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

Annotation of mandoc/roff.c, Revision 1.129

1.129   ! kristaps    1: /*     $Id: roff.c,v 1.128 2011/03/20 16:02:05 kristaps Exp $ */
1.1       kristaps    2: /*
1.119     schwarze    3:  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
                      4:  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
1.1       kristaps    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
1.66      kristaps    7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
1.1       kristaps    9:  *
1.106     kristaps   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66      kristaps   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106     kristaps   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66      kristaps   13:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     14:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     15:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     16:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1       kristaps   17:  */
1.66      kristaps   18: #ifdef HAVE_CONFIG_H
                     19: #include "config.h"
                     20: #endif
1.30      kristaps   21:
1.67      kristaps   22: #include <assert.h>
1.89      kristaps   23: #include <errno.h>
1.85      kristaps   24: #include <ctype.h>
1.89      kristaps   25: #include <limits.h>
1.1       kristaps   26: #include <stdlib.h>
1.67      kristaps   27: #include <string.h>
1.75      kristaps   28: #include <stdio.h>
1.1       kristaps   29:
1.67      kristaps   30: #include "mandoc.h"
1.109     kristaps   31: #include "libroff.h"
1.94      kristaps   32: #include "libmandoc.h"
1.33      kristaps   33:
1.82      kristaps   34: #define        RSTACK_MAX      128
                     35:
1.75      kristaps   36: #define        ROFF_CTL(c) \
                     37:        ('.' == (c) || '\'' == (c))
                     38:
1.67      kristaps   39: enum   rofft {
1.103     kristaps   40:        ROFF_ad,
1.80      kristaps   41:        ROFF_am,
                     42:        ROFF_ami,
                     43:        ROFF_am1,
                     44:        ROFF_de,
                     45:        ROFF_dei,
                     46:        ROFF_de1,
1.83      schwarze   47:        ROFF_ds,
1.82      kristaps   48:        ROFF_el,
1.103     kristaps   49:        ROFF_hy,
1.82      kristaps   50:        ROFF_ie,
1.75      kristaps   51:        ROFF_if,
1.76      kristaps   52:        ROFF_ig,
1.123     schwarze   53:        ROFF_it,
1.103     kristaps   54:        ROFF_ne,
                     55:        ROFF_nh,
1.104     kristaps   56:        ROFF_nr,
1.124     schwarze   57:        ROFF_ns,
                     58:        ROFF_ps,
1.83      schwarze   59:        ROFF_rm,
1.105     kristaps   60:        ROFF_so,
1.124     schwarze   61:        ROFF_ta,
1.83      schwarze   62:        ROFF_tr,
1.109     kristaps   63:        ROFF_TS,
                     64:        ROFF_TE,
1.112     kristaps   65:        ROFF_T_,
1.125     kristaps   66:        ROFF_EQ,
                     67:        ROFF_EN,
1.76      kristaps   68:        ROFF_cblock,
1.100     kristaps   69:        ROFF_ccond, /* FIXME: remove this. */
1.106     kristaps   70:        ROFF_USERDEF,
1.67      kristaps   71:        ROFF_MAX
                     72: };
                     73:
1.82      kristaps   74: enum   roffrule {
                     75:        ROFFRULE_ALLOW,
                     76:        ROFFRULE_DENY
                     77: };
                     78:
1.94      kristaps   79: struct roffstr {
                     80:        char            *name; /* key of symbol */
                     81:        char            *string; /* current value */
                     82:        struct roffstr  *next; /* next in list */
                     83: };
                     84:
1.67      kristaps   85: struct roff {
1.128     kristaps   86:        struct mparse   *parse; /* parse point */
1.67      kristaps   87:        struct roffnode *last; /* leaf of stack */
1.82      kristaps   88:        enum roffrule    rstack[RSTACK_MAX]; /* stack of !`ie' rules */
                     89:        int              rstackpos; /* position in rstack */
1.90      kristaps   90:        struct regset   *regs; /* read/writable registers */
1.106     kristaps   91:        struct roffstr  *first_string; /* user-defined strings & macros */
                     92:        const char      *current_string; /* value of last called user macro */
1.118     kristaps   93:        struct tbl_node *first_tbl; /* first table parsed */
                     94:        struct tbl_node *last_tbl; /* last table parsed */
                     95:        struct tbl_node *tbl; /* current table being parsed */
1.125     kristaps   96:        struct eqn_node *last_eqn; /* last equation parsed */
                     97:        struct eqn_node *first_eqn; /* first equation parsed */
                     98:        struct eqn_node *eqn; /* current equation being parsed */
1.79      kristaps   99: };
                    100:
1.67      kristaps  101: struct roffnode {
                    102:        enum rofft       tok; /* type of node */
                    103:        struct roffnode *parent; /* up one in stack */
                    104:        int              line; /* parse line */
                    105:        int              col; /* parse col */
1.106     kristaps  106:        char            *name; /* node name, e.g. macro name */
1.79      kristaps  107:        char            *end; /* end-rules: custom token */
                    108:        int              endspan; /* end-rules: next-line or infty */
1.82      kristaps  109:        enum roffrule    rule; /* current evaluation rule */
1.67      kristaps  110: };
                    111:
                    112: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.72      kristaps  113:                         enum rofft tok, /* tok of macro */ \
1.67      kristaps  114:                         char **bufp, /* input buffer */ \
                    115:                         size_t *szp, /* size of input buffer */ \
                    116:                         int ln, /* parse line */ \
1.75      kristaps  117:                         int ppos, /* original pos in buffer */ \
                    118:                         int pos, /* current pos in buffer */ \
1.74      kristaps  119:                         int *offs /* reset offset of buffer data */
1.67      kristaps  120:
                    121: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    122:
                    123: struct roffmac {
                    124:        const char      *name; /* macro name */
1.79      kristaps  125:        roffproc         proc; /* process new macro */
                    126:        roffproc         text; /* process as child text of macro */
                    127:        roffproc         sub; /* process as child of macro */
                    128:        int              flags;
                    129: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.85      kristaps  130:        struct roffmac  *next;
1.67      kristaps  131: };
                    132:
1.80      kristaps  133: static enum rofferr     roff_block(ROFF_ARGS);
                    134: static enum rofferr     roff_block_text(ROFF_ARGS);
                    135: static enum rofferr     roff_block_sub(ROFF_ARGS);
                    136: static enum rofferr     roff_cblock(ROFF_ARGS);
                    137: static enum rofferr     roff_ccond(ROFF_ARGS);
1.82      kristaps  138: static enum rofferr     roff_cond(ROFF_ARGS);
                    139: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    140: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.92      schwarze  141: static enum rofferr     roff_ds(ROFF_ARGS);
1.94      kristaps  142: static enum roffrule    roff_evalcond(const char *, int *);
                    143: static void             roff_freestr(struct roff *);
1.121     schwarze  144: static char            *roff_getname(struct roff *, char **, int, int);
1.94      kristaps  145: static const char      *roff_getstrn(const struct roff *,
                    146:                                const char *, size_t);
1.103     kristaps  147: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.89      kristaps  148: static enum rofferr     roff_nr(ROFF_ARGS);
1.95      kristaps  149: static int              roff_res(struct roff *,
                    150:                                char **, size_t *, int);
1.122     schwarze  151: static enum rofferr     roff_rm(ROFF_ARGS);
1.94      kristaps  152: static void             roff_setstr(struct roff *,
1.106     kristaps  153:                                const char *, const char *, int);
1.105     kristaps  154: static enum rofferr     roff_so(ROFF_ARGS);
1.109     kristaps  155: static enum rofferr     roff_TE(ROFF_ARGS);
                    156: static enum rofferr     roff_TS(ROFF_ARGS);
1.125     kristaps  157: static enum rofferr     roff_EQ(ROFF_ARGS);
                    158: static enum rofferr     roff_EN(ROFF_ARGS);
1.112     kristaps  159: static enum rofferr     roff_T_(ROFF_ARGS);
1.106     kristaps  160: static enum rofferr     roff_userdef(ROFF_ARGS);
1.67      kristaps  161:
1.85      kristaps  162: /* See roff_hash_find() */
                    163:
                    164: #define        ASCII_HI         126
                    165: #define        ASCII_LO         33
                    166: #define        HASHWIDTH       (ASCII_HI - ASCII_LO + 1)
                    167:
                    168: static struct roffmac  *hash[HASHWIDTH];
                    169:
                    170: static struct roffmac   roffs[ROFF_MAX] = {
1.103     kristaps  171:        { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.85      kristaps  172:        { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    173:        { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    174:        { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    175:        { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    176:        { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    177:        { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.92      schwarze  178:        { "ds", roff_ds, NULL, NULL, 0, NULL },
