[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 1.198

1.198   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.197 2018/04/09 02:31:37 schwarze Exp $ */
1.1       schwarze    2: /*
1.139     schwarze    3:  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.197     schwarze    4:  * Copyright (c) 2010-2015, 2017, 2018 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)
1.197     schwarze   47: #define        ROFFDEF_UNDEF   (1 << 5)  /* Completely undefined. */
1.187     schwarze   48:
1.138     schwarze   49: /* --- data types --------------------------------------------------------- */
                     50:
1.41      schwarze   51: /*
1.42      schwarze   52:  * An incredibly-simple string buffer.
                     53:  */
1.8       schwarze   54: struct roffstr {
1.42      schwarze   55:        char            *p; /* nil-terminated buffer */
                     56:        size_t           sz; /* saved strlen(p) */
                     57: };
                     58:
                     59: /*
                     60:  * A key-value roffstr pair as part of a singly-linked list.
                     61:  */
                     62: struct roffkv {
                     63:        struct roffstr   key;
                     64:        struct roffstr   val;
                     65:        struct roffkv   *next; /* next in list */
1.8       schwarze   66: };
                     67:
1.52      schwarze   68: /*
                     69:  * A single number register as part of a singly-linked list.
                     70:  */
                     71: struct roffreg {
                     72:        struct roffstr   key;
1.53      schwarze   73:        int              val;
1.52      schwarze   74:        struct roffreg  *next;
                     75: };
                     76:
1.167     schwarze   77: /*
                     78:  * Association of request and macro names with token IDs.
                     79:  */
                     80: struct roffreq {
                     81:        enum roff_tok    tok;
                     82:        char             name[];
                     83: };
                     84:
1.1       schwarze   85: struct roff {
1.35      schwarze   86:        struct mparse   *parse; /* parse point */
1.168     schwarze   87:        struct roff_man *man; /* mdoc or man parser */
1.1       schwarze   88:        struct roffnode *last; /* leaf of stack */
1.96      schwarze   89:        int             *rstack; /* stack of inverted `ie' values */
1.167     schwarze   90:        struct ohash    *reqtab; /* request lookup table */
1.52      schwarze   91:        struct roffreg  *regtab; /* number registers */
1.42      schwarze   92:        struct roffkv   *strtab; /* user-defined strings & macros */
1.178     schwarze   93:        struct roffkv   *rentab; /* renamed strings & macros */
1.42      schwarze   94:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     95:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze   96:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   97:        struct tbl_node *first_tbl; /* first table parsed */
                     98:        struct tbl_node *last_tbl; /* last table parsed */
                     99:        struct tbl_node *tbl; /* current table being parsed */
1.191     schwarze  100:        struct eqn_node *last_eqn; /* equation parser */
                    101:        struct eqn_node *eqn; /* active equation parser */
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);
1.193     schwarze  155: static void             roff_addtbl(struct roff_man *, struct tbl_node *);
1.183     schwarze  156: static enum rofferr     roff_als(ROFF_ARGS);
1.2       schwarze  157: static enum rofferr     roff_block(ROFF_ARGS);
                    158: static enum rofferr     roff_block_text(ROFF_ARGS);
                    159: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.168     schwarze  160: static enum rofferr     roff_br(ROFF_ARGS);
1.2       schwarze  161: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  162: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  163: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  164: static enum rofferr     roff_cond(ROFF_ARGS);
                    165: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    166: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  167: static enum rofferr     roff_ds(ROFF_ARGS);
1.175     schwarze  168: static enum rofferr     roff_ec(ROFF_ARGS);
                    169: static enum rofferr     roff_eo(ROFF_ARGS);
1.110     schwarze  170: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.143     schwarze  171: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.106     schwarze  172: static int              roff_evalnum(struct roff *, int,
                    173:                                const char *, int *, int *, int);
                    174: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  175:                                const char *, int *, int *, int);
1.71      schwarze  176: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  177: static void             roff_free1(struct roff *);
1.52      schwarze  178: static void             roff_freereg(struct roffreg *);
1.42      schwarze  179: static void             roff_freestr(struct roffkv *);
1.85      schwarze  180: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  181: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  182: static int              roff_getop(const char *, int *, char *);
1.198   ! schwarze  183: static int              roff_getregn(struct roff *, const char *, size_t);
1.145     schwarze  184: static int              roff_getregro(const struct roff *,
                    185:                                const char *name);
1.197     schwarze  186: static const char      *roff_getstrn(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);
1.198   ! schwarze  207: static void             roff_setregn(struct roff *, const char *,
        !           208:                                size_t, int, char);
1.8       schwarze  209: static void             roff_setstr(struct roff *,
1.16      schwarze  210:                                const char *, const char *, int);
1.80      schwarze  211: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  212:                                size_t, const char *, size_t, int);
1.14      schwarze  213: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  214: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  215: static enum rofferr     roff_Dd(ROFF_ARGS);
1.27      schwarze  216: static enum rofferr     roff_TE(ROFF_ARGS);
                    217: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  218: static enum rofferr     roff_EQ(ROFF_ARGS);
                    219: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  220: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  221: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  222: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  223:
1.138     schwarze  224: /* --- constant data ------------------------------------------------------ */
                    225:
1.133     schwarze  226: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    227: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    228:
1.166     schwarze  229: const char *__roff_name[MAN_MAX + 1] = {
1.177     schwarze  230:        "br",           "ce",           "ft",           "ll",
1.184     schwarze  231:        "mc",           "po",           "rj",           "sp",
                    232:        "ta",           "ti",           NULL,
1.166     schwarze  233:        "ab",           "ad",           "af",           "aln",
                    234:        "als",          "am",           "am1",          "ami",
                    235:        "ami1",         "as",           "as1",          "asciify",
                    236:        "backtrace",    "bd",           "bleedat",      "blm",
                    237:         "box",         "boxa",         "bp",           "BP",
                    238:        "break",        "breakchar",    "brnl",         "brp",
1.177     schwarze  239:        "brpnl",        "c2",           "cc",
1.166     schwarze  240:        "cf",           "cflags",       "ch",           "char",
                    241:        "chop",         "class",        "close",        "CL",
                    242:        "color",        "composite",    "continue",     "cp",
                    243:        "cropat",       "cs",           "cu",           "da",
                    244:        "dch",          "Dd",           "de",           "de1",
                    245:        "defcolor",     "dei",          "dei1",         "device",
                    246:        "devicem",      "di",           "do",           "ds",
                    247:        "ds1",          "dwh",          "dt",           "ec",
                    248:        "ecr",          "ecs",          "el",           "em",
                    249:        "EN",           "eo",           "EP",           "EQ",
                    250:        "errprint",     "ev",           "evc",          "ex",
                    251:        "fallback",     "fam",          "fc",           "fchar",
                    252:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    253:        "fl",           "flig",         "fp",           "fps",
                    254:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    255:        "fzoom",        "gcolor",       "hc",           "hcode",
                    256:        "hidechar",     "hla",          "hlm",          "hpf",
                    257:        "hpfa",         "hpfcode",      "hw",           "hy",
                    258:        "hylang",       "hylen",        "hym",          "hypp",
                    259:        "hys",          "ie",           "if",           "ig",
                    260:        "index",        "it",           "itc",          "IX",
                    261:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    262:        "lc",           "lc_ctype",     "lds",          "length",
                    263:        "letadj",       "lf",           "lg",           "lhang",
                    264:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.176     schwarze  265:        "ls",           "lsm",          "lt",
1.166     schwarze  266:        "mediasize",    "minss",        "mk",           "mso",
                    267:        "na",           "ne",           "nh",           "nhychar",
                    268:        "nm",           "nn",           "nop",          "nr",
                    269:        "nrf",          "nroff",        "ns",           "nx",
                    270:        "open",         "opena",        "os",           "output",
                    271:        "padj",         "papersize",    "pc",           "pev",
                    272:        "pi",           "PI",           "pl",           "pm",
1.184     schwarze  273:        "pn",           "pnr",          "ps",
1.166     schwarze  274:        "psbb",         "pshape",       "pso",          "ptr",
                    275:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.181     schwarze  276:        "return",       "rfschar",      "rhang",
1.166     schwarze  277:        "rm",           "rn",           "rnn",          "rr",
                    278:        "rs",           "rt",           "schar",        "sentchar",
                    279:        "shc",          "shift",        "sizes",        "so",
                    280:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    281:        "sty",          "substring",    "sv",           "sy",
1.172     schwarze  282:        "T&",           "tc",           "TE",
1.173     schwarze  283:        "TH",           "tkf",          "tl",
1.166     schwarze  284:        "tm",           "tm1",          "tmc",          "tr",
                    285:        "track",        "transchar",    "trf",          "trimat",
                    286:        "trin",         "trnt",         "troff",        "TS",
                    287:        "uf",           "ul",           "unformat",     "unwatch",
                    288:        "unwatchn",     "vpt",          "vs",           "warn",
                    289:        "warnscale",    "watch",        "watchlength",  "watchn",
                    290:        "wh",           "while",        "write",        "writec",
                    291:        "writem",       "xflag",        ".",            NULL,
1.178     schwarze  292:        NULL,           "text",
1.166     schwarze  293:        "Dd",           "Dt",           "Os",           "Sh",
                    294:        "Ss",           "Pp",           "D1",           "Dl",
                    295:        "Bd",           "Ed",           "Bl",           "El",
                    296:        "It",           "Ad",           "An",           "Ap",
                    297:        "Ar",           "Cd",           "Cm",           "Dv",
                    298:        "Er",           "Ev",           "Ex",           "Fa",
                    299:        "Fd",           "Fl",           "Fn",           "Ft",
                    300:        "Ic",           "In",           "Li",           "Nd",
                    301:        "Nm",           "Op",           "Ot",           "Pa",
                    302:        "Rv",           "St",           "Va",           "Vt",
                    303:        "Xr",           "%A",           "%B",           "%D",
                    304:        "%I",           "%J",           "%N",           "%O",
                    305:        "%P",           "%R",           "%T",           "%V",
                    306:        "Ac",           "Ao",           "Aq",           "At",
                    307:        "Bc",           "Bf",           "Bo",           "Bq",
                    308:        "Bsx",          "Bx",           "Db",           "Dc",
                    309:        "Do",           "Dq",           "Ec",           "Ef",
                    310:        "Em",           "Eo",           "Fx",           "Ms",
                    311:        "No",           "Ns",           "Nx",           "Ox",
                    312:        "Pc",           "Pf",           "Po",           "Pq",
                    313:        "Qc",           "Ql",           "Qo",           "Qq",
                    314:        "Re",           "Rs",           "Sc",           "So",
                    315:        "Sq",           "Sm",           "Sx",           "Sy",
                    316:        "Tn",           "Ux",           "Xc",           "Xo",
                    317:        "Fo",           "Fc",           "Oo",           "Oc",
                    318:        "Bk",           "Ek",           "Bt",           "Hf",
                    319:        "Fr",           "Ud",           "Lb",           "Lp",
                    320:        "Lk",           "Mt",           "Brq",          "Bro",
                    321:        "Brc",          "%C",           "Es",           "En",
1.171     schwarze  322:        "Dx",           "%Q",           "%U",           "Ta",
                    323:        NULL,
1.166     schwarze  324:        "TH",           "SH",           "SS",           "TP",
                    325:        "LP",           "PP",           "P",            "IP",
                    326:        "HP",           "SM",           "SB",           "BI",
                    327:        "IB",           "BR",           "RB",           "R",
                    328:        "B",            "I",            "IR",           "RI",
1.171     schwarze  329:        "nf",           "fi",
1.166     schwarze  330:        "RE",           "RS",           "DT",           "UC",
1.169     schwarze  331:        "PD",           "AT",           "in",
1.166     schwarze  332:        "OP",           "EX",           "EE",           "UR",
1.189     bentley   333:        "UE",           "MT",           "ME",           NULL
1.166     schwarze  334: };
                    335: const  char *const *roff_name = __roff_name;
                    336:
                    337: static struct roffmac   roffs[TOKEN_NONE] = {
1.168     schwarze  338:        { roff_br, NULL, NULL, 0 },  /* br */
1.177     schwarze  339:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.169     schwarze  340:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.170     schwarze  341:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.176     schwarze  342:        { roff_onearg, NULL, NULL, 0 },  /* mc */
1.184     schwarze  343:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.181     schwarze  344:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.171     schwarze  345:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.172     schwarze  346:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.173     schwarze  347:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.168     schwarze  348:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.167     schwarze  349:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    350:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    351:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    352:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.183     schwarze  353:        { roff_als, NULL, NULL, 0 },  /* als */
1.167     schwarze  354:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    355:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    356:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    357:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    358:        { roff_ds, NULL, NULL, 0 },  /* as */
                    359:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    360:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    361:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    362:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    363:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    364:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    365:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    366:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    367:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    368:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    369:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    370:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    371:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.168     schwarze  372:        { roff_br, NULL, NULL, 0 },  /* brp */
1.167     schwarze  373:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    374:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    375:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    376:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    377:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    378:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    379:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    380:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    381:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    382:        { roff_insec, NULL, NULL, 0 },  /* close */
                    383:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    384:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    385:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    386:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    387:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    388:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    389:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    390:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    391:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    392:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    393:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    394:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    395:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    396:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    397:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    398:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    400:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    401:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    402:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    403:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    404:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    405:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    406:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.175     schwarze  407:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.167     schwarze  408:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    409:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    410:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    411:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    412:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.175     schwarze  413:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.167     schwarze  414:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    415:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    416:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    417:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    418:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    419:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    421:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    422:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    423:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    425:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    429:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    430:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    431:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    432:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    438:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    439:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    440:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    442:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    444:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    445:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    446:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    447:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    448:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    449:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    450:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    451:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    452:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    453:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    454:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    455:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    456:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    457:        { roff_it, NULL, NULL, 0 },  /* it */
                    458:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    461:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    462:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    463:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    464:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    465:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    466:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    467:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    468:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    469:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    470:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    471:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    472:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    473:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    474:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    475:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    477:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    482:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    487:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    488:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    489:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    490:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    491:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    492:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    493:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    494:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    495:        { roff_insec, NULL, NULL, 0 },  /* open */
                    496:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    497:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    498:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    500:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    501:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    503:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    504:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    505:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    506:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    508:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    510:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    511:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    512:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    513:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    514:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    515:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    516:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    517:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    518:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    519:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    520:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    521:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.178     schwarze  522:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.167     schwarze  523:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    524:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    526:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    527:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    529:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    530:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    532:        { roff_so, NULL, NULL, 0 },  /* so */
                    533:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    534:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    535:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    536:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    537:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    538:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    539:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    540:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    541:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    542:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    543:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.187     schwarze  544:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.167     schwarze  545:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    546:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    549:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    550:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    551:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    552:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    553:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    554:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    555:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    556:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    557:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    558:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    559:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    560:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    561:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    562:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    563:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    565:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    566:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    568:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    569:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    570:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    571:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    572:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    573:        { roff_insec, NULL, NULL, 0 },  /* write */
                    574:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    575:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    576:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    577:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.178     schwarze  578:        { roff_renamed, NULL, NULL, 0 },
1.167     schwarze  579:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  580: };
                    581:
1.37      schwarze  582: /* Array of injected predefined strings. */
                    583: #define        PREDEFS_MAX      38
                    584: static const struct predef predefs[PREDEFS_MAX] = {
                    585: #include "predefs.in"
                    586: };
                    587:
1.177     schwarze  588: static int      roffce_lines;  /* number of input lines to center */
                    589: static struct roff_node *roffce_node;  /* active request */
1.51      schwarze  590: static int      roffit_lines;  /* number of lines to delay */
                    591: static char    *roffit_macro;  /* nil-terminated macro line */
                    592:
1.80      schwarze  593:
1.138     schwarze  594: /* --- request table ------------------------------------------------------ */
                    595:
1.167     schwarze  596: struct ohash *
                    597: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  598: {
1.167     schwarze  599:        struct ohash    *htab;
                    600:        struct roffreq  *req;
                    601:        enum roff_tok    tok;
                    602:        size_t           sz;
                    603:        unsigned int     slot;
1.3       schwarze  604:
1.167     schwarze  605:        htab = mandoc_malloc(sizeof(*htab));
                    606:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  607:
1.167     schwarze  608:        for (tok = mintok; tok < maxtok; tok++) {
1.168     schwarze  609:                if (roff_name[tok] == NULL)
                    610:                        continue;
1.167     schwarze  611:                sz = strlen(roff_name[tok]);
                    612:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    613:                req->tok = tok;
                    614:                memcpy(req->name, roff_name[tok], sz + 1);
                    615:                slot = ohash_qlookup(htab, req->name);
                    616:                ohash_insert(htab, slot, req);
1.3       schwarze  617:        }
1.167     schwarze  618:        return htab;
1.3       schwarze  619: }
                    620:
1.167     schwarze  621: void
                    622: roffhash_free(struct ohash *htab)
1.1       schwarze  623: {
1.167     schwarze  624:        struct roffreq  *req;
                    625:        unsigned int     slot;
1.1       schwarze  626:
1.167     schwarze  627:        if (htab == NULL)
                    628:                return;
                    629:        for (req = ohash_first(htab, &slot); req != NULL;
                    630:             req = ohash_next(htab, &slot))
                    631:                free(req);
                    632:        ohash_delete(htab);
                    633:        free(htab);
                    634: }
                    635:
                    636: enum roff_tok
                    637: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    638: {
                    639:        struct roffreq  *req;
                    640:        const char      *end;
                    641:
                    642:        if (sz) {
                    643:                end = name + sz;
                    644:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    645:        } else
                    646:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    647:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  648: }
                    649:
1.138     schwarze  650: /* --- stack of request blocks -------------------------------------------- */
                    651:
1.1       schwarze  652: /*
                    653:  * Pop the current node off of the stack of roff instructions currently
                    654:  * pending.
