } else {
                    560:                                curp->file = save_file;
1.52      schwarze  561:                                mandoc_vmsg(MANDOCERR_SO_FAIL,
                    562:                                    curp, curp->line, pos,
                    563:                                    ".so %s", ln.buf + of);
1.110     schwarze  564:                                ln.sz = mandoc_asprintf(&cp,
                    565:                                    ".sp\nSee the file %s.\n.sp",
                    566:                                    ln.buf + of);
                    567:                                free(ln.buf);
                    568:                                ln.buf = cp;
                    569:                                of = 0;
                    570:                                mparse_buf_r(curp, ln, of, 0);
1.52      schwarze  571:                        }
1.1       kristaps  572:                        pos = 0;
                    573:                        continue;
                    574:                default:
                    575:                        break;
                    576:                }
                    577:
1.150     schwarze  578:                if (curp->man->macroset == MACROSET_NONE)
1.84      schwarze  579:                        choose_parser(curp);
1.1       kristaps  580:
1.47      schwarze  581:                /*
1.84      schwarze  582:                 * Lastly, push down into the parsers themselves.
1.1       kristaps  583:                 * If libroff returns ROFF_TBL, then add it to the
                    584:                 * currently open parse.  Since we only get here if
                    585:                 * there does exist data (see tbl_data.c), we're
                    586:                 * guaranteed that something's been allocated.
                    587:                 * Do the same for ROFF_EQN.
                    588:                 */
                    589:
1.139     schwarze  590:                if (rr == ROFF_TBL)
1.100     schwarze  591:                        while ((span = roff_span(curp->roff)) != NULL)
1.139     schwarze  592:                                roff_addtbl(curp->man, span);
                    593:                else if (rr == ROFF_EQN)
                    594:                        roff_addeqn(curp->man, roff_eqn(curp->roff));
                    595:                else if ((curp->man->macroset == MACROSET_MDOC ?
1.135     schwarze  596:                    mdoc_parseln(curp->man, curp->line, ln.buf, of) :
1.100     schwarze  597:                    man_parseln(curp->man, curp->line, ln.buf, of)) == 2)
                    598:                                break;
1.1       kristaps  599:
                    600:                /* Temporary buffers typically are not full. */
                    601:
                    602:                if (0 == start && '\0' == blk.buf[i])
                    603:                        break;
                    604:
                    605:                /* Start the next input line. */
                    606:
                    607:                pos = 0;
                    608:        }
                    609:
                    610:        free(ln.buf);
1.163     schwarze  611:        return 1;
1.1       kristaps  612: }
                    613:
                    614: static int
1.40      schwarze  615: read_whole_file(struct mparse *curp, const char *file, int fd,
                    616:                struct buf *fb, int *with_mmap)
1.1       kristaps  617: {
1.161     schwarze  618:        struct stat      st;
1.140     schwarze  619:        gzFile           gz;
1.1       kristaps  620:        size_t           off;
                    621:        ssize_t          ssz;
1.143     schwarze  622:
                    623:        if (fstat(fd, &st) == -1)
                    624:                err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  625:
                    626:        /*
                    627:         * If we're a regular file, try just reading in the whole entry
                    628:         * via mmap().  This is faster than reading it into blocks, and
                    629:         * since each file is only a few bytes to begin with, I'm not
                    630:         * concerned that this is going to tank any machines.
                    631:         */
                    632:
1.140     schwarze  633:        if (curp->gzip == 0 && S_ISREG(st.st_mode)) {
1.131     schwarze  634:                if (st.st_size > 0x7fffffff) {
1.111     schwarze  635:                        mandoc_msg(MANDOCERR_TOOLARGE, curp, 0, 0, NULL);
1.142     schwarze  636:                        return 0;
1.1       kristaps  637:                }
                    638:                *with_mmap = 1;
                    639:                fb->sz = (size_t)st.st_size;
1.37      schwarze  640:                fb->buf = mmap(NULL, fb->sz, PROT_READ, MAP_SHARED, fd, 0);
1.1       kristaps  641:                if (fb->buf != MAP_FAILED)
1.142     schwarze  642:                        return 1;
1.1       kristaps  643:        }
                    644:
1.140     schwarze  645:        if (curp->gzip) {
1.143     schwarze  646:                if ((gz = gzdopen(fd, "rb")) == NULL)
                    647:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.140     schwarze  648:        } else
                    649:                gz = NULL;
                    650:
1.1       kristaps  651:        /*
                    652:         * If this isn't a regular file (like, say, stdin), then we must
                    653:         * go the old way and just read things in bit by bit.