1.85      kristaps  179:        { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.103     kristaps  180:        { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.85      kristaps  181:        { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    182:        { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    183:        { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  184:        { "it", roff_line_ignore, NULL, NULL, 0, NULL },
1.103     kristaps  185:        { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
                    186:        { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.104     kristaps  187:        { "nr", roff_nr, NULL, NULL, 0, NULL },
1.124     schwarze  188:        { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
                    189:        { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.122     schwarze  190:        { "rm", roff_rm, NULL, NULL, 0, NULL },
1.105     kristaps  191:        { "so", roff_so, NULL, NULL, 0, NULL },
1.124     schwarze  192:        { "ta", roff_line_ignore, NULL, NULL, 0, NULL },
1.103     kristaps  193:        { "tr", roff_line_ignore, NULL, NULL, 0, NULL },
1.109     kristaps  194:        { "TS", roff_TS, NULL, NULL, 0, NULL },
                    195:        { "TE", roff_TE, NULL, NULL, 0, NULL },
1.112     kristaps  196:        { "T&", roff_T_, NULL, NULL, 0, NULL },
1.125     kristaps  197:        { "EQ", roff_EQ, NULL, NULL, 0, NULL },
                    198:        { "EN", roff_EN, NULL, NULL, 0, NULL },
1.85      kristaps  199:        { ".", roff_cblock, NULL, NULL, 0, NULL },
                    200:        { "\\}", roff_ccond, NULL, NULL, 0, NULL },
1.106     kristaps  201:        { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.67      kristaps  202: };
                    203:
                    204: static void             roff_free1(struct roff *);
1.106     kristaps  205: static enum rofft       roff_hash_find(const char *, size_t);
1.85      kristaps  206: static void             roff_hash_init(void);
1.76      kristaps  207: static void             roffnode_cleanscope(struct roff *);
1.106     kristaps  208: static void             roffnode_push(struct roff *, enum rofft,
                    209:                                const char *, int, int);
1.67      kristaps  210: static void             roffnode_pop(struct roff *);
1.106     kristaps  211: static enum rofft       roff_parse(struct roff *, const char *, int *);
1.91      kristaps  212: static int              roff_parse_nat(const char *, unsigned int *);
1.67      kristaps  213:
1.85      kristaps  214: /* See roff_hash_find() */
                    215: #define        ROFF_HASH(p)    (p[0] - ASCII_LO)
                    216:
                    217: static void
                    218: roff_hash_init(void)
                    219: {
                    220:        struct roffmac   *n;
                    221:        int               buc, i;
                    222:
1.106     kristaps  223:        for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.85      kristaps  224:                assert(roffs[i].name[0] >= ASCII_LO);
                    225:                assert(roffs[i].name[0] <= ASCII_HI);
                    226:
                    227:                buc = ROFF_HASH(roffs[i].name);
                    228:
                    229:                if (NULL != (n = hash[buc])) {
                    230:                        for ( ; n->next; n = n->next)
                    231:                                /* Do nothing. */ ;
                    232:                        n->next = &roffs[i];
                    233:                } else
                    234:                        hash[buc] = &roffs[i];
                    235:        }
                    236: }
                    237:
1.67      kristaps  238:
                    239: /*
                    240:  * Look up a roff token by its name.  Returns ROFF_MAX if no macro by
                    241:  * the nil-terminated string name could be found.
                    242:  */
                    243: static enum rofft
1.106     kristaps  244: roff_hash_find(const char *p, size_t s)
1.67      kristaps  245: {
1.85      kristaps  246:        int              buc;
                    247:        struct roffmac  *n;
1.67      kristaps  248:
1.85      kristaps  249:        /*
                    250:         * libroff has an extremely simple hashtable, for the time
                    251:         * being, which simply keys on the first character, which must
                    252:         * be printable, then walks a chain.  It works well enough until
                    253:         * optimised.
                    254:         */
                    255:
                    256:        if (p[0] < ASCII_LO || p[0] > ASCII_HI)
                    257:                return(ROFF_MAX);
                    258:
                    259:        buc = ROFF_HASH(p);
                    260:
                    261:        if (NULL == (n = hash[buc]))
                    262:                return(ROFF_MAX);
                    263:        for ( ; n; n = n->next)
1.106     kristaps  264:                if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.85      kristaps  265:                        return((enum rofft)(n - roffs));
1.67      kristaps  266:
                    267:        return(ROFF_MAX);
                    268: }
                    269:
                    270:
                    271: /*
                    272:  * Pop the current node off of the stack of roff instructions currently
                    273:  * pending.
                    274:  */
                    275: static void
                    276: roffnode_pop(struct roff *r)
                    277: {
                    278:        struct roffnode *p;
                    279:
1.75      kristaps  280:        assert(r->last);
                    281:        p = r->last;
1.82      kristaps  282:
                    283:        if (ROFF_el == p->tok)
                    284:                if (r->rstackpos > -1)
                    285:                        r->rstackpos--;
                    286:
1.75      kristaps  287:        r->last = r->last->parent;
1.106     kristaps  288:        free(p->name);
                    289:        free(p->end);
1.67      kristaps  290:        free(p);
                    291: }
                    292:
                    293:
                    294: /*
                    295:  * Push a roff node onto the instruction stack.  This must later be
                    296:  * removed with roffnode_pop().
                    297:  */
1.98      schwarze  298: static void
1.106     kristaps  299: roffnode_push(struct roff *r, enum rofft tok, const char *name,
                    300:                int line, int col)
1.67      kristaps  301: {
                    302:        struct roffnode *p;
                    303:
1.98      schwarze  304:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.67      kristaps  305:        p->tok = tok;
1.106     kristaps  306:        if (name)
                    307:                p->name = mandoc_strdup(name);
1.67      kristaps  308:        p->parent = r->last;
                    309:        p->line = line;
                    310:        p->col = col;
1.79      kristaps  311:        p->rule = p->parent ? p->parent->rule : ROFFRULE_DENY;
1.67      kristaps  312:
                    313:        r->last = p;
                    314: }
                    315:
                    316:
                    317: static void
                    318: roff_free1(struct roff *r)
                    319: {
1.118     kristaps  320:        struct tbl_node *t;
1.125     kristaps  321:        struct eqn_node *e;
1.67      kristaps  322:
1.125     kristaps  323:        while (NULL != (t = r->first_tbl)) {
1.113     kristaps  324:                r->first_tbl = t->next;
                    325:                tbl_free(t);
1.109     kristaps  326:        }
                    327:
1.113     kristaps  328:        r->first_tbl = r->last_tbl = r->tbl = NULL;
                    329:
1.125     kristaps  330:        while (NULL != (e = r->first_eqn)) {
                    331:                r->first_eqn = e->next;
                    332:                eqn_free(e);
                    333:        }
                    334:
                    335:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    336:
1.67      kristaps  337:        while (r->last)
                    338:                roffnode_pop(r);
1.109     kristaps  339:
1.94      kristaps  340:        roff_freestr(r);
1.67      kristaps  341: }
                    342:
                    343:
                    344: void
                    345: roff_reset(struct roff *r)
                    346: {
                    347:
                    348:        roff_free1(r);
                    349: }
                    350:
                    351:
                    352: void
                    353: roff_free(struct roff *r)
                    354: {
                    355:
                    356:        roff_free1(r);
                    357:        free(r);
                    358: }
                    359:
                    360:
                    361: struct roff *
1.128     kristaps  362: roff_alloc(struct regset *regs, struct mparse *parse)
1.67      kristaps  363: {
                    364:        struct roff     *r;
                    365:
1.98      schwarze  366:        r = mandoc_calloc(1, sizeof(struct roff));
1.90      kristaps  367:        r->regs = regs;
1.128     kristaps  368:        r->parse = parse;
1.82      kristaps  369:        r->rstackpos = -1;
1.85      kristaps  370:
                    371:        roff_hash_init();
1.67      kristaps  372:        return(r);
                    373: }
                    374:
                    375:
1.94      kristaps  376: /*
                    377:  * Pre-filter each and every line for reserved words (one beginning with
                    378:  * `\*', e.g., `\*(ab').  These must be handled before the actual line
                    379:  * is processed.