                    655:  */
                    656: static void
                    657: roffnode_pop(struct roff *r)
                    658: {
                    659:        struct roffnode *p;
                    660:
1.2       schwarze  661:        assert(r->last);
1.80      schwarze  662:        p = r->last;
1.2       schwarze  663:
                    664:        r->last = r->last->parent;
1.16      schwarze  665:        free(p->name);
                    666:        free(p->end);
1.1       schwarze  667:        free(p);
                    668: }
                    669:
                    670: /*
                    671:  * Push a roff node onto the instruction stack.  This must later be
                    672:  * removed with roffnode_pop().
                    673:  */
1.11      schwarze  674: static void
1.166     schwarze  675: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  676:                int line, int col)
1.1       schwarze  677: {
                    678:        struct roffnode *p;
                    679:
1.11      schwarze  680:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  681:        p->tok = tok;
1.16      schwarze  682:        if (name)
                    683:                p->name = mandoc_strdup(name);
1.1       schwarze  684:        p->parent = r->last;
                    685:        p->line = line;
                    686:        p->col = col;
1.71      schwarze  687:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  688:
                    689:        r->last = p;
                    690: }
                    691:
1.138     schwarze  692: /* --- roff parser state data management ---------------------------------- */
                    693:
1.1       schwarze  694: static void
                    695: roff_free1(struct roff *r)
                    696: {
1.49      schwarze  697:        struct tbl_node *tbl;
1.42      schwarze  698:        int              i;
1.27      schwarze  699:
1.49      schwarze  700:        while (NULL != (tbl = r->first_tbl)) {
                    701:                r->first_tbl = tbl->next;
                    702:                tbl_free(tbl);
1.27      schwarze  703:        }
                    704:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  705:
1.191     schwarze  706:        if (r->last_eqn != NULL)
                    707:                eqn_free(r->last_eqn);
                    708:        r->last_eqn = r->eqn = NULL;
1.32      schwarze  709:
1.1       schwarze  710:        while (r->last)
                    711:                roffnode_pop(r);
1.27      schwarze  712:
1.96      schwarze  713:        free (r->rstack);
                    714:        r->rstack = NULL;
                    715:        r->rstacksz = 0;
                    716:        r->rstackpos = -1;
                    717:
                    718:        roff_freereg(r->regtab);
                    719:        r->regtab = NULL;
                    720:
1.42      schwarze  721:        roff_freestr(r->strtab);
1.178     schwarze  722:        roff_freestr(r->rentab);
1.42      schwarze  723:        roff_freestr(r->xmbtab);
1.178     schwarze  724:        r->strtab = r->rentab = r->xmbtab = NULL;
1.42      schwarze  725:
                    726:        if (r->xtab)
                    727:                for (i = 0; i < 128; i++)
                    728:                        free(r->xtab[i].p);
                    729:        free(r->xtab);
                    730:        r->xtab = NULL;
1.1       schwarze  731: }
                    732:
                    733: void
                    734: roff_reset(struct roff *r)
                    735: {
                    736:        roff_free1(r);
1.99      schwarze  737:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175     schwarze  738:        r->control = '\0';
                    739:        r->escape = '\\';
1.179     schwarze  740:        roffce_lines = 0;
                    741:        roffce_node = NULL;
                    742:        roffit_lines = 0;
                    743:        roffit_macro = NULL;
1.1       schwarze  744: }
                    745:
                    746: void
                    747: roff_free(struct roff *r)
                    748: {
                    749:        roff_free1(r);
1.167     schwarze  750:        roffhash_free(r->reqtab);
1.1       schwarze  751:        free(r);
                    752: }
                    753:
                    754: struct roff *
1.151     schwarze  755: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  756: {
                    757:        struct roff     *r;
                    758:
1.11      schwarze  759:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  760:        r->parse = parse;
1.167     schwarze  761:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72      schwarze  762:        r->options = options;
1.99      schwarze  763:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  764:        r->rstackpos = -1;
1.175     schwarze  765:        r->escape = '\\';
1.149     schwarze  766:        return r;
1.137     schwarze  767: }
                    768:
1.138     schwarze  769: /* --- syntax tree state data management ---------------------------------- */
                    770:
1.137     schwarze  771: static void
                    772: roff_man_free1(struct roff_man *man)
                    773: {
                    774:
1.138     schwarze  775:        if (man->first != NULL)
                    776:                roff_node_delete(man, man->first);
1.137     schwarze  777:        free(man->meta.msec);
                    778:        free(man->meta.vol);
                    779:        free(man->meta.os);
                    780:        free(man->meta.arch);
                    781:        free(man->meta.title);
                    782:        free(man->meta.name);
                    783:        free(man->meta.date);
                    784: }
                    785:
                    786: static void
                    787: roff_man_alloc1(struct roff_man *man)
                    788: {
                    789:
                    790:        memset(&man->meta, 0, sizeof(man->meta));
                    791:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    792:        man->first->type = ROFFT_ROOT;
                    793:        man->last = man->first;
                    794:        man->last_es = NULL;
                    795:        man->flags = 0;
                    796:        man->macroset = MACROSET_NONE;
                    797:        man->lastsec = man->lastnamed = SEC_NONE;
                    798:        man->next = ROFF_NEXT_CHILD;
                    799: }
                    800:
                    801: void
                    802: roff_man_reset(struct roff_man *man)
                    803: {
                    804:
                    805:        roff_man_free1(man);
                    806:        roff_man_alloc1(man);
                    807: }
                    808:
                    809: void
                    810: roff_man_free(struct roff_man *man)
                    811: {
                    812:
                    813:        roff_man_free1(man);
                    814:        free(man);
                    815: }
                    816:
                    817: struct roff_man *
                    818: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.188     schwarze  819:        const char *os_s, int quick)
1.137     schwarze  820: {
                    821:        struct roff_man *man;
                    822:
                    823:        man = mandoc_calloc(1, sizeof(*man));
                    824:        man->parse = parse;
                    825:        man->roff = roff;
1.188     schwarze  826:        man->os_s = os_s;
1.137     schwarze  827:        man->quick = quick;
                    828:        roff_man_alloc1(man);
1.168     schwarze  829:        roff->man = man;
1.149     schwarze  830:        return man;
1.1       schwarze  831: }
                    832:
1.138     schwarze  833: /* --- syntax tree handling ----------------------------------------------- */
                    834:
                    835: struct roff_node *
                    836: roff_node_alloc(struct roff_man *man, int line, int pos,
                    837:        enum roff_type type, int tok)
                    838: {
                    839:        struct roff_node        *n;
                    840:
                    841:        n = mandoc_calloc(1, sizeof(*n));
                    842:        n->line = line;
                    843:        n->pos = pos;
                    844:        n->tok = tok;
                    845:        n->type = type;
                    846:        n->sec = man->lastsec;
                    847:
                    848:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  849:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  850:        else
1.157     schwarze  851:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  852:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  853:                n->flags |= NODE_LINE;
1.138     schwarze  854:        man->flags &= ~MDOC_NEWLINE;
                    855:
1.149     schwarze  856:        return n;
1.138     schwarze  857: }
                    858:
                    859: void
                    860: roff_node_append(struct roff_man *man, struct roff_node *n)
                    861: {
                    862:
                    863:        switch (man->next) {
                    864:        case ROFF_NEXT_SIBLING:
1.153     schwarze  865:                if (man->last->next != NULL) {
                    866:                        n->next = man->last->next;
                    867:                        man->last->next->prev = n;
                    868:                } else
                    869:                        man->last->parent->last = n;
1.138     schwarze  870:                man->last->next = n;
                    871:                n->prev = man->last;
                    872:                n->parent = man->last->parent;
                    873:                break;
                    874:        case ROFF_NEXT_CHILD:
1.159     schwarze  875:                if (man->last->child != NULL) {
                    876:                        n->next = man->last->child;
                    877:                        man->last->child->prev = n;
                    878:                } else
                    879:                        man->last->last = n;
1.138     schwarze  880:                man->last->child = n;
                    881:                n->parent = man->last;
                    882:                break;
                    883:        default:
                    884:                abort();
                    885:        }
1.154     schwarze  886:        man->last = n;
1.138     schwarze  887:
                    888:        switch (n->type) {
                    889:        case ROFFT_HEAD:
                    890:                n->parent->head = n;
                    891:                break;
                    892:        case ROFFT_BODY:
1.154     schwarze  893:                if (n->end != ENDBODY_NOT)
                    894:                        return;
1.138     schwarze  895:                n->parent->body = n;
                    896:                break;
                    897:        case ROFFT_TAIL:
                    898:                n->parent->tail = n;
                    899:                break;
                    900:        default:
1.154     schwarze  901:                return;
1.138     schwarze  902:        }
1.154     schwarze  903:
                    904:        /*
                    905:         * Copy over the normalised-data pointer of our parent.  Not
                    906:         * everybody has one, but copying a null pointer is fine.
                    907:         */
                    908:
                    909:        n->norm = n->parent->norm;
                    910:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  911: }
                    912:
1.139     schwarze  913: void
                    914: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    915: {
                    916:        struct roff_node        *n;
                    917:
                    918:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    919:        n->string = roff_strdup(man->roff, word);
                    920:        roff_node_append(man, n);
1.158     schwarze  921:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  922:        man->next = ROFF_NEXT_SIBLING;
                    923: }
                    924:
                    925: void
                    926: roff_word_append(struct roff_man *man, const char *word)
                    927: {
                    928:        struct roff_node        *n;
                    929:        char                    *addstr, *newstr;
                    930:
                    931:        n = man->last;
                    932:        addstr = roff_strdup(man->roff, word);
                    933:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    934:        free(addstr);
                    935:        free(n->string);
                    936:        n->string = newstr;
                    937:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  938: }
                    939:
                    940: void
                    941: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    942: {
                    943:        struct roff_node        *n;
                    944:
                    945:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    946:        roff_node_append(man, n);
                    947:        man->next = ROFF_NEXT_CHILD;
                    948: }
                    949:
                    950: struct roff_node *
                    951: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    952: {
                    953:        struct roff_node        *n;
                    954:
                    955:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    956:        roff_node_append(man, n);
                    957:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  958:        return n;
1.139     schwarze  959: }
                    960:
1.138     schwarze  961: struct roff_node *
                    962: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    963: {
                    964:        struct roff_node        *n;
                    965:
                    966:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    967:        roff_node_append(man, n);
                    968:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  969:        return n;
1.138     schwarze  970: }
                    971:
                    972: struct roff_node *
                    973: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    974: {
                    975:        struct roff_node        *n;
                    976:
                    977:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    978:        roff_node_append(man, n);
                    979:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  980:        return n;
1.139     schwarze  981: }
                    982:
1.193     schwarze  983: static void
                    984: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.139     schwarze  985: {
                    986:        struct roff_node        *n;
1.193     schwarze  987:        const struct tbl_span   *span;
1.139     schwarze  988:
                    989:        if (man->macroset == MACROSET_MAN)
1.174     schwarze  990:                man_breakscope(man, ROFF_TS);
1.193     schwarze  991:        while ((span = tbl_span(tbl)) != NULL) {
                    992:                n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                    993:                n->span = span;
                    994:                roff_node_append(man, n);
                    995:                n->flags |= NODE_VALID | NODE_ENDED;
                    996:                man->next = ROFF_NEXT_SIBLING;
                    997:        }
1.138     schwarze  998: }
                    999:
                   1000: void
                   1001: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1002: {
                   1003:
                   1004:        /* Adjust siblings. */
                   1005:
                   1006:        if (n->prev)
                   1007:                n->prev->next = n->next;
                   1008:        if (n->next)
                   1009:                n->next->prev = n->prev;
                   1010:
                   1011:        /* Adjust parent. */
                   1012:
                   1013:        if (n->parent != NULL) {
                   1014:                if (n->parent->child == n)
                   1015:                        n->parent->child = n->next;
                   1016:                if (n->parent->last == n)
                   1017:                        n->parent->last = n->prev;
                   1018:        }
                   1019:
                   1020:        /* Adjust parse point. */
                   1021:
                   1022:        if (man == NULL)
                   1023:                return;
                   1024:        if (man->last == n) {
                   1025:                if (n->prev == NULL) {
                   1026:                        man->last = n->parent;
                   1027:                        man->next = ROFF_NEXT_CHILD;
                   1028:                } else {
                   1029:                        man->last = n->prev;
                   1030:                        man->next = ROFF_NEXT_SIBLING;
                   1031:                }
                   1032:        }
                   1033:        if (man->first == n)
                   1034:                man->first = NULL;
                   1035: }
                   1036:
                   1037: void
                   1038: roff_node_free(struct roff_node *n)
                   1039: {
                   1040:
                   1041:        if (n->args != NULL)
                   1042:                mdoc_argv_free(n->args);
                   1043:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1044:                free(n->norm);
1.191     schwarze 1045:        if (n->eqn != NULL)
                   1046:                eqn_box_free(n->eqn);
1.138     schwarze 1047:        free(n->string);
                   1048:        free(n);
                   1049: }
                   1050:
                   1051: void
                   1052: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1053: {
                   1054:
                   1055:        while (n->child != NULL)
                   1056:                roff_node_delete(man, n->child);
                   1057:        roff_node_unlink(man, n);
                   1058:        roff_node_free(n);
1.141     schwarze 1059: }
                   1060:
                   1061: void
                   1062: deroff(char **dest, const struct roff_node *n)
                   1063: {
                   1064:        char    *cp;
                   1065:        size_t   sz;
                   1066:
                   1067:        if (n->type != ROFFT_TEXT) {
                   1068:                for (n = n->child; n != NULL; n = n->next)
                   1069:                        deroff(dest, n);
                   1070:                return;
                   1071:        }
                   1072:
1.160     schwarze 1073:        /* Skip leading whitespace. */
1.141     schwarze 1074:
1.160     schwarze 1075:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1076:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1077:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1078:                        cp++;
1.160     schwarze 1079:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1080:                        break;
                   1081:        }
                   1082:
1.161     schwarze 1083:        /* Skip trailing backslash. */
                   1084:
                   1085:        sz = strlen(cp);
1.162     schwarze 1086:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1087:                sz--;
                   1088:
1.141     schwarze 1089:        /* Skip trailing whitespace. */
                   1090:
1.161     schwarze 1091:        for (; sz; sz--)
1.141     schwarze 1092:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1093:                        break;
                   1094:
                   1095:        /* Skip empty strings. */
                   1096:
                   1097:        if (sz == 0)
                   1098:                return;
                   1099:
                   1100:        if (*dest == NULL) {
                   1101:                *dest = mandoc_strndup(cp, sz);
                   1102:                return;
                   1103:        }
                   1104:
                   1105:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1106:        free(*dest);
                   1107:        *dest = cp;
1.138     schwarze 1108: }
                   1109:
                   1110: /* --- main functions of the roff parser ---------------------------------- */
                   1111:
1.8       schwarze 1112: /*
1.79      schwarze 1113:  * In the current line, expand escape sequences that tend to get
                   1114:  * used in numerical expressions and conditional requests.
