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

Annotation of mandoc/roff.c, Revision 1.238

1.238   ! schwarze    1: /*     $Id: roff.c,v 1.237 2014/10/28 17:36:19 schwarze Exp $ */
1.1       kristaps    2: /*
1.175     schwarze    3:  * Copyright (c) 2010, 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv>
1.190     schwarze    4:  * Copyright (c) 2010-2014 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: #include "config.h"
1.225     schwarze   19:
                     20: #include <sys/types.h>
1.30      kristaps   21:
1.67      kristaps   22: #include <assert.h>
1.85      kristaps   23: #include <ctype.h>
1.178     schwarze   24: #include <stdio.h>
1.1       kristaps   25: #include <stdlib.h>
1.67      kristaps   26: #include <string.h>
1.1       kristaps   27:
1.67      kristaps   28: #include "mandoc.h"
1.201     schwarze   29: #include "mandoc_aux.h"
1.226     schwarze   30: #include "libmandoc.h"
1.109     kristaps   31: #include "libroff.h"
1.33      kristaps   32:
1.141     kristaps   33: /* Maximum number of nested if-else conditionals. */
1.82      kristaps   34: #define        RSTACK_MAX      128
                     35:
1.170     schwarze   36: /* Maximum number of string expansions per line, to break infinite loops. */
                     37: #define        EXPAND_LIMIT    1000
                     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,
1.193     schwarze   44:        ROFF_as,
1.174     kristaps   45:        ROFF_cc,
1.194     schwarze   46:        ROFF_ce,
1.80      kristaps   47:        ROFF_de,
                     48:        ROFF_dei,
                     49:        ROFF_de1,
1.83      schwarze   50:        ROFF_ds,
1.82      kristaps   51:        ROFF_el,
1.185     schwarze   52:        ROFF_fam,
1.186     schwarze   53:        ROFF_hw,
1.103     kristaps   54:        ROFF_hy,
1.82      kristaps   55:        ROFF_ie,
1.75      kristaps   56:        ROFF_if,
1.76      kristaps   57:        ROFF_ig,
1.123     schwarze   58:        ROFF_it,
1.103     kristaps   59:        ROFF_ne,
                     60:        ROFF_nh,
1.104     kristaps   61:        ROFF_nr,
1.124     schwarze   62:        ROFF_ns,
1.229     schwarze   63:        ROFF_pl,
1.124     schwarze   64:        ROFF_ps,
1.83      schwarze   65:        ROFF_rm,
1.203     schwarze   66:        ROFF_rr,
1.105     kristaps   67:        ROFF_so,
1.124     schwarze   68:        ROFF_ta,
1.83      schwarze   69:        ROFF_tr,
1.175     schwarze   70:        ROFF_Dd,
                     71:        ROFF_TH,
1.109     kristaps   72:        ROFF_TS,
                     73:        ROFF_TE,
1.112     kristaps   74:        ROFF_T_,
1.125     kristaps   75:        ROFF_EQ,
                     76:        ROFF_EN,
1.76      kristaps   77:        ROFF_cblock,
1.106     kristaps   78:        ROFF_USERDEF,
1.67      kristaps   79:        ROFF_MAX
                     80: };
                     81:
1.147     kristaps   82: /*
1.167     kristaps   83:  * An incredibly-simple string buffer.
                     84:  */
1.94      kristaps   85: struct roffstr {
1.167     kristaps   86:        char            *p; /* nil-terminated buffer */
                     87:        size_t           sz; /* saved strlen(p) */
1.166     kristaps   88: };
                     89:
                     90: /*
1.167     kristaps   91:  * A key-value roffstr pair as part of a singly-linked list.
1.166     kristaps   92:  */
                     93: struct roffkv {
                     94:        struct roffstr   key;
                     95:        struct roffstr   val;
                     96:        struct roffkv   *next; /* next in list */
1.94      kristaps   97: };
                     98:
1.180     schwarze   99: /*
                    100:  * A single number register as part of a singly-linked list.
                    101:  */
                    102: struct roffreg {
                    103:        struct roffstr   key;
1.181     schwarze  104:        int              val;
1.180     schwarze  105:        struct roffreg  *next;
                    106: };
                    107:
1.67      kristaps  108: struct roff {
1.128     kristaps  109:        struct mparse   *parse; /* parse point */
1.237     schwarze  110:        const struct mchars *mchars; /* character table */
1.67      kristaps  111:        struct roffnode *last; /* leaf of stack */
1.223     schwarze  112:        int             *rstack; /* stack of inverted `ie' values */
1.180     schwarze  113:        struct roffreg  *regtab; /* number registers */
1.166     kristaps  114:        struct roffkv   *strtab; /* user-defined strings & macros */
1.167     kristaps  115:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                    116:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.106     kristaps  117:        const char      *current_string; /* value of last called user macro */
1.118     kristaps  118:        struct tbl_node *first_tbl; /* first table parsed */
                    119:        struct tbl_node *last_tbl; /* last table parsed */
                    120:        struct tbl_node *tbl; /* current table being parsed */
1.125     kristaps  121:        struct eqn_node *last_eqn; /* last equation parsed */
                    122:        struct eqn_node *first_eqn; /* first equation parsed */
                    123:        struct eqn_node *eqn; /* current equation being parsed */
1.230     schwarze  124:        int              eqn_inline; /* current equation is inline */
1.223     schwarze  125:        int              options; /* parse options */
                    126:        int              rstacksz; /* current size limit of rstack */
                    127:        int              rstackpos; /* position in rstack */
1.227     schwarze  128:        int              format; /* current file in mdoc or man format */
1.223     schwarze  129:        char             control; /* control character */
1.79      kristaps  130: };
                    131:
1.67      kristaps  132: struct roffnode {
                    133:        enum rofft       tok; /* type of node */
                    134:        struct roffnode *parent; /* up one in stack */
                    135:        int              line; /* parse line */
                    136:        int              col; /* parse col */
1.106     kristaps  137:        char            *name; /* node name, e.g. macro name */
1.79      kristaps  138:        char            *end; /* end-rules: custom token */
                    139:        int              endspan; /* end-rules: next-line or infty */
1.198     schwarze  140:        int              rule; /* current evaluation rule */
1.67      kristaps  141: };
                    142:
                    143: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.72      kristaps  144:                         enum rofft tok, /* tok of macro */ \
1.238   ! schwarze  145:                         struct buf *buf, /* input buffer */ \
1.67      kristaps  146:                         int ln, /* parse line */ \
1.75      kristaps  147:                         int ppos, /* original pos in buffer */ \
                    148:                         int pos, /* current pos in buffer */ \
1.74      kristaps  149:                         int *offs /* reset offset of buffer data */
1.67      kristaps  150:
                    151: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    152:
                    153: struct roffmac {
                    154:        const char      *name; /* macro name */
1.79      kristaps  155:        roffproc         proc; /* process new macro */
                    156:        roffproc         text; /* process as child text of macro */
                    157:        roffproc         sub; /* process as child of macro */
                    158:        int              flags;
                    159: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.85      kristaps  160:        struct roffmac  *next;
1.67      kristaps  161: };
                    162:
1.141     kristaps  163: struct predef {
                    164:        const char      *name; /* predefined input name */
                    165:        const char      *str; /* replacement symbol */
                    166: };
                    167:
                    168: #define        PREDEF(__name, __str) \
                    169:        { (__name), (__str) },
                    170:
1.155     kristaps  171: static enum rofft       roffhash_find(const char *, size_t);
                    172: static void             roffhash_init(void);
                    173: static void             roffnode_cleanscope(struct roff *);
                    174: static void             roffnode_pop(struct roff *);
                    175: static void             roffnode_push(struct roff *, enum rofft,
                    176:                                const char *, int, int);
1.80      kristaps  177: static enum rofferr     roff_block(ROFF_ARGS);
                    178: static enum rofferr     roff_block_text(ROFF_ARGS);
                    179: static enum rofferr     roff_block_sub(ROFF_ARGS);
                    180: static enum rofferr     roff_cblock(ROFF_ARGS);
1.174     kristaps  181: static enum rofferr     roff_cc(ROFF_ARGS);
1.195     schwarze  182: static void             roff_ccond(struct roff *, int, int);
1.82      kristaps  183: static enum rofferr     roff_cond(ROFF_ARGS);
                    184: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    185: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.92      schwarze  186: static enum rofferr     roff_ds(ROFF_ARGS);
1.238   ! schwarze  187: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.234     kristaps  188: static int              roff_evalcond(struct roff *r, int,
                    189:                                const char *, int *);
                    190: static int              roff_evalnum(struct roff *, int,
                    191:                                const char *, int *, int *, int);
                    192: static int              roff_evalpar(struct roff *, int,
                    193:                                const char *, int *, int *);
1.198     schwarze  194: static int              roff_evalstrcond(const char *, int *);
1.155     kristaps  195: static void             roff_free1(struct roff *);
1.180     schwarze  196: static void             roff_freereg(struct roffreg *);
1.167     kristaps  197: static void             roff_freestr(struct roffkv *);
1.212     schwarze  198: static size_t           roff_getname(struct roff *, char **, int, int);
1.184     schwarze  199: static int              roff_getnum(const char *, int *, int *);
                    200: static int              roff_getop(const char *, int *, char *);
1.181     schwarze  201: static int              roff_getregn(const struct roff *,
                    202:                                const char *, size_t);
1.192     schwarze  203: static int              roff_getregro(const char *name);
1.207     schwarze  204: static const char      *roff_getstrn(const struct roff *,
1.94      kristaps  205:                                const char *, size_t);
1.178     schwarze  206: static enum rofferr     roff_it(ROFF_ARGS);
1.103     kristaps  207: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.89      kristaps  208: static enum rofferr     roff_nr(ROFF_ARGS);
1.214     schwarze  209: static enum rofft       roff_parse(struct roff *, char *, int *,
                    210:                                int, int);
1.238   ! schwarze  211: static enum rofferr     roff_parsetext(struct buf *, int, int *);
        !           212: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.122     schwarze  213: static enum rofferr     roff_rm(ROFF_ARGS);
1.203     schwarze  214: static enum rofferr     roff_rr(ROFF_ARGS);
1.94      kristaps  215: static void             roff_setstr(struct roff *,
1.106     kristaps  216:                                const char *, const char *, int);
1.207     schwarze  217: static void             roff_setstrn(struct roffkv **, const char *,
1.164     kristaps  218:                                size_t, const char *, size_t, int);
1.105     kristaps  219: static enum rofferr     roff_so(ROFF_ARGS);
1.164     kristaps  220: static enum rofferr     roff_tr(ROFF_ARGS);
1.175     schwarze  221: static enum rofferr     roff_Dd(ROFF_ARGS);
                    222: static enum rofferr     roff_TH(ROFF_ARGS);
1.109     kristaps  223: static enum rofferr     roff_TE(ROFF_ARGS);
                    224: static enum rofferr     roff_TS(ROFF_ARGS);
1.125     kristaps  225: static enum rofferr     roff_EQ(ROFF_ARGS);
                    226: static enum rofferr     roff_EN(ROFF_ARGS);
1.112     kristaps  227: static enum rofferr     roff_T_(ROFF_ARGS);
1.106     kristaps  228: static enum rofferr     roff_userdef(ROFF_ARGS);
1.67      kristaps  229:
1.155     kristaps  230: /* See roffhash_find() */
1.85      kristaps  231:
                    232: #define        ASCII_HI         126
                    233: #define        ASCII_LO         33
                    234: #define        HASHWIDTH       (ASCII_HI - ASCII_LO + 1)
                    235:
                    236: static struct roffmac  *hash[HASHWIDTH];
                    237:
                    238: static struct roffmac   roffs[ROFF_MAX] = {
1.103     kristaps  239:        { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.85      kristaps  240:        { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    241:        { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    242:        { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.193     schwarze  243:        { "as", roff_ds, NULL, NULL, 0, NULL },
1.174     kristaps  244:        { "cc", roff_cc, NULL, NULL, 0, NULL },
1.194     schwarze  245:        { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.85      kristaps  246:        { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    247:        { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    248:        { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.92      schwarze  249:        { "ds", roff_ds, NULL, NULL, 0, NULL },
1.85      kristaps  250:        { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.185     schwarze  251:        { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.186     schwarze  252:        { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.103     kristaps  253:        { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.85      kristaps  254:        { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    255:        { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    256:        { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.178     schwarze  257:        { "it", roff_it, NULL, NULL, 0, NULL },
1.103     kristaps  258:        { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
                    259:        { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.104     kristaps  260:        { "nr", roff_nr, NULL, NULL, 0, NULL },
1.124     schwarze  261:        { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.229     schwarze  262:        { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.124     schwarze  263:        { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.122     schwarze  264:        { "rm", roff_rm, NULL, NULL, 0, NULL },
1.203     schwarze  265:        { "rr", roff_rr, NULL, NULL, 0, NULL },
1.105     kristaps  266:        { "so", roff_so, NULL, NULL, 0, NULL },
1.124     schwarze  267:        { "ta", roff_line_ignore, NULL, NULL, 0, NULL },
1.164     kristaps  268:        { "tr", roff_tr, NULL, NULL, 0, NULL },
1.175     schwarze  269:        { "Dd", roff_Dd, NULL, NULL, 0, NULL },
                    270:        { "TH", roff_TH, NULL, NULL, 0, NULL },
1.109     kristaps  271:        { "TS", roff_TS, NULL, NULL, 0, NULL },
                    272:        { "TE", roff_TE, NULL, NULL, 0, NULL },
1.112     kristaps  273:        { "T&", roff_T_, NULL, NULL, 0, NULL },
1.125     kristaps  274:        { "EQ", roff_EQ, NULL, NULL, 0, NULL },
                    275:        { "EN", roff_EN, NULL, NULL, 0, NULL },
1.85      kristaps  276:        { ".", roff_cblock, NULL, NULL, 0, NULL },
1.106     kristaps  277:        { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.67      kristaps  278: };
                    279:
1.200     schwarze  280: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175     schwarze  281: const  char *const __mdoc_reserved[] = {
                    282:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    283:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    284:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    285:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200     schwarze  286:        "Dt", "Dv", "Dx", "D1",
                    287:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    288:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175     schwarze  289:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200     schwarze  290:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    291:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175     schwarze  292:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200     schwarze  293:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    294:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    295:        "Sc", "Sh", "Sm", "So", "Sq",
1.175     schwarze  296:        "Ss", "St", "Sx", "Sy",
                    297:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200     schwarze  298:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175     schwarze  299:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    300:        NULL
                    301: };
                    302:
1.200     schwarze  303: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175     schwarze  304: const  char *const __man_reserved[] = {
1.200     schwarze  305:        "AT", "B", "BI", "BR", "DT",
                    306:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    307:        "LP", "OP", "P", "PD", "PP",
                    308:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    309:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175     schwarze  310:        NULL
                    311: };
                    312:
1.141     kristaps  313: /* Array of injected predefined strings. */
                    314: #define        PREDEFS_MAX      38
                    315: static const struct predef predefs[PREDEFS_MAX] = {
                    316: #include "predefs.in"
                    317: };
                    318:
1.155     kristaps  319: /* See roffhash_find() */
1.85      kristaps  320: #define        ROFF_HASH(p)    (p[0] - ASCII_LO)
                    321:
1.178     schwarze  322: static int      roffit_lines;  /* number of lines to delay */
                    323: static char    *roffit_macro;  /* nil-terminated macro line */
                    324:
1.207     schwarze  325:
1.85      kristaps  326: static void
1.155     kristaps  327: roffhash_init(void)
1.85      kristaps  328: {
                    329:        struct roffmac   *n;
                    330:        int               buc, i;
                    331:
1.106     kristaps  332:        for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.85      kristaps  333:                assert(roffs[i].name[0] >= ASCII_LO);
                    334:                assert(roffs[i].name[0] <= ASCII_HI);
                    335:
                    336:                buc = ROFF_HASH(roffs[i].name);
                    337:
                    338:                if (NULL != (n = hash[buc])) {
                    339:                        for ( ; n->next; n = n->next)
                    340:                                /* Do nothing. */ ;
                    341:                        n->next = &roffs[i];
                    342:                } else
                    343:                        hash[buc] = &roffs[i];
                    344:        }
                    345: }
                    346:
1.67      kristaps  347: /*
                    348:  * Look up a roff token by its name.  Returns ROFF_MAX if no macro by
                    349:  * the nil-terminated string name could be found.