                    380:  */
                    381: static int
1.95      kristaps  382: roff_res(struct roff *r, char **bufp, size_t *szp, int pos)
1.94      kristaps  383: {
1.108     schwarze  384:        const char      *stesc; /* start of an escape sequence ('\\') */
                    385:        const char      *stnam; /* start of the name, after "[(*" */
                    386:        const char      *cp;    /* end of the name, e.g. before ']' */
                    387:        const char      *res;   /* the string to be substituted */
1.94      kristaps  388:        int              i, maxl;
                    389:        size_t           nsz;
                    390:        char            *n;
                    391:
1.108     schwarze  392:        /* Search for a leading backslash and save a pointer to it. */
                    393:
                    394:        cp = *bufp + pos;
                    395:        while (NULL != (cp = strchr(cp, '\\'))) {
                    396:                stesc = cp++;
                    397:
                    398:                /*
                    399:                 * The second character must be an asterisk.
                    400:                 * If it isn't, skip it anyway:  It is escaped,
                    401:                 * so it can't start another escape sequence.
                    402:                 */
                    403:
                    404:                if ('\0' == *cp)
                    405:                        return(1);
                    406:                if ('*' != *cp++)
                    407:                        continue;
                    408:
                    409:                /*
                    410:                 * The third character decides the length
                    411:                 * of the name of the string.
                    412:                 * Save a pointer to the name.
                    413:                 */
                    414:
1.94      kristaps  415:                switch (*cp) {
1.108     schwarze  416:                case ('\0'):
                    417:                        return(1);
1.94      kristaps  418:                case ('('):
                    419:                        cp++;
                    420:                        maxl = 2;
                    421:                        break;
                    422:                case ('['):
                    423:                        cp++;
                    424:                        maxl = 0;
                    425:                        break;
                    426:                default:
                    427:                        maxl = 1;
                    428:                        break;
                    429:                }
1.108     schwarze  430:                stnam = cp;
1.94      kristaps  431:
1.108     schwarze  432:                /* Advance to the end of the name. */
1.94      kristaps  433:
                    434:                for (i = 0; 0 == maxl || i < maxl; i++, cp++) {
                    435:                        if ('\0' == *cp)
                    436:                                return(1); /* Error. */
                    437:                        if (0 == maxl && ']' == *cp)
                    438:                                break;
                    439:                }
                    440:
1.108     schwarze  441:                /*
                    442:                 * Retrieve the replacement string; if it is
                    443:                 * undefined, resume searching for escapes.
                    444:                 */
                    445:
                    446:                res = roff_getstrn(r, stnam, (size_t)i);
1.94      kristaps  447:
                    448:                if (NULL == res) {
                    449:                        cp -= maxl ? 1 : 0;
                    450:                        continue;
                    451:                }
                    452:
1.108     schwarze  453:                /* Replace the escape sequence by the string. */
                    454:
1.94      kristaps  455:                nsz = *szp + strlen(res) + 1;
                    456:                n = mandoc_malloc(nsz);
                    457:
1.108     schwarze  458:                strlcpy(n, *bufp, (size_t)(stesc - *bufp + 1));
1.94      kristaps  459:                strlcat(n, res, nsz);
                    460:                strlcat(n, cp + (maxl ? 0 : 1), nsz);
                    461:
                    462:                free(*bufp);
                    463:
                    464:                *bufp = n;
                    465:                *szp = nsz;
                    466:                return(0);
                    467:        }
                    468:
                    469:        return(1);
                    470: }
                    471:
                    472:
1.67      kristaps  473: enum rofferr
1.90      kristaps  474: roff_parseln(struct roff *r, int ln, char **bufp,
                    475:                size_t *szp, int pos, int *offs)
1.67      kristaps  476: {
                    477:        enum rofft       t;
1.109     kristaps  478:        enum rofferr     e;
1.79      kristaps  479:        int              ppos;
                    480:
                    481:        /*
1.94      kristaps  482:         * Run the reserved-word filter only if we have some reserved
                    483:         * words to fill in.
                    484:         */
                    485:
1.95      kristaps  486:        if (r->first_string && ! roff_res(r, bufp, szp, pos))
1.106     kristaps  487:                return(ROFF_REPARSE);
1.94      kristaps  488:
                    489:        /*
1.79      kristaps  490:         * First, if a scope is open and we're not a macro, pass the
                    491:         * text through the macro's filter.  If a scope isn't open and
                    492:         * we're not a macro, just let it through.
1.125     kristaps  493:         * Finally, if there's an equation scope open, divert it into it
                    494:         * no matter our state.
1.79      kristaps  495:         */
1.74      kristaps  496:
1.75      kristaps  497:        if (r->last && ! ROFF_CTL((*bufp)[pos])) {
1.78      kristaps  498:                t = r->last->tok;
                    499:                assert(roffs[t].text);
1.109     kristaps  500:                e = (*roffs[t].text)
                    501:                        (r, t, bufp, szp, ln, pos, pos, offs);
                    502:                assert(ROFF_IGN == e || ROFF_CONT == e);
1.125     kristaps  503:                if (ROFF_CONT != e)
                    504:                        return(e);
                    505:                if (r->eqn)
                    506:                        return(eqn_read(&r->eqn, ln, *bufp, *offs));
                    507:                if (r->tbl)
1.109     kristaps  508:                        return(tbl_read(r->tbl, ln, *bufp, *offs));
1.125     kristaps  509:                return(ROFF_CONT);
1.109     kristaps  510:        } else if ( ! ROFF_CTL((*bufp)[pos])) {
1.125     kristaps  511:                if (r->eqn)
                    512:                        return(eqn_read(&r->eqn, ln, *bufp, *offs));
1.109     kristaps  513:                if (r->tbl)
                    514:                        return(tbl_read(r->tbl, ln, *bufp, *offs));
1.67      kristaps  515:                return(ROFF_CONT);
1.125     kristaps  516:        } else if (r->eqn)
                    517:                return(eqn_read(&r->eqn, ln, *bufp, *offs));
1.67      kristaps  518:
1.79      kristaps  519:        /*
                    520:         * If a scope is open, go to the child handler for that macro,
                    521:         * as it may want to preprocess before doing anything with it.
1.125     kristaps  522:         * Don't do so if an equation is open.