                   1115:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1116:  */
1.45      schwarze 1117: static enum rofferr
1.110     schwarze 1118: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1119: {
1.81      schwarze 1120:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze 1121:        const char      *start; /* start of the string to process */
1.82      schwarze 1122:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze 1123:        const char      *stnam; /* start of the name, after "[(*" */
                   1124:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1125:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1126:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1127:        size_t           maxl;  /* expected length of the escape name */
                   1128:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1129:        enum mandoc_esc  esc;   /* type of the escape sequence */
                   1130:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1131:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1132:        int              npos;  /* position in numeric expression */
1.91      schwarze 1133:        int              arg_complete; /* argument not interrupted by eol */
1.175     schwarze 1134:        int              done;  /* no more input available */
1.187     schwarze 1135:        int              deftype; /* type of definition to paste */
1.195     schwarze 1136:        int              rcsid; /* kind of RCS id seen */
1.79      schwarze 1137:        char             term;  /* character terminating the escape */
1.8       schwarze 1138:
1.175     schwarze 1139:        /* Search forward for comments. */
                   1140:
                   1141:        done = 0;
                   1142:        start = buf->buf + pos;
                   1143:        for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
                   1144:                if (stesc[0] != r->escape || stesc[1] == '\0')
                   1145:                        continue;
                   1146:                stesc++;
                   1147:                if (*stesc != '"' && *stesc != '#')
                   1148:                        continue;
1.185     schwarze 1149:
                   1150:                /* Comment found, look for RCS id. */
                   1151:
1.195     schwarze 1152:                rcsid = 0;
1.185     schwarze 1153:                if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.195     schwarze 1154:                        rcsid = 1 << MANDOC_OS_OPENBSD;
1.185     schwarze 1155:                        cp += 8;
                   1156:                } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.195     schwarze 1157:                        rcsid = 1 << MANDOC_OS_NETBSD;
1.185     schwarze 1158:                        cp += 7;
                   1159:                }
                   1160:                if (cp != NULL &&
                   1161:                    isalnum((unsigned char)*cp) == 0 &&
1.186     schwarze 1162:                    strchr(cp, '$') != NULL) {
1.195     schwarze 1163:                        if (r->man->meta.rcsids & rcsid)
1.186     schwarze 1164:                                mandoc_msg(MANDOCERR_RCS_REP, r->parse,
                   1165:                                    ln, stesc + 1 - buf->buf, stesc + 1);
1.195     schwarze 1166:                        r->man->meta.rcsids |= rcsid;
1.186     schwarze 1167:                }
1.185     schwarze 1168:
                   1169:                /* Handle trailing whitespace. */
                   1170:
1.175     schwarze 1171:                cp = strchr(stesc--, '\0') - 1;
                   1172:                if (*cp == '\n') {
                   1173:                        done = 1;
                   1174:                        cp--;
                   1175:                }
                   1176:                if (*cp == ' ' || *cp == '\t')
                   1177:                        mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
                   1178:                            ln, cp - buf->buf, NULL);
                   1179:                while (stesc > start && stesc[-1] == ' ')
                   1180:                        stesc--;
                   1181:                *stesc = '\0';
                   1182:                break;
                   1183:        }
                   1184:        if (stesc == start)
                   1185:                return ROFF_CONT;
                   1186:        stesc--;
                   1187:
                   1188:        /* Notice the end of the input. */
                   1189:
                   1190:        if (*stesc == '\n') {
                   1191:                *stesc-- = '\0';
                   1192:                done = 1;
                   1193:        }
                   1194:
1.43      schwarze 1195:        expand_count = 0;
1.175     schwarze 1196:        while (stesc >= start) {
1.43      schwarze 1197:
1.78      schwarze 1198:                /* Search backwards for the next backslash. */
                   1199:
1.175     schwarze 1200:                if (*stesc != r->escape) {
                   1201:                        if (*stesc == '\\') {
                   1202:                                *stesc = '\0';
                   1203:                                buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
                   1204:                                    buf->buf, stesc + 1) + 1;
                   1205:                                start = nbuf + pos;
                   1206:                                stesc = nbuf + (stesc - buf->buf);
                   1207:                                free(buf->buf);
                   1208:                                buf->buf = nbuf;
                   1209:                        }
                   1210:                        stesc--;
1.78      schwarze 1211:                        continue;
1.175     schwarze 1212:                }
1.78      schwarze 1213:
                   1214:                /* If it is escaped, skip it. */
                   1215:
                   1216:                for (cp = stesc - 1; cp >= start; cp--)
1.175     schwarze 1217:                        if (*cp != r->escape)
1.78      schwarze 1218:                                break;
                   1219:
1.110     schwarze 1220:                if ((stesc - cp) % 2 == 0) {
1.175     schwarze 1221:                        while (stesc > cp)
                   1222:                                *stesc-- = '\\';
1.78      schwarze 1223:                        continue;
1.175     schwarze 1224:                } else if (stesc[1] != '\0') {
                   1225:                        *stesc = '\\';
                   1226:                } else {
                   1227:                        *stesc-- = '\0';
                   1228:                        if (done)
                   1229:                                continue;
                   1230:                        else
                   1231:                                return ROFF_APPEND;
1.78      schwarze 1232:                }
1.23      schwarze 1233:
1.79      schwarze 1234:                /* Decide whether to expand or to check only. */
1.23      schwarze 1235:
1.79      schwarze 1236:                term = '\0';
1.78      schwarze 1237:                cp = stesc + 1;
1.53      schwarze 1238:                switch (*cp) {
1.80      schwarze 1239:                case '*':
1.53      schwarze 1240:                        res = NULL;
                   1241:                        break;
1.80      schwarze 1242:                case 'B':
                   1243:                case 'w':
1.79      schwarze 1244:                        term = cp[1];
                   1245:                        /* FALLTHROUGH */
1.80      schwarze 1246:                case 'n':
1.53      schwarze 1247:                        res = ubuf;
                   1248:                        break;
                   1249:                default:
1.109     schwarze 1250:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1251:                        if (esc == ESCAPE_ERROR ||
                   1252:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1253:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1254:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1255:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1256:                                    "%.*s", (int)(cp - stesc), stesc);
1.175     schwarze 1257:                        stesc--;
1.78      schwarze 1258:                        continue;
1.42      schwarze 1259:                }
                   1260:
1.78      schwarze 1261:                if (EXPAND_LIMIT < ++expand_count) {
                   1262:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1263:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1264:                        return ROFF_IGN;
1.78      schwarze 1265:                }
1.23      schwarze 1266:
                   1267:                /*
                   1268:                 * The third character decides the length
1.53      schwarze 1269:                 * of the name of the string or register.
1.23      schwarze 1270:                 * Save a pointer to the name.
                   1271:                 */
                   1272:
1.110     schwarze 1273:                if (term == '\0') {
1.79      schwarze 1274:                        switch (*++cp) {
1.80      schwarze 1275:                        case '\0':
1.79      schwarze 1276:                                maxl = 0;
                   1277:                                break;
1.80      schwarze 1278:                        case '(':
1.79      schwarze 1279:                                cp++;
                   1280:                                maxl = 2;
                   1281:                                break;
1.80      schwarze 1282:                        case '[':
1.79      schwarze 1283:                                cp++;
                   1284:                                term = ']';
                   1285:                                maxl = 0;
                   1286:                                break;
                   1287:                        default:
                   1288:                                maxl = 1;
                   1289:                                break;
                   1290:                        }
                   1291:                } else {
                   1292:                        cp += 2;
1.8       schwarze 1293:                        maxl = 0;
                   1294:                }
1.23      schwarze 1295:                stnam = cp;
1.8       schwarze 1296:
1.23      schwarze 1297:                /* Advance to the end of the name. */
1.8       schwarze 1298:
1.125     schwarze 1299:                naml = 0;
1.91      schwarze 1300:                arg_complete = 1;
1.125     schwarze 1301:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1302:                        if (*cp == '\0') {
1.92      schwarze 1303:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1304:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1305:                                arg_complete = 0;
1.79      schwarze 1306:                                break;
1.42      schwarze 1307:                        }
1.110     schwarze 1308:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1309:                                cp++;
1.125     schwarze 1310:                                break;
                   1311:                        }
                   1312:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1313:                                naml++;
                   1314:                                continue;
                   1315:                        }
                   1316:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1317:                        case ESCAPE_SPECIAL:
                   1318:                        case ESCAPE_UNICODE:
                   1319:                        case ESCAPE_NUMBERED:
                   1320:                        case ESCAPE_OVERSTRIKE:
                   1321:                                naml++;
                   1322:                                break;
                   1323:                        default:
1.8       schwarze 1324:                                break;
1.79      schwarze 1325:                        }
1.8       schwarze 1326:                }
                   1327:
1.23      schwarze 1328:                /*
                   1329:                 * Retrieve the replacement string; if it is
                   1330:                 * undefined, resume searching for escapes.
                   1331:                 */
                   1332:
1.79      schwarze 1333:                switch (stesc[1]) {
1.80      schwarze 1334:                case '*':
1.187     schwarze 1335:                        if (arg_complete) {
                   1336:                                deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1337:                                res = roff_getstrn(r, stnam, naml, &deftype);
                   1338:                        }
1.79      schwarze 1339:                        break;
1.80      schwarze 1340:                case 'B':
1.79      schwarze 1341:                        npos = 0;
1.91      schwarze 1342:                        ubuf[0] = arg_complete &&
1.133     schwarze 1343:                            roff_evalnum(r, ln, stnam, &npos,
                   1344:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1345:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1346:                        ubuf[1] = '\0';
                   1347:                        break;
1.80      schwarze 1348:                case 'n':
1.91      schwarze 1349:                        if (arg_complete)
                   1350:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1351:                                    roff_getregn(r, stnam, naml));
                   1352:                        else
                   1353:                                ubuf[0] = '\0';
1.79      schwarze 1354:                        break;
1.80      schwarze 1355:                case 'w':
1.91      schwarze 1356:                        /* use even incomplete args */
1.81      schwarze 1357:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1358:                            24 * (int)naml);
                   1359:                        break;
                   1360:                }
1.8       schwarze 1361:
1.110     schwarze 1362:                if (res == NULL) {
1.92      schwarze 1363:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1364:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1365:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1366:                        res = "";
1.118     schwarze 1367:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1368:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1369:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1370:                        return ROFF_IGN;
1.8       schwarze 1371:                }
                   1372:
1.23      schwarze 1373:                /* Replace the escape sequence by the string. */
                   1374:
1.82      schwarze 1375:                *stesc = '\0';
1.110     schwarze 1376:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1377:                    buf->buf, res, cp) + 1;
1.8       schwarze 1378:
1.78      schwarze 1379:                /* Prepare for the next replacement. */
1.8       schwarze 1380:
1.78      schwarze 1381:                start = nbuf + pos;
1.110     schwarze 1382:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1383:                free(buf->buf);
                   1384:                buf->buf = nbuf;
1.42      schwarze 1385:        }
1.149     schwarze 1386:        return ROFF_CONT;
1.42      schwarze 1387: }
                   1388:
                   1389: /*
1.147     schwarze 1390:  * Process text streams.
1.42      schwarze 1391:  */
                   1392: static enum rofferr
1.177     schwarze 1393: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42      schwarze 1394: {
                   1395:        size_t           sz;
                   1396:        const char      *start;
1.51      schwarze 1397:        char            *p;
                   1398:        int              isz;
1.42      schwarze 1399:        enum mandoc_esc  esc;
                   1400:
1.147     schwarze 1401:        /* Spring the input line trap. */
                   1402:
                   1403:        if (roffit_lines == 1) {
                   1404:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1405:                free(buf->buf);
                   1406:                buf->buf = p;
                   1407:                buf->sz = isz + 1;
                   1408:                *offs = 0;
                   1409:                free(roffit_macro);
                   1410:                roffit_lines = 0;
1.149     schwarze 1411:                return ROFF_REPARSE;
1.147     schwarze 1412:        } else if (roffit_lines > 1)
                   1413:                --roffit_lines;
                   1414:
1.177     schwarze 1415:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1416:                if (roffce_lines < 1) {
                   1417:                        r->man->last = roffce_node;
                   1418:                        r->man->next = ROFF_NEXT_SIBLING;
                   1419:                        roffce_lines = 0;
                   1420:                        roffce_node = NULL;
                   1421:                } else
                   1422:                        roffce_lines--;
                   1423:        }
                   1424:
1.147     schwarze 1425:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1426:
1.110     schwarze 1427:        start = p = buf->buf + pos;
1.42      schwarze 1428:
1.110     schwarze 1429:        while (*p != '\0') {
1.42      schwarze 1430:                sz = strcspn(p, "-\\");
                   1431:                p += sz;
                   1432:
1.110     schwarze 1433:                if (*p == '\0')
1.42      schwarze 1434:                        break;
                   1435:
1.110     schwarze 1436:                if (*p == '\\') {
1.42      schwarze 1437:                        /* Skip over escapes. */
                   1438:                        p++;
1.62      schwarze 1439:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1440:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1441:                                break;
1.136     schwarze 1442:                        while (*p == '-')
                   1443:                                p++;
1.42      schwarze 1444:                        continue;
                   1445:                } else if (p == start) {
                   1446:                        p++;
                   1447:                        continue;
                   1448:                }
                   1449:
1.44      schwarze 1450:                if (isalpha((unsigned char)p[-1]) &&
                   1451:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1452:                        *p = ASCII_HYPH;
                   1453:                p++;
1.8       schwarze 1454:        }
1.149     schwarze 1455:        return ROFF_CONT;
1.8       schwarze 1456: }
                   1457:
1.1       schwarze 1458: enum rofferr
1.110     schwarze 1459: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1460: {
1.166     schwarze 1461:        enum roff_tok    t;
1.27      schwarze 1462:        enum rofferr     e;
1.110     schwarze 1463:        int              pos;   /* parse point */
1.115     schwarze 1464:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1465:        int              ppos;  /* original offset in buf->buf */
                   1466:        int              ctl;   /* macro line (boolean) */
                   1467:
                   1468:        ppos = pos = *offs;
1.1       schwarze 1469:
1.102     schwarze 1470:        /* Handle in-line equation delimiters. */
                   1471:
1.108     schwarze 1472:        if (r->tbl == NULL &&
                   1473:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1474:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1475:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1476:                if (e == ROFF_REPARSE)
1.149     schwarze 1477:                        return e;
1.102     schwarze 1478:                assert(e == ROFF_CONT);
                   1479:        }
                   1480:
                   1481:        /* Expand some escape sequences. */
1.8       schwarze 1482:
1.110     schwarze 1483:        e = roff_res(r, buf, ln, pos);
1.175     schwarze 1484:        if (e == ROFF_IGN || e == ROFF_APPEND)
1.149     schwarze 1485:                return e;
1.110     schwarze 1486:        assert(e == ROFF_CONT);
1.8       schwarze 1487:
1.110     schwarze 1488:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1489:
1.8       schwarze 1490:        /*
1.2       schwarze 1491:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1492:         * text through the macro's filter.
                   1493:         * Equations process all content themselves.
                   1494:         * Tables process almost all content themselves, but we want
                   1495:         * to warn about macros before passing it there.
1.2       schwarze 1496:         */
                   1497:
1.124     schwarze 1498:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1499:                t = r->last->tok;
1.110     schwarze 1500:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1501:                if (e == ROFF_IGN)
1.149     schwarze 1502:                        return e;
1.166     schwarze 1503:                assert(e == ROFF_CONT);
1.54      schwarze 1504:        }
1.191     schwarze 1505:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1506:                eqn_read(r->eqn, buf->buf + ppos);
                   1507:                return ROFF_IGN;
                   1508:        }
1.193     schwarze 1509:        if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
                   1510:                tbl_read(r->tbl, ln, buf->buf, ppos);
                   1511:                roff_addtbl(r->man, r->tbl);
                   1512:                return ROFF_IGN;
                   1513:        }
1.124     schwarze 1514:        if ( ! ctl)
1.177     schwarze 1515:                return roff_parsetext(r, buf, pos, offs);
1.100     schwarze 1516:
                   1517:        /* Skip empty request lines. */
                   1518:
1.110     schwarze 1519:        if (buf->buf[pos] == '"') {
1.100     schwarze 1520:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1521:                    ln, pos, NULL);
1.149     schwarze 1522:                return ROFF_IGN;
1.110     schwarze 1523:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1524:                return ROFF_IGN;
1.2       schwarze 1525:
                   1526:        /*
                   1527:         * If a scope is open, go to the child handler for that macro,
                   1528:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1529:         * Don't do so if an equation is open.
