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

Annotation of src/usr.bin/mandoc/roff.c, Revision 1.167

1.167   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.166 2017/04/24 23:06:09 schwarze Exp $ */
1.1       schwarze    2: /*
1.139     schwarze    3:  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.159     schwarze    4:  * Copyright (c) 2010-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
1.1       schwarze    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
                      7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
                      9:  *
1.16      schwarze   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1       schwarze   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16      schwarze   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1       schwarze   13:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     14:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     15:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     16:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
                     17:  */
1.99      schwarze   18: #include <sys/types.h>
                     19:
1.1       schwarze   20: #include <assert.h>
1.3       schwarze   21: #include <ctype.h>
1.117     schwarze   22: #include <limits.h>
1.167   ! schwarze   23: #include <stddef.h>
        !            24: #include <stdint.h>
1.51      schwarze   25: #include <stdio.h>
1.1       schwarze   26: #include <stdlib.h>
                     27: #include <string.h>
                     28:
                     29: #include "mandoc.h"
1.74      schwarze   30: #include "mandoc_aux.h"
1.167   ! schwarze   31: #include "mandoc_ohash.h"
1.137     schwarze   32: #include "roff.h"
1.98      daniel     33: #include "libmandoc.h"
1.138     schwarze   34: #include "roff_int.h"
1.27      schwarze   35: #include "libroff.h"
1.2       schwarze   36:
1.43      schwarze   37: /* Maximum number of string expansions per line, to break infinite loops. */
                     38: #define        EXPAND_LIMIT    1000
                     39:
1.138     schwarze   40: /* --- data types --------------------------------------------------------- */
                     41:
1.41      schwarze   42: /*
1.42      schwarze   43:  * An incredibly-simple string buffer.
                     44:  */
1.8       schwarze   45: struct roffstr {
1.42      schwarze   46:        char            *p; /* nil-terminated buffer */
                     47:        size_t           sz; /* saved strlen(p) */
                     48: };
                     49:
                     50: /*
                     51:  * A key-value roffstr pair as part of a singly-linked list.
                     52:  */
                     53: struct roffkv {
                     54:        struct roffstr   key;
                     55:        struct roffstr   val;
                     56:        struct roffkv   *next; /* next in list */
1.8       schwarze   57: };
                     58:
1.52      schwarze   59: /*
                     60:  * A single number register as part of a singly-linked list.
                     61:  */
                     62: struct roffreg {
                     63:        struct roffstr   key;
1.53      schwarze   64:        int              val;
1.52      schwarze   65:        struct roffreg  *next;
                     66: };
                     67:
1.167   ! schwarze   68: /*
        !            69:  * Association of request and macro names with token IDs.
        !            70:  */
        !            71: struct roffreq {
        !            72:        enum roff_tok    tok;
        !            73:        char             name[];
        !            74: };
        !            75:
1.1       schwarze   76: struct roff {
1.35      schwarze   77:        struct mparse   *parse; /* parse point */
1.1       schwarze   78:        struct roffnode *last; /* leaf of stack */
1.96      schwarze   79:        int             *rstack; /* stack of inverted `ie' values */
1.167   ! schwarze   80:        struct ohash    *reqtab; /* request lookup table */
1.52      schwarze   81:        struct roffreg  *regtab; /* number registers */
1.42      schwarze   82:        struct roffkv   *strtab; /* user-defined strings & macros */
                     83:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     84:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze   85:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   86:        struct tbl_node *first_tbl; /* first table parsed */
                     87:        struct tbl_node *last_tbl; /* last table parsed */
                     88:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze   89:        struct eqn_node *last_eqn; /* last equation parsed */
                     90:        struct eqn_node *first_eqn; /* first equation parsed */
                     91:        struct eqn_node *eqn; /* current equation being parsed */
1.102     schwarze   92:        int              eqn_inline; /* current equation is inline */
1.96      schwarze   93:        int              options; /* parse options */
                     94:        int              rstacksz; /* current size limit of rstack */
                     95:        int              rstackpos; /* position in rstack */
1.99      schwarze   96:        int              format; /* current file in mdoc or man format */
1.145     schwarze   97:        int              argc; /* number of args of the last macro */
1.96      schwarze   98:        char             control; /* control character */
1.1       schwarze   99: };
                    100:
                    101: struct roffnode {
1.166     schwarze  102:        enum roff_tok    tok; /* type of node */
1.1       schwarze  103:        struct roffnode *parent; /* up one in stack */
                    104:        int              line; /* parse line */
                    105:        int              col; /* parse col */
1.16      schwarze  106:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  107:        char            *end; /* end-rules: custom token */
                    108:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  109:        int              rule; /* current evaluation rule */
1.1       schwarze  110: };
                    111:
                    112: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.166     schwarze  113:                         enum roff_tok tok, /* tok of macro */ \
1.110     schwarze  114:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  115:                         int ln, /* parse line */ \
1.2       schwarze  116:                         int ppos, /* original pos in buffer */ \
                    117:                         int pos, /* current pos in buffer */ \
                    118:                         int *offs /* reset offset of buffer data */
1.1       schwarze  119:
                    120: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    121:
                    122: struct roffmac {
1.2       schwarze  123:        roffproc         proc; /* process new macro */
                    124:        roffproc         text; /* process as child text of macro */
                    125:        roffproc         sub; /* process as child of macro */
                    126:        int              flags;
                    127: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.1       schwarze  128: };
                    129:
1.37      schwarze  130: struct predef {
                    131:        const char      *name; /* predefined input name */
                    132:        const char      *str; /* replacement symbol */
                    133: };
                    134:
                    135: #define        PREDEF(__name, __str) \
                    136:        { (__name), (__str) },
                    137:
1.138     schwarze  138: /* --- function prototypes ------------------------------------------------ */
                    139:
1.42      schwarze  140: static void             roffnode_cleanscope(struct roff *);
                    141: static void             roffnode_pop(struct roff *);
1.166     schwarze  142: static void             roffnode_push(struct roff *, enum roff_tok,
1.42      schwarze  143:                                const char *, int, int);
1.2       schwarze  144: static enum rofferr     roff_block(ROFF_ARGS);
                    145: static enum rofferr     roff_block_text(ROFF_ARGS);
                    146: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.123     schwarze  147: static enum rofferr     roff_brp(ROFF_ARGS);
1.2       schwarze  148: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  149: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  150: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  151: static enum rofferr     roff_cond(ROFF_ARGS);
                    152: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    153: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  154: static enum rofferr     roff_ds(ROFF_ARGS);
1.110     schwarze  155: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.143     schwarze  156: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.106     schwarze  157: static int              roff_evalnum(struct roff *, int,
                    158:                                const char *, int *, int *, int);
                    159: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  160:                                const char *, int *, int *, int);
1.71      schwarze  161: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  162: static void             roff_free1(struct roff *);
1.52      schwarze  163: static void             roff_freereg(struct roffreg *);
1.42      schwarze  164: static void             roff_freestr(struct roffkv *);
1.85      schwarze  165: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  166: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  167: static int              roff_getop(const char *, int *, char *);
1.53      schwarze  168: static int              roff_getregn(const struct roff *,
                    169:                                const char *, size_t);
1.145     schwarze  170: static int              roff_getregro(const struct roff *,
                    171:                                const char *name);
1.80      schwarze  172: static const char      *roff_getstrn(const struct roff *,
1.8       schwarze  173:                                const char *, size_t);
1.143     schwarze  174: static int              roff_hasregn(const struct roff *,
                    175:                                const char *, size_t);
1.123     schwarze  176: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  177: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  178: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137     schwarze  179: static void             roff_man_alloc1(struct roff_man *);
                    180: static void             roff_man_free1(struct roff_man *);
1.6       schwarze  181: static enum rofferr     roff_nr(ROFF_ARGS);
1.166     schwarze  182: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.87      schwarze  183:                                int, int);
1.110     schwarze  184: static enum rofferr     roff_parsetext(struct buf *, int, int *);
                    185: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  186: static enum rofferr     roff_rm(ROFF_ARGS);
1.76      schwarze  187: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  188: static void             roff_setstr(struct roff *,
1.16      schwarze  189:                                const char *, const char *, int);
1.80      schwarze  190: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  191:                                size_t, const char *, size_t, int);
1.14      schwarze  192: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  193: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  194: static enum rofferr     roff_Dd(ROFF_ARGS);
                    195: static enum rofferr     roff_TH(ROFF_ARGS);
1.27      schwarze  196: static enum rofferr     roff_TE(ROFF_ARGS);
                    197: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  198: static enum rofferr     roff_EQ(ROFF_ARGS);
                    199: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  200: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  201: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  202: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  203:
1.138     schwarze  204: /* --- constant data ------------------------------------------------------ */
                    205:
1.133     schwarze  206: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    207: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    208:
1.166     schwarze  209: const char *__roff_name[MAN_MAX + 1] = {
                    210:        "ab",           "ad",           "af",           "aln",
                    211:        "als",          "am",           "am1",          "ami",
                    212:        "ami1",         "as",           "as1",          "asciify",
                    213:        "backtrace",    "bd",           "bleedat",      "blm",
                    214:         "box",         "boxa",         "bp",           "BP",
                    215:        "break",        "breakchar",    "brnl",         "brp",
                    216:        "brpnl",        "c2",           "cc",           "ce",
                    217:        "cf",           "cflags",       "ch",           "char",
                    218:        "chop",         "class",        "close",        "CL",
                    219:        "color",        "composite",    "continue",     "cp",
                    220:        "cropat",       "cs",           "cu",           "da",
                    221:        "dch",          "Dd",           "de",           "de1",
                    222:        "defcolor",     "dei",          "dei1",         "device",
                    223:        "devicem",      "di",           "do",           "ds",
                    224:        "ds1",          "dwh",          "dt",           "ec",
                    225:        "ecr",          "ecs",          "el",           "em",
                    226:        "EN",           "eo",           "EP",           "EQ",
                    227:        "errprint",     "ev",           "evc",          "ex",
                    228:        "fallback",     "fam",          "fc",           "fchar",
                    229:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    230:        "fl",           "flig",         "fp",           "fps",
                    231:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    232:        "fzoom",        "gcolor",       "hc",           "hcode",
                    233:        "hidechar",     "hla",          "hlm",          "hpf",
                    234:        "hpfa",         "hpfcode",      "hw",           "hy",
                    235:        "hylang",       "hylen",        "hym",          "hypp",
                    236:        "hys",          "ie",           "if",           "ig",
                    237:        "index",        "it",           "itc",          "IX",
                    238:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    239:        "lc",           "lc_ctype",     "lds",          "length",
                    240:        "letadj",       "lf",           "lg",           "lhang",
                    241:        "linetabs",     "lnr",          "lnrf",         "lpfx",
                    242:        "ls",           "lsm",          "lt",           "mc",
                    243:        "mediasize",    "minss",        "mk",           "mso",
                    244:        "na",           "ne",           "nh",           "nhychar",
                    245:        "nm",           "nn",           "nop",          "nr",
                    246:        "nrf",          "nroff",        "ns",           "nx",
                    247:        "open",         "opena",        "os",           "output",
                    248:        "padj",         "papersize",    "pc",           "pev",
                    249:        "pi",           "PI",           "pl",           "pm",
                    250:        "pn",           "pnr",          "po",           "ps",
                    251:        "psbb",         "pshape",       "pso",          "ptr",
                    252:        "pvs",          "rchar",        "rd",           "recursionlimit",
                    253:        "return",       "rfschar",      "rhang",        "rj",
                    254:        "rm",           "rn",           "rnn",          "rr",
                    255:        "rs",           "rt",           "schar",        "sentchar",
                    256:        "shc",          "shift",        "sizes",        "so",
                    257:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    258:        "sty",          "substring",    "sv",           "sy",
                    259:        "T&",           "ta",           "tc",           "TE",
                    260:        "TH",           "ti",           "tkf",          "tl",
                    261:        "tm",           "tm1",          "tmc",          "tr",
                    262:        "track",        "transchar",    "trf",          "trimat",
                    263:        "trin",         "trnt",         "troff",        "TS",
                    264:        "uf",           "ul",           "unformat",     "unwatch",
                    265:        "unwatchn",     "vpt",          "vs",           "warn",
                    266:        "warnscale",    "watch",        "watchlength",  "watchn",
                    267:        "wh",           "while",        "write",        "writec",
                    268:        "writem",       "xflag",        ".",            NULL,
                    269:        "text",
                    270:        "Dd",           "Dt",           "Os",           "Sh",
                    271:        "Ss",           "Pp",           "D1",           "Dl",
                    272:        "Bd",           "Ed",           "Bl",           "El",
                    273:        "It",           "Ad",           "An",           "Ap",
                    274:        "Ar",           "Cd",           "Cm",           "Dv",
                    275:        "Er",           "Ev",           "Ex",           "Fa",
                    276:        "Fd",           "Fl",           "Fn",           "Ft",
                    277:        "Ic",           "In",           "Li",           "Nd",
                    278:        "Nm",           "Op",           "Ot",           "Pa",
                    279:        "Rv",           "St",           "Va",           "Vt",
                    280:        "Xr",           "%A",           "%B",           "%D",
                    281:        "%I",           "%J",           "%N",           "%O",
                    282:        "%P",           "%R",           "%T",           "%V",
                    283:        "Ac",           "Ao",           "Aq",           "At",
                    284:        "Bc",           "Bf",           "Bo",           "Bq",
                    285:        "Bsx",          "Bx",           "Db",           "Dc",
                    286:        "Do",           "Dq",           "Ec",           "Ef",
                    287:        "Em",           "Eo",           "Fx",           "Ms",
                    288:        "No",           "Ns",           "Nx",           "Ox",
                    289:        "Pc",           "Pf",           "Po",           "Pq",
                    290:        "Qc",           "Ql",           "Qo",           "Qq",
                    291:        "Re",           "Rs",           "Sc",           "So",
                    292:        "Sq",           "Sm",           "Sx",           "Sy",
                    293:        "Tn",           "Ux",           "Xc",           "Xo",
                    294:        "Fo",           "Fc",           "Oo",           "Oc",
                    295:        "Bk",           "Ek",           "Bt",           "Hf",
                    296:        "Fr",           "Ud",           "Lb",           "Lp",
                    297:        "Lk",           "Mt",           "Brq",          "Bro",
                    298:        "Brc",          "%C",           "Es",           "En",
                    299:        "Dx",           "%Q",           "br",           "sp",
                    300:        "%U",           "Ta",           "ll",           NULL,
                    301:        "TH",           "SH",           "SS",           "TP",
                    302:        "LP",           "PP",           "P",            "IP",
                    303:        "HP",           "SM",           "SB",           "BI",
                    304:        "IB",           "BR",           "RB",           "R",
                    305:        "B",            "I",            "IR",           "RI",
                    306:        "br",           "sp",           "nf",           "fi",
                    307:        "RE",           "RS",           "DT",           "UC",
                    308:        "PD",           "AT",           "in",           "ft",
                    309:        "OP",           "EX",           "EE",           "UR",
                    310:        "UE",           "ll",           NULL
                    311: };
                    312: const  char *const *roff_name = __roff_name;
                    313:
                    314: static struct roffmac   roffs[TOKEN_NONE] = {
1.167   ! schwarze  315:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
        !           316:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
        !           317:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
        !           318:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
        !           319:        { roff_unsupp, NULL, NULL, 0 },  /* als */
        !           320:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
        !           321:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
        !           322:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
        !           323:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
        !           324:        { roff_ds, NULL, NULL, 0 },  /* as */
        !           325:        { roff_ds, NULL, NULL, 0 },  /* as1 */
        !           326:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
        !           327:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
        !           328:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
        !           329:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
        !           330:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
        !           331:        { roff_unsupp, NULL, NULL, 0 },  /* box */
        !           332:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
        !           333:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
        !           334:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
        !           335:        { roff_unsupp, NULL, NULL, 0 },  /* break */
        !           336:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
        !           337:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
        !           338:        { roff_brp, NULL, NULL, 0 },  /* brp */
        !           339:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
        !           340:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
        !           341:        { roff_cc, NULL, NULL, 0 },  /* cc */
        !           342:        { roff_line_ignore, NULL, NULL, 0 },  /* ce */
        !           343:        { roff_insec, NULL, NULL, 0 },  /* cf */
        !           344:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
        !           345:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
        !           346:        { roff_unsupp, NULL, NULL, 0 },  /* char */
        !           347:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
        !           348:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
        !           349:        { roff_insec, NULL, NULL, 0 },  /* close */
        !           350:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
        !           351:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
        !           352:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
        !           353:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
        !           354:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
        !           355:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
        !           356:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
        !           357:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
        !           358:        { roff_unsupp, NULL, NULL, 0 },  /* da */
        !           359:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
        !           360:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
        !           361:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
        !           362:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
        !           363:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
        !           364:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
        !           365:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
        !           366:        { roff_unsupp, NULL, NULL, 0 },  /* device */
        !           367:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
        !           368:        { roff_unsupp, NULL, NULL, 0 },  /* di */
        !           369:        { roff_unsupp, NULL, NULL, 0 },  /* do */
        !           370:        { roff_ds, NULL, NULL, 0 },  /* ds */
        !           371:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
        !           372:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
        !           373:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
        !           374:        { roff_unsupp, NULL, NULL, 0 },  /* ec */
        !           375:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
        !           376:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
        !           377:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
        !           378:        { roff_unsupp, NULL, NULL, 0 },  /* em */
        !           379:        { roff_EN, NULL, NULL, 0 },  /* EN */
        !           380:        { roff_unsupp, NULL, NULL, 0 },  /* eo */
        !           381:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
        !           382:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
        !           383:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
        !           384:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
        !           385:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
        !           386:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
        !           387:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
        !           388:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
        !           389:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
        !           390:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
        !           391:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
        !           392:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
        !           393:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
        !           394:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
        !           395:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
        !           396:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
        !           397:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
        !           398:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
        !           399:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
        !           400:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
        !           401:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
        !           402:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
        !           403:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