1.79      kristaps  523:         */
1.78      kristaps  524:
1.79      kristaps  525:        if (r->last) {
                    526:                t = r->last->tok;
                    527:                assert(roffs[t].sub);
                    528:                return((*roffs[t].sub)
1.90      kristaps  529:                                (r, t, bufp, szp,
                    530:                                 ln, pos, pos, offs));
1.79      kristaps  531:        }
1.78      kristaps  532:
1.79      kristaps  533:        /*
                    534:         * Lastly, as we've no scope open, try to look up and execute
                    535:         * the new macro.  If no macro is found, simply return and let
                    536:         * the compilers handle it.
                    537:         */
1.67      kristaps  538:
1.75      kristaps  539:        ppos = pos;
1.106     kristaps  540:        if (ROFF_MAX == (t = roff_parse(r, *bufp, &pos)))
1.79      kristaps  541:                return(ROFF_CONT);
1.67      kristaps  542:
1.75      kristaps  543:        assert(roffs[t].proc);
1.78      kristaps  544:        return((*roffs[t].proc)
1.90      kristaps  545:                        (r, t, bufp, szp,
                    546:                         ln, ppos, pos, offs));
1.74      kristaps  547: }
                    548:
                    549:
1.117     kristaps  550: void
1.74      kristaps  551: roff_endparse(struct roff *r)
                    552: {
                    553:
1.110     kristaps  554:        if (r->last)
1.128     kristaps  555:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.109     kristaps  556:                                r->last->line, r->last->col, NULL);
1.117     kristaps  557:
1.125     kristaps  558:        if (r->eqn) {
1.128     kristaps  559:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.126     kristaps  560:                                r->eqn->eqn.line, r->eqn->eqn.pos, NULL);
1.125     kristaps  561:                eqn_end(r->eqn);
                    562:                r->eqn = NULL;
                    563:        }
                    564:
1.117     kristaps  565:        if (r->tbl) {
1.128     kristaps  566:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.117     kristaps  567:                                r->tbl->line, r->tbl->pos, NULL);
                    568:                tbl_end(r->tbl);
                    569:                r->tbl = NULL;
                    570:        }
1.67      kristaps  571: }
                    572:
                    573:
                    574: /*
                    575:  * Parse a roff node's type from the input buffer.  This must be in the
                    576:  * form of ".foo xxx" in the usual way.
                    577:  */
                    578: static enum rofft
1.106     kristaps  579: roff_parse(struct roff *r, const char *buf, int *pos)
1.67      kristaps  580: {
1.106     kristaps  581:        const char      *mac;
                    582:        size_t           maclen;
1.67      kristaps  583:        enum rofft       t;
                    584:
1.75      kristaps  585:        assert(ROFF_CTL(buf[*pos]));
                    586:        (*pos)++;
1.67      kristaps  587:
1.106     kristaps  588:        while (' ' == buf[*pos] || '\t' == buf[*pos])
1.67      kristaps  589:                (*pos)++;
                    590:
                    591:        if ('\0' == buf[*pos])
                    592:                return(ROFF_MAX);
                    593:
1.106     kristaps  594:        mac = buf + *pos;
                    595:        maclen = strcspn(mac, " \\\t\0");
1.67      kristaps  596:
1.106     kristaps  597:        t = (r->current_string = roff_getstrn(r, mac, maclen))
                    598:            ? ROFF_USERDEF : roff_hash_find(mac, maclen);
1.67      kristaps  599:
1.127     kristaps  600:        *pos += (int)maclen;
1.67      kristaps  601:        while (buf[*pos] && ' ' == buf[*pos])
                    602:                (*pos)++;
                    603:
                    604:        return(t);
                    605: }
                    606:
                    607:
1.89      kristaps  608: static int
1.91      kristaps  609: roff_parse_nat(const char *buf, unsigned int *res)
1.89      kristaps  610: {
                    611:        char            *ep;
                    612:        long             lval;
                    613:
                    614:        errno = 0;
                    615:        lval = strtol(buf, &ep, 10);
                    616:        if (buf[0] == '\0' || *ep != '\0')
                    617:                return(0);
                    618:        if ((errno == ERANGE &&
                    619:                        (lval == LONG_MAX || lval == LONG_MIN)) ||
1.91      kristaps  620:                        (lval > INT_MAX || lval < 0))
1.89      kristaps  621:                return(0);
                    622:
1.91      kristaps  623:        *res = (unsigned int)lval;
1.89      kristaps  624:        return(1);
                    625: }
                    626:
                    627:
1.67      kristaps  628: /* ARGSUSED */
                    629: static enum rofferr
1.76      kristaps  630: roff_cblock(ROFF_ARGS)
1.67      kristaps  631: {
                    632:
1.79      kristaps  633:        /*
                    634:         * A block-close `..' should only be invoked as a child of an
                    635:         * ignore macro, otherwise raise a warning and just ignore it.
                    636:         */
                    637:
1.76      kristaps  638:        if (NULL == r->last) {
1.128     kristaps  639:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.76      kristaps  640:                return(ROFF_IGN);
                    641:        }
1.67      kristaps  642:
1.81      kristaps  643:        switch (r->last->tok) {
                    644:        case (ROFF_am):
                    645:                /* FALLTHROUGH */
                    646:        case (ROFF_ami):
                    647:                /* FALLTHROUGH */
                    648:        case (ROFF_am1):
                    649:                /* FALLTHROUGH */
                    650:        case (ROFF_de):
1.108     schwarze  651:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.81      kristaps  652:                /* FALLTHROUGH */
                    653:        case (ROFF_dei):
                    654:                /* FALLTHROUGH */
                    655:        case (ROFF_ig):
                    656:                break;
                    657:        default:
1.128     kristaps  658:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.67      kristaps  659:                return(ROFF_IGN);
1.76      kristaps  660:        }
1.67      kristaps  661:
1.76      kristaps  662:        if ((*bufp)[pos])
1.128     kristaps  663:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.71      kristaps  664:
                    665:        roffnode_pop(r);
1.76      kristaps  666:        roffnode_cleanscope(r);
                    667:        return(ROFF_IGN);
1.71      kristaps  668:
1.67      kristaps  669: }
                    670:
                    671:
1.76      kristaps  672: static void
                    673: roffnode_cleanscope(struct roff *r)
1.67      kristaps  674: {
                    675:
1.76      kristaps  676:        while (r->last) {
                    677:                if (--r->last->endspan < 0)
                    678:                        break;
                    679:                roffnode_pop(r);
                    680:        }
1.67      kristaps  681: }
                    682:
                    683:
1.75      kristaps  684: /* ARGSUSED */
1.74      kristaps  685: static enum rofferr
1.75      kristaps  686: roff_ccond(ROFF_ARGS)
1.74      kristaps  687: {
                    688:
1.76      kristaps  689:        if (NULL == r->last) {
1.128     kristaps  690:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.76      kristaps  691:                return(ROFF_IGN);
                    692:        }
                    693:
1.82      kristaps  694:        switch (r->last->tok) {
                    695:        case (ROFF_el):
                    696:                /* FALLTHROUGH */
                    697:        case (ROFF_ie):
                    698:                /* FALLTHROUGH */
                    699:        case (ROFF_if):
                    700:                break;
                    701:        default:
1.128     kristaps  702:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.75      kristaps  703:                return(ROFF_IGN);
                    704:        }
                    705:
1.76      kristaps  706:        if (r->last->endspan > -1) {
1.128     kristaps  707:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.76      kristaps  708:                return(ROFF_IGN);
                    709:        }
                    710:
                    711:        if ((*bufp)[pos])
1.128     kristaps  712:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.76      kristaps  713:
1.75      kristaps  714:        roffnode_pop(r);
1.76      kristaps  715:        roffnode_cleanscope(r);
                    716:        return(ROFF_IGN);
                    717: }
                    718:
1.75      kristaps  719:
1.76      kristaps  720: /* ARGSUSED */
                    721: static enum rofferr
1.80      kristaps  722: roff_block(ROFF_ARGS)
1.76      kristaps  723: {
1.78      kristaps  724:        int             sv;
                    725:        size_t          sz;
1.106     kristaps  726:        char            *name;
                    727:
                    728:        name = NULL;
1.76      kristaps  729:
1.106     kristaps  730:        if (ROFF_ig != tok) {
                    731:                if ('\0' == (*bufp)[pos]) {
1.128     kristaps  732:                        mandoc_msg(MANDOCERR_NOARGS, r->parse, ln, ppos, NULL);
1.106     kristaps  733:                        return(ROFF_IGN);
                    734:                }
1.107     kristaps  735:
                    736:                /*
                    737:                 * Re-write `de1', since we don't really care about
                    738:                 * groff's strange compatibility mode, into `de'.