                    350:  */
                    351: static enum rofft
1.155     kristaps  352: roffhash_find(const char *p, size_t s)
1.67      kristaps  353: {
1.85      kristaps  354:        int              buc;
                    355:        struct roffmac  *n;
1.67      kristaps  356:
1.85      kristaps  357:        /*
                    358:         * libroff has an extremely simple hashtable, for the time
                    359:         * being, which simply keys on the first character, which must
                    360:         * be printable, then walks a chain.  It works well enough until
                    361:         * optimised.
                    362:         */
                    363:
                    364:        if (p[0] < ASCII_LO || p[0] > ASCII_HI)
                    365:                return(ROFF_MAX);
                    366:
                    367:        buc = ROFF_HASH(p);
                    368:
                    369:        if (NULL == (n = hash[buc]))
                    370:                return(ROFF_MAX);
                    371:        for ( ; n; n = n->next)
1.106     kristaps  372:                if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.85      kristaps  373:                        return((enum rofft)(n - roffs));
1.67      kristaps  374:
                    375:        return(ROFF_MAX);
                    376: }
                    377:
                    378: /*
                    379:  * Pop the current node off of the stack of roff instructions currently
                    380:  * pending.
                    381:  */
                    382: static void
                    383: roffnode_pop(struct roff *r)
                    384: {
                    385:        struct roffnode *p;
                    386:
1.75      kristaps  387:        assert(r->last);
1.207     schwarze  388:        p = r->last;
1.82      kristaps  389:
1.75      kristaps  390:        r->last = r->last->parent;
1.106     kristaps  391:        free(p->name);
                    392:        free(p->end);
1.67      kristaps  393:        free(p);
                    394: }
                    395:
                    396: /*
                    397:  * Push a roff node onto the instruction stack.  This must later be
                    398:  * removed with roffnode_pop().
                    399:  */
1.98      schwarze  400: static void
1.106     kristaps  401: roffnode_push(struct roff *r, enum rofft tok, const char *name,
                    402:                int line, int col)
1.67      kristaps  403: {
                    404:        struct roffnode *p;
                    405:
1.98      schwarze  406:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.67      kristaps  407:        p->tok = tok;
1.106     kristaps  408:        if (name)
                    409:                p->name = mandoc_strdup(name);
1.67      kristaps  410:        p->parent = r->last;
                    411:        p->line = line;
                    412:        p->col = col;
1.198     schwarze  413:        p->rule = p->parent ? p->parent->rule : 0;
1.67      kristaps  414:
                    415:        r->last = p;
                    416: }
                    417:
                    418: static void
                    419: roff_free1(struct roff *r)
                    420: {
1.176     schwarze  421:        struct tbl_node *tbl;
1.125     kristaps  422:        struct eqn_node *e;
1.167     kristaps  423:        int              i;
1.67      kristaps  424:
1.176     schwarze  425:        while (NULL != (tbl = r->first_tbl)) {
                    426:                r->first_tbl = tbl->next;
                    427:                tbl_free(tbl);
1.109     kristaps  428:        }
1.113     kristaps  429:        r->first_tbl = r->last_tbl = r->tbl = NULL;
                    430:
1.125     kristaps  431:        while (NULL != (e = r->first_eqn)) {
                    432:                r->first_eqn = e->next;
                    433:                eqn_free(e);
                    434:        }
                    435:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    436:
1.67      kristaps  437:        while (r->last)
                    438:                roffnode_pop(r);
1.109     kristaps  439:
1.223     schwarze  440:        free (r->rstack);
                    441:        r->rstack = NULL;
                    442:        r->rstacksz = 0;
                    443:        r->rstackpos = -1;
                    444:
                    445:        roff_freereg(r->regtab);
                    446:        r->regtab = NULL;
                    447:
1.167     kristaps  448:        roff_freestr(r->strtab);
                    449:        roff_freestr(r->xmbtab);
                    450:        r->strtab = r->xmbtab = NULL;
                    451:
                    452:        if (r->xtab)
                    453:                for (i = 0; i < 128; i++)
                    454:                        free(r->xtab[i].p);
                    455:        free(r->xtab);
                    456:        r->xtab = NULL;
1.67      kristaps  457: }
                    458:
                    459: void
                    460: roff_reset(struct roff *r)
                    461: {
                    462:
                    463:        roff_free1(r);
1.227     schwarze  464:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174     kristaps  465:        r->control = 0;
1.67      kristaps  466: }
                    467:
                    468: void
                    469: roff_free(struct roff *r)
                    470: {
                    471:
                    472:        roff_free1(r);
                    473:        free(r);
                    474: }
                    475:
                    476: struct roff *
1.237     schwarze  477: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.67      kristaps  478: {
                    479:        struct roff     *r;
                    480:
1.98      schwarze  481:        r = mandoc_calloc(1, sizeof(struct roff));
1.128     kristaps  482:        r->parse = parse;
1.237     schwarze  483:        r->mchars = mchars;
1.199     schwarze  484:        r->options = options;
1.227     schwarze  485:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82      kristaps  486:        r->rstackpos = -1;
1.207     schwarze  487:
1.155     kristaps  488:        roffhash_init();
1.141     kristaps  489:
1.67      kristaps  490:        return(r);
                    491: }
                    492:
1.94      kristaps  493: /*
1.206     schwarze  494:  * In the current line, expand escape sequences that tend to get
                    495:  * used in numerical expressions and conditional requests.
                    496:  * Also check the syntax of the remaining escape sequences.
1.154     kristaps  497:  */
1.172     schwarze  498: static enum rofferr
1.238   ! schwarze  499: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94      kristaps  500: {
1.208     schwarze  501:        char             ubuf[24]; /* buffer to print the number */
1.205     schwarze  502:        const char      *start; /* start of the string to process */
1.209     schwarze  503:        char            *stesc; /* start of an escape sequence ('\\') */
1.108     schwarze  504:        const char      *stnam; /* start of the name, after "[(*" */
                    505:        const char      *cp;    /* end of the name, e.g. before ']' */
                    506:        const char      *res;   /* the string to be substituted */
1.238   ! schwarze  507:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.181     schwarze  508:        size_t           maxl;  /* expected length of the escape name */
                    509:        size_t           naml;  /* actual length of the escape name */
1.237     schwarze  510:        enum mandoc_esc  esc;   /* type of the escape sequence */
                    511:        int              inaml; /* length returned from mandoc_escape() */
1.181     schwarze  512:        int              expand_count;  /* to avoid infinite loops */
1.206     schwarze  513:        int              npos;  /* position in numeric expression */
1.218     schwarze  514:        int              arg_complete; /* argument not interrupted by eol */
1.206     schwarze  515:        char             term;  /* character terminating the escape */
1.94      kristaps  516:
1.170     schwarze  517:        expand_count = 0;
1.238   ! schwarze  518:        start = buf->buf + pos;
1.205     schwarze  519:        stesc = strchr(start, '\0') - 1;
                    520:        while (stesc-- > start) {
1.170     schwarze  521:
1.205     schwarze  522:                /* Search backwards for the next backslash. */
                    523:
1.238   ! schwarze  524:                if (*stesc != '\\')
1.205     schwarze  525:                        continue;
                    526:
                    527:                /* If it is escaped, skip it. */
                    528:
                    529:                for (cp = stesc - 1; cp >= start; cp--)
1.238   ! schwarze  530:                        if (*cp != '\\')
1.205     schwarze  531:                                break;
                    532:
1.238   ! schwarze  533:                if ((stesc - cp) % 2 == 0) {
1.209     schwarze  534:                        stesc = (char *)cp;
1.205     schwarze  535:                        continue;
                    536:                }
1.108     schwarze  537:
1.206     schwarze  538:                /* Decide whether to expand or to check only. */
1.108     schwarze  539:
1.206     schwarze  540:                term = '\0';
1.205     schwarze  541:                cp = stesc + 1;
1.181     schwarze  542:                switch (*cp) {
1.207     schwarze  543:                case '*':
1.181     schwarze  544:                        res = NULL;
                    545:                        break;
1.207     schwarze  546:                case 'B':
1.206     schwarze  547:                        /* FALLTHROUGH */
1.207     schwarze  548:                case 'w':
1.206     schwarze  549:                        term = cp[1];
                    550:                        /* FALLTHROUGH */
1.207     schwarze  551:                case 'n':
1.181     schwarze  552:                        res = ubuf;
                    553:                        break;
                    554:                default:
1.237     schwarze  555:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                    556:                        if (esc == ESCAPE_ERROR ||
                    557:                            (esc == ESCAPE_SPECIAL &&
                    558:                             mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.219     schwarze  559:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238   ! schwarze  560:                                    r->parse, ln, (int)(stesc - buf->buf),
1.219     schwarze  561:                                    "%.*s", (int)(cp - stesc), stesc);
1.205     schwarze  562:                        continue;
1.152     kristaps  563:                }
                    564:
1.205     schwarze  565:                if (EXPAND_LIMIT < ++expand_count) {
                    566:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238   ! schwarze  567:                            ln, (int)(stesc - buf->buf), NULL);
1.205     schwarze  568:                        return(ROFF_IGN);
                    569:                }
1.108     schwarze  570:
                    571:                /*
                    572:                 * The third character decides the length
1.181     schwarze  573:                 * of the name of the string or register.
1.108     schwarze  574:                 * Save a pointer to the name.
