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

Annotation of mandoc/read.c, Revision 1.159

1.159   ! schwarze    1: /*     $Id: read.c,v 1.158 2017/01/28 23:30:08 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>
1.28      joerg      33: #include <stdint.h>
1.1       kristaps   34: #include <stdio.h>
                     35: #include <stdlib.h>
                     36: #include <string.h>
                     37: #include <unistd.h>
1.140     schwarze   38: #include <zlib.h>
1.1       kristaps   39:
1.133     schwarze   40: #include "mandoc_aux.h"
1.1       kristaps   41: #include "mandoc.h"
1.133     schwarze   42: #include "roff.h"
1.1       kristaps   43: #include "mdoc.h"
                     44: #include "man.h"
1.133     schwarze   45: #include "libmandoc.h"
1.139     schwarze   46: #include "roff_int.h"
1.1       kristaps   47:
                     48: #define        REPARSE_LIMIT   1000
                     49:
                     50: struct mparse {
1.134     schwarze   51:        struct roff_man  *man; /* man parser */
1.1       kristaps   52:        struct roff      *roff; /* roff parser (!NULL) */
1.45      schwarze   53:        char             *sodest; /* filename pointed to by .so */
1.83      schwarze   54:        const char       *file; /* filename of current input file */
                     55:        struct buf       *primary; /* buffer currently being parsed */
                     56:        struct buf       *secondary; /* preprocessed copy of input */
                     57:        const char       *defos; /* default operating system */
                     58:        mandocmsg         mmsg; /* warning/error message handler */
                     59:        enum mandoclevel  file_status; /* status of current parse */
                     60:        enum mandoclevel  wlevel; /* ignore messages below this */
                     61:        int               options; /* parser options */
1.140     schwarze   62:        int               gzip; /* current input file is gzipped */
1.93      schwarze   63:        int               filenc; /* encoding of the current file */
1.1       kristaps   64:        int               reparse_count; /* finite interp. stack */
1.83      schwarze   65:        int               line; /* line number in the file */
1.1       kristaps   66: };
                     67:
1.84      schwarze   68: static void      choose_parser(struct mparse *);
1.1       kristaps   69: static void      resize_buf(struct buf *, size_t);
1.95      schwarze   70: static void      mparse_buf_r(struct mparse *, struct buf, size_t, int);
1.40      schwarze   71: static int       read_whole_file(struct mparse *, const char *, int,
                     72:                                struct buf *, int *);
1.1       kristaps   73: static void      mparse_end(struct mparse *);
1.37      schwarze   74: static void      mparse_parse_buffer(struct mparse *, struct buf,
                     75:                        const char *);
1.1       kristaps   76:
1.3       kristaps   77: static const enum mandocerr    mandoclimits[MANDOCLEVEL_MAX] = {
                     78:        MANDOCERR_OK,
                     79:        MANDOCERR_WARNING,
                     80:        MANDOCERR_WARNING,
                     81:        MANDOCERR_ERROR,
1.112     schwarze   82:        MANDOCERR_UNSUPP,
1.3       kristaps   83:        MANDOCERR_MAX,
                     84:        MANDOCERR_MAX
                     85: };
                     86:
1.7       kristaps   87: static const char * const      mandocerrs[MANDOCERR_MAX] = {
                     88:        "ok",
                     89:
                     90:        "generic warning",
                     91:
                     92:        /* related to the prologue */
1.79      schwarze   93:        "missing manual title, using UNTITLED",
                     94:        "missing manual title, using \"\"",
1.54      schwarze   95:        "lower case character in document title",
1.79      schwarze   96:        "missing manual section, using \"\"",
1.7       kristaps   97:        "unknown manual section",
1.54      schwarze   98:        "missing date, using today's date",
1.7       kristaps   99:        "cannot parse date, using it verbatim",
1.79      schwarze  100:        "missing Os macro, using \"\"",
                    101:        "duplicate prologue macro",
                    102:        "late prologue macro",
                    103:        "skipping late title macro",
1.7       kristaps  104:        "prologue macros out of order",
                    105:
                    106:        /* related to document structure */
                    107:        ".so is fragile, better use ln(1)",
1.50      schwarze  108:        "no document body",
1.54      schwarze  109:        "content before first section header",
                    110:        "first section is not \"NAME\"",
1.156     schwarze  111:        "NAME section without Nm before Nd",
1.128     schwarze  112:        "NAME section without description",
                    113:        "description not at the end of NAME",
                    114:        "bad NAME section content",
1.156     schwarze  115:        "missing comma before name",
1.121     schwarze  116:        "missing description line, using \"\"",
1.7       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) {
1.136     schwarze  296:                mdoc_hash_init();
1.137     schwarze  297:                curp->man->macroset = MACROSET_MDOC;
1.138     schwarze  298:                curp->man->first->tok = TOKEN_NONE;
1.137     schwarze  299:        } else {
                    300:                man_hash_init();
                    301:                curp->man->macroset = MACROSET_MAN;
1.138     schwarze  302:                curp->man->first->tok = TOKEN_NONE;
1.47      schwarze  303:        }
1.1       kristaps  304: }
                    305:
                    306: /*
1.95      schwarze  307:  * Main parse routine for a buffer.
