[BACK]Return to roff.c CVS log [TXT][DIR] Up to [local] / src / usr.bin / mandoc

Annotation of src/usr.bin/mandoc/roff.c, Revision 1.41

1.41    ! schwarze    1: /*     $Id: roff.c,v 1.40 2011/07/31 14:11:48 schwarze Exp $ */
1.1       schwarze    2: /*
1.27      schwarze    3:  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.25      schwarze    4:  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
1.1       schwarze    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
                      7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
                      9:  *
1.16      schwarze   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1       schwarze   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16      schwarze   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1       schwarze   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.
                     17:  */
                     18: #include <assert.h>
1.3       schwarze   19: #include <ctype.h>
1.1       schwarze   20: #include <stdlib.h>
                     21: #include <string.h>
                     22:
                     23: #include "mandoc.h"
1.27      schwarze   24: #include "libroff.h"
1.8       schwarze   25: #include "libmandoc.h"
1.1       schwarze   26:
1.37      schwarze   27: /* Maximum number of nested if-else conditionals. */
1.2       schwarze   28: #define        RSTACK_MAX      128
                     29:
1.1       schwarze   30: enum   rofft {
1.20      schwarze   31:        ROFF_ad,
1.2       schwarze   32:        ROFF_am,
                     33:        ROFF_ami,
                     34:        ROFF_am1,
1.1       schwarze   35:        ROFF_de,
                     36:        ROFF_dei,
1.2       schwarze   37:        ROFF_de1,
                     38:        ROFF_ds,
                     39:        ROFF_el,
1.20      schwarze   40:        ROFF_hy,
1.2       schwarze   41:        ROFF_ie,
                     42:        ROFF_if,
1.1       schwarze   43:        ROFF_ig,
1.30      schwarze   44:        ROFF_it,
1.20      schwarze   45:        ROFF_ne,
                     46:        ROFF_nh,
1.14      schwarze   47:        ROFF_nr,
1.31      schwarze   48:        ROFF_ns,
                     49:        ROFF_ps,
1.2       schwarze   50:        ROFF_rm,
1.14      schwarze   51:        ROFF_so,
1.31      schwarze   52:        ROFF_ta,
1.2       schwarze   53:        ROFF_tr,
1.27      schwarze   54:        ROFF_TS,
                     55:        ROFF_TE,
                     56:        ROFF_T_,
1.32      schwarze   57:        ROFF_EQ,
                     58:        ROFF_EN,
1.2       schwarze   59:        ROFF_cblock,
1.37      schwarze   60:        ROFF_ccond,
1.16      schwarze   61:        ROFF_USERDEF,
1.1       schwarze   62:        ROFF_MAX
                     63: };
                     64:
1.2       schwarze   65: enum   roffrule {
                     66:        ROFFRULE_ALLOW,
                     67:        ROFFRULE_DENY
                     68: };
                     69:
1.41    ! schwarze   70: /*
        !            71:  * A single register entity.  If "set" is zero, the value of the
        !            72:  * register should be the default one, which is per-register.
        !            73:  * Registers are assumed to be unsigned ints for now.
        !            74:  */
        !            75: struct reg {
        !            76:        int               set; /* whether set or not */
        !            77:        unsigned int      u; /* unsigned integer */
        !            78: };
        !            79:
1.8       schwarze   80: struct roffstr {
                     81:        char            *name; /* key of symbol */
                     82:        char            *string; /* current value */
                     83:        struct roffstr  *next; /* next in list */
                     84: };
                     85:
1.1       schwarze   86: struct roff {
1.35      schwarze   87:        struct mparse   *parse; /* parse point */
1.1       schwarze   88:        struct roffnode *last; /* leaf of stack */
1.2       schwarze   89:        enum roffrule    rstack[RSTACK_MAX]; /* stack of !`ie' rules */
                     90:        int              rstackpos; /* position in rstack */
1.41    ! schwarze   91:        struct reg       regs[REG__MAX];
1.16      schwarze   92:        struct roffstr  *first_string; /* user-defined strings & macros */
                     93:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   94:        struct tbl_node *first_tbl; /* first table parsed */
                     95:        struct tbl_node *last_tbl; /* last table parsed */
                     96:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze   97:        struct eqn_node *last_eqn; /* last equation parsed */
                     98:        struct eqn_node *first_eqn; /* first equation parsed */
                     99:        struct eqn_node *eqn; /* current equation being parsed */
1.1       schwarze  100: };
                    101:
                    102: struct roffnode {
                    103:        enum rofft       tok; /* type of node */
                    104:        struct roffnode *parent; /* up one in stack */
                    105:        int              line; /* parse line */
                    106:        int              col; /* parse col */
1.16      schwarze  107:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  108:        char            *end; /* end-rules: custom token */
                    109:        int              endspan; /* end-rules: next-line or infty */
                    110:        enum roffrule    rule; /* current evaluation rule */
1.1       schwarze  111: };
                    112:
                    113: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
                    114:                         enum rofft tok, /* tok of macro */ \
                    115:                         char **bufp, /* input buffer */ \
                    116:                         size_t *szp, /* size of input buffer */ \
                    117:                         int ln, /* parse line */ \
1.2       schwarze  118:                         int ppos, /* original pos in buffer */ \
                    119:                         int pos, /* current pos in buffer */ \
                    120:                         int *offs /* reset offset of buffer data */
1.1       schwarze  121:
                    122: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    123:
                    124: struct roffmac {
                    125:        const char      *name; /* macro name */
1.2       schwarze  126:        roffproc         proc; /* process new macro */
                    127:        roffproc         text; /* process as child text of macro */
                    128:        roffproc         sub; /* process as child of macro */
                    129:        int              flags;
                    130: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.3       schwarze  131:        struct roffmac  *next;
1.1       schwarze  132: };
                    133:
1.37      schwarze  134: struct predef {
                    135:        const char      *name; /* predefined input name */
                    136:        const char      *str; /* replacement symbol */
                    137: };
                    138:
                    139: #define        PREDEF(__name, __str) \
                    140:        { (__name), (__str) },
                    141:
1.2       schwarze  142: static enum rofferr     roff_block(ROFF_ARGS);
                    143: static enum rofferr     roff_block_text(ROFF_ARGS);
                    144: static enum rofferr     roff_block_sub(ROFF_ARGS);
                    145: static enum rofferr     roff_cblock(ROFF_ARGS);
                    146: static enum rofferr     roff_ccond(ROFF_ARGS);
                    147: static enum rofferr     roff_cond(ROFF_ARGS);
                    148: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    149: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  150: static enum rofferr     roff_ds(ROFF_ARGS);
1.8       schwarze  151: static enum roffrule    roff_evalcond(const char *, int *);
                    152: static void             roff_freestr(struct roff *);
1.28      schwarze  153: static char            *roff_getname(struct roff *, char **, int, int);
1.8       schwarze  154: static const char      *roff_getstrn(const struct roff *,
                    155:                                const char *, size_t);
1.21      schwarze  156: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.6       schwarze  157: static enum rofferr     roff_nr(ROFF_ARGS);
1.41    ! schwarze  158: static void             roff_openeqn(struct roff *, const char *,
        !           159:                                int, int, const char *);
1.9       schwarze  160: static int              roff_res(struct roff *,
1.37      schwarze  161:                                char **, size_t *, int, int);
1.29      schwarze  162: static enum rofferr     roff_rm(ROFF_ARGS);
1.8       schwarze  163: static void             roff_setstr(struct roff *,
1.16      schwarze  164:                                const char *, const char *, int);
1.14      schwarze  165: static enum rofferr     roff_so(ROFF_ARGS);
1.27      schwarze  166: static enum rofferr     roff_TE(ROFF_ARGS);
                    167: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  168: static enum rofferr     roff_EQ(ROFF_ARGS);
                    169: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  170: static enum rofferr     roff_T_(ROFF_ARGS);
1.16      schwarze  171: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  172:
1.3       schwarze  173: /* See roff_hash_find() */
                    174:
                    175: #define        ASCII_HI         126
                    176: #define        ASCII_LO         33
                    177: #define        HASHWIDTH       (ASCII_HI - ASCII_LO + 1)
                    178:
                    179: static struct roffmac  *hash[HASHWIDTH];
                    180:
                    181: static struct roffmac   roffs[ROFF_MAX] = {
1.21      schwarze  182:        { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  183:        { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    184:        { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    185:        { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    186:        { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    187:        { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    188:        { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.7       schwarze  189:        { "ds", roff_ds, NULL, NULL, 0, NULL },
