[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 1.187

1.187   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.186 2017/06/17 23:06:43 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:
1.187   ! schwarze   40: /* Types of definitions of macros and strings. */
        !            41: #define        ROFFDEF_USER    (1 << 1)  /* User-defined. */
        !            42: #define        ROFFDEF_PRE     (1 << 2)  /* Predefined. */
        !            43: #define        ROFFDEF_REN     (1 << 3)  /* Renamed standard macro. */
        !            44: #define        ROFFDEF_STD     (1 << 4)  /* mdoc(7) or man(7) macro. */
        !            45: #define        ROFFDEF_ANY     (ROFFDEF_USER | ROFFDEF_PRE | \
        !            46:                         ROFFDEF_REN | ROFFDEF_STD)
        !            47:
1.138     schwarze   48: /* --- data types --------------------------------------------------------- */
                     49:
1.41      schwarze   50: /*
1.42      schwarze   51:  * An incredibly-simple string buffer.
                     52:  */
1.8       schwarze   53: struct roffstr {
1.42      schwarze   54:        char            *p; /* nil-terminated buffer */
                     55:        size_t           sz; /* saved strlen(p) */
                     56: };
                     57:
                     58: /*
                     59:  * A key-value roffstr pair as part of a singly-linked list.
                     60:  */
                     61: struct roffkv {
                     62:        struct roffstr   key;
                     63:        struct roffstr   val;
                     64:        struct roffkv   *next; /* next in list */
1.8       schwarze   65: };
                     66:
1.52      schwarze   67: /*
                     68:  * A single number register as part of a singly-linked list.
                     69:  */
                     70: struct roffreg {
                     71:        struct roffstr   key;
1.53      schwarze   72:        int              val;
1.52      schwarze   73:        struct roffreg  *next;
                     74: };
                     75:
1.167     schwarze   76: /*
                     77:  * Association of request and macro names with token IDs.
                     78:  */
                     79: struct roffreq {
                     80:        enum roff_tok    tok;
                     81:        char             name[];
                     82: };
                     83:
1.1       schwarze   84: struct roff {
1.35      schwarze   85:        struct mparse   *parse; /* parse point */
1.168     schwarze   86:        struct roff_man *man; /* mdoc or man parser */
1.1       schwarze   87:        struct roffnode *last; /* leaf of stack */
1.96      schwarze   88:        int             *rstack; /* stack of inverted `ie' values */
1.167     schwarze   89:        struct ohash    *reqtab; /* request lookup table */
1.52      schwarze   90:        struct roffreg  *regtab; /* number registers */
1.42      schwarze   91:        struct roffkv   *strtab; /* user-defined strings & macros */
1.178     schwarze   92:        struct roffkv   *rentab; /* renamed strings & macros */
1.42      schwarze   93:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     94:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze   95:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   96:        struct tbl_node *first_tbl; /* first table parsed */
                     97:        struct tbl_node *last_tbl; /* last table parsed */
                     98:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze   99:        struct eqn_node *last_eqn; /* last equation parsed */
                    100:        struct eqn_node *first_eqn; /* first equation parsed */
                    101:        struct eqn_node *eqn; /* current equation being parsed */
1.102     schwarze  102:        int              eqn_inline; /* current equation is inline */
1.96      schwarze  103:        int              options; /* parse options */
                    104:        int              rstacksz; /* current size limit of rstack */
                    105:        int              rstackpos; /* position in rstack */
1.99      schwarze  106:        int              format; /* current file in mdoc or man format */
1.145     schwarze  107:        int              argc; /* number of args of the last macro */
1.96      schwarze  108:        char             control; /* control character */
1.175     schwarze  109:        char             escape; /* escape character */
1.1       schwarze  110: };
                    111:
                    112: struct roffnode {
1.166     schwarze  113:        enum roff_tok    tok; /* type of node */
1.1       schwarze  114:        struct roffnode *parent; /* up one in stack */
                    115:        int              line; /* parse line */
                    116:        int              col; /* parse col */
1.16      schwarze  117:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  118:        char            *end; /* end-rules: custom token */
                    119:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  120:        int              rule; /* current evaluation rule */
1.1       schwarze  121: };
                    122:
                    123: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.166     schwarze  124:                         enum roff_tok tok, /* tok of macro */ \
1.110     schwarze  125:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  126:                         int ln, /* parse line */ \
1.2       schwarze  127:                         int ppos, /* original pos in buffer */ \
                    128:                         int pos, /* current pos in buffer */ \
                    129:                         int *offs /* reset offset of buffer data */
1.1       schwarze  130:
                    131: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    132:
                    133: struct roffmac {
1.2       schwarze  134:        roffproc         proc; /* process new macro */
                    135:        roffproc         text; /* process as child text of macro */
                    136:        roffproc         sub; /* process as child of macro */
                    137:        int              flags;
                    138: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.1       schwarze  139: };
                    140:
1.37      schwarze  141: struct predef {
                    142:        const char      *name; /* predefined input name */
                    143:        const char      *str; /* replacement symbol */
                    144: };
                    145:
                    146: #define        PREDEF(__name, __str) \
                    147:        { (__name), (__str) },
                    148:
1.138     schwarze  149: /* --- function prototypes ------------------------------------------------ */
                    150:
1.42      schwarze  151: static void             roffnode_cleanscope(struct roff *);
                    152: static void             roffnode_pop(struct roff *);
1.166     schwarze  153: static void             roffnode_push(struct roff *, enum roff_tok,
1.42      schwarze  154:                                const char *, int, int);
1.183     schwarze  155: static enum rofferr     roff_als(ROFF_ARGS);
1.2       schwarze  156: static enum rofferr     roff_block(ROFF_ARGS);
                    157: static enum rofferr     roff_block_text(ROFF_ARGS);
                    158: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.168     schwarze  159: static enum rofferr     roff_br(ROFF_ARGS);
1.2       schwarze  160: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  161: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  162: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  163: static enum rofferr     roff_cond(ROFF_ARGS);
                    164: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    165: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  166: static enum rofferr     roff_ds(ROFF_ARGS);
1.175     schwarze  167: static enum rofferr     roff_ec(ROFF_ARGS);
                    168: static enum rofferr     roff_eo(ROFF_ARGS);
1.110     schwarze  169: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.143     schwarze  170: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.106     schwarze  171: static int              roff_evalnum(struct roff *, int,
                    172:                                const char *, int *, int *, int);
                    173: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  174:                                const char *, int *, int *, int);
1.71      schwarze  175: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  176: static void             roff_free1(struct roff *);
1.52      schwarze  177: static void             roff_freereg(struct roffreg *);
1.42      schwarze  178: static void             roff_freestr(struct roffkv *);
1.85      schwarze  179: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  180: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  181: static int              roff_getop(const char *, int *, char *);
1.53      schwarze  182: static int              roff_getregn(const struct roff *,
                    183:                                const char *, size_t);
1.145     schwarze  184: static int              roff_getregro(const struct roff *,
                    185:                                const char *name);
1.80      schwarze  186: static const char      *roff_getstrn(const struct roff *,
1.187   ! schwarze  187:                                const char *, size_t, int *);
1.143     schwarze  188: static int              roff_hasregn(const struct roff *,
                    189:                                const char *, size_t);
1.123     schwarze  190: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  191: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  192: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137     schwarze  193: static void             roff_man_alloc1(struct roff_man *);
                    194: static void             roff_man_free1(struct roff_man *);
1.172     schwarze  195: static enum rofferr     roff_manyarg(ROFF_ARGS);
1.6       schwarze  196: static enum rofferr     roff_nr(ROFF_ARGS);
1.169     schwarze  197: static enum rofferr     roff_onearg(ROFF_ARGS);
1.166     schwarze  198: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.87      schwarze  199:                                int, int);
1.177     schwarze  200: static enum rofferr     roff_parsetext(struct roff *, struct buf *,
                    201:                                int, int *);
1.178     schwarze  202: static enum rofferr     roff_renamed(ROFF_ARGS);
1.110     schwarze  203: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  204: static enum rofferr     roff_rm(ROFF_ARGS);
1.178     schwarze  205: static enum rofferr     roff_rn(ROFF_ARGS);
1.76      schwarze  206: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  207: static void             roff_setstr(struct roff *,
1.16      schwarze  208:                                const char *, const char *, int);
1.80      schwarze  209: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  210:                                size_t, const char *, size_t, int);
1.14      schwarze  211: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  212: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  213: static enum rofferr     roff_Dd(ROFF_ARGS);
1.27      schwarze  214: static enum rofferr     roff_TE(ROFF_ARGS);
                    215: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  216: static enum rofferr     roff_EQ(ROFF_ARGS);
                    217: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  218: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  219: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  220: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  221:
1.138     schwarze  222: /* --- constant data ------------------------------------------------------ */
                    223:
1.133     schwarze  224: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    225: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    226:
1.166     schwarze  227: const char *__roff_name[MAN_MAX + 1] = {
1.177     schwarze  228:        "br",           "ce",           "ft",           "ll",
1.184     schwarze  229:        "mc",           "po",           "rj",           "sp",
                    230:        "ta",           "ti",           NULL,
1.166     schwarze  231:        "ab",           "ad",           "af",           "aln",
                    232:        "als",          "am",           "am1",          "ami",
                    233:        "ami1",         "as",           "as1",          "asciify",
                    234:        "backtrace",    "bd",           "bleedat",      "blm",
                    235:         "box",         "boxa",         "bp",           "BP",
                    236:        "break",        "breakchar",    "brnl",         "brp",
1.177     schwarze  237:        "brpnl",        "c2",           "cc",
1.166     schwarze  238:        "cf",           "cflags",       "ch",           "char",
                    239:        "chop",         "class",        "close",        "CL",
                    240:        "color",        "composite",    "continue",     "cp",
                    241:        "cropat",       "cs",           "cu",           "da",
                    242:        "dch",          "Dd",           "de",           "de1",
                    243:        "defcolor",     "dei",          "dei1",         "device",
                    244:        "devicem",      "di",           "do",           "ds",
                    245:        "ds1",          "dwh",          "dt",           "ec",
                    246:        "ecr",          "ecs",          "el",           "em",
                    247:        "EN",           "eo",           "EP",           "EQ",
                    248:        "errprint",     "ev",           "evc",          "ex",
                    249:        "fallback",     "fam",          "fc",           "fchar",
                    250:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    251:        "fl",           "flig",         "fp",           "fps",
                    252:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    253:        "fzoom",        "gcolor",       "hc",           "hcode",
                    254:        "hidechar",     "hla",          "hlm",          "hpf",
                    255:        "hpfa",         "hpfcode",      "hw",           "hy",
                    256:        "hylang",       "hylen",        "hym",          "hypp",
                    257:        "hys",          "ie",           "if",           "ig",
                    258:        "index",        "it",           "itc",          "IX",
                    259:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    260:        "lc",           "lc_ctype",     "lds",          "length",
                    261:        "letadj",       "lf",           "lg",           "lhang",
                    262:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.176     schwarze  263:        "ls",           "lsm",          "lt",
1.166     schwarze  264:        "mediasize",    "minss",        "mk",           "mso",
                    265:        "na",           "ne",           "nh",           "nhychar",
                    266:        "nm",           "nn",           "nop",          "nr",
                    267:        "nrf",          "nroff",        "ns",           "nx",
                    268:        "open",         "opena",        "os",           "output",
                    269:        "padj",         "papersize",    "pc",           "pev",
                    270:        "pi",           "PI",           "pl",           "pm",
1.184     schwarze  271:        "pn",           "pnr",          "ps",
1.166     schwarze  272:        "psbb",         "pshape",       "pso",          "ptr",
                    273:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.181     schwarze  274:        "return",       "rfschar",      "rhang",
1.166     schwarze  275:        "rm",           "rn",           "rnn",          "rr",
                    276:        "rs",           "rt",           "schar",        "sentchar",
                    277:        "shc",          "shift",        "sizes",        "so",
                    278:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    279:        "sty",          "substring",    "sv",           "sy",
1.172     schwarze  280:        "T&",           "tc",           "TE",
1.173     schwarze  281:        "TH",           "tkf",          "tl",
1.166     schwarze  282:        "tm",           "tm1",          "tmc",          "tr",
                    283:        "track",        "transchar",    "trf",          "trimat",
                    284:        "trin",         "trnt",         "troff",        "TS",
                    285:        "uf",           "ul",           "unformat",     "unwatch",
                    286:        "unwatchn",     "vpt",          "vs",           "warn",
                    287:        "warnscale",    "watch",        "watchlength",  "watchn",
                    288:        "wh",           "while",        "write",        "writec",
                    289:        "writem",       "xflag",        ".",            NULL,
1.178     schwarze  290:        NULL,           "text",
1.166     schwarze  291:        "Dd",           "Dt",           "Os",           "Sh",
                    292:        "Ss",           "Pp",           "D1",           "Dl",
                    293:        "Bd",           "Ed",           "Bl",           "El",
                    294:        "It",           "Ad",           "An",           "Ap",
                    295:        "Ar",           "Cd",           "Cm",           "Dv",
                    296:        "Er",           "Ev",           "Ex",           "Fa",
                    297:        "Fd",           "Fl",           "Fn",           "Ft",
                    298:        "Ic",           "In",           "Li",           "Nd",
                    299:        "Nm",           "Op",           "Ot",           "Pa",
                    300:        "Rv",           "St",           "Va",           "Vt",
                    301:        "Xr",           "%A",           "%B",           "%D",
                    302:        "%I",           "%J",           "%N",           "%O",
                    303:        "%P",           "%R",           "%T",           "%V",
                    304:        "Ac",           "Ao",           "Aq",           "At",
                    305:        "Bc",           "Bf",           "Bo",           "Bq",
                    306:        "Bsx",          "Bx",           "Db",           "Dc",
                    307:        "Do",           "Dq",           "Ec",           "Ef",
                    308:        "Em",           "Eo",           "Fx",           "Ms",
                    309:        "No",           "Ns",           "Nx",           "Ox",
                    310:        "Pc",           "Pf",           "Po",           "Pq",
                    311:        "Qc",           "Ql",           "Qo",           "Qq",
                    312:        "Re",           "Rs",           "Sc",           "So",
                    313:        "Sq",           "Sm",           "Sx",           "Sy",
                    314:        "Tn",           "Ux",           "Xc",           "Xo",
                    315:        "Fo",           "Fc",           "Oo",           "Oc",
                    316:        "Bk",           "Ek",           "Bt",           "Hf",
                    317:        "Fr",           "Ud",           "Lb",           "Lp",
                    318:        "Lk",           "Mt",           "Brq",          "Bro",
                    319:        "Brc",          "%C",           "Es",           "En",
1.171     schwarze  320:        "Dx",           "%Q",           "%U",           "Ta",
                    321:        NULL,
1.166     schwarze  322:        "TH",           "SH",           "SS",           "TP",
                    323:        "LP",           "PP",           "P",            "IP",
                    324:        "HP",           "SM",           "SB",           "BI",
                    325:        "IB",           "BR",           "RB",           "R",
                    326:        "B",            "I",            "IR",           "RI",
1.171     schwarze  327:        "nf",           "fi",
1.166     schwarze  328:        "RE",           "RS",           "DT",           "UC",
1.169     schwarze  329:        "PD",           "AT",           "in",
1.166     schwarze  330:        "OP",           "EX",           "EE",           "UR",
1.170     schwarze  331:        "UE",           NULL
1.166     schwarze  332: };
                    333: const  char *const *roff_name = __roff_name;
                    334:
                    335: static struct roffmac   roffs[TOKEN_NONE] = {
1.168     schwarze  336:        { roff_br, NULL, NULL, 0 },  /* br */
1.177     schwarze  337:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.169     schwarze  338:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.170     schwarze  339:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.176     schwarze  340:        { roff_onearg, NULL, NULL, 0 },  /* mc */
1.184     schwarze  341:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.181     schwarze  342:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.171     schwarze  343:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.172     schwarze  344:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.173     schwarze  345:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.168     schwarze  346:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.167     schwarze  347:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    348:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    349:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    350:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.183     schwarze  351:        { roff_als, NULL, NULL, 0 },  /* als */
1.167     schwarze  352:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    353:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    354:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    355:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    356:        { roff_ds, NULL, NULL, 0 },  /* as */
                    357:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    358:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    359:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    360:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    361:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    362:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    363:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    364:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    365:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    366:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    367:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    368:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    369:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.168     schwarze  370:        { roff_br, NULL, NULL, 0 },  /* brp */
1.167     schwarze  371:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    372:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    373:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    374:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    375:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    376:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    377:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    378:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    379:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    380:        { roff_insec, NULL, NULL, 0 },  /* close */
                    381:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    382:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    383:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    384:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    385:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    386:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    387:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    388:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    389:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    390:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    391:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    392:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    393:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    394:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    395:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    396:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    397:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    398:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    400:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    401:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    402:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    403:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    404:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.175     schwarze  405:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.167     schwarze  406:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    407:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    408:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    409:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    410:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.175     schwarze  411:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.167     schwarze  412:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    413:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    414:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    415:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    416:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    417:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    419:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    420:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    421:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    422:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    425:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    429:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    430:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    431:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    432:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    438:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    439:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    440:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    442:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    444:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    445:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    446:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    447:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    448:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    449:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    450:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    451:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    452:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    453:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    454:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    455:        { roff_it, NULL, NULL, 0 },  /* it */
                    456:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    457:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    458:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    461:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    462:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    463:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    464:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    465:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    466:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    467:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    468:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    469:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    470:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    471:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    472:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    473:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    474:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    475:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    480:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    485:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    486:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    487:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    488:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    489:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    491:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    492:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    493:        { roff_insec, NULL, NULL, 0 },  /* open */
                    494:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    495:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    496:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    497:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    500:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    501:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    502:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    503:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    504:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    505:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    506:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    508:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    509:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    510:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    511:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    512:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    513:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    514:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    516:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    517:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    518:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    519:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.178     schwarze  520:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.167     schwarze  521:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    522:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    523:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    524:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    525:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    526:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    528:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    529:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    530:        { roff_so, NULL, NULL, 0 },  /* so */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    533:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    534:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    535:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    536:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    537:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    538:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    539:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    540:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    541:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.187   ! schwarze  542:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.167     schwarze  543:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    544:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    548:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    549:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    550:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    551:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    552:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    553:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    554:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    555:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    556:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    557:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    558:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    559:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    560:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    561:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    562:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    563:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    565:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    566:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    568:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    569:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    570:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    571:        { roff_insec, NULL, NULL, 0 },  /* write */
                    572:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    573:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    574:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    575:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.178     schwarze  576:        { roff_renamed, NULL, NULL, 0 },
1.167     schwarze  577:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  578: };
                    579:
1.37      schwarze  580: /* Array of injected predefined strings. */
                    581: #define        PREDEFS_MAX      38
                    582: static const struct predef predefs[PREDEFS_MAX] = {
                    583: #include "predefs.in"
                    584: };
                    585:
1.177     schwarze  586: static int      roffce_lines;  /* number of input lines to center */
                    587: static struct roff_node *roffce_node;  /* active request */
1.51      schwarze  588: static int      roffit_lines;  /* number of lines to delay */
                    589: static char    *roffit_macro;  /* nil-terminated macro line */
                    590:
1.80      schwarze  591:
1.138     schwarze  592: /* --- request table ------------------------------------------------------ */
                    593:
1.167     schwarze  594: struct ohash *
                    595: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  596: {
1.167     schwarze  597:        struct ohash    *htab;
                    598:        struct roffreq  *req;
                    599:        enum roff_tok    tok;
                    600:        size_t           sz;
                    601:        unsigned int     slot;
1.3       schwarze  602:
1.167     schwarze  603:        htab = mandoc_malloc(sizeof(*htab));
                    604:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  605:
1.167     schwarze  606:        for (tok = mintok; tok < maxtok; tok++) {
1.168     schwarze  607:                if (roff_name[tok] == NULL)
                    608:                        continue;
1.167     schwarze  609:                sz = strlen(roff_name[tok]);
                    610:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    611:                req->tok = tok;
                    612:                memcpy(req->name, roff_name[tok], sz + 1);
                    613:                slot = ohash_qlookup(htab, req->name);
                    614:                ohash_insert(htab, slot, req);
1.3       schwarze  615:        }
1.167     schwarze  616:        return htab;
1.3       schwarze  617: }
                    618:
1.167     schwarze  619: void
                    620: roffhash_free(struct ohash *htab)
1.1       schwarze  621: {
1.167     schwarze  622:        struct roffreq  *req;
                    623:        unsigned int     slot;
1.1       schwarze  624:
1.167     schwarze  625:        if (htab == NULL)
                    626:                return;
                    627:        for (req = ohash_first(htab, &slot); req != NULL;
                    628:             req = ohash_next(htab, &slot))
                    629:                free(req);
                    630:        ohash_delete(htab);
                    631:        free(htab);
                    632: }
                    633:
                    634: enum roff_tok
                    635: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    636: {
                    637:        struct roffreq  *req;
                    638:        const char      *end;
                    639:
                    640:        if (sz) {
                    641:                end = name + sz;
                    642:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    643:        } else
                    644:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    645:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  646: }
                    647:
1.138     schwarze  648: /* --- stack of request blocks -------------------------------------------- */
                    649:
1.1       schwarze  650: /*
                    651:  * Pop the current node off of the stack of roff instructions currently
                    652:  * pending.