1.2       schwarze 1530:         */
                   1531:
                   1532:        if (r->last) {
1.1       schwarze 1533:                t = r->last->tok;
1.149     schwarze 1534:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1535:        }
                   1536:
1.115     schwarze 1537:        /* No scope is open.  This is a new request or macro. */
                   1538:
                   1539:        spos = pos;
                   1540:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1541:
                   1542:        /* Tables ignore most macros. */
                   1543:
1.180     schwarze 1544:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181     schwarze 1545:            t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.115     schwarze 1546:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1547:                    ln, pos, buf->buf + spos);
1.180     schwarze 1548:                if (t != TOKEN_NONE)
1.149     schwarze 1549:                        return ROFF_IGN;
1.129     schwarze 1550:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1551:                        pos++;
1.163     schwarze 1552:                while (buf->buf[pos] == ' ')
1.129     schwarze 1553:                        pos++;
1.193     schwarze 1554:                tbl_read(r->tbl, ln, buf->buf, pos);
                   1555:                roff_addtbl(r->man, r->tbl);
                   1556:                return ROFF_IGN;
1.115     schwarze 1557:        }
                   1558:
1.177     schwarze 1559:        /* For now, let high level macros abort .ce mode. */
                   1560:
                   1561:        if (ctl && roffce_node != NULL &&
1.196     schwarze 1562:            (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
                   1563:             t == ROFF_TH || t == ROFF_TS)) {
1.177     schwarze 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: {
1.193     schwarze 1586:        if (r->last != NULL)
1.94      schwarze 1587:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1588:                    r->last->line, r->last->col,
1.167     schwarze 1589:                    roff_name[r->last->tok]);
1.27      schwarze 1590:
1.193     schwarze 1591:        if (r->eqn != NULL) {
1.94      schwarze 1592:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.191     schwarze 1593:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1594:                eqn_parse(r->eqn);
                   1595:                r->eqn = NULL;
1.32      schwarze 1596:        }
                   1597:
1.193     schwarze 1598:        if (r->tbl != NULL) {
1.94      schwarze 1599:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1600:                    r->tbl->line, r->tbl->pos, "TS");
1.193     schwarze 1601:                tbl_end(r->tbl);
                   1602:                r->tbl = NULL;
1.27      schwarze 1603:        }
1.1       schwarze 1604: }
                   1605:
                   1606: /*
                   1607:  * Parse a roff node's type from the input buffer.  This must be in the
                   1608:  * form of ".foo xxx" in the usual way.
                   1609:  */
1.166     schwarze 1610: static enum roff_tok
1.87      schwarze 1611: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1612: {
1.87      schwarze 1613:        char            *cp;
1.16      schwarze 1614:        const char      *mac;
                   1615:        size_t           maclen;
1.187     schwarze 1616:        int              deftype;
1.166     schwarze 1617:        enum roff_tok    t;
1.1       schwarze 1618:
1.87      schwarze 1619:        cp = buf + *pos;
                   1620:
                   1621:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1622:                return TOKEN_NONE;
1.1       schwarze 1623:
1.87      schwarze 1624:        mac = cp;
                   1625:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1626:
1.187     schwarze 1627:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1628:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1629:        switch (deftype) {
                   1630:        case ROFFDEF_USER:
                   1631:                t = ROFF_USERDEF;
                   1632:                break;
                   1633:        case ROFFDEF_REN:
                   1634:                t = ROFF_RENAMED;
                   1635:                break;
                   1636:        default:
                   1637:                t = roffhash_find(r->reqtab, mac, maclen);
                   1638:                break;
                   1639:        }
1.166     schwarze 1640:        if (t != TOKEN_NONE)
1.87      schwarze 1641:                *pos = cp - buf;
1.197     schwarze 1642:        else if (deftype == ROFFDEF_UNDEF) {
                   1643:                /* Using an undefined macro defines it to be empty. */
                   1644:                roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
                   1645:                roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
                   1646:        }
1.149     schwarze 1647:        return t;
1.1       schwarze 1648: }
                   1649:
1.138     schwarze 1650: /* --- handling of request blocks ----------------------------------------- */
                   1651:
1.1       schwarze 1652: static enum rofferr
1.2       schwarze 1653: roff_cblock(ROFF_ARGS)
1.1       schwarze 1654: {
                   1655:
1.2       schwarze 1656:        /*
                   1657:         * A block-close `..' should only be invoked as a child of an
                   1658:         * ignore macro, otherwise raise a warning and just ignore it.
                   1659:         */
                   1660:
1.110     schwarze 1661:        if (r->last == NULL) {
1.94      schwarze 1662:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1663:                    ln, ppos, "..");
1.149     schwarze 1664:                return ROFF_IGN;
1.2       schwarze 1665:        }
1.1       schwarze 1666:
1.2       schwarze 1667:        switch (r->last->tok) {
1.80      schwarze 1668:        case ROFF_am:
1.93      schwarze 1669:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1670:        case ROFF_ami:
                   1671:        case ROFF_de:
1.23      schwarze 1672:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1673:        case ROFF_dei:
                   1674:        case ROFF_ig:
1.2       schwarze 1675:                break;
                   1676:        default:
1.94      schwarze 1677:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1678:                    ln, ppos, "..");
1.149     schwarze 1679:                return ROFF_IGN;
1.2       schwarze 1680:        }
                   1681:
1.110     schwarze 1682:        if (buf->buf[pos] != '\0')
1.90      schwarze 1683:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1684:                    ".. %s", buf->buf + pos);
1.2       schwarze 1685:
                   1686:        roffnode_pop(r);
                   1687:        roffnode_cleanscope(r);
1.149     schwarze 1688:        return ROFF_IGN;
1.2       schwarze 1689:
                   1690: }
1.1       schwarze 1691:
1.2       schwarze 1692: static void
                   1693: roffnode_cleanscope(struct roff *r)
                   1694: {
1.1       schwarze 1695:
1.2       schwarze 1696:        while (r->last) {
1.46      schwarze 1697:                if (--r->last->endspan != 0)
1.2       schwarze 1698:                        break;
                   1699:                roffnode_pop(r);
                   1700:        }
                   1701: }
1.1       schwarze 1702:
1.68      schwarze 1703: static void
                   1704: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1705: {
1.1       schwarze 1706:
1.2       schwarze 1707:        if (NULL == r->last) {
1.94      schwarze 1708:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1709:                    ln, ppos, "\\}");
1.68      schwarze 1710:                return;
1.2       schwarze 1711:        }
1.1       schwarze 1712:
1.2       schwarze 1713:        switch (r->last->tok) {
1.80      schwarze 1714:        case ROFF_el:
                   1715:        case ROFF_ie:
                   1716:        case ROFF_if:
1.2       schwarze 1717:                break;
                   1718:        default:
1.94      schwarze 1719:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1720:                    ln, ppos, "\\}");
1.68      schwarze 1721:                return;
1.2       schwarze 1722:        }
1.1       schwarze 1723:
1.2       schwarze 1724:        if (r->last->endspan > -1) {
1.94      schwarze 1725:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1726:                    ln, ppos, "\\}");
1.68      schwarze 1727:                return;
1.2       schwarze 1728:        }
                   1729:
                   1730:        roffnode_pop(r);
                   1731:        roffnode_cleanscope(r);
1.68      schwarze 1732:        return;
1.1       schwarze 1733: }
                   1734:
                   1735: static enum rofferr
1.2       schwarze 1736: roff_block(ROFF_ARGS)
1.1       schwarze 1737: {
1.187     schwarze 1738:        const char      *name, *value;
                   1739:        char            *call, *cp, *iname, *rname;
                   1740:        size_t           csz, namesz, rsz;
                   1741:        int              deftype;
1.16      schwarze 1742:
1.93      schwarze 1743:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1744:
1.110     schwarze 1745:        if (tok == ROFF_de1)
1.93      schwarze 1746:                tok = ROFF_de;
1.123     schwarze 1747:        else if (tok == ROFF_dei1)
                   1748:                tok = ROFF_dei;
1.110     schwarze 1749:        else if (tok == ROFF_am1)
1.93      schwarze 1750:                tok = ROFF_am;
1.123     schwarze 1751:        else if (tok == ROFF_ami1)
                   1752:                tok = ROFF_ami;
1.93      schwarze 1753:
                   1754:        /* Parse the macro name argument. */
                   1755:
1.110     schwarze 1756:        cp = buf->buf + pos;
                   1757:        if (tok == ROFF_ig) {
1.93      schwarze 1758:                iname = NULL;
                   1759:                namesz = 0;
                   1760:        } else {
                   1761:                iname = cp;
                   1762:                namesz = roff_getname(r, &cp, ln, ppos);
                   1763:                iname[namesz] = '\0';
                   1764:        }
1.22      schwarze 1765:
1.93      schwarze 1766:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1767:
1.110     schwarze 1768:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1769:                deftype = ROFFDEF_USER;
                   1770:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1771:                if (name == NULL) {
1.93      schwarze 1772:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1773:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1774:                            "%.*s", (int)namesz, iname);
                   1775:                        namesz = 0;
                   1776:                } else
                   1777:                        namesz = strlen(name);
                   1778:        } else
                   1779:                name = iname;
1.22      schwarze 1780:
1.110     schwarze 1781:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1782:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167     schwarze 1783:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1784:                return ROFF_IGN;
1.93      schwarze 1785:        }
1.2       schwarze 1786:
1.16      schwarze 1787:        roffnode_push(r, tok, name, ln, ppos);
                   1788:
                   1789:        /*
                   1790:         * At the beginning of a `de' macro, clear the existing string
                   1791:         * with the same name, if there is one.  New content will be
1.66      schwarze 1792:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1793:         */
1.22      schwarze 1794:
1.183     schwarze 1795:        if (tok == ROFF_de || tok == ROFF_dei) {
1.86      schwarze 1796:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183     schwarze 1797:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.187     schwarze 1798:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   1799:                deftype = ROFFDEF_ANY;
                   1800:                value = roff_getstrn(r, iname, namesz, &deftype);
                   1801:                switch (deftype) {  /* Before appending, ... */
                   1802:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   1803:                        roff_setstrn(&r->strtab, name, namesz,
                   1804:                            value, strlen(value), 0);
                   1805:                        break;
                   1806:                case ROFFDEF_REN: /* call original standard macro. */
                   1807:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1808:                            (int)strlen(value), value);
                   1809:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1810:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   1811:                        free(call);
                   1812:                        break;
                   1813:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   1814:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   1815:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   1816:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1817:                            (int)rsz, rname);
                   1818:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1819:                        free(call);
                   1820:                        free(rname);
                   1821:                        break;
                   1822:                default:
                   1823:                        break;
                   1824:                }
1.183     schwarze 1825:        }
1.2       schwarze 1826:
1.110     schwarze 1827:        if (*cp == '\0')
1.149     schwarze 1828:                return ROFF_IGN;
1.1       schwarze 1829:
1.93      schwarze 1830:        /* Get the custom end marker. */
1.22      schwarze 1831:
1.93      schwarze 1832:        iname = cp;
1.86      schwarze 1833:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1834:
                   1835:        /* Resolve the end marker if it is indirect. */
                   1836:
1.110     schwarze 1837:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1838:                deftype = ROFFDEF_USER;
                   1839:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1840:                if (name == NULL) {
1.93      schwarze 1841:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1842:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1843:                            "%.*s", (int)namesz, iname);
                   1844:                        namesz = 0;
                   1845:                } else
                   1846:                        namesz = strlen(name);
                   1847:        } else
                   1848:                name = iname;
                   1849:
1.86      schwarze 1850:        if (namesz)
                   1851:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1852:
1.110     schwarze 1853:        if (*cp != '\0')
1.90      schwarze 1854:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167     schwarze 1855:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1856:
1.149     schwarze 1857:        return ROFF_IGN;
1.1       schwarze 1858: }
                   1859:
                   1860: static enum rofferr
1.2       schwarze 1861: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1862: {
1.166     schwarze 1863:        enum roff_tok   t;
1.2       schwarze 1864:        int             i, j;
                   1865:
                   1866:        /*
                   1867:         * First check whether a custom macro exists at this level.  If
                   1868:         * it does, then check against it.  This is some of groff's
                   1869:         * stranger behaviours.  If we encountered a custom end-scope
                   1870:         * tag and that tag also happens to be a "real" macro, then we
                   1871:         * need to try interpreting it again as a real macro.  If it's
                   1872:         * not, then return ignore.  Else continue.
                   1873:         */
                   1874:
                   1875:        if (r->last->end) {
1.35      schwarze 1876:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1877:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1878:                                break;
1.1       schwarze 1879:
1.110     schwarze 1880:                if (r->last->end[j] == '\0' &&
                   1881:                    (buf->buf[i] == '\0' ||
                   1882:                     buf->buf[i] == ' ' ||
                   1883:                     buf->buf[i] == '\t')) {
1.2       schwarze 1884:                        roffnode_pop(r);
                   1885:                        roffnode_cleanscope(r);
1.1       schwarze 1886:
1.110     schwarze 1887:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1888:                                i++;
                   1889:
                   1890:                        pos = i;
1.110     schwarze 1891:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1892:                            TOKEN_NONE)
1.149     schwarze 1893:                                return ROFF_RERUN;
                   1894:                        return ROFF_IGN;
1.2       schwarze 1895:                }
1.1       schwarze 1896:        }
                   1897:
1.2       schwarze 1898:        /*
                   1899:         * If we have no custom end-query or lookup failed, then try
                   1900:         * pulling it out of the hashtable.
                   1901:         */
1.1       schwarze 1902:
1.110     schwarze 1903:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1904:
1.110     schwarze 1905:        if (t != ROFF_cblock) {
                   1906:                if (tok != ROFF_ig)
                   1907:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1908:                return ROFF_IGN;
1.16      schwarze 1909:        }
1.1       schwarze 1910:
1.149     schwarze 1911:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1912: }
                   1913:
                   1914: static enum rofferr
                   1915: roff_block_text(ROFF_ARGS)
                   1916: {
                   1917:
1.110     schwarze 1918:        if (tok != ROFF_ig)
                   1919:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1920:
1.149     schwarze 1921:        return ROFF_IGN;
1.2       schwarze 1922: }
                   1923:
                   1924: static enum rofferr
                   1925: roff_cond_sub(ROFF_ARGS)
                   1926: {
1.166     schwarze 1927:        enum roff_tok    t;
1.37      schwarze 1928:        char            *ep;
1.71      schwarze 1929:        int              rr;
1.2       schwarze 1930:
                   1931:        rr = r->last->rule;
1.37      schwarze 1932:        roffnode_cleanscope(r);
1.39      schwarze 1933:
1.69      schwarze 1934:        /*
                   1935:         * If `\}' occurs on a macro line without a preceding macro,
                   1936:         * drop the line completely.
                   1937:         */
                   1938:
1.110     schwarze 1939:        ep = buf->buf + pos;
                   1940:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1941:                rr = 0;
1.69      schwarze 1942:
1.50      schwarze 1943:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1944:
1.110     schwarze 1945:        while ((ep = strchr(ep, '\\')) != NULL) {
1.190     schwarze 1946:                switch (ep[1]) {
                   1947:                case '}':
                   1948:                        memmove(ep, ep + 2, strlen(ep + 2) + 1);
                   1949:                        roff_ccond(r, ln, ep - buf->buf);
                   1950:                        break;
                   1951:                case '\0':
                   1952:                        ++ep;
                   1953:                        break;
                   1954:                default:
                   1955:                        ep += 2;
                   1956:                        break;
1.70      schwarze 1957:                }
1.50      schwarze 1958:        }
1.190     schwarze 1959:
                   1960:        /*
                   1961:         * Fully handle known macros when they are structurally
                   1962:         * required or when the conditional evaluated to true.
                   1963:         */
                   1964:
                   1965:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1966:        return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
                   1967:            ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
                   1968:            ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1969: }
                   1970:
                   1971: static enum rofferr
                   1972: roff_cond_text(ROFF_ARGS)
                   1973: {
1.37      schwarze 1974:        char            *ep;
1.71      schwarze 1975:        int              rr;
1.2       schwarze 1976:
                   1977:        rr = r->last->rule;
1.37      schwarze 1978:        roffnode_cleanscope(r);
1.1       schwarze 1979:
1.110     schwarze 1980:        ep = buf->buf + pos;
                   1981:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1982:                if (*(++ep) == '}') {
1.70      schwarze 1983:                        *ep = '&';
1.110     schwarze 1984:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1985:                }
1.119     schwarze 1986:                if (*ep != '\0')
                   1987:                        ++ep;
1.2       schwarze 1988:        }
1.149     schwarze 1989:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1990: }
                   1991:
1.138     schwarze 1992: /* --- handling of numeric and conditional expressions -------------------- */
                   1993:
1.77      schwarze 1994: /*
                   1995:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1996:  * If there is at least one digit, return success and advance the
                   1997:  * parse point, else return failure and let the parse point unchanged.
                   1998:  * Ignore overflows, treat them just like the C language.