                    654:         */
                    655:
                    656:        *with_mmap = 0;
                    657:        off = 0;
                    658:        fb->sz = 0;
                    659:        fb->buf = NULL;
                    660:        for (;;) {
                    661:                if (off == fb->sz) {
                    662:                        if (fb->sz == (1U << 31)) {
1.111     schwarze  663:                                mandoc_msg(MANDOCERR_TOOLARGE, curp,
                    664:                                    0, 0, NULL);
1.1       kristaps  665:                                break;
                    666:                        }
                    667:                        resize_buf(fb, 65536);
                    668:                }
1.140     schwarze  669:                ssz = curp->gzip ?
                    670:                    gzread(gz, fb->buf + (int)off, fb->sz - off) :
                    671:                    read(fd, fb->buf + (int)off, fb->sz - off);
1.1       kristaps  672:                if (ssz == 0) {
                    673:                        fb->sz = off;
1.142     schwarze  674:                        return 1;
1.1       kristaps  675:                }
1.143     schwarze  676:                if (ssz == -1)
                    677:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  678:                off += (size_t)ssz;
                    679:        }
                    680:
                    681:        free(fb->buf);
                    682:        fb->buf = NULL;
1.142     schwarze  683:        return 0;
1.1       kristaps  684: }
                    685:
                    686: static void
                    687: mparse_end(struct mparse *curp)
                    688: {
1.135     schwarze  689:        if (curp->man->macroset == MACROSET_NONE)
                    690:                curp->man->macroset = MACROSET_MAN;
                    691:        if (curp->man->macroset == MACROSET_MDOC)
                    692:                mdoc_endparse(curp->man);
                    693:        else
1.111     schwarze  694:                man_endparse(curp->man);
1.1       kristaps  695:        roff_endparse(curp->roff);
                    696: }
                    697:
                    698: static void
1.36      schwarze  699: mparse_parse_buffer(struct mparse *curp, struct buf blk, const char *file)
1.28      joerg     700: {
1.85      schwarze  701:        struct buf      *svprimary;
1.28      joerg     702:        const char      *svfile;
1.95      schwarze  703:        size_t           offset;
1.36      schwarze  704:        static int       recursion_depth;
                    705:
                    706:        if (64 < recursion_depth) {
                    707:                mandoc_msg(MANDOCERR_ROFFLOOP, curp, curp->line, 0, NULL);
                    708:                return;
                    709:        }
1.28      joerg     710:
                    711:        /* Line number is per-file. */
                    712:        svfile = curp->file;
                    713:        curp->file = file;
1.85      schwarze  714:        svprimary = curp->primary;
1.83      schwarze  715:        curp->primary = &blk;
1.28      joerg     716:        curp->line = 1;
1.36      schwarze  717:        recursion_depth++;
1.28      joerg     718:
1.93      schwarze  719:        /* Skip an UTF-8 byte order mark. */
                    720:        if (curp->filenc & MPARSE_UTF8 && blk.sz > 2 &&
                    721:            (unsigned char)blk.buf[0] == 0xef &&
                    722:            (unsigned char)blk.buf[1] == 0xbb &&
                    723:            (unsigned char)blk.buf[2] == 0xbf) {
1.95      schwarze  724:                offset = 3;
1.93      schwarze  725:                curp->filenc &= ~MPARSE_LATIN1;
1.95      schwarze  726:        } else
                    727:                offset = 0;
1.93      schwarze  728:
1.95      schwarze  729:        mparse_buf_r(curp, blk, offset, 1);
1.28      joerg     730:
1.111     schwarze  731:        if (--recursion_depth == 0)
1.28      joerg     732:                mparse_end(curp);
                    733:
1.85      schwarze  734:        curp->primary = svprimary;
1.28      joerg     735:        curp->file = svfile;
                    736: }
                    737:
                    738: enum mandoclevel
1.104     schwarze  739: mparse_readmem(struct mparse *curp, void *buf, size_t len,
1.28      joerg     740:                const char *file)
                    741: {
                    742:        struct buf blk;
                    743:
1.104     schwarze  744:        blk.buf = buf;
1.28      joerg     745:        blk.sz = len;
                    746:
1.36      schwarze  747:        mparse_parse_buffer(curp, blk, file);
1.142     schwarze  748:        return curp->file_status;
1.28      joerg     749: }
                    750:
1.98      schwarze  751: /*
                    752:  * Read the whole file into memory and call the parsers.