                    653:  */
                    654: static void
                    655: roffnode_pop(struct roff *r)
                    656: {
                    657:        struct roffnode *p;
                    658:
1.2       schwarze  659:        assert(r->last);
1.80      schwarze  660:        p = r->last;
1.2       schwarze  661:
                    662:        r->last = r->last->parent;
1.16      schwarze  663:        free(p->name);
                    664:        free(p->end);
1.1       schwarze  665:        free(p);
                    666: }
                    667:
                    668: /*
                    669:  * Push a roff node onto the instruction stack.  This must later be
                    670:  * removed with roffnode_pop().
                    671:  */
1.11      schwarze  672: static void
1.166     schwarze  673: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  674:                int line, int col)
1.1       schwarze  675: {
                    676:        struct roffnode *p;
                    677:
1.11      schwarze  678:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  679:        p->tok = tok;
1.16      schwarze  680:        if (name)
                    681:                p->name = mandoc_strdup(name);
1.1       schwarze  682:        p->parent = r->last;
                    683:        p->line = line;
                    684:        p->col = col;
1.71      schwarze  685:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  686:
                    687:        r->last = p;
                    688: }
                    689:
1.138     schwarze  690: /* --- roff parser state data management ---------------------------------- */
                    691:
1.1       schwarze  692: static void
                    693: roff_free1(struct roff *r)
                    694: {
1.49      schwarze  695:        struct tbl_node *tbl;
1.32      schwarze  696:        struct eqn_node *e;
1.42      schwarze  697:        int              i;
1.27      schwarze  698:
1.49      schwarze  699:        while (NULL != (tbl = r->first_tbl)) {
                    700:                r->first_tbl = tbl->next;
                    701:                tbl_free(tbl);
1.27      schwarze  702:        }
                    703:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  704:
1.32      schwarze  705:        while (NULL != (e = r->first_eqn)) {
                    706:                r->first_eqn = e->next;
                    707:                eqn_free(e);
                    708:        }
                    709:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    710:
1.1       schwarze  711:        while (r->last)
                    712:                roffnode_pop(r);
1.27      schwarze  713:
1.96      schwarze  714:        free (r->rstack);
                    715:        r->rstack = NULL;
                    716:        r->rstacksz = 0;
                    717:        r->rstackpos = -1;
                    718:
                    719:        roff_freereg(r->regtab);
                    720:        r->regtab = NULL;
                    721:
1.42      schwarze  722:        roff_freestr(r->strtab);
1.178     schwarze  723:        roff_freestr(r->rentab);
1.42      schwarze  724:        roff_freestr(r->xmbtab);
1.178     schwarze  725:        r->strtab = r->rentab = r->xmbtab = NULL;
1.42      schwarze  726:
                    727:        if (r->xtab)
                    728:                for (i = 0; i < 128; i++)
                    729:                        free(r->xtab[i].p);
                    730:        free(r->xtab);
                    731:        r->xtab = NULL;
1.1       schwarze  732: }
                    733:
                    734: void
                    735: roff_reset(struct roff *r)
                    736: {
                    737:        roff_free1(r);
1.99      schwarze  738:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175     schwarze  739:        r->control = '\0';
                    740:        r->escape = '\\';
1.179     schwarze  741:        roffce_lines = 0;
                    742:        roffce_node = NULL;
                    743:        roffit_lines = 0;
                    744:        roffit_macro = NULL;
1.1       schwarze  745: }
                    746:
                    747: void
                    748: roff_free(struct roff *r)
                    749: {
                    750:        roff_free1(r);
1.167     schwarze  751:        roffhash_free(r->reqtab);
1.1       schwarze  752:        free(r);
                    753: }
                    754:
                    755: struct roff *
1.151     schwarze  756: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  757: {
                    758:        struct roff     *r;
                    759:
1.11      schwarze  760:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  761:        r->parse = parse;
1.167     schwarze  762:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72      schwarze  763:        r->options = options;
1.99      schwarze  764:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  765:        r->rstackpos = -1;
1.175     schwarze  766:        r->escape = '\\';
1.149     schwarze  767:        return r;
1.137     schwarze  768: }
                    769:
1.138     schwarze  770: /* --- syntax tree state data management ---------------------------------- */
                    771:
1.137     schwarze  772: static void
                    773: roff_man_free1(struct roff_man *man)
                    774: {
                    775:
1.138     schwarze  776:        if (man->first != NULL)
                    777:                roff_node_delete(man, man->first);
1.137     schwarze  778:        free(man->meta.msec);
                    779:        free(man->meta.vol);
                    780:        free(man->meta.os);
                    781:        free(man->meta.arch);
                    782:        free(man->meta.title);
                    783:        free(man->meta.name);
                    784:        free(man->meta.date);
                    785: }
                    786:
                    787: static void
                    788: roff_man_alloc1(struct roff_man *man)
                    789: {
                    790:
                    791:        memset(&man->meta, 0, sizeof(man->meta));
                    792:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    793:        man->first->type = ROFFT_ROOT;
                    794:        man->last = man->first;
                    795:        man->last_es = NULL;
                    796:        man->flags = 0;
                    797:        man->macroset = MACROSET_NONE;
                    798:        man->lastsec = man->lastnamed = SEC_NONE;
                    799:        man->next = ROFF_NEXT_CHILD;
                    800: }
                    801:
                    802: void
                    803: roff_man_reset(struct roff_man *man)
                    804: {
                    805:
                    806:        roff_man_free1(man);
                    807:        roff_man_alloc1(man);
                    808: }
                    809:
                    810: void
                    811: roff_man_free(struct roff_man *man)
                    812: {
                    813:
                    814:        roff_man_free1(man);
                    815:        free(man);
                    816: }
                    817:
                    818: struct roff_man *
                    819: roff_man_alloc(struct roff *roff, struct mparse *parse,
                    820:        const char *defos, int quick)
                    821: {
                    822:        struct roff_man *man;
                    823:
                    824:        man = mandoc_calloc(1, sizeof(*man));
                    825:        man->parse = parse;
                    826:        man->roff = roff;
                    827:        man->defos = defos;
                    828:        man->quick = quick;
                    829:        roff_man_alloc1(man);
1.168     schwarze  830:        roff->man = man;
1.149     schwarze  831:        return man;
1.1       schwarze  832: }
                    833:
1.138     schwarze  834: /* --- syntax tree handling ----------------------------------------------- */
                    835:
                    836: struct roff_node *
                    837: roff_node_alloc(struct roff_man *man, int line, int pos,
                    838:        enum roff_type type, int tok)
                    839: {
                    840:        struct roff_node        *n;
                    841:
                    842:        n = mandoc_calloc(1, sizeof(*n));
                    843:        n->line = line;
                    844:        n->pos = pos;
                    845:        n->tok = tok;
                    846:        n->type = type;
                    847:        n->sec = man->lastsec;
                    848:
                    849:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  850:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  851:        else
1.157     schwarze  852:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  853:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  854:                n->flags |= NODE_LINE;
1.138     schwarze  855:        man->flags &= ~MDOC_NEWLINE;
                    856:
1.149     schwarze  857:        return n;
1.138     schwarze  858: }
                    859:
                    860: void
                    861: roff_node_append(struct roff_man *man, struct roff_node *n)
                    862: {
                    863:
                    864:        switch (man->next) {
                    865:        case ROFF_NEXT_SIBLING:
1.153     schwarze  866:                if (man->last->next != NULL) {
                    867:                        n->next = man->last->next;
                    868:                        man->last->next->prev = n;
                    869:                } else
                    870:                        man->last->parent->last = n;
1.138     schwarze  871:                man->last->next = n;
                    872:                n->prev = man->last;
                    873:                n->parent = man->last->parent;
                    874:                break;
                    875:        case ROFF_NEXT_CHILD:
1.159     schwarze  876:                if (man->last->child != NULL) {
                    877:                        n->next = man->last->child;
                    878:                        man->last->child->prev = n;
                    879:                } else
                    880:                        man->last->last = n;
1.138     schwarze  881:                man->last->child = n;
                    882:                n->parent = man->last;
                    883:                break;
                    884:        default:
                    885:                abort();
                    886:        }
1.154     schwarze  887:        man->last = n;
1.138     schwarze  888:
                    889:        switch (n->type) {
                    890:        case ROFFT_HEAD:
                    891:                n->parent->head = n;
                    892:                break;
                    893:        case ROFFT_BODY:
1.154     schwarze  894:                if (n->end != ENDBODY_NOT)
                    895:                        return;
1.138     schwarze  896:                n->parent->body = n;
                    897:                break;
                    898:        case ROFFT_TAIL:
                    899:                n->parent->tail = n;
                    900:                break;
                    901:        default:
1.154     schwarze  902:                return;
1.138     schwarze  903:        }
1.154     schwarze  904:
                    905:        /*
                    906:         * Copy over the normalised-data pointer of our parent.  Not
                    907:         * everybody has one, but copying a null pointer is fine.
                    908:         */
                    909:
                    910:        n->norm = n->parent->norm;
                    911:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  912: }
                    913:
1.139     schwarze  914: void
                    915: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    916: {
                    917:        struct roff_node        *n;
                    918:
                    919:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    920:        n->string = roff_strdup(man->roff, word);
                    921:        roff_node_append(man, n);
1.158     schwarze  922:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  923:        man->next = ROFF_NEXT_SIBLING;
                    924: }
                    925:
                    926: void
                    927: roff_word_append(struct roff_man *man, const char *word)
                    928: {
                    929:        struct roff_node        *n;
                    930:        char                    *addstr, *newstr;
                    931:
                    932:        n = man->last;
                    933:        addstr = roff_strdup(man->roff, word);
                    934:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    935:        free(addstr);
                    936:        free(n->string);
                    937:        n->string = newstr;
                    938:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  939: }
                    940:
                    941: void
                    942: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    943: {
                    944:        struct roff_node        *n;
                    945:
                    946:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    947:        roff_node_append(man, n);
                    948:        man->next = ROFF_NEXT_CHILD;
                    949: }
                    950:
                    951: struct roff_node *
                    952: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    953: {
                    954:        struct roff_node        *n;
                    955:
                    956:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    957:        roff_node_append(man, n);
                    958:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  959:        return n;
1.139     schwarze  960: }
                    961:
1.138     schwarze  962: struct roff_node *
                    963: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    964: {
                    965:        struct roff_node        *n;
                    966:
                    967:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    968:        roff_node_append(man, n);
                    969:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  970:        return n;
1.138     schwarze  971: }
                    972:
                    973: struct roff_node *
                    974: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    975: {
                    976:        struct roff_node        *n;
                    977:
                    978:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    979:        roff_node_append(man, n);
                    980:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  981:        return n;
1.139     schwarze  982: }
                    983:
                    984: void
                    985: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
                    986: {
                    987:        struct roff_node        *n;
                    988:
                    989:        n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
                    990:        n->eqn = eqn;
                    991:        if (eqn->ln > man->last->line)
1.157     schwarze  992:                n->flags |= NODE_LINE;
1.139     schwarze  993:        roff_node_append(man, n);
                    994:        man->next = ROFF_NEXT_SIBLING;
                    995: }
                    996:
                    997: void
                    998: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
                    999: {
                   1000:        struct roff_node        *n;
                   1001:
                   1002:        if (man->macroset == MACROSET_MAN)
1.174     schwarze 1003:                man_breakscope(man, ROFF_TS);
1.139     schwarze 1004:        n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                   1005:        n->span = tbl;
                   1006:        roff_node_append(man, n);
1.158     schwarze 1007:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze 1008:        man->next = ROFF_NEXT_SIBLING;
1.138     schwarze 1009: }
                   1010:
                   1011: void
                   1012: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1013: {
                   1014:
                   1015:        /* Adjust siblings. */
                   1016:
                   1017:        if (n->prev)
                   1018:                n->prev->next = n->next;
                   1019:        if (n->next)
                   1020:                n->next->prev = n->prev;
                   1021:
                   1022:        /* Adjust parent. */
                   1023:
                   1024:        if (n->parent != NULL) {
                   1025:                if (n->parent->child == n)
                   1026:                        n->parent->child = n->next;
                   1027:                if (n->parent->last == n)
                   1028:                        n->parent->last = n->prev;
                   1029:        }
                   1030:
                   1031:        /* Adjust parse point. */
                   1032:
                   1033:        if (man == NULL)
                   1034:                return;
                   1035:        if (man->last == n) {
                   1036:                if (n->prev == NULL) {
                   1037:                        man->last = n->parent;
                   1038:                        man->next = ROFF_NEXT_CHILD;
                   1039:                } else {
                   1040:                        man->last = n->prev;
                   1041:                        man->next = ROFF_NEXT_SIBLING;
                   1042:                }
                   1043:        }
                   1044:        if (man->first == n)
                   1045:                man->first = NULL;
                   1046: }
                   1047:
                   1048: void
                   1049: roff_node_free(struct roff_node *n)
                   1050: {
                   1051:
                   1052:        if (n->args != NULL)
                   1053:                mdoc_argv_free(n->args);
                   1054:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1055:                free(n->norm);
                   1056:        free(n->string);
                   1057:        free(n);
                   1058: }
                   1059:
                   1060: void
                   1061: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1062: {
                   1063:
                   1064:        while (n->child != NULL)
                   1065:                roff_node_delete(man, n->child);
                   1066:        roff_node_unlink(man, n);
                   1067:        roff_node_free(n);
1.141     schwarze 1068: }
                   1069:
                   1070: void
                   1071: deroff(char **dest, const struct roff_node *n)
                   1072: {
                   1073:        char    *cp;
                   1074:        size_t   sz;
                   1075:
                   1076:        if (n->type != ROFFT_TEXT) {
                   1077:                for (n = n->child; n != NULL; n = n->next)
                   1078:                        deroff(dest, n);
                   1079:                return;
                   1080:        }
                   1081:
1.160     schwarze 1082:        /* Skip leading whitespace. */
1.141     schwarze 1083:
1.160     schwarze 1084:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1085:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1086:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1087:                        cp++;
1.160     schwarze 1088:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1089:                        break;
                   1090:        }
                   1091:
1.161     schwarze 1092:        /* Skip trailing backslash. */
                   1093:
                   1094:        sz = strlen(cp);
1.162     schwarze 1095:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1096:                sz--;
                   1097:
1.141     schwarze 1098:        /* Skip trailing whitespace. */
                   1099:
1.161     schwarze 1100:        for (; sz; sz--)
1.141     schwarze 1101:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1102:                        break;
                   1103:
                   1104:        /* Skip empty strings. */
                   1105:
                   1106:        if (sz == 0)
                   1107:                return;
                   1108:
                   1109:        if (*dest == NULL) {
                   1110:                *dest = mandoc_strndup(cp, sz);
                   1111:                return;
                   1112:        }
                   1113:
                   1114:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1115:        free(*dest);
                   1116:        *dest = cp;
1.138     schwarze 1117: }
                   1118:
                   1119: /* --- main functions of the roff parser ---------------------------------- */
                   1120:
1.8       schwarze 1121: /*
1.79      schwarze 1122:  * In the current line, expand escape sequences that tend to get
                   1123:  * used in numerical expressions and conditional requests.