                    308:  * It assumes encoding and line numbering are already set up.
                    309:  * It can recurse directly (for invocations of user-defined
                    310:  * macros, inline equations, and input line traps)
                    311:  * and indirectly (for .so file inclusion).
1.1       kristaps  312:  */
                    313: static void
1.95      schwarze  314: mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
1.1       kristaps  315: {
                    316:        const struct tbl_span   *span;
                    317:        struct buf       ln;
1.113     schwarze  318:        const char      *save_file;
1.110     schwarze  319:        char            *cp;
1.95      schwarze  320:        size_t           pos; /* byte number in the ln buffer */
1.154     schwarze  321:        size_t           j;  /* auxiliary byte number in the blk buffer */
1.1       kristaps  322:        enum rofferr     rr;
1.100     schwarze  323:        int              of;
1.1       kristaps  324:        int              lnn; /* line number in the real file */
1.110     schwarze  325:        int              fd;
1.1       kristaps  326:        unsigned char    c;
                    327:
1.95      schwarze  328:        memset(&ln, 0, sizeof(ln));
1.1       kristaps  329:
1.47      schwarze  330:        lnn = curp->line;
                    331:        pos = 0;
1.1       kristaps  332:
1.95      schwarze  333:        while (i < blk.sz) {
1.1       kristaps  334:                if (0 == pos && '\0' == blk.buf[i])
                    335:                        break;
                    336:
                    337:                if (start) {
                    338:                        curp->line = lnn;
                    339:                        curp->reparse_count = 0;
1.93      schwarze  340:
                    341:                        if (lnn < 3 &&
                    342:                            curp->filenc & MPARSE_UTF8 &&
1.95      schwarze  343:                            curp->filenc & MPARSE_LATIN1)
                    344:                                curp->filenc = preconv_cue(&blk, i);
1.1       kristaps  345:                }
                    346:
1.95      schwarze  347:                while (i < blk.sz && (start || blk.buf[i] != '\0')) {
1.1       kristaps  348:
                    349:                        /*
                    350:                         * When finding an unescaped newline character,
                    351:                         * leave the character loop to process the line.
                    352:                         * Skip a preceding carriage return, if any.
                    353:                         */
                    354:
1.95      schwarze  355:                        if ('\r' == blk.buf[i] && i + 1 < blk.sz &&
1.1       kristaps  356:                            '\n' == blk.buf[i + 1])
                    357:                                ++i;
                    358:                        if ('\n' == blk.buf[i]) {
                    359:                                ++i;
                    360:                                ++lnn;
                    361:                                break;
                    362:                        }
                    363:
1.35      schwarze  364:                        /*
1.93      schwarze  365:                         * Make sure we have space for the worst
                    366:                         * case of 11 bytes: "\\[u10ffff]\0"
1.35      schwarze  367:                         */
                    368:
1.95      schwarze  369:                        if (pos + 11 > ln.sz)
1.35      schwarze  370:                                resize_buf(&ln, 256);
                    371:
1.47      schwarze  372:                        /*
1.93      schwarze  373:                         * Encode 8-bit input.