1.3       schwarze  190:        { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.21      schwarze  191:        { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  192:        { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    193:        { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    194:        { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.30      schwarze  195:        { "it", roff_line_ignore, NULL, NULL, 0, NULL },
1.21      schwarze  196:        { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
                    197:        { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.14      schwarze  198:        { "nr", roff_nr, NULL, NULL, 0, NULL },
1.31      schwarze  199:        { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
                    200:        { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.29      schwarze  201:        { "rm", roff_rm, NULL, NULL, 0, NULL },
1.14      schwarze  202:        { "so", roff_so, NULL, NULL, 0, NULL },
1.31      schwarze  203:        { "ta", roff_line_ignore, NULL, NULL, 0, NULL },
1.21      schwarze  204:        { "tr", roff_line_ignore, NULL, NULL, 0, NULL },
1.27      schwarze  205:        { "TS", roff_TS, NULL, NULL, 0, NULL },
                    206:        { "TE", roff_TE, NULL, NULL, 0, NULL },
                    207:        { "T&", roff_T_, NULL, NULL, 0, NULL },
1.32      schwarze  208:        { "EQ", roff_EQ, NULL, NULL, 0, NULL },
                    209:        { "EN", roff_EN, NULL, NULL, 0, NULL },
1.3       schwarze  210:        { ".", roff_cblock, NULL, NULL, 0, NULL },
                    211:        { "\\}", roff_ccond, NULL, NULL, 0, NULL },
1.16      schwarze  212:        { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1       schwarze  213: };
                    214:
1.37      schwarze  215: /* Array of injected predefined strings. */
                    216: #define        PREDEFS_MAX      38
                    217: static const struct predef predefs[PREDEFS_MAX] = {
                    218: #include "predefs.in"
                    219: };
                    220:
1.1       schwarze  221: static void             roff_free1(struct roff *);
1.16      schwarze  222: static enum rofft       roff_hash_find(const char *, size_t);
1.3       schwarze  223: static void             roff_hash_init(void);
1.2       schwarze  224: static void             roffnode_cleanscope(struct roff *);
1.16      schwarze  225: static void             roffnode_push(struct roff *, enum rofft,
                    226:                                const char *, int, int);
1.1       schwarze  227: static void             roffnode_pop(struct roff *);
1.16      schwarze  228: static enum rofft       roff_parse(struct roff *, const char *, int *);
1.1       schwarze  229:
1.3       schwarze  230: /* See roff_hash_find() */
                    231: #define        ROFF_HASH(p)    (p[0] - ASCII_LO)
                    232:
                    233: static void
                    234: roff_hash_init(void)
                    235: {
                    236:        struct roffmac   *n;
                    237:        int               buc, i;
                    238:
1.16      schwarze  239:        for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3       schwarze  240:                assert(roffs[i].name[0] >= ASCII_LO);
                    241:                assert(roffs[i].name[0] <= ASCII_HI);
                    242:
                    243:                buc = ROFF_HASH(roffs[i].name);
                    244:
                    245:                if (NULL != (n = hash[buc])) {
                    246:                        for ( ; n->next; n = n->next)
                    247:                                /* Do nothing. */ ;
                    248:                        n->next = &roffs[i];
                    249:                } else
                    250:                        hash[buc] = &roffs[i];
                    251:        }
                    252: }
                    253:
1.1       schwarze  254: /*
                    255:  * Look up a roff token by its name.  Returns ROFF_MAX if no macro by
                    256:  * the nil-terminated string name could be found.
                    257:  */
                    258: static enum rofft
1.16      schwarze  259: roff_hash_find(const char *p, size_t s)
1.1       schwarze  260: {
1.3       schwarze  261:        int              buc;
                    262:        struct roffmac  *n;
1.1       schwarze  263:
1.3       schwarze  264:        /*
                    265:         * libroff has an extremely simple hashtable, for the time
                    266:         * being, which simply keys on the first character, which must
                    267:         * be printable, then walks a chain.  It works well enough until
                    268:         * optimised.
                    269:         */
                    270:
                    271:        if (p[0] < ASCII_LO || p[0] > ASCII_HI)
                    272:                return(ROFF_MAX);
                    273:
                    274:        buc = ROFF_HASH(p);
                    275:
                    276:        if (NULL == (n = hash[buc]))
                    277:                return(ROFF_MAX);
                    278:        for ( ; n; n = n->next)
1.16      schwarze  279:                if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3       schwarze  280:                        return((enum rofft)(n - roffs));
1.1       schwarze  281:
                    282:        return(ROFF_MAX);
                    283: }
                    284:
                    285:
                    286: /*
                    287:  * Pop the current node off of the stack of roff instructions currently
                    288:  * pending.
                    289:  */
                    290: static void
                    291: roffnode_pop(struct roff *r)
                    292: {
                    293:        struct roffnode *p;
                    294:
1.2       schwarze  295:        assert(r->last);
                    296:        p = r->last;
                    297:
                    298:        r->last = r->last->parent;
1.16      schwarze  299:        free(p->name);
                    300:        free(p->end);
1.1       schwarze  301:        free(p);
                    302: }
                    303:
                    304:
                    305: /*
                    306:  * Push a roff node onto the instruction stack.  This must later be
                    307:  * removed with roffnode_pop().
                    308:  */
1.11      schwarze  309: static void
1.16      schwarze  310: roffnode_push(struct roff *r, enum rofft tok, const char *name,
                    311:                int line, int col)
1.1       schwarze  312: {
                    313:        struct roffnode *p;
                    314:
1.11      schwarze  315:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  316:        p->tok = tok;
1.16      schwarze  317:        if (name)
                    318:                p->name = mandoc_strdup(name);
1.1       schwarze  319:        p->parent = r->last;
                    320:        p->line = line;
                    321:        p->col = col;
1.2       schwarze  322:        p->rule = p->parent ? p->parent->rule : ROFFRULE_DENY;
1.1       schwarze  323:
                    324:        r->last = p;
                    325: }
                    326:
                    327:
                    328: static void
                    329: roff_free1(struct roff *r)
                    330: {
1.27      schwarze  331:        struct tbl_node *t;
1.32      schwarze  332:        struct eqn_node *e;
1.27      schwarze  333:
1.32      schwarze  334:        while (NULL != (t = r->first_tbl)) {
1.27      schwarze  335:                r->first_tbl = t->next;
                    336:                tbl_free(t);
                    337:        }
                    338:
                    339:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  340:
1.32      schwarze  341:        while (NULL != (e = r->first_eqn)) {
                    342:                r->first_eqn = e->next;
                    343:                eqn_free(e);
                    344:        }
                    345:
                    346:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    347:
1.1       schwarze  348:        while (r->last)
                    349:                roffnode_pop(r);
1.27      schwarze  350:
1.8       schwarze  351:        roff_freestr(r);
1.1       schwarze  352: }
                    353:
                    354:
                    355: void
                    356: roff_reset(struct roff *r)
                    357: {
1.38      schwarze  358:        int              i;
1.1       schwarze  359:
                    360:        roff_free1(r);
1.38      schwarze  361:
1.41    ! schwarze  362:        memset(&r->regs, 0, sizeof(struct reg) * REG__MAX);
        !           363:
1.38      schwarze  364:        for (i = 0; i < PREDEFS_MAX; i++)
                    365:                roff_setstr(r, predefs[i].name, predefs[i].str, 0);
1.1       schwarze  366: }
                    367:
                    368:
                    369: void
                    370: roff_free(struct roff *r)
                    371: {
                    372:
                    373:        roff_free1(r);
                    374:        free(r);
                    375: }
                    376:
                    377:
                    378: struct roff *
1.41    ! schwarze  379: roff_alloc(struct mparse *parse)
1.1       schwarze  380: {
                    381:        struct roff     *r;
1.37      schwarze  382:        int              i;
1.1       schwarze  383:
1.11      schwarze  384:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  385:        r->parse = parse;
1.2       schwarze  386:        r->rstackpos = -1;
1.3       schwarze  387:
                    388:        roff_hash_init();
1.37      schwarze  389:
                    390:        for (i = 0; i < PREDEFS_MAX; i++)
                    391:                roff_setstr(r, predefs[i].name, predefs[i].str, 0);
                    392:
1.1       schwarze  393:        return(r);
                    394: }
                    395:
                    396:
1.8       schwarze  397: /*
                    398:  * Pre-filter each and every line for reserved words (one beginning with
                    399:  * `\*', e.g., `\*(ab').  These must be handled before the actual line
                    400:  * is processed.