                   1124:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1125:  */
1.45      schwarze 1126: static enum rofferr
1.110     schwarze 1127: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1128: {
1.81      schwarze 1129:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze 1130:        const char      *start; /* start of the string to process */
1.82      schwarze 1131:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze 1132:        const char      *stnam; /* start of the name, after "[(*" */
                   1133:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1134:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1135:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1136:        size_t           maxl;  /* expected length of the escape name */
                   1137:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1138:        enum mandoc_esc  esc;   /* type of the escape sequence */
1.185     schwarze 1139:        enum mdoc_os     os_e;  /* kind of RCS id seen */
1.109     schwarze 1140:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1141:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1142:        int              npos;  /* position in numeric expression */
1.91      schwarze 1143:        int              arg_complete; /* argument not interrupted by eol */
1.175     schwarze 1144:        int              done;  /* no more input available */
1.187   ! schwarze 1145:        int              deftype; /* type of definition to paste */
1.79      schwarze 1146:        char             term;  /* character terminating the escape */
1.8       schwarze 1147:
1.175     schwarze 1148:        /* Search forward for comments. */
                   1149:
                   1150:        done = 0;
                   1151:        start = buf->buf + pos;
                   1152:        for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
                   1153:                if (stesc[0] != r->escape || stesc[1] == '\0')
                   1154:                        continue;
                   1155:                stesc++;
                   1156:                if (*stesc != '"' && *stesc != '#')
                   1157:                        continue;
1.185     schwarze 1158:
                   1159:                /* Comment found, look for RCS id. */
                   1160:
                   1161:                if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
                   1162:                        os_e = MDOC_OS_OPENBSD;
                   1163:                        cp += 8;
                   1164:                } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
                   1165:                        os_e = MDOC_OS_NETBSD;
                   1166:                        cp += 7;
                   1167:                }
                   1168:                if (cp != NULL &&
                   1169:                    isalnum((unsigned char)*cp) == 0 &&
1.186     schwarze 1170:                    strchr(cp, '$') != NULL) {
                   1171:                        if (r->man->meta.rcsids & (1 << os_e))
                   1172:                                mandoc_msg(MANDOCERR_RCS_REP, r->parse,
                   1173:                                    ln, stesc + 1 - buf->buf, stesc + 1);
1.185     schwarze 1174:                        r->man->meta.rcsids |= 1 << os_e;
1.186     schwarze 1175:                }
1.185     schwarze 1176:
                   1177:                /* Handle trailing whitespace. */
                   1178:
1.175     schwarze 1179:                cp = strchr(stesc--, '\0') - 1;
                   1180:                if (*cp == '\n') {
                   1181:                        done = 1;
                   1182:                        cp--;
                   1183:                }
                   1184:                if (*cp == ' ' || *cp == '\t')
                   1185:                        mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
                   1186:                            ln, cp - buf->buf, NULL);
                   1187:                while (stesc > start && stesc[-1] == ' ')
                   1188:                        stesc--;
                   1189:                *stesc = '\0';
                   1190:                break;
                   1191:        }
                   1192:        if (stesc == start)
                   1193:                return ROFF_CONT;
                   1194:        stesc--;
                   1195:
                   1196:        /* Notice the end of the input. */
                   1197:
                   1198:        if (*stesc == '\n') {
                   1199:                *stesc-- = '\0';
                   1200:                done = 1;
                   1201:        }
                   1202:
1.43      schwarze 1203:        expand_count = 0;
1.175     schwarze 1204:        while (stesc >= start) {
1.43      schwarze 1205:
1.78      schwarze 1206:                /* Search backwards for the next backslash. */
                   1207:
1.175     schwarze 1208:                if (*stesc != r->escape) {
                   1209:                        if (*stesc == '\\') {
                   1210:                                *stesc = '\0';
                   1211:                                buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
                   1212:                                    buf->buf, stesc + 1) + 1;
                   1213:                                start = nbuf + pos;
                   1214:                                stesc = nbuf + (stesc - buf->buf);
                   1215:                                free(buf->buf);
                   1216:                                buf->buf = nbuf;
                   1217:                        }
                   1218:                        stesc--;
1.78      schwarze 1219:                        continue;
1.175     schwarze 1220:                }
1.78      schwarze 1221:
                   1222:                /* If it is escaped, skip it. */
                   1223:
                   1224:                for (cp = stesc - 1; cp >= start; cp--)
1.175     schwarze 1225:                        if (*cp != r->escape)
1.78      schwarze 1226:                                break;
                   1227:
1.110     schwarze 1228:                if ((stesc - cp) % 2 == 0) {
1.175     schwarze 1229:                        while (stesc > cp)
                   1230:                                *stesc-- = '\\';
1.78      schwarze 1231:                        continue;
1.175     schwarze 1232:                } else if (stesc[1] != '\0') {
                   1233:                        *stesc = '\\';
                   1234:                } else {
                   1235:                        *stesc-- = '\0';
                   1236:                        if (done)
                   1237:                                continue;
                   1238:                        else
                   1239:                                return ROFF_APPEND;
1.78      schwarze 1240:                }
1.23      schwarze 1241:
1.79      schwarze 1242:                /* Decide whether to expand or to check only. */
1.23      schwarze 1243:
1.79      schwarze 1244:                term = '\0';
1.78      schwarze 1245:                cp = stesc + 1;
1.53      schwarze 1246:                switch (*cp) {
1.80      schwarze 1247:                case '*':
1.53      schwarze 1248:                        res = NULL;
                   1249:                        break;
1.80      schwarze 1250:                case 'B':
                   1251:                case 'w':
1.79      schwarze 1252:                        term = cp[1];
                   1253:                        /* FALLTHROUGH */
1.80      schwarze 1254:                case 'n':
1.53      schwarze 1255:                        res = ubuf;
                   1256:                        break;
                   1257:                default:
1.109     schwarze 1258:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1259:                        if (esc == ESCAPE_ERROR ||
                   1260:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1261:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1262:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1263:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1264:                                    "%.*s", (int)(cp - stesc), stesc);
1.175     schwarze 1265:                        stesc--;
1.78      schwarze 1266:                        continue;
1.42      schwarze 1267:                }
                   1268:
1.78      schwarze 1269:                if (EXPAND_LIMIT < ++expand_count) {
                   1270:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1271:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1272:                        return ROFF_IGN;
1.78      schwarze 1273:                }
1.23      schwarze 1274:
                   1275:                /*
                   1276:                 * The third character decides the length
1.53      schwarze 1277:                 * of the name of the string or register.
1.23      schwarze 1278:                 * Save a pointer to the name.
                   1279:                 */
                   1280:
1.110     schwarze 1281:                if (term == '\0') {
1.79      schwarze 1282:                        switch (*++cp) {
1.80      schwarze 1283:                        case '\0':
1.79      schwarze 1284:                                maxl = 0;
                   1285:                                break;
1.80      schwarze 1286:                        case '(':
1.79      schwarze 1287:                                cp++;
                   1288:                                maxl = 2;
                   1289:                                break;
1.80      schwarze 1290:                        case '[':
1.79      schwarze 1291:                                cp++;
                   1292:                                term = ']';
                   1293:                                maxl = 0;
                   1294:                                break;
                   1295:                        default:
                   1296:                                maxl = 1;
                   1297:                                break;
                   1298:                        }
                   1299:                } else {
                   1300:                        cp += 2;
1.8       schwarze 1301:                        maxl = 0;
                   1302:                }
1.23      schwarze 1303:                stnam = cp;
1.8       schwarze 1304:
1.23      schwarze 1305:                /* Advance to the end of the name. */
1.8       schwarze 1306:
1.125     schwarze 1307:                naml = 0;
1.91      schwarze 1308:                arg_complete = 1;
1.125     schwarze 1309:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1310:                        if (*cp == '\0') {
1.92      schwarze 1311:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1312:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1313:                                arg_complete = 0;
1.79      schwarze 1314:                                break;
1.42      schwarze 1315:                        }
1.110     schwarze 1316:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1317:                                cp++;
1.125     schwarze 1318:                                break;
                   1319:                        }
                   1320:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1321:                                naml++;
                   1322:                                continue;
                   1323:                        }
                   1324:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1325:                        case ESCAPE_SPECIAL:
                   1326:                        case ESCAPE_UNICODE:
                   1327:                        case ESCAPE_NUMBERED:
                   1328:                        case ESCAPE_OVERSTRIKE:
                   1329:                                naml++;
                   1330:                                break;
                   1331:                        default:
1.8       schwarze 1332:                                break;
1.79      schwarze 1333:                        }
1.8       schwarze 1334:                }
                   1335:
1.23      schwarze 1336:                /*
                   1337:                 * Retrieve the replacement string; if it is
                   1338:                 * undefined, resume searching for escapes.
                   1339:                 */
                   1340:
1.79      schwarze 1341:                switch (stesc[1]) {
1.80      schwarze 1342:                case '*':
1.187   ! schwarze 1343:                        if (arg_complete) {
        !          1344:                                deftype = ROFFDEF_USER | ROFFDEF_PRE;
        !          1345:                                res = roff_getstrn(r, stnam, naml, &deftype);
        !          1346:                        }
1.79      schwarze 1347:                        break;
1.80      schwarze 1348:                case 'B':
1.79      schwarze 1349:                        npos = 0;
1.91      schwarze 1350:                        ubuf[0] = arg_complete &&
1.133     schwarze 1351:                            roff_evalnum(r, ln, stnam, &npos,
                   1352:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1353:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1354:                        ubuf[1] = '\0';
                   1355:                        break;
1.80      schwarze 1356:                case 'n':
1.91      schwarze 1357:                        if (arg_complete)
                   1358:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1359:                                    roff_getregn(r, stnam, naml));
                   1360:                        else
                   1361:                                ubuf[0] = '\0';
1.79      schwarze 1362:                        break;
1.80      schwarze 1363:                case 'w':
1.91      schwarze 1364:                        /* use even incomplete args */
1.81      schwarze 1365:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1366:                            24 * (int)naml);
                   1367:                        break;
                   1368:                }
1.8       schwarze 1369:
1.110     schwarze 1370:                if (res == NULL) {
1.92      schwarze 1371:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1372:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1373:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1374:                        res = "";
1.118     schwarze 1375:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1376:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1377:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1378:                        return ROFF_IGN;
1.8       schwarze 1379:                }
                   1380:
1.23      schwarze 1381:                /* Replace the escape sequence by the string. */
                   1382:
1.82      schwarze 1383:                *stesc = '\0';
1.110     schwarze 1384:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1385:                    buf->buf, res, cp) + 1;
1.8       schwarze 1386:
1.78      schwarze 1387:                /* Prepare for the next replacement. */
1.8       schwarze 1388:
1.78      schwarze 1389:                start = nbuf + pos;
1.110     schwarze 1390:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1391:                free(buf->buf);
                   1392:                buf->buf = nbuf;
1.42      schwarze 1393:        }
1.149     schwarze 1394:        return ROFF_CONT;
1.42      schwarze 1395: }
                   1396:
                   1397: /*
1.147     schwarze 1398:  * Process text streams.
1.42      schwarze 1399:  */
                   1400: static enum rofferr
1.177     schwarze 1401: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42      schwarze 1402: {
                   1403:        size_t           sz;
                   1404:        const char      *start;
1.51      schwarze 1405:        char            *p;
                   1406:        int              isz;
1.42      schwarze 1407:        enum mandoc_esc  esc;
                   1408:
1.147     schwarze 1409:        /* Spring the input line trap. */
                   1410:
                   1411:        if (roffit_lines == 1) {
                   1412:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1413:                free(buf->buf);
                   1414:                buf->buf = p;
                   1415:                buf->sz = isz + 1;
                   1416:                *offs = 0;
                   1417:                free(roffit_macro);
                   1418:                roffit_lines = 0;
1.149     schwarze 1419:                return ROFF_REPARSE;
1.147     schwarze 1420:        } else if (roffit_lines > 1)
                   1421:                --roffit_lines;
                   1422:
1.177     schwarze 1423:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1424:                if (roffce_lines < 1) {
                   1425:                        r->man->last = roffce_node;
                   1426:                        r->man->next = ROFF_NEXT_SIBLING;
                   1427:                        roffce_lines = 0;
                   1428:                        roffce_node = NULL;
                   1429:                } else
                   1430:                        roffce_lines--;
                   1431:        }
                   1432:
1.147     schwarze 1433:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1434:
1.110     schwarze 1435:        start = p = buf->buf + pos;
1.42      schwarze 1436:
1.110     schwarze 1437:        while (*p != '\0') {
1.42      schwarze 1438:                sz = strcspn(p, "-\\");
                   1439:                p += sz;
                   1440:
1.110     schwarze 1441:                if (*p == '\0')
1.42      schwarze 1442:                        break;
                   1443:
1.110     schwarze 1444:                if (*p == '\\') {
1.42      schwarze 1445:                        /* Skip over escapes. */
                   1446:                        p++;
1.62      schwarze 1447:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1448:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1449:                                break;
1.136     schwarze 1450:                        while (*p == '-')
                   1451:                                p++;
1.42      schwarze 1452:                        continue;
                   1453:                } else if (p == start) {
                   1454:                        p++;
                   1455:                        continue;
                   1456:                }
                   1457:
1.44      schwarze 1458:                if (isalpha((unsigned char)p[-1]) &&
                   1459:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1460:                        *p = ASCII_HYPH;
                   1461:                p++;
1.8       schwarze 1462:        }
1.149     schwarze 1463:        return ROFF_CONT;
1.8       schwarze 1464: }
                   1465:
1.1       schwarze 1466: enum rofferr
1.110     schwarze 1467: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1468: {
1.166     schwarze 1469:        enum roff_tok    t;
1.27      schwarze 1470:        enum rofferr     e;
1.110     schwarze 1471:        int              pos;   /* parse point */
1.115     schwarze 1472:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1473:        int              ppos;  /* original offset in buf->buf */
                   1474:        int              ctl;   /* macro line (boolean) */
                   1475:
                   1476:        ppos = pos = *offs;
1.1       schwarze 1477:
1.102     schwarze 1478:        /* Handle in-line equation delimiters. */
                   1479:
1.108     schwarze 1480:        if (r->tbl == NULL &&
                   1481:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1482:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1483:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1484:                if (e == ROFF_REPARSE)
1.149     schwarze 1485:                        return e;
1.102     schwarze 1486:                assert(e == ROFF_CONT);
                   1487:        }
                   1488:
                   1489:        /* Expand some escape sequences. */
1.8       schwarze 1490:
1.110     schwarze 1491:        e = roff_res(r, buf, ln, pos);
1.175     schwarze 1492:        if (e == ROFF_IGN || e == ROFF_APPEND)
1.149     schwarze 1493:                return e;
1.110     schwarze 1494:        assert(e == ROFF_CONT);
1.8       schwarze 1495:
1.110     schwarze 1496:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1497:
1.8       schwarze 1498:        /*
1.2       schwarze 1499:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1500:         * text through the macro's filter.
