[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 1.197

1.197   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.196 2017/07/14 17:16:13 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 *);
1.53      schwarze  183: static int              roff_getregn(const struct roff *,
                    184:                                const char *, size_t);
1.145     schwarze  185: static int              roff_getregro(const struct roff *,
                    186:                                const char *name);
1.197   ! schwarze  187: static const char      *roff_getstrn(struct roff *,
1.187     schwarze  188:                                const char *, size_t, int *);
1.143     schwarze  189: static int              roff_hasregn(const struct roff *,
                    190:                                const char *, size_t);
1.123     schwarze  191: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  192: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  193: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137     schwarze  194: static void             roff_man_alloc1(struct roff_man *);
                    195: static void             roff_man_free1(struct roff_man *);
1.172     schwarze  196: static enum rofferr     roff_manyarg(ROFF_ARGS);
1.6       schwarze  197: static enum rofferr     roff_nr(ROFF_ARGS);
1.169     schwarze  198: static enum rofferr     roff_onearg(ROFF_ARGS);
1.166     schwarze  199: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.87      schwarze  200:                                int, int);
1.177     schwarze  201: static enum rofferr     roff_parsetext(struct roff *, struct buf *,
                    202:                                int, int *);
1.178     schwarze  203: static enum rofferr     roff_renamed(ROFF_ARGS);
1.110     schwarze  204: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  205: static enum rofferr     roff_rm(ROFF_ARGS);
1.178     schwarze  206: static enum rofferr     roff_rn(ROFF_ARGS);
1.76      schwarze  207: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  208: static void             roff_setstr(struct roff *,
1.16      schwarze  209:                                const char *, const char *, int);
1.80      schwarze  210: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  211:                                size_t, const char *, size_t, int);
1.14      schwarze  212: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  213: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  214: static enum rofferr     roff_Dd(ROFF_ARGS);
1.27      schwarze  215: static enum rofferr     roff_TE(ROFF_ARGS);
                    216: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  217: static enum rofferr     roff_EQ(ROFF_ARGS);
                    218: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  219: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  220: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  221: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  222:
1.138     schwarze  223: /* --- constant data ------------------------------------------------------ */
                    224:
1.133     schwarze  225: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    226: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    227:
1.166     schwarze  228: const char *__roff_name[MAN_MAX + 1] = {
1.177     schwarze  229:        "br",           "ce",           "ft",           "ll",
1.184     schwarze  230:        "mc",           "po",           "rj",           "sp",
                    231:        "ta",           "ti",           NULL,
1.166     schwarze  232:        "ab",           "ad",           "af",           "aln",
                    233:        "als",          "am",           "am1",          "ami",
                    234:        "ami1",         "as",           "as1",          "asciify",
                    235:        "backtrace",    "bd",           "bleedat",      "blm",
                    236:         "box",         "boxa",         "bp",           "BP",
                    237:        "break",        "breakchar",    "brnl",         "brp",
1.177     schwarze  238:        "brpnl",        "c2",           "cc",
1.166     schwarze  239:        "cf",           "cflags",       "ch",           "char",
                    240:        "chop",         "class",        "close",        "CL",
                    241:        "color",        "composite",    "continue",     "cp",
                    242:        "cropat",       "cs",           "cu",           "da",
                    243:        "dch",          "Dd",           "de",           "de1",
                    244:        "defcolor",     "dei",          "dei1",         "device",
                    245:        "devicem",      "di",           "do",           "ds",
                    246:        "ds1",          "dwh",          "dt",           "ec",
                    247:        "ecr",          "ecs",          "el",           "em",
                    248:        "EN",           "eo",           "EP",           "EQ",
                    249:        "errprint",     "ev",           "evc",          "ex",
                    250:        "fallback",     "fam",          "fc",           "fchar",
                    251:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    252:        "fl",           "flig",         "fp",           "fps",
                    253:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    254:        "fzoom",        "gcolor",       "hc",           "hcode",
                    255:        "hidechar",     "hla",          "hlm",          "hpf",
                    256:        "hpfa",         "hpfcode",      "hw",           "hy",
                    257:        "hylang",       "hylen",        "hym",          "hypp",
                    258:        "hys",          "ie",           "if",           "ig",
                    259:        "index",        "it",           "itc",          "IX",
                    260:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    261:        "lc",           "lc_ctype",     "lds",          "length",
                    262:        "letadj",       "lf",           "lg",           "lhang",
                    263:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.176     schwarze  264:        "ls",           "lsm",          "lt",
1.166     schwarze  265:        "mediasize",    "minss",        "mk",           "mso",
                    266:        "na",           "ne",           "nh",           "nhychar",
                    267:        "nm",           "nn",           "nop",          "nr",
                    268:        "nrf",          "nroff",        "ns",           "nx",
                    269:        "open",         "opena",        "os",           "output",
                    270:        "padj",         "papersize",    "pc",           "pev",
                    271:        "pi",           "PI",           "pl",           "pm",
1.184     schwarze  272:        "pn",           "pnr",          "ps",
1.166     schwarze  273:        "psbb",         "pshape",       "pso",          "ptr",
                    274:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.181     schwarze  275:        "return",       "rfschar",      "rhang",
1.166     schwarze  276:        "rm",           "rn",           "rnn",          "rr",
                    277:        "rs",           "rt",           "schar",        "sentchar",
                    278:        "shc",          "shift",        "sizes",        "so",
                    279:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    280:        "sty",          "substring",    "sv",           "sy",
1.172     schwarze  281:        "T&",           "tc",           "TE",
1.173     schwarze  282:        "TH",           "tkf",          "tl",
1.166     schwarze  283:        "tm",           "tm1",          "tmc",          "tr",
                    284:        "track",        "transchar",    "trf",          "trimat",
                    285:        "trin",         "trnt",         "troff",        "TS",
                    286:        "uf",           "ul",           "unformat",     "unwatch",
                    287:        "unwatchn",     "vpt",          "vs",           "warn",
                    288:        "warnscale",    "watch",        "watchlength",  "watchn",
                    289:        "wh",           "while",        "write",        "writec",
                    290:        "writem",       "xflag",        ".",            NULL,
1.178     schwarze  291:        NULL,           "text",
1.166     schwarze  292:        "Dd",           "Dt",           "Os",           "Sh",
                    293:        "Ss",           "Pp",           "D1",           "Dl",
                    294:        "Bd",           "Ed",           "Bl",           "El",
                    295:        "It",           "Ad",           "An",           "Ap",
                    296:        "Ar",           "Cd",           "Cm",           "Dv",
                    297:        "Er",           "Ev",           "Ex",           "Fa",
                    298:        "Fd",           "Fl",           "Fn",           "Ft",
                    299:        "Ic",           "In",           "Li",           "Nd",
                    300:        "Nm",           "Op",           "Ot",           "Pa",
                    301:        "Rv",           "St",           "Va",           "Vt",
                    302:        "Xr",           "%A",           "%B",           "%D",
                    303:        "%I",           "%J",           "%N",           "%O",
                    304:        "%P",           "%R",           "%T",           "%V",
                    305:        "Ac",           "Ao",           "Aq",           "At",
                    306:        "Bc",           "Bf",           "Bo",           "Bq",