        !           404:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
        !           405:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
        !           406:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
        !           407:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
        !           408:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
        !           409:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
        !           410:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
        !           411:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
        !           412:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
        !           413:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
        !           414:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
        !           415:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
        !           416:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
        !           417:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
        !           418:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
        !           419:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
        !           420:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
        !           421:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
        !           422:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
        !           423:        { roff_unsupp, NULL, NULL, 0 },  /* index */
        !           424:        { roff_it, NULL, NULL, 0 },  /* it */
        !           425:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
        !           426:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
        !           427:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
        !           428:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
        !           429:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
        !           430:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
        !           431:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
        !           432:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
        !           433:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
        !           434:        { roff_unsupp, NULL, NULL, 0 },  /* length */
        !           435:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
        !           436:        { roff_insec, NULL, NULL, 0 },  /* lf */
        !           437:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
        !           438:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
        !           439:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
        !           440:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
        !           441:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
        !           442:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
        !           443:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
        !           444:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
        !           445:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
        !           446:        { roff_line_ignore, NULL, NULL, 0 },  /* mc */
        !           447:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
        !           448:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
        !           449:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
        !           450:        { roff_insec, NULL, NULL, 0 },  /* mso */
        !           451:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
        !           452:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
        !           453:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
        !           454:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
        !           455:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
        !           456:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
        !           457:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
        !           458:        { roff_nr, NULL, NULL, 0 },  /* nr */
        !           459:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
        !           460:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
        !           461:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
        !           462:        { roff_insec, NULL, NULL, 0 },  /* nx */
        !           463:        { roff_insec, NULL, NULL, 0 },  /* open */
        !           464:        { roff_insec, NULL, NULL, 0 },  /* opena */
        !           465:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
        !           466:        { roff_unsupp, NULL, NULL, 0 },  /* output */
        !           467:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
        !           468:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
        !           469:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
        !           470:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
        !           471:        { roff_insec, NULL, NULL, 0 },  /* pi */
        !           472:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
        !           473:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
        !           474:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
        !           475:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
        !           476:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
        !           477:        { roff_line_ignore, NULL, NULL, 0 },  /* po */
        !           478:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
        !           479:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
        !           480:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
        !           481:        { roff_insec, NULL, NULL, 0 },  /* pso */
        !           482:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
        !           483:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
        !           484:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
        !           485:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
        !           486:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
        !           487:        { roff_unsupp, NULL, NULL, 0 },  /* return */
        !           488:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
        !           489:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
        !           490:        { roff_line_ignore, NULL, NULL, 0 },  /* rj */
        !           491:        { roff_rm, NULL, NULL, 0 },  /* rm */
        !           492:        { roff_unsupp, NULL, NULL, 0 },  /* rn */
        !           493:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
        !           494:        { roff_rr, NULL, NULL, 0 },  /* rr */
        !           495:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
        !           496:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
        !           497:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
        !           498:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
        !           499:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
        !           500:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
        !           501:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
        !           502:        { roff_so, NULL, NULL, 0 },  /* so */
        !           503:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
        !           504:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
        !           505:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
        !           506:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
        !           507:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
        !           508:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
        !           509:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
        !           510:        { roff_insec, NULL, NULL, 0 },  /* sy */
        !           511:        { roff_T_, NULL, NULL, 0 },  /* T& */
        !           512:        { roff_unsupp, NULL, NULL, 0 },  /* ta */
        !           513:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
        !           514:        { roff_TE, NULL, NULL, 0 },  /* TE */
        !           515:        { roff_TH, NULL, NULL, 0 },  /* TH */
        !           516:        { roff_unsupp, NULL, NULL, 0 },  /* ti */
        !           517:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
        !           518:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
        !           519:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
        !           520:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
        !           521:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
        !           522:        { roff_tr, NULL, NULL, 0 },  /* tr */
        !           523:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
        !           524:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
        !           525:        { roff_insec, NULL, NULL, 0 },  /* trf */
        !           526:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
        !           527:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
        !           528:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
        !           529:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
        !           530:        { roff_TS, NULL, NULL, 0 },  /* TS */
        !           531:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
        !           532:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
        !           533:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
        !           534:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
        !           535:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
        !           536:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
        !           537:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
        !           538:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
        !           539:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
        !           540:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
        !           541:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
        !           542:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
        !           543:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
        !           544:        { roff_unsupp, NULL, NULL, 0 },  /* while */
        !           545:        { roff_insec, NULL, NULL, 0 },  /* write */
        !           546:        { roff_insec, NULL, NULL, 0 },  /* writec */
        !           547:        { roff_insec, NULL, NULL, 0 },  /* writem */
        !           548:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
        !           549:        { roff_cblock, NULL, NULL, 0 },  /* . */
        !           550:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  551: };
                    552:
1.73      schwarze  553: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47      schwarze  554: const  char *const __mdoc_reserved[] = {
                    555:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    556:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    557:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    558:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73      schwarze  559:        "Dt", "Dv", "Dx", "D1",
                    560:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    561:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47      schwarze  562:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73      schwarze  563:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    564:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47      schwarze  565:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73      schwarze  566:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    567:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    568:        "Sc", "Sh", "Sm", "So", "Sq",
1.47      schwarze  569:        "Ss", "St", "Sx", "Sy",
                    570:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73      schwarze  571:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47      schwarze  572:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    573:        NULL
                    574: };
                    575:
1.73      schwarze  576: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47      schwarze  577: const  char *const __man_reserved[] = {
1.73      schwarze  578:        "AT", "B", "BI", "BR", "DT",
                    579:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    580:        "LP", "OP", "P", "PD", "PP",
                    581:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    582:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47      schwarze  583:        NULL
                    584: };
                    585:
1.37      schwarze  586: /* Array of injected predefined strings. */
                    587: #define        PREDEFS_MAX      38
                    588: static const struct predef predefs[PREDEFS_MAX] = {
                    589: #include "predefs.in"
                    590: };
                    591:
1.51      schwarze  592: static int      roffit_lines;  /* number of lines to delay */
                    593: static char    *roffit_macro;  /* nil-terminated macro line */
                    594:
1.80      schwarze  595:
1.138     schwarze  596: /* --- request table ------------------------------------------------------ */
                    597:
1.167   ! schwarze  598: struct ohash *
        !           599: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  600: {
1.167   ! schwarze  601:        struct ohash    *htab;
        !           602:        struct roffreq  *req;
        !           603:        enum roff_tok    tok;
        !           604:        size_t           sz;
        !           605:        unsigned int     slot;
1.3       schwarze  606:
1.167   ! schwarze  607:        htab = mandoc_malloc(sizeof(*htab));
        !           608:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  609:
1.167   ! schwarze  610:        for (tok = mintok; tok < maxtok; tok++) {
        !           611:                sz = strlen(roff_name[tok]);
        !           612:                req = mandoc_malloc(sizeof(*req) + sz + 1);
        !           613:                req->tok = tok;
        !           614:                memcpy(req->name, roff_name[tok], sz + 1);
        !           615:                slot = ohash_qlookup(htab, req->name);
        !           616:                ohash_insert(htab, slot, req);
1.3       schwarze  617:        }
1.167   ! schwarze  618:        return htab;
1.3       schwarze  619: }
                    620:
1.167   ! schwarze  621: void
        !           622: roffhash_free(struct ohash *htab)
1.1       schwarze  623: {
1.167   ! schwarze  624:        struct roffreq  *req;
        !           625:        unsigned int     slot;
1.1       schwarze  626:
1.167   ! schwarze  627:        if (htab == NULL)
        !           628:                return;
        !           629:        for (req = ohash_first(htab, &slot); req != NULL;
        !           630:             req = ohash_next(htab, &slot))
        !           631:                free(req);
        !           632:        ohash_delete(htab);
        !           633:        free(htab);
        !           634: }
        !           635:
        !           636: enum roff_tok
        !           637: roffhash_find(struct ohash *htab, const char *name, size_t sz)
        !           638: {
        !           639:        struct roffreq  *req;
        !           640:        const char      *end;
        !           641:
        !           642:        if (sz) {
        !           643:                end = name + sz;
        !           644:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
        !           645:        } else
        !           646:                req = ohash_find(htab, ohash_qlookup(htab, name));
        !           647:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  648: }
                    649:
1.138     schwarze  650: /* --- stack of request blocks -------------------------------------------- */
                    651:
1.1       schwarze  652: /*
                    653:  * Pop the current node off of the stack of roff instructions currently
                    654:  * pending.
                    655:  */
                    656: static void
                    657: roffnode_pop(struct roff *r)
                    658: {
                    659:        struct roffnode *p;
                    660:
1.2       schwarze  661:        assert(r->last);
1.80      schwarze  662:        p = r->last;
1.2       schwarze  663:
                    664:        r->last = r->last->parent;
1.16      schwarze  665:        free(p->name);
                    666:        free(p->end);
1.1       schwarze  667:        free(p);
                    668: }
                    669:
                    670: /*
                    671:  * Push a roff node onto the instruction stack.  This must later be
                    672:  * removed with roffnode_pop().
                    673:  */
1.11      schwarze  674: static void
1.166     schwarze  675: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  676:                int line, int col)
1.1       schwarze  677: {
                    678:        struct roffnode *p;
                    679:
1.11      schwarze  680:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  681:        p->tok = tok;
1.16      schwarze  682:        if (name)
                    683:                p->name = mandoc_strdup(name);
1.1       schwarze  684:        p->parent = r->last;
                    685:        p->line = line;
                    686:        p->col = col;
1.71      schwarze  687:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  688:
                    689:        r->last = p;
                    690: }
                    691:
1.138     schwarze  692: /* --- roff parser state data management ---------------------------------- */
                    693:
1.1       schwarze  694: static void
                    695: roff_free1(struct roff *r)
                    696: {
1.49      schwarze  697:        struct tbl_node *tbl;
1.32      schwarze  698:        struct eqn_node *e;
1.42      schwarze  699:        int              i;
1.27      schwarze  700:
1.49      schwarze  701:        while (NULL != (tbl = r->first_tbl)) {
                    702:                r->first_tbl = tbl->next;
                    703:                tbl_free(tbl);
1.27      schwarze  704:        }
                    705:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  706:
1.32      schwarze  707:        while (NULL != (e = r->first_eqn)) {
                    708:                r->first_eqn = e->next;
                    709:                eqn_free(e);
                    710:        }
                    711:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    712:
1.1       schwarze  713:        while (r->last)
                    714:                roffnode_pop(r);
1.27      schwarze  715:
1.96      schwarze  716:        free (r->rstack);
                    717:        r->rstack = NULL;
                    718:        r->rstacksz = 0;
                    719:        r->rstackpos = -1;
                    720:
                    721:        roff_freereg(r->regtab);
                    722:        r->regtab = NULL;
                    723:
1.42      schwarze  724:        roff_freestr(r->strtab);
                    725:        roff_freestr(r->xmbtab);
                    726:        r->strtab = r->xmbtab = NULL;
                    727:
                    728:        if (r->xtab)
                    729:                for (i = 0; i < 128; i++)
                    730:                        free(r->xtab[i].p);
                    731:        free(r->xtab);
                    732:        r->xtab = NULL;
1.1       schwarze  733: }
                    734:
                    735: void
                    736: roff_reset(struct roff *r)
                    737: {
                    738:        roff_free1(r);
1.99      schwarze  739:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48      schwarze  740:        r->control = 0;
1.1       schwarze  741: }
                    742:
                    743: void
                    744: roff_free(struct roff *r)
                    745: {
                    746:        roff_free1(r);
1.167   ! schwarze  747:        roffhash_free(r->reqtab);
1.1       schwarze  748:        free(r);
                    749: }
                    750:
                    751: struct roff *
1.151     schwarze  752: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  753: {
                    754:        struct roff     *r;
                    755:
1.11      schwarze  756:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  757:        r->parse = parse;
1.167   ! schwarze  758:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72      schwarze  759:        r->options = options;
1.99      schwarze  760:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  761:        r->rstackpos = -1;
1.149     schwarze  762:        return r;
1.137     schwarze  763: }
                    764:
1.138     schwarze  765: /* --- syntax tree state data management ---------------------------------- */
                    766:
1.137     schwarze  767: static void
                    768: roff_man_free1(struct roff_man *man)
                    769: {
                    770:
1.138     schwarze  771:        if (man->first != NULL)
                    772:                roff_node_delete(man, man->first);
1.137     schwarze  773:        free(man->meta.msec);
                    774:        free(man->meta.vol);
                    775:        free(man->meta.os);
                    776:        free(man->meta.arch);
                    777:        free(man->meta.title);
                    778:        free(man->meta.name);
                    779:        free(man->meta.date);
                    780: }
                    781:
                    782: static void
                    783: roff_man_alloc1(struct roff_man *man)
                    784: {
                    785:
                    786:        memset(&man->meta, 0, sizeof(man->meta));
                    787:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    788:        man->first->type = ROFFT_ROOT;
                    789:        man->last = man->first;
                    790:        man->last_es = NULL;
                    791:        man->flags = 0;
                    792:        man->macroset = MACROSET_NONE;
                    793:        man->lastsec = man->lastnamed = SEC_NONE;
                    794:        man->next = ROFF_NEXT_CHILD;
                    795: }
                    796:
                    797: void
                    798: roff_man_reset(struct roff_man *man)
                    799: {
                    800:
                    801:        roff_man_free1(man);
                    802:        roff_man_alloc1(man);
                    803: }
                    804:
                    805: void
                    806: roff_man_free(struct roff_man *man)
                    807: {
                    808:
                    809:        roff_man_free1(man);
                    810:        free(man);
                    811: }
                    812:
                    813: struct roff_man *
                    814: roff_man_alloc(struct roff *roff, struct mparse *parse,
                    815:        const char *defos, int quick)
                    816: {
                    817:        struct roff_man *man;
                    818:
                    819:        man = mandoc_calloc(1, sizeof(*man));
                    820:        man->parse = parse;
                    821:        man->roff = roff;
                    822:        man->defos = defos;
                    823:        man->quick = quick;
                    824:        roff_man_alloc1(man);
1.149     schwarze  825:        return man;
1.1       schwarze  826: }
                    827:
1.138     schwarze  828: /* --- syntax tree handling ----------------------------------------------- */
                    829:
                    830: struct roff_node *
                    831: roff_node_alloc(struct roff_man *man, int line, int pos,
                    832:        enum roff_type type, int tok)
                    833: {
                    834:        struct roff_node        *n;
                    835:
                    836:        n = mandoc_calloc(1, sizeof(*n));
                    837:        n->line = line;
                    838:        n->pos = pos;
                    839:        n->tok = tok;
                    840:        n->type = type;
                    841:        n->sec = man->lastsec;
                    842:
                    843:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  844:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  845:        else
1.157     schwarze  846:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  847:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  848:                n->flags |= NODE_LINE;
1.138     schwarze  849:        man->flags &= ~MDOC_NEWLINE;
                    850:
1.149     schwarze  851:        return n;
1.138     schwarze  852: }
                    853:
                    854: void
                    855: roff_node_append(struct roff_man *man, struct roff_node *n)
                    856: {
                    857:
                    858:        switch (man->next) {
                    859:        case ROFF_NEXT_SIBLING:
1.153     schwarze  860:                if (man->last->next != NULL) {
                    861:                        n->next = man->last->next;
                    862:                        man->last->next->prev = n;
                    863:                } else
                    864:                        man->last->parent->last = n;
1.138     schwarze  865:                man->last->next = n;
                    866:                n->prev = man->last;
                    867:                n->parent = man->last->parent;
                    868:                break;
                    869:        case ROFF_NEXT_CHILD:
1.159     schwarze  870:                if (man->last->child != NULL) {
                    871:                        n->next = man->last->child;
                    872:                        man->last->child->prev = n;
                    873:                } else
                    874:                        man->last->last = n;
1.138     schwarze  875:                man->last->child = n;
                    876:                n->parent = man->last;
                    877:                break;
                    878:        default:
                    879:                abort();
                    880:        }
1.154     schwarze  881:        man->last = n;
1.138     schwarze  882:
                    883:        switch (n->type) {
                    884:        case ROFFT_HEAD:
                    885:                n->parent->head = n;
                    886:                break;
                    887:        case ROFFT_BODY:
1.154     schwarze  888:                if (n->end != ENDBODY_NOT)
                    889:                        return;
1.138     schwarze  890:                n->parent->body = n;
                    891:                break;
                    892:        case ROFFT_TAIL:
                    893:                n->parent->tail = n;
                    894:                break;
                    895:        default:
1.154     schwarze  896:                return;
1.138     schwarze  897:        }
1.154     schwarze  898:
                    899:        /*
                    900:         * Copy over the normalised-data pointer of our parent.  Not
                    901:         * everybody has one, but copying a null pointer is fine.