                    739:                 */
                    740:
1.106     kristaps  741:                if (ROFF_de1 == tok)
                    742:                        tok = ROFF_de;
                    743:                if (ROFF_de == tok)
                    744:                        name = *bufp + pos;
                    745:                else
1.128     kristaps  746:                        mandoc_msg(MANDOCERR_REQUEST, r->parse, ln, ppos,
1.106     kristaps  747:                            roffs[tok].name);
1.107     kristaps  748:
1.80      kristaps  749:                while ((*bufp)[pos] && ' ' != (*bufp)[pos])
                    750:                        pos++;
1.107     kristaps  751:
1.80      kristaps  752:                while (' ' == (*bufp)[pos])
1.106     kristaps  753:                        (*bufp)[pos++] = '\0';
1.80      kristaps  754:        }
                    755:
1.106     kristaps  756:        roffnode_push(r, tok, name, ln, ppos);
                    757:
                    758:        /*
                    759:         * At the beginning of a `de' macro, clear the existing string
                    760:         * with the same name, if there is one.  New content will be
                    761:         * added from roff_block_text() in multiline mode.
                    762:         */
1.107     kristaps  763:
1.106     kristaps  764:        if (ROFF_de == tok)
1.108     schwarze  765:                roff_setstr(r, name, "", 0);
1.76      kristaps  766:
1.79      kristaps  767:        if ('\0' == (*bufp)[pos])
1.78      kristaps  768:                return(ROFF_IGN);
                    769:
1.107     kristaps  770:        /* If present, process the custom end-of-line marker. */
                    771:
1.78      kristaps  772:        sv = pos;
1.108     schwarze  773:        while ((*bufp)[pos] &&
1.107     kristaps  774:                        ' ' != (*bufp)[pos] &&
1.78      kristaps  775:                        '\t' != (*bufp)[pos])
                    776:                pos++;
                    777:
                    778:        /*
                    779:         * Note: groff does NOT like escape characters in the input.
                    780:         * Instead of detecting this, we're just going to let it fly and
                    781:         * to hell with it.
                    782:         */
                    783:
                    784:        assert(pos > sv);
                    785:        sz = (size_t)(pos - sv);
                    786:
1.79      kristaps  787:        if (1 == sz && '.' == (*bufp)[sv])
                    788:                return(ROFF_IGN);
                    789:
1.98      schwarze  790:        r->last->end = mandoc_malloc(sz + 1);
1.78      kristaps  791:
                    792:        memcpy(r->last->end, *bufp + sv, sz);
                    793:        r->last->end[(int)sz] = '\0';
                    794:
1.77      kristaps  795:        if ((*bufp)[pos])
1.128     kristaps  796:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.74      kristaps  797:
1.78      kristaps  798:        return(ROFF_IGN);
                    799: }
                    800:
                    801:
                    802: /* ARGSUSED */
                    803: static enum rofferr
1.80      kristaps  804: roff_block_sub(ROFF_ARGS)
1.79      kristaps  805: {
                    806:        enum rofft      t;
                    807:        int             i, j;
                    808:
                    809:        /*
                    810:         * First check whether a custom macro exists at this level.  If
                    811:         * it does, then check against it.  This is some of groff's
                    812:         * stranger behaviours.  If we encountered a custom end-scope
                    813:         * tag and that tag also happens to be a "real" macro, then we
                    814:         * need to try interpreting it again as a real macro.  If it's
                    815:         * not, then return ignore.  Else continue.
                    816:         */
                    817:
                    818:        if (r->last->end) {
                    819:                i = pos + 1;
                    820:                while (' ' == (*bufp)[i] || '\t' == (*bufp)[i])
                    821:                        i++;
                    822:
                    823:                for (j = 0; r->last->end[j]; j++, i++)
                    824:                        if ((*bufp)[i] != r->last->end[j])
                    825:                                break;
                    826:
                    827:                if ('\0' == r->last->end[j] &&
                    828:                                ('\0' == (*bufp)[i] ||
                    829:                                 ' ' == (*bufp)[i] ||
                    830:                                 '\t' == (*bufp)[i])) {
                    831:                        roffnode_pop(r);
                    832:                        roffnode_cleanscope(r);
                    833:
1.106     kristaps  834:                        if (ROFF_MAX != roff_parse(r, *bufp, &pos))
1.79      kristaps  835:                                return(ROFF_RERUN);
                    836:                        return(ROFF_IGN);
                    837:                }
                    838:        }
                    839:
                    840:        /*
                    841:         * If we have no custom end-query or lookup failed, then try
                    842:         * pulling it out of the hashtable.
                    843:         */
                    844:
                    845:        ppos = pos;
1.106     kristaps  846:        t = roff_parse(r, *bufp, &pos);
1.79      kristaps  847:
1.106     kristaps  848:        /*
                    849:         * Macros other than block-end are only significant
                    850:         * in `de' blocks; elsewhere, simply throw them away.
                    851:         */
                    852:        if (ROFF_cblock != t) {
                    853:                if (ROFF_de == tok)
                    854:                        roff_setstr(r, r->last->name, *bufp + ppos, 1);
1.79      kristaps  855:                return(ROFF_IGN);
1.106     kristaps  856:        }
1.79      kristaps  857:
                    858:        assert(roffs[t].proc);
1.90      kristaps  859:        return((*roffs[t].proc)(r, t, bufp, szp,
                    860:                                ln, ppos, pos, offs));
1.79      kristaps  861: }
                    862:
                    863:
                    864: /* ARGSUSED */
                    865: static enum rofferr
1.80      kristaps  866: roff_block_text(ROFF_ARGS)
1.78      kristaps  867: {
                    868:
1.106     kristaps  869:        if (ROFF_de == tok)
                    870:                roff_setstr(r, r->last->name, *bufp + pos, 1);
                    871:
1.78      kristaps  872:        return(ROFF_IGN);
                    873: }
                    874:
                    875:
                    876: /* ARGSUSED */
                    877: static enum rofferr
1.82      kristaps  878: roff_cond_sub(ROFF_ARGS)
                    879: {
                    880:        enum rofft       t;
                    881:        enum roffrule    rr;
                    882:
                    883:        ppos = pos;
                    884:        rr = r->last->rule;
                    885:
1.87      kristaps  886:        /*
                    887:         * Clean out scope.  If we've closed ourselves, then don't
                    888:         * continue.
                    889:         */
                    890:
                    891:        roffnode_cleanscope(r);
1.82      kristaps  892:
1.106     kristaps  893:        if (ROFF_MAX == (t = roff_parse(r, *bufp, &pos))) {
1.100     kristaps  894:                if ('\\' == (*bufp)[pos] && '}' == (*bufp)[pos + 1])
                    895:                        return(roff_ccond
1.106     kristaps  896:                                (r, ROFF_ccond, bufp, szp,
1.100     kristaps  897:                                 ln, pos, pos + 2, offs));
1.82      kristaps  898:                return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
1.100     kristaps  899:        }
1.82      kristaps  900:
                    901:        /*
                    902:         * A denied conditional must evaluate its children if and only
                    903:         * if they're either structurally required (such as loops and
                    904:         * conditionals) or a closing macro.