                    307:        "Bsx",          "Bx",           "Db",           "Dc",
                    308:        "Do",           "Dq",           "Ec",           "Ef",
                    309:        "Em",           "Eo",           "Fx",           "Ms",
                    310:        "No",           "Ns",           "Nx",           "Ox",
                    311:        "Pc",           "Pf",           "Po",           "Pq",
                    312:        "Qc",           "Ql",           "Qo",           "Qq",
                    313:        "Re",           "Rs",           "Sc",           "So",
                    314:        "Sq",           "Sm",           "Sx",           "Sy",
                    315:        "Tn",           "Ux",           "Xc",           "Xo",
                    316:        "Fo",           "Fc",           "Oo",           "Oc",
                    317:        "Bk",           "Ek",           "Bt",           "Hf",
                    318:        "Fr",           "Ud",           "Lb",           "Lp",
                    319:        "Lk",           "Mt",           "Brq",          "Bro",
                    320:        "Brc",          "%C",           "Es",           "En",
1.171     schwarze  321:        "Dx",           "%Q",           "%U",           "Ta",
                    322:        NULL,
1.166     schwarze  323:        "TH",           "SH",           "SS",           "TP",
                    324:        "LP",           "PP",           "P",            "IP",
                    325:        "HP",           "SM",           "SB",           "BI",
                    326:        "IB",           "BR",           "RB",           "R",
                    327:        "B",            "I",            "IR",           "RI",
1.171     schwarze  328:        "nf",           "fi",
1.166     schwarze  329:        "RE",           "RS",           "DT",           "UC",
1.169     schwarze  330:        "PD",           "AT",           "in",
1.166     schwarze  331:        "OP",           "EX",           "EE",           "UR",
1.189     bentley   332:        "UE",           "MT",           "ME",           NULL
1.166     schwarze  333: };
                    334: const  char *const *roff_name = __roff_name;
                    335:
                    336: static struct roffmac   roffs[TOKEN_NONE] = {
1.168     schwarze  337:        { roff_br, NULL, NULL, 0 },  /* br */
1.177     schwarze  338:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.169     schwarze  339:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.170     schwarze  340:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.176     schwarze  341:        { roff_onearg, NULL, NULL, 0 },  /* mc */
1.184     schwarze  342:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.181     schwarze  343:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.171     schwarze  344:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.172     schwarze  345:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.173     schwarze  346:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.168     schwarze  347:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.167     schwarze  348:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    349:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    350:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    351:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.183     schwarze  352:        { roff_als, NULL, NULL, 0 },  /* als */
1.167     schwarze  353:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    354:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    355:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    356:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    357:        { roff_ds, NULL, NULL, 0 },  /* as */
                    358:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    359:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    360:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    361:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    362:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    363:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    364:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    365:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    366:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    367:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    368:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    369:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    370:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.168     schwarze  371:        { roff_br, NULL, NULL, 0 },  /* brp */
1.167     schwarze  372:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    373:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    374:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    375:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    376:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    377:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    378:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    379:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    380:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    381:        { roff_insec, NULL, NULL, 0 },  /* close */
                    382:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    383:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    384:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    385:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    386:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    387:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    388:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    389:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    390:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    391:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    392:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    393:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    394:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    395:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    396:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    397:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    398:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    400:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    401:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    402:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    403:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    404:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    405:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.175     schwarze  406:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.167     schwarze  407:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    408:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    409:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    410:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    411:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.175     schwarze  412:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.167     schwarze  413:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    414:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    415:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    416:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    417:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    418:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    419:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    421:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    422:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    425:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    429:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    430:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    431:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    432:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    438:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    439:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    440:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    442:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    444:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    445:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    446:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    447:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    448:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    449:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    450:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    451:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    452:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    453:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    454:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    455:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    456:        { roff_it, NULL, NULL, 0 },  /* it */