                    902:         */
                    903:
                    904:        n->norm = n->parent->norm;
                    905:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  906: }
                    907:
1.139     schwarze  908: void
                    909: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    910: {
                    911:        struct roff_node        *n;
                    912:
                    913:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    914:        n->string = roff_strdup(man->roff, word);
                    915:        roff_node_append(man, n);
1.158     schwarze  916:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  917:        man->next = ROFF_NEXT_SIBLING;
                    918: }
                    919:
                    920: void
                    921: roff_word_append(struct roff_man *man, const char *word)
                    922: {
                    923:        struct roff_node        *n;
                    924:        char                    *addstr, *newstr;
                    925:
                    926:        n = man->last;
                    927:        addstr = roff_strdup(man->roff, word);
                    928:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    929:        free(addstr);
                    930:        free(n->string);
                    931:        n->string = newstr;
                    932:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  933: }
                    934:
                    935: void
                    936: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    937: {
                    938:        struct roff_node        *n;
                    939:
                    940:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    941:        roff_node_append(man, n);
                    942:        man->next = ROFF_NEXT_CHILD;
                    943: }
                    944:
                    945: struct roff_node *
                    946: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    947: {
                    948:        struct roff_node        *n;
                    949:
                    950:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    951:        roff_node_append(man, n);
                    952:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  953:        return n;
1.139     schwarze  954: }
                    955:
1.138     schwarze  956: struct roff_node *
                    957: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    958: {
                    959:        struct roff_node        *n;
                    960:
                    961:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    962:        roff_node_append(man, n);
                    963:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  964:        return n;
1.138     schwarze  965: }
                    966:
                    967: struct roff_node *
                    968: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    969: {
                    970:        struct roff_node        *n;
                    971:
                    972:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    973:        roff_node_append(man, n);
                    974:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  975:        return n;
1.139     schwarze  976: }
                    977:
                    978: void
                    979: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
                    980: {
                    981:        struct roff_node        *n;
                    982:
                    983:        n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
                    984:        n->eqn = eqn;
                    985:        if (eqn->ln > man->last->line)
1.157     schwarze  986:                n->flags |= NODE_LINE;
1.139     schwarze  987:        roff_node_append(man, n);
                    988:        man->next = ROFF_NEXT_SIBLING;
                    989: }
                    990:
                    991: void
                    992: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
                    993: {
                    994:        struct roff_node        *n;
                    995:
                    996:        if (man->macroset == MACROSET_MAN)
                    997:                man_breakscope(man, TOKEN_NONE);
                    998:        n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                    999:        n->span = tbl;
                   1000:        roff_node_append(man, n);
1.158     schwarze 1001:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze 1002:        man->next = ROFF_NEXT_SIBLING;
1.138     schwarze 1003: }
                   1004:
                   1005: void
                   1006: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1007: {
                   1008:
                   1009:        /* Adjust siblings. */
                   1010:
                   1011:        if (n->prev)
                   1012:                n->prev->next = n->next;
                   1013:        if (n->next)
                   1014:                n->next->prev = n->prev;
                   1015:
                   1016:        /* Adjust parent. */
                   1017:
                   1018:        if (n->parent != NULL) {
                   1019:                if (n->parent->child == n)
                   1020:                        n->parent->child = n->next;
                   1021:                if (n->parent->last == n)
                   1022:                        n->parent->last = n->prev;
                   1023:        }
                   1024:
                   1025:        /* Adjust parse point. */
                   1026:
                   1027:        if (man == NULL)
                   1028:                return;
                   1029:        if (man->last == n) {
                   1030:                if (n->prev == NULL) {
                   1031:                        man->last = n->parent;
                   1032:                        man->next = ROFF_NEXT_CHILD;
                   1033:                } else {
                   1034:                        man->last = n->prev;
                   1035:                        man->next = ROFF_NEXT_SIBLING;
                   1036:                }
                   1037:        }
                   1038:        if (man->first == n)
                   1039:                man->first = NULL;
                   1040: }
                   1041:
                   1042: void
                   1043: roff_node_free(struct roff_node *n)
                   1044: {
                   1045:
                   1046:        if (n->args != NULL)
                   1047:                mdoc_argv_free(n->args);
                   1048:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1049:                free(n->norm);
                   1050:        free(n->string);
                   1051:        free(n);
                   1052: }
                   1053:
                   1054: void
                   1055: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1056: {
                   1057:
                   1058:        while (n->child != NULL)
                   1059:                roff_node_delete(man, n->child);
                   1060:        roff_node_unlink(man, n);
                   1061:        roff_node_free(n);
1.141     schwarze 1062: }
                   1063:
                   1064: void
                   1065: deroff(char **dest, const struct roff_node *n)
                   1066: {
                   1067:        char    *cp;
                   1068:        size_t   sz;
                   1069:
                   1070:        if (n->type != ROFFT_TEXT) {
                   1071:                for (n = n->child; n != NULL; n = n->next)
                   1072:                        deroff(dest, n);
                   1073:                return;
                   1074:        }
                   1075:
1.160     schwarze 1076:        /* Skip leading whitespace. */
1.141     schwarze 1077:
1.160     schwarze 1078:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1079:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1080:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1081:                        cp++;
1.160     schwarze 1082:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1083:                        break;
                   1084:        }
                   1085:
1.161     schwarze 1086:        /* Skip trailing backslash. */
                   1087:
                   1088:        sz = strlen(cp);
1.162     schwarze 1089:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1090:                sz--;
                   1091:
1.141     schwarze 1092:        /* Skip trailing whitespace. */
                   1093:
1.161     schwarze 1094:        for (; sz; sz--)
1.141     schwarze 1095:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1096:                        break;
                   1097:
                   1098:        /* Skip empty strings. */
                   1099:
                   1100:        if (sz == 0)
                   1101:                return;
                   1102:
                   1103:        if (*dest == NULL) {
                   1104:                *dest = mandoc_strndup(cp, sz);
                   1105:                return;
                   1106:        }
                   1107:
                   1108:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1109:        free(*dest);
                   1110:        *dest = cp;
1.138     schwarze 1111: }
                   1112:
                   1113: /* --- main functions of the roff parser ---------------------------------- */
                   1114:
1.8       schwarze 1115: /*
1.79      schwarze 1116:  * In the current line, expand escape sequences that tend to get
                   1117:  * used in numerical expressions and conditional requests.
                   1118:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1119:  */
1.45      schwarze 1120: static enum rofferr
1.110     schwarze 1121: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1122: {
1.81      schwarze 1123:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze 1124:        const char      *start; /* start of the string to process */
1.82      schwarze 1125:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze 1126:        const char      *stnam; /* start of the name, after "[(*" */
                   1127:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1128:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1129:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1130:        size_t           maxl;  /* expected length of the escape name */
                   1131:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1132:        enum mandoc_esc  esc;   /* type of the escape sequence */
                   1133:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1134:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1135:        int              npos;  /* position in numeric expression */
1.91      schwarze 1136:        int              arg_complete; /* argument not interrupted by eol */
1.79      schwarze 1137:        char             term;  /* character terminating the escape */
1.8       schwarze 1138:
1.43      schwarze 1139:        expand_count = 0;
1.110     schwarze 1140:        start = buf->buf + pos;
1.78      schwarze 1141:        stesc = strchr(start, '\0') - 1;
                   1142:        while (stesc-- > start) {
1.43      schwarze 1143:
1.78      schwarze 1144:                /* Search backwards for the next backslash. */
                   1145:
1.110     schwarze 1146:                if (*stesc != '\\')
1.78      schwarze 1147:                        continue;
                   1148:
                   1149:                /* If it is escaped, skip it. */
                   1150:
                   1151:                for (cp = stesc - 1; cp >= start; cp--)
1.110     schwarze 1152:                        if (*cp != '\\')
1.78      schwarze 1153:                                break;
                   1154:
1.110     schwarze 1155:                if ((stesc - cp) % 2 == 0) {
1.82      schwarze 1156:                        stesc = (char *)cp;
1.78      schwarze 1157:                        continue;
                   1158:                }
1.23      schwarze 1159:
1.79      schwarze 1160:                /* Decide whether to expand or to check only. */
1.23      schwarze 1161:
1.79      schwarze 1162:                term = '\0';
1.78      schwarze 1163:                cp = stesc + 1;
1.53      schwarze 1164:                switch (*cp) {
1.80      schwarze 1165:                case '*':
1.53      schwarze 1166:                        res = NULL;
                   1167:                        break;
1.80      schwarze 1168:                case 'B':
                   1169:                case 'w':
1.79      schwarze 1170:                        term = cp[1];
                   1171:                        /* FALLTHROUGH */
1.80      schwarze 1172:                case 'n':
1.53      schwarze 1173:                        res = ubuf;
                   1174:                        break;
                   1175:                default:
1.109     schwarze 1176:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1177:                        if (esc == ESCAPE_ERROR ||
                   1178:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1179:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1180:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1181:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1182:                                    "%.*s", (int)(cp - stesc), stesc);
1.78      schwarze 1183:                        continue;
1.42      schwarze 1184:                }
                   1185:
1.78      schwarze 1186:                if (EXPAND_LIMIT < ++expand_count) {
                   1187:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1188:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1189:                        return ROFF_IGN;
1.78      schwarze 1190:                }
1.23      schwarze 1191:
                   1192:                /*
                   1193:                 * The third character decides the length
1.53      schwarze 1194:                 * of the name of the string or register.
1.23      schwarze 1195:                 * Save a pointer to the name.
                   1196:                 */
                   1197:
1.110     schwarze 1198:                if (term == '\0') {
1.79      schwarze 1199:                        switch (*++cp) {
1.80      schwarze 1200:                        case '\0':
1.79      schwarze 1201:                                maxl = 0;
                   1202:                                break;
1.80      schwarze 1203:                        case '(':
1.79      schwarze 1204:                                cp++;
                   1205:                                maxl = 2;
                   1206:                                break;
1.80      schwarze 1207:                        case '[':
1.79      schwarze 1208:                                cp++;
                   1209:                                term = ']';
                   1210:                                maxl = 0;
                   1211:                                break;
                   1212:                        default:
                   1213:                                maxl = 1;
                   1214:                                break;
                   1215:                        }
                   1216:                } else {
                   1217:                        cp += 2;
1.8       schwarze 1218:                        maxl = 0;
                   1219:                }
1.23      schwarze 1220:                stnam = cp;
1.8       schwarze 1221:
1.23      schwarze 1222:                /* Advance to the end of the name. */
1.8       schwarze 1223:
1.125     schwarze 1224:                naml = 0;
1.91      schwarze 1225:                arg_complete = 1;
1.125     schwarze 1226:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1227:                        if (*cp == '\0') {
1.92      schwarze 1228:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1229:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1230:                                arg_complete = 0;
1.79      schwarze 1231:                                break;
1.42      schwarze 1232:                        }
1.110     schwarze 1233:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1234:                                cp++;
1.125     schwarze 1235:                                break;
                   1236:                        }
                   1237:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1238:                                naml++;
                   1239:                                continue;
                   1240:                        }
                   1241:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1242:                        case ESCAPE_SPECIAL:
                   1243:                        case ESCAPE_UNICODE:
                   1244:                        case ESCAPE_NUMBERED:
                   1245:                        case ESCAPE_OVERSTRIKE:
                   1246:                                naml++;
                   1247:                                break;
                   1248:                        default:
1.8       schwarze 1249:                                break;
1.79      schwarze 1250:                        }
1.8       schwarze 1251:                }
                   1252:
1.23      schwarze 1253:                /*
                   1254:                 * Retrieve the replacement string; if it is
                   1255:                 * undefined, resume searching for escapes.