                   1501:         * Equations process all content themselves.
                   1502:         * Tables process almost all content themselves, but we want
                   1503:         * to warn about macros before passing it there.
1.2       schwarze 1504:         */
                   1505:
1.124     schwarze 1506:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1507:                t = r->last->tok;
1.110     schwarze 1508:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1509:                if (e == ROFF_IGN)
1.149     schwarze 1510:                        return e;
1.166     schwarze 1511:                assert(e == ROFF_CONT);
1.54      schwarze 1512:        }
1.124     schwarze 1513:        if (r->eqn != NULL)
1.149     schwarze 1514:                return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.124     schwarze 1515:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.149     schwarze 1516:                return tbl_read(r->tbl, ln, buf->buf, ppos);
1.124     schwarze 1517:        if ( ! ctl)
1.177     schwarze 1518:                return roff_parsetext(r, buf, pos, offs);
1.100     schwarze 1519:
                   1520:        /* Skip empty request lines. */
                   1521:
1.110     schwarze 1522:        if (buf->buf[pos] == '"') {
1.100     schwarze 1523:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1524:                    ln, pos, NULL);
1.149     schwarze 1525:                return ROFF_IGN;
1.110     schwarze 1526:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1527:                return ROFF_IGN;
1.2       schwarze 1528:
                   1529:        /*
                   1530:         * If a scope is open, go to the child handler for that macro,
                   1531:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1532:         * Don't do so if an equation is open.
1.2       schwarze 1533:         */
                   1534:
                   1535:        if (r->last) {
1.1       schwarze 1536:                t = r->last->tok;
1.149     schwarze 1537:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1538:        }
                   1539:
1.115     schwarze 1540:        /* No scope is open.  This is a new request or macro. */
                   1541:
                   1542:        spos = pos;
                   1543:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1544:
                   1545:        /* Tables ignore most macros. */
                   1546:
1.180     schwarze 1547:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181     schwarze 1548:            t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.115     schwarze 1549:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1550:                    ln, pos, buf->buf + spos);
1.180     schwarze 1551:                if (t != TOKEN_NONE)
1.149     schwarze 1552:                        return ROFF_IGN;
1.129     schwarze 1553:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1554:                        pos++;
1.163     schwarze 1555:                while (buf->buf[pos] == ' ')
1.129     schwarze 1556:                        pos++;
1.149     schwarze 1557:                return tbl_read(r->tbl, ln, buf->buf, pos);
1.115     schwarze 1558:        }
                   1559:
1.177     schwarze 1560:        /* For now, let high level macros abort .ce mode. */
                   1561:
                   1562:        if (ctl && roffce_node != NULL &&
                   1563:            (t == TOKEN_NONE || t == ROFF_EQ || t == ROFF_TS)) {
                   1564:                r->man->last = roffce_node;
                   1565:                r->man->next = ROFF_NEXT_SIBLING;
                   1566:                roffce_lines = 0;
                   1567:                roffce_node = NULL;
                   1568:        }
                   1569:
1.2       schwarze 1570:        /*
1.115     schwarze 1571:         * This is neither a roff request nor a user-defined macro.
                   1572:         * Let the standard macro set parsers handle it.
1.2       schwarze 1573:         */
                   1574:
1.166     schwarze 1575:        if (t == TOKEN_NONE)
1.149     schwarze 1576:                return ROFF_CONT;
1.115     schwarze 1577:
                   1578:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1579:
1.168     schwarze 1580:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2       schwarze 1581: }
                   1582:
1.27      schwarze 1583: void
1.2       schwarze 1584: roff_endparse(struct roff *r)
                   1585: {
1.1       schwarze 1586:
1.27      schwarze 1587:        if (r->last)
1.94      schwarze 1588:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1589:                    r->last->line, r->last->col,
1.167     schwarze 1590:                    roff_name[r->last->tok]);
1.27      schwarze 1591:
1.32      schwarze 1592:        if (r->eqn) {
1.94      schwarze 1593:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1594:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41      schwarze 1595:                eqn_end(&r->eqn);
1.32      schwarze 1596:        }
                   1597:
1.27      schwarze 1598:        if (r->tbl) {
1.94      schwarze 1599:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1600:                    r->tbl->line, r->tbl->pos, "TS");
1.41      schwarze 1601:                tbl_end(&r->tbl);
1.27      schwarze 1602:        }
1.1       schwarze 1603: }
                   1604:
                   1605: /*
                   1606:  * Parse a roff node's type from the input buffer.  This must be in the
                   1607:  * form of ".foo xxx" in the usual way.
                   1608:  */
1.166     schwarze 1609: static enum roff_tok
1.87      schwarze 1610: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1611: {
1.87      schwarze 1612:        char            *cp;
1.16      schwarze 1613:        const char      *mac;
                   1614:        size_t           maclen;
1.187   ! schwarze 1615:        int              deftype;
1.166     schwarze 1616:        enum roff_tok    t;
1.1       schwarze 1617:
1.87      schwarze 1618:        cp = buf + *pos;
                   1619:
                   1620:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1621:                return TOKEN_NONE;
1.1       schwarze 1622:
1.87      schwarze 1623:        mac = cp;
                   1624:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1625:
1.187   ! schwarze 1626:        deftype = ROFFDEF_USER | ROFFDEF_REN;
        !          1627:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
        !          1628:        switch (deftype) {
        !          1629:        case ROFFDEF_USER:
        !          1630:                t = ROFF_USERDEF;
        !          1631:                break;
        !          1632:        case ROFFDEF_REN:
        !          1633:                t = ROFF_RENAMED;
        !          1634:                break;
        !          1635:        default:
        !          1636:                t = roffhash_find(r->reqtab, mac, maclen);
        !          1637:                break;
        !          1638:        }
1.166     schwarze 1639:        if (t != TOKEN_NONE)
1.87      schwarze 1640:                *pos = cp - buf;
1.149     schwarze 1641:        return t;
1.1       schwarze 1642: }
                   1643:
1.138     schwarze 1644: /* --- handling of request blocks ----------------------------------------- */
                   1645:
1.1       schwarze 1646: static enum rofferr
1.2       schwarze 1647: roff_cblock(ROFF_ARGS)
1.1       schwarze 1648: {
                   1649:
1.2       schwarze 1650:        /*
                   1651:         * A block-close `..' should only be invoked as a child of an
                   1652:         * ignore macro, otherwise raise a warning and just ignore it.
                   1653:         */
                   1654:
1.110     schwarze 1655:        if (r->last == NULL) {
1.94      schwarze 1656:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1657:                    ln, ppos, "..");
1.149     schwarze 1658:                return ROFF_IGN;
1.2       schwarze 1659:        }
1.1       schwarze 1660:
1.2       schwarze 1661:        switch (r->last->tok) {
1.80      schwarze 1662:        case ROFF_am:
1.93      schwarze 1663:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1664:        case ROFF_ami:
                   1665:        case ROFF_de:
1.23      schwarze 1666:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1667:        case ROFF_dei:
                   1668:        case ROFF_ig:
1.2       schwarze 1669:                break;
                   1670:        default:
1.94      schwarze 1671:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1672:                    ln, ppos, "..");
1.149     schwarze 1673:                return ROFF_IGN;
1.2       schwarze 1674:        }
                   1675:
1.110     schwarze 1676:        if (buf->buf[pos] != '\0')
1.90      schwarze 1677:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1678:                    ".. %s", buf->buf + pos);
1.2       schwarze 1679:
                   1680:        roffnode_pop(r);
                   1681:        roffnode_cleanscope(r);
1.149     schwarze 1682:        return ROFF_IGN;
1.2       schwarze 1683:
                   1684: }
1.1       schwarze 1685:
1.2       schwarze 1686: static void
                   1687: roffnode_cleanscope(struct roff *r)
                   1688: {
1.1       schwarze 1689:
1.2       schwarze 1690:        while (r->last) {
1.46      schwarze 1691:                if (--r->last->endspan != 0)
1.2       schwarze 1692:                        break;
                   1693:                roffnode_pop(r);
                   1694:        }
                   1695: }
1.1       schwarze 1696:
1.68      schwarze 1697: static void
                   1698: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1699: {
1.1       schwarze 1700:
1.2       schwarze 1701:        if (NULL == r->last) {
1.94      schwarze 1702:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1703:                    ln, ppos, "\\}");
1.68      schwarze 1704:                return;
1.2       schwarze 1705:        }
1.1       schwarze 1706:
1.2       schwarze 1707:        switch (r->last->tok) {
1.80      schwarze 1708:        case ROFF_el:
                   1709:        case ROFF_ie:
                   1710:        case ROFF_if:
1.2       schwarze 1711:                break;
                   1712:        default:
1.94      schwarze 1713:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1714:                    ln, ppos, "\\}");
1.68      schwarze 1715:                return;
1.2       schwarze 1716:        }
1.1       schwarze 1717:
1.2       schwarze 1718:        if (r->last->endspan > -1) {
1.94      schwarze 1719:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1720:                    ln, ppos, "\\}");
1.68      schwarze 1721:                return;
1.2       schwarze 1722:        }
                   1723:
                   1724:        roffnode_pop(r);
                   1725:        roffnode_cleanscope(r);
1.68      schwarze 1726:        return;
1.1       schwarze 1727: }
                   1728:
                   1729: static enum rofferr
1.2       schwarze 1730: roff_block(ROFF_ARGS)
1.1       schwarze 1731: {
1.187   ! schwarze 1732:        const char      *name, *value;
        !          1733:        char            *call, *cp, *iname, *rname;
        !          1734:        size_t           csz, namesz, rsz;
        !          1735:        int              deftype;
1.16      schwarze 1736:
1.93      schwarze 1737:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1738:
1.110     schwarze 1739:        if (tok == ROFF_de1)
1.93      schwarze 1740:                tok = ROFF_de;
1.123     schwarze 1741:        else if (tok == ROFF_dei1)
                   1742:                tok = ROFF_dei;
1.110     schwarze 1743:        else if (tok == ROFF_am1)
1.93      schwarze 1744:                tok = ROFF_am;
1.123     schwarze 1745:        else if (tok == ROFF_ami1)
                   1746:                tok = ROFF_ami;
1.93      schwarze 1747:
                   1748:        /* Parse the macro name argument. */
                   1749:
1.110     schwarze 1750:        cp = buf->buf + pos;
                   1751:        if (tok == ROFF_ig) {
1.93      schwarze 1752:                iname = NULL;
                   1753:                namesz = 0;
                   1754:        } else {
                   1755:                iname = cp;
                   1756:                namesz = roff_getname(r, &cp, ln, ppos);
                   1757:                iname[namesz] = '\0';
                   1758:        }
1.22      schwarze 1759:
1.93      schwarze 1760:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1761:
1.110     schwarze 1762:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187   ! schwarze 1763:                deftype = ROFFDEF_USER;
        !          1764:                name = roff_getstrn(r, iname, namesz, &deftype);
        !          1765:                if (name == NULL) {
1.93      schwarze 1766:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1767:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1768:                            "%.*s", (int)namesz, iname);
                   1769:                        namesz = 0;
                   1770:                } else
                   1771:                        namesz = strlen(name);
                   1772:        } else
                   1773:                name = iname;
1.22      schwarze 1774:
1.110     schwarze 1775:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1776:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167     schwarze 1777:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1778:                return ROFF_IGN;
1.93      schwarze 1779:        }
1.2       schwarze 1780:
1.16      schwarze 1781:        roffnode_push(r, tok, name, ln, ppos);
                   1782:
                   1783:        /*
                   1784:         * At the beginning of a `de' macro, clear the existing string
                   1785:         * with the same name, if there is one.  New content will be
1.66      schwarze 1786:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1787:         */
1.22      schwarze 1788:
1.183     schwarze 1789:        if (tok == ROFF_de || tok == ROFF_dei) {
1.86      schwarze 1790:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183     schwarze 1791:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.187   ! schwarze 1792:        } else if (tok == ROFF_am || tok == ROFF_ami) {
        !          1793:                deftype = ROFFDEF_ANY;
        !          1794:                value = roff_getstrn(r, iname, namesz, &deftype);
        !          1795:                switch (deftype) {  /* Before appending, ... */
        !          1796:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
        !          1797:                        roff_setstrn(&r->strtab, name, namesz,
        !          1798:                            value, strlen(value), 0);
        !          1799:                        break;
        !          1800:                case ROFFDEF_REN: /* call original standard macro. */
        !          1801:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
        !          1802:                            (int)strlen(value), value);
        !          1803:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
        !          1804:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
        !          1805:                        free(call);
        !          1806:                        break;
        !          1807:                case ROFFDEF_STD:  /* rename and call standard macro. */
        !          1808:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
        !          1809:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
        !          1810:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
        !          1811:                            (int)rsz, rname);
        !          1812:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
        !          1813:                        free(call);
        !          1814:                        free(rname);
        !          1815:                        break;
        !          1816:                default:
        !          1817:                        break;
        !          1818:                }
1.183     schwarze 1819:        }
1.2       schwarze 1820:
1.110     schwarze 1821:        if (*cp == '\0')
1.149     schwarze 1822:                return ROFF_IGN;
1.1       schwarze 1823:
1.93      schwarze 1824:        /* Get the custom end marker. */
1.22      schwarze 1825:
1.93      schwarze 1826:        iname = cp;
1.86      schwarze 1827:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1828:
                   1829:        /* Resolve the end marker if it is indirect. */
                   1830:
1.110     schwarze 1831:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187   ! schwarze 1832:                deftype = ROFFDEF_USER;
        !          1833:                name = roff_getstrn(r, iname, namesz, &deftype);
        !          1834:                if (name == NULL) {
1.93      schwarze 1835:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1836:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1837:                            "%.*s", (int)namesz, iname);
                   1838:                        namesz = 0;
                   1839:                } else
                   1840:                        namesz = strlen(name);
                   1841:        } else
                   1842:                name = iname;
                   1843:
1.86      schwarze 1844:        if (namesz)
                   1845:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1846:
1.110     schwarze 1847:        if (*cp != '\0')
1.90      schwarze 1848:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167     schwarze 1849:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1850:
1.149     schwarze 1851:        return ROFF_IGN;
1.1       schwarze 1852: }
                   1853:
                   1854: static enum rofferr
1.2       schwarze 1855: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1856: {
1.166     schwarze 1857:        enum roff_tok   t;
1.2       schwarze 1858:        int             i, j;
                   1859:
                   1860:        /*
                   1861:         * First check whether a custom macro exists at this level.  If
                   1862:         * it does, then check against it.  This is some of groff's
                   1863:         * stranger behaviours.  If we encountered a custom end-scope
                   1864:         * tag and that tag also happens to be a "real" macro, then we
                   1865:         * need to try interpreting it again as a real macro.  If it's
                   1866:         * not, then return ignore.  Else continue.
                   1867:         */
                   1868:
                   1869:        if (r->last->end) {
1.35      schwarze 1870:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1871:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1872:                                break;
1.1       schwarze 1873:
1.110     schwarze 1874:                if (r->last->end[j] == '\0' &&
                   1875:                    (buf->buf[i] == '\0' ||
                   1876:                     buf->buf[i] == ' ' ||
                   1877:                     buf->buf[i] == '\t')) {
1.2       schwarze 1878:                        roffnode_pop(r);
                   1879:                        roffnode_cleanscope(r);
1.1       schwarze 1880:
1.110     schwarze 1881:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1882:                                i++;
                   1883:
                   1884:                        pos = i;
1.110     schwarze 1885:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1886:                            TOKEN_NONE)
1.149     schwarze 1887:                                return ROFF_RERUN;
                   1888:                        return ROFF_IGN;
1.2       schwarze 1889:                }
1.1       schwarze 1890:        }
                   1891:
1.2       schwarze 1892:        /*
                   1893:         * If we have no custom end-query or lookup failed, then try
                   1894:         * pulling it out of the hashtable.
                   1895:         */
1.1       schwarze 1896:
1.110     schwarze 1897:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1898:
1.110     schwarze 1899:        if (t != ROFF_cblock) {
                   1900:                if (tok != ROFF_ig)
                   1901:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1902:                return ROFF_IGN;
1.16      schwarze 1903:        }
1.1       schwarze 1904:
1.149     schwarze 1905:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1906: }
                   1907:
                   1908: static enum rofferr
                   1909: roff_block_text(ROFF_ARGS)
                   1910: {
                   1911:
1.110     schwarze 1912:        if (tok != ROFF_ig)
                   1913:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1914:
1.149     schwarze 1915:        return ROFF_IGN;
1.2       schwarze 1916: }
                   1917:
                   1918: static enum rofferr
                   1919: roff_cond_sub(ROFF_ARGS)
                   1920: {
1.166     schwarze 1921:        enum roff_tok    t;
1.37      schwarze 1922:        char            *ep;
1.71      schwarze 1923:        int              rr;
1.2       schwarze 1924:
                   1925:        rr = r->last->rule;
1.37      schwarze 1926:        roffnode_cleanscope(r);
1.110     schwarze 1927:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2       schwarze 1928:
1.37      schwarze 1929:        /*
1.50      schwarze 1930:         * Fully handle known macros when they are structurally
                   1931:         * required or when the conditional evaluated to true.