                    457:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    458:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    461:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    462:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    463:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    464:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    465:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    466:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    467:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    468:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    469:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    470:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    471:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    472:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    473:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    474:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    475:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    476:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    481:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    486:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    487:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    488:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    489:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    490:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    491:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    492:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    493:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    494:        { roff_insec, NULL, NULL, 0 },  /* open */
                    495:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    496:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    497:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    500:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    501:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    502:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    503:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    504:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    505:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    506:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    508:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    509:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    510:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    511:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    512:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    513:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    514:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    516:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    517:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    518:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    519:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    520:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.178     schwarze  521:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.167     schwarze  522:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    523:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    524:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    526:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    529:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    530:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    531:        { roff_so, NULL, NULL, 0 },  /* so */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    533:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    534:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    535:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    536:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    537:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    538:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    539:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    540:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    541:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    542:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.187     schwarze  543:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.167     schwarze  544:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    545:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    549:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    550:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    551:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    552:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    553:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    554:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    555:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    557:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    558:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    559:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    560:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    561:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    562:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    563:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    565:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    566:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    568:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    569:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    570:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    571:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    572:        { roff_insec, NULL, NULL, 0 },  /* write */
                    573:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    574:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    575:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    576:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.178     schwarze  577:        { roff_renamed, NULL, NULL, 0 },
1.167     schwarze  578:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  579: };
                    580:
1.37      schwarze  581: /* Array of injected predefined strings. */
                    582: #define        PREDEFS_MAX      38
                    583: static const struct predef predefs[PREDEFS_MAX] = {
                    584: #include "predefs.in"
                    585: };
                    586:
1.177     schwarze  587: static int      roffce_lines;  /* number of input lines to center */
                    588: static struct roff_node *roffce_node;  /* active request */
1.51      schwarze  589: static int      roffit_lines;  /* number of lines to delay */
                    590: static char    *roffit_macro;  /* nil-terminated macro line */
                    591:
1.80      schwarze  592:
1.138     schwarze  593: /* --- request table ------------------------------------------------------ */
                    594:
1.167     schwarze  595: struct ohash *
                    596: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  597: {
1.167     schwarze  598:        struct ohash    *htab;
                    599:        struct roffreq  *req;
                    600:        enum roff_tok    tok;
                    601:        size_t           sz;
                    602:        unsigned int     slot;
1.3       schwarze  603:
1.167     schwarze  604:        htab = mandoc_malloc(sizeof(*htab));
                    605:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  606:
1.167     schwarze  607:        for (tok = mintok; tok < maxtok; tok++) {
1.168     schwarze  608:                if (roff_name[tok] == NULL)
                    609:                        continue;
1.167     schwarze  610:                sz = strlen(roff_name[tok]);
                    611:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    612:                req->tok = tok;
                    613:                memcpy(req->name, roff_name[tok], sz + 1);
                    614:                slot = ohash_qlookup(htab, req->name);
                    615:                ohash_insert(htab, slot, req);
1.3       schwarze  616:        }
1.167     schwarze  617:        return htab;
1.3       schwarze  618: }
                    619:
1.167     schwarze  620: void
                    621: roffhash_free(struct ohash *htab)
1.1       schwarze  622: {
1.167     schwarze  623:        struct roffreq  *req;
                    624:        unsigned int     slot;
1.1       schwarze  625:
1.167     schwarze  626:        if (htab == NULL)
                    627:                return;
                    628:        for (req = ohash_first(htab, &slot); req != NULL;
                    629:             req = ohash_next(htab, &slot))
                    630:                free(req);
                    631:        ohash_delete(htab);
                    632:        free(htab);
                    633: }
                    634:
                    635: enum roff_tok
                    636: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    637: {
                    638:        struct roffreq  *req;
                    639:        const char      *end;
                    640:
                    641:        if (sz) {
                    642:                end = name + sz;
                    643:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    644:        } else
                    645:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    646:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  647: }
                    648:
1.138     schwarze  649: /* --- stack of request blocks -------------------------------------------- */
                    650:
1.1       schwarze  651: /*
                    652:  * Pop the current node off of the stack of roff instructions currently
                    653:  * pending.
                    654:  */
                    655: static void
                    656: roffnode_pop(struct roff *r)
                    657: {
                    658:        struct roffnode *p;
                    659:
1.2       schwarze  660:        assert(r->last);
1.80      schwarze  661:        p = r->last;
1.2       schwarze  662:
                    663:        r->last = r->last->parent;
1.16      schwarze  664:        free(p->name);
                    665:        free(p->end);
1.1       schwarze  666:        free(p);
                    667: }
                    668:
                    669: /*
                    670:  * Push a roff node onto the instruction stack.  This must later be
                    671:  * removed with roffnode_pop().