                   1256:                 */
                   1257:
1.79      schwarze 1258:                switch (stesc[1]) {
1.80      schwarze 1259:                case '*':
1.91      schwarze 1260:                        if (arg_complete)
                   1261:                                res = roff_getstrn(r, stnam, naml);
1.79      schwarze 1262:                        break;
1.80      schwarze 1263:                case 'B':
1.79      schwarze 1264:                        npos = 0;
1.91      schwarze 1265:                        ubuf[0] = arg_complete &&
1.133     schwarze 1266:                            roff_evalnum(r, ln, stnam, &npos,
                   1267:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1268:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1269:                        ubuf[1] = '\0';
                   1270:                        break;
1.80      schwarze 1271:                case 'n':
1.91      schwarze 1272:                        if (arg_complete)
                   1273:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1274:                                    roff_getregn(r, stnam, naml));
                   1275:                        else
                   1276:                                ubuf[0] = '\0';
1.79      schwarze 1277:                        break;
1.80      schwarze 1278:                case 'w':
1.91      schwarze 1279:                        /* use even incomplete args */
1.81      schwarze 1280:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1281:                            24 * (int)naml);
                   1282:                        break;
                   1283:                }
1.8       schwarze 1284:
1.110     schwarze 1285:                if (res == NULL) {
1.92      schwarze 1286:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1287:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1288:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1289:                        res = "";
1.118     schwarze 1290:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1291:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1292:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1293:                        return ROFF_IGN;
1.8       schwarze 1294:                }
                   1295:
1.23      schwarze 1296:                /* Replace the escape sequence by the string. */
                   1297:
1.82      schwarze 1298:                *stesc = '\0';
1.110     schwarze 1299:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1300:                    buf->buf, res, cp) + 1;
1.8       schwarze 1301:
1.78      schwarze 1302:                /* Prepare for the next replacement. */
1.8       schwarze 1303:
1.78      schwarze 1304:                start = nbuf + pos;
1.110     schwarze 1305:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1306:                free(buf->buf);
                   1307:                buf->buf = nbuf;
1.42      schwarze 1308:        }
1.149     schwarze 1309:        return ROFF_CONT;
1.42      schwarze 1310: }
                   1311:
                   1312: /*
1.147     schwarze 1313:  * Process text streams.
1.42      schwarze 1314:  */
                   1315: static enum rofferr
1.110     schwarze 1316: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42      schwarze 1317: {
                   1318:        size_t           sz;
                   1319:        const char      *start;
1.51      schwarze 1320:        char            *p;
                   1321:        int              isz;
1.42      schwarze 1322:        enum mandoc_esc  esc;
                   1323:
1.147     schwarze 1324:        /* Spring the input line trap. */
                   1325:
                   1326:        if (roffit_lines == 1) {
                   1327:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1328:                free(buf->buf);
                   1329:                buf->buf = p;
                   1330:                buf->sz = isz + 1;
                   1331:                *offs = 0;
                   1332:                free(roffit_macro);
                   1333:                roffit_lines = 0;
1.149     schwarze 1334:                return ROFF_REPARSE;
1.147     schwarze 1335:        } else if (roffit_lines > 1)
                   1336:                --roffit_lines;
                   1337:
                   1338:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1339:
1.110     schwarze 1340:        start = p = buf->buf + pos;
1.42      schwarze 1341:
1.110     schwarze 1342:        while (*p != '\0') {
1.42      schwarze 1343:                sz = strcspn(p, "-\\");
                   1344:                p += sz;
                   1345:
1.110     schwarze 1346:                if (*p == '\0')
1.42      schwarze 1347:                        break;
                   1348:
1.110     schwarze 1349:                if (*p == '\\') {
1.42      schwarze 1350:                        /* Skip over escapes. */
                   1351:                        p++;
1.62      schwarze 1352:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1353:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1354:                                break;
1.136     schwarze 1355:                        while (*p == '-')
                   1356:                                p++;
1.42      schwarze 1357:                        continue;
                   1358:                } else if (p == start) {
                   1359:                        p++;
                   1360:                        continue;
                   1361:                }
                   1362:
1.44      schwarze 1363:                if (isalpha((unsigned char)p[-1]) &&
                   1364:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1365:                        *p = ASCII_HYPH;
                   1366:                p++;
1.8       schwarze 1367:        }
1.149     schwarze 1368:        return ROFF_CONT;
1.8       schwarze 1369: }
                   1370:
1.1       schwarze 1371: enum rofferr
1.110     schwarze 1372: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1373: {
1.166     schwarze 1374:        enum roff_tok    t;
1.27      schwarze 1375:        enum rofferr     e;
1.110     schwarze 1376:        int              pos;   /* parse point */
1.115     schwarze 1377:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1378:        int              ppos;  /* original offset in buf->buf */
                   1379:        int              ctl;   /* macro line (boolean) */
                   1380:
                   1381:        ppos = pos = *offs;
1.1       schwarze 1382:
1.102     schwarze 1383:        /* Handle in-line equation delimiters. */
                   1384:
1.108     schwarze 1385:        if (r->tbl == NULL &&
                   1386:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1387:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1388:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1389:                if (e == ROFF_REPARSE)
1.149     schwarze 1390:                        return e;
1.102     schwarze 1391:                assert(e == ROFF_CONT);
                   1392:        }
                   1393:
                   1394:        /* Expand some escape sequences. */
1.8       schwarze 1395:
1.110     schwarze 1396:        e = roff_res(r, buf, ln, pos);
                   1397:        if (e == ROFF_IGN)
1.149     schwarze 1398:                return e;
1.110     schwarze 1399:        assert(e == ROFF_CONT);
1.8       schwarze 1400:
1.110     schwarze 1401:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1402:
1.8       schwarze 1403:        /*
1.2       schwarze 1404:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1405:         * text through the macro's filter.
                   1406:         * Equations process all content themselves.
                   1407:         * Tables process almost all content themselves, but we want
                   1408:         * to warn about macros before passing it there.
1.2       schwarze 1409:         */
                   1410:
1.124     schwarze 1411:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1412:                t = r->last->tok;
1.110     schwarze 1413:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1414:                if (e == ROFF_IGN)
1.149     schwarze 1415:                        return e;
1.166     schwarze 1416:                assert(e == ROFF_CONT);
1.54      schwarze 1417:        }
1.124     schwarze 1418:        if (r->eqn != NULL)
1.149     schwarze 1419:                return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.124     schwarze 1420:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.149     schwarze 1421:                return tbl_read(r->tbl, ln, buf->buf, ppos);
1.124     schwarze 1422:        if ( ! ctl)
1.149     schwarze 1423:                return roff_parsetext(buf, pos, offs);
1.100     schwarze 1424:
                   1425:        /* Skip empty request lines. */
                   1426:
1.110     schwarze 1427:        if (buf->buf[pos] == '"') {
1.100     schwarze 1428:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1429:                    ln, pos, NULL);
1.149     schwarze 1430:                return ROFF_IGN;
1.110     schwarze 1431:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1432:                return ROFF_IGN;
1.2       schwarze 1433:
                   1434:        /*
                   1435:         * If a scope is open, go to the child handler for that macro,
                   1436:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1437:         * Don't do so if an equation is open.
1.2       schwarze 1438:         */
                   1439:
                   1440:        if (r->last) {
1.1       schwarze 1441:                t = r->last->tok;
1.149     schwarze 1442:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1443:        }
                   1444:
1.115     schwarze 1445:        /* No scope is open.  This is a new request or macro. */
                   1446:
                   1447:        spos = pos;
                   1448:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1449:
                   1450:        /* Tables ignore most macros. */
                   1451:
1.166     schwarze 1452:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS)) {
1.115     schwarze 1453:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1454:                    ln, pos, buf->buf + spos);
1.129     schwarze 1455:                if (t == ROFF_TS)
1.149     schwarze 1456:                        return ROFF_IGN;
1.129     schwarze 1457:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1458:                        pos++;
1.163     schwarze 1459:                while (buf->buf[pos] == ' ')
1.129     schwarze 1460:                        pos++;
1.149     schwarze 1461:                return tbl_read(r->tbl, ln, buf->buf, pos);
1.115     schwarze 1462:        }
                   1463:
1.2       schwarze 1464:        /*
1.115     schwarze 1465:         * This is neither a roff request nor a user-defined macro.
                   1466:         * Let the standard macro set parsers handle it.
1.2       schwarze 1467:         */
                   1468:
1.166     schwarze 1469:        if (t == TOKEN_NONE)
1.149     schwarze 1470:                return ROFF_CONT;
1.115     schwarze 1471:
                   1472:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1473:
1.149     schwarze 1474:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1475: }
                   1476:
1.27      schwarze 1477: void
1.2       schwarze 1478: roff_endparse(struct roff *r)
                   1479: {
1.1       schwarze 1480:
1.27      schwarze 1481:        if (r->last)
1.94      schwarze 1482:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1483:                    r->last->line, r->last->col,
1.167   ! schwarze 1484:                    roff_name[r->last->tok]);
1.27      schwarze 1485:
1.32      schwarze 1486:        if (r->eqn) {
1.94      schwarze 1487:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1488:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41      schwarze 1489:                eqn_end(&r->eqn);
1.32      schwarze 1490:        }
                   1491:
1.27      schwarze 1492:        if (r->tbl) {
1.94      schwarze 1493:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1494:                    r->tbl->line, r->tbl->pos, "TS");
1.41      schwarze 1495:                tbl_end(&r->tbl);
1.27      schwarze 1496:        }
1.1       schwarze 1497: }
                   1498:
                   1499: /*
                   1500:  * Parse a roff node's type from the input buffer.  This must be in the
                   1501:  * form of ".foo xxx" in the usual way.
                   1502:  */
1.166     schwarze 1503: static enum roff_tok
1.87      schwarze 1504: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1505: {
1.87      schwarze 1506:        char            *cp;
1.16      schwarze 1507:        const char      *mac;
                   1508:        size_t           maclen;
1.166     schwarze 1509:        enum roff_tok    t;
1.1       schwarze 1510:
1.87      schwarze 1511:        cp = buf + *pos;
                   1512:
                   1513:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1514:                return TOKEN_NONE;
1.1       schwarze 1515:
1.87      schwarze 1516:        mac = cp;
                   1517:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1518:
1.16      schwarze 1519:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.167   ! schwarze 1520:            ? ROFF_USERDEF : roffhash_find(r->reqtab, mac, maclen);
1.1       schwarze 1521:
1.166     schwarze 1522:        if (t != TOKEN_NONE)
1.87      schwarze 1523:                *pos = cp - buf;
1.1       schwarze 1524:
1.149     schwarze 1525:        return t;
1.1       schwarze 1526: }
                   1527:
1.138     schwarze 1528: /* --- handling of request blocks ----------------------------------------- */
                   1529:
1.1       schwarze 1530: static enum rofferr
1.2       schwarze 1531: roff_cblock(ROFF_ARGS)
1.1       schwarze 1532: {
                   1533:
1.2       schwarze 1534:        /*
                   1535:         * A block-close `..' should only be invoked as a child of an
                   1536:         * ignore macro, otherwise raise a warning and just ignore it.
                   1537:         */
                   1538:
1.110     schwarze 1539:        if (r->last == NULL) {
1.94      schwarze 1540:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1541:                    ln, ppos, "..");
1.149     schwarze 1542:                return ROFF_IGN;
1.2       schwarze 1543:        }
1.1       schwarze 1544:
1.2       schwarze 1545:        switch (r->last->tok) {
1.80      schwarze 1546:        case ROFF_am:
1.93      schwarze 1547:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1548:        case ROFF_ami:
                   1549:        case ROFF_de:
1.23      schwarze 1550:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1551:        case ROFF_dei:
                   1552:        case ROFF_ig:
1.2       schwarze 1553:                break;
                   1554:        default:
1.94      schwarze 1555:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1556:                    ln, ppos, "..");
1.149     schwarze 1557:                return ROFF_IGN;
1.2       schwarze 1558:        }
                   1559:
1.110     schwarze 1560:        if (buf->buf[pos] != '\0')
1.90      schwarze 1561:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1562:                    ".. %s", buf->buf + pos);
1.2       schwarze 1563:
                   1564:        roffnode_pop(r);
                   1565:        roffnode_cleanscope(r);
1.149     schwarze 1566:        return ROFF_IGN;
1.2       schwarze 1567:
                   1568: }
1.1       schwarze 1569:
1.2       schwarze 1570: static void
                   1571: roffnode_cleanscope(struct roff *r)
                   1572: {
1.1       schwarze 1573:
1.2       schwarze 1574:        while (r->last) {
1.46      schwarze 1575:                if (--r->last->endspan != 0)
1.2       schwarze 1576:                        break;
                   1577:                roffnode_pop(r);
                   1578:        }
                   1579: }
1.1       schwarze 1580:
1.68      schwarze 1581: static void
                   1582: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1583: {
1.1       schwarze 1584:
1.2       schwarze 1585:        if (NULL == r->last) {
1.94      schwarze 1586:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1587:                    ln, ppos, "\\}");
1.68      schwarze 1588:                return;
1.2       schwarze 1589:        }
1.1       schwarze 1590:
1.2       schwarze 1591:        switch (r->last->tok) {
1.80      schwarze 1592:        case ROFF_el:
                   1593:        case ROFF_ie:
                   1594:        case ROFF_if:
1.2       schwarze 1595:                break;
                   1596:        default:
1.94      schwarze 1597:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1598:                    ln, ppos, "\\}");
1.68      schwarze 1599:                return;
1.2       schwarze 1600:        }
1.1       schwarze 1601:
1.2       schwarze 1602:        if (r->last->endspan > -1) {
1.94      schwarze 1603:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1604:                    ln, ppos, "\\}");
1.68      schwarze 1605:                return;
1.2       schwarze 1606:        }
                   1607:
                   1608:        roffnode_pop(r);
                   1609:        roffnode_cleanscope(r);
1.68      schwarze 1610:        return;
1.1       schwarze 1611: }
                   1612:
                   1613: static enum rofferr
1.2       schwarze 1614: roff_block(ROFF_ARGS)
1.1       schwarze 1615: {
1.93      schwarze 1616:        const char      *name;
                   1617:        char            *iname, *cp;
1.86      schwarze 1618:        size_t           namesz;
1.16      schwarze 1619:
1.93      schwarze 1620:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1621:
1.110     schwarze 1622:        if (tok == ROFF_de1)
1.93      schwarze 1623:                tok = ROFF_de;
1.123     schwarze 1624:        else if (tok == ROFF_dei1)
                   1625:                tok = ROFF_dei;
1.110     schwarze 1626:        else if (tok == ROFF_am1)
1.93      schwarze 1627:                tok = ROFF_am;
1.123     schwarze 1628:        else if (tok == ROFF_ami1)
                   1629:                tok = ROFF_ami;
1.93      schwarze 1630:
                   1631:        /* Parse the macro name argument. */
                   1632:
1.110     schwarze 1633:        cp = buf->buf + pos;
                   1634:        if (tok == ROFF_ig) {
1.93      schwarze 1635:                iname = NULL;
                   1636:                namesz = 0;
                   1637:        } else {
                   1638:                iname = cp;
                   1639:                namesz = roff_getname(r, &cp, ln, ppos);
                   1640:                iname[namesz] = '\0';
                   1641:        }
1.22      schwarze 1642:
1.93      schwarze 1643:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1644:
1.110     schwarze 1645:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1646:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1647:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1648:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1649:                            "%.*s", (int)namesz, iname);
                   1650:                        namesz = 0;
                   1651:                } else
                   1652:                        namesz = strlen(name);
                   1653:        } else
                   1654:                name = iname;
1.22      schwarze 1655:
1.110     schwarze 1656:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1657:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167   ! schwarze 1658:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1659:                return ROFF_IGN;
1.93      schwarze 1660:        }
1.2       schwarze 1661:
1.16      schwarze 1662:        roffnode_push(r, tok, name, ln, ppos);
                   1663:
                   1664:        /*
                   1665:         * At the beginning of a `de' macro, clear the existing string
                   1666:         * with the same name, if there is one.  New content will be
1.66      schwarze 1667:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1668:         */
1.22      schwarze 1669:
1.110     schwarze 1670:        if (tok == ROFF_de || tok == ROFF_dei)
1.86      schwarze 1671:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2       schwarze 1672:
1.110     schwarze 1673:        if (*cp == '\0')
1.149     schwarze 1674:                return ROFF_IGN;
1.1       schwarze 1675:
1.93      schwarze 1676:        /* Get the custom end marker. */
1.22      schwarze 1677:
1.93      schwarze 1678:        iname = cp;
1.86      schwarze 1679:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1680:
                   1681:        /* Resolve the end marker if it is indirect. */
                   1682:
1.110     schwarze 1683:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1684:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1685:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1686:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1687:                            "%.*s", (int)namesz, iname);
                   1688:                        namesz = 0;
                   1689:                } else
                   1690:                        namesz = strlen(name);
                   1691:        } else
                   1692:                name = iname;
                   1693:
1.86      schwarze 1694:        if (namesz)
                   1695:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1696:
1.110     schwarze 1697:        if (*cp != '\0')
1.90      schwarze 1698:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167   ! schwarze 1699:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1700:
1.149     schwarze 1701:        return ROFF_IGN;
1.1       schwarze 1702: }
                   1703:
                   1704: static enum rofferr
1.2       schwarze 1705: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1706: {
1.166     schwarze 1707:        enum roff_tok   t;
1.2       schwarze 1708:        int             i, j;
                   1709:
                   1710:        /*
                   1711:         * First check whether a custom macro exists at this level.  If
                   1712:         * it does, then check against it.  This is some of groff's
                   1713:         * stranger behaviours.  If we encountered a custom end-scope
                   1714:         * tag and that tag also happens to be a "real" macro, then we
                   1715:         * need to try interpreting it again as a real macro.  If it's
                   1716:         * not, then return ignore.  Else continue.