                    575:                 */
                    576:
1.238   ! schwarze  577:                if (term == '\0') {
1.206     schwarze  578:                        switch (*++cp) {
1.207     schwarze  579:                        case '\0':
1.206     schwarze  580:                                maxl = 0;
                    581:                                break;
1.207     schwarze  582:                        case '(':
1.206     schwarze  583:                                cp++;
                    584:                                maxl = 2;
                    585:                                break;
1.207     schwarze  586:                        case '[':
1.206     schwarze  587:                                cp++;
                    588:                                term = ']';
                    589:                                maxl = 0;
                    590:                                break;
                    591:                        default:
                    592:                                maxl = 1;
                    593:                                break;
                    594:                        }
                    595:                } else {
                    596:                        cp += 2;
1.94      kristaps  597:                        maxl = 0;
                    598:                }
1.108     schwarze  599:                stnam = cp;
1.94      kristaps  600:
1.108     schwarze  601:                /* Advance to the end of the name. */
1.94      kristaps  602:
1.218     schwarze  603:                arg_complete = 1;
1.238   ! schwarze  604:                for (naml = 0; maxl == 0 || naml < maxl; naml++, cp++) {
        !           605:                        if (*cp == '\0') {
1.219     schwarze  606:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238   ! schwarze  607:                                    ln, (int)(stesc - buf->buf), stesc);
1.218     schwarze  608:                                arg_complete = 0;
1.206     schwarze  609:                                break;
1.153     kristaps  610:                        }
1.238   ! schwarze  611:                        if (maxl == 0 && *cp == term) {
1.206     schwarze  612:                                cp++;
1.94      kristaps  613:                                break;
1.206     schwarze  614:                        }
1.94      kristaps  615:                }
                    616:
1.108     schwarze  617:                /*
                    618:                 * Retrieve the replacement string; if it is
                    619:                 * undefined, resume searching for escapes.
                    620:                 */
                    621:
1.206     schwarze  622:                switch (stesc[1]) {
1.207     schwarze  623:                case '*':
1.218     schwarze  624:                        if (arg_complete)
                    625:                                res = roff_getstrn(r, stnam, naml);
1.206     schwarze  626:                        break;
1.207     schwarze  627:                case 'B':
1.206     schwarze  628:                        npos = 0;
1.218     schwarze  629:                        ubuf[0] = arg_complete &&
1.234     kristaps  630:                            roff_evalnum(r, ln, stnam, &npos, NULL, 0) &&
1.218     schwarze  631:                            stnam + npos + 1 == cp ? '1' : '0';
1.206     schwarze  632:                        ubuf[1] = '\0';
                    633:                        break;
1.207     schwarze  634:                case 'n':
1.218     schwarze  635:                        if (arg_complete)
                    636:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                    637:                                    roff_getregn(r, stnam, naml));
                    638:                        else
                    639:                                ubuf[0] = '\0';
1.206     schwarze  640:                        break;
1.207     schwarze  641:                case 'w':
1.218     schwarze  642:                        /* use even incomplete args */
1.208     schwarze  643:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206     schwarze  644:                            24 * (int)naml);
                    645:                        break;
                    646:                }
1.94      kristaps  647:
1.238   ! schwarze  648:                if (res == NULL) {
1.219     schwarze  649:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238   ! schwarze  650:                            r->parse, ln, (int)(stesc - buf->buf),
1.219     schwarze  651:                            "%.*s", (int)naml, stnam);
1.142     kristaps  652:                        res = "";
1.94      kristaps  653:                }
                    654:
1.108     schwarze  655:                /* Replace the escape sequence by the string. */
                    656:
1.209     schwarze  657:                *stesc = '\0';
1.238   ! schwarze  658:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
        !           659:                    buf->buf, res, cp) + 1;
1.94      kristaps  660:
1.205     schwarze  661:                /* Prepare for the next replacement. */
1.94      kristaps  662:
1.205     schwarze  663:                start = nbuf + pos;
1.238   ! schwarze  664:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
        !           665:                free(buf->buf);
        !           666:                buf->buf = nbuf;
1.154     kristaps  667:        }
1.172     schwarze  668:        return(ROFF_CONT);
1.154     kristaps  669: }
                    670:
                    671: /*
1.178     schwarze  672:  * Process text streams:
                    673:  * Convert all breakable hyphens into ASCII_HYPH.
                    674:  * Decrement and spring input line trap.
1.154     kristaps  675:  */
                    676: static enum rofferr
1.238   ! schwarze  677: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154     kristaps  678: {
                    679:        size_t           sz;
                    680:        const char      *start;
1.178     schwarze  681:        char            *p;
                    682:        int              isz;
1.154     kristaps  683:        enum mandoc_esc  esc;
                    684:
1.238   ! schwarze  685:        start = p = buf->buf + pos;
1.154     kristaps  686:
1.238   ! schwarze  687:        while (*p != '\0') {
1.154     kristaps  688:                sz = strcspn(p, "-\\");
                    689:                p += sz;
                    690:
1.238   ! schwarze  691:                if (*p == '\0')
1.159     kristaps  692:                        break;
                    693:
1.238   ! schwarze  694:                if (*p == '\\') {
1.154     kristaps  695:                        /* Skip over escapes. */
                    696:                        p++;
1.189     schwarze  697:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238   ! schwarze  698:                        if (esc == ESCAPE_ERROR)
1.154     kristaps  699:                                break;
1.155     kristaps  700:                        continue;
1.159     kristaps  701:                } else if (p == start) {
1.158     kristaps  702:                        p++;
1.155     kristaps  703:                        continue;
1.158     kristaps  704:                }
1.155     kristaps  705:
1.171     schwarze  706:                if (isalpha((unsigned char)p[-1]) &&
                    707:                    isalpha((unsigned char)p[1]))
1.155     kristaps  708:                        *p = ASCII_HYPH;
                    709:                p++;
1.94      kristaps  710:        }
                    711:
1.178     schwarze  712:        /* Spring the input line trap. */
1.238   ! schwarze  713:        if (roffit_lines == 1) {
        !           714:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
        !           715:                free(buf->buf);
        !           716:                buf->buf = p;
        !           717:                buf->sz = isz + 1;
1.178     schwarze  718:                *offs = 0;
                    719:                free(roffit_macro);
                    720:                roffit_lines = 0;
                    721:                return(ROFF_REPARSE);
1.238   ! schwarze  722:        } else if (roffit_lines > 1)
1.178     schwarze  723:                --roffit_lines;
1.154     kristaps  724:        return(ROFF_CONT);
1.94      kristaps  725: }
                    726:
1.67      kristaps  727: enum rofferr
1.238   ! schwarze  728: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67      kristaps  729: {
                    730:        enum rofft       t;
1.109     kristaps  731:        enum rofferr     e;
1.238   ! schwarze  732:        int              pos;   /* parse point */
        !           733:        int              ppos;  /* original offset in buf->buf */
        !           734:        int              ctl;   /* macro line (boolean) */
        !           735:
        !           736:        ppos = pos = *offs;
1.79      kristaps  737:
1.230     schwarze  738:        /* Handle in-line equation delimiters. */
                    739:
1.236     schwarze  740:        if (r->tbl == NULL &&
                    741:            r->last_eqn != NULL && r->last_eqn->delim &&
1.230     schwarze  742:            (r->eqn == NULL || r->eqn_inline)) {
1.238   ! schwarze  743:                e = roff_eqndelim(r, buf, pos);
1.230     schwarze  744:                if (e == ROFF_REPARSE)
                    745:                        return(e);
                    746:                assert(e == ROFF_CONT);
                    747:        }
                    748:
                    749:        /* Expand some escape sequences. */
1.94      kristaps  750:
1.238   ! schwarze  751:        e = roff_res(r, buf, ln, pos);
        !           752:        if (e == ROFF_IGN)
1.172     schwarze  753:                return(e);
1.238   ! schwarze  754:        assert(e == ROFF_CONT);
1.94      kristaps  755:
1.238   ! schwarze  756:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.130     kristaps  757:
1.94      kristaps  758:        /*
1.79      kristaps  759:         * First, if a scope is open and we're not a macro, pass the
                    760:         * text through the macro's filter.  If a scope isn't open and
                    761:         * we're not a macro, just let it through.
1.125     kristaps  762:         * Finally, if there's an equation scope open, divert it into it
                    763:         * no matter our state.
1.79      kristaps  764:         */
1.74      kristaps  765:
1.130     kristaps  766:        if (r->last && ! ctl) {
1.78      kristaps  767:                t = r->last->tok;
                    768:                assert(roffs[t].text);
1.238   ! schwarze  769:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
        !           770:                assert(e == ROFF_IGN || e == ROFF_CONT);
        !           771:                if (e != ROFF_CONT)
1.125     kristaps  772:                        return(e);
1.182     schwarze  773:        }
                    774:        if (r->eqn)
1.238   ! schwarze  775:                return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.182     schwarze  776:        if ( ! ctl) {
1.125     kristaps  777:                if (r->tbl)
1.238   ! schwarze  778:                        return(tbl_read(r->tbl, ln, buf->buf, pos));
        !           779:                return(roff_parsetext(buf, pos, offs));
1.182     schwarze  780:        }
1.228     schwarze  781:
                    782:        /* Skip empty request lines. */
                    783:
1.238   ! schwarze  784:        if (buf->buf[pos] == '"') {
1.228     schwarze  785:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                    786:                    ln, pos, NULL);
                    787:                return(ROFF_IGN);
1.238   ! schwarze  788:        } else if (buf->buf[pos] == '\0')
1.228     schwarze  789:                return(ROFF_IGN);
1.67      kristaps  790:
1.79      kristaps  791:        /*
                    792:         * If a scope is open, go to the child handler for that macro,
                    793:         * as it may want to preprocess before doing anything with it.
1.125     kristaps  794:         * Don't do so if an equation is open.
1.79      kristaps  795:         */
1.78      kristaps  796:
1.79      kristaps  797:        if (r->last) {
                    798:                t = r->last->tok;
                    799:                assert(roffs[t].sub);
1.238   ! schwarze  800:                return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.79      kristaps  801:        }
1.78      kristaps  802:
1.79      kristaps  803:        /*
                    804:         * Lastly, as we've no scope open, try to look up and execute
                    805:         * the new macro.  If no macro is found, simply return and let
                    806:         * the compilers handle it.
                    807:         */
1.67      kristaps  808:
1.238   ! schwarze  809:        if ((t = roff_parse(r, buf->buf, &pos, ln, ppos)) == ROFF_MAX)
1.79      kristaps  810:                return(ROFF_CONT);
1.67      kristaps  811:
1.75      kristaps  812:        assert(roffs[t].proc);
1.238   ! schwarze  813:        return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.74      kristaps  814: }
                    815:
1.117     kristaps  816: void
1.74      kristaps  817: roff_endparse(struct roff *r)
                    818: {
                    819:
1.110     kristaps  820:        if (r->last)
1.221     schwarze  821:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                    822:                    r->last->line, r->last->col,
                    823:                    roffs[r->last->tok].name);
1.117     kristaps  824:
1.125     kristaps  825:        if (r->eqn) {
1.221     schwarze  826:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                    827:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151     kristaps  828:                eqn_end(&r->eqn);
1.125     kristaps  829:        }
                    830:
1.117     kristaps  831:        if (r->tbl) {
1.221     schwarze  832:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                    833:                    r->tbl->line, r->tbl->pos, "TS");
1.151     kristaps  834:                tbl_end(&r->tbl);
1.117     kristaps  835:        }
1.67      kristaps  836: }
                    837:
                    838: /*
                    839:  * Parse a roff node's type from the input buffer.  This must be in the
                    840:  * form of ".foo xxx" in the usual way.
                    841:  */
                    842: static enum rofft
1.214     schwarze  843: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67      kristaps  844: {
1.214     schwarze  845:        char            *cp;
1.106     kristaps  846:        const char      *mac;
                    847:        size_t           maclen;
1.67      kristaps  848:        enum rofft       t;
                    849:
1.214     schwarze  850:        cp = buf + *pos;
                    851:
                    852:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.67      kristaps  853:                return(ROFF_MAX);
                    854:
1.214     schwarze  855:        mac = cp;
                    856:        maclen = roff_getname(r, &cp, ln, ppos);
1.67      kristaps  857:
1.106     kristaps  858:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.155     kristaps  859:            ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.67      kristaps  860:
1.214     schwarze  861:        if (ROFF_MAX != t)
                    862:                *pos = cp - buf;
1.67      kristaps  863:
                    864:        return(t);
                    865: }
                    866:
                    867: static enum rofferr
1.76      kristaps  868: roff_cblock(ROFF_ARGS)
1.67      kristaps  869: {
                    870:
1.79      kristaps  871:        /*
                    872:         * A block-close `..' should only be invoked as a child of an
                    873:         * ignore macro, otherwise raise a warning and just ignore it.