                    672:  */
1.11      schwarze  673: static void
1.166     schwarze  674: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  675:                int line, int col)
1.1       schwarze  676: {
                    677:        struct roffnode *p;
                    678:
1.11      schwarze  679:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  680:        p->tok = tok;
1.16      schwarze  681:        if (name)
                    682:                p->name = mandoc_strdup(name);
1.1       schwarze  683:        p->parent = r->last;
                    684:        p->line = line;
                    685:        p->col = col;
1.71      schwarze  686:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  687:
                    688:        r->last = p;
                    689: }
                    690:
1.138     schwarze  691: /* --- roff parser state data management ---------------------------------- */
                    692:
1.1       schwarze  693: static void
                    694: roff_free1(struct roff *r)
                    695: {
1.49      schwarze  696:        struct tbl_node *tbl;
1.42      schwarze  697:        int              i;
1.27      schwarze  698:
1.49      schwarze  699:        while (NULL != (tbl = r->first_tbl)) {
                    700:                r->first_tbl = tbl->next;
                    701:                tbl_free(tbl);
1.27      schwarze  702:        }
                    703:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  704:
1.191     schwarze  705:        if (r->last_eqn != NULL)
                    706:                eqn_free(r->last_eqn);
                    707:        r->last_eqn = r->eqn = NULL;
1.32      schwarze  708:
1.1       schwarze  709:        while (r->last)
                    710:                roffnode_pop(r);
1.27      schwarze  711:
1.96      schwarze  712:        free (r->rstack);
                    713:        r->rstack = NULL;
                    714:        r->rstacksz = 0;
                    715:        r->rstackpos = -1;
                    716:
                    717:        roff_freereg(r->regtab);
                    718:        r->regtab = NULL;
                    719:
1.42      schwarze  720:        roff_freestr(r->strtab);
1.178     schwarze  721:        roff_freestr(r->rentab);
1.42      schwarze  722:        roff_freestr(r->xmbtab);
1.178     schwarze  723:        r->strtab = r->rentab = r->xmbtab = NULL;
1.42      schwarze  724:
                    725:        if (r->xtab)
                    726:                for (i = 0; i < 128; i++)
                    727:                        free(r->xtab[i].p);
                    728:        free(r->xtab);
                    729:        r->xtab = NULL;
1.1       schwarze  730: }
                    731:
                    732: void
                    733: roff_reset(struct roff *r)
                    734: {
                    735:        roff_free1(r);
1.99      schwarze  736:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175     schwarze  737:        r->control = '\0';
                    738:        r->escape = '\\';
1.179     schwarze  739:        roffce_lines = 0;
                    740:        roffce_node = NULL;
                    741:        roffit_lines = 0;
                    742:        roffit_macro = NULL;
1.1       schwarze  743: }
                    744:
                    745: void
                    746: roff_free(struct roff *r)
                    747: {
                    748:        roff_free1(r);
1.167     schwarze  749:        roffhash_free(r->reqtab);
1.1       schwarze  750:        free(r);
                    751: }
                    752:
                    753: struct roff *
1.151     schwarze  754: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  755: {
                    756:        struct roff     *r;
                    757:
1.11      schwarze  758:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  759:        r->parse = parse;
1.167     schwarze  760:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72      schwarze  761:        r->options = options;
1.99      schwarze  762:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  763:        r->rstackpos = -1;
1.175     schwarze  764:        r->escape = '\\';
1.149     schwarze  765:        return r;
1.137     schwarze  766: }
                    767:
1.138     schwarze  768: /* --- syntax tree state data management ---------------------------------- */
                    769:
1.137     schwarze  770: static void
                    771: roff_man_free1(struct roff_man *man)
                    772: {
                    773:
1.138     schwarze  774:        if (man->first != NULL)
                    775:                roff_node_delete(man, man->first);
1.137     schwarze  776:        free(man->meta.msec);
                    777:        free(man->meta.vol);
                    778:        free(man->meta.os);
                    779:        free(man->meta.arch);
                    780:        free(man->meta.title);
                    781:        free(man->meta.name);
                    782:        free(man->meta.date);
                    783: }
                    784:
                    785: static void
                    786: roff_man_alloc1(struct roff_man *man)
                    787: {
                    788:
                    789:        memset(&man->meta, 0, sizeof(man->meta));
                    790:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    791:        man->first->type = ROFFT_ROOT;
                    792:        man->last = man->first;
                    793:        man->last_es = NULL;
                    794:        man->flags = 0;
                    795:        man->macroset = MACROSET_NONE;
                    796:        man->lastsec = man->lastnamed = SEC_NONE;
                    797:        man->next = ROFF_NEXT_CHILD;
                    798: }
                    799:
                    800: void
                    801: roff_man_reset(struct roff_man *man)
                    802: {
                    803:
                    804:        roff_man_free1(man);
                    805:        roff_man_alloc1(man);
                    806: }
                    807:
                    808: void
                    809: roff_man_free(struct roff_man *man)
                    810: {
                    811:
                    812:        roff_man_free1(man);
                    813:        free(man);
                    814: }
                    815:
                    816: struct roff_man *
                    817: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.188     schwarze  818:        const char *os_s, int quick)
1.137     schwarze  819: {
                    820:        struct roff_man *man;
                    821:
                    822:        man = mandoc_calloc(1, sizeof(*man));
                    823:        man->parse = parse;
                    824:        man->roff = roff;
1.188     schwarze  825:        man->os_s = os_s;
1.137     schwarze  826:        man->quick = quick;
                    827:        roff_man_alloc1(man);
1.168     schwarze  828:        roff->man = man;
1.149     schwarze  829:        return man;
1.1       schwarze  830: }
                    831:
1.138     schwarze  832: /* --- syntax tree handling ----------------------------------------------- */
                    833:
                    834: struct roff_node *
                    835: roff_node_alloc(struct roff_man *man, int line, int pos,
                    836:        enum roff_type type, int tok)
                    837: {
                    838:        struct roff_node        *n;
                    839:
                    840:        n = mandoc_calloc(1, sizeof(*n));
                    841:        n->line = line;
                    842:        n->pos = pos;
                    843:        n->tok = tok;
                    844:        n->type = type;
                    845:        n->sec = man->lastsec;
                    846:
                    847:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  848:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  849:        else
1.157     schwarze  850:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  851:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  852:                n->flags |= NODE_LINE;
1.138     schwarze  853:        man->flags &= ~MDOC_NEWLINE;
                    854:
1.149     schwarze  855:        return n;
1.138     schwarze  856: }
                    857:
                    858: void
                    859: roff_node_append(struct roff_man *man, struct roff_node *n)
                    860: {
                    861:
                    862:        switch (man->next) {
                    863:        case ROFF_NEXT_SIBLING:
1.153     schwarze  864:                if (man->last->next != NULL) {
                    865:                        n->next = man->last->next;
                    866:                        man->last->next->prev = n;
                    867:                } else
                    868:                        man->last->parent->last = n;
1.138     schwarze  869:                man->last->next = n;
                    870:                n->prev = man->last;
                    871:                n->parent = man->last->parent;
                    872:                break;
                    873:        case ROFF_NEXT_CHILD:
1.159     schwarze  874:                if (man->last->child != NULL) {
                    875:                        n->next = man->last->child;
                    876:                        man->last->child->prev = n;
                    877:                } else
                    878:                        man->last->last = n;
1.138     schwarze  879:                man->last->child = n;
                    880:                n->parent = man->last;
                    881:                break;
                    882:        default:
                    883:                abort();
                    884:        }
1.154     schwarze  885:        man->last = n;
1.138     schwarze  886:
                    887:        switch (n->type) {
                    888:        case ROFFT_HEAD:
                    889:                n->parent->head = n;
                    890:                break;
                    891:        case ROFFT_BODY:
1.154     schwarze  892:                if (n->end != ENDBODY_NOT)
                    893:                        return;
1.138     schwarze  894:                n->parent->body = n;
                    895:                break;
                    896:        case ROFFT_TAIL:
                    897:                n->parent->tail = n;
                    898:                break;
                    899:        default:
1.154     schwarze  900:                return;
1.138     schwarze  901:        }
1.154     schwarze  902:
                    903:        /*
                    904:         * Copy over the normalised-data pointer of our parent.  Not
                    905:         * everybody has one, but copying a null pointer is fine.