                   1717:         */
                   1718:
                   1719:        if (r->last->end) {
1.35      schwarze 1720:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1721:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1722:                                break;
1.1       schwarze 1723:
1.110     schwarze 1724:                if (r->last->end[j] == '\0' &&
                   1725:                    (buf->buf[i] == '\0' ||
                   1726:                     buf->buf[i] == ' ' ||
                   1727:                     buf->buf[i] == '\t')) {
1.2       schwarze 1728:                        roffnode_pop(r);
                   1729:                        roffnode_cleanscope(r);
1.1       schwarze 1730:
1.110     schwarze 1731:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1732:                                i++;
                   1733:
                   1734:                        pos = i;
1.110     schwarze 1735:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1736:                            TOKEN_NONE)
1.149     schwarze 1737:                                return ROFF_RERUN;
                   1738:                        return ROFF_IGN;
1.2       schwarze 1739:                }
1.1       schwarze 1740:        }
                   1741:
1.2       schwarze 1742:        /*
                   1743:         * If we have no custom end-query or lookup failed, then try
                   1744:         * pulling it out of the hashtable.
                   1745:         */
1.1       schwarze 1746:
1.110     schwarze 1747:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1748:
1.110     schwarze 1749:        if (t != ROFF_cblock) {
                   1750:                if (tok != ROFF_ig)
                   1751:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1752:                return ROFF_IGN;
1.16      schwarze 1753:        }
1.1       schwarze 1754:
1.149     schwarze 1755:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1756: }
                   1757:
                   1758: static enum rofferr
                   1759: roff_block_text(ROFF_ARGS)
                   1760: {
                   1761:
1.110     schwarze 1762:        if (tok != ROFF_ig)
                   1763:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1764:
1.149     schwarze 1765:        return ROFF_IGN;
1.2       schwarze 1766: }
                   1767:
                   1768: static enum rofferr
                   1769: roff_cond_sub(ROFF_ARGS)
                   1770: {
1.166     schwarze 1771:        enum roff_tok    t;
1.37      schwarze 1772:        char            *ep;
1.71      schwarze 1773:        int              rr;
1.2       schwarze 1774:
                   1775:        rr = r->last->rule;
1.37      schwarze 1776:        roffnode_cleanscope(r);
1.110     schwarze 1777:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2       schwarze 1778:
1.37      schwarze 1779:        /*
1.50      schwarze 1780:         * Fully handle known macros when they are structurally
                   1781:         * required or when the conditional evaluated to true.
1.5       schwarze 1782:         */
                   1783:
1.166     schwarze 1784:        if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.149     schwarze 1785:                return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.39      schwarze 1786:
1.69      schwarze 1787:        /*
                   1788:         * If `\}' occurs on a macro line without a preceding macro,
                   1789:         * drop the line completely.
                   1790:         */
                   1791:
1.110     schwarze 1792:        ep = buf->buf + pos;
                   1793:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1794:                rr = 0;
1.69      schwarze 1795:
1.50      schwarze 1796:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1797:
1.110     schwarze 1798:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1799:                if (*(++ep) == '}') {
1.70      schwarze 1800:                        *ep = '&';
1.110     schwarze 1801:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1802:                }
1.119     schwarze 1803:                if (*ep != '\0')
                   1804:                        ++ep;
1.50      schwarze 1805:        }
1.149     schwarze 1806:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1807: }
                   1808:
                   1809: static enum rofferr
                   1810: roff_cond_text(ROFF_ARGS)
                   1811: {
1.37      schwarze 1812:        char            *ep;
1.71      schwarze 1813:        int              rr;
1.2       schwarze 1814:
                   1815:        rr = r->last->rule;
1.37      schwarze 1816:        roffnode_cleanscope(r);
1.1       schwarze 1817:
1.110     schwarze 1818:        ep = buf->buf + pos;
                   1819:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1820:                if (*(++ep) == '}') {
1.70      schwarze 1821:                        *ep = '&';
1.110     schwarze 1822:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1823:                }
1.119     schwarze 1824:                if (*ep != '\0')
                   1825:                        ++ep;
1.2       schwarze 1826:        }
1.149     schwarze 1827:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1828: }
                   1829:
1.138     schwarze 1830: /* --- handling of numeric and conditional expressions -------------------- */
                   1831:
1.77      schwarze 1832: /*
                   1833:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1834:  * If there is at least one digit, return success and advance the
                   1835:  * parse point, else return failure and let the parse point unchanged.
                   1836:  * Ignore overflows, treat them just like the C language.
                   1837:  */
1.56      schwarze 1838: static int
1.133     schwarze 1839: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 1840: {
1.133     schwarze 1841:        int      myres, scaled, n, p;
1.79      schwarze 1842:
                   1843:        if (NULL == res)
                   1844:                res = &myres;
1.56      schwarze 1845:
                   1846:        p = *pos;
                   1847:        n = v[p] == '-';
1.133     schwarze 1848:        if (n || v[p] == '+')
1.56      schwarze 1849:                p++;
                   1850:
1.133     schwarze 1851:        if (flags & ROFFNUM_WHITE)
                   1852:                while (isspace((unsigned char)v[p]))
                   1853:                        p++;
                   1854:
1.56      schwarze 1855:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 1856:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 1857:        if (p == *pos + n)
                   1858:                return 0;
                   1859:
                   1860:        if (n)
                   1861:                *res = -*res;
                   1862:
1.126     schwarze 1863:        /* Each number may be followed by one optional scaling unit. */
                   1864:
                   1865:        switch (v[p]) {
                   1866:        case 'f':
1.133     schwarze 1867:                scaled = *res * 65536;
1.126     schwarze 1868:                break;
                   1869:        case 'i':
1.133     schwarze 1870:                scaled = *res * 240;
1.126     schwarze 1871:                break;
                   1872:        case 'c':
1.133     schwarze 1873:                scaled = *res * 240 / 2.54;
1.126     schwarze 1874:                break;
                   1875:        case 'v':
                   1876:        case 'P':
1.133     schwarze 1877:                scaled = *res * 40;
1.126     schwarze 1878:                break;
                   1879:        case 'm':
                   1880:        case 'n':
1.133     schwarze 1881:                scaled = *res * 24;
1.126     schwarze 1882:                break;
                   1883:        case 'p':
1.133     schwarze 1884:                scaled = *res * 10 / 3;
1.126     schwarze 1885:                break;
                   1886:        case 'u':
1.133     schwarze 1887:                scaled = *res;
1.126     schwarze 1888:                break;
                   1889:        case 'M':
1.133     schwarze 1890:                scaled = *res * 6 / 25;
1.126     schwarze 1891:                break;
                   1892:        default:
1.133     schwarze 1893:                scaled = *res;
1.126     schwarze 1894:                p--;
                   1895:                break;
                   1896:        }
1.133     schwarze 1897:        if (flags & ROFFNUM_SCALE)
                   1898:                *res = scaled;
1.126     schwarze 1899:
                   1900:        *pos = p + 1;
1.149     schwarze 1901:        return 1;
1.56      schwarze 1902: }
                   1903:
1.71      schwarze 1904: /*
                   1905:  * Evaluate a string comparison condition.
                   1906:  * The first character is the delimiter.
                   1907:  * Succeed if the string up to its second occurrence
                   1908:  * matches the string up to its third occurence.
                   1909:  * Advance the cursor after the third occurrence
                   1910:  * or lacking that, to the end of the line.
                   1911:  */
                   1912: static int
                   1913: roff_evalstrcond(const char *v, int *pos)
                   1914: {
                   1915:        const char      *s1, *s2, *s3;
                   1916:        int              match;
                   1917:
                   1918:        match = 0;
                   1919:        s1 = v + *pos;          /* initial delimiter */
                   1920:        s2 = s1 + 1;            /* for scanning the first string */
                   1921:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   1922:
                   1923:        if (NULL == s3)         /* found no middle delimiter */
                   1924:                goto out;
                   1925:
                   1926:        while ('\0' != *++s3) {
                   1927:                if (*s2 != *s3) {  /* mismatch */
                   1928:                        s3 = strchr(s3, *s1);
                   1929:                        break;
                   1930:                }
                   1931:                if (*s3 == *s1) {  /* found the final delimiter */
                   1932:                        match = 1;
                   1933:                        break;
                   1934:                }
                   1935:                s2++;
                   1936:        }
                   1937:
                   1938: out:
                   1939:        if (NULL == s3)
                   1940:                s3 = strchr(s2, '\0');
1.114     schwarze 1941:        else if (*s3 != '\0')
1.71      schwarze 1942:                s3++;
                   1943:        *pos = s3 - v;
1.149     schwarze 1944:        return match;
1.71      schwarze 1945: }
                   1946:
1.77      schwarze 1947: /*
                   1948:  * Evaluate an optionally negated single character, numerical,
                   1949:  * or string condition.
                   1950:  */
1.71      schwarze 1951: static int
1.143     schwarze 1952: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 1953: {
1.143     schwarze 1954:        char    *cp, *name;
                   1955:        size_t   sz;
1.113     schwarze 1956:        int      number, savepos, wanttrue;
1.5       schwarze 1957:
1.71      schwarze 1958:        if ('!' == v[*pos]) {
                   1959:                wanttrue = 0;
                   1960:                (*pos)++;
                   1961:        } else
                   1962:                wanttrue = 1;
                   1963:
1.5       schwarze 1964:        switch (v[*pos]) {
1.112     schwarze 1965:        case '\0':
1.149     schwarze 1966:                return 0;
1.80      schwarze 1967:        case 'n':
                   1968:        case 'o':
1.5       schwarze 1969:                (*pos)++;
1.149     schwarze 1970:                return wanttrue;
1.80      schwarze 1971:        case 'c':
                   1972:        case 'd':
                   1973:        case 'e':
                   1974:        case 't':
1.111     schwarze 1975:        case 'v':
1.5       schwarze 1976:                (*pos)++;
1.149     schwarze 1977:                return !wanttrue;
1.143     schwarze 1978:        case 'r':
                   1979:                cp = name = v + ++*pos;
                   1980:                sz = roff_getname(r, &cp, ln, *pos);
                   1981:                *pos = cp - v;
1.149     schwarze 1982:                return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.5       schwarze 1983:        default:
                   1984:                break;
                   1985:        }
                   1986:
1.113     schwarze 1987:        savepos = *pos;
1.133     schwarze 1988:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 1989:                return (number > 0) == wanttrue;
1.113     schwarze 1990:        else if (*pos == savepos)
1.149     schwarze 1991:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 1992:        else
1.149     schwarze 1993:                return 0;
1.5       schwarze 1994: }
                   1995:
1.2       schwarze 1996: static enum rofferr
1.21      schwarze 1997: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 1998: {
1.30      schwarze 1999:
1.149     schwarze 2000:        return ROFF_IGN;
1.21      schwarze 2001: }
                   2002:
                   2003: static enum rofferr
1.123     schwarze 2004: roff_insec(ROFF_ARGS)
                   2005: {
                   2006:
                   2007:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167   ! schwarze 2008:            ln, ppos, roff_name[tok]);
1.149     schwarze 2009:        return ROFF_IGN;
1.123     schwarze 2010: }
                   2011:
                   2012: static enum rofferr
                   2013: roff_unsupp(ROFF_ARGS)
                   2014: {
                   2015:
                   2016:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167   ! schwarze 2017:            ln, ppos, roff_name[tok]);
1.149     schwarze 2018:        return ROFF_IGN;
1.123     schwarze 2019: }
                   2020:
                   2021: static enum rofferr
1.2       schwarze 2022: roff_cond(ROFF_ARGS)
                   2023: {
1.46      schwarze 2024:
                   2025:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2026:
1.80      schwarze 2027:        /*
1.35      schwarze 2028:         * An `.el' has no conditional body: it will consume the value
                   2029:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2030:         * defaults to DENY.
1.35      schwarze 2031:         *
                   2032:         * If we're not an `el', however, then evaluate the conditional.
                   2033:         */
1.1       schwarze 2034:
1.110     schwarze 2035:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2036:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2037:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2038:
1.35      schwarze 2039:        /*
                   2040:         * An if-else will put the NEGATION of the current evaluated
                   2041:         * conditional into the stack of rules.
                   2042:         */
                   2043:
1.110     schwarze 2044:        if (tok == ROFF_ie) {
1.96      schwarze 2045:                if (r->rstackpos + 1 == r->rstacksz) {
                   2046:                        r->rstacksz += 16;
                   2047:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2048:                            r->rstacksz, sizeof(int));
1.35      schwarze 2049:                }
1.71      schwarze 2050:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2051:        }
1.5       schwarze 2052:
                   2053:        /* If the parent has false as its rule, then so do we. */
                   2054:
1.71      schwarze 2055:        if (r->last->parent && !r->last->parent->rule)
                   2056:                r->last->rule = 0;
1.5       schwarze 2057:
                   2058:        /*
1.46      schwarze 2059:         * Determine scope.
                   2060:         * If there is nothing on the line after the conditional,
                   2061:         * not even whitespace, use next-line scope.
1.5       schwarze 2062:         */
1.2       schwarze 2063:
1.110     schwarze 2064:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2065:                r->last->endspan = 2;
                   2066:                goto out;
                   2067:        }
                   2068:
1.110     schwarze 2069:        while (buf->buf[pos] == ' ')
1.46      schwarze 2070:                pos++;
                   2071:
                   2072:        /* An opening brace requests multiline scope. */
1.2       schwarze 2073:
1.110     schwarze 2074:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2075:                r->last->endspan = -1;
                   2076:                pos += 2;
1.144     schwarze 2077:                while (buf->buf[pos] == ' ')
                   2078:                        pos++;
1.46      schwarze 2079:                goto out;
1.80      schwarze 2080:        }
1.2       schwarze 2081:
                   2082:        /*
1.46      schwarze 2083:         * Anything else following the conditional causes
                   2084:         * single-line scope.  Warn if the scope contains
                   2085:         * nothing but trailing whitespace.