1.1       kristaps  374:                         */
                    375:
1.93      schwarze  376:                        c = blk.buf[i];
                    377:                        if (c & 0x80) {
1.95      schwarze  378:                                if ( ! (curp->filenc && preconv_encode(
                    379:                                    &blk, &i, &ln, &pos, &curp->filenc))) {
1.114     schwarze  380:                                        mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
                    381:                                            curp->line, pos, "0x%x", c);
1.93      schwarze  382:                                        ln.buf[pos++] = '?';
                    383:                                        i++;
                    384:                                }
                    385:                                continue;
                    386:                        }
                    387:
                    388:                        /*
                    389:                         * Exclude control characters.
                    390:                         */
1.1       kristaps  391:
1.93      schwarze  392:                        if (c == 0x7f || (c < 0x20 && c != 0x09)) {
1.114     schwarze  393:                                mandoc_vmsg(c == 0x00 || c == 0x04 ||
                    394:                                    c > 0x0a ? MANDOCERR_CHAR_BAD :
                    395:                                    MANDOCERR_CHAR_UNSUPP,
                    396:                                    curp, curp->line, pos, "0x%x", c);
1.1       kristaps  397:                                i++;
1.127     schwarze  398:                                if (c != '\r')
                    399:                                        ln.buf[pos++] = '?';
1.1       kristaps  400:                                continue;
                    401:                        }
                    402:
                    403:                        /* Trailing backslash = a plain char. */
                    404:
1.95      schwarze  405:                        if (blk.buf[i] != '\\' || i + 1 == blk.sz) {
1.1       kristaps  406:                                ln.buf[pos++] = blk.buf[i++];
                    407:                                continue;
                    408:                        }
                    409:
                    410:                        /*
                    411:                         * Found escape and at least one other character.
                    412:                         * When it's a newline character, skip it.
                    413:                         * When there is a carriage return in between,
                    414:                         * skip that one as well.
                    415:                         */
                    416:
1.95      schwarze  417:                        if ('\r' == blk.buf[i + 1] && i + 2 < blk.sz &&
1.1       kristaps  418:                            '\n' == blk.buf[i + 2])
                    419:                                ++i;
                    420:                        if ('\n' == blk.buf[i + 1]) {
                    421:                                i += 2;
                    422:                                ++lnn;
                    423:                                continue;
                    424:                        }
                    425:
1.13      kristaps  426:                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {
1.154     schwarze  427:                                j = i;
1.1       kristaps  428:                                i += 2;
                    429:                                /* Comment, skip to end of line */
1.95      schwarze  430:                                for (; i < blk.sz; ++i) {
1.153     schwarze  431:                                        if (blk.buf[i] != '\n')
                    432:                                                continue;
                    433:                                        if (blk.buf[i - 1] == ' ' ||
                    434:                                            blk.buf[i - 1] == '\t')
                    435:                                                mandoc_msg(
                    436:                                                    MANDOCERR_SPACE_EOL,
                    437:                                                    curp, curp->line,
1.154     schwarze  438:                                                    pos + i-1 - j, NULL);
1.153     schwarze  439:                                        ++i;
                    440:                                        ++lnn;
                    441:                                        break;
1.1       kristaps  442:                                }
                    443:
                    444:                                /* Backout trailing whitespaces */
                    445:                                for (; pos > 0; --pos) {
                    446:                                        if (ln.buf[pos - 1] != ' ')
                    447:                                                break;
                    448:                                        if (pos > 2 && ln.buf[pos - 2] == '\\')
                    449:                                                break;
                    450:                                }
                    451:                                break;
                    452:                        }
                    453:
1.35      schwarze  454:                        /* Catch escaped bogus characters. */
                    455:
                    456:                        c = (unsigned char) blk.buf[i+1];
                    457:
1.47      schwarze  458:                        if ( ! (isascii(c) &&
                    459:                            (isgraph(c) || isblank(c)))) {
1.114     schwarze  460:                                mandoc_vmsg(MANDOCERR_CHAR_BAD, curp,
1.78      schwarze  461:                                    curp->line, pos, "0x%x", c);
1.35      schwarze  462:                                i += 2;
                    463:                                ln.buf[pos++] = '?';
                    464:                                continue;
                    465:                        }
                    466:
1.1       kristaps  467:                        /* Some other escape sequence, copy & cont. */
                    468:
                    469:                        ln.buf[pos++] = blk.buf[i++];
                    470:                        ln.buf[pos++] = blk.buf[i++];
                    471:                }
                    472:
1.95      schwarze  473:                if (pos >= ln.sz)
1.1       kristaps  474:                        resize_buf(&ln, 256);
                    475:
                    476:                ln.buf[pos] = '\0';
                    477:
                    478:                /*
                    479:                 * A significant amount of complexity is contained by
                    480:                 * the roff preprocessor.  It's line-oriented but can be
                    481:                 * expressed on one line, so we need at times to
                    482:                 * readjust our starting point and re-run it.  The roff
                    483:                 * preprocessor can also readjust the buffers with new
                    484:                 * data, so we pass them in wholesale.