1.5       schwarze 1932:         */
                   1933:
1.166     schwarze 1934:        if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.149     schwarze 1935:                return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.39      schwarze 1936:
1.69      schwarze 1937:        /*
                   1938:         * If `\}' occurs on a macro line without a preceding macro,
                   1939:         * drop the line completely.
                   1940:         */
                   1941:
1.110     schwarze 1942:        ep = buf->buf + pos;
                   1943:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1944:                rr = 0;
1.69      schwarze 1945:
1.50      schwarze 1946:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1947:
1.110     schwarze 1948:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1949:                if (*(++ep) == '}') {
1.70      schwarze 1950:                        *ep = '&';
1.110     schwarze 1951:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1952:                }
1.119     schwarze 1953:                if (*ep != '\0')
                   1954:                        ++ep;
1.50      schwarze 1955:        }
1.149     schwarze 1956:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1957: }
                   1958:
                   1959: static enum rofferr
                   1960: roff_cond_text(ROFF_ARGS)
                   1961: {
1.37      schwarze 1962:        char            *ep;
1.71      schwarze 1963:        int              rr;
1.2       schwarze 1964:
                   1965:        rr = r->last->rule;
1.37      schwarze 1966:        roffnode_cleanscope(r);
1.1       schwarze 1967:
1.110     schwarze 1968:        ep = buf->buf + pos;
                   1969:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1970:                if (*(++ep) == '}') {
1.70      schwarze 1971:                        *ep = '&';
1.110     schwarze 1972:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1973:                }
1.119     schwarze 1974:                if (*ep != '\0')
                   1975:                        ++ep;
1.2       schwarze 1976:        }
1.149     schwarze 1977:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1978: }
                   1979:
1.138     schwarze 1980: /* --- handling of numeric and conditional expressions -------------------- */
                   1981:
1.77      schwarze 1982: /*
                   1983:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1984:  * If there is at least one digit, return success and advance the
                   1985:  * parse point, else return failure and let the parse point unchanged.
                   1986:  * Ignore overflows, treat them just like the C language.
                   1987:  */
1.56      schwarze 1988: static int
1.133     schwarze 1989: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 1990: {
1.133     schwarze 1991:        int      myres, scaled, n, p;
1.79      schwarze 1992:
                   1993:        if (NULL == res)
                   1994:                res = &myres;
1.56      schwarze 1995:
                   1996:        p = *pos;
                   1997:        n = v[p] == '-';
1.133     schwarze 1998:        if (n || v[p] == '+')
1.56      schwarze 1999:                p++;
                   2000:
1.133     schwarze 2001:        if (flags & ROFFNUM_WHITE)
                   2002:                while (isspace((unsigned char)v[p]))
                   2003:                        p++;
                   2004:
1.56      schwarze 2005:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 2006:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 2007:        if (p == *pos + n)
                   2008:                return 0;
                   2009:
                   2010:        if (n)
                   2011:                *res = -*res;
                   2012:
1.126     schwarze 2013:        /* Each number may be followed by one optional scaling unit. */
                   2014:
                   2015:        switch (v[p]) {
                   2016:        case 'f':
1.133     schwarze 2017:                scaled = *res * 65536;
1.126     schwarze 2018:                break;
                   2019:        case 'i':
1.133     schwarze 2020:                scaled = *res * 240;
1.126     schwarze 2021:                break;
                   2022:        case 'c':
1.133     schwarze 2023:                scaled = *res * 240 / 2.54;
1.126     schwarze 2024:                break;
                   2025:        case 'v':
                   2026:        case 'P':
1.133     schwarze 2027:                scaled = *res * 40;
1.126     schwarze 2028:                break;
                   2029:        case 'm':
                   2030:        case 'n':
1.133     schwarze 2031:                scaled = *res * 24;
1.126     schwarze 2032:                break;
                   2033:        case 'p':
1.133     schwarze 2034:                scaled = *res * 10 / 3;
1.126     schwarze 2035:                break;
                   2036:        case 'u':
1.133     schwarze 2037:                scaled = *res;
1.126     schwarze 2038:                break;
                   2039:        case 'M':
1.133     schwarze 2040:                scaled = *res * 6 / 25;
1.126     schwarze 2041:                break;
                   2042:        default:
1.133     schwarze 2043:                scaled = *res;
1.126     schwarze 2044:                p--;
                   2045:                break;
                   2046:        }
1.133     schwarze 2047:        if (flags & ROFFNUM_SCALE)
                   2048:                *res = scaled;
1.126     schwarze 2049:
                   2050:        *pos = p + 1;
1.149     schwarze 2051:        return 1;
1.56      schwarze 2052: }
                   2053:
1.71      schwarze 2054: /*
                   2055:  * Evaluate a string comparison condition.
                   2056:  * The first character is the delimiter.
                   2057:  * Succeed if the string up to its second occurrence
                   2058:  * matches the string up to its third occurence.
                   2059:  * Advance the cursor after the third occurrence
                   2060:  * or lacking that, to the end of the line.
                   2061:  */
                   2062: static int
                   2063: roff_evalstrcond(const char *v, int *pos)
                   2064: {
                   2065:        const char      *s1, *s2, *s3;
                   2066:        int              match;
                   2067:
                   2068:        match = 0;
                   2069:        s1 = v + *pos;          /* initial delimiter */
                   2070:        s2 = s1 + 1;            /* for scanning the first string */
                   2071:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2072:
                   2073:        if (NULL == s3)         /* found no middle delimiter */
                   2074:                goto out;
                   2075:
                   2076:        while ('\0' != *++s3) {
                   2077:                if (*s2 != *s3) {  /* mismatch */
                   2078:                        s3 = strchr(s3, *s1);
                   2079:                        break;
                   2080:                }
                   2081:                if (*s3 == *s1) {  /* found the final delimiter */
                   2082:                        match = 1;
                   2083:                        break;
                   2084:                }
                   2085:                s2++;
                   2086:        }
                   2087:
                   2088: out:
                   2089:        if (NULL == s3)
                   2090:                s3 = strchr(s2, '\0');
1.114     schwarze 2091:        else if (*s3 != '\0')
1.71      schwarze 2092:                s3++;
                   2093:        *pos = s3 - v;
1.149     schwarze 2094:        return match;
1.71      schwarze 2095: }
                   2096:
1.77      schwarze 2097: /*
                   2098:  * Evaluate an optionally negated single character, numerical,
                   2099:  * or string condition.
                   2100:  */
1.71      schwarze 2101: static int
1.143     schwarze 2102: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 2103: {
1.143     schwarze 2104:        char    *cp, *name;
                   2105:        size_t   sz;
1.187   ! schwarze 2106:        int      deftype, number, savepos, istrue, wanttrue;
1.5       schwarze 2107:
1.71      schwarze 2108:        if ('!' == v[*pos]) {
                   2109:                wanttrue = 0;
                   2110:                (*pos)++;
                   2111:        } else
                   2112:                wanttrue = 1;
                   2113:
1.5       schwarze 2114:        switch (v[*pos]) {
1.112     schwarze 2115:        case '\0':
1.149     schwarze 2116:                return 0;
1.80      schwarze 2117:        case 'n':
                   2118:        case 'o':
1.5       schwarze 2119:                (*pos)++;
1.149     schwarze 2120:                return wanttrue;
1.80      schwarze 2121:        case 'c':
                   2122:        case 'e':
                   2123:        case 't':
1.111     schwarze 2124:        case 'v':
1.5       schwarze 2125:                (*pos)++;
1.149     schwarze 2126:                return !wanttrue;
1.182     schwarze 2127:        case 'd':
1.143     schwarze 2128:        case 'r':
1.182     schwarze 2129:                cp = v + *pos + 1;
                   2130:                while (*cp == ' ')
                   2131:                        cp++;
                   2132:                name = cp;
                   2133:                sz = roff_getname(r, &cp, ln, cp - v);
1.187   ! schwarze 2134:                if (sz == 0)
        !          2135:                        istrue = 0;
        !          2136:                else if (v[*pos] == 'r')
        !          2137:                        istrue = roff_hasregn(r, name, sz);
        !          2138:                else {
        !          2139:                        deftype = ROFFDEF_ANY;
        !          2140:                        roff_getstrn(r, name, sz, &deftype);
        !          2141:                        istrue = !!deftype;
        !          2142:                }
1.143     schwarze 2143:                *pos = cp - v;
1.182     schwarze 2144:                return istrue == wanttrue;
1.5       schwarze 2145:        default:
                   2146:                break;
                   2147:        }
                   2148:
1.113     schwarze 2149:        savepos = *pos;
1.133     schwarze 2150:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 2151:                return (number > 0) == wanttrue;
1.113     schwarze 2152:        else if (*pos == savepos)
1.149     schwarze 2153:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 2154:        else
1.149     schwarze 2155:                return 0;
1.5       schwarze 2156: }
                   2157:
1.2       schwarze 2158: static enum rofferr
1.21      schwarze 2159: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 2160: {
1.30      schwarze 2161:
1.149     schwarze 2162:        return ROFF_IGN;
1.21      schwarze 2163: }
                   2164:
                   2165: static enum rofferr
1.123     schwarze 2166: roff_insec(ROFF_ARGS)
                   2167: {
                   2168:
                   2169:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167     schwarze 2170:            ln, ppos, roff_name[tok]);
1.149     schwarze 2171:        return ROFF_IGN;
1.123     schwarze 2172: }
                   2173:
                   2174: static enum rofferr
                   2175: roff_unsupp(ROFF_ARGS)
                   2176: {
                   2177:
                   2178:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167     schwarze 2179:            ln, ppos, roff_name[tok]);
1.149     schwarze 2180:        return ROFF_IGN;
1.123     schwarze 2181: }
                   2182:
                   2183: static enum rofferr
1.2       schwarze 2184: roff_cond(ROFF_ARGS)
                   2185: {
1.46      schwarze 2186:
                   2187:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2188:
1.80      schwarze 2189:        /*
1.35      schwarze 2190:         * An `.el' has no conditional body: it will consume the value
                   2191:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2192:         * defaults to DENY.
1.35      schwarze 2193:         *
                   2194:         * If we're not an `el', however, then evaluate the conditional.
                   2195:         */
1.1       schwarze 2196:
1.110     schwarze 2197:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2198:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2199:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2200:
1.35      schwarze 2201:        /*
                   2202:         * An if-else will put the NEGATION of the current evaluated
                   2203:         * conditional into the stack of rules.
                   2204:         */
                   2205:
1.110     schwarze 2206:        if (tok == ROFF_ie) {
1.96      schwarze 2207:                if (r->rstackpos + 1 == r->rstacksz) {
                   2208:                        r->rstacksz += 16;
                   2209:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2210:                            r->rstacksz, sizeof(int));
1.35      schwarze 2211:                }
1.71      schwarze 2212:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2213:        }
1.5       schwarze 2214:
                   2215:        /* If the parent has false as its rule, then so do we. */
                   2216:
1.71      schwarze 2217:        if (r->last->parent && !r->last->parent->rule)
                   2218:                r->last->rule = 0;
1.5       schwarze 2219:
                   2220:        /*
1.46      schwarze 2221:         * Determine scope.
                   2222:         * If there is nothing on the line after the conditional,
                   2223:         * not even whitespace, use next-line scope.
1.5       schwarze 2224:         */
1.2       schwarze 2225:
1.110     schwarze 2226:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2227:                r->last->endspan = 2;
                   2228:                goto out;
                   2229:        }
                   2230:
1.110     schwarze 2231:        while (buf->buf[pos] == ' ')
1.46      schwarze 2232:                pos++;
                   2233:
                   2234:        /* An opening brace requests multiline scope. */
1.2       schwarze 2235:
1.110     schwarze 2236:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2237:                r->last->endspan = -1;
                   2238:                pos += 2;
1.144     schwarze 2239:                while (buf->buf[pos] == ' ')
                   2240:                        pos++;
1.46      schwarze 2241:                goto out;
1.80      schwarze 2242:        }
1.2       schwarze 2243:
                   2244:        /*
1.46      schwarze 2245:         * Anything else following the conditional causes
                   2246:         * single-line scope.  Warn if the scope contains
                   2247:         * nothing but trailing whitespace.
1.2       schwarze 2248:         */
                   2249:
1.110     schwarze 2250:        if (buf->buf[pos] == '\0')
1.89      schwarze 2251:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167     schwarze 2252:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2253:
1.46      schwarze 2254:        r->last->endspan = 1;
1.1       schwarze 2255:
1.46      schwarze 2256: out:
1.2       schwarze 2257:        *offs = pos;
1.149     schwarze 2258:        return ROFF_RERUN;
1.1       schwarze 2259: }
                   2260:
1.2       schwarze 2261: static enum rofferr
1.7       schwarze 2262: roff_ds(ROFF_ARGS)
                   2263: {
1.85      schwarze 2264:        char            *string;
                   2265:        const char      *name;
                   2266:        size_t           namesz;
1.10      schwarze 2267:
1.123     schwarze 2268:        /* Ignore groff compatibility mode for now. */
                   2269:
                   2270:        if (tok == ROFF_ds1)
                   2271:                tok = ROFF_ds;
                   2272:        else if (tok == ROFF_as1)
                   2273:                tok = ROFF_as;
                   2274:
1.10      schwarze 2275:        /*
1.85      schwarze 2276:         * The first word is the name of the string.
                   2277:         * If it is empty or terminated by an escape sequence,
                   2278:         * abort the `ds' request without defining anything.
1.10      schwarze 2279:         */
1.7       schwarze 2280:
1.110     schwarze 2281:        name = string = buf->buf + pos;
                   2282:        if (*name == '\0')
1.149     schwarze 2283:                return ROFF_IGN;
1.7       schwarze 2284:
1.85      schwarze 2285:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2286:        if (name[namesz] == '\\')
1.149     schwarze 2287:                return ROFF_IGN;
1.85      schwarze 2288:
                   2289:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2290:        if (*string == '"')
1.7       schwarze 2291:                string++;
                   2292:
1.10      schwarze 2293:        /* The rest is the value. */
1.85      schwarze 2294:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2295:            ROFF_as == tok);
1.183     schwarze 2296:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149     schwarze 2297:        return ROFF_IGN;
1.7       schwarze 2298: }
                   2299:
1.77      schwarze 2300: /*
                   2301:  * Parse a single operator, one or two characters long.
                   2302:  * If the operator is recognized, return success and advance the
                   2303:  * parse point, else return failure and let the parse point unchanged.
                   2304:  */
                   2305: static int
                   2306: roff_getop(const char *v, int *pos, char *res)
                   2307: {
                   2308:
                   2309:        *res = v[*pos];
                   2310:
                   2311:        switch (*res) {
1.80      schwarze 2312:        case '+':
                   2313:        case '-':
                   2314:        case '*':
                   2315:        case '/':
                   2316:        case '%':
                   2317:        case '&':
                   2318:        case ':':
1.77      schwarze 2319:                break;
                   2320:        case '<':
                   2321:                switch (v[*pos + 1]) {
1.80      schwarze 2322:                case '=':
1.77      schwarze 2323:                        *res = 'l';
                   2324:                        (*pos)++;
                   2325:                        break;
1.80      schwarze 2326:                case '>':
1.77      schwarze 2327:                        *res = '!';
                   2328:                        (*pos)++;
                   2329:                        break;
1.80      schwarze 2330:                case '?':
1.77      schwarze 2331:                        *res = 'i';
                   2332:                        (*pos)++;
                   2333:                        break;
                   2334:                default:
                   2335:                        break;
                   2336:                }
                   2337:                break;
                   2338:        case '>':
                   2339:                switch (v[*pos + 1]) {
1.80      schwarze 2340:                case '=':
1.77      schwarze 2341:                        *res = 'g';
                   2342:                        (*pos)++;
                   2343:                        break;
1.80      schwarze 2344:                case '?':
1.77      schwarze 2345:                        *res = 'a';
                   2346:                        (*pos)++;
                   2347:                        break;
                   2348:                default:
                   2349:                        break;
                   2350:                }
                   2351:                break;
                   2352:        case '=':
                   2353:                if ('=' == v[*pos + 1])
                   2354:                        (*pos)++;
                   2355:                break;
                   2356:        default:
1.149     schwarze 2357:                return 0;
1.77      schwarze 2358:        }
                   2359:        (*pos)++;
                   2360:
1.149     schwarze 2361:        return *res;
1.77      schwarze 2362: }
                   2363:
                   2364: /*
                   2365:  * Evaluate either a parenthesized numeric expression
                   2366:  * or a single signed integer number.
                   2367:  */
                   2368: static int
1.106     schwarze 2369: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2370:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2371: {
                   2372:
                   2373:        if ('(' != v[*pos])
1.149     schwarze 2374:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2375:
                   2376:        (*pos)++;
1.133     schwarze 2377:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2378:                return 0;
1.77      schwarze 2379:
1.79      schwarze 2380:        /*
                   2381:         * Omission of the closing parenthesis
                   2382:         * is an error in validation mode,
                   2383:         * but ignored in evaluation mode.