1.2       schwarze 2086:         */
                   2087:
1.110     schwarze 2088:        if (buf->buf[pos] == '\0')
1.89      schwarze 2089:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167   ! schwarze 2090:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2091:
1.46      schwarze 2092:        r->last->endspan = 1;
1.1       schwarze 2093:
1.46      schwarze 2094: out:
1.2       schwarze 2095:        *offs = pos;
1.149     schwarze 2096:        return ROFF_RERUN;
1.1       schwarze 2097: }
                   2098:
1.2       schwarze 2099: static enum rofferr
1.7       schwarze 2100: roff_ds(ROFF_ARGS)
                   2101: {
1.85      schwarze 2102:        char            *string;
                   2103:        const char      *name;
                   2104:        size_t           namesz;
1.10      schwarze 2105:
1.123     schwarze 2106:        /* Ignore groff compatibility mode for now. */
                   2107:
                   2108:        if (tok == ROFF_ds1)
                   2109:                tok = ROFF_ds;
                   2110:        else if (tok == ROFF_as1)
                   2111:                tok = ROFF_as;
                   2112:
1.10      schwarze 2113:        /*
1.85      schwarze 2114:         * The first word is the name of the string.
                   2115:         * If it is empty or terminated by an escape sequence,
                   2116:         * abort the `ds' request without defining anything.
1.10      schwarze 2117:         */
1.7       schwarze 2118:
1.110     schwarze 2119:        name = string = buf->buf + pos;
                   2120:        if (*name == '\0')
1.149     schwarze 2121:                return ROFF_IGN;
1.7       schwarze 2122:
1.85      schwarze 2123:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2124:        if (name[namesz] == '\\')
1.149     schwarze 2125:                return ROFF_IGN;
1.85      schwarze 2126:
                   2127:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2128:        if (*string == '"')
1.7       schwarze 2129:                string++;
                   2130:
1.10      schwarze 2131:        /* The rest is the value. */
1.85      schwarze 2132:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2133:            ROFF_as == tok);
1.149     schwarze 2134:        return ROFF_IGN;
1.7       schwarze 2135: }
                   2136:
1.77      schwarze 2137: /*
                   2138:  * Parse a single operator, one or two characters long.
                   2139:  * If the operator is recognized, return success and advance the
                   2140:  * parse point, else return failure and let the parse point unchanged.
                   2141:  */
                   2142: static int
                   2143: roff_getop(const char *v, int *pos, char *res)
                   2144: {
                   2145:
                   2146:        *res = v[*pos];
                   2147:
                   2148:        switch (*res) {
1.80      schwarze 2149:        case '+':
                   2150:        case '-':
                   2151:        case '*':
                   2152:        case '/':
                   2153:        case '%':
                   2154:        case '&':
                   2155:        case ':':
1.77      schwarze 2156:                break;
                   2157:        case '<':
                   2158:                switch (v[*pos + 1]) {
1.80      schwarze 2159:                case '=':
1.77      schwarze 2160:                        *res = 'l';
                   2161:                        (*pos)++;
                   2162:                        break;
1.80      schwarze 2163:                case '>':
1.77      schwarze 2164:                        *res = '!';
                   2165:                        (*pos)++;
                   2166:                        break;
1.80      schwarze 2167:                case '?':
1.77      schwarze 2168:                        *res = 'i';
                   2169:                        (*pos)++;
                   2170:                        break;
                   2171:                default:
                   2172:                        break;
                   2173:                }
                   2174:                break;
                   2175:        case '>':
                   2176:                switch (v[*pos + 1]) {
1.80      schwarze 2177:                case '=':
1.77      schwarze 2178:                        *res = 'g';
                   2179:                        (*pos)++;
                   2180:                        break;
1.80      schwarze 2181:                case '?':
1.77      schwarze 2182:                        *res = 'a';
                   2183:                        (*pos)++;
                   2184:                        break;
                   2185:                default:
                   2186:                        break;
                   2187:                }
                   2188:                break;
                   2189:        case '=':
                   2190:                if ('=' == v[*pos + 1])
                   2191:                        (*pos)++;
                   2192:                break;
                   2193:        default:
1.149     schwarze 2194:                return 0;
1.77      schwarze 2195:        }
                   2196:        (*pos)++;
                   2197:
1.149     schwarze 2198:        return *res;
1.77      schwarze 2199: }
                   2200:
                   2201: /*
                   2202:  * Evaluate either a parenthesized numeric expression
                   2203:  * or a single signed integer number.
                   2204:  */
                   2205: static int
1.106     schwarze 2206: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2207:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2208: {
                   2209:
                   2210:        if ('(' != v[*pos])
1.149     schwarze 2211:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2212:
                   2213:        (*pos)++;
1.133     schwarze 2214:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2215:                return 0;
1.77      schwarze 2216:
1.79      schwarze 2217:        /*
                   2218:         * Omission of the closing parenthesis
                   2219:         * is an error in validation mode,
                   2220:         * but ignored in evaluation mode.
                   2221:         */
                   2222:
1.77      schwarze 2223:        if (')' == v[*pos])
                   2224:                (*pos)++;
1.79      schwarze 2225:        else if (NULL == res)
1.149     schwarze 2226:                return 0;
1.77      schwarze 2227:
1.149     schwarze 2228:        return 1;
1.77      schwarze 2229: }
                   2230:
                   2231: /*
                   2232:  * Evaluate a complete numeric expression.
                   2233:  * Proceed left to right, there is no concept of precedence.
                   2234:  */
                   2235: static int
1.106     schwarze 2236: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2237:        int *pos, int *res, int flags)
1.77      schwarze 2238: {
                   2239:        int              mypos, operand2;
                   2240:        char             operator;
                   2241:
                   2242:        if (NULL == pos) {
                   2243:                mypos = 0;
                   2244:                pos = &mypos;
                   2245:        }
                   2246:
1.133     schwarze 2247:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2248:                while (isspace((unsigned char)v[*pos]))
                   2249:                        (*pos)++;
                   2250:
1.133     schwarze 2251:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2252:                return 0;
1.77      schwarze 2253:
                   2254:        while (1) {
1.133     schwarze 2255:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2256:                        while (isspace((unsigned char)v[*pos]))
                   2257:                                (*pos)++;
                   2258:
                   2259:                if ( ! roff_getop(v, pos, &operator))
                   2260:                        break;
                   2261:
1.133     schwarze 2262:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2263:                        while (isspace((unsigned char)v[*pos]))
                   2264:                                (*pos)++;
                   2265:
1.133     schwarze 2266:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2267:                        return 0;
1.77      schwarze 2268:
1.133     schwarze 2269:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2270:                        while (isspace((unsigned char)v[*pos]))
                   2271:                                (*pos)++;
1.79      schwarze 2272:
                   2273:                if (NULL == res)
                   2274:                        continue;
1.77      schwarze 2275:
                   2276:                switch (operator) {
1.80      schwarze 2277:                case '+':
1.77      schwarze 2278:                        *res += operand2;
                   2279:                        break;
1.80      schwarze 2280:                case '-':
1.77      schwarze 2281:                        *res -= operand2;
                   2282:                        break;
1.80      schwarze 2283:                case '*':
1.77      schwarze 2284:                        *res *= operand2;
                   2285:                        break;
1.80      schwarze 2286:                case '/':
1.116     schwarze 2287:                        if (operand2 == 0) {
1.106     schwarze 2288:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2289:                                        r->parse, ln, *pos, v);
                   2290:                                *res = 0;
                   2291:                                break;
                   2292:                        }
1.77      schwarze 2293:                        *res /= operand2;
                   2294:                        break;
1.80      schwarze 2295:                case '%':
1.116     schwarze 2296:                        if (operand2 == 0) {
                   2297:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2298:                                        r->parse, ln, *pos, v);
                   2299:                                *res = 0;
                   2300:                                break;
                   2301:                        }
1.77      schwarze 2302:                        *res %= operand2;
                   2303:                        break;
1.80      schwarze 2304:                case '<':
1.77      schwarze 2305:                        *res = *res < operand2;
                   2306:                        break;
1.80      schwarze 2307:                case '>':
1.77      schwarze 2308:                        *res = *res > operand2;
                   2309:                        break;
1.80      schwarze 2310:                case 'l':
1.77      schwarze 2311:                        *res = *res <= operand2;
                   2312:                        break;
1.80      schwarze 2313:                case 'g':
1.77      schwarze 2314:                        *res = *res >= operand2;
                   2315:                        break;
1.80      schwarze 2316:                case '=':
1.77      schwarze 2317:                        *res = *res == operand2;
                   2318:                        break;
1.80      schwarze 2319:                case '!':
1.77      schwarze 2320:                        *res = *res != operand2;
                   2321:                        break;
1.80      schwarze 2322:                case '&':
1.77      schwarze 2323:                        *res = *res && operand2;
                   2324:                        break;
1.80      schwarze 2325:                case ':':
1.77      schwarze 2326:                        *res = *res || operand2;
                   2327:                        break;
1.80      schwarze 2328:                case 'i':
1.77      schwarze 2329:                        if (operand2 < *res)
                   2330:                                *res = operand2;
                   2331:                        break;
1.80      schwarze 2332:                case 'a':
1.77      schwarze 2333:                        if (operand2 > *res)
                   2334:                                *res = operand2;
                   2335:                        break;
                   2336:                default:
                   2337:                        abort();
                   2338:                }
                   2339:        }
1.149     schwarze 2340:        return 1;
1.77      schwarze 2341: }
                   2342:
1.138     schwarze 2343: /* --- register management ------------------------------------------------ */
                   2344:
1.52      schwarze 2345: void
1.60      schwarze 2346: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2347: {
1.52      schwarze 2348:        struct roffreg  *reg;
                   2349:
                   2350:        /* Search for an existing register with the same name. */
                   2351:        reg = r->regtab;
                   2352:
                   2353:        while (reg && strcmp(name, reg->key.p))
                   2354:                reg = reg->next;
1.41      schwarze 2355:
1.52      schwarze 2356:        if (NULL == reg) {
                   2357:                /* Create a new register. */
                   2358:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2359:                reg->key.p = mandoc_strdup(name);
                   2360:                reg->key.sz = strlen(name);
1.60      schwarze 2361:                reg->val = 0;
1.52      schwarze 2362:                reg->next = r->regtab;
                   2363:                r->regtab = reg;
                   2364:        }
                   2365:
1.60      schwarze 2366:        if ('+' == sign)
                   2367:                reg->val += val;
                   2368:        else if ('-' == sign)
                   2369:                reg->val -= val;
                   2370:        else
                   2371:                reg->val = val;
1.41      schwarze 2372: }
                   2373:
1.65      schwarze 2374: /*
                   2375:  * Handle some predefined read-only number registers.
                   2376:  * For now, return -1 if the requested register is not predefined;
                   2377:  * in case a predefined read-only register having the value -1
                   2378:  * were to turn up, another special value would have to be chosen.
                   2379:  */
                   2380: static int
1.145     schwarze 2381: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2382: {
                   2383:
                   2384:        switch (*name) {
1.145     schwarze 2385:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2386:                return r->argc;
1.80      schwarze 2387:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2388:                return 0;
1.80      schwarze 2389:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2390:                return 1;
1.80      schwarze 2391:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2392:                return 24;
1.80      schwarze 2393:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2394:                return 0;
1.80      schwarze 2395:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2396:                return 1;
1.80      schwarze 2397:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2398:                return 40;
1.65      schwarze 2399:        default:
1.149     schwarze 2400:                return -1;
1.65      schwarze 2401:        }
                   2402: }
                   2403:
1.53      schwarze 2404: int
1.52      schwarze 2405: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2406: {
1.52      schwarze 2407:        struct roffreg  *reg;
1.65      schwarze 2408:        int              val;
                   2409:
                   2410:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145     schwarze 2411:                val = roff_getregro(r, name + 1);
1.65      schwarze 2412:                if (-1 != val)
1.149     schwarze 2413:                        return val;
1.65      schwarze 2414:        }
1.52      schwarze 2415:
                   2416:        for (reg = r->regtab; reg; reg = reg->next)
                   2417:                if (0 == strcmp(name, reg->key.p))
1.149     schwarze 2418:                        return reg->val;
1.53      schwarze 2419:
1.149     schwarze 2420:        return 0;
1.53      schwarze 2421: }
                   2422:
                   2423: static int
                   2424: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2425: {
                   2426:        struct roffreg  *reg;
1.65      schwarze 2427:        int              val;
                   2428:
                   2429:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2430:                val = roff_getregro(r, name + 1);
1.65      schwarze 2431:                if (-1 != val)
1.149     schwarze 2432:                        return val;
1.65      schwarze 2433:        }
1.53      schwarze 2434:
                   2435:        for (reg = r->regtab; reg; reg = reg->next)
                   2436:                if (len == reg->key.sz &&
                   2437:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2438:                        return reg->val;
1.143     schwarze 2439:
1.149     schwarze 2440:        return 0;
1.143     schwarze 2441: }
                   2442:
                   2443: static int
                   2444: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2445: {
                   2446:        struct roffreg  *reg;
                   2447:        int              val;
                   2448:
                   2449:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2450:                val = roff_getregro(r, name + 1);
1.143     schwarze 2451:                if (-1 != val)
1.149     schwarze 2452:                        return 1;
1.143     schwarze 2453:        }
                   2454:
                   2455:        for (reg = r->regtab; reg; reg = reg->next)
                   2456:                if (len == reg->key.sz &&
                   2457:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2458:                        return 1;
1.41      schwarze 2459:
1.149     schwarze 2460:        return 0;
1.41      schwarze 2461: }
                   2462:
1.52      schwarze 2463: static void
                   2464: roff_freereg(struct roffreg *reg)
1.41      schwarze 2465: {
1.52      schwarze 2466:        struct roffreg  *old_reg;
1.41      schwarze 2467:
1.52      schwarze 2468:        while (NULL != reg) {
                   2469:                free(reg->key.p);
                   2470:                old_reg = reg;
                   2471:                reg = reg->next;
                   2472:                free(old_reg);
                   2473:        }
1.41      schwarze 2474: }
1.7       schwarze 2475:
                   2476: static enum rofferr
1.6       schwarze 2477: roff_nr(ROFF_ARGS)
1.1       schwarze 2478: {
1.85      schwarze 2479:        char            *key, *val;
                   2480:        size_t           keysz;
1.37      schwarze 2481:        int              iv;
1.60      schwarze 2482:        char             sign;
1.6       schwarze 2483:
1.110     schwarze 2484:        key = val = buf->buf + pos;
                   2485:        if (*key == '\0')
1.149     schwarze 2486:                return ROFF_IGN;
1.85      schwarze 2487:
                   2488:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2489:        if (key[keysz] == '\\')
1.149     schwarze 2490:                return ROFF_IGN;
1.85      schwarze 2491:        key[keysz] = '\0';
1.6       schwarze 2492:
1.60      schwarze 2493:        sign = *val;
1.110     schwarze 2494:        if (sign == '+' || sign == '-')
1.60      schwarze 2495:                val++;
                   2496:
1.133     schwarze 2497:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77      schwarze 2498:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2499:
1.149     schwarze 2500:        return ROFF_IGN;
1.76      schwarze 2501: }
                   2502:
                   2503: static enum rofferr
                   2504: roff_rr(ROFF_ARGS)
                   2505: {
                   2506:        struct roffreg  *reg, **prev;
1.85      schwarze 2507:        char            *name, *cp;
                   2508:        size_t           namesz;
1.76      schwarze 2509:
1.110     schwarze 2510:        name = cp = buf->buf + pos;
                   2511:        if (*name == '\0')
1.149     schwarze 2512:                return ROFF_IGN;
1.85      schwarze 2513:        namesz = roff_getname(r, &cp, ln, pos);
                   2514:        name[namesz] = '\0';
1.76      schwarze 2515:
                   2516:        prev = &r->regtab;
                   2517:        while (1) {
                   2518:                reg = *prev;
1.110     schwarze 2519:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2520:                        break;
                   2521:                prev = &reg->next;
                   2522:        }
1.110     schwarze 2523:        if (reg != NULL) {
1.76      schwarze 2524:                *prev = reg->next;
                   2525:                free(reg->key.p);
                   2526:                free(reg);
                   2527:        }
1.149     schwarze 2528:        return ROFF_IGN;
1.29      schwarze 2529: }
                   2530:
1.138     schwarze 2531: /* --- handler functions for roff requests -------------------------------- */
                   2532:
1.29      schwarze 2533: static enum rofferr
                   2534: roff_rm(ROFF_ARGS)
                   2535: {
                   2536:        const char       *name;
                   2537:        char             *cp;
1.85      schwarze 2538:        size_t            namesz;
1.29      schwarze 2539:
1.110     schwarze 2540:        cp = buf->buf + pos;
                   2541:        while (*cp != '\0') {
1.85      schwarze 2542:                name = cp;
1.110     schwarze 2543:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2544:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110     schwarze 2545:                if (name[namesz] == '\\')
1.85      schwarze 2546:                        break;
1.29      schwarze 2547:        }
1.149     schwarze 2548:        return ROFF_IGN;
1.51      schwarze 2549: }
                   2550:
                   2551: static enum rofferr
                   2552: roff_it(ROFF_ARGS)
                   2553: {
                   2554:        int              iv;
                   2555:
                   2556:        /* Parse the number of lines. */
1.133     schwarze 2557:
                   2558:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2559:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2560:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2561:                return ROFF_IGN;
1.51      schwarze 2562:        }
                   2563:
1.134     schwarze 2564:        while (isspace((unsigned char)buf->buf[pos]))
                   2565:                pos++;
                   2566:
                   2567:        /*
                   2568:         * Arm the input line trap.