                    485:                 */
                    486:
                    487:                of = 0;
                    488:
1.24      kristaps  489:                /*
                    490:                 * Maintain a lookaside buffer of all parsed lines.  We
                    491:                 * only do this if mparse_keep() has been invoked (the
                    492:                 * buffer may be accessed with mparse_getkeep()).
                    493:                 */
                    494:
                    495:                if (curp->secondary) {
1.47      schwarze  496:                        curp->secondary->buf = mandoc_realloc(
                    497:                            curp->secondary->buf,
                    498:                            curp->secondary->sz + pos + 2);
                    499:                        memcpy(curp->secondary->buf +
                    500:                            curp->secondary->sz,
                    501:                            ln.buf, pos);
1.24      kristaps  502:                        curp->secondary->sz += pos;
                    503:                        curp->secondary->buf
                    504:                                [curp->secondary->sz] = '\n';
                    505:                        curp->secondary->sz++;
                    506:                        curp->secondary->buf
                    507:                                [curp->secondary->sz] = '\0';
                    508:                }
1.1       kristaps  509: rerun:
1.96      schwarze  510:                rr = roff_parseln(curp->roff, curp->line, &ln, &of);
1.1       kristaps  511:
                    512:                switch (rr) {
1.47      schwarze  513:                case ROFF_REPARSE:
1.1       kristaps  514:                        if (REPARSE_LIMIT >= ++curp->reparse_count)
1.95      schwarze  515:                                mparse_buf_r(curp, ln, of, 0);
1.1       kristaps  516:                        else
1.3       kristaps  517:                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
1.47      schwarze  518:                                    curp->line, pos, NULL);
1.1       kristaps  519:                        pos = 0;
                    520:                        continue;
1.47      schwarze  521:                case ROFF_APPEND:
1.95      schwarze  522:                        pos = strlen(ln.buf);
1.1       kristaps  523:                        continue;
1.47      schwarze  524:                case ROFF_RERUN:
1.1       kristaps  525:                        goto rerun;
1.47      schwarze  526:                case ROFF_IGN:
1.1       kristaps  527:                        pos = 0;
                    528:                        continue;
1.47      schwarze  529:                case ROFF_SO:
1.95      schwarze  530:                        if ( ! (curp->options & MPARSE_SO) &&
                    531:                            (i >= blk.sz || blk.buf[i] == '\0')) {
1.45      schwarze  532:                                curp->sodest = mandoc_strdup(ln.buf + of);
                    533:                                free(ln.buf);
                    534:                                return;
                    535:                        }
1.24      kristaps  536:                        /*
                    537:                         * We remove `so' clauses from our lookaside
                    538:                         * buffer because we're going to descend into
                    539:                         * the file recursively.
                    540:                         */
1.47      schwarze  541:                        if (curp->secondary)
1.25      kristaps  542:                                curp->secondary->sz -= pos + 1;
1.113     schwarze  543:                        save_file = curp->file;
1.148     schwarze  544:                        if ((fd = mparse_open(curp, ln.buf + of)) != -1) {
1.110     schwarze  545:                                mparse_readfd(curp, fd, ln.buf + of);
1.147     schwarze  546:                                close(fd);
1.113     schwarze  547:                                curp->file = save_file;
                    548:                        } else {
                    549:                                curp->file = save_file;
1.52      schwarze  550:                                mandoc_vmsg(MANDOCERR_SO_FAIL,
                    551:                                    curp, curp->line, pos,
                    552:                                    ".so %s", ln.buf + of);
1.110     schwarze  553:                                ln.sz = mandoc_asprintf(&cp,
                    554:                                    ".sp\nSee the file %s.\n.sp",
                    555:                                    ln.buf + of);
                    556:                                free(ln.buf);
                    557:                                ln.buf = cp;
                    558:                                of = 0;
                    559:                                mparse_buf_r(curp, ln, of, 0);
1.52      schwarze  560:                        }
1.1       kristaps  561:                        pos = 0;
                    562:                        continue;
                    563:                default:
                    564:                        break;
                    565:                }
                    566:
1.150     schwarze  567:                if (curp->man->macroset == MACROSET_NONE)
1.84      schwarze  568:                        choose_parser(curp);
1.1       kristaps  569:
1.47      schwarze  570:                /*
1.84      schwarze  571:                 * Lastly, push down into the parsers themselves.