                    401:  */
                    402: static int
1.37      schwarze  403: roff_res(struct roff *r, char **bufp, size_t *szp, int ln, int pos)
1.8       schwarze  404: {
1.23      schwarze  405:        const char      *stesc; /* start of an escape sequence ('\\') */
                    406:        const char      *stnam; /* start of the name, after "[(*" */
                    407:        const char      *cp;    /* end of the name, e.g. before ']' */
                    408:        const char      *res;   /* the string to be substituted */
1.8       schwarze  409:        int              i, maxl;
                    410:        size_t           nsz;
                    411:        char            *n;
                    412:
1.24      schwarze  413:        /* Search for a leading backslash and save a pointer to it. */
1.23      schwarze  414:
1.24      schwarze  415:        cp = *bufp + pos;
                    416:        while (NULL != (cp = strchr(cp, '\\'))) {
                    417:                stesc = cp++;
1.23      schwarze  418:
                    419:                /*
                    420:                 * The second character must be an asterisk.
                    421:                 * If it isn't, skip it anyway:  It is escaped,
                    422:                 * so it can't start another escape sequence.
                    423:                 */
                    424:
1.24      schwarze  425:                if ('\0' == *cp)
                    426:                        return(1);
                    427:                if ('*' != *cp++)
1.23      schwarze  428:                        continue;
                    429:
                    430:                /*
                    431:                 * The third character decides the length
                    432:                 * of the name of the string.
                    433:                 * Save a pointer to the name.
                    434:                 */
                    435:
1.24      schwarze  436:                switch (*cp) {
                    437:                case ('\0'):
                    438:                        return(1);
1.8       schwarze  439:                case ('('):
                    440:                        cp++;
                    441:                        maxl = 2;
                    442:                        break;
                    443:                case ('['):
                    444:                        cp++;
                    445:                        maxl = 0;
                    446:                        break;
                    447:                default:
                    448:                        maxl = 1;
                    449:                        break;
                    450:                }
1.23      schwarze  451:                stnam = cp;
1.8       schwarze  452:
1.23      schwarze  453:                /* Advance to the end of the name. */
1.8       schwarze  454:
                    455:                for (i = 0; 0 == maxl || i < maxl; i++, cp++) {
                    456:                        if ('\0' == *cp)
                    457:                                return(1); /* Error. */
                    458:                        if (0 == maxl && ']' == *cp)
                    459:                                break;
                    460:                }
                    461:
1.23      schwarze  462:                /*
                    463:                 * Retrieve the replacement string; if it is
                    464:                 * undefined, resume searching for escapes.
                    465:                 */
                    466:
                    467:                res = roff_getstrn(r, stnam, (size_t)i);
1.8       schwarze  468:
                    469:                if (NULL == res) {
1.37      schwarze  470:                        /* TODO: keep track of the correct position. */
                    471:                        mandoc_msg(MANDOCERR_BADESCAPE, r->parse, ln, pos, NULL);
                    472:                        res = "";
1.8       schwarze  473:                }
                    474:
1.23      schwarze  475:                /* Replace the escape sequence by the string. */
                    476:
1.8       schwarze  477:                nsz = *szp + strlen(res) + 1;
                    478:                n = mandoc_malloc(nsz);
                    479:
1.23      schwarze  480:                strlcpy(n, *bufp, (size_t)(stesc - *bufp + 1));
1.8       schwarze  481:                strlcat(n, res, nsz);
                    482:                strlcat(n, cp + (maxl ? 0 : 1), nsz);
                    483:
                    484:                free(*bufp);
                    485:
                    486:                *bufp = n;
                    487:                *szp = nsz;
                    488:                return(0);
                    489:        }
                    490:
                    491:        return(1);
                    492: }
                    493:
                    494:
1.1       schwarze  495: enum rofferr
1.6       schwarze  496: roff_parseln(struct roff *r, int ln, char **bufp,
                    497:                size_t *szp, int pos, int *offs)
1.1       schwarze  498: {
                    499:        enum rofft       t;
1.27      schwarze  500:        enum rofferr     e;
1.35      schwarze  501:        int              ppos, ctl;
1.1       schwarze  502:
1.2       schwarze  503:        /*
1.8       schwarze  504:         * Run the reserved-word filter only if we have some reserved
                    505:         * words to fill in.
                    506:         */
                    507:
1.37      schwarze  508:        if (r->first_string && ! roff_res(r, bufp, szp, ln, pos))
1.16      schwarze  509:                return(ROFF_REPARSE);
1.8       schwarze  510:
1.35      schwarze  511:        ppos = pos;
                    512:        ctl = mandoc_getcontrol(*bufp, &pos);
                    513:
1.8       schwarze  514:        /*
1.2       schwarze  515:         * First, if a scope is open and we're not a macro, pass the
                    516:         * text through the macro's filter.  If a scope isn't open and
                    517:         * we're not a macro, just let it through.
1.32      schwarze  518:         * Finally, if there's an equation scope open, divert it into it
                    519:         * no matter our state.
1.2       schwarze  520:         */
                    521:
1.35      schwarze  522:        if (r->last && ! ctl) {
1.2       schwarze  523:                t = r->last->tok;
                    524:                assert(roffs[t].text);
1.27      schwarze  525:                e = (*roffs[t].text)
                    526:                        (r, t, bufp, szp, ln, pos, pos, offs);
                    527:                assert(ROFF_IGN == e || ROFF_CONT == e);
1.32      schwarze  528:                if (ROFF_CONT != e)
                    529:                        return(e);
                    530:                if (r->eqn)
1.41    ! schwarze  531:                        return(eqn_read(&r->eqn, ln, *bufp, pos, offs));
1.32      schwarze  532:                if (r->tbl)
1.35      schwarze  533:                        return(tbl_read(r->tbl, ln, *bufp, pos));
1.32      schwarze  534:                return(ROFF_CONT);
1.35      schwarze  535:        } else if ( ! ctl) {
1.32      schwarze  536:                if (r->eqn)
1.41    ! schwarze  537:                        return(eqn_read(&r->eqn, ln, *bufp, pos, offs));
1.27      schwarze  538:                if (r->tbl)
1.35      schwarze  539:                        return(tbl_read(r->tbl, ln, *bufp, pos));
1.2       schwarze  540:                return(ROFF_CONT);
1.32      schwarze  541:        } else if (r->eqn)
1.41    ! schwarze  542:                return(eqn_read(&r->eqn, ln, *bufp, ppos, offs));
1.2       schwarze  543:
                    544:        /*
                    545:         * If a scope is open, go to the child handler for that macro,
                    546:         * as it may want to preprocess before doing anything with it.