                    905:         */
                    906:        if (ROFFRULE_DENY == rr)
                    907:                if ( ! (ROFFMAC_STRUCT & roffs[t].flags))
                    908:                        if (ROFF_ccond != t)
                    909:                                return(ROFF_IGN);
                    910:
                    911:        assert(roffs[t].proc);
1.90      kristaps  912:        return((*roffs[t].proc)(r, t, bufp, szp,
                    913:                                ln, ppos, pos, offs));
1.82      kristaps  914: }
                    915:
                    916:
                    917: /* ARGSUSED */
                    918: static enum rofferr
                    919: roff_cond_text(ROFF_ARGS)
1.78      kristaps  920: {
                    921:        char            *ep, *st;
1.82      kristaps  922:        enum roffrule    rr;
                    923:
                    924:        rr = r->last->rule;
                    925:
                    926:        /*
                    927:         * We display the value of the text if out current evaluation
                    928:         * scope permits us to do so.
                    929:         */
1.100     kristaps  930:
                    931:        /* FIXME: use roff_ccond? */
1.78      kristaps  932:
                    933:        st = &(*bufp)[pos];
                    934:        if (NULL == (ep = strstr(st, "\\}"))) {
                    935:                roffnode_cleanscope(r);
1.82      kristaps  936:                return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
1.78      kristaps  937:        }
                    938:
1.86      kristaps  939:        if (ep == st || (ep > st && '\\' != *(ep - 1)))
1.78      kristaps  940:                roffnode_pop(r);
                    941:
                    942:        roffnode_cleanscope(r);
1.82      kristaps  943:        return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
1.74      kristaps  944: }
                    945:
                    946:
1.88      kristaps  947: static enum roffrule
                    948: roff_evalcond(const char *v, int *pos)
                    949: {
                    950:
                    951:        switch (v[*pos]) {
                    952:        case ('n'):
                    953:                (*pos)++;
                    954:                return(ROFFRULE_ALLOW);
                    955:        case ('e'):
                    956:                /* FALLTHROUGH */
                    957:        case ('o'):
                    958:                /* FALLTHROUGH */
                    959:        case ('t'):
                    960:                (*pos)++;
                    961:                return(ROFFRULE_DENY);
                    962:        default:
                    963:                break;
                    964:        }
                    965:
                    966:        while (v[*pos] && ' ' != v[*pos])
                    967:                (*pos)++;
                    968:        return(ROFFRULE_DENY);
                    969: }
                    970:
1.75      kristaps  971: /* ARGSUSED */
1.74      kristaps  972: static enum rofferr
1.103     kristaps  973: roff_line_ignore(ROFF_ARGS)
1.89      kristaps  974: {
1.123     schwarze  975:
                    976:        if (ROFF_it == tok)
1.128     kristaps  977:                mandoc_msg(MANDOCERR_REQUEST, r->parse, ln, ppos, "it");
1.89      kristaps  978:
                    979:        return(ROFF_IGN);
                    980: }
                    981:
1.104     kristaps  982: /* ARGSUSED */
                    983: static enum rofferr
1.82      kristaps  984: roff_cond(ROFF_ARGS)
1.74      kristaps  985: {
1.77      kristaps  986:        int              sv;
1.88      kristaps  987:        enum roffrule    rule;
1.74      kristaps  988:
1.82      kristaps  989:        /* Stack overflow! */
                    990:
                    991:        if (ROFF_ie == tok && r->rstackpos == RSTACK_MAX - 1) {
1.128     kristaps  992:                mandoc_msg(MANDOCERR_MEM, r->parse, ln, ppos, NULL);
1.82      kristaps  993:                return(ROFF_ERR);
                    994:        }
1.74      kristaps  995:
1.88      kristaps  996:        /* First, evaluate the conditional. */
1.84      schwarze  997:
1.88      kristaps  998:        if (ROFF_el == tok) {
                    999:                /*
                   1000:                 * An `.el' will get the value of the current rstack
                   1001:                 * entry set in prior `ie' calls or defaults to DENY.
                   1002:                 */
                   1003:                if (r->rstackpos < 0)
                   1004:                        rule = ROFFRULE_DENY;
                   1005:                else
                   1006:                        rule = r->rstack[r->rstackpos];
                   1007:        } else
                   1008:                rule = roff_evalcond(*bufp, &pos);
1.77      kristaps 1009:
                   1010:        sv = pos;
1.88      kristaps 1011:
1.75      kristaps 1012:        while (' ' == (*bufp)[pos])
                   1013:                pos++;
1.74      kristaps 1014:
1.77      kristaps 1015:        /*
                   1016:         * Roff is weird.  If we have just white-space after the
                   1017:         * conditional, it's considered the BODY and we exit without
                   1018:         * really doing anything.  Warn about this.  It's probably
                   1019:         * wrong.
                   1020:         */
1.88      kristaps 1021:
1.77      kristaps 1022:        if ('\0' == (*bufp)[pos] && sv != pos) {
1.128     kristaps 1023:                mandoc_msg(MANDOCERR_NOARGS, r->parse, ln, ppos, NULL);
1.107     kristaps 1024:                return(ROFF_IGN);
1.77      kristaps 1025:        }
                   1026:
1.106     kristaps 1027:        roffnode_push(r, tok, NULL, ln, ppos);
1.77      kristaps 1028:
1.88      kristaps 1029:        r->last->rule = rule;
                   1030:
1.84      schwarze 1031:        if (ROFF_ie == tok) {
1.82      kristaps 1032:                /*
                   1033:                 * An if-else will put the NEGATION of the current
                   1034:                 * evaluated conditional into the stack.
                   1035:                 */
                   1036:                r->rstackpos++;
                   1037:                if (ROFFRULE_DENY == r->last->rule)
                   1038:                        r->rstack[r->rstackpos] = ROFFRULE_ALLOW;
                   1039:                else
                   1040:                        r->rstack[r->rstackpos] = ROFFRULE_DENY;
                   1041:        }
1.88      kristaps 1042:
                   1043:        /* If the parent has false as its rule, then so do we. */
                   1044:
1.109     kristaps 1045:        if (r->last->parent && ROFFRULE_DENY == r->last->parent->rule)
1.84      schwarze 1046:                r->last->rule = ROFFRULE_DENY;
1.88      kristaps 1047:
                   1048:        /*
                   1049:         * Determine scope.  If we're invoked with "\{" trailing the
                   1050:         * conditional, then we're in a multiline scope.  Else our scope
                   1051:         * expires on the next line.
                   1052:         */
1.74      kristaps 1053:
1.75      kristaps 1054:        r->last->endspan = 1;
                   1055:
                   1056:        if ('\\' == (*bufp)[pos] && '{' == (*bufp)[pos + 1]) {
                   1057:                r->last->endspan = -1;
                   1058:                pos += 2;
1.109     kristaps 1059:        }
1.74      kristaps 1060:
1.77      kristaps 1061:        /*
                   1062:         * If there are no arguments on the line, the next-line scope is
                   1063:         * assumed.