                    906:         */
                    907:
                    908:        n->norm = n->parent->norm;
                    909:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  910: }
                    911:
1.139     schwarze  912: void
                    913: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    914: {
                    915:        struct roff_node        *n;
                    916:
                    917:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    918:        n->string = roff_strdup(man->roff, word);
                    919:        roff_node_append(man, n);
1.158     schwarze  920:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  921:        man->next = ROFF_NEXT_SIBLING;
                    922: }
                    923:
                    924: void
                    925: roff_word_append(struct roff_man *man, const char *word)
                    926: {
                    927:        struct roff_node        *n;
                    928:        char                    *addstr, *newstr;
                    929:
                    930:        n = man->last;
                    931:        addstr = roff_strdup(man->roff, word);
                    932:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    933:        free(addstr);
                    934:        free(n->string);
                    935:        n->string = newstr;
                    936:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  937: }
                    938:
                    939: void
                    940: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    941: {
                    942:        struct roff_node        *n;
                    943:
                    944:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    945:        roff_node_append(man, n);
                    946:        man->next = ROFF_NEXT_CHILD;
                    947: }
                    948:
                    949: struct roff_node *
                    950: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    951: {
                    952:        struct roff_node        *n;
                    953:
                    954:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    955:        roff_node_append(man, n);
                    956:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  957:        return n;
1.139     schwarze  958: }
                    959:
1.138     schwarze  960: struct roff_node *
                    961: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    962: {
                    963:        struct roff_node        *n;
                    964:
                    965:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    966:        roff_node_append(man, n);
                    967:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  968:        return n;
1.138     schwarze  969: }
                    970:
                    971: struct roff_node *
                    972: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    973: {
                    974:        struct roff_node        *n;
                    975:
                    976:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    977:        roff_node_append(man, n);
                    978:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  979:        return n;
1.139     schwarze  980: }
                    981:
1.193     schwarze  982: static void
                    983: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.139     schwarze  984: {
                    985:        struct roff_node        *n;
1.193     schwarze  986:        const struct tbl_span   *span;
1.139     schwarze  987:
                    988:        if (man->macroset == MACROSET_MAN)
1.174     schwarze  989:                man_breakscope(man, ROFF_TS);
1.193     schwarze  990:        while ((span = tbl_span(tbl)) != NULL) {
                    991:                n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                    992:                n->span = span;
                    993:                roff_node_append(man, n);
                    994:                n->flags |= NODE_VALID | NODE_ENDED;
                    995:                man->next = ROFF_NEXT_SIBLING;
                    996:        }
1.138     schwarze  997: }
                    998:
                    999: void
                   1000: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1001: {
                   1002:
                   1003:        /* Adjust siblings. */
                   1004:
                   1005:        if (n->prev)
                   1006:                n->prev->next = n->next;
                   1007:        if (n->next)
                   1008:                n->next->prev = n->prev;
                   1009:
                   1010:        /* Adjust parent. */
                   1011:
                   1012:        if (n->parent != NULL) {
                   1013:                if (n->parent->child == n)
                   1014:                        n->parent->child = n->next;
                   1015:                if (n->parent->last == n)
                   1016:                        n->parent->last = n->prev;
                   1017:        }
                   1018:
                   1019:        /* Adjust parse point. */
                   1020:
                   1021:        if (man == NULL)
                   1022:                return;
                   1023:        if (man->last == n) {
                   1024:                if (n->prev == NULL) {
                   1025:                        man->last = n->parent;
                   1026:                        man->next = ROFF_NEXT_CHILD;
                   1027:                } else {
                   1028:                        man->last = n->prev;
                   1029:                        man->next = ROFF_NEXT_SIBLING;
                   1030:                }
                   1031:        }
                   1032:        if (man->first == n)
                   1033:                man->first = NULL;
                   1034: }
                   1035:
                   1036: void
                   1037: roff_node_free(struct roff_node *n)
                   1038: {
                   1039:
                   1040:        if (n->args != NULL)
                   1041:                mdoc_argv_free(n->args);
                   1042:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1043:                free(n->norm);
1.191     schwarze 1044:        if (n->eqn != NULL)
                   1045:                eqn_box_free(n->eqn);
1.138     schwarze 1046:        free(n->string);
                   1047:        free(n);
                   1048: }
                   1049:
                   1050: void
                   1051: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1052: {
                   1053:
                   1054:        while (n->child != NULL)
                   1055:                roff_node_delete(man, n->child);
                   1056:        roff_node_unlink(man, n);
                   1057:        roff_node_free(n);
1.141     schwarze 1058: }
                   1059:
                   1060: void
                   1061: deroff(char **dest, const struct roff_node *n)
                   1062: {
                   1063:        char    *cp;
                   1064:        size_t   sz;
                   1065:
                   1066:        if (n->type != ROFFT_TEXT) {