1.32      schwarze  547:         * Don't do so if an equation is open.
1.2       schwarze  548:         */
                    549:
                    550:        if (r->last) {
1.1       schwarze  551:                t = r->last->tok;
                    552:                assert(roffs[t].sub);
1.2       schwarze  553:                return((*roffs[t].sub)
1.8       schwarze  554:                                (r, t, bufp, szp,
1.35      schwarze  555:                                 ln, ppos, pos, offs));
1.2       schwarze  556:        }
                    557:
                    558:        /*
                    559:         * Lastly, as we've no scope open, try to look up and execute
                    560:         * the new macro.  If no macro is found, simply return and let
                    561:         * the compilers handle it.
                    562:         */
                    563:
1.16      schwarze  564:        if (ROFF_MAX == (t = roff_parse(r, *bufp, &pos)))
1.1       schwarze  565:                return(ROFF_CONT);
                    566:
1.2       schwarze  567:        assert(roffs[t].proc);
                    568:        return((*roffs[t].proc)
1.8       schwarze  569:                        (r, t, bufp, szp,
                    570:                         ln, ppos, pos, offs));
1.2       schwarze  571: }
                    572:
1.1       schwarze  573:
1.27      schwarze  574: void
1.2       schwarze  575: roff_endparse(struct roff *r)
                    576: {
1.1       schwarze  577:
1.27      schwarze  578:        if (r->last)
1.35      schwarze  579:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.27      schwarze  580:                                r->last->line, r->last->col, NULL);
                    581:
1.32      schwarze  582:        if (r->eqn) {
1.35      schwarze  583:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.41    ! schwarze  584:                                r->eqn->eqn.ln, r->eqn->eqn.pos, NULL);
        !           585:                eqn_end(&r->eqn);
1.32      schwarze  586:        }
                    587:
1.27      schwarze  588:        if (r->tbl) {
1.35      schwarze  589:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.27      schwarze  590:                                r->tbl->line, r->tbl->pos, NULL);
1.41    ! schwarze  591:                tbl_end(&r->tbl);
1.27      schwarze  592:        }
1.1       schwarze  593: }
                    594:
                    595: /*
                    596:  * Parse a roff node's type from the input buffer.  This must be in the
                    597:  * form of ".foo xxx" in the usual way.
                    598:  */
                    599: static enum rofft
1.16      schwarze  600: roff_parse(struct roff *r, const char *buf, int *pos)
1.1       schwarze  601: {
1.16      schwarze  602:        const char      *mac;
                    603:        size_t           maclen;
1.1       schwarze  604:        enum rofft       t;
                    605:
1.39      schwarze  606:        if ('\0' == buf[*pos] || '"' == buf[*pos] ||
                    607:                        '\t' == buf[*pos] || ' ' == buf[*pos])
1.1       schwarze  608:                return(ROFF_MAX);
                    609:
1.39      schwarze  610:        /*
                    611:         * We stop the macro parse at an escape, tab, space, or nil.
                    612:         * However, `\}' is also a valid macro, so make sure we don't
                    613:         * clobber it by seeing the `\' as the end of token.
                    614:         */
                    615:
1.16      schwarze  616:        mac = buf + *pos;
1.39      schwarze  617:        maclen = strcspn(mac + 1, " \\\t\0") + 1;
1.1       schwarze  618:
1.16      schwarze  619:        t = (r->current_string = roff_getstrn(r, mac, maclen))
                    620:            ? ROFF_USERDEF : roff_hash_find(mac, maclen);
1.1       schwarze  621:
1.34      schwarze  622:        *pos += (int)maclen;
1.35      schwarze  623:
1.1       schwarze  624:        while (buf[*pos] && ' ' == buf[*pos])
                    625:                (*pos)++;
                    626:
                    627:        return(t);
                    628: }
                    629:
                    630: /* ARGSUSED */
                    631: static enum rofferr
1.2       schwarze  632: roff_cblock(ROFF_ARGS)
1.1       schwarze  633: {
                    634:
1.2       schwarze  635:        /*
                    636:         * A block-close `..' should only be invoked as a child of an
                    637:         * ignore macro, otherwise raise a warning and just ignore it.
                    638:         */
                    639:
                    640:        if (NULL == r->last) {
1.35      schwarze  641:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.2       schwarze  642:                return(ROFF_IGN);
                    643:        }
1.1       schwarze  644:
1.2       schwarze  645:        switch (r->last->tok) {
                    646:        case (ROFF_am):
                    647:                /* FALLTHROUGH */
                    648:        case (ROFF_ami):
                    649:                /* FALLTHROUGH */
                    650:        case (ROFF_am1):
                    651:                /* FALLTHROUGH */
                    652:        case (ROFF_de):
1.23      schwarze  653:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2       schwarze  654:                /* FALLTHROUGH */
                    655:        case (ROFF_dei):
                    656:                /* FALLTHROUGH */
                    657:        case (ROFF_ig):
                    658:                break;
                    659:        default:
1.35      schwarze  660:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.1       schwarze  661:                return(ROFF_IGN);
1.2       schwarze  662:        }
                    663:
                    664:        if ((*bufp)[pos])
1.35      schwarze  665:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.2       schwarze  666:
                    667:        roffnode_pop(r);
                    668:        roffnode_cleanscope(r);
                    669:        return(ROFF_IGN);
                    670:
                    671: }
1.1       schwarze  672:
                    673:
1.2       schwarze  674: static void
                    675: roffnode_cleanscope(struct roff *r)
                    676: {
1.1       schwarze  677:
1.2       schwarze  678:        while (r->last) {
                    679:                if (--r->last->endspan < 0)
                    680:                        break;
                    681:                roffnode_pop(r);
                    682:        }
                    683: }
1.1       schwarze  684:
                    685:
1.2       schwarze  686: /* ARGSUSED */
                    687: static enum rofferr
                    688: roff_ccond(ROFF_ARGS)
                    689: {
1.1       schwarze  690:
1.2       schwarze  691:        if (NULL == r->last) {
1.35      schwarze  692:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.1       schwarze  693:                return(ROFF_IGN);
1.2       schwarze  694:        }
1.1       schwarze  695:
1.2       schwarze  696:        switch (r->last->tok) {
                    697:        case (ROFF_el):
                    698:                /* FALLTHROUGH */
                    699:        case (ROFF_ie):
                    700:                /* FALLTHROUGH */
                    701:        case (ROFF_if):
                    702:                break;
                    703:        default:
1.35      schwarze  704:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.2       schwarze  705:                return(ROFF_IGN);
                    706:        }
1.1       schwarze  707:
1.2       schwarze  708:        if (r->last->endspan > -1) {
1.35      schwarze  709:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.1       schwarze  710:                return(ROFF_IGN);
1.2       schwarze  711:        }
                    712:
                    713:        if ((*bufp)[pos])
1.35      schwarze  714:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.1       schwarze  715:
1.2       schwarze  716:        roffnode_pop(r);
                    717:        roffnode_cleanscope(r);
1.1       schwarze  718:        return(ROFF_IGN);
                    719: }
                    720:
                    721:
                    722: /* ARGSUSED */
                    723: static enum rofferr
1.2       schwarze  724: roff_block(ROFF_ARGS)
1.1       schwarze  725: {
1.2       schwarze  726:        int             sv;
                    727:        size_t          sz;
1.16      schwarze  728:        char            *name;
                    729:
                    730:        name = NULL;
1.2       schwarze  731:
1.16      schwarze  732:        if (ROFF_ig != tok) {
                    733:                if ('\0' == (*bufp)[pos]) {
1.35      schwarze  734:                        mandoc_msg(MANDOCERR_NOARGS, r->parse, ln, ppos, NULL);
1.16      schwarze  735:                        return(ROFF_IGN);
                    736:                }
1.22      schwarze  737:
                    738:                /*
                    739:                 * Re-write `de1', since we don't really care about
                    740:                 * groff's strange compatibility mode, into `de'.