                    874:         */
                    875:
1.238   ! schwarze  876:        if (r->last == NULL) {
1.221     schwarze  877:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                    878:                    ln, ppos, "..");
1.76      kristaps  879:                return(ROFF_IGN);
                    880:        }
1.67      kristaps  881:
1.81      kristaps  882:        switch (r->last->tok) {
1.207     schwarze  883:        case ROFF_am:
1.220     schwarze  884:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.81      kristaps  885:                /* FALLTHROUGH */
1.207     schwarze  886:        case ROFF_ami:
1.81      kristaps  887:                /* FALLTHROUGH */
1.207     schwarze  888:        case ROFF_de:
1.108     schwarze  889:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.81      kristaps  890:                /* FALLTHROUGH */
1.207     schwarze  891:        case ROFF_dei:
1.81      kristaps  892:                /* FALLTHROUGH */
1.207     schwarze  893:        case ROFF_ig:
1.81      kristaps  894:                break;
                    895:        default:
1.221     schwarze  896:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                    897:                    ln, ppos, "..");
1.67      kristaps  898:                return(ROFF_IGN);
1.76      kristaps  899:        }
1.67      kristaps  900:
1.238   ! schwarze  901:        if (buf->buf[pos] != '\0')
1.217     schwarze  902:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238   ! schwarze  903:                    ".. %s", buf->buf + pos);
1.71      kristaps  904:
                    905:        roffnode_pop(r);
1.76      kristaps  906:        roffnode_cleanscope(r);
                    907:        return(ROFF_IGN);
1.71      kristaps  908:
1.67      kristaps  909: }
                    910:
1.76      kristaps  911: static void
                    912: roffnode_cleanscope(struct roff *r)
1.67      kristaps  913: {
                    914:
1.76      kristaps  915:        while (r->last) {
1.173     schwarze  916:                if (--r->last->endspan != 0)
1.76      kristaps  917:                        break;
                    918:                roffnode_pop(r);
                    919:        }
1.67      kristaps  920: }
                    921:
1.195     schwarze  922: static void
                    923: roff_ccond(struct roff *r, int ln, int ppos)
1.74      kristaps  924: {
                    925:
1.76      kristaps  926:        if (NULL == r->last) {
1.221     schwarze  927:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                    928:                    ln, ppos, "\\}");
1.195     schwarze  929:                return;
1.76      kristaps  930:        }
                    931:
1.82      kristaps  932:        switch (r->last->tok) {
1.207     schwarze  933:        case ROFF_el:
1.82      kristaps  934:                /* FALLTHROUGH */
1.207     schwarze  935:        case ROFF_ie:
1.82      kristaps  936:                /* FALLTHROUGH */
1.207     schwarze  937:        case ROFF_if:
1.82      kristaps  938:                break;
                    939:        default:
1.221     schwarze  940:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                    941:                    ln, ppos, "\\}");
1.195     schwarze  942:                return;
1.75      kristaps  943:        }
                    944:
1.76      kristaps  945:        if (r->last->endspan > -1) {
1.221     schwarze  946:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                    947:                    ln, ppos, "\\}");
1.195     schwarze  948:                return;
1.76      kristaps  949:        }
                    950:
1.75      kristaps  951:        roffnode_pop(r);
1.76      kristaps  952:        roffnode_cleanscope(r);
1.195     schwarze  953:        return;
1.76      kristaps  954: }
                    955:
                    956: static enum rofferr
1.80      kristaps  957: roff_block(ROFF_ARGS)
1.76      kristaps  958: {
1.220     schwarze  959:        const char      *name;
                    960:        char            *iname, *cp;
1.213     schwarze  961:        size_t           namesz;
1.106     kristaps  962:
1.220     schwarze  963:        /* Ignore groff compatibility mode for now. */
1.76      kristaps  964:
1.238   ! schwarze  965:        if (tok == ROFF_de1)
1.220     schwarze  966:                tok = ROFF_de;
1.238   ! schwarze  967:        else if (tok == ROFF_am1)
1.220     schwarze  968:                tok = ROFF_am;
                    969:
                    970:        /* Parse the macro name argument. */
                    971:
1.238   ! schwarze  972:        cp = buf->buf + pos;
        !           973:        if (tok == ROFF_ig) {
1.220     schwarze  974:                iname = NULL;
                    975:                namesz = 0;
                    976:        } else {
                    977:                iname = cp;
                    978:                namesz = roff_getname(r, &cp, ln, ppos);
                    979:                iname[namesz] = '\0';
                    980:        }
1.107     kristaps  981:
1.220     schwarze  982:        /* Resolve the macro name argument if it is indirect. */
1.107     kristaps  983:
1.238   ! schwarze  984:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
        !           985:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220     schwarze  986:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238   ! schwarze  987:                            r->parse, ln, (int)(iname - buf->buf),
1.220     schwarze  988:                            "%.*s", (int)namesz, iname);
                    989:                        namesz = 0;
                    990:                } else
                    991:                        namesz = strlen(name);
                    992:        } else
                    993:                name = iname;
1.107     kristaps  994:
1.238   ! schwarze  995:        if (namesz == 0 && tok != ROFF_ig) {
1.220     schwarze  996:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
                    997:                    ln, ppos, roffs[tok].name);
                    998:                return(ROFF_IGN);
                    999:        }
1.80      kristaps 1000:
1.106     kristaps 1001:        roffnode_push(r, tok, name, ln, ppos);
                   1002:
                   1003:        /*
                   1004:         * At the beginning of a `de' macro, clear the existing string
                   1005:         * with the same name, if there is one.  New content will be
1.193     schwarze 1006:         * appended from roff_block_text() in multiline mode.
1.106     kristaps 1007:         */
1.107     kristaps 1008:
1.238   ! schwarze 1009:        if (tok == ROFF_de || tok == ROFF_dei)
1.213     schwarze 1010:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76      kristaps 1011:
1.238   ! schwarze 1012:        if (*cp == '\0')
1.78      kristaps 1013:                return(ROFF_IGN);
                   1014:
1.220     schwarze 1015:        /* Get the custom end marker. */
1.107     kristaps 1016:
1.220     schwarze 1017:        iname = cp;
1.213     schwarze 1018:        namesz = roff_getname(r, &cp, ln, ppos);
1.220     schwarze 1019:
                   1020:        /* Resolve the end marker if it is indirect. */
                   1021:
1.238   ! schwarze 1022:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
        !          1023:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220     schwarze 1024:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238   ! schwarze 1025:                            r->parse, ln, (int)(iname - buf->buf),
1.220     schwarze 1026:                            "%.*s", (int)namesz, iname);
                   1027:                        namesz = 0;
                   1028:                } else
                   1029:                        namesz = strlen(name);
                   1030:        } else
                   1031:                name = iname;
                   1032:
1.213     schwarze 1033:        if (namesz)
                   1034:                r->last->end = mandoc_strndup(name, namesz);
1.78      kristaps 1035:
1.238   ! schwarze 1036:        if (*cp != '\0')
1.217     schwarze 1037:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   1038:                    ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.74      kristaps 1039:
1.78      kristaps 1040:        return(ROFF_IGN);
                   1041: }
                   1042:
                   1043: static enum rofferr
1.80      kristaps 1044: roff_block_sub(ROFF_ARGS)
1.79      kristaps 1045: {
                   1046:        enum rofft      t;
                   1047:        int             i, j;
                   1048:
                   1049:        /*
                   1050:         * First check whether a custom macro exists at this level.  If
                   1051:         * it does, then check against it.  This is some of groff's
                   1052:         * stranger behaviours.  If we encountered a custom end-scope
                   1053:         * tag and that tag also happens to be a "real" macro, then we
                   1054:         * need to try interpreting it again as a real macro.  If it's
                   1055:         * not, then return ignore.  Else continue.
                   1056:         */
                   1057:
                   1058:        if (r->last->end) {
1.130     kristaps 1059:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238   ! schwarze 1060:                        if (buf->buf[i] != r->last->end[j])
1.79      kristaps 1061:                                break;
                   1062:
1.238   ! schwarze 1063:                if (r->last->end[j] == '\0' &&
        !          1064:                    (buf->buf[i] == '\0' ||
        !          1065:                     buf->buf[i] == ' ' ||
        !          1066:                     buf->buf[i] == '\t')) {
1.79      kristaps 1067:                        roffnode_pop(r);
                   1068:                        roffnode_cleanscope(r);
                   1069:
1.238   ! schwarze 1070:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130     kristaps 1071:                                i++;
                   1072:
                   1073:                        pos = i;
1.238   ! schwarze 1074:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
        !          1075:                            ROFF_MAX)
1.79      kristaps 1076:                                return(ROFF_RERUN);
                   1077:                        return(ROFF_IGN);
                   1078:                }
                   1079:        }
                   1080:
                   1081:        /*
                   1082:         * If we have no custom end-query or lookup failed, then try
                   1083:         * pulling it out of the hashtable.
                   1084:         */
                   1085:
1.238   ! schwarze 1086:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79      kristaps 1087:
1.238   ! schwarze 1088:        if (t != ROFF_cblock) {
        !          1089:                if (tok != ROFF_ig)
        !          1090:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.79      kristaps 1091:                return(ROFF_IGN);
1.106     kristaps 1092:        }
1.79      kristaps 1093:
                   1094:        assert(roffs[t].proc);
1.238   ! schwarze 1095:        return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.79      kristaps 1096: }
                   1097:
                   1098: static enum rofferr
1.80      kristaps 1099: roff_block_text(ROFF_ARGS)
1.78      kristaps 1100: {
                   1101:
1.238   ! schwarze 1102:        if (tok != ROFF_ig)
        !          1103:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106     kristaps 1104:
1.78      kristaps 1105:        return(ROFF_IGN);
                   1106: }
                   1107:
                   1108: static enum rofferr
1.82      kristaps 1109: roff_cond_sub(ROFF_ARGS)
                   1110: {
                   1111:        enum rofft       t;
1.139     kristaps 1112:        char            *ep;
1.198     schwarze 1113:        int              rr;
1.82      kristaps 1114:
                   1115:        rr = r->last->rule;
1.139     kristaps 1116:        roffnode_cleanscope(r);
1.238   ! schwarze 1117:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82      kristaps 1118:
1.139     kristaps 1119:        /*
1.177     schwarze 1120:         * Fully handle known macros when they are structurally
                   1121:         * required or when the conditional evaluated to true.
1.87      kristaps 1122:         */
                   1123:
1.238   ! schwarze 1124:        if ((t != ROFF_MAX) &&
        !          1125:            (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.177     schwarze 1126:                assert(roffs[t].proc);
1.238   ! schwarze 1127:                return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.177     schwarze 1128:        }
1.144     kristaps 1129:
1.196     schwarze 1130:        /*
                   1131:         * If `\}' occurs on a macro line without a preceding macro,
                   1132:         * drop the line completely.
                   1133:         */
                   1134:
1.238   ! schwarze 1135:        ep = buf->buf + pos;
        !          1136:        if (ep[0] == '\\' && ep[1] == '}')
1.198     schwarze 1137:                rr = 0;
1.196     schwarze 1138:
1.177     schwarze 1139:        /* Always check for the closing delimiter `\}'. */
1.144     kristaps 1140:
1.238   ! schwarze 1141:        while ((ep = strchr(ep, '\\')) != NULL) {
        !          1142:                if (*(++ep) == '}') {
1.197     schwarze 1143:                        *ep = '&';
1.238   ! schwarze 1144:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.197     schwarze 1145:                }
                   1146:                ++ep;
1.177     schwarze 1147:        }
1.198     schwarze 1148:        return(rr ? ROFF_CONT : ROFF_IGN);
1.82      kristaps 1149: }
                   1150:
                   1151: static enum rofferr
                   1152: roff_cond_text(ROFF_ARGS)
1.78      kristaps 1153: {
1.140     kristaps 1154:        char            *ep;
1.198     schwarze 1155:        int              rr;
1.82      kristaps 1156:
                   1157:        rr = r->last->rule;
1.140     kristaps 1158:        roffnode_cleanscope(r);
1.82      kristaps 1159:
1.238   ! schwarze 1160:        ep = buf->buf + pos;
        !          1161:        while ((ep = strchr(ep, '\\')) != NULL) {
        !          1162:                if (*(++ep) == '}') {
1.197     schwarze 1163:                        *ep = '&';
1.238   ! schwarze 1164:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.197     schwarze 1165:                }
                   1166:                ++ep;
1.78      kristaps 1167:        }
1.198     schwarze 1168:        return(rr ? ROFF_CONT : ROFF_IGN);
1.74      kristaps 1169: }
                   1170:
1.204     schwarze 1171: /*
                   1172:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1173:  * If there is at least one digit, return success and advance the
                   1174:  * parse point, else return failure and let the parse point unchanged.
                   1175:  * Ignore overflows, treat them just like the C language.
                   1176:  */
1.184     schwarze 1177: static int
                   1178: roff_getnum(const char *v, int *pos, int *res)
                   1179: {
1.206     schwarze 1180:        int      myres, n, p;
                   1181:
                   1182:        if (NULL == res)
                   1183:                res = &myres;
1.184     schwarze 1184:
                   1185:        p = *pos;
                   1186:        n = v[p] == '-';
                   1187:        if (n)
                   1188:                p++;
                   1189:
                   1190:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204     schwarze 1191:                *res = 10 * *res + v[p] - '0';
1.184     schwarze 1192:        if (p == *pos + n)
                   1193:                return 0;
                   1194:
                   1195:        if (n)
                   1196:                *res = -*res;
                   1197:
                   1198:        *pos = p;
                   1199:        return 1;
                   1200: }
                   1201:
1.198     schwarze 1202: /*
                   1203:  * Evaluate a string comparison condition.
                   1204:  * The first character is the delimiter.