                   2384:         */
                   2385:
1.77      schwarze 2386:        if (')' == v[*pos])
                   2387:                (*pos)++;
1.79      schwarze 2388:        else if (NULL == res)
1.149     schwarze 2389:                return 0;
1.77      schwarze 2390:
1.149     schwarze 2391:        return 1;
1.77      schwarze 2392: }
                   2393:
                   2394: /*
                   2395:  * Evaluate a complete numeric expression.
                   2396:  * Proceed left to right, there is no concept of precedence.
                   2397:  */
                   2398: static int
1.106     schwarze 2399: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2400:        int *pos, int *res, int flags)
1.77      schwarze 2401: {
                   2402:        int              mypos, operand2;
                   2403:        char             operator;
                   2404:
                   2405:        if (NULL == pos) {
                   2406:                mypos = 0;
                   2407:                pos = &mypos;
                   2408:        }
                   2409:
1.133     schwarze 2410:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2411:                while (isspace((unsigned char)v[*pos]))
                   2412:                        (*pos)++;
                   2413:
1.133     schwarze 2414:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2415:                return 0;
1.77      schwarze 2416:
                   2417:        while (1) {
1.133     schwarze 2418:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2419:                        while (isspace((unsigned char)v[*pos]))
                   2420:                                (*pos)++;
                   2421:
                   2422:                if ( ! roff_getop(v, pos, &operator))
                   2423:                        break;
                   2424:
1.133     schwarze 2425:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2426:                        while (isspace((unsigned char)v[*pos]))
                   2427:                                (*pos)++;
                   2428:
1.133     schwarze 2429:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2430:                        return 0;
1.77      schwarze 2431:
1.133     schwarze 2432:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2433:                        while (isspace((unsigned char)v[*pos]))
                   2434:                                (*pos)++;
1.79      schwarze 2435:
                   2436:                if (NULL == res)
                   2437:                        continue;
1.77      schwarze 2438:
                   2439:                switch (operator) {
1.80      schwarze 2440:                case '+':
1.77      schwarze 2441:                        *res += operand2;
                   2442:                        break;
1.80      schwarze 2443:                case '-':
1.77      schwarze 2444:                        *res -= operand2;
                   2445:                        break;
1.80      schwarze 2446:                case '*':
1.77      schwarze 2447:                        *res *= operand2;
                   2448:                        break;
1.80      schwarze 2449:                case '/':
1.116     schwarze 2450:                        if (operand2 == 0) {
1.106     schwarze 2451:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2452:                                        r->parse, ln, *pos, v);
                   2453:                                *res = 0;
                   2454:                                break;
                   2455:                        }
1.77      schwarze 2456:                        *res /= operand2;
                   2457:                        break;
1.80      schwarze 2458:                case '%':
1.116     schwarze 2459:                        if (operand2 == 0) {
                   2460:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2461:                                        r->parse, ln, *pos, v);
                   2462:                                *res = 0;
                   2463:                                break;
                   2464:                        }
1.77      schwarze 2465:                        *res %= operand2;
                   2466:                        break;
1.80      schwarze 2467:                case '<':
1.77      schwarze 2468:                        *res = *res < operand2;
                   2469:                        break;
1.80      schwarze 2470:                case '>':
1.77      schwarze 2471:                        *res = *res > operand2;
                   2472:                        break;
1.80      schwarze 2473:                case 'l':
1.77      schwarze 2474:                        *res = *res <= operand2;
                   2475:                        break;
1.80      schwarze 2476:                case 'g':
1.77      schwarze 2477:                        *res = *res >= operand2;
                   2478:                        break;
1.80      schwarze 2479:                case '=':
1.77      schwarze 2480:                        *res = *res == operand2;
                   2481:                        break;
1.80      schwarze 2482:                case '!':
1.77      schwarze 2483:                        *res = *res != operand2;
                   2484:                        break;
1.80      schwarze 2485:                case '&':
1.77      schwarze 2486:                        *res = *res && operand2;
                   2487:                        break;
1.80      schwarze 2488:                case ':':
1.77      schwarze 2489:                        *res = *res || operand2;
                   2490:                        break;
1.80      schwarze 2491:                case 'i':
1.77      schwarze 2492:                        if (operand2 < *res)
                   2493:                                *res = operand2;
                   2494:                        break;
1.80      schwarze 2495:                case 'a':
1.77      schwarze 2496:                        if (operand2 > *res)
                   2497:                                *res = operand2;
                   2498:                        break;
                   2499:                default:
                   2500:                        abort();
                   2501:                }
                   2502:        }
1.149     schwarze 2503:        return 1;
1.77      schwarze 2504: }
                   2505:
1.138     schwarze 2506: /* --- register management ------------------------------------------------ */
                   2507:
1.52      schwarze 2508: void
1.60      schwarze 2509: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2510: {
1.52      schwarze 2511:        struct roffreg  *reg;
                   2512:
                   2513:        /* Search for an existing register with the same name. */
                   2514:        reg = r->regtab;
                   2515:
                   2516:        while (reg && strcmp(name, reg->key.p))
                   2517:                reg = reg->next;
1.41      schwarze 2518:
1.52      schwarze 2519:        if (NULL == reg) {
                   2520:                /* Create a new register. */
                   2521:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2522:                reg->key.p = mandoc_strdup(name);
                   2523:                reg->key.sz = strlen(name);
1.60      schwarze 2524:                reg->val = 0;
1.52      schwarze 2525:                reg->next = r->regtab;
                   2526:                r->regtab = reg;
                   2527:        }
                   2528:
1.60      schwarze 2529:        if ('+' == sign)
                   2530:                reg->val += val;
                   2531:        else if ('-' == sign)
                   2532:                reg->val -= val;
                   2533:        else
                   2534:                reg->val = val;
1.41      schwarze 2535: }
                   2536:
1.65      schwarze 2537: /*
                   2538:  * Handle some predefined read-only number registers.
                   2539:  * For now, return -1 if the requested register is not predefined;
                   2540:  * in case a predefined read-only register having the value -1
                   2541:  * were to turn up, another special value would have to be chosen.
                   2542:  */
                   2543: static int
1.145     schwarze 2544: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2545: {
                   2546:
                   2547:        switch (*name) {
1.145     schwarze 2548:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2549:                return r->argc;
1.80      schwarze 2550:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2551:                return 0;
1.80      schwarze 2552:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2553:                return 1;
1.80      schwarze 2554:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2555:                return 24;
1.80      schwarze 2556:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2557:                return 0;
1.80      schwarze 2558:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2559:                return 1;
1.80      schwarze 2560:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2561:                return 40;
1.65      schwarze 2562:        default:
1.149     schwarze 2563:                return -1;
1.65      schwarze 2564:        }
                   2565: }
                   2566:
1.53      schwarze 2567: int
1.52      schwarze 2568: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2569: {
1.52      schwarze 2570:        struct roffreg  *reg;
1.65      schwarze 2571:        int              val;
                   2572:
                   2573:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145     schwarze 2574:                val = roff_getregro(r, name + 1);
1.65      schwarze 2575:                if (-1 != val)
1.149     schwarze 2576:                        return val;
1.65      schwarze 2577:        }
1.52      schwarze 2578:
                   2579:        for (reg = r->regtab; reg; reg = reg->next)
                   2580:                if (0 == strcmp(name, reg->key.p))
1.149     schwarze 2581:                        return reg->val;
1.53      schwarze 2582:
1.149     schwarze 2583:        return 0;
1.53      schwarze 2584: }
                   2585:
                   2586: static int
                   2587: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2588: {
                   2589:        struct roffreg  *reg;
1.65      schwarze 2590:        int              val;
                   2591:
                   2592:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2593:                val = roff_getregro(r, name + 1);
1.65      schwarze 2594:                if (-1 != val)
1.149     schwarze 2595:                        return val;
1.65      schwarze 2596:        }
1.53      schwarze 2597:
                   2598:        for (reg = r->regtab; reg; reg = reg->next)
                   2599:                if (len == reg->key.sz &&
                   2600:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2601:                        return reg->val;
1.143     schwarze 2602:
1.149     schwarze 2603:        return 0;
1.143     schwarze 2604: }
                   2605:
                   2606: static int
                   2607: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2608: {
                   2609:        struct roffreg  *reg;
                   2610:        int              val;
                   2611:
                   2612:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2613:                val = roff_getregro(r, name + 1);
1.143     schwarze 2614:                if (-1 != val)
1.149     schwarze 2615:                        return 1;
1.143     schwarze 2616:        }
                   2617:
                   2618:        for (reg = r->regtab; reg; reg = reg->next)
                   2619:                if (len == reg->key.sz &&
                   2620:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2621:                        return 1;
1.41      schwarze 2622:
1.149     schwarze 2623:        return 0;
1.41      schwarze 2624: }
                   2625:
1.52      schwarze 2626: static void
                   2627: roff_freereg(struct roffreg *reg)
1.41      schwarze 2628: {
1.52      schwarze 2629:        struct roffreg  *old_reg;
1.41      schwarze 2630:
1.52      schwarze 2631:        while (NULL != reg) {
                   2632:                free(reg->key.p);
                   2633:                old_reg = reg;
                   2634:                reg = reg->next;
                   2635:                free(old_reg);
                   2636:        }
1.41      schwarze 2637: }
1.7       schwarze 2638:
                   2639: static enum rofferr
1.6       schwarze 2640: roff_nr(ROFF_ARGS)
1.1       schwarze 2641: {
1.85      schwarze 2642:        char            *key, *val;
                   2643:        size_t           keysz;
1.37      schwarze 2644:        int              iv;
1.60      schwarze 2645:        char             sign;
1.6       schwarze 2646:
1.110     schwarze 2647:        key = val = buf->buf + pos;
                   2648:        if (*key == '\0')
1.149     schwarze 2649:                return ROFF_IGN;
1.85      schwarze 2650:
                   2651:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2652:        if (key[keysz] == '\\')
1.149     schwarze 2653:                return ROFF_IGN;
1.85      schwarze 2654:        key[keysz] = '\0';
1.6       schwarze 2655:
1.60      schwarze 2656:        sign = *val;
1.110     schwarze 2657:        if (sign == '+' || sign == '-')
1.60      schwarze 2658:                val++;
                   2659:
1.133     schwarze 2660:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77      schwarze 2661:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2662:
1.149     schwarze 2663:        return ROFF_IGN;
1.76      schwarze 2664: }
                   2665:
                   2666: static enum rofferr
                   2667: roff_rr(ROFF_ARGS)
                   2668: {
                   2669:        struct roffreg  *reg, **prev;
1.85      schwarze 2670:        char            *name, *cp;
                   2671:        size_t           namesz;
1.76      schwarze 2672:
1.110     schwarze 2673:        name = cp = buf->buf + pos;
                   2674:        if (*name == '\0')
1.149     schwarze 2675:                return ROFF_IGN;
1.85      schwarze 2676:        namesz = roff_getname(r, &cp, ln, pos);
                   2677:        name[namesz] = '\0';
1.76      schwarze 2678:
                   2679:        prev = &r->regtab;
                   2680:        while (1) {
                   2681:                reg = *prev;
1.110     schwarze 2682:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2683:                        break;
                   2684:                prev = &reg->next;
                   2685:        }
1.110     schwarze 2686:        if (reg != NULL) {
1.76      schwarze 2687:                *prev = reg->next;
                   2688:                free(reg->key.p);
                   2689:                free(reg);
                   2690:        }
1.149     schwarze 2691:        return ROFF_IGN;
1.29      schwarze 2692: }
                   2693:
1.138     schwarze 2694: /* --- handler functions for roff requests -------------------------------- */
                   2695:
1.29      schwarze 2696: static enum rofferr
                   2697: roff_rm(ROFF_ARGS)
                   2698: {
                   2699:        const char       *name;
                   2700:        char             *cp;
1.85      schwarze 2701:        size_t            namesz;
1.29      schwarze 2702:
1.110     schwarze 2703:        cp = buf->buf + pos;
                   2704:        while (*cp != '\0') {
1.85      schwarze 2705:                name = cp;
1.110     schwarze 2706:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2707:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183     schwarze 2708:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.110     schwarze 2709:                if (name[namesz] == '\\')
1.85      schwarze 2710:                        break;
1.29      schwarze 2711:        }
1.149     schwarze 2712:        return ROFF_IGN;
1.51      schwarze 2713: }
                   2714:
                   2715: static enum rofferr
                   2716: roff_it(ROFF_ARGS)
                   2717: {
                   2718:        int              iv;
                   2719:
                   2720:        /* Parse the number of lines. */
1.133     schwarze 2721:
                   2722:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2723:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2724:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2725:                return ROFF_IGN;
1.51      schwarze 2726:        }
                   2727:
1.134     schwarze 2728:        while (isspace((unsigned char)buf->buf[pos]))
                   2729:                pos++;
                   2730:
                   2731:        /*
                   2732:         * Arm the input line trap.
                   2733:         * Special-casing "an-trap" is an ugly workaround to cope
                   2734:         * with DocBook stupidly fiddling with man(7) internals.
                   2735:         */
1.133     schwarze 2736:
1.51      schwarze 2737:        roffit_lines = iv;
1.134     schwarze 2738:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2739:            strcmp(buf->buf + pos, "an-trap") ?
                   2740:            buf->buf + pos : "br");
1.149     schwarze 2741:        return ROFF_IGN;
1.47      schwarze 2742: }
                   2743:
                   2744: static enum rofferr
                   2745: roff_Dd(ROFF_ARGS)
                   2746: {
1.187   ! schwarze 2747:        int              mask;
        !          2748:        enum roff_tok    t, te;
1.99      schwarze 2749:
1.187   ! schwarze 2750:        switch (tok) {
        !          2751:        case ROFF_Dd:
        !          2752:                tok = MDOC_Dd;
        !          2753:                te = MDOC_MAX;
        !          2754:                if (r->format == 0)
        !          2755:                        r->format = MPARSE_MDOC;
        !          2756:                mask = MPARSE_MDOC | MPARSE_QUICK;
        !          2757:                break;
        !          2758:        case ROFF_TH:
        !          2759:                tok = MAN_TH;
        !          2760:                te = MAN_MAX;
        !          2761:                if (r->format == 0)
        !          2762:                        r->format = MPARSE_MAN;
        !          2763:                mask = MPARSE_QUICK;
        !          2764:                break;
        !          2765:        default:
        !          2766:                abort();
        !          2767:        }
        !          2768:        if ((r->options & mask) == 0)
        !          2769:                for (t = tok; t < te; t++)
        !          2770:                        roff_setstr(r, roff_name[t], NULL, 0);
1.149     schwarze 2771:        return ROFF_CONT;
1.14      schwarze 2772: }
                   2773:
                   2774: static enum rofferr
1.27      schwarze 2775: roff_TE(ROFF_ARGS)
                   2776: {
                   2777:
                   2778:        if (NULL == r->tbl)
1.94      schwarze 2779:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2780:                    ln, ppos, "TE");
1.130     schwarze 2781:        else if ( ! tbl_end(&r->tbl)) {
                   2782:                free(buf->buf);
                   2783:                buf->buf = mandoc_strdup(".sp");
                   2784:                buf->sz = 4;
1.149     schwarze 2785:                return ROFF_REPARSE;
1.130     schwarze 2786:        }
1.149     schwarze 2787:        return ROFF_IGN;
1.27      schwarze 2788: }
                   2789:
                   2790: static enum rofferr
                   2791: roff_T_(ROFF_ARGS)
                   2792: {
                   2793:
                   2794:        if (NULL == r->tbl)
1.94      schwarze 2795:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2796:                    ln, ppos, "T&");
1.27      schwarze 2797:        else
1.168     schwarze 2798:                tbl_restart(ln, ppos, r->tbl);
1.27      schwarze 2799:
1.149     schwarze 2800:        return ROFF_IGN;
1.27      schwarze 2801: }
                   2802:
1.102     schwarze 2803: /*
                   2804:  * Handle in-line equation delimiters.
                   2805:  */
                   2806: static enum rofferr
1.110     schwarze 2807: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2808: {
1.105     schwarze 2809:        char            *cp1, *cp2;
                   2810:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2811:
1.102     schwarze 2812:        /*
                   2813:         * Outside equations, look for an opening delimiter.
                   2814:         * If we are inside an equation, we already know it is
                   2815:         * in-line, or this function wouldn't have been called;
                   2816:         * so look for a closing delimiter.
                   2817:         */
                   2818:
1.110     schwarze 2819:        cp1 = buf->buf + pos;
1.102     schwarze 2820:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2821:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2822:        if (cp2 == NULL)
1.149     schwarze 2823:                return ROFF_CONT;
1.102     schwarze 2824:
1.105     schwarze 2825:        *cp2++ = '\0';
                   2826:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2827:
                   2828:        /* Handle preceding text, protecting whitespace. */
                   2829:
1.110     schwarze 2830:        if (*buf->buf != '\0') {
1.105     schwarze 2831:                if (r->eqn == NULL)
                   2832:                        bef_pr = "\\&";
                   2833:                bef_nl = "\n";
                   2834:        }
                   2835:
                   2836:        /*
                   2837:         * Prepare replacing the delimiter with an equation macro
                   2838:         * and drop leading white space from the equation.