                   2569:         * Special-casing "an-trap" is an ugly workaround to cope
                   2570:         * with DocBook stupidly fiddling with man(7) internals.
                   2571:         */
1.133     schwarze 2572:
1.51      schwarze 2573:        roffit_lines = iv;
1.134     schwarze 2574:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2575:            strcmp(buf->buf + pos, "an-trap") ?
                   2576:            buf->buf + pos : "br");
1.149     schwarze 2577:        return ROFF_IGN;
1.47      schwarze 2578: }
                   2579:
                   2580: static enum rofferr
                   2581: roff_Dd(ROFF_ARGS)
                   2582: {
                   2583:        const char *const       *cp;
                   2584:
1.99      schwarze 2585:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47      schwarze 2586:                for (cp = __mdoc_reserved; *cp; cp++)
                   2587:                        roff_setstr(r, *cp, NULL, 0);
                   2588:
1.99      schwarze 2589:        if (r->format == 0)
                   2590:                r->format = MPARSE_MDOC;
                   2591:
1.149     schwarze 2592:        return ROFF_CONT;
1.47      schwarze 2593: }
                   2594:
                   2595: static enum rofferr
                   2596: roff_TH(ROFF_ARGS)
                   2597: {
                   2598:        const char *const       *cp;
                   2599:
1.99      schwarze 2600:        if ((r->options & MPARSE_QUICK) == 0)
1.47      schwarze 2601:                for (cp = __man_reserved; *cp; cp++)
                   2602:                        roff_setstr(r, *cp, NULL, 0);
                   2603:
1.99      schwarze 2604:        if (r->format == 0)
                   2605:                r->format = MPARSE_MAN;
                   2606:
1.149     schwarze 2607:        return ROFF_CONT;
1.14      schwarze 2608: }
                   2609:
                   2610: static enum rofferr
1.27      schwarze 2611: roff_TE(ROFF_ARGS)
                   2612: {
                   2613:
                   2614:        if (NULL == r->tbl)
1.94      schwarze 2615:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2616:                    ln, ppos, "TE");
1.130     schwarze 2617:        else if ( ! tbl_end(&r->tbl)) {
                   2618:                free(buf->buf);
                   2619:                buf->buf = mandoc_strdup(".sp");
                   2620:                buf->sz = 4;
1.149     schwarze 2621:                return ROFF_REPARSE;
1.130     schwarze 2622:        }
1.149     schwarze 2623:        return ROFF_IGN;
1.27      schwarze 2624: }
                   2625:
                   2626: static enum rofferr
                   2627: roff_T_(ROFF_ARGS)
                   2628: {
                   2629:
                   2630:        if (NULL == r->tbl)
1.94      schwarze 2631:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2632:                    ln, ppos, "T&");
1.27      schwarze 2633:        else
                   2634:                tbl_restart(ppos, ln, r->tbl);
                   2635:
1.149     schwarze 2636:        return ROFF_IGN;
1.27      schwarze 2637: }
                   2638:
1.102     schwarze 2639: /*
                   2640:  * Handle in-line equation delimiters.
                   2641:  */
                   2642: static enum rofferr
1.110     schwarze 2643: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2644: {
1.105     schwarze 2645:        char            *cp1, *cp2;
                   2646:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2647:
1.102     schwarze 2648:        /*
                   2649:         * Outside equations, look for an opening delimiter.
                   2650:         * If we are inside an equation, we already know it is
                   2651:         * in-line, or this function wouldn't have been called;
                   2652:         * so look for a closing delimiter.
                   2653:         */
                   2654:
1.110     schwarze 2655:        cp1 = buf->buf + pos;
1.102     schwarze 2656:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2657:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2658:        if (cp2 == NULL)
1.149     schwarze 2659:                return ROFF_CONT;
1.102     schwarze 2660:
1.105     schwarze 2661:        *cp2++ = '\0';
                   2662:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2663:
                   2664:        /* Handle preceding text, protecting whitespace. */
                   2665:
1.110     schwarze 2666:        if (*buf->buf != '\0') {
1.105     schwarze 2667:                if (r->eqn == NULL)
                   2668:                        bef_pr = "\\&";
                   2669:                bef_nl = "\n";
                   2670:        }
                   2671:
                   2672:        /*
                   2673:         * Prepare replacing the delimiter with an equation macro
                   2674:         * and drop leading white space from the equation.
                   2675:         */
1.102     schwarze 2676:
1.105     schwarze 2677:        if (r->eqn == NULL) {
                   2678:                while (*cp2 == ' ')
                   2679:                        cp2++;
                   2680:                mac = ".EQ";
                   2681:        } else
                   2682:                mac = ".EN";
                   2683:
                   2684:        /* Handle following text, protecting whitespace. */
                   2685:
                   2686:        if (*cp2 != '\0') {
                   2687:                aft_nl = "\n";
                   2688:                if (r->eqn != NULL)
                   2689:                        aft_pr = "\\&";
                   2690:        }
                   2691:
                   2692:        /* Do the actual replacement. */
                   2693:
1.110     schwarze 2694:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2695:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2696:        free(buf->buf);
                   2697:        buf->buf = cp1;
1.102     schwarze 2698:
                   2699:        /* Toggle the in-line state of the eqn subsystem. */
                   2700:
                   2701:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2702:        return ROFF_REPARSE;
1.41      schwarze 2703: }
                   2704:
1.107     schwarze 2705: static enum rofferr
                   2706: roff_EQ(ROFF_ARGS)
1.32      schwarze 2707: {
1.41      schwarze 2708:        struct eqn_node *e;
1.32      schwarze 2709:
1.110     schwarze 2710:        assert(r->eqn == NULL);
1.107     schwarze 2711:        e = eqn_alloc(ppos, ln, r->parse);
1.32      schwarze 2712:
1.102     schwarze 2713:        if (r->last_eqn) {
1.32      schwarze 2714:                r->last_eqn->next = e;
1.102     schwarze 2715:                e->delim = r->last_eqn->delim;
                   2716:                e->odelim = r->last_eqn->odelim;
                   2717:                e->cdelim = r->last_eqn->cdelim;
                   2718:        } else
1.32      schwarze 2719:                r->first_eqn = r->last_eqn = e;
                   2720:
                   2721:        r->eqn = r->last_eqn = e;
1.41      schwarze 2722:
1.110     schwarze 2723:        if (buf->buf[pos] != '\0')
1.107     schwarze 2724:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2725:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2726:
1.149     schwarze 2727:        return ROFF_IGN;
1.32      schwarze 2728: }
                   2729:
                   2730: static enum rofferr
                   2731: roff_EN(ROFF_ARGS)
                   2732: {
                   2733:
1.94      schwarze 2734:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.149     schwarze 2735:        return ROFF_IGN;
1.32      schwarze 2736: }
                   2737:
                   2738: static enum rofferr
1.27      schwarze 2739: roff_TS(ROFF_ARGS)
                   2740: {
1.49      schwarze 2741:        struct tbl_node *tbl;
1.27      schwarze 2742:
                   2743:        if (r->tbl) {
1.94      schwarze 2744:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2745:                    ln, ppos, "TS breaks TS");
1.41      schwarze 2746:                tbl_end(&r->tbl);
1.27      schwarze 2747:        }
                   2748:
1.49      schwarze 2749:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2750:
                   2751:        if (r->last_tbl)
1.49      schwarze 2752:                r->last_tbl->next = tbl;
1.27      schwarze 2753:        else
1.49      schwarze 2754:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 2755:
1.49      schwarze 2756:        r->tbl = r->last_tbl = tbl;
1.149     schwarze 2757:        return ROFF_IGN;
1.123     schwarze 2758: }
                   2759:
                   2760: static enum rofferr
                   2761: roff_brp(ROFF_ARGS)
                   2762: {
                   2763:
                   2764:        buf->buf[pos - 1] = '\0';
1.149     schwarze 2765:        return ROFF_CONT;
1.27      schwarze 2766: }
                   2767:
                   2768: static enum rofferr
1.48      schwarze 2769: roff_cc(ROFF_ARGS)
                   2770: {
                   2771:        const char      *p;
                   2772:
1.110     schwarze 2773:        p = buf->buf + pos;
1.48      schwarze 2774:
1.110     schwarze 2775:        if (*p == '\0' || (r->control = *p++) == '.')
1.48      schwarze 2776:                r->control = 0;
                   2777:
1.110     schwarze 2778:        if (*p != '\0')
1.132     schwarze 2779:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   2780:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 2781:
1.149     schwarze 2782:        return ROFF_IGN;
1.48      schwarze 2783: }
                   2784:
                   2785: static enum rofferr
1.42      schwarze 2786: roff_tr(ROFF_ARGS)
                   2787: {
                   2788:        const char      *p, *first, *second;
                   2789:        size_t           fsz, ssz;
                   2790:        enum mandoc_esc  esc;
                   2791:
1.110     schwarze 2792:        p = buf->buf + pos;
1.42      schwarze 2793:
1.110     schwarze 2794:        if (*p == '\0') {
1.132     schwarze 2795:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 2796:                return ROFF_IGN;
1.42      schwarze 2797:        }
                   2798:
1.110     schwarze 2799:        while (*p != '\0') {
1.42      schwarze 2800:                fsz = ssz = 1;
                   2801:
                   2802:                first = p++;
1.110     schwarze 2803:                if (*first == '\\') {
1.42      schwarze 2804:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2805:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2806:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2807:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 2808:                                return ROFF_IGN;
1.42      schwarze 2809:                        }
                   2810:                        fsz = (size_t)(p - first);
                   2811:                }
                   2812:
                   2813:                second = p++;
1.110     schwarze 2814:                if (*second == '\\') {
1.42      schwarze 2815:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2816:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2817:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2818:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 2819:                                return ROFF_IGN;
1.42      schwarze 2820:                        }
                   2821:                        ssz = (size_t)(p - second);
1.110     schwarze 2822:                } else if (*second == '\0') {
1.132     schwarze 2823:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   2824:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 2825:                        second = " ";
                   2826:                        p--;
                   2827:                }
                   2828:
                   2829:                if (fsz > 1) {
1.80      schwarze 2830:                        roff_setstrn(&r->xmbtab, first, fsz,
                   2831:                            second, ssz, 0);
1.42      schwarze 2832:                        continue;
                   2833:                }
                   2834:
1.110     schwarze 2835:                if (r->xtab == NULL)
1.80      schwarze 2836:                        r->xtab = mandoc_calloc(128,
                   2837:                            sizeof(struct roffstr));
1.42      schwarze 2838:
                   2839:                free(r->xtab[(int)*first].p);
                   2840:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   2841:                r->xtab[(int)*first].sz = ssz;
                   2842:        }
                   2843:
1.149     schwarze 2844:        return ROFF_IGN;
1.42      schwarze 2845: }
                   2846:
                   2847: static enum rofferr
1.14      schwarze 2848: roff_so(ROFF_ARGS)
                   2849: {
1.121     schwarze 2850:        char *name, *cp;
1.15      schwarze 2851:
1.110     schwarze 2852:        name = buf->buf + pos;
1.97      schwarze 2853:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 2854:
1.22      schwarze 2855:        /*
                   2856:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   2857:         * opening anything that's not in our cwd or anything beneath
                   2858:         * it.  Thus, explicitly disallow traversing up the file-system
                   2859:         * or using absolute paths.
                   2860:         */
                   2861:
1.110     schwarze 2862:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 2863:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   2864:                    ".so %s", name);
1.121     schwarze 2865:                buf->sz = mandoc_asprintf(&cp,
                   2866:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   2867:                free(buf->buf);
                   2868:                buf->buf = cp;
                   2869:                *offs = 0;
1.149     schwarze 2870:                return ROFF_REPARSE;
1.14      schwarze 2871:        }
                   2872:
                   2873:        *offs = pos;
1.149     schwarze 2874:        return ROFF_SO;
1.7       schwarze 2875: }
                   2876:
1.138     schwarze 2877: /* --- user defined strings and macros ------------------------------------ */
                   2878:
1.16      schwarze 2879: static enum rofferr
                   2880: roff_userdef(ROFF_ARGS)
1.12      schwarze 2881: {
1.135     schwarze 2882:        const char       *arg[9], *ap;
1.16      schwarze 2883:        char             *cp, *n1, *n2;
1.164     schwarze 2884:        int               expand_count, i, ib, ie;
1.135     schwarze 2885:        size_t            asz, rsz;
1.12      schwarze 2886:
1.16      schwarze 2887:        /*
                   2888:         * Collect pointers to macro argument strings
1.61      schwarze 2889:         * and NUL-terminate them.
1.16      schwarze 2890:         */
1.135     schwarze 2891:
1.145     schwarze 2892:        r->argc = 0;
1.110     schwarze 2893:        cp = buf->buf + pos;
1.145     schwarze 2894:        for (i = 0; i < 9; i++) {
                   2895:                if (*cp == '\0')
                   2896:                        arg[i] = "";
                   2897:                else {
                   2898:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   2899:                        r->argc = i + 1;
                   2900:                }
                   2901:        }
1.16      schwarze 2902:
                   2903:        /*
                   2904:         * Expand macro arguments.