                   1205:  * Succeed if the string up to its second occurrence
                   1206:  * matches the string up to its third occurence.
                   1207:  * Advance the cursor after the third occurrence
                   1208:  * or lacking that, to the end of the line.
                   1209:  */
                   1210: static int
                   1211: roff_evalstrcond(const char *v, int *pos)
                   1212: {
                   1213:        const char      *s1, *s2, *s3;
                   1214:        int              match;
                   1215:
                   1216:        match = 0;
                   1217:        s1 = v + *pos;          /* initial delimiter */
                   1218:        s2 = s1 + 1;            /* for scanning the first string */
                   1219:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   1220:
                   1221:        if (NULL == s3)         /* found no middle delimiter */
                   1222:                goto out;
                   1223:
                   1224:        while ('\0' != *++s3) {
                   1225:                if (*s2 != *s3) {  /* mismatch */
                   1226:                        s3 = strchr(s3, *s1);
                   1227:                        break;
                   1228:                }
                   1229:                if (*s3 == *s1) {  /* found the final delimiter */
                   1230:                        match = 1;
                   1231:                        break;
                   1232:                }
                   1233:                s2++;
                   1234:        }
                   1235:
                   1236: out:
                   1237:        if (NULL == s3)
                   1238:                s3 = strchr(s2, '\0');
                   1239:        else
                   1240:                s3++;
                   1241:        *pos = s3 - v;
                   1242:        return(match);
                   1243: }
                   1244:
1.204     schwarze 1245: /*
                   1246:  * Evaluate an optionally negated single character, numerical,
                   1247:  * or string condition.
                   1248:  */
1.198     schwarze 1249: static int
1.234     kristaps 1250: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.88      kristaps 1251: {
1.204     schwarze 1252:        int      wanttrue, number;
1.88      kristaps 1253:
1.198     schwarze 1254:        if ('!' == v[*pos]) {
                   1255:                wanttrue = 0;
                   1256:                (*pos)++;
                   1257:        } else
                   1258:                wanttrue = 1;
                   1259:
1.88      kristaps 1260:        switch (v[*pos]) {
1.207     schwarze 1261:        case 'n':
1.198     schwarze 1262:                /* FALLTHROUGH */
1.207     schwarze 1263:        case 'o':
1.88      kristaps 1264:                (*pos)++;
1.198     schwarze 1265:                return(wanttrue);
1.207     schwarze 1266:        case 'c':
1.198     schwarze 1267:                /* FALLTHROUGH */
1.207     schwarze 1268:        case 'd':
1.198     schwarze 1269:                /* FALLTHROUGH */
1.207     schwarze 1270:        case 'e':
1.88      kristaps 1271:                /* FALLTHROUGH */
1.207     schwarze 1272:        case 'r':
1.88      kristaps 1273:                /* FALLTHROUGH */
1.207     schwarze 1274:        case 't':
1.88      kristaps 1275:                (*pos)++;
1.198     schwarze 1276:                return(!wanttrue);
1.88      kristaps 1277:        default:
                   1278:                break;
                   1279:        }
                   1280:
1.234     kristaps 1281:        if (roff_evalnum(r, ln, v, pos, &number, 0))
1.204     schwarze 1282:                return((number > 0) == wanttrue);
                   1283:        else
1.198     schwarze 1284:                return(roff_evalstrcond(v, pos) == wanttrue);
1.88      kristaps 1285: }
                   1286:
1.74      kristaps 1287: static enum rofferr
1.103     kristaps 1288: roff_line_ignore(ROFF_ARGS)
1.89      kristaps 1289: {
1.123     schwarze 1290:
1.89      kristaps 1291:        return(ROFF_IGN);
                   1292: }
                   1293:
1.104     kristaps 1294: static enum rofferr
1.82      kristaps 1295: roff_cond(ROFF_ARGS)
1.74      kristaps 1296: {
1.173     schwarze 1297:
                   1298:        roffnode_push(r, tok, NULL, ln, ppos);
1.74      kristaps 1299:
1.207     schwarze 1300:        /*
1.134     kristaps 1301:         * An `.el' has no conditional body: it will consume the value
                   1302:         * of the current rstack entry set in prior `ie' calls or
1.207     schwarze 1303:         * defaults to DENY.
1.134     kristaps 1304:         *
                   1305:         * If we're not an `el', however, then evaluate the conditional.
                   1306:         */
1.133     kristaps 1307:
1.238   ! schwarze 1308:        r->last->rule = tok == ROFF_el ?
1.207     schwarze 1309:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238   ! schwarze 1310:            roff_evalcond(r, ln, buf->buf, &pos);
1.77      kristaps 1311:
1.134     kristaps 1312:        /*
                   1313:         * An if-else will put the NEGATION of the current evaluated
                   1314:         * conditional into the stack of rules.
                   1315:         */
                   1316:
1.238   ! schwarze 1317:        if (tok == ROFF_ie) {
1.223     schwarze 1318:                if (r->rstackpos + 1 == r->rstacksz) {
                   1319:                        r->rstacksz += 16;
                   1320:                        r->rstack = mandoc_reallocarray(r->rstack,
                   1321:                            r->rstacksz, sizeof(int));
1.134     kristaps 1322:                }
1.198     schwarze 1323:                r->rstack[++r->rstackpos] = !r->last->rule;
1.82      kristaps 1324:        }
1.88      kristaps 1325:
                   1326:        /* If the parent has false as its rule, then so do we. */
                   1327:
1.198     schwarze 1328:        if (r->last->parent && !r->last->parent->rule)
                   1329:                r->last->rule = 0;
1.88      kristaps 1330:
                   1331:        /*
1.173     schwarze 1332:         * Determine scope.
                   1333:         * If there is nothing on the line after the conditional,
                   1334:         * not even whitespace, use next-line scope.
1.88      kristaps 1335:         */
1.74      kristaps 1336:
1.238   ! schwarze 1337:        if (buf->buf[pos] == '\0') {
1.173     schwarze 1338:                r->last->endspan = 2;
                   1339:                goto out;
                   1340:        }
                   1341:
1.238   ! schwarze 1342:        while (buf->buf[pos] == ' ')
1.173     schwarze 1343:                pos++;
                   1344:
                   1345:        /* An opening brace requests multiline scope. */
1.75      kristaps 1346:
1.238   ! schwarze 1347:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75      kristaps 1348:                r->last->endspan = -1;
                   1349:                pos += 2;
1.173     schwarze 1350:                goto out;
1.207     schwarze 1351:        }
1.74      kristaps 1352:
1.77      kristaps 1353:        /*
1.173     schwarze 1354:         * Anything else following the conditional causes
                   1355:         * single-line scope.  Warn if the scope contains
                   1356:         * nothing but trailing whitespace.
1.77      kristaps 1357:         */
                   1358:
1.238   ! schwarze 1359:        if (buf->buf[pos] == '\0')
1.216     schwarze 1360:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
                   1361:                    ln, ppos, roffs[tok].name);
1.77      kristaps 1362:
1.173     schwarze 1363:        r->last->endspan = 1;
1.74      kristaps 1364:
1.173     schwarze 1365: out:
1.75      kristaps 1366:        *offs = pos;
                   1367:        return(ROFF_RERUN);
1.83      schwarze 1368: }
                   1369:
                   1370: static enum rofferr
1.92      schwarze 1371: roff_ds(ROFF_ARGS)
                   1372: {
1.212     schwarze 1373:        char            *string;
                   1374:        const char      *name;
                   1375:        size_t           namesz;
1.96      kristaps 1376:
                   1377:        /*
1.212     schwarze 1378:         * The first word is the name of the string.
                   1379:         * If it is empty or terminated by an escape sequence,
                   1380:         * abort the `ds' request without defining anything.
1.96      kristaps 1381:         */
1.92      schwarze 1382:
1.238   ! schwarze 1383:        name = string = buf->buf + pos;
        !          1384:        if (*name == '\0')
1.92      schwarze 1385:                return(ROFF_IGN);
                   1386:
1.212     schwarze 1387:        namesz = roff_getname(r, &string, ln, pos);
1.238   ! schwarze 1388:        if (name[namesz] == '\\')
1.212     schwarze 1389:                return(ROFF_IGN);
                   1390:
                   1391:        /* Read past the initial double-quote, if any. */
1.238   ! schwarze 1392:        if (*string == '"')
1.92      schwarze 1393:                string++;
                   1394:
1.96      kristaps 1395:        /* The rest is the value. */
1.212     schwarze 1396:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   1397:            ROFF_as == tok);
1.92      schwarze 1398:        return(ROFF_IGN);
                   1399: }
                   1400:
1.204     schwarze 1401: /*
                   1402:  * Parse a single operator, one or two characters long.
                   1403:  * If the operator is recognized, return success and advance the
                   1404:  * parse point, else return failure and let the parse point unchanged.
                   1405:  */
                   1406: static int
                   1407: roff_getop(const char *v, int *pos, char *res)
                   1408: {
                   1409:
                   1410:        *res = v[*pos];
                   1411:
                   1412:        switch (*res) {
1.207     schwarze 1413:        case '+':
1.204     schwarze 1414:                /* FALLTHROUGH */
1.207     schwarze 1415:        case '-':
1.204     schwarze 1416:                /* FALLTHROUGH */
1.207     schwarze 1417:        case '*':
1.204     schwarze 1418:                /* FALLTHROUGH */
1.207     schwarze 1419:        case '/':
1.204     schwarze 1420:                /* FALLTHROUGH */
1.207     schwarze 1421:        case '%':
1.204     schwarze 1422:                /* FALLTHROUGH */
1.207     schwarze 1423:        case '&':
1.204     schwarze 1424:                /* FALLTHROUGH */
1.207     schwarze 1425:        case ':':
1.204     schwarze 1426:                break;
                   1427:        case '<':
                   1428:                switch (v[*pos + 1]) {
1.207     schwarze 1429:                case '=':
1.204     schwarze 1430:                        *res = 'l';
                   1431:                        (*pos)++;
                   1432:                        break;
1.207     schwarze 1433:                case '>':
1.204     schwarze 1434:                        *res = '!';
                   1435:                        (*pos)++;
                   1436:                        break;
1.207     schwarze 1437:                case '?':
1.204     schwarze 1438:                        *res = 'i';
                   1439:                        (*pos)++;
                   1440:                        break;
                   1441:                default:
                   1442:                        break;
                   1443:                }
                   1444:                break;
                   1445:        case '>':
                   1446:                switch (v[*pos + 1]) {
1.207     schwarze 1447:                case '=':
1.204     schwarze 1448:                        *res = 'g';
                   1449:                        (*pos)++;
                   1450:                        break;
1.207     schwarze 1451:                case '?':
1.204     schwarze 1452:                        *res = 'a';
                   1453:                        (*pos)++;
                   1454:                        break;
                   1455:                default:
                   1456:                        break;
                   1457:                }
                   1458:                break;
                   1459:        case '=':
                   1460:                if ('=' == v[*pos + 1])
                   1461:                        (*pos)++;
                   1462:                break;
                   1463:        default:
                   1464:                return(0);
                   1465:        }
                   1466:        (*pos)++;
                   1467:
                   1468:        return(*res);
                   1469: }
                   1470:
                   1471: /*
                   1472:  * Evaluate either a parenthesized numeric expression
                   1473:  * or a single signed integer number.
                   1474:  */
                   1475: static int
1.235     schwarze 1476: roff_evalpar(struct roff *r, int ln,
1.234     kristaps 1477:        const char *v, int *pos, int *res)
1.204     schwarze 1478: {
                   1479:
                   1480:        if ('(' != v[*pos])
                   1481:                return(roff_getnum(v, pos, res));
                   1482:
                   1483:        (*pos)++;
1.234     kristaps 1484:        if ( ! roff_evalnum(r, ln, v, pos, res, 1))
1.204     schwarze 1485:                return(0);
                   1486:
1.206     schwarze 1487:        /*
                   1488:         * Omission of the closing parenthesis
                   1489:         * is an error in validation mode,
                   1490:         * but ignored in evaluation mode.
                   1491:         */
                   1492:
1.204     schwarze 1493:        if (')' == v[*pos])
                   1494:                (*pos)++;
1.206     schwarze 1495:        else if (NULL == res)
                   1496:                return(0);
1.204     schwarze 1497:
                   1498:        return(1);
                   1499: }
                   1500:
                   1501: /*
                   1502:  * Evaluate a complete numeric expression.
                   1503:  * Proceed left to right, there is no concept of precedence.