                    741:                 */
                    742:
1.18      schwarze  743:                if (ROFF_de1 == tok)
                    744:                        tok = ROFF_de;
1.16      schwarze  745:                if (ROFF_de == tok)
                    746:                        name = *bufp + pos;
1.21      schwarze  747:                else
1.35      schwarze  748:                        mandoc_msg(MANDOCERR_REQUEST, r->parse, ln, ppos,
1.21      schwarze  749:                            roffs[tok].name);
1.22      schwarze  750:
1.33      schwarze  751:                while ((*bufp)[pos] && ! isspace((unsigned char)(*bufp)[pos]))
1.2       schwarze  752:                        pos++;
1.22      schwarze  753:
1.33      schwarze  754:                while (isspace((unsigned char)(*bufp)[pos]))
1.16      schwarze  755:                        (*bufp)[pos++] = '\0';
1.2       schwarze  756:        }
                    757:
1.16      schwarze  758:        roffnode_push(r, tok, name, ln, ppos);
                    759:
                    760:        /*
                    761:         * At the beginning of a `de' macro, clear the existing string
                    762:         * with the same name, if there is one.  New content will be
                    763:         * added from roff_block_text() in multiline mode.
                    764:         */
1.22      schwarze  765:
1.16      schwarze  766:        if (ROFF_de == tok)
1.19      schwarze  767:                roff_setstr(r, name, "", 0);
1.2       schwarze  768:
                    769:        if ('\0' == (*bufp)[pos])
                    770:                return(ROFF_IGN);
1.1       schwarze  771:
1.22      schwarze  772:        /* If present, process the custom end-of-line marker. */
                    773:
1.2       schwarze  774:        sv = pos;
1.33      schwarze  775:        while ((*bufp)[pos] && ! isspace((unsigned char)(*bufp)[pos]))
1.2       schwarze  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:
                    787:        if (1 == sz && '.' == (*bufp)[sv])
                    788:                return(ROFF_IGN);
                    789:
1.11      schwarze  790:        r->last->end = mandoc_malloc(sz + 1);
1.2       schwarze  791:
                    792:        memcpy(r->last->end, *bufp + sv, sz);
                    793:        r->last->end[(int)sz] = '\0';
                    794:
                    795:        if ((*bufp)[pos])
1.35      schwarze  796:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.1       schwarze  797:
                    798:        return(ROFF_IGN);
                    799: }
                    800:
                    801:
                    802: /* ARGSUSED */
                    803: static enum rofferr
1.2       schwarze  804: roff_block_sub(ROFF_ARGS)
1.1       schwarze  805: {
1.2       schwarze  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) {
1.35      schwarze  819:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.2       schwarze  820:                        if ((*bufp)[i] != r->last->end[j])
                    821:                                break;
1.1       schwarze  822:
1.2       schwarze  823:                if ('\0' == r->last->end[j] &&
                    824:                                ('\0' == (*bufp)[i] ||
                    825:                                 ' ' == (*bufp)[i] ||
                    826:                                 '\t' == (*bufp)[i])) {
                    827:                        roffnode_pop(r);
                    828:                        roffnode_cleanscope(r);
1.1       schwarze  829:
1.35      schwarze  830:                        while (' ' == (*bufp)[i] || '\t' == (*bufp)[i])
                    831:                                i++;
                    832:
                    833:                        pos = i;
1.16      schwarze  834:                        if (ROFF_MAX != roff_parse(r, *bufp, &pos))
1.2       schwarze  835:                                return(ROFF_RERUN);
                    836:                        return(ROFF_IGN);
                    837:                }
1.1       schwarze  838:        }
                    839:
1.2       schwarze  840:        /*
                    841:         * If we have no custom end-query or lookup failed, then try
                    842:         * pulling it out of the hashtable.
                    843:         */
1.1       schwarze  844:
1.36      schwarze  845:        t = roff_parse(r, *bufp, &pos);
1.1       schwarze  846:
1.16      schwarze  847:        /*
                    848:         * Macros other than block-end are only significant
                    849:         * in `de' blocks; elsewhere, simply throw them away.
                    850:         */
                    851:        if (ROFF_cblock != t) {
                    852:                if (ROFF_de == tok)
                    853:                        roff_setstr(r, r->last->name, *bufp + ppos, 1);
1.1       schwarze  854:                return(ROFF_IGN);
1.16      schwarze  855:        }
1.1       schwarze  856:
1.2       schwarze  857:        assert(roffs[t].proc);
1.6       schwarze  858:        return((*roffs[t].proc)(r, t, bufp, szp,
                    859:                                ln, ppos, pos, offs));
1.2       schwarze  860: }
                    861:
                    862:
                    863: /* ARGSUSED */
                    864: static enum rofferr
                    865: roff_block_text(ROFF_ARGS)
                    866: {
                    867:
1.16      schwarze  868:        if (ROFF_de == tok)
                    869:                roff_setstr(r, r->last->name, *bufp + pos, 1);
                    870:
1.2       schwarze  871:        return(ROFF_IGN);
                    872: }
                    873:
                    874:
                    875: /* ARGSUSED */
                    876: static enum rofferr
                    877: roff_cond_sub(ROFF_ARGS)
                    878: {
                    879:        enum rofft       t;
                    880:        enum roffrule    rr;
1.37      schwarze  881:        char            *ep;
1.2       schwarze  882:
                    883:        rr = r->last->rule;
1.37      schwarze  884:        roffnode_cleanscope(r);
1.2       schwarze  885:
1.37      schwarze  886:        /*
                    887:         * If the macro is unknown, first check if it contains a closing
                    888:         * delimiter `\}'.  If it does, close out our scope and return
                    889:         * the currently-scoped rule (ignore or continue).  Else, drop
                    890:         * into the currently-scoped rule.
1.5       schwarze  891:         */
                    892:
1.16      schwarze  893:        if (ROFF_MAX == (t = roff_parse(r, *bufp, &pos))) {
1.37      schwarze  894:                ep = &(*bufp)[pos];
                    895:                for ( ; NULL != (ep = strchr(ep, '\\')); ep++) {
                    896:                        ep++;
                    897:                        if ('}' != *ep)
                    898:                                continue;
1.39      schwarze  899:
                    900:                        /*
                    901:                         * Make the \} go away.
                    902:                         * This is a little haphazard, as it's not quite
                    903:                         * clear how nroff does this.
                    904:                         * If we're at the end of line, then just chop
                    905:                         * off the \} and resize the buffer.
                    906:                         * If we aren't, then conver it to spaces.
                    907:                         */
                    908:
                    909:                        if ('\0' == *(ep + 1)) {
                    910:                                *--ep = '\0';
                    911:                                *szp -= 2;
                    912:                        } else
                    913:                                *(ep - 1) = *ep = ' ';
                    914:
1.37      schwarze  915:                        roff_ccond(r, ROFF_ccond, bufp, szp,
                    916:                                        ln, pos, pos + 2, offs);
                    917:                        break;
                    918:                }
1.2       schwarze  919:                return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
1.12      schwarze  920:        }
1.2       schwarze  921:
                    922:        /*
                    923:         * A denied conditional must evaluate its children if and only
                    924:         * if they're either structurally required (such as loops and
                    925:         * conditionals) or a closing macro.