                   1999:  */
1.56      schwarze 2000: static int
1.133     schwarze 2001: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 2002: {
1.133     schwarze 2003:        int      myres, scaled, n, p;
1.79      schwarze 2004:
                   2005:        if (NULL == res)
                   2006:                res = &myres;
1.56      schwarze 2007:
                   2008:        p = *pos;
                   2009:        n = v[p] == '-';
1.133     schwarze 2010:        if (n || v[p] == '+')
1.56      schwarze 2011:                p++;
                   2012:
1.133     schwarze 2013:        if (flags & ROFFNUM_WHITE)
                   2014:                while (isspace((unsigned char)v[p]))
                   2015:                        p++;
                   2016:
1.56      schwarze 2017:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 2018:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 2019:        if (p == *pos + n)
                   2020:                return 0;
                   2021:
                   2022:        if (n)
                   2023:                *res = -*res;
                   2024:
1.126     schwarze 2025:        /* Each number may be followed by one optional scaling unit. */
                   2026:
                   2027:        switch (v[p]) {
                   2028:        case 'f':
1.133     schwarze 2029:                scaled = *res * 65536;
1.126     schwarze 2030:                break;
                   2031:        case 'i':
1.133     schwarze 2032:                scaled = *res * 240;
1.126     schwarze 2033:                break;
                   2034:        case 'c':
1.133     schwarze 2035:                scaled = *res * 240 / 2.54;
1.126     schwarze 2036:                break;
                   2037:        case 'v':
                   2038:        case 'P':
1.133     schwarze 2039:                scaled = *res * 40;
1.126     schwarze 2040:                break;
                   2041:        case 'm':
                   2042:        case 'n':
1.133     schwarze 2043:                scaled = *res * 24;
1.126     schwarze 2044:                break;
                   2045:        case 'p':
1.133     schwarze 2046:                scaled = *res * 10 / 3;
1.126     schwarze 2047:                break;
                   2048:        case 'u':
1.133     schwarze 2049:                scaled = *res;
1.126     schwarze 2050:                break;
                   2051:        case 'M':
1.133     schwarze 2052:                scaled = *res * 6 / 25;
1.126     schwarze 2053:                break;
                   2054:        default:
1.133     schwarze 2055:                scaled = *res;
1.126     schwarze 2056:                p--;
                   2057:                break;
                   2058:        }
1.133     schwarze 2059:        if (flags & ROFFNUM_SCALE)
                   2060:                *res = scaled;
1.126     schwarze 2061:
                   2062:        *pos = p + 1;
1.149     schwarze 2063:        return 1;
1.56      schwarze 2064: }
                   2065:
1.71      schwarze 2066: /*
                   2067:  * Evaluate a string comparison condition.
                   2068:  * The first character is the delimiter.
                   2069:  * Succeed if the string up to its second occurrence
                   2070:  * matches the string up to its third occurence.
                   2071:  * Advance the cursor after the third occurrence
                   2072:  * or lacking that, to the end of the line.
                   2073:  */
                   2074: static int
                   2075: roff_evalstrcond(const char *v, int *pos)
                   2076: {
                   2077:        const char      *s1, *s2, *s3;
                   2078:        int              match;
                   2079:
                   2080:        match = 0;
                   2081:        s1 = v + *pos;          /* initial delimiter */
                   2082:        s2 = s1 + 1;            /* for scanning the first string */
                   2083:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2084:
                   2085:        if (NULL == s3)         /* found no middle delimiter */
                   2086:                goto out;
                   2087:
                   2088:        while ('\0' != *++s3) {
                   2089:                if (*s2 != *s3) {  /* mismatch */
                   2090:                        s3 = strchr(s3, *s1);
                   2091:                        break;
                   2092:                }
                   2093:                if (*s3 == *s1) {  /* found the final delimiter */
                   2094:                        match = 1;
                   2095:                        break;
                   2096:                }
                   2097:                s2++;
                   2098:        }
                   2099:
                   2100: out:
                   2101:        if (NULL == s3)
                   2102:                s3 = strchr(s2, '\0');
1.114     schwarze 2103:        else if (*s3 != '\0')
1.71      schwarze 2104:                s3++;
                   2105:        *pos = s3 - v;
1.149     schwarze 2106:        return match;
1.71      schwarze 2107: }
                   2108:
1.77      schwarze 2109: /*
                   2110:  * Evaluate an optionally negated single character, numerical,
                   2111:  * or string condition.
                   2112:  */
1.71      schwarze 2113: static int
1.143     schwarze 2114: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 2115: {
1.143     schwarze 2116:        char    *cp, *name;
                   2117:        size_t   sz;
1.187     schwarze 2118:        int      deftype, number, savepos, istrue, wanttrue;
1.5       schwarze 2119:
1.71      schwarze 2120:        if ('!' == v[*pos]) {
                   2121:                wanttrue = 0;
                   2122:                (*pos)++;
                   2123:        } else
                   2124:                wanttrue = 1;
                   2125:
1.5       schwarze 2126:        switch (v[*pos]) {
1.112     schwarze 2127:        case '\0':
1.149     schwarze 2128:                return 0;
1.80      schwarze 2129:        case 'n':
                   2130:        case 'o':
1.5       schwarze 2131:                (*pos)++;
1.149     schwarze 2132:                return wanttrue;
1.80      schwarze 2133:        case 'c':
                   2134:        case 'e':
                   2135:        case 't':
1.111     schwarze 2136:        case 'v':
1.5       schwarze 2137:                (*pos)++;
1.149     schwarze 2138:                return !wanttrue;
1.182     schwarze 2139:        case 'd':
1.143     schwarze 2140:        case 'r':
1.182     schwarze 2141:                cp = v + *pos + 1;
                   2142:                while (*cp == ' ')
                   2143:                        cp++;
                   2144:                name = cp;
                   2145:                sz = roff_getname(r, &cp, ln, cp - v);
1.187     schwarze 2146:                if (sz == 0)
                   2147:                        istrue = 0;
                   2148:                else if (v[*pos] == 'r')
                   2149:                        istrue = roff_hasregn(r, name, sz);
                   2150:                else {
                   2151:                        deftype = ROFFDEF_ANY;
                   2152:                        roff_getstrn(r, name, sz, &deftype);
                   2153:                        istrue = !!deftype;
                   2154:                }
1.143     schwarze 2155:                *pos = cp - v;
1.182     schwarze 2156:                return istrue == wanttrue;
1.5       schwarze 2157:        default:
                   2158:                break;
                   2159:        }
                   2160:
1.113     schwarze 2161:        savepos = *pos;
1.133     schwarze 2162:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 2163:                return (number > 0) == wanttrue;
1.113     schwarze 2164:        else if (*pos == savepos)
1.149     schwarze 2165:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 2166:        else
1.149     schwarze 2167:                return 0;
1.5       schwarze 2168: }
                   2169:
1.2       schwarze 2170: static enum rofferr
1.21      schwarze 2171: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 2172: {
1.30      schwarze 2173:
1.149     schwarze 2174:        return ROFF_IGN;
1.21      schwarze 2175: }
                   2176:
                   2177: static enum rofferr
1.123     schwarze 2178: roff_insec(ROFF_ARGS)
                   2179: {
                   2180:
                   2181:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167     schwarze 2182:            ln, ppos, roff_name[tok]);
1.149     schwarze 2183:        return ROFF_IGN;
1.123     schwarze 2184: }
                   2185:
                   2186: static enum rofferr
                   2187: roff_unsupp(ROFF_ARGS)
                   2188: {
                   2189:
                   2190:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167     schwarze 2191:            ln, ppos, roff_name[tok]);
1.149     schwarze 2192:        return ROFF_IGN;
1.123     schwarze 2193: }
                   2194:
                   2195: static enum rofferr
1.2       schwarze 2196: roff_cond(ROFF_ARGS)
                   2197: {
1.46      schwarze 2198:
                   2199:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2200:
1.80      schwarze 2201:        /*
1.35      schwarze 2202:         * An `.el' has no conditional body: it will consume the value
                   2203:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2204:         * defaults to DENY.
1.35      schwarze 2205:         *
                   2206:         * If we're not an `el', however, then evaluate the conditional.
                   2207:         */
1.1       schwarze 2208:
1.110     schwarze 2209:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2210:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2211:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2212:
1.35      schwarze 2213:        /*
                   2214:         * An if-else will put the NEGATION of the current evaluated
                   2215:         * conditional into the stack of rules.
                   2216:         */
                   2217:
1.110     schwarze 2218:        if (tok == ROFF_ie) {
1.96      schwarze 2219:                if (r->rstackpos + 1 == r->rstacksz) {
                   2220:                        r->rstacksz += 16;
                   2221:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2222:                            r->rstacksz, sizeof(int));
1.35      schwarze 2223:                }
1.71      schwarze 2224:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2225:        }
1.5       schwarze 2226:
                   2227:        /* If the parent has false as its rule, then so do we. */
                   2228:
1.71      schwarze 2229:        if (r->last->parent && !r->last->parent->rule)
                   2230:                r->last->rule = 0;
1.5       schwarze 2231:
                   2232:        /*
1.46      schwarze 2233:         * Determine scope.
                   2234:         * If there is nothing on the line after the conditional,
                   2235:         * not even whitespace, use next-line scope.
1.5       schwarze 2236:         */
1.2       schwarze 2237:
1.110     schwarze 2238:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2239:                r->last->endspan = 2;
                   2240:                goto out;
                   2241:        }
                   2242:
1.110     schwarze 2243:        while (buf->buf[pos] == ' ')
1.46      schwarze 2244:                pos++;
                   2245:
                   2246:        /* An opening brace requests multiline scope. */
1.2       schwarze 2247:
1.110     schwarze 2248:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2249:                r->last->endspan = -1;
                   2250:                pos += 2;
1.144     schwarze 2251:                while (buf->buf[pos] == ' ')
                   2252:                        pos++;
1.46      schwarze 2253:                goto out;
1.80      schwarze 2254:        }
1.2       schwarze 2255:
                   2256:        /*
1.46      schwarze 2257:         * Anything else following the conditional causes
                   2258:         * single-line scope.  Warn if the scope contains
                   2259:         * nothing but trailing whitespace.
1.2       schwarze 2260:         */
                   2261:
1.110     schwarze 2262:        if (buf->buf[pos] == '\0')
1.89      schwarze 2263:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167     schwarze 2264:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2265:
1.46      schwarze 2266:        r->last->endspan = 1;
1.1       schwarze 2267:
1.46      schwarze 2268: out:
1.2       schwarze 2269:        *offs = pos;
1.149     schwarze 2270:        return ROFF_RERUN;
1.1       schwarze 2271: }
                   2272:
1.2       schwarze 2273: static enum rofferr
1.7       schwarze 2274: roff_ds(ROFF_ARGS)
                   2275: {
1.85      schwarze 2276:        char            *string;
                   2277:        const char      *name;
                   2278:        size_t           namesz;
1.10      schwarze 2279:
1.123     schwarze 2280:        /* Ignore groff compatibility mode for now. */
                   2281:
                   2282:        if (tok == ROFF_ds1)
                   2283:                tok = ROFF_ds;
                   2284:        else if (tok == ROFF_as1)
                   2285:                tok = ROFF_as;
                   2286:
1.10      schwarze 2287:        /*
1.85      schwarze 2288:         * The first word is the name of the string.
                   2289:         * If it is empty or terminated by an escape sequence,
                   2290:         * abort the `ds' request without defining anything.
1.10      schwarze 2291:         */
1.7       schwarze 2292:
1.110     schwarze 2293:        name = string = buf->buf + pos;
                   2294:        if (*name == '\0')
1.149     schwarze 2295:                return ROFF_IGN;
1.7       schwarze 2296:
1.85      schwarze 2297:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2298:        if (name[namesz] == '\\')
1.149     schwarze 2299:                return ROFF_IGN;
1.85      schwarze 2300:
                   2301:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2302:        if (*string == '"')
1.7       schwarze 2303:                string++;
                   2304:
1.10      schwarze 2305:        /* The rest is the value. */
1.85      schwarze 2306:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2307:            ROFF_as == tok);
1.183     schwarze 2308:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149     schwarze 2309:        return ROFF_IGN;
1.7       schwarze 2310: }
                   2311:
1.77      schwarze 2312: /*
                   2313:  * Parse a single operator, one or two characters long.
                   2314:  * If the operator is recognized, return success and advance the
                   2315:  * parse point, else return failure and let the parse point unchanged.
                   2316:  */
                   2317: static int
                   2318: roff_getop(const char *v, int *pos, char *res)
                   2319: {
                   2320:
                   2321:        *res = v[*pos];
                   2322:
                   2323:        switch (*res) {
1.80      schwarze 2324:        case '+':
                   2325:        case '-':
                   2326:        case '*':
                   2327:        case '/':
                   2328:        case '%':
                   2329:        case '&':
                   2330:        case ':':
1.77      schwarze 2331:                break;
                   2332:        case '<':
                   2333:                switch (v[*pos + 1]) {
1.80      schwarze 2334:                case '=':
1.77      schwarze 2335:                        *res = 'l';
                   2336:                        (*pos)++;
                   2337:                        break;
1.80      schwarze 2338:                case '>':
1.77      schwarze 2339:                        *res = '!';
                   2340:                        (*pos)++;
                   2341:                        break;
1.80      schwarze 2342:                case '?':
1.77      schwarze 2343:                        *res = 'i';
                   2344:                        (*pos)++;
                   2345:                        break;
                   2346:                default:
                   2347:                        break;
                   2348:                }
                   2349:                break;
                   2350:        case '>':
                   2351:                switch (v[*pos + 1]) {
1.80      schwarze 2352:                case '=':
1.77      schwarze 2353:                        *res = 'g';
                   2354:                        (*pos)++;
                   2355:                        break;
1.80      schwarze 2356:                case '?':
1.77      schwarze 2357:                        *res = 'a';
                   2358:                        (*pos)++;
                   2359:                        break;
                   2360:                default:
                   2361:                        break;
                   2362:                }
                   2363:                break;
                   2364:        case '=':
                   2365:                if ('=' == v[*pos + 1])
                   2366:                        (*pos)++;
                   2367:                break;
                   2368:        default:
1.149     schwarze 2369:                return 0;
1.77      schwarze 2370:        }
                   2371:        (*pos)++;
                   2372:
1.149     schwarze 2373:        return *res;
1.77      schwarze 2374: }
                   2375:
                   2376: /*
                   2377:  * Evaluate either a parenthesized numeric expression
                   2378:  * or a single signed integer number.
                   2379:  */
                   2380: static int
1.106     schwarze 2381: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2382:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2383: {
                   2384:
                   2385:        if ('(' != v[*pos])
1.149     schwarze 2386:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2387:
                   2388:        (*pos)++;
1.133     schwarze 2389:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2390:                return 0;
1.77      schwarze 2391:
1.79      schwarze 2392:        /*
                   2393:         * Omission of the closing parenthesis
                   2394:         * is an error in validation mode,
                   2395:         * but ignored in evaluation mode.
                   2396:         */
                   2397:
1.77      schwarze 2398:        if (')' == v[*pos])
                   2399:                (*pos)++;
1.79      schwarze 2400:        else if (NULL == res)
1.149     schwarze 2401:                return 0;
1.77      schwarze 2402:
1.149     schwarze 2403:        return 1;
1.77      schwarze 2404: }
                   2405:
                   2406: /*
                   2407:  * Evaluate a complete numeric expression.
                   2408:  * Proceed left to right, there is no concept of precedence.