                   2839:         */
1.102     schwarze 2840:
1.105     schwarze 2841:        if (r->eqn == NULL) {
                   2842:                while (*cp2 == ' ')
                   2843:                        cp2++;
                   2844:                mac = ".EQ";
                   2845:        } else
                   2846:                mac = ".EN";
                   2847:
                   2848:        /* Handle following text, protecting whitespace. */
                   2849:
                   2850:        if (*cp2 != '\0') {
                   2851:                aft_nl = "\n";
                   2852:                if (r->eqn != NULL)
                   2853:                        aft_pr = "\\&";
                   2854:        }
                   2855:
                   2856:        /* Do the actual replacement. */
                   2857:
1.110     schwarze 2858:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2859:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2860:        free(buf->buf);
                   2861:        buf->buf = cp1;
1.102     schwarze 2862:
                   2863:        /* Toggle the in-line state of the eqn subsystem. */
                   2864:
                   2865:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2866:        return ROFF_REPARSE;
1.41      schwarze 2867: }
                   2868:
1.107     schwarze 2869: static enum rofferr
                   2870: roff_EQ(ROFF_ARGS)
1.32      schwarze 2871: {
1.41      schwarze 2872:        struct eqn_node *e;
1.32      schwarze 2873:
1.110     schwarze 2874:        assert(r->eqn == NULL);
1.107     schwarze 2875:        e = eqn_alloc(ppos, ln, r->parse);
1.32      schwarze 2876:
1.102     schwarze 2877:        if (r->last_eqn) {
1.32      schwarze 2878:                r->last_eqn->next = e;
1.102     schwarze 2879:                e->delim = r->last_eqn->delim;
                   2880:                e->odelim = r->last_eqn->odelim;
                   2881:                e->cdelim = r->last_eqn->cdelim;
                   2882:        } else
1.32      schwarze 2883:                r->first_eqn = r->last_eqn = e;
                   2884:
                   2885:        r->eqn = r->last_eqn = e;
1.41      schwarze 2886:
1.110     schwarze 2887:        if (buf->buf[pos] != '\0')
1.107     schwarze 2888:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2889:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2890:
1.149     schwarze 2891:        return ROFF_IGN;
1.32      schwarze 2892: }
                   2893:
                   2894: static enum rofferr
                   2895: roff_EN(ROFF_ARGS)
                   2896: {
                   2897:
1.94      schwarze 2898:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.149     schwarze 2899:        return ROFF_IGN;
1.32      schwarze 2900: }
                   2901:
                   2902: static enum rofferr
1.27      schwarze 2903: roff_TS(ROFF_ARGS)
                   2904: {
1.49      schwarze 2905:        struct tbl_node *tbl;
1.27      schwarze 2906:
                   2907:        if (r->tbl) {
1.94      schwarze 2908:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2909:                    ln, ppos, "TS breaks TS");
1.41      schwarze 2910:                tbl_end(&r->tbl);
1.27      schwarze 2911:        }
                   2912:
1.49      schwarze 2913:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2914:
                   2915:        if (r->last_tbl)
1.49      schwarze 2916:                r->last_tbl->next = tbl;
1.27      schwarze 2917:        else
1.49      schwarze 2918:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 2919:
1.49      schwarze 2920:        r->tbl = r->last_tbl = tbl;
1.169     schwarze 2921:        return ROFF_IGN;
                   2922: }
                   2923:
                   2924: static enum rofferr
                   2925: roff_onearg(ROFF_ARGS)
                   2926: {
                   2927:        struct roff_node        *n;
                   2928:        char                    *cp;
1.177     schwarze 2929:        int                      npos;
1.169     schwarze 2930:
1.174     schwarze 2931:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
                   2932:            (tok == ROFF_sp || tok == ROFF_ti))
                   2933:                man_breakscope(r->man, tok);
                   2934:
1.181     schwarze 2935:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177     schwarze 2936:                r->man->last = roffce_node;
                   2937:                r->man->next = ROFF_NEXT_SIBLING;
                   2938:        }
                   2939:
1.169     schwarze 2940:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2941:        n = r->man->last;
                   2942:
                   2943:        cp = buf->buf + pos;
                   2944:        if (*cp != '\0') {
                   2945:                while (*cp != '\0' && *cp != ' ')
                   2946:                        cp++;
                   2947:                while (*cp == ' ')
                   2948:                        *cp++ = '\0';
                   2949:                if (*cp != '\0')
                   2950:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS,
                   2951:                            r->parse, ln, cp - buf->buf,
                   2952:                            "%s ... %s", roff_name[tok], cp);
                   2953:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172     schwarze 2954:        }
                   2955:
1.181     schwarze 2956:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   2957:                if (r->man->last->type == ROFFT_ELEM) {
1.177     schwarze 2958:                        roff_word_alloc(r->man, ln, pos, "1");
                   2959:                        r->man->last->flags |= NODE_NOSRC;
                   2960:                }
                   2961:                npos = 0;
                   2962:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   2963:                    &roffce_lines, 0) == 0) {
                   2964:                        mandoc_vmsg(MANDOCERR_CE_NONUM,
                   2965:                            r->parse, ln, pos, "ce %s", buf->buf + pos);
                   2966:                        roffce_lines = 1;
                   2967:                }
                   2968:                if (roffce_lines < 1) {
                   2969:                        r->man->last = r->man->last->parent;
                   2970:                        roffce_node = NULL;
                   2971:                        roffce_lines = 0;
                   2972:                } else
                   2973:                        roffce_node = r->man->last->parent;
                   2974:        } else {
                   2975:                n->flags |= NODE_VALID | NODE_ENDED;
                   2976:                r->man->last = n;
                   2977:        }
                   2978:        n->flags |= NODE_LINE;
1.172     schwarze 2979:        r->man->next = ROFF_NEXT_SIBLING;
                   2980:        return ROFF_IGN;
                   2981: }
                   2982:
                   2983: static enum rofferr
                   2984: roff_manyarg(ROFF_ARGS)
                   2985: {
                   2986:        struct roff_node        *n;
                   2987:        char                    *sp, *ep;
                   2988:
                   2989:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2990:        n = r->man->last;
                   2991:
                   2992:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   2993:                while (*ep != '\0' && *ep != ' ')
                   2994:                        ep++;
                   2995:                while (*ep == ' ')
                   2996:                        *ep++ = '\0';
                   2997:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169     schwarze 2998:        }
                   2999:
                   3000:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3001:        r->man->last = n;
                   3002:        r->man->next = ROFF_NEXT_SIBLING;
1.149     schwarze 3003:        return ROFF_IGN;
1.123     schwarze 3004: }
                   3005:
                   3006: static enum rofferr
1.183     schwarze 3007: roff_als(ROFF_ARGS)
                   3008: {
                   3009:        char            *oldn, *newn, *end, *value;
                   3010:        size_t           oldsz, newsz, valsz;
                   3011:
                   3012:        newn = oldn = buf->buf + pos;
                   3013:        if (*newn == '\0')
                   3014:                return ROFF_IGN;
                   3015:
                   3016:        newsz = roff_getname(r, &oldn, ln, pos);
                   3017:        if (newn[newsz] == '\\' || *oldn == '\0')
                   3018:                return ROFF_IGN;
                   3019:
                   3020:        end = oldn;
                   3021:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3022:        if (oldsz == 0)
                   3023:                return ROFF_IGN;
                   3024:
1.187   ! schwarze 3025:        valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
        !          3026:            (int)oldsz, oldn);
1.183     schwarze 3027:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3028:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3029:        free(value);
                   3030:        return ROFF_IGN;
                   3031: }
                   3032:
                   3033: static enum rofferr
1.168     schwarze 3034: roff_br(ROFF_ARGS)
1.123     schwarze 3035: {
1.174     schwarze 3036:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3037:                man_breakscope(r->man, ROFF_br);
1.168     schwarze 3038:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
                   3039:        if (buf->buf[pos] != '\0')
                   3040:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   3041:                    "%s %s", roff_name[tok], buf->buf + pos);
                   3042:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3043:        r->man->next = ROFF_NEXT_SIBLING;
                   3044:        return ROFF_IGN;
1.27      schwarze 3045: }
                   3046:
                   3047: static enum rofferr
1.48      schwarze 3048: roff_cc(ROFF_ARGS)
                   3049: {
                   3050:        const char      *p;
                   3051:
1.110     schwarze 3052:        p = buf->buf + pos;
1.48      schwarze 3053:
1.110     schwarze 3054:        if (*p == '\0' || (r->control = *p++) == '.')
1.175     schwarze 3055:                r->control = '\0';
1.48      schwarze 3056:
1.110     schwarze 3057:        if (*p != '\0')
1.132     schwarze 3058:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3059:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 3060:
1.149     schwarze 3061:        return ROFF_IGN;
1.48      schwarze 3062: }
                   3063:
                   3064: static enum rofferr
1.175     schwarze 3065: roff_ec(ROFF_ARGS)
                   3066: {
                   3067:        const char      *p;
                   3068:
                   3069:        p = buf->buf + pos;
                   3070:        if (*p == '\0')
                   3071:                r->escape = '\\';
                   3072:        else {
                   3073:                r->escape = *p;
                   3074:                if (*++p != '\0')
                   3075:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3076:                            ln, p - buf->buf, "ec ... %s", p);
                   3077:        }
                   3078:        return ROFF_IGN;
                   3079: }
                   3080:
                   3081: static enum rofferr
                   3082: roff_eo(ROFF_ARGS)
                   3083: {
                   3084:        r->escape = '\0';
                   3085:        if (buf->buf[pos] != '\0')
                   3086:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
                   3087:                    ln, pos, "eo %s", buf->buf + pos);
                   3088:        return ROFF_IGN;
                   3089: }
                   3090:
                   3091: static enum rofferr
1.42      schwarze 3092: roff_tr(ROFF_ARGS)
                   3093: {
                   3094:        const char      *p, *first, *second;
                   3095:        size_t           fsz, ssz;
                   3096:        enum mandoc_esc  esc;
                   3097:
1.110     schwarze 3098:        p = buf->buf + pos;
1.42      schwarze 3099:
1.110     schwarze 3100:        if (*p == '\0') {
1.132     schwarze 3101:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 3102:                return ROFF_IGN;
1.42      schwarze 3103:        }
                   3104:
1.110     schwarze 3105:        while (*p != '\0') {
1.42      schwarze 3106:                fsz = ssz = 1;
                   3107:
                   3108:                first = p++;
1.110     schwarze 3109:                if (*first == '\\') {
1.42      schwarze 3110:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3111:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3112:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3113:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 3114:                                return ROFF_IGN;
1.42      schwarze 3115:                        }
                   3116:                        fsz = (size_t)(p - first);
                   3117:                }
                   3118:
                   3119:                second = p++;
1.110     schwarze 3120:                if (*second == '\\') {
1.42      schwarze 3121:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3122:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3123:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3124:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 3125:                                return ROFF_IGN;
1.42      schwarze 3126:                        }
                   3127:                        ssz = (size_t)(p - second);
1.110     schwarze 3128:                } else if (*second == '\0') {
1.132     schwarze 3129:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   3130:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 3131:                        second = " ";
                   3132:                        p--;
                   3133:                }
                   3134:
                   3135:                if (fsz > 1) {
1.80      schwarze 3136:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3137:                            second, ssz, 0);
1.42      schwarze 3138:                        continue;
                   3139:                }
                   3140:
1.110     schwarze 3141:                if (r->xtab == NULL)
1.80      schwarze 3142:                        r->xtab = mandoc_calloc(128,
                   3143:                            sizeof(struct roffstr));
1.42      schwarze 3144:
                   3145:                free(r->xtab[(int)*first].p);
                   3146:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3147:                r->xtab[(int)*first].sz = ssz;
                   3148:        }
                   3149:
1.149     schwarze 3150:        return ROFF_IGN;
1.42      schwarze 3151: }
                   3152:
                   3153: static enum rofferr
1.178     schwarze 3154: roff_rn(ROFF_ARGS)
                   3155: {
                   3156:        const char      *value;
                   3157:        char            *oldn, *newn, *end;
                   3158:        size_t           oldsz, newsz;
1.187   ! schwarze 3159:        int              deftype;
1.178     schwarze 3160:
                   3161:        oldn = newn = buf->buf + pos;
                   3162:        if (*oldn == '\0')
                   3163:                return ROFF_IGN;
                   3164:
                   3165:        oldsz = roff_getname(r, &newn, ln, pos);
                   3166:        if (oldn[oldsz] == '\\' || *newn == '\0')
                   3167:                return ROFF_IGN;
                   3168:
                   3169:        end = newn;
                   3170:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3171:        if (newsz == 0)
                   3172:                return ROFF_IGN;
                   3173:
1.187   ! schwarze 3174:        deftype = ROFFDEF_ANY;
        !          3175:        value = roff_getstrn(r, oldn, oldsz, &deftype);
        !          3176:        switch (deftype) {
        !          3177:        case ROFFDEF_USER:
1.178     schwarze 3178:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3179:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3180:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.187   ! schwarze 3181:                break;
        !          3182:        case ROFFDEF_PRE:
        !          3183:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
        !          3184:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
        !          3185:                break;
        !          3186:        case ROFFDEF_REN:
1.178     schwarze 3187:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3188:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.187   ! schwarze 3189:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
        !          3190:                break;
        !          3191:        case ROFFDEF_STD:
1.178     schwarze 3192:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.187   ! schwarze 3193:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
        !          3194:                break;
        !          3195:        default:
        !          3196:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
        !          3197:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
        !          3198:                break;
        !          3199:        }
1.178     schwarze 3200:        return ROFF_IGN;
                   3201: }
                   3202:
                   3203: static enum rofferr
1.14      schwarze 3204: roff_so(ROFF_ARGS)
                   3205: {
1.121     schwarze 3206:        char *name, *cp;
1.15      schwarze 3207:
1.110     schwarze 3208:        name = buf->buf + pos;
1.97      schwarze 3209:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 3210:
1.22      schwarze 3211:        /*
                   3212:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3213:         * opening anything that's not in our cwd or anything beneath
                   3214:         * it.  Thus, explicitly disallow traversing up the file-system
                   3215:         * or using absolute paths.
                   3216:         */
                   3217:
1.110     schwarze 3218:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 3219:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   3220:                    ".so %s", name);
1.121     schwarze 3221:                buf->sz = mandoc_asprintf(&cp,
                   3222:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3223:                free(buf->buf);
                   3224:                buf->buf = cp;
                   3225:                *offs = 0;
1.149     schwarze 3226:                return ROFF_REPARSE;
1.14      schwarze 3227:        }
                   3228:
                   3229:        *offs = pos;
1.149     schwarze 3230:        return ROFF_SO;
1.7       schwarze 3231: }
                   3232:
1.138     schwarze 3233: /* --- user defined strings and macros ------------------------------------ */
                   3234:
1.16      schwarze 3235: static enum rofferr
                   3236: roff_userdef(ROFF_ARGS)
1.12      schwarze 3237: {
1.183     schwarze 3238:        const char       *arg[16], *ap;
1.16      schwarze 3239:        char             *cp, *n1, *n2;
1.164     schwarze 3240:        int               expand_count, i, ib, ie;
1.135     schwarze 3241:        size_t            asz, rsz;
1.12      schwarze 3242:
1.16      schwarze 3243:        /*
                   3244:         * Collect pointers to macro argument strings
1.61      schwarze 3245:         * and NUL-terminate them.
1.16      schwarze 3246:         */
1.135     schwarze 3247:
1.145     schwarze 3248:        r->argc = 0;
1.110     schwarze 3249:        cp = buf->buf + pos;
1.183     schwarze 3250:        for (i = 0; i < 16; i++) {
1.145     schwarze 3251:                if (*cp == '\0')
                   3252:                        arg[i] = "";
                   3253:                else {
                   3254:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   3255:                        r->argc = i + 1;
                   3256:                }
                   3257:        }
1.16      schwarze 3258:
                   3259:        /*
                   3260:         * Expand macro arguments.
1.12      schwarze 3261:         */
1.135     schwarze 3262:
                   3263:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 3264:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 3265:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 3266:        expand_count = 0;
1.135     schwarze 3267:        while (*cp != '\0') {
                   3268:
                   3269:                /* Scan ahead for the next argument invocation. */
                   3270:
                   3271:                if (*cp++ != '\\')
                   3272:                        continue;
                   3273:                if (*cp++ != '$')
                   3274:                        continue;
1.146     schwarze 3275:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   3276:                        ib = 0;
                   3277:                        ie = r->argc - 1;
                   3278:                } else {  /* \\$1 .. \\$9 insert one argument */
                   3279:                        ib = ie = *cp - '1';
                   3280:                        if (ib < 0 || ib > 8)
                   3281:                                continue;
                   3282:                }
1.135     schwarze 3283:                cp -= 2;
1.164     schwarze 3284:
                   3285:                /*
                   3286:                 * Prevent infinite recursion.
                   3287:                 */
                   3288:
                   3289:                if (cp >= n2)
                   3290:                        expand_count = 1;
                   3291:                else if (++expand_count > EXPAND_LIMIT) {
                   3292:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   3293:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 3294:                        free(buf->buf);
                   3295:                        buf->buf = n1;
1.164     schwarze 3296:                        return ROFF_IGN;
                   3297:                }
1.135     schwarze 3298:
                   3299:                /*
                   3300:                 * Determine the size of the expanded argument,
                   3301:                 * taking escaping of quotes into account.