                   1067:                for (n = n->child; n != NULL; n = n->next)
                   1068:                        deroff(dest, n);
                   1069:                return;
                   1070:        }
                   1071:
1.160     schwarze 1072:        /* Skip leading whitespace. */
1.141     schwarze 1073:
1.160     schwarze 1074:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1075:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1076:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1077:                        cp++;
1.160     schwarze 1078:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1079:                        break;
                   1080:        }
                   1081:
1.161     schwarze 1082:        /* Skip trailing backslash. */
                   1083:
                   1084:        sz = strlen(cp);
1.162     schwarze 1085:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1086:                sz--;
                   1087:
1.141     schwarze 1088:        /* Skip trailing whitespace. */
                   1089:
1.161     schwarze 1090:        for (; sz; sz--)
1.141     schwarze 1091:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1092:                        break;
                   1093:
                   1094:        /* Skip empty strings. */
                   1095:
                   1096:        if (sz == 0)
                   1097:                return;
                   1098:
                   1099:        if (*dest == NULL) {
                   1100:                *dest = mandoc_strndup(cp, sz);
                   1101:                return;
                   1102:        }
                   1103:
                   1104:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1105:        free(*dest);
                   1106:        *dest = cp;
1.138     schwarze 1107: }
                   1108:
                   1109: /* --- main functions of the roff parser ---------------------------------- */
                   1110:
1.8       schwarze 1111: /*
1.79      schwarze 1112:  * In the current line, expand escape sequences that tend to get
                   1113:  * used in numerical expressions and conditional requests.
                   1114:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1115:  */
1.45      schwarze 1116: static enum rofferr
1.110     schwarze 1117: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1118: {
1.81      schwarze 1119:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze 1120:        const char      *start; /* start of the string to process */
1.82      schwarze 1121:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze 1122:        const char      *stnam; /* start of the name, after "[(*" */
                   1123:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1124:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1125:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1126:        size_t           maxl;  /* expected length of the escape name */
                   1127:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1128:        enum mandoc_esc  esc;   /* type of the escape sequence */
                   1129:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1130:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1131:        int              npos;  /* position in numeric expression */
1.91      schwarze 1132:        int              arg_complete; /* argument not interrupted by eol */
1.175     schwarze 1133:        int              done;  /* no more input available */
1.187     schwarze 1134:        int              deftype; /* type of definition to paste */
1.195     schwarze 1135:        int              rcsid; /* kind of RCS id seen */
1.79      schwarze 1136:        char             term;  /* character terminating the escape */
1.8       schwarze 1137:
1.175     schwarze 1138:        /* Search forward for comments. */
                   1139:
                   1140:        done = 0;
                   1141:        start = buf->buf + pos;
                   1142:        for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
                   1143:                if (stesc[0] != r->escape || stesc[1] == '\0')
                   1144:                        continue;
                   1145:                stesc++;
                   1146:                if (*stesc != '"' && *stesc != '#')
                   1147:                        continue;
1.185     schwarze 1148:
                   1149:                /* Comment found, look for RCS id. */
                   1150:
1.195     schwarze 1151:                rcsid = 0;
1.185     schwarze 1152:                if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.195     schwarze 1153:                        rcsid = 1 << MANDOC_OS_OPENBSD;
1.185     schwarze 1154:                        cp += 8;
                   1155:                } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.195     schwarze 1156:                        rcsid = 1 << MANDOC_OS_NETBSD;
1.185     schwarze 1157:                        cp += 7;
                   1158:                }
                   1159:                if (cp != NULL &&
                   1160:                    isalnum((unsigned char)*cp) == 0 &&
1.186     schwarze 1161:                    strchr(cp, '$') != NULL) {
1.195     schwarze 1162:                        if (r->man->meta.rcsids & rcsid)
1.186     schwarze 1163:                                mandoc_msg(MANDOCERR_RCS_REP, r->parse,
                   1164:                                    ln, stesc + 1 - buf->buf, stesc + 1);
1.195     schwarze 1165:                        r->man->meta.rcsids |= rcsid;
1.186     schwarze 1166:                }
1.185     schwarze 1167:
                   1168:                /* Handle trailing whitespace. */
                   1169:
1.175     schwarze 1170:                cp = strchr(stesc--, '\0') - 1;
                   1171:                if (*cp == '\n') {
                   1172:                        done = 1;
                   1173:                        cp--;
                   1174:                }
                   1175:                if (*cp == ' ' || *cp == '\t')
                   1176:                        mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
                   1177:                            ln, cp - buf->buf, NULL);
                   1178:                while (stesc > start && stesc[-1] == ' ')
                   1179:                        stesc--;
                   1180:                *stesc = '\0';
                   1181:                break;
                   1182:        }
                   1183:        if (stesc == start)
                   1184:                return ROFF_CONT;
                   1185:        stesc--;
                   1186:
                   1187:        /* Notice the end of the input. */
                   1188:
                   1189:        if (*stesc == '\n') {
                   1190:                *stesc-- = '\0';
                   1191:                done = 1;
                   1192:        }
                   1193:
1.43      schwarze 1194:        expand_count = 0;
1.175     schwarze 1195:        while (stesc >= start) {
1.43      schwarze 1196:
1.78      schwarze 1197:                /* Search backwards for the next backslash. */
                   1198:
1.175     schwarze 1199:                if (*stesc != r->escape) {
                   1200:                        if (*stesc == '\\') {