                   2409:  */
                   2410: static int
1.106     schwarze 2411: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2412:        int *pos, int *res, int flags)
1.77      schwarze 2413: {
                   2414:        int              mypos, operand2;
                   2415:        char             operator;
                   2416:
                   2417:        if (NULL == pos) {
                   2418:                mypos = 0;
                   2419:                pos = &mypos;
                   2420:        }
                   2421:
1.133     schwarze 2422:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2423:                while (isspace((unsigned char)v[*pos]))
                   2424:                        (*pos)++;
                   2425:
1.133     schwarze 2426:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2427:                return 0;
1.77      schwarze 2428:
                   2429:        while (1) {
1.133     schwarze 2430:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2431:                        while (isspace((unsigned char)v[*pos]))
                   2432:                                (*pos)++;
                   2433:
                   2434:                if ( ! roff_getop(v, pos, &operator))
                   2435:                        break;
                   2436:
1.133     schwarze 2437:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2438:                        while (isspace((unsigned char)v[*pos]))
                   2439:                                (*pos)++;
                   2440:
1.133     schwarze 2441:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2442:                        return 0;
1.77      schwarze 2443:
1.133     schwarze 2444:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2445:                        while (isspace((unsigned char)v[*pos]))
                   2446:                                (*pos)++;
1.79      schwarze 2447:
                   2448:                if (NULL == res)
                   2449:                        continue;
1.77      schwarze 2450:
                   2451:                switch (operator) {
1.80      schwarze 2452:                case '+':
1.77      schwarze 2453:                        *res += operand2;
                   2454:                        break;
1.80      schwarze 2455:                case '-':
1.77      schwarze 2456:                        *res -= operand2;
                   2457:                        break;
1.80      schwarze 2458:                case '*':
1.77      schwarze 2459:                        *res *= operand2;
                   2460:                        break;
1.80      schwarze 2461:                case '/':
1.116     schwarze 2462:                        if (operand2 == 0) {
1.106     schwarze 2463:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2464:                                        r->parse, ln, *pos, v);
                   2465:                                *res = 0;
                   2466:                                break;
                   2467:                        }
1.77      schwarze 2468:                        *res /= operand2;
                   2469:                        break;
1.80      schwarze 2470:                case '%':
1.116     schwarze 2471:                        if (operand2 == 0) {
                   2472:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2473:                                        r->parse, ln, *pos, v);
                   2474:                                *res = 0;
                   2475:                                break;
                   2476:                        }
1.77      schwarze 2477:                        *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 'l':
1.77      schwarze 2486:                        *res = *res <= operand2;
                   2487:                        break;
1.80      schwarze 2488:                case 'g':
1.77      schwarze 2489:                        *res = *res >= operand2;
                   2490:                        break;
1.80      schwarze 2491:                case '=':
1.77      schwarze 2492:                        *res = *res == operand2;
                   2493:                        break;
1.80      schwarze 2494:                case '!':
1.77      schwarze 2495:                        *res = *res != operand2;
                   2496:                        break;
1.80      schwarze 2497:                case '&':
1.77      schwarze 2498:                        *res = *res && operand2;
                   2499:                        break;
1.80      schwarze 2500:                case ':':
1.77      schwarze 2501:                        *res = *res || operand2;
                   2502:                        break;
1.80      schwarze 2503:                case 'i':
1.77      schwarze 2504:                        if (operand2 < *res)
                   2505:                                *res = operand2;
                   2506:                        break;
1.80      schwarze 2507:                case 'a':
1.77      schwarze 2508:                        if (operand2 > *res)
                   2509:                                *res = operand2;
                   2510:                        break;
                   2511:                default:
                   2512:                        abort();
                   2513:                }
                   2514:        }
1.149     schwarze 2515:        return 1;
1.77      schwarze 2516: }
                   2517:
1.138     schwarze 2518: /* --- register management ------------------------------------------------ */
                   2519:
1.52      schwarze 2520: void
1.60      schwarze 2521: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2522: {
1.198   ! schwarze 2523:        roff_setregn(r, name, strlen(name), val, sign);
        !          2524: }
        !          2525:
        !          2526: static void
        !          2527: roff_setregn(struct roff *r, const char *name, size_t len,
        !          2528:     int val, char sign)
        !          2529: {
1.52      schwarze 2530:        struct roffreg  *reg;
                   2531:
                   2532:        /* Search for an existing register with the same name. */
                   2533:        reg = r->regtab;
                   2534:
1.198   ! schwarze 2535:        while (reg != NULL && (reg->key.sz != len ||
        !          2536:            strncmp(reg->key.p, name, len) != 0))
1.52      schwarze 2537:                reg = reg->next;
1.41      schwarze 2538:
1.52      schwarze 2539:        if (NULL == reg) {
                   2540:                /* Create a new register. */
                   2541:                reg = mandoc_malloc(sizeof(struct roffreg));
1.198   ! schwarze 2542:                reg->key.p = mandoc_strndup(name, len);
        !          2543:                reg->key.sz = len;
1.60      schwarze 2544:                reg->val = 0;
1.52      schwarze 2545:                reg->next = r->regtab;
                   2546:                r->regtab = reg;
                   2547:        }
                   2548:
1.60      schwarze 2549:        if ('+' == sign)
                   2550:                reg->val += val;
                   2551:        else if ('-' == sign)
                   2552:                reg->val -= val;
                   2553:        else
                   2554:                reg->val = val;
1.41      schwarze 2555: }
                   2556:
1.65      schwarze 2557: /*
                   2558:  * Handle some predefined read-only number registers.
                   2559:  * For now, return -1 if the requested register is not predefined;
                   2560:  * in case a predefined read-only register having the value -1
                   2561:  * were to turn up, another special value would have to be chosen.
                   2562:  */
                   2563: static int
1.145     schwarze 2564: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2565: {
                   2566:
                   2567:        switch (*name) {
1.145     schwarze 2568:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2569:                return r->argc;
1.80      schwarze 2570:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2571:                return 0;
1.80      schwarze 2572:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2573:                return 1;
1.80      schwarze 2574:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2575:                return 24;
1.80      schwarze 2576:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2577:                return 0;
1.80      schwarze 2578:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2579:                return 1;
1.80      schwarze 2580:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2581:                return 40;
1.65      schwarze 2582:        default:
1.149     schwarze 2583:                return -1;
1.65      schwarze 2584:        }
                   2585: }
                   2586:
1.53      schwarze 2587: int
1.198   ! schwarze 2588: roff_getreg(struct roff *r, const char *name)
1.41      schwarze 2589: {
1.198   ! schwarze 2590:        return roff_getregn(r, name, strlen(name));
1.53      schwarze 2591: }
                   2592:
                   2593: static int
1.198   ! schwarze 2594: roff_getregn(struct roff *r, const char *name, size_t len)
1.53      schwarze 2595: {
                   2596:        struct roffreg  *reg;
1.65      schwarze 2597:        int              val;
                   2598:
                   2599:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2600:                val = roff_getregro(r, name + 1);
1.65      schwarze 2601:                if (-1 != val)
1.149     schwarze 2602:                        return val;
1.65      schwarze 2603:        }
1.53      schwarze 2604:
                   2605:        for (reg = r->regtab; reg; reg = reg->next)
                   2606:                if (len == reg->key.sz &&
                   2607:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2608:                        return reg->val;
1.143     schwarze 2609:
1.198   ! schwarze 2610:        roff_setregn(r, name, len, 0, '\0');
1.149     schwarze 2611:        return 0;
1.143     schwarze 2612: }
                   2613:
                   2614: static int
                   2615: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2616: {
                   2617:        struct roffreg  *reg;
                   2618:        int              val;
                   2619:
                   2620:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2621:                val = roff_getregro(r, name + 1);
1.143     schwarze 2622:                if (-1 != val)
1.149     schwarze 2623:                        return 1;
1.143     schwarze 2624:        }
                   2625:
                   2626:        for (reg = r->regtab; reg; reg = reg->next)
                   2627:                if (len == reg->key.sz &&
                   2628:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2629:                        return 1;
1.41      schwarze 2630:
1.149     schwarze 2631:        return 0;
1.41      schwarze 2632: }
                   2633:
1.52      schwarze 2634: static void
                   2635: roff_freereg(struct roffreg *reg)
1.41      schwarze 2636: {
1.52      schwarze 2637:        struct roffreg  *old_reg;
1.41      schwarze 2638:
1.52      schwarze 2639:        while (NULL != reg) {
                   2640:                free(reg->key.p);
                   2641:                old_reg = reg;
                   2642:                reg = reg->next;
                   2643:                free(old_reg);
                   2644:        }
1.41      schwarze 2645: }
1.7       schwarze 2646:
                   2647: static enum rofferr
1.6       schwarze 2648: roff_nr(ROFF_ARGS)
1.1       schwarze 2649: {
1.85      schwarze 2650:        char            *key, *val;
                   2651:        size_t           keysz;
1.37      schwarze 2652:        int              iv;
1.60      schwarze 2653:        char             sign;
1.6       schwarze 2654:
1.110     schwarze 2655:        key = val = buf->buf + pos;
                   2656:        if (*key == '\0')
1.149     schwarze 2657:                return ROFF_IGN;
1.85      schwarze 2658:
                   2659:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2660:        if (key[keysz] == '\\')
1.149     schwarze 2661:                return ROFF_IGN;
1.6       schwarze 2662:
1.60      schwarze 2663:        sign = *val;
1.110     schwarze 2664:        if (sign == '+' || sign == '-')
1.60      schwarze 2665:                val++;
                   2666:
1.133     schwarze 2667:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.198   ! schwarze 2668:                roff_setregn(r, key, keysz, iv, sign);
1.1       schwarze 2669:
1.149     schwarze 2670:        return ROFF_IGN;
1.76      schwarze 2671: }
                   2672:
                   2673: static enum rofferr
                   2674: roff_rr(ROFF_ARGS)
                   2675: {
                   2676:        struct roffreg  *reg, **prev;
1.85      schwarze 2677:        char            *name, *cp;
                   2678:        size_t           namesz;
1.76      schwarze 2679:
1.110     schwarze 2680:        name = cp = buf->buf + pos;
                   2681:        if (*name == '\0')
1.149     schwarze 2682:                return ROFF_IGN;
1.85      schwarze 2683:        namesz = roff_getname(r, &cp, ln, pos);
                   2684:        name[namesz] = '\0';
1.76      schwarze 2685:
                   2686:        prev = &r->regtab;
                   2687:        while (1) {
                   2688:                reg = *prev;
1.110     schwarze 2689:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2690:                        break;
                   2691:                prev = &reg->next;
                   2692:        }
1.110     schwarze 2693:        if (reg != NULL) {
1.76      schwarze 2694:                *prev = reg->next;
                   2695:                free(reg->key.p);
                   2696:                free(reg);
                   2697:        }
1.149     schwarze 2698:        return ROFF_IGN;
1.29      schwarze 2699: }
                   2700:
1.138     schwarze 2701: /* --- handler functions for roff requests -------------------------------- */
                   2702:
1.29      schwarze 2703: static enum rofferr
                   2704: roff_rm(ROFF_ARGS)
                   2705: {
                   2706:        const char       *name;
                   2707:        char             *cp;
1.85      schwarze 2708:        size_t            namesz;
1.29      schwarze 2709:
1.110     schwarze 2710:        cp = buf->buf + pos;
                   2711:        while (*cp != '\0') {
1.85      schwarze 2712:                name = cp;
1.110     schwarze 2713:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2714:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183     schwarze 2715:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.110     schwarze 2716:                if (name[namesz] == '\\')
1.85      schwarze 2717:                        break;
1.29      schwarze 2718:        }
1.149     schwarze 2719:        return ROFF_IGN;
1.51      schwarze 2720: }
                   2721:
                   2722: static enum rofferr
                   2723: roff_it(ROFF_ARGS)
                   2724: {
                   2725:        int              iv;
                   2726:
                   2727:        /* Parse the number of lines. */
1.133     schwarze 2728:
                   2729:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2730:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2731:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2732:                return ROFF_IGN;
1.51      schwarze 2733:        }
                   2734:
1.134     schwarze 2735:        while (isspace((unsigned char)buf->buf[pos]))
                   2736:                pos++;
                   2737:
                   2738:        /*
                   2739:         * Arm the input line trap.
                   2740:         * Special-casing "an-trap" is an ugly workaround to cope
                   2741:         * with DocBook stupidly fiddling with man(7) internals.
                   2742:         */
1.133     schwarze 2743:
1.51      schwarze 2744:        roffit_lines = iv;
1.134     schwarze 2745:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2746:            strcmp(buf->buf + pos, "an-trap") ?
                   2747:            buf->buf + pos : "br");
1.149     schwarze 2748:        return ROFF_IGN;
1.47      schwarze 2749: }
                   2750:
                   2751: static enum rofferr
                   2752: roff_Dd(ROFF_ARGS)
                   2753: {
1.187     schwarze 2754:        int              mask;
                   2755:        enum roff_tok    t, te;
1.99      schwarze 2756:
1.187     schwarze 2757:        switch (tok) {
                   2758:        case ROFF_Dd:
                   2759:                tok = MDOC_Dd;
                   2760:                te = MDOC_MAX;
                   2761:                if (r->format == 0)
                   2762:                        r->format = MPARSE_MDOC;
                   2763:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   2764:                break;
                   2765:        case ROFF_TH:
                   2766:                tok = MAN_TH;
                   2767:                te = MAN_MAX;
                   2768:                if (r->format == 0)
                   2769:                        r->format = MPARSE_MAN;
                   2770:                mask = MPARSE_QUICK;
                   2771:                break;
                   2772:        default:
                   2773:                abort();
                   2774:        }
                   2775:        if ((r->options & mask) == 0)
                   2776:                for (t = tok; t < te; t++)
                   2777:                        roff_setstr(r, roff_name[t], NULL, 0);
1.149     schwarze 2778:        return ROFF_CONT;
1.14      schwarze 2779: }
                   2780:
                   2781: static enum rofferr
1.27      schwarze 2782: roff_TE(ROFF_ARGS)
                   2783: {
1.193     schwarze 2784:        if (r->tbl == NULL) {
1.94      schwarze 2785:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2786:                    ln, ppos, "TE");
1.193     schwarze 2787:                return ROFF_IGN;
                   2788:        }
                   2789:        if (tbl_end(r->tbl) == 0) {
                   2790:                r->tbl = NULL;
1.130     schwarze 2791:                free(buf->buf);
                   2792:                buf->buf = mandoc_strdup(".sp");
                   2793:                buf->sz = 4;
1.149     schwarze 2794:                return ROFF_REPARSE;
1.130     schwarze 2795:        }
1.193     schwarze 2796:        r->tbl = NULL;
1.149     schwarze 2797:        return ROFF_IGN;
1.27      schwarze 2798: }
                   2799:
                   2800: static enum rofferr
                   2801: roff_T_(ROFF_ARGS)
                   2802: {
                   2803:
                   2804:        if (NULL == r->tbl)
1.94      schwarze 2805:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2806:                    ln, ppos, "T&");
1.27      schwarze 2807:        else
1.168     schwarze 2808:                tbl_restart(ln, ppos, r->tbl);
1.27      schwarze 2809:
1.149     schwarze 2810:        return ROFF_IGN;
1.27      schwarze 2811: }
                   2812:
1.102     schwarze 2813: /*
                   2814:  * Handle in-line equation delimiters.
                   2815:  */
                   2816: static enum rofferr
1.110     schwarze 2817: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2818: {
1.105     schwarze 2819:        char            *cp1, *cp2;
                   2820:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2821:
1.102     schwarze 2822:        /*
                   2823:         * Outside equations, look for an opening delimiter.
                   2824:         * If we are inside an equation, we already know it is
                   2825:         * in-line, or this function wouldn't have been called;
                   2826:         * so look for a closing delimiter.
                   2827:         */
                   2828:
1.110     schwarze 2829:        cp1 = buf->buf + pos;
1.102     schwarze 2830:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2831:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2832:        if (cp2 == NULL)
1.149     schwarze 2833:                return ROFF_CONT;
1.102     schwarze 2834:
1.105     schwarze 2835:        *cp2++ = '\0';
                   2836:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2837:
                   2838:        /* Handle preceding text, protecting whitespace. */
                   2839:
1.110     schwarze 2840:        if (*buf->buf != '\0') {
1.105     schwarze 2841:                if (r->eqn == NULL)
                   2842:                        bef_pr = "\\&";
                   2843:                bef_nl = "\n";
                   2844:        }
                   2845:
                   2846:        /*
                   2847:         * Prepare replacing the delimiter with an equation macro
                   2848:         * and drop leading white space from the equation.