                   1064:         */
                   1065:
1.75      kristaps 1066:        if ('\0' == (*bufp)[pos])
                   1067:                return(ROFF_IGN);
1.77      kristaps 1068:
                   1069:        /* Otherwise re-run the roff parser after recalculating. */
1.74      kristaps 1070:
1.75      kristaps 1071:        *offs = pos;
                   1072:        return(ROFF_RERUN);
1.83      schwarze 1073: }
                   1074:
                   1075:
                   1076: /* ARGSUSED */
                   1077: static enum rofferr
1.92      schwarze 1078: roff_ds(ROFF_ARGS)
                   1079: {
1.96      kristaps 1080:        char            *name, *string;
                   1081:
                   1082:        /*
                   1083:         * A symbol is named by the first word following the macro
                   1084:         * invocation up to a space.  Its value is anything after the
                   1085:         * name's trailing whitespace and optional double-quote.  Thus,
                   1086:         *
                   1087:         *  [.ds foo "bar  "     ]
                   1088:         *
                   1089:         * will have `bar  "     ' as its value.
                   1090:         */
1.92      schwarze 1091:
1.121     schwarze 1092:        string = *bufp + pos;
                   1093:        name = roff_getname(r, &string, ln, pos);
1.92      schwarze 1094:        if ('\0' == *name)
                   1095:                return(ROFF_IGN);
                   1096:
1.121     schwarze 1097:        /* Read past initial double-quote. */
                   1098:        if ('"' == *string)
1.92      schwarze 1099:                string++;
                   1100:
1.96      kristaps 1101:        /* The rest is the value. */
1.106     kristaps 1102:        roff_setstr(r, name, string, 0);
1.92      schwarze 1103:        return(ROFF_IGN);
                   1104: }
                   1105:
                   1106:
                   1107: /* ARGSUSED */
                   1108: static enum rofferr
1.89      kristaps 1109: roff_nr(ROFF_ARGS)
1.83      schwarze 1110: {
1.121     schwarze 1111:        const char      *key;
                   1112:        char            *val;
1.91      kristaps 1113:        struct reg      *rg;
1.89      kristaps 1114:
1.121     schwarze 1115:        val = *bufp + pos;
                   1116:        key = roff_getname(r, &val, ln, pos);
1.91      kristaps 1117:        rg = r->regs->regs;
1.89      kristaps 1118:
                   1119:        if (0 == strcmp(key, "nS")) {
1.91      kristaps 1120:                rg[(int)REG_nS].set = 1;
                   1121:                if ( ! roff_parse_nat(val, &rg[(int)REG_nS].v.u))
                   1122:                        rg[(int)REG_nS].v.u = 0;
1.109     kristaps 1123:        }
                   1124:
1.122     schwarze 1125:        return(ROFF_IGN);
                   1126: }
                   1127:
                   1128: /* ARGSUSED */
                   1129: static enum rofferr
                   1130: roff_rm(ROFF_ARGS)
                   1131: {
                   1132:        const char       *name;
                   1133:        char             *cp;
                   1134:
                   1135:        cp = *bufp + pos;
                   1136:        while ('\0' != *cp) {
1.127     kristaps 1137:                name = roff_getname(r, &cp, ln, (int)(cp - *bufp));
1.122     schwarze 1138:                if ('\0' != *name)
                   1139:                        roff_setstr(r, name, NULL, 0);
                   1140:        }
1.109     kristaps 1141:        return(ROFF_IGN);
                   1142: }
                   1143:
                   1144: /* ARGSUSED */
                   1145: static enum rofferr
                   1146: roff_TE(ROFF_ARGS)
                   1147: {
                   1148:
                   1149:        if (NULL == r->tbl)
1.128     kristaps 1150:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.115     kristaps 1151:        else
                   1152:                tbl_end(r->tbl);
1.109     kristaps 1153:
                   1154:        r->tbl = NULL;
1.112     kristaps 1155:        return(ROFF_IGN);
                   1156: }
                   1157:
                   1158: /* ARGSUSED */
                   1159: static enum rofferr
                   1160: roff_T_(ROFF_ARGS)
                   1161: {
                   1162:
                   1163:        if (NULL == r->tbl)
1.128     kristaps 1164:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.112     kristaps 1165:        else
1.116     kristaps 1166:                tbl_restart(ppos, ln, r->tbl);
1.112     kristaps 1167:
1.109     kristaps 1168:        return(ROFF_IGN);
                   1169: }
                   1170:
                   1171: /* ARGSUSED */
                   1172: static enum rofferr
1.125     kristaps 1173: roff_EQ(ROFF_ARGS)
                   1174: {
                   1175:        struct eqn_node *e;
                   1176:
                   1177:        assert(NULL == r->eqn);
                   1178:        e = eqn_alloc(ppos, ln);
                   1179:
                   1180:        if (r->last_eqn)
                   1181:                r->last_eqn->next = e;
                   1182:        else
                   1183:                r->first_eqn = r->last_eqn = e;
                   1184:
                   1185:        r->eqn = r->last_eqn = e;
                   1186:        return(ROFF_IGN);
                   1187: }
                   1188:
                   1189: /* ARGSUSED */
                   1190: static enum rofferr
                   1191: roff_EN(ROFF_ARGS)
                   1192: {
                   1193:
1.128     kristaps 1194:        mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.125     kristaps 1195:        return(ROFF_IGN);
                   1196: }
                   1197:
                   1198: /* ARGSUSED */
                   1199: static enum rofferr
1.109     kristaps 1200: roff_TS(ROFF_ARGS)
                   1201: {
1.118     kristaps 1202:        struct tbl_node *t;
1.89      kristaps 1203:
1.115     kristaps 1204:        if (r->tbl) {
1.128     kristaps 1205:                mandoc_msg(MANDOCERR_SCOPEBROKEN, r->parse, ln, ppos, NULL);
1.115     kristaps 1206:                tbl_end(r->tbl);
                   1207:        }
1.83      schwarze 1208:
1.128     kristaps 1209:        t = tbl_alloc(ppos, ln, r->parse);
1.113     kristaps 1210:
                   1211:        if (r->last_tbl)
                   1212:                r->last_tbl->next = t;
                   1213:        else
                   1214:                r->first_tbl = r->last_tbl = t;
                   1215:
                   1216:        r->tbl = r->last_tbl = t;
1.83      schwarze 1217:        return(ROFF_IGN);
1.92      schwarze 1218: }
                   1219:
1.105     kristaps 1220: /* ARGSUSED */
                   1221: static enum rofferr
                   1222: roff_so(ROFF_ARGS)
                   1223: {
                   1224:        char *name;
                   1225:
1.128     kristaps 1226:        mandoc_msg(MANDOCERR_SO, r->parse, ln, ppos, NULL);
1.105     kristaps 1227:
                   1228:        /*
                   1229:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   1230:         * opening anything that's not in our cwd or anything beneath
                   1231:         * it.  Thus, explicitly disallow traversing up the file-system
                   1232:         * or using absolute paths.
                   1233:         */
                   1234:
                   1235:        name = *bufp + pos;
                   1236:        if ('/' == *name || strstr(name, "../") || strstr(name, "/..")) {
1.128     kristaps 1237:                mandoc_msg(MANDOCERR_SOPATH, r->parse, ln, pos, NULL);
1.105     kristaps 1238:                return(ROFF_ERR);
                   1239:        }
                   1240:
                   1241:        *offs = pos;
                   1242:        return(ROFF_SO);
                   1243: }
1.92      schwarze 1244:
1.106     kristaps 1245: /* ARGSUSED */
                   1246: static enum rofferr
                   1247: roff_userdef(ROFF_ARGS)
1.99      kristaps 1248: {
1.106     kristaps 1249:        const char       *arg[9];
                   1250:        char             *cp, *n1, *n2;
1.119     schwarze 1251:        int               i;
1.106     kristaps 1252:
                   1253:        /*
                   1254:         * Collect pointers to macro argument strings
                   1255:         * and null-terminate them.