                   1504:  */
                   1505: static int
1.235     schwarze 1506: roff_evalnum(struct roff *r, int ln, const char *v,
1.234     kristaps 1507:        int *pos, int *res, int skipwhite)
1.204     schwarze 1508: {
                   1509:        int              mypos, operand2;
                   1510:        char             operator;
                   1511:
                   1512:        if (NULL == pos) {
                   1513:                mypos = 0;
                   1514:                pos = &mypos;
                   1515:        }
                   1516:
                   1517:        if (skipwhite)
                   1518:                while (isspace((unsigned char)v[*pos]))
                   1519:                        (*pos)++;
                   1520:
1.234     kristaps 1521:        if ( ! roff_evalpar(r, ln, v, pos, res))
1.204     schwarze 1522:                return(0);
                   1523:
                   1524:        while (1) {
                   1525:                if (skipwhite)
                   1526:                        while (isspace((unsigned char)v[*pos]))
                   1527:                                (*pos)++;
                   1528:
                   1529:                if ( ! roff_getop(v, pos, &operator))
                   1530:                        break;
                   1531:
                   1532:                if (skipwhite)
                   1533:                        while (isspace((unsigned char)v[*pos]))
                   1534:                                (*pos)++;
                   1535:
1.234     kristaps 1536:                if ( ! roff_evalpar(r, ln, v, pos, &operand2))
1.204     schwarze 1537:                        return(0);
                   1538:
                   1539:                if (skipwhite)
                   1540:                        while (isspace((unsigned char)v[*pos]))
                   1541:                                (*pos)++;
1.206     schwarze 1542:
                   1543:                if (NULL == res)
                   1544:                        continue;
1.204     schwarze 1545:
                   1546:                switch (operator) {
1.207     schwarze 1547:                case '+':
1.204     schwarze 1548:                        *res += operand2;
                   1549:                        break;
1.207     schwarze 1550:                case '-':
1.204     schwarze 1551:                        *res -= operand2;
                   1552:                        break;
1.207     schwarze 1553:                case '*':
1.204     schwarze 1554:                        *res *= operand2;
                   1555:                        break;
1.207     schwarze 1556:                case '/':
1.234     kristaps 1557:                        if (0 == operand2) {
1.235     schwarze 1558:                                mandoc_msg(MANDOCERR_DIVZERO,
1.234     kristaps 1559:                                        r->parse, ln, *pos, v);
                   1560:                                *res = 0;
                   1561:                                break;
                   1562:                        }
1.204     schwarze 1563:                        *res /= operand2;
                   1564:                        break;
1.207     schwarze 1565:                case '%':
1.204     schwarze 1566:                        *res %= operand2;
                   1567:                        break;
1.207     schwarze 1568:                case '<':
1.204     schwarze 1569:                        *res = *res < operand2;
                   1570:                        break;
1.207     schwarze 1571:                case '>':
1.204     schwarze 1572:                        *res = *res > operand2;
                   1573:                        break;
1.207     schwarze 1574:                case 'l':
1.204     schwarze 1575:                        *res = *res <= operand2;
                   1576:                        break;
1.207     schwarze 1577:                case 'g':
1.204     schwarze 1578:                        *res = *res >= operand2;
                   1579:                        break;
1.207     schwarze 1580:                case '=':
1.204     schwarze 1581:                        *res = *res == operand2;
                   1582:                        break;
1.207     schwarze 1583:                case '!':
1.204     schwarze 1584:                        *res = *res != operand2;
                   1585:                        break;
1.207     schwarze 1586:                case '&':
1.204     schwarze 1587:                        *res = *res && operand2;
                   1588:                        break;
1.207     schwarze 1589:                case ':':
1.204     schwarze 1590:                        *res = *res || operand2;
                   1591:                        break;
1.207     schwarze 1592:                case 'i':
1.204     schwarze 1593:                        if (operand2 < *res)
                   1594:                                *res = operand2;
                   1595:                        break;
1.207     schwarze 1596:                case 'a':
1.204     schwarze 1597:                        if (operand2 > *res)
                   1598:                                *res = operand2;
                   1599:                        break;
                   1600:                default:
                   1601:                        abort();
                   1602:                }
                   1603:        }
                   1604:        return(1);
                   1605: }
                   1606:
1.180     schwarze 1607: void
1.187     schwarze 1608: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147     kristaps 1609: {
1.180     schwarze 1610:        struct roffreg  *reg;
                   1611:
                   1612:        /* Search for an existing register with the same name. */
                   1613:        reg = r->regtab;
                   1614:
                   1615:        while (reg && strcmp(name, reg->key.p))
                   1616:                reg = reg->next;
1.147     kristaps 1617:
1.180     schwarze 1618:        if (NULL == reg) {
                   1619:                /* Create a new register. */
                   1620:                reg = mandoc_malloc(sizeof(struct roffreg));
                   1621:                reg->key.p = mandoc_strdup(name);
                   1622:                reg->key.sz = strlen(name);
1.187     schwarze 1623:                reg->val = 0;
1.180     schwarze 1624:                reg->next = r->regtab;
                   1625:                r->regtab = reg;
                   1626:        }
                   1627:
1.187     schwarze 1628:        if ('+' == sign)
                   1629:                reg->val += val;
                   1630:        else if ('-' == sign)
                   1631:                reg->val -= val;
                   1632:        else
                   1633:                reg->val = val;
1.147     kristaps 1634: }
                   1635:
1.192     schwarze 1636: /*
                   1637:  * Handle some predefined read-only number registers.
                   1638:  * For now, return -1 if the requested register is not predefined;
                   1639:  * in case a predefined read-only register having the value -1
                   1640:  * were to turn up, another special value would have to be chosen.
                   1641:  */
                   1642: static int
                   1643: roff_getregro(const char *name)
                   1644: {
                   1645:
                   1646:        switch (*name) {
1.207     schwarze 1647:        case 'A':  /* ASCII approximation mode is always off. */
1.192     schwarze 1648:                return(0);
1.207     schwarze 1649:        case 'g':  /* Groff compatibility mode is always on. */
1.192     schwarze 1650:                return(1);
1.207     schwarze 1651:        case 'H':  /* Fixed horizontal resolution. */
1.192     schwarze 1652:                return (24);
1.207     schwarze 1653:        case 'j':  /* Always adjust left margin only. */
1.192     schwarze 1654:                return(0);
1.207     schwarze 1655:        case 'T':  /* Some output device is always defined. */
1.192     schwarze 1656:                return(1);
1.207     schwarze 1657:        case 'V':  /* Fixed vertical resolution. */
1.192     schwarze 1658:                return (40);
                   1659:        default:
                   1660:                return (-1);
                   1661:        }
                   1662: }
                   1663:
1.181     schwarze 1664: int
1.180     schwarze 1665: roff_getreg(const struct roff *r, const char *name)
1.147     kristaps 1666: {
1.180     schwarze 1667:        struct roffreg  *reg;
1.192     schwarze 1668:        int              val;
                   1669:
                   1670:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
                   1671:                val = roff_getregro(name + 1);
                   1672:                if (-1 != val)
                   1673:                        return (val);
                   1674:        }
1.180     schwarze 1675:
                   1676:        for (reg = r->regtab; reg; reg = reg->next)
                   1677:                if (0 == strcmp(name, reg->key.p))
1.181     schwarze 1678:                        return(reg->val);
                   1679:
                   1680:        return(0);
                   1681: }
                   1682:
                   1683: static int
                   1684: roff_getregn(const struct roff *r, const char *name, size_t len)
                   1685: {
                   1686:        struct roffreg  *reg;
1.192     schwarze 1687:        int              val;
                   1688:
                   1689:        if ('.' == name[0] && 2 == len) {
                   1690:                val = roff_getregro(name + 1);
                   1691:                if (-1 != val)
                   1692:                        return (val);
                   1693:        }
1.181     schwarze 1694:
                   1695:        for (reg = r->regtab; reg; reg = reg->next)
                   1696:                if (len == reg->key.sz &&
                   1697:                    0 == strncmp(name, reg->key.p, len))
                   1698:                        return(reg->val);
1.147     kristaps 1699:
1.180     schwarze 1700:        return(0);
1.147     kristaps 1701: }
                   1702:
1.180     schwarze 1703: static void
                   1704: roff_freereg(struct roffreg *reg)
1.147     kristaps 1705: {
1.180     schwarze 1706:        struct roffreg  *old_reg;
1.147     kristaps 1707:
1.180     schwarze 1708:        while (NULL != reg) {
                   1709:                free(reg->key.p);
                   1710:                old_reg = reg;
                   1711:                reg = reg->next;
                   1712:                free(old_reg);
                   1713:        }
1.147     kristaps 1714: }
1.92      schwarze 1715:
                   1716: static enum rofferr
1.89      kristaps 1717: roff_nr(ROFF_ARGS)
1.83      schwarze 1718: {
1.212     schwarze 1719:        char            *key, *val;
                   1720:        size_t           keysz;
1.138     kristaps 1721:        int              iv;
1.187     schwarze 1722:        char             sign;
1.89      kristaps 1723:
1.238   ! schwarze 1724:        key = val = buf->buf + pos;
        !          1725:        if (*key == '\0')
1.212     schwarze 1726:                return(ROFF_IGN);
                   1727:
                   1728:        keysz = roff_getname(r, &val, ln, pos);
1.238   ! schwarze 1729:        if (key[keysz] == '\\')
1.212     schwarze 1730:                return(ROFF_IGN);
                   1731:        key[keysz] = '\0';
1.89      kristaps 1732:
1.187     schwarze 1733:        sign = *val;
1.238   ! schwarze 1734:        if (sign == '+' || sign == '-')
1.187     schwarze 1735:                val++;
                   1736:
1.234     kristaps 1737:        if (roff_evalnum(r, ln, val, NULL, &iv, 0))
1.204     schwarze 1738:                roff_setreg(r, key, iv, sign);
1.109     kristaps 1739:
1.203     schwarze 1740:        return(ROFF_IGN);
                   1741: }
                   1742:
                   1743: static enum rofferr
                   1744: roff_rr(ROFF_ARGS)
                   1745: {
                   1746:        struct roffreg  *reg, **prev;
1.212     schwarze 1747:        char            *name, *cp;
                   1748:        size_t           namesz;
1.203     schwarze 1749:
1.238   ! schwarze 1750:        name = cp = buf->buf + pos;
        !          1751:        if (*name == '\0')
1.212     schwarze 1752:                return(ROFF_IGN);
                   1753:        namesz = roff_getname(r, &cp, ln, pos);
                   1754:        name[namesz] = '\0';
1.203     schwarze 1755:
                   1756:        prev = &r->regtab;
                   1757:        while (1) {
                   1758:                reg = *prev;
1.238   ! schwarze 1759:                if (reg == NULL || !strcmp(name, reg->key.p))
1.203     schwarze 1760:                        break;
                   1761:                prev = &reg->next;
                   1762:        }
1.238   ! schwarze 1763:        if (reg != NULL) {
1.203     schwarze 1764:                *prev = reg->next;
                   1765:                free(reg->key.p);
                   1766:                free(reg);
                   1767:        }
1.122     schwarze 1768:        return(ROFF_IGN);
                   1769: }
                   1770:
                   1771: static enum rofferr
                   1772: roff_rm(ROFF_ARGS)
                   1773: {
                   1774:        const char       *name;
                   1775:        char             *cp;
1.212     schwarze 1776:        size_t            namesz;
1.122     schwarze 1777:
1.238   ! schwarze 1778:        cp = buf->buf + pos;
        !          1779:        while (*cp != '\0') {
1.212     schwarze 1780:                name = cp;
1.238   ! schwarze 1781:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212     schwarze 1782:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238   ! schwarze 1783:                if (name[namesz] == '\\')
1.212     schwarze 1784:                        break;
1.122     schwarze 1785:        }
1.178     schwarze 1786:        return(ROFF_IGN);
                   1787: }
                   1788:
                   1789: static enum rofferr
                   1790: roff_it(ROFF_ARGS)
                   1791: {
                   1792:        char            *cp;
                   1793:        size_t           len;
                   1794:        int              iv;
                   1795:
                   1796:        /* Parse the number of lines. */
1.238   ! schwarze 1797:        cp = buf->buf + pos;
1.178     schwarze 1798:        len = strcspn(cp, " \t");
                   1799:        cp[len] = '\0';
                   1800:        if ((iv = mandoc_strntoi(cp, len, 10)) <= 0) {
1.222     schwarze 1801:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238   ! schwarze 1802:                    ln, ppos, buf->buf + 1);
1.178     schwarze 1803:                return(ROFF_IGN);
                   1804:        }
                   1805:        cp += len + 1;
                   1806:
                   1807:        /* Arm the input line trap. */
                   1808:        roffit_lines = iv;
                   1809:        roffit_macro = mandoc_strdup(cp);
1.109     kristaps 1810:        return(ROFF_IGN);
1.175     schwarze 1811: }
                   1812:
                   1813: static enum rofferr
                   1814: roff_Dd(ROFF_ARGS)
                   1815: {
                   1816:        const char *const       *cp;
                   1817:
1.227     schwarze 1818:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175     schwarze 1819:                for (cp = __mdoc_reserved; *cp; cp++)
                   1820:                        roff_setstr(r, *cp, NULL, 0);
                   1821:
1.227     schwarze 1822:        if (r->format == 0)
                   1823:                r->format = MPARSE_MDOC;
                   1824:
1.175     schwarze 1825:        return(ROFF_CONT);
                   1826: }
                   1827:
                   1828: static enum rofferr
                   1829: roff_TH(ROFF_ARGS)
                   1830: {
                   1831:        const char *const       *cp;
                   1832:
1.227     schwarze 1833:        if ((r->options & MPARSE_QUICK) == 0)
1.175     schwarze 1834:                for (cp = __man_reserved; *cp; cp++)
                   1835:                        roff_setstr(r, *cp, NULL, 0);
                   1836:
1.227     schwarze 1837:        if (r->format == 0)
                   1838:                r->format = MPARSE_MAN;
                   1839:
1.175     schwarze 1840:        return(ROFF_CONT);
1.109     kristaps 1841: }
                   1842:
                   1843: static enum rofferr
                   1844: roff_TE(ROFF_ARGS)
                   1845: {
                   1846:
                   1847:        if (NULL == r->tbl)
1.221     schwarze 1848:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1849:                    ln, ppos, "TE");
1.115     kristaps 1850:        else
1.151     kristaps 1851:                tbl_end(&r->tbl);
1.109     kristaps 1852:
1.112     kristaps 1853:        return(ROFF_IGN);
                   1854: }
                   1855:
                   1856: static enum rofferr
                   1857: roff_T_(ROFF_ARGS)
                   1858: {
                   1859:
                   1860:        if (NULL == r->tbl)
1.221     schwarze 1861:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1862:                    ln, ppos, "T&");
1.112     kristaps 1863:        else
1.116     kristaps 1864:                tbl_restart(ppos, ln, r->tbl);
1.112     kristaps 1865:
1.109     kristaps 1866:        return(ROFF_IGN);
                   1867: }
                   1868:
1.230     schwarze 1869: /*
                   1870:  * Handle in-line equation delimiters.