1.12      schwarze 2905:         */
1.135     schwarze 2906:
                   2907:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 2908:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 2909:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 2910:        expand_count = 0;
1.135     schwarze 2911:        while (*cp != '\0') {
                   2912:
                   2913:                /* Scan ahead for the next argument invocation. */
                   2914:
                   2915:                if (*cp++ != '\\')
                   2916:                        continue;
                   2917:                if (*cp++ != '$')
                   2918:                        continue;
1.146     schwarze 2919:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   2920:                        ib = 0;
                   2921:                        ie = r->argc - 1;
                   2922:                } else {  /* \\$1 .. \\$9 insert one argument */
                   2923:                        ib = ie = *cp - '1';
                   2924:                        if (ib < 0 || ib > 8)
                   2925:                                continue;
                   2926:                }
1.135     schwarze 2927:                cp -= 2;
1.164     schwarze 2928:
                   2929:                /*
                   2930:                 * Prevent infinite recursion.
                   2931:                 */
                   2932:
                   2933:                if (cp >= n2)
                   2934:                        expand_count = 1;
                   2935:                else if (++expand_count > EXPAND_LIMIT) {
                   2936:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   2937:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 2938:                        free(buf->buf);
                   2939:                        buf->buf = n1;
1.164     schwarze 2940:                        return ROFF_IGN;
                   2941:                }
1.135     schwarze 2942:
                   2943:                /*
                   2944:                 * Determine the size of the expanded argument,
                   2945:                 * taking escaping of quotes into account.
                   2946:                 */
                   2947:
1.146     schwarze 2948:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   2949:                for (i = ib; i <= ie; i++) {
                   2950:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   2951:                                asz++;
                   2952:                                if (*ap == '"')
                   2953:                                        asz += 3;
                   2954:                        }
1.135     schwarze 2955:                }
                   2956:                if (asz != 3) {
                   2957:
                   2958:                        /*
                   2959:                         * Determine the size of the rest of the
                   2960:                         * unexpanded macro, including the NUL.
                   2961:                         */
                   2962:
                   2963:                        rsz = buf->sz - (cp - n1) - 3;
                   2964:
                   2965:                        /*
                   2966:                         * When shrinking, move before
                   2967:                         * releasing the storage.
                   2968:                         */
                   2969:
                   2970:                        if (asz < 3)
                   2971:                                memmove(cp + asz, cp + 3, rsz);
                   2972:
                   2973:                        /*
                   2974:                         * Resize the storage for the macro
                   2975:                         * and readjust the parse pointer.
                   2976:                         */
                   2977:
                   2978:                        buf->sz += asz - 3;
                   2979:                        n2 = mandoc_realloc(n1, buf->sz);
                   2980:                        cp = n2 + (cp - n1);
                   2981:                        n1 = n2;
                   2982:
                   2983:                        /*
                   2984:                         * When growing, make room
                   2985:                         * for the expanded argument.
                   2986:                         */
                   2987:
                   2988:                        if (asz > 3)
                   2989:                                memmove(cp + asz, cp + 3, rsz);
                   2990:                }
                   2991:
                   2992:                /* Copy the expanded argument, escaping quotes. */
                   2993:
                   2994:                n2 = cp;
1.146     schwarze 2995:                for (i = ib; i <= ie; i++) {
                   2996:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   2997:                                if (*ap == '"') {
                   2998:                                        memcpy(n2, "\\(dq", 4);
                   2999:                                        n2 += 4;
                   3000:                                } else
                   3001:                                        *n2++ = *ap;
                   3002:                        }
                   3003:                        if (i < ie)
                   3004:                                *n2++ = ' ';
1.16      schwarze 3005:                }
1.12      schwarze 3006:        }
                   3007:
1.16      schwarze 3008:        /*
                   3009:         * Replace the macro invocation
                   3010:         * by the expanded macro.
                   3011:         */
1.135     schwarze 3012:
1.110     schwarze 3013:        free(buf->buf);
                   3014:        buf->buf = n1;
1.120     schwarze 3015:        *offs = 0;
1.16      schwarze 3016:
1.149     schwarze 3017:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3018:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3019: }
1.28      schwarze 3020:
1.85      schwarze 3021: static size_t
1.28      schwarze 3022: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3023: {
                   3024:        char     *name, *cp;
1.85      schwarze 3025:        size_t    namesz;
1.28      schwarze 3026:
                   3027:        name = *cpp;
                   3028:        if ('\0' == *name)
1.149     schwarze 3029:                return 0;
1.28      schwarze 3030:
1.85      schwarze 3031:        /* Read until end of name and terminate it with NUL. */
                   3032:        for (cp = name; 1; cp++) {
                   3033:                if ('\0' == *cp || ' ' == *cp) {
                   3034:                        namesz = cp - name;
                   3035:                        break;
                   3036:                }
1.28      schwarze 3037:                if ('\\' != *cp)
                   3038:                        continue;
1.88      schwarze 3039:                namesz = cp - name;
                   3040:                if ('{' == cp[1] || '}' == cp[1])
                   3041:                        break;
1.28      schwarze 3042:                cp++;
                   3043:                if ('\\' == *cp)
                   3044:                        continue;
1.97      schwarze 3045:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3046:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3047:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3048:                break;
1.28      schwarze 3049:        }
                   3050:
                   3051:        /* Read past spaces. */
                   3052:        while (' ' == *cp)
                   3053:                cp++;
                   3054:
                   3055:        *cpp = cp;
1.149     schwarze 3056:        return namesz;
1.28      schwarze 3057: }
                   3058:
1.16      schwarze 3059: /*
                   3060:  * Store *string into the user-defined string called *name.
                   3061:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3062:  * append == 0: replace mode
                   3063:  * append == 1: single-line append mode
                   3064:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3065:  */
1.8       schwarze 3066: static void
1.16      schwarze 3067: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3068:        int append)
1.7       schwarze 3069: {
1.42      schwarze 3070:
                   3071:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.80      schwarze 3072:            string ? strlen(string) : 0, append);
1.42      schwarze 3073: }
                   3074:
                   3075: static void
                   3076: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3077:                const char *string, size_t stringsz, int append)
1.42      schwarze 3078: {
                   3079:        struct roffkv   *n;
                   3080:        char            *c;
                   3081:        int              i;
                   3082:        size_t           oldch, newch;
1.7       schwarze 3083:
1.16      schwarze 3084:        /* Search for an existing string with the same name. */
1.42      schwarze 3085:        n = *r;
                   3086:
1.84      schwarze 3087:        while (n && (namesz != n->key.sz ||
                   3088:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3089:                n = n->next;
1.8       schwarze 3090:
                   3091:        if (NULL == n) {
1.16      schwarze 3092:                /* Create a new string table entry. */
1.42      schwarze 3093:                n = mandoc_malloc(sizeof(struct roffkv));
                   3094:                n->key.p = mandoc_strndup(name, namesz);
                   3095:                n->key.sz = namesz;
                   3096:                n->val.p = NULL;
                   3097:                n->val.sz = 0;
                   3098:                n->next = *r;
                   3099:                *r = n;
1.66      schwarze 3100:        } else if (0 == append) {
1.42      schwarze 3101:                free(n->val.p);
                   3102:                n->val.p = NULL;
                   3103:                n->val.sz = 0;
1.16      schwarze 3104:        }
                   3105:
                   3106:        if (NULL == string)
                   3107:                return;
                   3108:
                   3109:        /*
                   3110:         * One additional byte for the '\n' in multiline mode,
                   3111:         * and one for the terminating '\0'.
                   3112:         */
1.66      schwarze 3113:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3114:
                   3115:        if (NULL == n->val.p) {
                   3116:                n->val.p = mandoc_malloc(newch);
                   3117:                *n->val.p = '\0';
1.16      schwarze 3118:                oldch = 0;
                   3119:        } else {
1.42      schwarze 3120:                oldch = n->val.sz;
                   3121:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3122:        }
                   3123:
                   3124:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3125:        c = n->val.p + (int)oldch;
1.16      schwarze 3126:
                   3127:        /* Append new content to the destination buffer. */
1.42      schwarze 3128:        i = 0;
                   3129:        while (i < (int)stringsz) {
1.16      schwarze 3130:                /*
                   3131:                 * Rudimentary roff copy mode:
                   3132:                 * Handle escaped backslashes.
                   3133:                 */
1.42      schwarze 3134:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3135:                        i++;
                   3136:                *c++ = string[i++];
1.16      schwarze 3137:        }
1.8       schwarze 3138:
1.16      schwarze 3139:        /* Append terminating bytes. */
1.66      schwarze 3140:        if (1 < append)
1.16      schwarze 3141:                *c++ = '\n';
1.42      schwarze 3142:
1.16      schwarze 3143:        *c = '\0';
1.42      schwarze 3144:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3145: }
                   3146:
1.8       schwarze 3147: static const char *
                   3148: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7       schwarze 3149: {
1.42      schwarze 3150:        const struct roffkv *n;
1.64      schwarze 3151:        int i;
1.7       schwarze 3152:
1.42      schwarze 3153:        for (n = r->strtab; n; n = n->next)
1.80      schwarze 3154:                if (0 == strncmp(name, n->key.p, len) &&
                   3155:                    '\0' == n->key.p[(int)len])
1.149     schwarze 3156:                        return n->val.p;
1.64      schwarze 3157:
                   3158:        for (i = 0; i < PREDEFS_MAX; i++)
                   3159:                if (0 == strncmp(name, predefs[i].name, len) &&
                   3160:                                '\0' == predefs[i].name[(int)len])
1.149     schwarze 3161:                        return predefs[i].str;
1.8       schwarze 3162:
1.149     schwarze 3163:        return NULL;
1.7       schwarze 3164: }
                   3165:
1.8       schwarze 3166: static void
1.42      schwarze 3167: roff_freestr(struct roffkv *r)
1.7       schwarze 3168: {
1.42      schwarze 3169:        struct roffkv    *n, *nn;
1.7       schwarze 3170:
1.42      schwarze 3171:        for (n = r; n; n = nn) {
                   3172:                free(n->key.p);
                   3173:                free(n->val.p);
1.7       schwarze 3174:                nn = n->next;
                   3175:                free(n);
                   3176:        }
1.27      schwarze 3177: }
1.138     schwarze 3178:
                   3179: /* --- accessors and utility functions ------------------------------------ */
1.27      schwarze 3180:
                   3181: const struct tbl_span *
                   3182: roff_span(const struct roff *r)
                   3183: {
1.80      schwarze 3184:
1.149     schwarze 3185:        return r->tbl ? tbl_span(r->tbl) : NULL;
1.32      schwarze 3186: }
                   3187:
                   3188: const struct eqn *
                   3189: roff_eqn(const struct roff *r)
                   3190: {
1.80      schwarze 3191:
1.149     schwarze 3192:        return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.42      schwarze 3193: }
                   3194:
                   3195: /*
                   3196:  * Duplicate an input string, making the appropriate character
                   3197:  * conversations (as stipulated by `tr') along the way.
                   3198:  * Returns a heap-allocated string with all the replacements made.
                   3199:  */
                   3200: char *
                   3201: roff_strdup(const struct roff *r, const char *p)
                   3202: {
                   3203:        const struct roffkv *cp;
                   3204:        char            *res;
                   3205:        const char      *pp;
                   3206:        size_t           ssz, sz;
                   3207:        enum mandoc_esc  esc;
                   3208:
                   3209:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3210:                return mandoc_strdup(p);
1.42      schwarze 3211:        else if ('\0' == *p)
1.149     schwarze 3212:                return mandoc_strdup("");
1.42      schwarze 3213:
                   3214:        /*
                   3215:         * Step through each character looking for term matches
                   3216:         * (remember that a `tr' can be invoked with an escape, which is
                   3217:         * a glyph but the escape is multi-character).
                   3218:         * We only do this if the character hash has been initialised
                   3219:         * and the string is >0 length.
                   3220:         */
                   3221:
                   3222:        res = NULL;
                   3223:        ssz = 0;
                   3224:
                   3225:        while ('\0' != *p) {
1.161     schwarze 3226:                assert((unsigned int)*p < 128);
                   3227:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3228:                        sz = r->xtab[(int)*p].sz;
                   3229:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3230:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3231:                        ssz += sz;
                   3232:                        p++;
                   3233:                        continue;
                   3234:                } else if ('\\' != *p) {
                   3235:                        res = mandoc_realloc(res, ssz + 2);
                   3236:                        res[ssz++] = *p++;
                   3237:                        continue;
                   3238:                }
                   3239:
                   3240:                /* Search for term matches. */
                   3241:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3242:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3243:                                break;
                   3244:
                   3245:                if (NULL != cp) {
                   3246:                        /*
                   3247:                         * A match has been found.
                   3248:                         * Append the match to the array and move
                   3249:                         * forward by its keysize.
                   3250:                         */
1.80      schwarze 3251:                        res = mandoc_realloc(res,
                   3252:                            ssz + cp->val.sz + 1);
1.42      schwarze 3253:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3254:                        ssz += cp->val.sz;
                   3255:                        p += (int)cp->key.sz;
                   3256:                        continue;
                   3257:                }
                   3258:
                   3259:                /*
                   3260:                 * Handle escapes carefully: we need to copy
                   3261:                 * over just the escape itself, or else we might
                   3262:                 * do replacements within the escape itself.
                   3263:                 * Make sure to pass along the bogus string.
                   3264:                 */
                   3265:                pp = p++;
                   3266:                esc = mandoc_escape(&p, NULL, NULL);
                   3267:                if (ESCAPE_ERROR == esc) {
                   3268:                        sz = strlen(pp);
                   3269:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3270:                        memcpy(res + ssz, pp, sz);
                   3271:                        break;
                   3272:                }
1.80      schwarze 3273:                /*
                   3274:                 * We bail out on bad escapes.
1.42      schwarze 3275:                 * No need to warn: we already did so when
                   3276:                 * roff_res() was called.
                   3277:                 */
                   3278:                sz = (int)(p - pp);
                   3279:                res = mandoc_realloc(res, ssz + sz + 1);
                   3280:                memcpy(res + ssz, pp, sz);
                   3281:                ssz += sz;
                   3282:        }
                   3283:
                   3284:        res[(int)ssz] = '\0';
1.149     schwarze 3285:        return res;
1.99      schwarze 3286: }
                   3287:
                   3288: int
                   3289: roff_getformat(const struct roff *r)
                   3290: {
                   3291:
1.149     schwarze 3292:        return r->format;
1.48      schwarze 3293: }
                   3294:
                   3295: /*
1.80      schwarze 3296:  * Find out whether a line is a macro line or not.
1.48      schwarze 3297:  * If it is, adjust the current position and return one; if it isn't,
                   3298:  * return zero and don't change the current position.
                   3299:  * If the control character has been set with `.cc', then let that grain
                   3300:  * precedence.
                   3301:  * This is slighly contrary to groff, where using the non-breaking
                   3302:  * control character when `cc' has been invoked will cause the
                   3303:  * non-breaking macro contents to be printed verbatim.
                   3304:  */
                   3305: int
                   3306: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3307: {
                   3308:        int             pos;
                   3309:
                   3310:        pos = *ppos;
                   3311:
                   3312:        if (0 != r->control && cp[pos] == r->control)
                   3313:                pos++;
                   3314:        else if (0 != r->control)
1.149     schwarze 3315:                return 0;
1.48      schwarze 3316:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3317:                pos += 2;
                   3318:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3319:                pos++;
                   3320:        else
1.149     schwarze 3321:                return 0;
1.48      schwarze 3322:
                   3323:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3324:                pos++;
                   3325:
                   3326:        *ppos = pos;
1.149     schwarze 3327:        return 1;
1.1       schwarze 3328: }