                    926:         */
1.37      schwarze  927:
1.2       schwarze  928:        if (ROFFRULE_DENY == rr)
                    929:                if ( ! (ROFFMAC_STRUCT & roffs[t].flags))
                    930:                        if (ROFF_ccond != t)
                    931:                                return(ROFF_IGN);
                    932:
                    933:        assert(roffs[t].proc);
1.6       schwarze  934:        return((*roffs[t].proc)(r, t, bufp, szp,
                    935:                                ln, ppos, pos, offs));
1.2       schwarze  936: }
                    937:
                    938: /* ARGSUSED */
                    939: static enum rofferr
                    940: roff_cond_text(ROFF_ARGS)
                    941: {
1.37      schwarze  942:        char            *ep;
1.2       schwarze  943:        enum roffrule    rr;
                    944:
                    945:        rr = r->last->rule;
1.37      schwarze  946:        roffnode_cleanscope(r);
1.1       schwarze  947:
1.37      schwarze  948:        ep = &(*bufp)[pos];
                    949:        for ( ; NULL != (ep = strchr(ep, '\\')); ep++) {
                    950:                ep++;
                    951:                if ('}' != *ep)
                    952:                        continue;
                    953:                *ep = '&';
                    954:                roff_ccond(r, ROFF_ccond, bufp, szp,
                    955:                                ln, pos, pos + 2, offs);
1.2       schwarze  956:        }
                    957:        return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
                    958: }
                    959:
1.5       schwarze  960: static enum roffrule
                    961: roff_evalcond(const char *v, int *pos)
                    962: {
                    963:
                    964:        switch (v[*pos]) {
                    965:        case ('n'):
                    966:                (*pos)++;
                    967:                return(ROFFRULE_ALLOW);
                    968:        case ('e'):
                    969:                /* FALLTHROUGH */
                    970:        case ('o'):
                    971:                /* FALLTHROUGH */
                    972:        case ('t'):
                    973:                (*pos)++;
                    974:                return(ROFFRULE_DENY);
                    975:        default:
                    976:                break;
                    977:        }
                    978:
                    979:        while (v[*pos] && ' ' != v[*pos])
                    980:                (*pos)++;
                    981:        return(ROFFRULE_DENY);
                    982: }
                    983:
1.2       schwarze  984: /* ARGSUSED */
                    985: static enum rofferr
1.21      schwarze  986: roff_line_ignore(ROFF_ARGS)
1.6       schwarze  987: {
1.30      schwarze  988:
                    989:        if (ROFF_it == tok)
1.35      schwarze  990:                mandoc_msg(MANDOCERR_REQUEST, r->parse, ln, ppos, "it");
1.6       schwarze  991:
1.21      schwarze  992:        return(ROFF_IGN);
                    993: }
                    994:
                    995: /* ARGSUSED */
                    996: static enum rofferr
1.2       schwarze  997: roff_cond(ROFF_ARGS)
                    998: {
                    999:        int              sv;
1.5       schwarze 1000:        enum roffrule    rule;
1.2       schwarze 1001:
1.35      schwarze 1002:        /*
                   1003:         * An `.el' has no conditional body: it will consume the value
                   1004:         * of the current rstack entry set in prior `ie' calls or
                   1005:         * defaults to DENY.
                   1006:         *
                   1007:         * If we're not an `el', however, then evaluate the conditional.
                   1008:         */
1.1       schwarze 1009:
1.35      schwarze 1010:        rule = ROFF_el == tok ?
                   1011:                (r->rstackpos < 0 ?
                   1012:                 ROFFRULE_DENY : r->rstack[r->rstackpos--]) :
                   1013:                roff_evalcond(*bufp, &pos);
1.2       schwarze 1014:
                   1015:        sv = pos;
                   1016:        while (' ' == (*bufp)[pos])
                   1017:                pos++;
                   1018:
                   1019:        /*
                   1020:         * Roff is weird.  If we have just white-space after the
                   1021:         * conditional, it's considered the BODY and we exit without
                   1022:         * really doing anything.  Warn about this.  It's probably
                   1023:         * wrong.
                   1024:         */
1.5       schwarze 1025:
1.2       schwarze 1026:        if ('\0' == (*bufp)[pos] && sv != pos) {
1.35      schwarze 1027:                mandoc_msg(MANDOCERR_NOARGS, r->parse, ln, ppos, NULL);
1.22      schwarze 1028:                return(ROFF_IGN);
1.2       schwarze 1029:        }
                   1030:
1.16      schwarze 1031:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 1032:
1.5       schwarze 1033:        r->last->rule = rule;
1.2       schwarze 1034:
1.35      schwarze 1035:        /*
                   1036:         * An if-else will put the NEGATION of the current evaluated
                   1037:         * conditional into the stack of rules.
                   1038:         */
                   1039:
1.2       schwarze 1040:        if (ROFF_ie == tok) {
1.35      schwarze 1041:                if (r->rstackpos == RSTACK_MAX - 1) {
                   1042:                        mandoc_msg(MANDOCERR_MEM,
                   1043:                                r->parse, ln, ppos, NULL);
                   1044:                        return(ROFF_ERR);
                   1045:                }
                   1046:                r->rstack[++r->rstackpos] =
                   1047:                        ROFFRULE_DENY == r->last->rule ?
                   1048:                        ROFFRULE_ALLOW : ROFFRULE_DENY;
1.2       schwarze 1049:        }
1.5       schwarze 1050:
                   1051:        /* If the parent has false as its rule, then so do we. */
                   1052:
1.2       schwarze 1053:        if (r->last->parent && ROFFRULE_DENY == r->last->parent->rule)
                   1054:                r->last->rule = ROFFRULE_DENY;
1.5       schwarze 1055:
                   1056:        /*
                   1057:         * Determine scope.  If we're invoked with "\{" trailing the
                   1058:         * conditional, then we're in a multiline scope.  Else our scope
                   1059:         * expires on the next line.
                   1060:         */
1.2       schwarze 1061:
                   1062:        r->last->endspan = 1;
                   1063:
                   1064:        if ('\\' == (*bufp)[pos] && '{' == (*bufp)[pos + 1]) {
                   1065:                r->last->endspan = -1;
                   1066:                pos += 2;
                   1067:        }
                   1068:
                   1069:        /*
                   1070:         * If there are no arguments on the line, the next-line scope is
                   1071:         * assumed.
                   1072:         */
                   1073:
                   1074:        if ('\0' == (*bufp)[pos])
                   1075:                return(ROFF_IGN);
                   1076:
                   1077:        /* Otherwise re-run the roff parser after recalculating. */
1.1       schwarze 1078:
1.2       schwarze 1079:        *offs = pos;
                   1080:        return(ROFF_RERUN);
1.1       schwarze 1081: }
                   1082:
                   1083:
1.2       schwarze 1084: /* ARGSUSED */
                   1085: static enum rofferr
1.7       schwarze 1086: roff_ds(ROFF_ARGS)
                   1087: {
1.10      schwarze 1088:        char            *name, *string;
                   1089:
                   1090:        /*
                   1091:         * A symbol is named by the first word following the macro
                   1092:         * invocation up to a space.  Its value is anything after the
                   1093:         * name's trailing whitespace and optional double-quote.  Thus,
                   1094:         *
                   1095:         *  [.ds foo "bar  "     ]
                   1096:         *
                   1097:         * will have `bar  "     ' as its value.