                   2849:         */
1.102     schwarze 2850:
1.105     schwarze 2851:        if (r->eqn == NULL) {
                   2852:                while (*cp2 == ' ')
                   2853:                        cp2++;
                   2854:                mac = ".EQ";
                   2855:        } else
                   2856:                mac = ".EN";
                   2857:
                   2858:        /* Handle following text, protecting whitespace. */
                   2859:
                   2860:        if (*cp2 != '\0') {
                   2861:                aft_nl = "\n";
                   2862:                if (r->eqn != NULL)
                   2863:                        aft_pr = "\\&";
                   2864:        }
                   2865:
                   2866:        /* Do the actual replacement. */
                   2867:
1.110     schwarze 2868:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2869:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2870:        free(buf->buf);
                   2871:        buf->buf = cp1;
1.102     schwarze 2872:
                   2873:        /* Toggle the in-line state of the eqn subsystem. */
                   2874:
                   2875:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2876:        return ROFF_REPARSE;
1.41      schwarze 2877: }
                   2878:
1.107     schwarze 2879: static enum rofferr
                   2880: roff_EQ(ROFF_ARGS)
1.32      schwarze 2881: {
1.191     schwarze 2882:        struct roff_node        *n;
                   2883:
1.194     schwarze 2884:        if (r->man->macroset == MACROSET_MAN)
                   2885:                man_breakscope(r->man, ROFF_EQ);
1.191     schwarze 2886:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   2887:        if (ln > r->man->last->line)
                   2888:                n->flags |= NODE_LINE;
                   2889:        n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
                   2890:        n->eqn->expectargs = UINT_MAX;
                   2891:        roff_node_append(r->man, n);
                   2892:        r->man->next = ROFF_NEXT_SIBLING;
1.32      schwarze 2893:
1.110     schwarze 2894:        assert(r->eqn == NULL);
1.191     schwarze 2895:        if (r->last_eqn == NULL)
                   2896:                r->last_eqn = eqn_alloc(r->parse);
                   2897:        else
                   2898:                eqn_reset(r->last_eqn);
                   2899:        r->eqn = r->last_eqn;
                   2900:        r->eqn->node = n;
1.41      schwarze 2901:
1.110     schwarze 2902:        if (buf->buf[pos] != '\0')
1.107     schwarze 2903:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2904:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2905:
1.149     schwarze 2906:        return ROFF_IGN;
1.32      schwarze 2907: }
                   2908:
                   2909: static enum rofferr
                   2910: roff_EN(ROFF_ARGS)
                   2911: {
1.191     schwarze 2912:        if (r->eqn != NULL) {
                   2913:                eqn_parse(r->eqn);
                   2914:                r->eqn = NULL;
                   2915:        } else
                   2916:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
                   2917:        if (buf->buf[pos] != '\0')
                   2918:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   2919:                    "EN %s", buf->buf + pos);
1.149     schwarze 2920:        return ROFF_IGN;
1.32      schwarze 2921: }
                   2922:
                   2923: static enum rofferr
1.27      schwarze 2924: roff_TS(ROFF_ARGS)
                   2925: {
1.193     schwarze 2926:        if (r->tbl != NULL) {
1.94      schwarze 2927:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2928:                    ln, ppos, "TS breaks TS");
1.193     schwarze 2929:                tbl_end(r->tbl);
1.27      schwarze 2930:        }
1.193     schwarze 2931:        r->tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2932:        if (r->last_tbl)
1.193     schwarze 2933:                r->last_tbl->next = r->tbl;
1.27      schwarze 2934:        else
1.193     schwarze 2935:                r->first_tbl = r->tbl;
                   2936:        r->last_tbl = r->tbl;
1.169     schwarze 2937:        return ROFF_IGN;
                   2938: }
                   2939:
                   2940: static enum rofferr
                   2941: roff_onearg(ROFF_ARGS)
                   2942: {
                   2943:        struct roff_node        *n;
                   2944:        char                    *cp;
1.177     schwarze 2945:        int                      npos;
1.169     schwarze 2946:
1.174     schwarze 2947:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.192     schwarze 2948:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
                   2949:             tok == ROFF_ti))
1.174     schwarze 2950:                man_breakscope(r->man, tok);
                   2951:
1.181     schwarze 2952:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177     schwarze 2953:                r->man->last = roffce_node;
                   2954:                r->man->next = ROFF_NEXT_SIBLING;
                   2955:        }
                   2956:
1.169     schwarze 2957:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2958:        n = r->man->last;
                   2959:
                   2960:        cp = buf->buf + pos;
                   2961:        if (*cp != '\0') {
                   2962:                while (*cp != '\0' && *cp != ' ')
                   2963:                        cp++;
                   2964:                while (*cp == ' ')
                   2965:                        *cp++ = '\0';
                   2966:                if (*cp != '\0')
                   2967:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS,
                   2968:                            r->parse, ln, cp - buf->buf,
                   2969:                            "%s ... %s", roff_name[tok], cp);
                   2970:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172     schwarze 2971:        }
                   2972:
1.181     schwarze 2973:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   2974:                if (r->man->last->type == ROFFT_ELEM) {
1.177     schwarze 2975:                        roff_word_alloc(r->man, ln, pos, "1");
                   2976:                        r->man->last->flags |= NODE_NOSRC;
                   2977:                }
                   2978:                npos = 0;
                   2979:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   2980:                    &roffce_lines, 0) == 0) {
                   2981:                        mandoc_vmsg(MANDOCERR_CE_NONUM,
                   2982:                            r->parse, ln, pos, "ce %s", buf->buf + pos);
                   2983:                        roffce_lines = 1;
                   2984:                }
                   2985:                if (roffce_lines < 1) {
                   2986:                        r->man->last = r->man->last->parent;
                   2987:                        roffce_node = NULL;
                   2988:                        roffce_lines = 0;
                   2989:                } else
                   2990:                        roffce_node = r->man->last->parent;
                   2991:        } else {
                   2992:                n->flags |= NODE_VALID | NODE_ENDED;
                   2993:                r->man->last = n;
                   2994:        }
                   2995:        n->flags |= NODE_LINE;
1.172     schwarze 2996:        r->man->next = ROFF_NEXT_SIBLING;
                   2997:        return ROFF_IGN;
                   2998: }
                   2999:
                   3000: static enum rofferr
                   3001: roff_manyarg(ROFF_ARGS)
                   3002: {
                   3003:        struct roff_node        *n;
                   3004:        char                    *sp, *ep;
                   3005:
                   3006:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3007:        n = r->man->last;
                   3008:
                   3009:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   3010:                while (*ep != '\0' && *ep != ' ')
                   3011:                        ep++;
                   3012:                while (*ep == ' ')
                   3013:                        *ep++ = '\0';
                   3014:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169     schwarze 3015:        }
                   3016:
                   3017:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3018:        r->man->last = n;
                   3019:        r->man->next = ROFF_NEXT_SIBLING;
1.149     schwarze 3020:        return ROFF_IGN;
1.123     schwarze 3021: }
                   3022:
                   3023: static enum rofferr
1.183     schwarze 3024: roff_als(ROFF_ARGS)
                   3025: {
                   3026:        char            *oldn, *newn, *end, *value;
                   3027:        size_t           oldsz, newsz, valsz;
                   3028:
                   3029:        newn = oldn = buf->buf + pos;
                   3030:        if (*newn == '\0')
                   3031:                return ROFF_IGN;
                   3032:
                   3033:        newsz = roff_getname(r, &oldn, ln, pos);
                   3034:        if (newn[newsz] == '\\' || *oldn == '\0')
                   3035:                return ROFF_IGN;
                   3036:
                   3037:        end = oldn;
                   3038:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3039:        if (oldsz == 0)
                   3040:                return ROFF_IGN;
                   3041:
1.187     schwarze 3042:        valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
                   3043:            (int)oldsz, oldn);
1.183     schwarze 3044:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3045:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3046:        free(value);
                   3047:        return ROFF_IGN;
                   3048: }
                   3049:
                   3050: static enum rofferr
1.168     schwarze 3051: roff_br(ROFF_ARGS)
1.123     schwarze 3052: {
1.174     schwarze 3053:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3054:                man_breakscope(r->man, ROFF_br);
1.168     schwarze 3055:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
                   3056:        if (buf->buf[pos] != '\0')
                   3057:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   3058:                    "%s %s", roff_name[tok], buf->buf + pos);
                   3059:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3060:        r->man->next = ROFF_NEXT_SIBLING;
                   3061:        return ROFF_IGN;
1.27      schwarze 3062: }
                   3063:
                   3064: static enum rofferr
1.48      schwarze 3065: roff_cc(ROFF_ARGS)
                   3066: {
                   3067:        const char      *p;
                   3068:
1.110     schwarze 3069:        p = buf->buf + pos;
1.48      schwarze 3070:
1.110     schwarze 3071:        if (*p == '\0' || (r->control = *p++) == '.')
1.175     schwarze 3072:                r->control = '\0';
1.48      schwarze 3073:
1.110     schwarze 3074:        if (*p != '\0')
1.132     schwarze 3075:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3076:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 3077:
1.149     schwarze 3078:        return ROFF_IGN;
1.48      schwarze 3079: }
                   3080:
                   3081: static enum rofferr
1.175     schwarze 3082: roff_ec(ROFF_ARGS)
                   3083: {
                   3084:        const char      *p;
                   3085:
                   3086:        p = buf->buf + pos;
                   3087:        if (*p == '\0')
                   3088:                r->escape = '\\';
                   3089:        else {
                   3090:                r->escape = *p;
                   3091:                if (*++p != '\0')
                   3092:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3093:                            ln, p - buf->buf, "ec ... %s", p);
                   3094:        }
                   3095:        return ROFF_IGN;
                   3096: }
                   3097:
                   3098: static enum rofferr
                   3099: roff_eo(ROFF_ARGS)
                   3100: {
                   3101:        r->escape = '\0';
                   3102:        if (buf->buf[pos] != '\0')
                   3103:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
                   3104:                    ln, pos, "eo %s", buf->buf + pos);
                   3105:        return ROFF_IGN;
                   3106: }
                   3107:
                   3108: static enum rofferr
1.42      schwarze 3109: roff_tr(ROFF_ARGS)
                   3110: {
                   3111:        const char      *p, *first, *second;
                   3112:        size_t           fsz, ssz;
                   3113:        enum mandoc_esc  esc;
                   3114:
1.110     schwarze 3115:        p = buf->buf + pos;
1.42      schwarze 3116:
1.110     schwarze 3117:        if (*p == '\0') {
1.132     schwarze 3118:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 3119:                return ROFF_IGN;
1.42      schwarze 3120:        }
                   3121:
1.110     schwarze 3122:        while (*p != '\0') {
1.42      schwarze 3123:                fsz = ssz = 1;
                   3124:
                   3125:                first = p++;
1.110     schwarze 3126:                if (*first == '\\') {
1.42      schwarze 3127:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3128:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3129:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3130:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 3131:                                return ROFF_IGN;
1.42      schwarze 3132:                        }
                   3133:                        fsz = (size_t)(p - first);
                   3134:                }
                   3135:
                   3136:                second = p++;
1.110     schwarze 3137:                if (*second == '\\') {
1.42      schwarze 3138:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3139:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3140:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3141:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 3142:                                return ROFF_IGN;
1.42      schwarze 3143:                        }
                   3144:                        ssz = (size_t)(p - second);
1.110     schwarze 3145:                } else if (*second == '\0') {
1.132     schwarze 3146:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   3147:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 3148:                        second = " ";
                   3149:                        p--;
                   3150:                }
                   3151:
                   3152:                if (fsz > 1) {
1.80      schwarze 3153:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3154:                            second, ssz, 0);
1.42      schwarze 3155:                        continue;
                   3156:                }
                   3157:
1.110     schwarze 3158:                if (r->xtab == NULL)
1.80      schwarze 3159:                        r->xtab = mandoc_calloc(128,
                   3160:                            sizeof(struct roffstr));
1.42      schwarze 3161:
                   3162:                free(r->xtab[(int)*first].p);
                   3163:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3164:                r->xtab[(int)*first].sz = ssz;
                   3165:        }
                   3166:
1.149     schwarze 3167:        return ROFF_IGN;
1.42      schwarze 3168: }
                   3169:
                   3170: static enum rofferr
1.178     schwarze 3171: roff_rn(ROFF_ARGS)
                   3172: {
                   3173:        const char      *value;
                   3174:        char            *oldn, *newn, *end;
                   3175:        size_t           oldsz, newsz;
1.187     schwarze 3176:        int              deftype;
1.178     schwarze 3177:
                   3178:        oldn = newn = buf->buf + pos;
                   3179:        if (*oldn == '\0')
                   3180:                return ROFF_IGN;
                   3181:
                   3182:        oldsz = roff_getname(r, &newn, ln, pos);
                   3183:        if (oldn[oldsz] == '\\' || *newn == '\0')
                   3184:                return ROFF_IGN;
                   3185:
                   3186:        end = newn;
                   3187:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3188:        if (newsz == 0)
                   3189:                return ROFF_IGN;
                   3190:
1.187     schwarze 3191:        deftype = ROFFDEF_ANY;
                   3192:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3193:        switch (deftype) {
                   3194:        case ROFFDEF_USER:
1.178     schwarze 3195:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3196:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3197:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.187     schwarze 3198:                break;
                   3199:        case ROFFDEF_PRE:
                   3200:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3201:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3202:                break;
                   3203:        case ROFFDEF_REN:
1.178     schwarze 3204:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3205:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.187     schwarze 3206:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3207:                break;
                   3208:        case ROFFDEF_STD:
1.178     schwarze 3209:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.187     schwarze 3210:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3211:                break;
                   3212:        default:
                   3213:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3214:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3215:                break;
                   3216:        }
1.178     schwarze 3217:        return ROFF_IGN;
                   3218: }
                   3219:
                   3220: static enum rofferr
1.14      schwarze 3221: roff_so(ROFF_ARGS)
                   3222: {
1.121     schwarze 3223:        char *name, *cp;
1.15      schwarze 3224:
1.110     schwarze 3225:        name = buf->buf + pos;
1.97      schwarze 3226:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 3227:
1.22      schwarze 3228:        /*
                   3229:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3230:         * opening anything that's not in our cwd or anything beneath
                   3231:         * it.  Thus, explicitly disallow traversing up the file-system
                   3232:         * or using absolute paths.
                   3233:         */
                   3234:
1.110     schwarze 3235:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 3236:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   3237:                    ".so %s", name);
1.121     schwarze 3238:                buf->sz = mandoc_asprintf(&cp,
                   3239:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3240:                free(buf->buf);
                   3241:                buf->buf = cp;
                   3242:                *offs = 0;
1.149     schwarze 3243:                return ROFF_REPARSE;
1.14      schwarze 3244:        }
                   3245:
                   3246:        *offs = pos;
1.149     schwarze 3247:        return ROFF_SO;
1.7       schwarze 3248: }
                   3249:
1.138     schwarze 3250: /* --- user defined strings and macros ------------------------------------ */
                   3251:
1.16      schwarze 3252: static enum rofferr
                   3253: roff_userdef(ROFF_ARGS)
1.12      schwarze 3254: {
1.183     schwarze 3255:        const char       *arg[16], *ap;
1.16      schwarze 3256:        char             *cp, *n1, *n2;
1.164     schwarze 3257:        int               expand_count, i, ib, ie;
1.135     schwarze 3258:        size_t            asz, rsz;
1.12      schwarze 3259:
1.16      schwarze 3260:        /*
                   3261:         * Collect pointers to macro argument strings
1.61      schwarze 3262:         * and NUL-terminate them.
1.16      schwarze 3263:         */
1.135     schwarze 3264:
1.145     schwarze 3265:        r->argc = 0;
1.110     schwarze 3266:        cp = buf->buf + pos;
1.183     schwarze 3267:        for (i = 0; i < 16; i++) {
1.145     schwarze 3268:                if (*cp == '\0')
                   3269:                        arg[i] = "";
                   3270:                else {
                   3271:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   3272:                        r->argc = i + 1;
                   3273:                }
                   3274:        }
1.16      schwarze 3275:
                   3276:        /*
                   3277:         * Expand macro arguments.
1.12      schwarze 3278:         */
1.135     schwarze 3279:
                   3280:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 3281:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 3282:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 3283:        expand_count = 0;
1.135     schwarze 3284:        while (*cp != '\0') {
                   3285:
                   3286:                /* Scan ahead for the next argument invocation. */
                   3287:
                   3288:                if (*cp++ != '\\')
                   3289:                        continue;
                   3290:                if (*cp++ != '$')
                   3291:                        continue;
1.146     schwarze 3292:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   3293:                        ib = 0;
                   3294:                        ie = r->argc - 1;
                   3295:                } else {  /* \\$1 .. \\$9 insert one argument */
                   3296:                        ib = ie = *cp - '1';
                   3297:                        if (ib < 0 || ib > 8)
                   3298:                                continue;
                   3299:                }
1.135     schwarze 3300:                cp -= 2;
1.164     schwarze 3301:
                   3302:                /*
                   3303:                 * Prevent infinite recursion.
                   3304:                 */
                   3305:
                   3306:                if (cp >= n2)
                   3307:                        expand_count = 1;
                   3308:                else if (++expand_count > EXPAND_LIMIT) {
                   3309:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   3310:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 3311:                        free(buf->buf);
                   3312:                        buf->buf = n1;
1.164     schwarze 3313:                        return ROFF_IGN;
                   3314:                }
1.135     schwarze 3315:
                   3316:                /*
                   3317:                 * Determine the size of the expanded argument,
                   3318:                 * taking escaping of quotes into account.
                   3319:                 */
                   3320:
1.146     schwarze 3321:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   3322:                for (i = ib; i <= ie; i++) {
                   3323:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3324:                                asz++;
                   3325:                                if (*ap == '"')
                   3326:                                        asz += 3;
                   3327:                        }
1.135     schwarze 3328:                }
                   3329:                if (asz != 3) {
                   3330:
                   3331:                        /*
                   3332:                         * Determine the size of the rest of the
                   3333:                         * unexpanded macro, including the NUL.
                   3334:                         */
                   3335:
                   3336:                        rsz = buf->sz - (cp - n1) - 3;
                   3337:
                   3338:                        /*
                   3339:                         * When shrinking, move before
                   3340:                         * releasing the storage.