1.1       kristaps  572:                 * If libroff returns ROFF_TBL, then add it to the
                    573:                 * currently open parse.  Since we only get here if
                    574:                 * there does exist data (see tbl_data.c), we're
                    575:                 * guaranteed that something's been allocated.
                    576:                 * Do the same for ROFF_EQN.
                    577:                 */
                    578:
1.139     schwarze  579:                if (rr == ROFF_TBL)
1.100     schwarze  580:                        while ((span = roff_span(curp->roff)) != NULL)
1.139     schwarze  581:                                roff_addtbl(curp->man, span);
                    582:                else if (rr == ROFF_EQN)
                    583:                        roff_addeqn(curp->man, roff_eqn(curp->roff));
                    584:                else if ((curp->man->macroset == MACROSET_MDOC ?
1.135     schwarze  585:                    mdoc_parseln(curp->man, curp->line, ln.buf, of) :
1.100     schwarze  586:                    man_parseln(curp->man, curp->line, ln.buf, of)) == 2)
                    587:                                break;
1.1       kristaps  588:
                    589:                /* Temporary buffers typically are not full. */
                    590:
                    591:                if (0 == start && '\0' == blk.buf[i])
                    592:                        break;
                    593:
                    594:                /* Start the next input line. */
                    595:
                    596:                pos = 0;
                    597:        }
                    598:
                    599:        free(ln.buf);
                    600: }
                    601:
                    602: static int
1.40      schwarze  603: read_whole_file(struct mparse *curp, const char *file, int fd,
                    604:                struct buf *fb, int *with_mmap)
1.1       kristaps  605: {
1.140     schwarze  606:        gzFile           gz;
1.1       kristaps  607:        size_t           off;
                    608:        ssize_t          ssz;
                    609:
1.15      kristaps  610:        struct stat      st;
1.143     schwarze  611:
                    612:        if (fstat(fd, &st) == -1)
                    613:                err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  614:
                    615:        /*
                    616:         * If we're a regular file, try just reading in the whole entry
                    617:         * via mmap().  This is faster than reading it into blocks, and
                    618:         * since each file is only a few bytes to begin with, I'm not
                    619:         * concerned that this is going to tank any machines.
                    620:         */
                    621:
1.140     schwarze  622:        if (curp->gzip == 0 && S_ISREG(st.st_mode)) {
1.131     schwarze  623:                if (st.st_size > 0x7fffffff) {
1.111     schwarze  624:                        mandoc_msg(MANDOCERR_TOOLARGE, curp, 0, 0, NULL);
1.142     schwarze  625:                        return 0;
1.1       kristaps  626:                }
                    627:                *with_mmap = 1;
                    628:                fb->sz = (size_t)st.st_size;
1.37      schwarze  629:                fb->buf = mmap(NULL, fb->sz, PROT_READ, MAP_SHARED, fd, 0);
1.1       kristaps  630:                if (fb->buf != MAP_FAILED)
1.142     schwarze  631:                        return 1;
1.1       kristaps  632:        }
                    633:
1.140     schwarze  634:        if (curp->gzip) {
1.143     schwarze  635:                if ((gz = gzdopen(fd, "rb")) == NULL)
                    636:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.140     schwarze  637:        } else
                    638:                gz = NULL;
                    639:
1.1       kristaps  640:        /*
                    641:         * If this isn't a regular file (like, say, stdin), then we must
                    642:         * go the old way and just read things in bit by bit.