                   1098:         */
1.7       schwarze 1099:
1.28      schwarze 1100:        string = *bufp + pos;
                   1101:        name = roff_getname(r, &string, ln, pos);
1.7       schwarze 1102:        if ('\0' == *name)
                   1103:                return(ROFF_IGN);
                   1104:
1.28      schwarze 1105:        /* Read past initial double-quote. */
                   1106:        if ('"' == *string)
1.7       schwarze 1107:                string++;
                   1108:
1.10      schwarze 1109:        /* The rest is the value. */
1.16      schwarze 1110:        roff_setstr(r, name, string, 0);
1.7       schwarze 1111:        return(ROFF_IGN);
                   1112: }
                   1113:
1.41    ! schwarze 1114: int
        !          1115: roff_regisset(const struct roff *r, enum regs reg)
        !          1116: {
        !          1117:
        !          1118:        return(r->regs[(int)reg].set);
        !          1119: }
        !          1120:
        !          1121: unsigned int
        !          1122: roff_regget(const struct roff *r, enum regs reg)
        !          1123: {
        !          1124:
        !          1125:        return(r->regs[(int)reg].u);
        !          1126: }
        !          1127:
        !          1128: void
        !          1129: roff_regunset(struct roff *r, enum regs reg)
        !          1130: {
        !          1131:
        !          1132:        r->regs[(int)reg].set = 0;
        !          1133: }
1.7       schwarze 1134:
                   1135: /* ARGSUSED */
                   1136: static enum rofferr
1.6       schwarze 1137: roff_nr(ROFF_ARGS)
1.1       schwarze 1138: {
1.28      schwarze 1139:        const char      *key;
                   1140:        char            *val;
1.37      schwarze 1141:        int              iv;
1.6       schwarze 1142:
1.28      schwarze 1143:        val = *bufp + pos;
                   1144:        key = roff_getname(r, &val, ln, pos);
1.6       schwarze 1145:
                   1146:        if (0 == strcmp(key, "nS")) {
1.41    ! schwarze 1147:                r->regs[(int)REG_nS].set = 1;
        !          1148:                if ((iv = mandoc_strntoi(val, strlen(val), 10)) >= 0)
        !          1149:                        r->regs[(int)REG_nS].u = (unsigned)iv;
1.37      schwarze 1150:                else
1.41    ! schwarze 1151:                        r->regs[(int)REG_nS].u = 0u;
1.6       schwarze 1152:        }
1.1       schwarze 1153:
1.29      schwarze 1154:        return(ROFF_IGN);
                   1155: }
                   1156:
                   1157: /* ARGSUSED */
                   1158: static enum rofferr
                   1159: roff_rm(ROFF_ARGS)
                   1160: {
                   1161:        const char       *name;
                   1162:        char             *cp;
                   1163:
                   1164:        cp = *bufp + pos;
                   1165:        while ('\0' != *cp) {
1.34      schwarze 1166:                name = roff_getname(r, &cp, ln, (int)(cp - *bufp));
1.29      schwarze 1167:                if ('\0' != *name)
                   1168:                        roff_setstr(r, name, NULL, 0);
                   1169:        }
1.2       schwarze 1170:        return(ROFF_IGN);
1.14      schwarze 1171: }
                   1172:
                   1173: /* ARGSUSED */
                   1174: static enum rofferr
1.27      schwarze 1175: roff_TE(ROFF_ARGS)
                   1176: {
                   1177:
                   1178:        if (NULL == r->tbl)
1.35      schwarze 1179:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.27      schwarze 1180:        else
1.41    ! schwarze 1181:                tbl_end(&r->tbl);
1.27      schwarze 1182:
                   1183:        return(ROFF_IGN);
                   1184: }
                   1185:
                   1186: /* ARGSUSED */
                   1187: static enum rofferr
                   1188: roff_T_(ROFF_ARGS)
                   1189: {
                   1190:
                   1191:        if (NULL == r->tbl)
1.35      schwarze 1192:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.27      schwarze 1193:        else
                   1194:                tbl_restart(ppos, ln, r->tbl);
                   1195:
                   1196:        return(ROFF_IGN);
                   1197: }
                   1198:
1.41    ! schwarze 1199: #if 0
        !          1200: static int
        !          1201: roff_closeeqn(struct roff *r)
        !          1202: {
        !          1203:
        !          1204:        return(r->eqn && ROFF_EQN == eqn_end(&r->eqn) ? 1 : 0);
        !          1205: }
        !          1206: #endif
        !          1207:
        !          1208: static void
        !          1209: roff_openeqn(struct roff *r, const char *name, int line,
        !          1210:                int offs, const char *buf)
1.32      schwarze 1211: {
1.41    ! schwarze 1212:        struct eqn_node *e;
        !          1213:        int              poff;
1.32      schwarze 1214:
                   1215:        assert(NULL == r->eqn);
1.41    ! schwarze 1216:        e = eqn_alloc(name, offs, line, r->parse);
1.32      schwarze 1217:
                   1218:        if (r->last_eqn)
                   1219:                r->last_eqn->next = e;
                   1220:        else
                   1221:                r->first_eqn = r->last_eqn = e;
                   1222:
                   1223:        r->eqn = r->last_eqn = e;
1.41    ! schwarze 1224:
        !          1225:        if (buf) {
        !          1226:                poff = 0;
        !          1227:                eqn_read(&r->eqn, line, buf, offs, &poff);
        !          1228:        }
        !          1229: }
        !          1230:
        !          1231: /* ARGSUSED */
        !          1232: static enum rofferr
        !          1233: roff_EQ(ROFF_ARGS)
        !          1234: {
        !          1235:
        !          1236:        roff_openeqn(r, *bufp + pos, ln, ppos, NULL);
1.32      schwarze 1237:        return(ROFF_IGN);
                   1238: }
                   1239:
                   1240: /* ARGSUSED */
                   1241: static enum rofferr
                   1242: roff_EN(ROFF_ARGS)
                   1243: {
                   1244:
1.35      schwarze 1245:        mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.32      schwarze 1246:        return(ROFF_IGN);
                   1247: }
                   1248:
                   1249: /* ARGSUSED */
                   1250: static enum rofferr
1.27      schwarze 1251: roff_TS(ROFF_ARGS)
                   1252: {
                   1253:        struct tbl_node *t;
                   1254:
                   1255:        if (r->tbl) {
1.35      schwarze 1256:                mandoc_msg(MANDOCERR_SCOPEBROKEN, r->parse, ln, ppos, NULL);
1.41    ! schwarze 1257:                tbl_end(&r->tbl);
1.27      schwarze 1258:        }
                   1259:
1.35      schwarze 1260:        t = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 1261:
                   1262:        if (r->last_tbl)
                   1263:                r->last_tbl->next = t;
                   1264:        else
                   1265:                r->first_tbl = r->last_tbl = t;
                   1266:
                   1267:        r->tbl = r->last_tbl = t;
                   1268:        return(ROFF_IGN);
                   1269: }
                   1270:
                   1271: /* ARGSUSED */
                   1272: static enum rofferr
1.14      schwarze 1273: roff_so(ROFF_ARGS)
                   1274: {
                   1275:        char *name;
1.15      schwarze 1276:
1.35      schwarze 1277:        mandoc_msg(MANDOCERR_SO, r->parse, ln, ppos, NULL);
1.14      schwarze 1278:
1.22      schwarze 1279:        /*
                   1280:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   1281:         * opening anything that's not in our cwd or anything beneath
                   1282:         * it.  Thus, explicitly disallow traversing up the file-system
                   1283:         * or using absolute paths.
                   1284:         */
                   1285:
1.14      schwarze 1286:        name = *bufp + pos;
                   1287:        if ('/' == *name || strstr(name, "../") || strstr(name, "/..")) {
1.35      schwarze 1288:                mandoc_msg(MANDOCERR_SOPATH, r->parse, ln, pos, NULL);
1.14      schwarze 1289:                return(ROFF_ERR);
                   1290:        }
                   1291:
                   1292:        *offs = pos;
                   1293:        return(ROFF_SO);
1.7       schwarze 1294: }
                   1295:
1.16      schwarze 1296: /* ARGSUSED */
                   1297: static enum rofferr
                   1298: roff_userdef(ROFF_ARGS)
1.12      schwarze 1299: {
1.16      schwarze 1300:        const char       *arg[9];
                   1301:        char             *cp, *n1, *n2;
1.25      schwarze 1302:        int               i;
1.12      schwarze 1303:
1.16      schwarze 1304:        /*
                   1305:         * Collect pointers to macro argument strings
                   1306:         * and null-terminate them.