                   1201:                                *stesc = '\0';
                   1202:                                buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
                   1203:                                    buf->buf, stesc + 1) + 1;
                   1204:                                start = nbuf + pos;
                   1205:                                stesc = nbuf + (stesc - buf->buf);
                   1206:                                free(buf->buf);
                   1207:                                buf->buf = nbuf;
                   1208:                        }
                   1209:                        stesc--;
1.78      schwarze 1210:                        continue;
1.175     schwarze 1211:                }
1.78      schwarze 1212:
                   1213:                /* If it is escaped, skip it. */
                   1214:
                   1215:                for (cp = stesc - 1; cp >= start; cp--)
1.175     schwarze 1216:                        if (*cp != r->escape)
1.78      schwarze 1217:                                break;
                   1218:
1.110     schwarze 1219:                if ((stesc - cp) % 2 == 0) {
1.175     schwarze 1220:                        while (stesc > cp)
                   1221:                                *stesc-- = '\\';
1.78      schwarze 1222:                        continue;
1.175     schwarze 1223:                } else if (stesc[1] != '\0') {
                   1224:                        *stesc = '\\';
                   1225:                } else {
                   1226:                        *stesc-- = '\0';
                   1227:                        if (done)
                   1228:                                continue;
                   1229:                        else
                   1230:                                return ROFF_APPEND;
1.78      schwarze 1231:                }
1.23      schwarze 1232:
1.79      schwarze 1233:                /* Decide whether to expand or to check only. */
1.23      schwarze 1234:
1.79      schwarze 1235:                term = '\0';
1.78      schwarze 1236:                cp = stesc + 1;
1.53      schwarze 1237:                switch (*cp) {
1.80      schwarze 1238:                case '*':
1.53      schwarze 1239:                        res = NULL;
                   1240:                        break;
1.80      schwarze 1241:                case 'B':
                   1242:                case 'w':
1.79      schwarze 1243:                        term = cp[1];
                   1244:                        /* FALLTHROUGH */
1.80      schwarze 1245:                case 'n':
1.53      schwarze 1246:                        res = ubuf;
                   1247:                        break;
                   1248:                default:
1.109     schwarze 1249:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1250:                        if (esc == ESCAPE_ERROR ||
                   1251:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1252:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1253:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1254:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1255:                                    "%.*s", (int)(cp - stesc), stesc);
1.175     schwarze 1256:                        stesc--;
1.78      schwarze 1257:                        continue;
1.42      schwarze 1258:                }
                   1259:
1.78      schwarze 1260:                if (EXPAND_LIMIT < ++expand_count) {
                   1261:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1262:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1263:                        return ROFF_IGN;
1.78      schwarze 1264:                }
1.23      schwarze 1265:
                   1266:                /*
                   1267:                 * The third character decides the length
1.53      schwarze 1268:                 * of the name of the string or register.
1.23      schwarze 1269:                 * Save a pointer to the name.
                   1270:                 */
                   1271:
1.110     schwarze 1272:                if (term == '\0') {
1.79      schwarze 1273:                        switch (*++cp) {
1.80      schwarze 1274:                        case '\0':
1.79      schwarze 1275:                                maxl = 0;
                   1276:                                break;
1.80      schwarze 1277:                        case '(':
1.79      schwarze 1278:                                cp++;
                   1279:                                maxl = 2;
                   1280:                                break;
1.80      schwarze 1281:                        case '[':
1.79      schwarze 1282:                                cp++;
                   1283:                                term = ']';
                   1284:                                maxl = 0;
                   1285:                                break;
                   1286:                        default:
                   1287:                                maxl = 1;
                   1288:                                break;
                   1289:                        }
                   1290:                } else {
                   1291:                        cp += 2;
1.8       schwarze 1292:                        maxl = 0;
                   1293:                }
1.23      schwarze 1294:                stnam = cp;
1.8       schwarze 1295:
1.23      schwarze 1296:                /* Advance to the end of the name. */
1.8       schwarze 1297:
1.125     schwarze 1298:                naml = 0;
1.91      schwarze 1299:                arg_complete = 1;
1.125     schwarze 1300:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1301:                        if (*cp == '\0') {
1.92      schwarze 1302:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1303:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1304:                                arg_complete = 0;
1.79      schwarze 1305:                                break;
1.42      schwarze 1306:                        }
1.110     schwarze 1307:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1308:                                cp++;
1.125     schwarze 1309:                                break;
                   1310:                        }
                   1311:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1312:                                naml++;
                   1313:                                continue;
                   1314:                        }
                   1315:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1316:                        case ESCAPE_SPECIAL:
                   1317:                        case ESCAPE_UNICODE:
                   1318:                        case ESCAPE_NUMBERED:
                   1319:                        case ESCAPE_OVERSTRIKE:
                   1320:                                naml++;
                   1321:                                break;
                   1322:                        default:
1.8       schwarze 1323:                                break;
1.79      schwarze 1324:                        }
1.8       schwarze 1325:                }
                   1326:
1.23      schwarze 1327:                /*
                   1328:                 * Retrieve the replacement string; if it is
                   1329:                 * undefined, resume searching for escapes.