                    643:         */
                    644:
                    645:        *with_mmap = 0;
                    646:        off = 0;
                    647:        fb->sz = 0;
                    648:        fb->buf = NULL;
                    649:        for (;;) {
                    650:                if (off == fb->sz) {
                    651:                        if (fb->sz == (1U << 31)) {
1.111     schwarze  652:                                mandoc_msg(MANDOCERR_TOOLARGE, curp,
                    653:                                    0, 0, NULL);
1.1       kristaps  654:                                break;
                    655:                        }
                    656:                        resize_buf(fb, 65536);
                    657:                }
1.140     schwarze  658:                ssz = curp->gzip ?
                    659:                    gzread(gz, fb->buf + (int)off, fb->sz - off) :
                    660:                    read(fd, fb->buf + (int)off, fb->sz - off);
1.1       kristaps  661:                if (ssz == 0) {
                    662:                        fb->sz = off;
1.142     schwarze  663:                        return 1;
1.1       kristaps  664:                }
1.143     schwarze  665:                if (ssz == -1)
                    666:                        err((int)MANDOCLEVEL_SYSERR, "%s", file);
1.1       kristaps  667:                off += (size_t)ssz;
                    668:        }
                    669:
                    670:        free(fb->buf);
                    671:        fb->buf = NULL;
1.142     schwarze  672:        return 0;
1.1       kristaps  673: }
                    674:
                    675: static void
                    676: mparse_end(struct mparse *curp)
                    677: {
1.135     schwarze  678:        if (curp->man->macroset == MACROSET_NONE)
                    679:                curp->man->macroset = MACROSET_MAN;
                    680:        if (curp->man->macroset == MACROSET_MDOC)
                    681:                mdoc_endparse(curp->man);
                    682:        else
1.111     schwarze  683:                man_endparse(curp->man);
1.1       kristaps  684:        roff_endparse(curp->roff);
                    685: }
                    686:
                    687: static void
1.36      schwarze  688: mparse_parse_buffer(struct mparse *curp, struct buf blk, const char *file)
1.28      joerg     689: {
1.85      schwarze  690:        struct buf      *svprimary;
1.28      joerg     691:        const char      *svfile;
1.95      schwarze  692:        size_t           offset;
1.36      schwarze  693:        static int       recursion_depth;
                    694:
                    695:        if (64 < recursion_depth) {
                    696:                mandoc_msg(MANDOCERR_ROFFLOOP, curp, curp->line, 0, NULL);
                    697:                return;
                    698:        }
1.28      joerg     699:
                    700:        /* Line number is per-file. */
                    701:        svfile = curp->file;
                    702:        curp->file = file;
1.85      schwarze  703:        svprimary = curp->primary;
1.83      schwarze  704:        curp->primary = &blk;
1.28      joerg     705:        curp->line = 1;
1.36      schwarze  706:        recursion_depth++;
1.28      joerg     707:
1.93      schwarze  708:        /* Skip an UTF-8 byte order mark. */
                    709:        if (curp->filenc & MPARSE_UTF8 && blk.sz > 2 &&
                    710:            (unsigned char)blk.buf[0] == 0xef &&
                    711:            (unsigned char)blk.buf[1] == 0xbb &&
                    712:            (unsigned char)blk.buf[2] == 0xbf) {
1.95      schwarze  713:                offset = 3;
1.93      schwarze  714:                curp->filenc &= ~MPARSE_LATIN1;
1.95      schwarze  715:        } else
                    716:                offset = 0;
1.93      schwarze  717:
1.95      schwarze  718:        mparse_buf_r(curp, blk, offset, 1);
1.28      joerg     719:
1.111     schwarze  720:        if (--recursion_depth == 0)
1.28      joerg     721:                mparse_end(curp);
                    722:
1.85      schwarze  723:        curp->primary = svprimary;
1.28      joerg     724:        curp->file = svfile;
                    725: }
                    726:
                    727: enum mandoclevel
1.104     schwarze  728: mparse_readmem(struct mparse *curp, void *buf, size_t len,
1.28      joerg     729:                const char *file)
                    730: {
                    731:        struct buf blk;
                    732:
1.104     schwarze  733:        blk.buf = buf;
1.28      joerg     734:        blk.sz = len;
                    735:
1.36      schwarze  736:        mparse_parse_buffer(curp, blk, file);
1.142     schwarze  737:        return curp->file_status;
1.28      joerg     738: }
                    739:
1.98      schwarze  740: /*
                    741:  * Read the whole file into memory and call the parsers.