                   1256:         */
                   1257:        cp = *bufp + pos;
1.119     schwarze 1258:        for (i = 0; i < 9; i++)
1.120     schwarze 1259:                arg[i] = '\0' == *cp ? "" :
1.128     kristaps 1260:                    mandoc_getarg(r->parse, &cp, ln, &pos);
1.99      kristaps 1261:
1.106     kristaps 1262:        /*
                   1263:         * Expand macro arguments.
1.99      kristaps 1264:         */
1.106     kristaps 1265:        *szp = 0;
                   1266:        n1 = cp = mandoc_strdup(r->current_string);
                   1267:        while (NULL != (cp = strstr(cp, "\\$"))) {
                   1268:                i = cp[2] - '1';
                   1269:                if (0 > i || 8 < i) {
                   1270:                        /* Not an argument invocation. */
                   1271:                        cp += 2;
                   1272:                        continue;
                   1273:                }
                   1274:
                   1275:                *szp = strlen(n1) - 3 + strlen(arg[i]) + 1;
                   1276:                n2 = mandoc_malloc(*szp);
                   1277:
                   1278:                strlcpy(n2, n1, (size_t)(cp - n1 + 1));
                   1279:                strlcat(n2, arg[i], *szp);
                   1280:                strlcat(n2, cp + 3, *szp);
                   1281:
                   1282:                cp = n2 + (cp - n1);
                   1283:                free(n1);
                   1284:                n1 = n2;
1.99      kristaps 1285:        }
                   1286:
1.106     kristaps 1287:        /*
                   1288:         * Replace the macro invocation
                   1289:         * by the expanded macro.
                   1290:         */
                   1291:        free(*bufp);
                   1292:        *bufp = n1;
                   1293:        if (0 == *szp)
                   1294:                *szp = strlen(*bufp) + 1;
                   1295:
                   1296:        return(*szp > 1 && '\n' == (*bufp)[(int)*szp - 2] ?
                   1297:           ROFF_REPARSE : ROFF_APPEND);
1.99      kristaps 1298: }
1.121     schwarze 1299:
                   1300: static char *
                   1301: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   1302: {
                   1303:        char     *name, *cp;
                   1304:
                   1305:        name = *cpp;
                   1306:        if ('\0' == *name)
                   1307:                return(name);
                   1308:
                   1309:        /* Read until end of name. */
                   1310:        for (cp = name; '\0' != *cp && ' ' != *cp; cp++) {
                   1311:                if ('\\' != *cp)
                   1312:                        continue;
                   1313:                cp++;
                   1314:                if ('\\' == *cp)
                   1315:                        continue;
1.128     kristaps 1316:                mandoc_msg(MANDOCERR_NAMESC, r->parse, ln, pos, NULL);
1.121     schwarze 1317:                *cp = '\0';
                   1318:                name = cp;
                   1319:        }
                   1320:
                   1321:        /* Nil-terminate name. */
                   1322:        if ('\0' != *cp)
                   1323:                *(cp++) = '\0';
                   1324:
                   1325:        /* Read past spaces. */
                   1326:        while (' ' == *cp)
                   1327:                cp++;
                   1328:
                   1329:        *cpp = cp;
                   1330:        return(name);
                   1331: }
                   1332:
1.106     kristaps 1333: /*
                   1334:  * Store *string into the user-defined string called *name.
                   1335:  * In multiline mode, append to an existing entry and append '\n';
                   1336:  * else replace the existing entry, if there is one.
                   1337:  * To clear an existing entry, call with (*r, *name, NULL, 0).
                   1338:  */
1.94      kristaps 1339: static void
1.106     kristaps 1340: roff_setstr(struct roff *r, const char *name, const char *string,
                   1341:        int multiline)
1.92      schwarze 1342: {
                   1343:        struct roffstr   *n;
1.106     kristaps 1344:        char             *c;
                   1345:        size_t            oldch, newch;
1.92      schwarze 1346:
1.106     kristaps 1347:        /* Search for an existing string with the same name. */
1.94      kristaps 1348:        n = r->first_string;
1.92      schwarze 1349:        while (n && strcmp(name, n->name))
                   1350:                n = n->next;
1.94      kristaps 1351:
                   1352:        if (NULL == n) {
1.106     kristaps 1353:                /* Create a new string table entry. */
1.94      kristaps 1354:                n = mandoc_malloc(sizeof(struct roffstr));
1.106     kristaps 1355:                n->name = mandoc_strdup(name);
                   1356:                n->string = NULL;
1.94      kristaps 1357:                n->next = r->first_string;
                   1358:                r->first_string = n;
1.106     kristaps 1359:        } else if (0 == multiline) {
                   1360:                /* In multiline mode, append; else replace. */
1.92      schwarze 1361:                free(n->string);
1.106     kristaps 1362:                n->string = NULL;
                   1363:        }
                   1364:
                   1365:        if (NULL == string)
                   1366:                return;
                   1367:
                   1368:        /*
                   1369:         * One additional byte for the '\n' in multiline mode,
                   1370:         * and one for the terminating '\0'.
                   1371:         */
1.127     kristaps 1372:        newch = strlen(string) + (multiline ? 2u : 1u);
1.106     kristaps 1373:        if (NULL == n->string) {
                   1374:                n->string = mandoc_malloc(newch);
                   1375:                *n->string = '\0';
                   1376:                oldch = 0;
                   1377:        } else {
                   1378:                oldch = strlen(n->string);
                   1379:                n->string = mandoc_realloc(n->string, oldch + newch);
                   1380:        }
                   1381:
                   1382:        /* Skip existing content in the destination buffer. */
1.127     kristaps 1383:        c = n->string + (int)oldch;
1.106     kristaps 1384:
                   1385:        /* Append new content to the destination buffer. */
                   1386:        while (*string) {
                   1387:                /*
                   1388:                 * Rudimentary roff copy mode:
                   1389:                 * Handle escaped backslashes.
                   1390:                 */
                   1391:                if ('\\' == *string && '\\' == *(string + 1))
                   1392:                        string++;
                   1393:                *c++ = *string++;
                   1394:        }
1.94      kristaps 1395:
1.106     kristaps 1396:        /* Append terminating bytes. */
                   1397:        if (multiline)
                   1398:                *c++ = '\n';
                   1399:        *c = '\0';
1.92      schwarze 1400: }
                   1401:
1.94      kristaps 1402: static const char *
                   1403: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92      schwarze 1404: {
1.94      kristaps 1405:        const struct roffstr *n;
1.92      schwarze 1406:
1.94      kristaps 1407:        n = r->first_string;
1.97      kristaps 1408:        while (n && (strncmp(name, n->name, len) || '\0' != n->name[(int)len]))
1.92      schwarze 1409:                n = n->next;
1.94      kristaps 1410:
                   1411:        return(n ? n->string : NULL);
1.92      schwarze 1412: }
                   1413:
1.94      kristaps 1414: static void
                   1415: roff_freestr(struct roff *r)
1.92      schwarze 1416: {
                   1417:        struct roffstr   *n, *nn;
                   1418:
1.94      kristaps 1419:        for (n = r->first_string; n; n = nn) {
1.92      schwarze 1420:                free(n->name);
                   1421:                free(n->string);
                   1422:                nn = n->next;
                   1423:                free(n);
                   1424:        }
1.94      kristaps 1425:
                   1426:        r->first_string = NULL;
1.114     kristaps 1427: }
                   1428:
                   1429: const struct tbl_span *
                   1430: roff_span(const struct roff *r)
                   1431: {
                   1432:
                   1433:        return(r->tbl ? tbl_span(r->tbl) : NULL);
1.125     kristaps 1434: }
                   1435:
                   1436: const struct eqn *
                   1437: roff_eqn(const struct roff *r)
                   1438: {
                   1439:
                   1440:        return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.74      kristaps 1441: }

CVSweb