                    753:  * Called recursively when an .so request is encountered.
                    754:  */
1.36      schwarze  755: enum mandoclevel
                    756: mparse_readfd(struct mparse *curp, int fd, const char *file)
1.1       kristaps  757: {
1.28      joerg     758:        struct buf       blk;
                    759:        int              with_mmap;
1.93      schwarze  760:        int              save_filenc;
1.1       kristaps  761:
1.91      schwarze  762:        if (read_whole_file(curp, file, fd, &blk, &with_mmap)) {
1.93      schwarze  763:                save_filenc = curp->filenc;
                    764:                curp->filenc = curp->options &
                    765:                    (MPARSE_UTF8 | MPARSE_LATIN1);
1.91      schwarze  766:                mparse_parse_buffer(curp, blk, file);
1.93      schwarze  767:                curp->filenc = save_filenc;
1.91      schwarze  768:                if (with_mmap)
                    769:                        munmap(blk.buf, blk.sz);
                    770:                else
                    771:                        free(blk.buf);
                    772:        }
1.142     schwarze  773:        return curp->file_status;
1.82      schwarze  774: }
                    775:
1.148     schwarze  776: int
                    777: mparse_open(struct mparse *curp, const char *file)
1.82      schwarze  778: {
                    779:        char             *cp;
1.148     schwarze  780:        int               fd;
1.82      schwarze  781:
                    782:        curp->file = file;
1.140     schwarze  783:        cp = strrchr(file, '.');
                    784:        curp->gzip = (cp != NULL && ! strcmp(cp + 1, "gz"));
1.98      schwarze  785:
1.140     schwarze  786:        /* First try to use the filename as it is. */
1.98      schwarze  787:
1.148     schwarze  788:        if ((fd = open(file, O_RDONLY)) != -1)
                    789:                return fd;
1.98      schwarze  790:
1.140     schwarze  791:        /*
                    792:         * If that doesn't work and the filename doesn't
                    793:         * already  end in .gz, try appending .gz.
                    794:         */
1.98      schwarze  795:
1.140     schwarze  796:        if ( ! curp->gzip) {
1.98      schwarze  797:                mandoc_asprintf(&cp, "%s.gz", file);
1.149     schwarze  798:                fd = open(cp, O_RDONLY);
1.108     schwarze  799:                free(cp);
1.148     schwarze  800:                if (fd != -1) {
1.140     schwarze  801:                        curp->gzip = 1;
1.148     schwarze  802:                        return fd;
1.82      schwarze  803:                }
                    804:        }
                    805:
1.140     schwarze  806:        /* Neither worked, give up. */
1.97      schwarze  807:
1.140     schwarze  808:        mandoc_msg(MANDOCERR_FILE, curp, 0, 0, strerror(errno));
1.148     schwarze  809:        return -1;
1.1       kristaps  810: }
                    811:
                    812: struct mparse *
1.94      schwarze  813: mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,
1.144     schwarze  814:     const char *defos)
1.1       kristaps  815: {
                    816:        struct mparse   *curp;
1.10      kristaps  817:
1.1       kristaps  818:        curp = mandoc_calloc(1, sizeof(struct mparse));
                    819:
1.44      schwarze  820:        curp->options = options;
1.3       kristaps  821:        curp->wlevel = wlevel;
1.1       kristaps  822:        curp->mmsg = mmsg;
1.29      schwarze  823:        curp->defos = defos;
1.1       kristaps  824:
1.144     schwarze  825:        curp->roff = roff_alloc(curp, options);
1.137     schwarze  826:        curp->man = roff_man_alloc( curp->roff, curp, curp->defos,
                    827:                curp->options & MPARSE_QUICK ? 1 : 0);
1.136     schwarze  828:        if (curp->options & MPARSE_MDOC) {
1.137     schwarze  829:                curp->man->macroset = MACROSET_MDOC;
1.164     schwarze  830:                if (curp->man->mdocmac == NULL)
                    831:                        curp->man->mdocmac = roffhash_alloc(MDOC_Dd, MDOC_MAX);
1.137     schwarze  832:        } else if (curp->options & MPARSE_MAN) {
                    833:                curp->man->macroset = MACROSET_MAN;
1.164     schwarze  834:                if (curp->man->manmac == NULL)
                    835:                        curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX);
1.136     schwarze  836:        }
1.138     schwarze  837:        curp->man->first->tok = TOKEN_NONE;
1.142     schwarze  838:        return curp;
1.1       kristaps  839: }
                    840:
                    841: void
                    842: mparse_reset(struct mparse *curp)
                    843: {
                    844:        roff_reset(curp->roff);