                   1307:         */
                   1308:        cp = *bufp + pos;
1.25      schwarze 1309:        for (i = 0; i < 9; i++)
1.26      schwarze 1310:                arg[i] = '\0' == *cp ? "" :
1.35      schwarze 1311:                    mandoc_getarg(r->parse, &cp, ln, &pos);
1.16      schwarze 1312:
                   1313:        /*
                   1314:         * Expand macro arguments.
1.12      schwarze 1315:         */
1.16      schwarze 1316:        *szp = 0;
                   1317:        n1 = cp = mandoc_strdup(r->current_string);
                   1318:        while (NULL != (cp = strstr(cp, "\\$"))) {
                   1319:                i = cp[2] - '1';
                   1320:                if (0 > i || 8 < i) {
                   1321:                        /* Not an argument invocation. */
                   1322:                        cp += 2;
                   1323:                        continue;
                   1324:                }
                   1325:
                   1326:                *szp = strlen(n1) - 3 + strlen(arg[i]) + 1;
                   1327:                n2 = mandoc_malloc(*szp);
                   1328:
                   1329:                strlcpy(n2, n1, (size_t)(cp - n1 + 1));
                   1330:                strlcat(n2, arg[i], *szp);
                   1331:                strlcat(n2, cp + 3, *szp);
                   1332:
                   1333:                cp = n2 + (cp - n1);
                   1334:                free(n1);
                   1335:                n1 = n2;
1.12      schwarze 1336:        }
                   1337:
1.16      schwarze 1338:        /*
                   1339:         * Replace the macro invocation
                   1340:         * by the expanded macro.
                   1341:         */
                   1342:        free(*bufp);
                   1343:        *bufp = n1;
                   1344:        if (0 == *szp)
                   1345:                *szp = strlen(*bufp) + 1;
                   1346:
1.19      schwarze 1347:        return(*szp > 1 && '\n' == (*bufp)[(int)*szp - 2] ?
1.16      schwarze 1348:           ROFF_REPARSE : ROFF_APPEND);
1.12      schwarze 1349: }
1.28      schwarze 1350:
                   1351: static char *
                   1352: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   1353: {
                   1354:        char     *name, *cp;
                   1355:
                   1356:        name = *cpp;
                   1357:        if ('\0' == *name)
                   1358:                return(name);
                   1359:
                   1360:        /* Read until end of name. */
                   1361:        for (cp = name; '\0' != *cp && ' ' != *cp; cp++) {
                   1362:                if ('\\' != *cp)
                   1363:                        continue;
                   1364:                cp++;
                   1365:                if ('\\' == *cp)
                   1366:                        continue;
1.35      schwarze 1367:                mandoc_msg(MANDOCERR_NAMESC, r->parse, ln, pos, NULL);
1.28      schwarze 1368:                *cp = '\0';
                   1369:                name = cp;
                   1370:        }
                   1371:
                   1372:        /* Nil-terminate name. */
                   1373:        if ('\0' != *cp)
                   1374:                *(cp++) = '\0';
                   1375:
                   1376:        /* Read past spaces. */
                   1377:        while (' ' == *cp)
                   1378:                cp++;
                   1379:
                   1380:        *cpp = cp;
                   1381:        return(name);
                   1382: }
                   1383:
1.16      schwarze 1384: /*
                   1385:  * Store *string into the user-defined string called *name.
                   1386:  * In multiline mode, append to an existing entry and append '\n';
                   1387:  * else replace the existing entry, if there is one.
                   1388:  * To clear an existing entry, call with (*r, *name, NULL, 0).
                   1389:  */
1.8       schwarze 1390: static void
1.16      schwarze 1391: roff_setstr(struct roff *r, const char *name, const char *string,
                   1392:        int multiline)
1.7       schwarze 1393: {
                   1394:        struct roffstr   *n;
1.16      schwarze 1395:        char             *c;
                   1396:        size_t            oldch, newch;
1.40      schwarze 1397:
                   1398:        /* XXX workaround for the Perl preamble until we get .tr */
                   1399:        if ( ! strcmp(name, "--")) {
                   1400:                string = "--";
                   1401:                multiline = 0;
                   1402:        }
1.7       schwarze 1403:
1.16      schwarze 1404:        /* Search for an existing string with the same name. */
1.8       schwarze 1405:        n = r->first_string;
1.7       schwarze 1406:        while (n && strcmp(name, n->name))
                   1407:                n = n->next;
1.8       schwarze 1408:
                   1409:        if (NULL == n) {
1.16      schwarze 1410:                /* Create a new string table entry. */
1.8       schwarze 1411:                n = mandoc_malloc(sizeof(struct roffstr));
1.16      schwarze 1412:                n->name = mandoc_strdup(name);
                   1413:                n->string = NULL;
1.8       schwarze 1414:                n->next = r->first_string;
                   1415:                r->first_string = n;
1.16      schwarze 1416:        } else if (0 == multiline) {
                   1417:                /* In multiline mode, append; else replace. */
1.7       schwarze 1418:                free(n->string);
1.16      schwarze 1419:                n->string = NULL;
                   1420:        }
                   1421:
                   1422:        if (NULL == string)
                   1423:                return;
                   1424:
                   1425:        /*
                   1426:         * One additional byte for the '\n' in multiline mode,
                   1427:         * and one for the terminating '\0'.
                   1428:         */
1.34      schwarze 1429:        newch = strlen(string) + (multiline ? 2u : 1u);
1.16      schwarze 1430:        if (NULL == n->string) {
                   1431:                n->string = mandoc_malloc(newch);
                   1432:                *n->string = '\0';
                   1433:                oldch = 0;
                   1434:        } else {
                   1435:                oldch = strlen(n->string);
                   1436:                n->string = mandoc_realloc(n->string, oldch + newch);
                   1437:        }
                   1438:
                   1439:        /* Skip existing content in the destination buffer. */
1.34      schwarze 1440:        c = n->string + (int)oldch;
1.16      schwarze 1441:
                   1442:        /* Append new content to the destination buffer. */
                   1443:        while (*string) {
                   1444:                /*
                   1445:                 * Rudimentary roff copy mode:
                   1446:                 * Handle escaped backslashes.
                   1447:                 */
                   1448:                if ('\\' == *string && '\\' == *(string + 1))
                   1449:                        string++;
                   1450:                *c++ = *string++;
                   1451:        }
1.8       schwarze 1452:
1.16      schwarze 1453:        /* Append terminating bytes. */
                   1454:        if (multiline)
                   1455:                *c++ = '\n';
                   1456:        *c = '\0';
1.7       schwarze 1457: }
                   1458:
1.8       schwarze 1459: static const char *
                   1460: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7       schwarze 1461: {
1.8       schwarze 1462:        const struct roffstr *n;
1.7       schwarze 1463:
1.8       schwarze 1464:        n = r->first_string;
1.10      schwarze 1465:        while (n && (strncmp(name, n->name, len) || '\0' != n->name[(int)len]))
1.7       schwarze 1466:                n = n->next;
1.8       schwarze 1467:
                   1468:        return(n ? n->string : NULL);
1.7       schwarze 1469: }
                   1470:
1.8       schwarze 1471: static void
                   1472: roff_freestr(struct roff *r)
1.7       schwarze 1473: {
                   1474:        struct roffstr   *n, *nn;
                   1475:
1.8       schwarze 1476:        for (n = r->first_string; n; n = nn) {
1.7       schwarze 1477:                free(n->name);
                   1478:                free(n->string);
                   1479:                nn = n->next;
                   1480:                free(n);
                   1481:        }
1.8       schwarze 1482:
                   1483:        r->first_string = NULL;
1.27      schwarze 1484: }
                   1485:
                   1486: const struct tbl_span *
                   1487: roff_span(const struct roff *r)
                   1488: {
                   1489:
                   1490:        return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32      schwarze 1491: }
                   1492:
                   1493: const struct eqn *
                   1494: roff_eqn(const struct roff *r)
                   1495: {
                   1496:
                   1497:        return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.1       schwarze 1498: }