                    742:  * Called recursively when an .so request is encountered.
                    743:  */
1.36      schwarze  744: enum mandoclevel
                    745: mparse_readfd(struct mparse *curp, int fd, const char *file)
1.1       kristaps  746: {
1.28      joerg     747:        struct buf       blk;
                    748:        int              with_mmap;
1.93      schwarze  749:        int              save_filenc;
1.1       kristaps  750:
1.91      schwarze  751:        if (read_whole_file(curp, file, fd, &blk, &with_mmap)) {
1.93      schwarze  752:                save_filenc = curp->filenc;
                    753:                curp->filenc = curp->options &
                    754:                    (MPARSE_UTF8 | MPARSE_LATIN1);
1.91      schwarze  755:                mparse_parse_buffer(curp, blk, file);
1.93      schwarze  756:                curp->filenc = save_filenc;
1.91      schwarze  757:                if (with_mmap)
                    758:                        munmap(blk.buf, blk.sz);
                    759:                else
                    760:                        free(blk.buf);
                    761:        }
1.142     schwarze  762:        return curp->file_status;
1.82      schwarze  763: }
                    764:
1.148     schwarze  765: int
                    766: mparse_open(struct mparse *curp, const char *file)
1.82      schwarze  767: {
                    768:        char             *cp;
1.148     schwarze  769:        int               fd;
1.82      schwarze  770:
                    771:        curp->file = file;
1.140     schwarze  772:        cp = strrchr(file, '.');
                    773:        curp->gzip = (cp != NULL && ! strcmp(cp + 1, "gz"));
1.98      schwarze  774:
1.140     schwarze  775:        /* First try to use the filename as it is. */
1.98      schwarze  776:
1.148     schwarze  777:        if ((fd = open(file, O_RDONLY)) != -1)
                    778:                return fd;
1.98      schwarze  779:
1.140     schwarze  780:        /*
                    781:         * If that doesn't work and the filename doesn't
                    782:         * already  end in .gz, try appending .gz.
                    783:         */
1.98      schwarze  784:
1.140     schwarze  785:        if ( ! curp->gzip) {
1.98      schwarze  786:                mandoc_asprintf(&cp, "%s.gz", file);
1.149     schwarze  787:                fd = open(cp, O_RDONLY);
1.108     schwarze  788:                free(cp);
1.148     schwarze  789:                if (fd != -1) {
1.140     schwarze  790:                        curp->gzip = 1;
1.148     schwarze  791:                        return fd;
1.82      schwarze  792:                }
                    793:        }
                    794:
1.140     schwarze  795:        /* Neither worked, give up. */
1.97      schwarze  796:
1.140     schwarze  797:        mandoc_msg(MANDOCERR_FILE, curp, 0, 0, strerror(errno));
1.148     schwarze  798:        return -1;
1.1       kristaps  799: }
                    800:
                    801: struct mparse *
1.94      schwarze  802: mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,
1.144     schwarze  803:     const char *defos)
1.1       kristaps  804: {
                    805:        struct mparse   *curp;
1.10      kristaps  806:
1.1       kristaps  807:        curp = mandoc_calloc(1, sizeof(struct mparse));
                    808:
1.44      schwarze  809:        curp->options = options;
1.3       kristaps  810:        curp->wlevel = wlevel;
1.1       kristaps  811:        curp->mmsg = mmsg;
1.29      schwarze  812:        curp->defos = defos;
1.1       kristaps  813:
1.144     schwarze  814:        curp->roff = roff_alloc(curp, options);
1.137     schwarze  815:        curp->man = roff_man_alloc( curp->roff, curp, curp->defos,
                    816:                curp->options & MPARSE_QUICK ? 1 : 0);
1.136     schwarze  817:        if (curp->options & MPARSE_MDOC) {
                    818:                mdoc_hash_init();
1.137     schwarze  819:                curp->man->macroset = MACROSET_MDOC;
                    820:        } else if (curp->options & MPARSE_MAN) {
1.136     schwarze  821:                man_hash_init();
1.137     schwarze  822:                curp->man->macroset = MACROSET_MAN;
1.136     schwarze  823:        }
1.138     schwarze  824:        curp->man->first->tok = TOKEN_NONE;
1.142     schwarze  825:        return curp;
1.1       kristaps  826: }
                    827:
                    828: void
                    829: mparse_reset(struct mparse *curp)
                    830: {
                    831:        roff_reset(curp->roff);
1.150     schwarze  832:        roff_man_reset(curp->man);