                   3302:                 */
                   3303:
1.146     schwarze 3304:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   3305:                for (i = ib; i <= ie; i++) {
                   3306:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3307:                                asz++;
                   3308:                                if (*ap == '"')
                   3309:                                        asz += 3;
                   3310:                        }
1.135     schwarze 3311:                }
                   3312:                if (asz != 3) {
                   3313:
                   3314:                        /*
                   3315:                         * Determine the size of the rest of the
                   3316:                         * unexpanded macro, including the NUL.
                   3317:                         */
                   3318:
                   3319:                        rsz = buf->sz - (cp - n1) - 3;
                   3320:
                   3321:                        /*
                   3322:                         * When shrinking, move before
                   3323:                         * releasing the storage.
                   3324:                         */
                   3325:
                   3326:                        if (asz < 3)
                   3327:                                memmove(cp + asz, cp + 3, rsz);
                   3328:
                   3329:                        /*
                   3330:                         * Resize the storage for the macro
                   3331:                         * and readjust the parse pointer.
                   3332:                         */
                   3333:
                   3334:                        buf->sz += asz - 3;
                   3335:                        n2 = mandoc_realloc(n1, buf->sz);
                   3336:                        cp = n2 + (cp - n1);
                   3337:                        n1 = n2;
                   3338:
                   3339:                        /*
                   3340:                         * When growing, make room
                   3341:                         * for the expanded argument.
                   3342:                         */
                   3343:
                   3344:                        if (asz > 3)
                   3345:                                memmove(cp + asz, cp + 3, rsz);
                   3346:                }
                   3347:
                   3348:                /* Copy the expanded argument, escaping quotes. */
                   3349:
                   3350:                n2 = cp;
1.146     schwarze 3351:                for (i = ib; i <= ie; i++) {
                   3352:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3353:                                if (*ap == '"') {
                   3354:                                        memcpy(n2, "\\(dq", 4);
                   3355:                                        n2 += 4;
                   3356:                                } else
                   3357:                                        *n2++ = *ap;
                   3358:                        }
                   3359:                        if (i < ie)
                   3360:                                *n2++ = ' ';
1.16      schwarze 3361:                }
1.12      schwarze 3362:        }
                   3363:
1.16      schwarze 3364:        /*
                   3365:         * Replace the macro invocation
                   3366:         * by the expanded macro.
                   3367:         */
1.135     schwarze 3368:
1.110     schwarze 3369:        free(buf->buf);
                   3370:        buf->buf = n1;
1.120     schwarze 3371:        *offs = 0;
1.16      schwarze 3372:
1.149     schwarze 3373:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3374:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3375: }
1.28      schwarze 3376:
1.178     schwarze 3377: /*
                   3378:  * Calling a high-level macro that was renamed with .rn.
                   3379:  * r->current_string has already been set up by roff_parse().
                   3380:  */
                   3381: static enum rofferr
                   3382: roff_renamed(ROFF_ARGS)
                   3383: {
                   3384:        char    *nbuf;
                   3385:
1.187   ! schwarze 3386:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
        !          3387:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.178     schwarze 3388:        free(buf->buf);
                   3389:        buf->buf = nbuf;
                   3390:        return ROFF_CONT;
                   3391: }
                   3392:
1.85      schwarze 3393: static size_t
1.28      schwarze 3394: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3395: {
                   3396:        char     *name, *cp;
1.85      schwarze 3397:        size_t    namesz;
1.28      schwarze 3398:
                   3399:        name = *cpp;
                   3400:        if ('\0' == *name)
1.149     schwarze 3401:                return 0;
1.28      schwarze 3402:
1.85      schwarze 3403:        /* Read until end of name and terminate it with NUL. */
                   3404:        for (cp = name; 1; cp++) {
                   3405:                if ('\0' == *cp || ' ' == *cp) {
                   3406:                        namesz = cp - name;
                   3407:                        break;
                   3408:                }
1.28      schwarze 3409:                if ('\\' != *cp)
                   3410:                        continue;
1.88      schwarze 3411:                namesz = cp - name;
                   3412:                if ('{' == cp[1] || '}' == cp[1])
                   3413:                        break;
1.28      schwarze 3414:                cp++;
                   3415:                if ('\\' == *cp)
                   3416:                        continue;
1.97      schwarze 3417:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3418:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3419:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3420:                break;
1.28      schwarze 3421:        }
                   3422:
                   3423:        /* Read past spaces. */
                   3424:        while (' ' == *cp)
                   3425:                cp++;
                   3426:
                   3427:        *cpp = cp;
1.149     schwarze 3428:        return namesz;
1.28      schwarze 3429: }
                   3430:
1.16      schwarze 3431: /*
                   3432:  * Store *string into the user-defined string called *name.
                   3433:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3434:  * append == 0: replace mode
                   3435:  * append == 1: single-line append mode
                   3436:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3437:  */
1.8       schwarze 3438: static void
1.16      schwarze 3439: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3440:        int append)
1.7       schwarze 3441: {
1.187   ! schwarze 3442:        size_t   namesz;
1.42      schwarze 3443:
1.187   ! schwarze 3444:        namesz = strlen(name);
        !          3445:        roff_setstrn(&r->strtab, name, namesz, string,
1.80      schwarze 3446:            string ? strlen(string) : 0, append);
1.187   ! schwarze 3447:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.42      schwarze 3448: }
                   3449:
                   3450: static void
                   3451: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3452:                const char *string, size_t stringsz, int append)
1.42      schwarze 3453: {
                   3454:        struct roffkv   *n;
                   3455:        char            *c;
                   3456:        int              i;
                   3457:        size_t           oldch, newch;
1.7       schwarze 3458:
1.16      schwarze 3459:        /* Search for an existing string with the same name. */
1.42      schwarze 3460:        n = *r;
                   3461:
1.84      schwarze 3462:        while (n && (namesz != n->key.sz ||
                   3463:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3464:                n = n->next;
1.8       schwarze 3465:
                   3466:        if (NULL == n) {
1.16      schwarze 3467:                /* Create a new string table entry. */
1.42      schwarze 3468:                n = mandoc_malloc(sizeof(struct roffkv));
                   3469:                n->key.p = mandoc_strndup(name, namesz);
                   3470:                n->key.sz = namesz;
                   3471:                n->val.p = NULL;
                   3472:                n->val.sz = 0;
                   3473:                n->next = *r;
                   3474:                *r = n;
1.66      schwarze 3475:        } else if (0 == append) {
1.42      schwarze 3476:                free(n->val.p);
                   3477:                n->val.p = NULL;
                   3478:                n->val.sz = 0;
1.16      schwarze 3479:        }
                   3480:
                   3481:        if (NULL == string)
                   3482:                return;
                   3483:
                   3484:        /*
                   3485:         * One additional byte for the '\n' in multiline mode,
                   3486:         * and one for the terminating '\0'.
                   3487:         */
1.66      schwarze 3488:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3489:
                   3490:        if (NULL == n->val.p) {
                   3491:                n->val.p = mandoc_malloc(newch);
                   3492:                *n->val.p = '\0';
1.16      schwarze 3493:                oldch = 0;
                   3494:        } else {
1.42      schwarze 3495:                oldch = n->val.sz;
                   3496:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3497:        }
                   3498:
                   3499:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3500:        c = n->val.p + (int)oldch;
1.16      schwarze 3501:
                   3502:        /* Append new content to the destination buffer. */
1.42      schwarze 3503:        i = 0;
                   3504:        while (i < (int)stringsz) {
1.16      schwarze 3505:                /*
                   3506:                 * Rudimentary roff copy mode:
                   3507:                 * Handle escaped backslashes.
                   3508:                 */
1.42      schwarze 3509:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3510:                        i++;
                   3511:                *c++ = string[i++];
1.16      schwarze 3512:        }
1.8       schwarze 3513:
1.16      schwarze 3514:        /* Append terminating bytes. */
1.66      schwarze 3515:        if (1 < append)
1.16      schwarze 3516:                *c++ = '\n';
1.42      schwarze 3517:
1.16      schwarze 3518:        *c = '\0';
1.42      schwarze 3519:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3520: }
                   3521:
1.8       schwarze 3522: static const char *
1.187   ! schwarze 3523: roff_getstrn(const struct roff *r, const char *name, size_t len,
        !          3524:     int *deftype)
1.7       schwarze 3525: {
1.187   ! schwarze 3526:        const struct roffkv     *n;
        !          3527:        int                      i;
        !          3528:        enum roff_tok            tok;
        !          3529:
        !          3530:        if (*deftype & ROFFDEF_USER) {
        !          3531:                for (n = r->strtab; n != NULL; n = n->next) {
        !          3532:                        if (strncmp(name, n->key.p, len) == 0 &&
        !          3533:                            n->key.p[len] == '\0' &&
        !          3534:                            n->val.p != NULL) {
        !          3535:                                *deftype = ROFFDEF_USER;
        !          3536:                                return n->val.p;
        !          3537:                        }
        !          3538:                }
        !          3539:        }
        !          3540:        if (*deftype & ROFFDEF_PRE) {
        !          3541:                for (i = 0; i < PREDEFS_MAX; i++) {
        !          3542:                        if (strncmp(name, predefs[i].name, len) == 0 &&
        !          3543:                            predefs[i].name[len] == '\0') {
        !          3544:                                *deftype = ROFFDEF_PRE;
        !          3545:                                return predefs[i].str;
        !          3546:                        }
        !          3547:                }
        !          3548:        }
        !          3549:        if (*deftype & ROFFDEF_REN) {
        !          3550:                for (n = r->rentab; n != NULL; n = n->next) {
        !          3551:                        if (strncmp(name, n->key.p, len) == 0 &&
        !          3552:                            n->key.p[len] == '\0' &&
        !          3553:                            n->val.p != NULL) {
        !          3554:                                *deftype = ROFFDEF_REN;
        !          3555:                                return n->val.p;
        !          3556:                        }
        !          3557:                }
        !          3558:        }
        !          3559:        if (*deftype & ROFFDEF_STD) {
        !          3560:                if (r->man->macroset != MACROSET_MAN) {
        !          3561:                        for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
        !          3562:                                if (strncmp(name, roff_name[tok], len) == 0 &&
        !          3563:                                    roff_name[tok][len] == '\0') {
        !          3564:                                        *deftype = ROFFDEF_STD;
        !          3565:                                        return NULL;
        !          3566:                                }
        !          3567:                        }
        !          3568:                }
        !          3569:                if (r->man->macroset != MACROSET_MDOC) {
        !          3570:                        for (tok = MAN_TH; tok < MAN_MAX; tok++) {
        !          3571:                                if (strncmp(name, roff_name[tok], len) == 0 &&
        !          3572:                                    roff_name[tok][len] == '\0') {
        !          3573:                                        *deftype = ROFFDEF_STD;
        !          3574:                                        return NULL;
        !          3575:                                }
        !          3576:                        }
        !          3577:                }
        !          3578:        }
        !          3579:        *deftype = 0;
1.149     schwarze 3580:        return NULL;
1.7       schwarze 3581: }
                   3582:
1.8       schwarze 3583: static void
1.42      schwarze 3584: roff_freestr(struct roffkv *r)
1.7       schwarze 3585: {
1.42      schwarze 3586:        struct roffkv    *n, *nn;
1.7       schwarze 3587:
1.42      schwarze 3588:        for (n = r; n; n = nn) {
                   3589:                free(n->key.p);
                   3590:                free(n->val.p);
1.7       schwarze 3591:                nn = n->next;
                   3592:                free(n);
                   3593:        }
1.27      schwarze 3594: }
1.138     schwarze 3595:
                   3596: /* --- accessors and utility functions ------------------------------------ */
1.27      schwarze 3597:
                   3598: const struct tbl_span *
                   3599: roff_span(const struct roff *r)
                   3600: {
1.80      schwarze 3601:
1.149     schwarze 3602:        return r->tbl ? tbl_span(r->tbl) : NULL;
1.32      schwarze 3603: }
                   3604:
                   3605: const struct eqn *
                   3606: roff_eqn(const struct roff *r)
                   3607: {
1.80      schwarze 3608:
1.149     schwarze 3609:        return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.42      schwarze 3610: }
                   3611:
                   3612: /*
                   3613:  * Duplicate an input string, making the appropriate character
                   3614:  * conversations (as stipulated by `tr') along the way.
                   3615:  * Returns a heap-allocated string with all the replacements made.
                   3616:  */
                   3617: char *
                   3618: roff_strdup(const struct roff *r, const char *p)
                   3619: {
                   3620:        const struct roffkv *cp;
                   3621:        char            *res;
                   3622:        const char      *pp;
                   3623:        size_t           ssz, sz;
                   3624:        enum mandoc_esc  esc;
                   3625:
                   3626:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3627:                return mandoc_strdup(p);
1.42      schwarze 3628:        else if ('\0' == *p)
1.149     schwarze 3629:                return mandoc_strdup("");
1.42      schwarze 3630:
                   3631:        /*
                   3632:         * Step through each character looking for term matches
                   3633:         * (remember that a `tr' can be invoked with an escape, which is
                   3634:         * a glyph but the escape is multi-character).
                   3635:         * We only do this if the character hash has been initialised
                   3636:         * and the string is >0 length.
                   3637:         */
                   3638:
                   3639:        res = NULL;
                   3640:        ssz = 0;
                   3641:
                   3642:        while ('\0' != *p) {
1.161     schwarze 3643:                assert((unsigned int)*p < 128);
                   3644:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3645:                        sz = r->xtab[(int)*p].sz;
                   3646:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3647:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3648:                        ssz += sz;
                   3649:                        p++;
                   3650:                        continue;
                   3651:                } else if ('\\' != *p) {
                   3652:                        res = mandoc_realloc(res, ssz + 2);
                   3653:                        res[ssz++] = *p++;
                   3654:                        continue;
                   3655:                }
                   3656:
                   3657:                /* Search for term matches. */
                   3658:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3659:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3660:                                break;
                   3661:
                   3662:                if (NULL != cp) {
                   3663:                        /*
                   3664:                         * A match has been found.
                   3665:                         * Append the match to the array and move
                   3666:                         * forward by its keysize.
                   3667:                         */
1.80      schwarze 3668:                        res = mandoc_realloc(res,
                   3669:                            ssz + cp->val.sz + 1);
1.42      schwarze 3670:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3671:                        ssz += cp->val.sz;
                   3672:                        p += (int)cp->key.sz;
                   3673:                        continue;
                   3674:                }
                   3675:
                   3676:                /*
                   3677:                 * Handle escapes carefully: we need to copy
                   3678:                 * over just the escape itself, or else we might
                   3679:                 * do replacements within the escape itself.
                   3680:                 * Make sure to pass along the bogus string.
                   3681:                 */
                   3682:                pp = p++;
                   3683:                esc = mandoc_escape(&p, NULL, NULL);
                   3684:                if (ESCAPE_ERROR == esc) {
                   3685:                        sz = strlen(pp);
                   3686:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3687:                        memcpy(res + ssz, pp, sz);
                   3688:                        break;
                   3689:                }
1.80      schwarze 3690:                /*
                   3691:                 * We bail out on bad escapes.
1.42      schwarze 3692:                 * No need to warn: we already did so when
                   3693:                 * roff_res() was called.
                   3694:                 */
                   3695:                sz = (int)(p - pp);
                   3696:                res = mandoc_realloc(res, ssz + sz + 1);
                   3697:                memcpy(res + ssz, pp, sz);
                   3698:                ssz += sz;
                   3699:        }
                   3700:
                   3701:        res[(int)ssz] = '\0';
1.149     schwarze 3702:        return res;
1.99      schwarze 3703: }
                   3704:
                   3705: int
                   3706: roff_getformat(const struct roff *r)
                   3707: {
                   3708:
1.149     schwarze 3709:        return r->format;
1.48      schwarze 3710: }
                   3711:
                   3712: /*
1.80      schwarze 3713:  * Find out whether a line is a macro line or not.
1.48      schwarze 3714:  * If it is, adjust the current position and return one; if it isn't,
                   3715:  * return zero and don't change the current position.
                   3716:  * If the control character has been set with `.cc', then let that grain
                   3717:  * precedence.
                   3718:  * This is slighly contrary to groff, where using the non-breaking
                   3719:  * control character when `cc' has been invoked will cause the
                   3720:  * non-breaking macro contents to be printed verbatim.
                   3721:  */
                   3722: int
                   3723: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3724: {
                   3725:        int             pos;
                   3726:
                   3727:        pos = *ppos;
                   3728:
1.175     schwarze 3729:        if (r->control != '\0' && cp[pos] == r->control)
1.48      schwarze 3730:                pos++;
1.175     schwarze 3731:        else if (r->control != '\0')
1.149     schwarze 3732:                return 0;
1.48      schwarze 3733:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3734:                pos += 2;
                   3735:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3736:                pos++;
                   3737:        else
1.149     schwarze 3738:                return 0;
1.48      schwarze 3739:
                   3740:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3741:                pos++;
                   3742:
                   3743:        *ppos = pos;
1.149     schwarze 3744:        return 1;
1.1       schwarze 3745: }