                   1871:  */
                   1872: static enum rofferr
1.238   ! schwarze 1873: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151     kristaps 1874: {
1.233     schwarze 1875:        char            *cp1, *cp2;
                   1876:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151     kristaps 1877:
1.230     schwarze 1878:        /*
                   1879:         * Outside equations, look for an opening delimiter.
                   1880:         * If we are inside an equation, we already know it is
                   1881:         * in-line, or this function wouldn't have been called;
                   1882:         * so look for a closing delimiter.
                   1883:         */
                   1884:
1.238   ! schwarze 1885:        cp1 = buf->buf + pos;
1.230     schwarze 1886:        cp2 = strchr(cp1, r->eqn == NULL ?
                   1887:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   1888:        if (cp2 == NULL)
                   1889:                return(ROFF_CONT);
                   1890:
1.233     schwarze 1891:        *cp2++ = '\0';
                   1892:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   1893:
                   1894:        /* Handle preceding text, protecting whitespace. */
                   1895:
1.238   ! schwarze 1896:        if (*buf->buf != '\0') {
1.233     schwarze 1897:                if (r->eqn == NULL)
                   1898:                        bef_pr = "\\&";
                   1899:                bef_nl = "\n";
                   1900:        }
                   1901:
                   1902:        /*
                   1903:         * Prepare replacing the delimiter with an equation macro
                   1904:         * and drop leading white space from the equation.
                   1905:         */
1.230     schwarze 1906:
1.233     schwarze 1907:        if (r->eqn == NULL) {
                   1908:                while (*cp2 == ' ')
                   1909:                        cp2++;
                   1910:                mac = ".EQ";
                   1911:        } else
                   1912:                mac = ".EN";
                   1913:
                   1914:        /* Handle following text, protecting whitespace. */
                   1915:
                   1916:        if (*cp2 != '\0') {
                   1917:                aft_nl = "\n";
                   1918:                if (r->eqn != NULL)
                   1919:                        aft_pr = "\\&";
                   1920:        }
                   1921:
                   1922:        /* Do the actual replacement. */
                   1923:
1.238   ! schwarze 1924:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233     schwarze 1925:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238   ! schwarze 1926:        free(buf->buf);
        !          1927:        buf->buf = cp1;
1.230     schwarze 1928:
                   1929:        /* Toggle the in-line state of the eqn subsystem. */
                   1930:
                   1931:        r->eqn_inline = r->eqn == NULL;
                   1932:        return(ROFF_REPARSE);
1.151     kristaps 1933: }
                   1934:
1.235     schwarze 1935: static enum rofferr
                   1936: roff_EQ(ROFF_ARGS)
1.125     kristaps 1937: {
1.151     kristaps 1938:        struct eqn_node *e;
1.125     kristaps 1939:
1.238   ! schwarze 1940:        assert(r->eqn == NULL);
1.235     schwarze 1941:        e = eqn_alloc(ppos, ln, r->parse);
1.125     kristaps 1942:
1.230     schwarze 1943:        if (r->last_eqn) {
1.125     kristaps 1944:                r->last_eqn->next = e;
1.230     schwarze 1945:                e->delim = r->last_eqn->delim;
                   1946:                e->odelim = r->last_eqn->odelim;
                   1947:                e->cdelim = r->last_eqn->cdelim;
                   1948:        } else
1.125     kristaps 1949:                r->first_eqn = r->last_eqn = e;
                   1950:
                   1951:        r->eqn = r->last_eqn = e;
1.151     kristaps 1952:
1.238   ! schwarze 1953:        if (buf->buf[pos] != '\0')
1.235     schwarze 1954:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238   ! schwarze 1955:                    ".EQ %s", buf->buf + pos);
1.151     kristaps 1956:
1.125     kristaps 1957:        return(ROFF_IGN);
                   1958: }
                   1959:
                   1960: static enum rofferr
                   1961: roff_EN(ROFF_ARGS)
                   1962: {
                   1963:
1.221     schwarze 1964:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.125     kristaps 1965:        return(ROFF_IGN);
                   1966: }
                   1967:
                   1968: static enum rofferr
1.109     kristaps 1969: roff_TS(ROFF_ARGS)
                   1970: {
1.176     schwarze 1971:        struct tbl_node *tbl;
1.89      kristaps 1972:
1.115     kristaps 1973:        if (r->tbl) {
1.221     schwarze 1974:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   1975:                    ln, ppos, "TS breaks TS");
1.151     kristaps 1976:                tbl_end(&r->tbl);
1.115     kristaps 1977:        }
1.83      schwarze 1978:
1.176     schwarze 1979:        tbl = tbl_alloc(ppos, ln, r->parse);
1.113     kristaps 1980:
                   1981:        if (r->last_tbl)
1.176     schwarze 1982:                r->last_tbl->next = tbl;
1.113     kristaps 1983:        else
1.176     schwarze 1984:                r->first_tbl = r->last_tbl = tbl;
1.113     kristaps 1985:
1.176     schwarze 1986:        r->tbl = r->last_tbl = tbl;
1.83      schwarze 1987:        return(ROFF_IGN);
1.92      schwarze 1988: }
                   1989:
1.105     kristaps 1990: static enum rofferr
1.174     kristaps 1991: roff_cc(ROFF_ARGS)
                   1992: {
                   1993:        const char      *p;
                   1994:
1.238   ! schwarze 1995:        p = buf->buf + pos;
1.174     kristaps 1996:
1.238   ! schwarze 1997:        if (*p == '\0' || (r->control = *p++) == '.')
1.174     kristaps 1998:                r->control = 0;
                   1999:
1.238   ! schwarze 2000:        if (*p != '\0')
1.174     kristaps 2001:                mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
                   2002:
                   2003:        return(ROFF_IGN);
                   2004: }
                   2005:
                   2006: static enum rofferr
1.164     kristaps 2007: roff_tr(ROFF_ARGS)
                   2008: {
                   2009:        const char      *p, *first, *second;
                   2010:        size_t           fsz, ssz;
                   2011:        enum mandoc_esc  esc;
                   2012:
1.238   ! schwarze 2013:        p = buf->buf + pos;
1.164     kristaps 2014:
1.238   ! schwarze 2015:        if (*p == '\0') {
1.164     kristaps 2016:                mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
                   2017:                return(ROFF_IGN);
                   2018:        }
                   2019:
1.238   ! schwarze 2020:        while (*p != '\0') {
1.164     kristaps 2021:                fsz = ssz = 1;
                   2022:
                   2023:                first = p++;
1.238   ! schwarze 2024:                if (*first == '\\') {
1.164     kristaps 2025:                        esc = mandoc_escape(&p, NULL, NULL);
1.238   ! schwarze 2026:                        if (esc == ESCAPE_ERROR) {
1.219     schwarze 2027:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238   ! schwarze 2028:                                    ln, (int)(p - buf->buf), first);
1.164     kristaps 2029:                                return(ROFF_IGN);
                   2030:                        }
                   2031:                        fsz = (size_t)(p - first);
                   2032:                }
                   2033:
                   2034:                second = p++;
1.238   ! schwarze 2035:                if (*second == '\\') {
1.164     kristaps 2036:                        esc = mandoc_escape(&p, NULL, NULL);
1.238   ! schwarze 2037:                        if (esc == ESCAPE_ERROR) {
1.219     schwarze 2038:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238   ! schwarze 2039:                                    ln, (int)(p - buf->buf), second);
1.164     kristaps 2040:                                return(ROFF_IGN);
                   2041:                        }
                   2042:                        ssz = (size_t)(p - second);
1.238   ! schwarze 2043:                } else if (*second == '\0') {
1.207     schwarze 2044:                        mandoc_msg(MANDOCERR_ARGCOUNT, r->parse,
1.238   ! schwarze 2045:                            ln, (int)(p - buf->buf), NULL);
1.164     kristaps 2046:                        second = " ";
1.165     kristaps 2047:                        p--;
1.164     kristaps 2048:                }
                   2049:
1.167     kristaps 2050:                if (fsz > 1) {
1.207     schwarze 2051:                        roff_setstrn(&r->xmbtab, first, fsz,
                   2052:                            second, ssz, 0);
1.167     kristaps 2053:                        continue;
                   2054:                }
                   2055:
1.238   ! schwarze 2056:                if (r->xtab == NULL)
1.207     schwarze 2057:                        r->xtab = mandoc_calloc(128,
                   2058:                            sizeof(struct roffstr));
1.167     kristaps 2059:
                   2060:                free(r->xtab[(int)*first].p);
                   2061:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   2062:                r->xtab[(int)*first].sz = ssz;
1.164     kristaps 2063:        }
                   2064:
                   2065:        return(ROFF_IGN);
                   2066: }
                   2067:
                   2068: static enum rofferr
1.105     kristaps 2069: roff_so(ROFF_ARGS)
                   2070: {
                   2071:        char *name;
                   2072:
1.238   ! schwarze 2073:        name = buf->buf + pos;
1.224     schwarze 2074:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105     kristaps 2075:
                   2076:        /*
                   2077:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   2078:         * opening anything that's not in our cwd or anything beneath
                   2079:         * it.  Thus, explicitly disallow traversing up the file-system
                   2080:         * or using absolute paths.
                   2081:         */
                   2082:
1.238   ! schwarze 2083:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210     schwarze 2084:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   2085:                    ".so %s", name);
1.105     kristaps 2086:                return(ROFF_ERR);
                   2087:        }
                   2088:
                   2089:        *offs = pos;
                   2090:        return(ROFF_SO);
                   2091: }
1.92      schwarze 2092:
1.106     kristaps 2093: static enum rofferr
                   2094: roff_userdef(ROFF_ARGS)
1.99      kristaps 2095: {
1.106     kristaps 2096:        const char       *arg[9];
                   2097:        char             *cp, *n1, *n2;
1.119     schwarze 2098:        int               i;
1.106     kristaps 2099:
                   2100:        /*
                   2101:         * Collect pointers to macro argument strings
1.188     schwarze 2102:         * and NUL-terminate them.
1.106     kristaps 2103:         */
1.238   ! schwarze 2104:        cp = buf->buf + pos;
1.119     schwarze 2105:        for (i = 0; i < 9; i++)
1.238   ! schwarze 2106:                arg[i] = *cp == '\0' ? "" :
1.136     kristaps 2107:                    mandoc_getarg(r->parse, &cp, ln, &pos);
1.99      kristaps 2108:
1.106     kristaps 2109:        /*
                   2110:         * Expand macro arguments.
1.99      kristaps 2111:         */
1.238   ! schwarze 2112:        buf->sz = 0;
1.106     kristaps 2113:        n1 = cp = mandoc_strdup(r->current_string);
1.238   ! schwarze 2114:        while ((cp = strstr(cp, "\\$")) != NULL) {
1.106     kristaps 2115:                i = cp[2] - '1';
                   2116:                if (0 > i || 8 < i) {
                   2117:                        /* Not an argument invocation. */
                   2118:                        cp += 2;
                   2119:                        continue;
                   2120:                }
1.209     schwarze 2121:                *cp = '\0';
1.238   ! schwarze 2122:                buf->sz = mandoc_asprintf(&n2, "%s%s%s",
1.209     schwarze 2123:                    n1, arg[i], cp + 3) + 1;
1.106     kristaps 2124:                cp = n2 + (cp - n1);
                   2125:                free(n1);
                   2126:                n1 = n2;
1.99      kristaps 2127:        }
                   2128:
1.106     kristaps 2129:        /*
                   2130:         * Replace the macro invocation
                   2131:         * by the expanded macro.