1.24      kristaps  833:        if (curp->secondary)
                    834:                curp->secondary->sz = 0;
1.1       kristaps  835:
                    836:        curp->file_status = MANDOCLEVEL_OK;
1.45      schwarze  837:
                    838:        free(curp->sodest);
                    839:        curp->sodest = NULL;
1.159   ! schwarze  840:        curp->gzip = 0;
1.1       kristaps  841: }
                    842:
                    843: void
                    844: mparse_free(struct mparse *curp)
                    845: {
                    846:
1.137     schwarze  847:        roff_man_free(curp->man);
1.1       kristaps  848:        if (curp->roff)
                    849:                roff_free(curp->roff);
1.24      kristaps  850:        if (curp->secondary)
                    851:                free(curp->secondary->buf);
1.1       kristaps  852:
1.24      kristaps  853:        free(curp->secondary);
1.45      schwarze  854:        free(curp->sodest);
1.1       kristaps  855:        free(curp);
                    856: }
                    857:
                    858: void
1.135     schwarze  859: mparse_result(struct mparse *curp, struct roff_man **man,
                    860:        char **sodest)
1.1       kristaps  861: {
                    862:
1.45      schwarze  863:        if (sodest && NULL != (*sodest = curp->sodest)) {
                    864:                *man = NULL;
                    865:                return;
                    866:        }
1.9       kristaps  867:        if (man)
                    868:                *man = curp->man;
1.157     schwarze  869: }
                    870:
                    871: void
                    872: mparse_updaterc(struct mparse *curp, enum mandoclevel *rc)
                    873: {
                    874:        if (curp->file_status > *rc)
                    875:                *rc = curp->file_status;
1.3       kristaps  876: }
                    877:
                    878: void
                    879: mandoc_vmsg(enum mandocerr t, struct mparse *m,
                    880:                int ln, int pos, const char *fmt, ...)
                    881: {
                    882:        char             buf[256];
                    883:        va_list          ap;
                    884:
                    885:        va_start(ap, fmt);
1.48      schwarze  886:        (void)vsnprintf(buf, sizeof(buf), fmt, ap);
1.3       kristaps  887:        va_end(ap);
                    888:
                    889:        mandoc_msg(t, m, ln, pos, buf);
                    890: }
                    891:
                    892: void
1.47      schwarze  893: mandoc_msg(enum mandocerr er, struct mparse *m,
1.3       kristaps  894:                int ln, int col, const char *msg)
                    895: {
                    896:        enum mandoclevel level;
                    897:
1.112     schwarze  898:        level = MANDOCLEVEL_UNSUPP;
1.3       kristaps  899:        while (er < mandoclimits[level])
                    900:                level--;
                    901:
1.108     schwarze  902:        if (level < m->wlevel && er != MANDOCERR_FILE)
1.3       kristaps  903:                return;
                    904:
1.8       kristaps  905:        if (m->mmsg)
                    906:                (*m->mmsg)(er, level, m->file, ln, col, msg);
1.3       kristaps  907:
                    908:        if (m->file_status < level)
                    909:                m->file_status = level;
1.7       kristaps  910: }
                    911:
                    912: const char *
                    913: mparse_strerror(enum mandocerr er)
                    914: {
                    915:
1.142     schwarze  916:        return mandocerrs[er];
1.7       kristaps  917: }
                    918:
                    919: const char *
                    920: mparse_strlevel(enum mandoclevel lvl)
                    921: {
1.142     schwarze  922:        return mandoclevels[lvl];
1.24      kristaps  923: }
                    924:
                    925: void
                    926: mparse_keep(struct mparse *p)
                    927: {
                    928:
                    929:        assert(NULL == p->secondary);
                    930:        p->secondary = mandoc_calloc(1, sizeof(struct buf));
                    931: }
                    932:
                    933: const char *
                    934: mparse_getkeep(const struct mparse *p)
                    935: {
                    936:
                    937:        assert(p->secondary);
1.142     schwarze  938:        return p->secondary->sz ? p->secondary->buf : NULL;
1.1       kristaps  939: }

CVSweb