1.150     schwarze  845:        roff_man_reset(curp->man);
1.160     schwarze  846:
                    847:        free(curp->sodest);
                    848:        curp->sodest = NULL;
                    849:
1.24      kristaps  850:        if (curp->secondary)
                    851:                curp->secondary->sz = 0;
1.1       kristaps  852:
                    853:        curp->file_status = MANDOCLEVEL_OK;
1.159     schwarze  854:        curp->gzip = 0;
1.1       kristaps  855: }
                    856:
                    857: void
                    858: mparse_free(struct mparse *curp)
                    859: {
                    860:
1.164     schwarze  861:        roffhash_free(curp->man->mdocmac);
                    862:        roffhash_free(curp->man->manmac);
1.137     schwarze  863:        roff_man_free(curp->man);
1.160     schwarze  864:        roff_free(curp->roff);
1.24      kristaps  865:        if (curp->secondary)
                    866:                free(curp->secondary->buf);
1.1       kristaps  867:
1.24      kristaps  868:        free(curp->secondary);
1.45      schwarze  869:        free(curp->sodest);
1.1       kristaps  870:        free(curp);
                    871: }
                    872:
                    873: void
1.135     schwarze  874: mparse_result(struct mparse *curp, struct roff_man **man,
                    875:        char **sodest)
1.1       kristaps  876: {
                    877:
1.45      schwarze  878:        if (sodest && NULL != (*sodest = curp->sodest)) {
                    879:                *man = NULL;
                    880:                return;
                    881:        }
1.9       kristaps  882:        if (man)
                    883:                *man = curp->man;
1.157     schwarze  884: }
                    885:
                    886: void
                    887: mparse_updaterc(struct mparse *curp, enum mandoclevel *rc)
                    888: {
                    889:        if (curp->file_status > *rc)
                    890:                *rc = curp->file_status;
1.3       kristaps  891: }
                    892:
                    893: void
                    894: mandoc_vmsg(enum mandocerr t, struct mparse *m,
                    895:                int ln, int pos, const char *fmt, ...)
                    896: {
                    897:        char             buf[256];
                    898:        va_list          ap;
                    899:
                    900:        va_start(ap, fmt);
1.48      schwarze  901:        (void)vsnprintf(buf, sizeof(buf), fmt, ap);
1.3       kristaps  902:        va_end(ap);
                    903:
                    904:        mandoc_msg(t, m, ln, pos, buf);
                    905: }
                    906:
                    907: void
1.47      schwarze  908: mandoc_msg(enum mandocerr er, struct mparse *m,
1.3       kristaps  909:                int ln, int col, const char *msg)
                    910: {
                    911:        enum mandoclevel level;
                    912:
1.112     schwarze  913:        level = MANDOCLEVEL_UNSUPP;
1.3       kristaps  914:        while (er < mandoclimits[level])
                    915:                level--;
                    916:
1.108     schwarze  917:        if (level < m->wlevel && er != MANDOCERR_FILE)
1.3       kristaps  918:                return;
                    919:
1.8       kristaps  920:        if (m->mmsg)
                    921:                (*m->mmsg)(er, level, m->file, ln, col, msg);
1.3       kristaps  922:
                    923:        if (m->file_status < level)
                    924:                m->file_status = level;
1.7       kristaps  925: }
                    926:
                    927: const char *
                    928: mparse_strerror(enum mandocerr er)
                    929: {
                    930:
1.142     schwarze  931:        return mandocerrs[er];
1.7       kristaps  932: }
                    933:
                    934: const char *
                    935: mparse_strlevel(enum mandoclevel lvl)
                    936: {
1.142     schwarze  937:        return mandoclevels[lvl];
1.24      kristaps  938: }
                    939:
                    940: void
                    941: mparse_keep(struct mparse *p)
                    942: {
                    943:
                    944:        assert(NULL == p->secondary);
                    945:        p->secondary = mandoc_calloc(1, sizeof(struct buf));
                    946: }
                    947:
                    948: const char *
                    949: mparse_getkeep(const struct mparse *p)
                    950: {
                    951:
                    952:        assert(p->secondary);
1.142     schwarze  953:        return p->secondary->sz ? p->secondary->buf : NULL;
1.1       kristaps  954: }

CVSweb