                   2132:         */
1.238   ! schwarze 2133:        free(buf->buf);
        !          2134:        buf->buf = n1;
        !          2135:        if (buf->sz == 0)
        !          2136:                buf->sz = strlen(buf->buf) + 1;
1.106     kristaps 2137:
1.238   ! schwarze 2138:        return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.106     kristaps 2139:           ROFF_REPARSE : ROFF_APPEND);
1.99      kristaps 2140: }
1.121     schwarze 2141:
1.212     schwarze 2142: static size_t
1.121     schwarze 2143: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   2144: {
                   2145:        char     *name, *cp;
1.212     schwarze 2146:        size_t    namesz;
1.121     schwarze 2147:
                   2148:        name = *cpp;
                   2149:        if ('\0' == *name)
1.212     schwarze 2150:                return(0);
1.121     schwarze 2151:
1.212     schwarze 2152:        /* Read until end of name and terminate it with NUL. */
                   2153:        for (cp = name; 1; cp++) {
                   2154:                if ('\0' == *cp || ' ' == *cp) {
                   2155:                        namesz = cp - name;
                   2156:                        break;
                   2157:                }
1.121     schwarze 2158:                if ('\\' != *cp)
                   2159:                        continue;
1.215     schwarze 2160:                namesz = cp - name;
                   2161:                if ('{' == cp[1] || '}' == cp[1])
                   2162:                        break;
1.121     schwarze 2163:                cp++;
                   2164:                if ('\\' == *cp)
                   2165:                        continue;
1.224     schwarze 2166:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   2167:                    "%.*s", (int)(cp - name + 1), name);
1.212     schwarze 2168:                mandoc_escape((const char **)&cp, NULL, NULL);
                   2169:                break;
1.121     schwarze 2170:        }
                   2171:
                   2172:        /* Read past spaces. */
                   2173:        while (' ' == *cp)
                   2174:                cp++;
                   2175:
                   2176:        *cpp = cp;
1.212     schwarze 2177:        return(namesz);
1.121     schwarze 2178: }
                   2179:
1.106     kristaps 2180: /*
                   2181:  * Store *string into the user-defined string called *name.
                   2182:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193     schwarze 2183:  * append == 0: replace mode
                   2184:  * append == 1: single-line append mode
                   2185:  * append == 2: multiline append mode, append '\n' after each call
1.106     kristaps 2186:  */
1.94      kristaps 2187: static void
1.106     kristaps 2188: roff_setstr(struct roff *r, const char *name, const char *string,
1.193     schwarze 2189:        int append)
1.92      schwarze 2190: {
1.164     kristaps 2191:
                   2192:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.207     schwarze 2193:            string ? strlen(string) : 0, append);
1.164     kristaps 2194: }
                   2195:
                   2196: static void
1.166     kristaps 2197: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193     schwarze 2198:                const char *string, size_t stringsz, int append)
1.164     kristaps 2199: {
1.166     kristaps 2200:        struct roffkv   *n;
1.164     kristaps 2201:        char            *c;
                   2202:        int              i;
                   2203:        size_t           oldch, newch;
1.92      schwarze 2204:
1.106     kristaps 2205:        /* Search for an existing string with the same name. */
1.164     kristaps 2206:        n = *r;
                   2207:
1.211     schwarze 2208:        while (n && (namesz != n->key.sz ||
                   2209:                        strncmp(n->key.p, name, namesz)))
1.92      schwarze 2210:                n = n->next;
1.94      kristaps 2211:
                   2212:        if (NULL == n) {
1.106     kristaps 2213:                /* Create a new string table entry. */
1.166     kristaps 2214:                n = mandoc_malloc(sizeof(struct roffkv));
                   2215:                n->key.p = mandoc_strndup(name, namesz);
                   2216:                n->key.sz = namesz;
                   2217:                n->val.p = NULL;
                   2218:                n->val.sz = 0;
1.164     kristaps 2219:                n->next = *r;
                   2220:                *r = n;
1.193     schwarze 2221:        } else if (0 == append) {
1.166     kristaps 2222:                free(n->val.p);
                   2223:                n->val.p = NULL;
                   2224:                n->val.sz = 0;
1.106     kristaps 2225:        }
                   2226:
                   2227:        if (NULL == string)
                   2228:                return;
                   2229:
                   2230:        /*
                   2231:         * One additional byte for the '\n' in multiline mode,
                   2232:         * and one for the terminating '\0'.
                   2233:         */
1.193     schwarze 2234:        newch = stringsz + (1 < append ? 2u : 1u);
1.164     kristaps 2235:
1.166     kristaps 2236:        if (NULL == n->val.p) {
                   2237:                n->val.p = mandoc_malloc(newch);
                   2238:                *n->val.p = '\0';
1.106     kristaps 2239:                oldch = 0;
                   2240:        } else {
1.166     kristaps 2241:                oldch = n->val.sz;
                   2242:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106     kristaps 2243:        }
                   2244:
                   2245:        /* Skip existing content in the destination buffer. */
1.166     kristaps 2246:        c = n->val.p + (int)oldch;
1.106     kristaps 2247:
                   2248:        /* Append new content to the destination buffer. */
1.164     kristaps 2249:        i = 0;
                   2250:        while (i < (int)stringsz) {
1.106     kristaps 2251:                /*
                   2252:                 * Rudimentary roff copy mode:
                   2253:                 * Handle escaped backslashes.
                   2254:                 */
1.164     kristaps 2255:                if ('\\' == string[i] && '\\' == string[i + 1])
                   2256:                        i++;
                   2257:                *c++ = string[i++];
1.106     kristaps 2258:        }
1.94      kristaps 2259:
1.106     kristaps 2260:        /* Append terminating bytes. */
1.193     schwarze 2261:        if (1 < append)
1.106     kristaps 2262:                *c++ = '\n';
1.163     kristaps 2263:
1.106     kristaps 2264:        *c = '\0';
1.166     kristaps 2265:        n->val.sz = (int)(c - n->val.p);
1.92      schwarze 2266: }
                   2267:
1.94      kristaps 2268: static const char *
                   2269: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92      schwarze 2270: {
1.166     kristaps 2271:        const struct roffkv *n;
1.191     schwarze 2272:        int i;
1.92      schwarze 2273:
1.164     kristaps 2274:        for (n = r->strtab; n; n = n->next)
1.207     schwarze 2275:                if (0 == strncmp(name, n->key.p, len) &&
                   2276:                    '\0' == n->key.p[(int)len])
1.166     kristaps 2277:                        return(n->val.p);
1.191     schwarze 2278:
                   2279:        for (i = 0; i < PREDEFS_MAX; i++)
                   2280:                if (0 == strncmp(name, predefs[i].name, len) &&
                   2281:                                '\0' == predefs[i].name[(int)len])
                   2282:                        return(predefs[i].str);
1.94      kristaps 2283:
1.157     kristaps 2284:        return(NULL);
1.92      schwarze 2285: }
                   2286:
1.94      kristaps 2287: static void
1.167     kristaps 2288: roff_freestr(struct roffkv *r)
1.92      schwarze 2289: {
1.166     kristaps 2290:        struct roffkv    *n, *nn;
1.92      schwarze 2291:
1.167     kristaps 2292:        for (n = r; n; n = nn) {
1.166     kristaps 2293:                free(n->key.p);
                   2294:                free(n->val.p);
1.92      schwarze 2295:                nn = n->next;
                   2296:                free(n);
                   2297:        }
1.114     kristaps 2298: }
                   2299:
                   2300: const struct tbl_span *
                   2301: roff_span(const struct roff *r)
                   2302: {
1.207     schwarze 2303:
1.114     kristaps 2304:        return(r->tbl ? tbl_span(r->tbl) : NULL);
1.125     kristaps 2305: }
                   2306:
                   2307: const struct eqn *
                   2308: roff_eqn(const struct roff *r)
                   2309: {
1.207     schwarze 2310:
1.125     kristaps 2311:        return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.164     kristaps 2312: }
                   2313:
                   2314: /*
                   2315:  * Duplicate an input string, making the appropriate character
                   2316:  * conversations (as stipulated by `tr') along the way.
                   2317:  * Returns a heap-allocated string with all the replacements made.
                   2318:  */
                   2319: char *
                   2320: roff_strdup(const struct roff *r, const char *p)
                   2321: {
1.166     kristaps 2322:        const struct roffkv *cp;
1.164     kristaps 2323:        char            *res;
                   2324:        const char      *pp;
                   2325:        size_t           ssz, sz;
                   2326:        enum mandoc_esc  esc;
                   2327:
1.167     kristaps 2328:        if (NULL == r->xmbtab && NULL == r->xtab)
1.164     kristaps 2329:                return(mandoc_strdup(p));
                   2330:        else if ('\0' == *p)
                   2331:                return(mandoc_strdup(""));
                   2332:
                   2333:        /*
                   2334:         * Step through each character looking for term matches
                   2335:         * (remember that a `tr' can be invoked with an escape, which is
                   2336:         * a glyph but the escape is multi-character).
                   2337:         * We only do this if the character hash has been initialised
                   2338:         * and the string is >0 length.
                   2339:         */
                   2340:
                   2341:        res = NULL;
                   2342:        ssz = 0;
                   2343:
                   2344:        while ('\0' != *p) {
1.167     kristaps 2345:                if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
                   2346:                        sz = r->xtab[(int)*p].sz;
                   2347:                        res = mandoc_realloc(res, ssz + sz + 1);
                   2348:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   2349:                        ssz += sz;
                   2350:                        p++;
                   2351:                        continue;
                   2352:                } else if ('\\' != *p) {
                   2353:                        res = mandoc_realloc(res, ssz + 2);
                   2354:                        res[ssz++] = *p++;
                   2355:                        continue;
                   2356:                }
                   2357:
1.164     kristaps 2358:                /* Search for term matches. */
1.167     kristaps 2359:                for (cp = r->xmbtab; cp; cp = cp->next)
1.166     kristaps 2360:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164     kristaps 2361:                                break;
                   2362:
                   2363:                if (NULL != cp) {
                   2364:                        /*
                   2365:                         * A match has been found.
                   2366:                         * Append the match to the array and move
                   2367:                         * forward by its keysize.
                   2368:                         */
1.207     schwarze 2369:                        res = mandoc_realloc(res,
                   2370:                            ssz + cp->val.sz + 1);
1.166     kristaps 2371:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   2372:                        ssz += cp->val.sz;
                   2373:                        p += (int)cp->key.sz;
1.164     kristaps 2374:                        continue;
                   2375:                }
                   2376:
1.167     kristaps 2377:                /*
                   2378:                 * Handle escapes carefully: we need to copy
                   2379:                 * over just the escape itself, or else we might
                   2380:                 * do replacements within the escape itself.
                   2381:                 * Make sure to pass along the bogus string.
                   2382:                 */
                   2383:                pp = p++;
                   2384:                esc = mandoc_escape(&p, NULL, NULL);
                   2385:                if (ESCAPE_ERROR == esc) {
                   2386:                        sz = strlen(pp);
1.164     kristaps 2387:                        res = mandoc_realloc(res, ssz + sz + 1);
                   2388:                        memcpy(res + ssz, pp, sz);
1.167     kristaps 2389:                        break;
1.164     kristaps 2390:                }
1.207     schwarze 2391:                /*
                   2392:                 * We bail out on bad escapes.
1.167     kristaps 2393:                 * No need to warn: we already did so when
                   2394:                 * roff_res() was called.
                   2395:                 */
                   2396:                sz = (int)(p - pp);
                   2397:                res = mandoc_realloc(res, ssz + sz + 1);
                   2398:                memcpy(res + ssz, pp, sz);
                   2399:                ssz += sz;
1.164     kristaps 2400:        }
                   2401:
                   2402:        res[(int)ssz] = '\0';
                   2403:        return(res);
1.227     schwarze 2404: }
                   2405:
                   2406: int
                   2407: roff_getformat(const struct roff *r)
                   2408: {
                   2409:
                   2410:        return(r->format);
1.174     kristaps 2411: }
                   2412:
                   2413: /*
1.207     schwarze 2414:  * Find out whether a line is a macro line or not.
1.174     kristaps 2415:  * If it is, adjust the current position and return one; if it isn't,
                   2416:  * return zero and don't change the current position.
                   2417:  * If the control character has been set with `.cc', then let that grain
                   2418:  * precedence.
                   2419:  * This is slighly contrary to groff, where using the non-breaking
                   2420:  * control character when `cc' has been invoked will cause the
                   2421:  * non-breaking macro contents to be printed verbatim.
                   2422:  */
                   2423: int
                   2424: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   2425: {
                   2426:        int             pos;
                   2427:
                   2428:        pos = *ppos;
                   2429:
                   2430:        if (0 != r->control && cp[pos] == r->control)
                   2431:                pos++;
                   2432:        else if (0 != r->control)
                   2433:                return(0);
                   2434:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   2435:                pos += 2;
                   2436:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   2437:                pos++;
                   2438:        else
                   2439:                return(0);
                   2440:
                   2441:        while (' ' == cp[pos] || '\t' == cp[pos])
                   2442:                pos++;
                   2443:
                   2444:        *ppos = pos;
                   2445:        return(1);
1.74      kristaps 2446: }

CVSweb