                   3341:                         */
                   3342:
                   3343:                        if (asz < 3)
                   3344:                                memmove(cp + asz, cp + 3, rsz);
                   3345:
                   3346:                        /*
                   3347:                         * Resize the storage for the macro
                   3348:                         * and readjust the parse pointer.
                   3349:                         */
                   3350:
                   3351:                        buf->sz += asz - 3;
                   3352:                        n2 = mandoc_realloc(n1, buf->sz);
                   3353:                        cp = n2 + (cp - n1);
                   3354:                        n1 = n2;
                   3355:
                   3356:                        /*
                   3357:                         * When growing, make room
                   3358:                         * for the expanded argument.
                   3359:                         */
                   3360:
                   3361:                        if (asz > 3)
                   3362:                                memmove(cp + asz, cp + 3, rsz);
                   3363:                }
                   3364:
                   3365:                /* Copy the expanded argument, escaping quotes. */
                   3366:
                   3367:                n2 = cp;
1.146     schwarze 3368:                for (i = ib; i <= ie; i++) {
                   3369:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3370:                                if (*ap == '"') {
                   3371:                                        memcpy(n2, "\\(dq", 4);
                   3372:                                        n2 += 4;
                   3373:                                } else
                   3374:                                        *n2++ = *ap;
                   3375:                        }
                   3376:                        if (i < ie)
                   3377:                                *n2++ = ' ';
1.16      schwarze 3378:                }
1.12      schwarze 3379:        }
                   3380:
1.16      schwarze 3381:        /*
                   3382:         * Replace the macro invocation
                   3383:         * by the expanded macro.
                   3384:         */
1.135     schwarze 3385:
1.110     schwarze 3386:        free(buf->buf);
                   3387:        buf->buf = n1;
1.120     schwarze 3388:        *offs = 0;
1.16      schwarze 3389:
1.149     schwarze 3390:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3391:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3392: }
1.28      schwarze 3393:
1.178     schwarze 3394: /*
                   3395:  * Calling a high-level macro that was renamed with .rn.
                   3396:  * r->current_string has already been set up by roff_parse().
                   3397:  */
                   3398: static enum rofferr
                   3399: roff_renamed(ROFF_ARGS)
                   3400: {
                   3401:        char    *nbuf;
                   3402:
1.187     schwarze 3403:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3404:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.178     schwarze 3405:        free(buf->buf);
                   3406:        buf->buf = nbuf;
                   3407:        return ROFF_CONT;
                   3408: }
                   3409:
1.85      schwarze 3410: static size_t
1.28      schwarze 3411: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3412: {
                   3413:        char     *name, *cp;
1.85      schwarze 3414:        size_t    namesz;
1.28      schwarze 3415:
                   3416:        name = *cpp;
                   3417:        if ('\0' == *name)
1.149     schwarze 3418:                return 0;
1.28      schwarze 3419:
1.85      schwarze 3420:        /* Read until end of name and terminate it with NUL. */
                   3421:        for (cp = name; 1; cp++) {
                   3422:                if ('\0' == *cp || ' ' == *cp) {
                   3423:                        namesz = cp - name;
                   3424:                        break;
                   3425:                }
1.28      schwarze 3426:                if ('\\' != *cp)
                   3427:                        continue;
1.88      schwarze 3428:                namesz = cp - name;
                   3429:                if ('{' == cp[1] || '}' == cp[1])
                   3430:                        break;
1.28      schwarze 3431:                cp++;
                   3432:                if ('\\' == *cp)
                   3433:                        continue;
1.97      schwarze 3434:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3435:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3436:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3437:                break;
1.28      schwarze 3438:        }
                   3439:
                   3440:        /* Read past spaces. */
                   3441:        while (' ' == *cp)
                   3442:                cp++;
                   3443:
                   3444:        *cpp = cp;
1.149     schwarze 3445:        return namesz;
1.28      schwarze 3446: }
                   3447:
1.16      schwarze 3448: /*
                   3449:  * Store *string into the user-defined string called *name.
                   3450:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3451:  * append == 0: replace mode
                   3452:  * append == 1: single-line append mode
                   3453:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3454:  */
1.8       schwarze 3455: static void
1.16      schwarze 3456: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3457:        int append)
1.7       schwarze 3458: {
1.187     schwarze 3459:        size_t   namesz;
1.42      schwarze 3460:
1.187     schwarze 3461:        namesz = strlen(name);
                   3462:        roff_setstrn(&r->strtab, name, namesz, string,
1.80      schwarze 3463:            string ? strlen(string) : 0, append);
1.187     schwarze 3464:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.42      schwarze 3465: }
                   3466:
                   3467: static void
                   3468: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3469:                const char *string, size_t stringsz, int append)
1.42      schwarze 3470: {
                   3471:        struct roffkv   *n;
                   3472:        char            *c;
                   3473:        int              i;
                   3474:        size_t           oldch, newch;
1.7       schwarze 3475:
1.16      schwarze 3476:        /* Search for an existing string with the same name. */
1.42      schwarze 3477:        n = *r;
                   3478:
1.84      schwarze 3479:        while (n && (namesz != n->key.sz ||
                   3480:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3481:                n = n->next;
1.8       schwarze 3482:
                   3483:        if (NULL == n) {
1.16      schwarze 3484:                /* Create a new string table entry. */
1.42      schwarze 3485:                n = mandoc_malloc(sizeof(struct roffkv));
                   3486:                n->key.p = mandoc_strndup(name, namesz);
                   3487:                n->key.sz = namesz;
                   3488:                n->val.p = NULL;
                   3489:                n->val.sz = 0;
                   3490:                n->next = *r;
                   3491:                *r = n;
1.66      schwarze 3492:        } else if (0 == append) {
1.42      schwarze 3493:                free(n->val.p);
                   3494:                n->val.p = NULL;
                   3495:                n->val.sz = 0;
1.16      schwarze 3496:        }
                   3497:
                   3498:        if (NULL == string)
                   3499:                return;
                   3500:
                   3501:        /*
                   3502:         * One additional byte for the '\n' in multiline mode,
                   3503:         * and one for the terminating '\0'.
                   3504:         */
1.66      schwarze 3505:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3506:
                   3507:        if (NULL == n->val.p) {
                   3508:                n->val.p = mandoc_malloc(newch);
                   3509:                *n->val.p = '\0';
1.16      schwarze 3510:                oldch = 0;
                   3511:        } else {
1.42      schwarze 3512:                oldch = n->val.sz;
                   3513:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3514:        }
                   3515:
                   3516:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3517:        c = n->val.p + (int)oldch;
1.16      schwarze 3518:
                   3519:        /* Append new content to the destination buffer. */
1.42      schwarze 3520:        i = 0;
                   3521:        while (i < (int)stringsz) {
1.16      schwarze 3522:                /*
                   3523:                 * Rudimentary roff copy mode:
                   3524:                 * Handle escaped backslashes.
                   3525:                 */
1.42      schwarze 3526:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3527:                        i++;
                   3528:                *c++ = string[i++];
1.16      schwarze 3529:        }
1.8       schwarze 3530:
1.16      schwarze 3531:        /* Append terminating bytes. */
1.66      schwarze 3532:        if (1 < append)
1.16      schwarze 3533:                *c++ = '\n';
1.42      schwarze 3534:
1.16      schwarze 3535:        *c = '\0';
1.42      schwarze 3536:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3537: }
                   3538:
1.8       schwarze 3539: static const char *
1.197     schwarze 3540: roff_getstrn(struct roff *r, const char *name, size_t len,
1.187     schwarze 3541:     int *deftype)
1.7       schwarze 3542: {
1.187     schwarze 3543:        const struct roffkv     *n;
1.197     schwarze 3544:        int                      found, i;
1.187     schwarze 3545:        enum roff_tok            tok;
                   3546:
1.197     schwarze 3547:        found = 0;
                   3548:        for (n = r->strtab; n != NULL; n = n->next) {
                   3549:                if (strncmp(name, n->key.p, len) != 0 ||
                   3550:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   3551:                        continue;
                   3552:                if (*deftype & ROFFDEF_USER) {
                   3553:                        *deftype = ROFFDEF_USER;
                   3554:                        return n->val.p;
                   3555:                } else {
                   3556:                        found = 1;
                   3557:                        break;
                   3558:                }
                   3559:        }
                   3560:        for (n = r->rentab; n != NULL; n = n->next) {
                   3561:                if (strncmp(name, n->key.p, len) != 0 ||
                   3562:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   3563:                        continue;
                   3564:                if (*deftype & ROFFDEF_REN) {
                   3565:                        *deftype = ROFFDEF_REN;
                   3566:                        return n->val.p;
                   3567:                } else {
                   3568:                        found = 1;
                   3569:                        break;
1.187     schwarze 3570:                }
                   3571:        }
1.197     schwarze 3572:        for (i = 0; i < PREDEFS_MAX; i++) {
                   3573:                if (strncmp(name, predefs[i].name, len) != 0 ||
                   3574:                    predefs[i].name[len] != '\0')
                   3575:                        continue;
                   3576:                if (*deftype & ROFFDEF_PRE) {
                   3577:                        *deftype = ROFFDEF_PRE;
                   3578:                        return predefs[i].str;
                   3579:                } else {
                   3580:                        found = 1;
                   3581:                        break;
1.187     schwarze 3582:                }
                   3583:        }
1.197     schwarze 3584:        if (r->man->macroset != MACROSET_MAN) {
                   3585:                for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
                   3586:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   3587:                            roff_name[tok][len] != '\0')
                   3588:                                continue;
                   3589:                        if (*deftype & ROFFDEF_STD) {
                   3590:                                *deftype = ROFFDEF_STD;
                   3591:                                return NULL;
                   3592:                        } else {
                   3593:                                found = 1;
                   3594:                                break;
1.187     schwarze 3595:                        }
                   3596:                }
                   3597:        }
1.197     schwarze 3598:        if (r->man->macroset != MACROSET_MDOC) {
                   3599:                for (tok = MAN_TH; tok < MAN_MAX; tok++) {
                   3600:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   3601:                            roff_name[tok][len] != '\0')
                   3602:                                continue;
                   3603:                        if (*deftype & ROFFDEF_STD) {
                   3604:                                *deftype = ROFFDEF_STD;
                   3605:                                return NULL;
                   3606:                        } else {
                   3607:                                found = 1;
                   3608:                                break;
1.187     schwarze 3609:                        }
                   3610:                }
1.197     schwarze 3611:        }
                   3612:
                   3613:        if (found == 0 && *deftype != ROFFDEF_ANY) {
                   3614:                if (*deftype & ROFFDEF_REN) {
                   3615:                        /*
                   3616:                         * This might still be a request,
                   3617:                         * so do not treat it as undefined yet.
                   3618:                         */
                   3619:                        *deftype = ROFFDEF_UNDEF;
                   3620:                        return NULL;
1.187     schwarze 3621:                }
1.197     schwarze 3622:
                   3623:                /* Using an undefined string defines it to be empty. */
                   3624:
                   3625:                roff_setstrn(&r->strtab, name, len, "", 0, 0);
                   3626:                roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.187     schwarze 3627:        }
1.197     schwarze 3628:
1.187     schwarze 3629:        *deftype = 0;
1.149     schwarze 3630:        return NULL;
1.7       schwarze 3631: }
                   3632:
1.8       schwarze 3633: static void
1.42      schwarze 3634: roff_freestr(struct roffkv *r)
1.7       schwarze 3635: {
1.42      schwarze 3636:        struct roffkv    *n, *nn;
1.7       schwarze 3637:
1.42      schwarze 3638:        for (n = r; n; n = nn) {
                   3639:                free(n->key.p);
                   3640:                free(n->val.p);
1.7       schwarze 3641:                nn = n->next;
                   3642:                free(n);
                   3643:        }
1.27      schwarze 3644: }
1.138     schwarze 3645:
                   3646: /* --- accessors and utility functions ------------------------------------ */
1.42      schwarze 3647:
                   3648: /*
                   3649:  * Duplicate an input string, making the appropriate character
                   3650:  * conversations (as stipulated by `tr') along the way.
                   3651:  * Returns a heap-allocated string with all the replacements made.
                   3652:  */
                   3653: char *
                   3654: roff_strdup(const struct roff *r, const char *p)
                   3655: {
                   3656:        const struct roffkv *cp;
                   3657:        char            *res;
                   3658:        const char      *pp;
                   3659:        size_t           ssz, sz;
                   3660:        enum mandoc_esc  esc;
                   3661:
                   3662:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3663:                return mandoc_strdup(p);
1.42      schwarze 3664:        else if ('\0' == *p)
1.149     schwarze 3665:                return mandoc_strdup("");
1.42      schwarze 3666:
                   3667:        /*
                   3668:         * Step through each character looking for term matches
                   3669:         * (remember that a `tr' can be invoked with an escape, which is
                   3670:         * a glyph but the escape is multi-character).
                   3671:         * We only do this if the character hash has been initialised
                   3672:         * and the string is >0 length.
                   3673:         */
                   3674:
                   3675:        res = NULL;
                   3676:        ssz = 0;
                   3677:
                   3678:        while ('\0' != *p) {
1.161     schwarze 3679:                assert((unsigned int)*p < 128);
                   3680:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3681:                        sz = r->xtab[(int)*p].sz;
                   3682:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3683:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3684:                        ssz += sz;
                   3685:                        p++;
                   3686:                        continue;
                   3687:                } else if ('\\' != *p) {
                   3688:                        res = mandoc_realloc(res, ssz + 2);
                   3689:                        res[ssz++] = *p++;
                   3690:                        continue;
                   3691:                }
                   3692:
                   3693:                /* Search for term matches. */
                   3694:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3695:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3696:                                break;
                   3697:
                   3698:                if (NULL != cp) {
                   3699:                        /*
                   3700:                         * A match has been found.
                   3701:                         * Append the match to the array and move
                   3702:                         * forward by its keysize.
                   3703:                         */
1.80      schwarze 3704:                        res = mandoc_realloc(res,
                   3705:                            ssz + cp->val.sz + 1);
1.42      schwarze 3706:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3707:                        ssz += cp->val.sz;
                   3708:                        p += (int)cp->key.sz;
                   3709:                        continue;
                   3710:                }
                   3711:
                   3712:                /*
                   3713:                 * Handle escapes carefully: we need to copy
                   3714:                 * over just the escape itself, or else we might
                   3715:                 * do replacements within the escape itself.
                   3716:                 * Make sure to pass along the bogus string.
                   3717:                 */
                   3718:                pp = p++;
                   3719:                esc = mandoc_escape(&p, NULL, NULL);
                   3720:                if (ESCAPE_ERROR == esc) {
                   3721:                        sz = strlen(pp);
                   3722:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3723:                        memcpy(res + ssz, pp, sz);
                   3724:                        break;
                   3725:                }
1.80      schwarze 3726:                /*
                   3727:                 * We bail out on bad escapes.
1.42      schwarze 3728:                 * No need to warn: we already did so when
                   3729:                 * roff_res() was called.
                   3730:                 */
                   3731:                sz = (int)(p - pp);
                   3732:                res = mandoc_realloc(res, ssz + sz + 1);
                   3733:                memcpy(res + ssz, pp, sz);
                   3734:                ssz += sz;
                   3735:        }
                   3736:
                   3737:        res[(int)ssz] = '\0';
1.149     schwarze 3738:        return res;
1.99      schwarze 3739: }
                   3740:
                   3741: int
                   3742: roff_getformat(const struct roff *r)
                   3743: {
                   3744:
1.149     schwarze 3745:        return r->format;
1.48      schwarze 3746: }
                   3747:
                   3748: /*
1.80      schwarze 3749:  * Find out whether a line is a macro line or not.
1.48      schwarze 3750:  * If it is, adjust the current position and return one; if it isn't,
                   3751:  * return zero and don't change the current position.
                   3752:  * If the control character has been set with `.cc', then let that grain
                   3753:  * precedence.
                   3754:  * This is slighly contrary to groff, where using the non-breaking
                   3755:  * control character when `cc' has been invoked will cause the
                   3756:  * non-breaking macro contents to be printed verbatim.
                   3757:  */
                   3758: int
                   3759: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3760: {
                   3761:        int             pos;
                   3762:
                   3763:        pos = *ppos;
                   3764:
1.175     schwarze 3765:        if (r->control != '\0' && cp[pos] == r->control)
1.48      schwarze 3766:                pos++;
1.175     schwarze 3767:        else if (r->control != '\0')
1.149     schwarze 3768:                return 0;
1.48      schwarze 3769:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3770:                pos += 2;
                   3771:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3772:                pos++;
                   3773:        else
1.149     schwarze 3774:                return 0;
1.48      schwarze 3775:
                   3776:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3777:                pos++;
                   3778:
                   3779:        *ppos = pos;
1.149     schwarze 3780:        return 1;
1.1       schwarze 3781: }