                   1330:                 */
                   1331:
1.79      schwarze 1332:                switch (stesc[1]) {
1.80      schwarze 1333:                case '*':
1.187     schwarze 1334:                        if (arg_complete) {
                   1335:                                deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1336:                                res = roff_getstrn(r, stnam, naml, &deftype);
                   1337:                        }
1.79      schwarze 1338:                        break;
1.80      schwarze 1339:                case 'B':
1.79      schwarze 1340:                        npos = 0;
1.91      schwarze 1341:                        ubuf[0] = arg_complete &&
1.133     schwarze 1342:                            roff_evalnum(r, ln, stnam, &npos,
                   1343:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1344:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1345:                        ubuf[1] = '\0';
                   1346:                        break;
1.80      schwarze 1347:                case 'n':
1.91      schwarze 1348:                        if (arg_complete)
                   1349:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1350:                                    roff_getregn(r, stnam, naml));
                   1351:                        else
                   1352:                                ubuf[0] = '\0';
1.79      schwarze 1353:                        break;
1.80      schwarze 1354:                case 'w':
1.91      schwarze 1355:                        /* use even incomplete args */
1.81      schwarze 1356:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1357:                            24 * (int)naml);
                   1358:                        break;
                   1359:                }
1.8       schwarze 1360:
1.110     schwarze 1361:                if (res == NULL) {
1.92      schwarze 1362:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1363:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1364:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1365:                        res = "";
1.118     schwarze 1366:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1367:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1368:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1369:                        return ROFF_IGN;
1.8       schwarze 1370:                }
                   1371:
1.23      schwarze 1372:                /* Replace the escape sequence by the string. */
                   1373:
1.82      schwarze 1374:                *stesc = '\0';
1.110     schwarze 1375:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1376:                    buf->buf, res, cp) + 1;
1.8       schwarze 1377:
1.78      schwarze 1378:                /* Prepare for the next replacement. */
1.8       schwarze 1379:
1.78      schwarze 1380:                start = nbuf + pos;
1.110     schwarze 1381:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1382:                free(buf->buf);
                   1383:                buf->buf = nbuf;
1.42      schwarze 1384:        }
1.149     schwarze 1385:        return ROFF_CONT;
1.42      schwarze 1386: }
                   1387:
                   1388: /*
1.147     schwarze 1389:  * Process text streams.
1.42      schwarze 1390:  */
                   1391: static enum rofferr
1.177     schwarze 1392: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42      schwarze 1393: {
                   1394:        size_t           sz;
                   1395:        const char      *start;
1.51      schwarze 1396:        char            *p;
                   1397:        int              isz;
1.42      schwarze 1398:        enum mandoc_esc  esc;
                   1399:
1.147     schwarze 1400:        /* Spring the input line trap. */
                   1401:
                   1402:        if (roffit_lines == 1) {
                   1403:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1404:                free(buf->buf);
                   1405:                buf->buf = p;
                   1406:                buf->sz = isz + 1;
                   1407:                *offs = 0;
                   1408:                free(roffit_macro);
                   1409:                roffit_lines = 0;
1.149     schwarze 1410:                return ROFF_REPARSE;
1.147     schwarze 1411:        } else if (roffit_lines > 1)
                   1412:                --roffit_lines;
                   1413:
1.177     schwarze 1414:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1415:                if (roffce_lines < 1) {
                   1416:                        r->man->last = roffce_node;
                   1417:                        r->man->next = ROFF_NEXT_SIBLING;
                   1418:                        roffce_lines = 0;
                   1419:                        roffce_node = NULL;
                   1420:                } else
                   1421:                        roffce_lines--;
                   1422:        }
                   1423:
1.147     schwarze 1424:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1425:
1.110     schwarze 1426:        start = p = buf->buf + pos;
1.42      schwarze 1427:
1.110     schwarze 1428:        while (*p != '\0') {
1.42      schwarze 1429:                sz = strcspn(p, "-\\");
                   1430:                p += sz;
                   1431:
1.110     schwarze 1432:                if (*p == '\0')
1.42      schwarze 1433:                        break;
                   1434:
1.110     schwarze 1435:                if (*p == '\\') {
1.42      schwarze 1436:                        /* Skip over escapes. */
                   1437:                        p++;
1.62      schwarze 1438:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1439:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1440:                                break;
1.136     schwarze 1441:                        while (*p == '-')
                   1442:                                p++;
1.42      schwarze 1443:                        continue;
                   1444:                } else if (p == start) {
                   1445:                        p++;
                   1446:                        continue;
                   1447:                }
                   1448:
1.44      schwarze 1449:                if (isalpha((unsigned char)p[-1]) &&
                   1450:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1451:                        *p = ASCII_HYPH;
                   1452:                p++;
1.8       schwarze 1453:        }
1.149     schwarze 1454:        return ROFF_CONT;
1.8       schwarze 1455: }
                   1456:
1.1       schwarze 1457: enum rofferr
1.110     schwarze 1458: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1459: {
1.166     schwarze 1460:        enum roff_tok    t;
1.27      schwarze 1461:        enum rofferr     e;
1.110     schwarze 1462:        int              pos;   /* parse point */
1.115     schwarze 1463:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1464:        int              ppos;  /* original offset in buf->buf */
                   1465:        int              ctl;   /* macro line (boolean) */
                   1466:
                   1467:        ppos = pos = *offs;
1.1       schwarze 1468:
1.102     schwarze 1469:        /* Handle in-line equation delimiters. */
                   1470:
1.108     schwarze 1471:        if (r->tbl == NULL &&
                   1472:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1473:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1474:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1475:                if (e == ROFF_REPARSE)
1.149     schwarze 1476:                        return e;
1.102     schwarze 1477:                assert(e == ROFF_CONT);
                   1478:        }
                   1479:
                   1480:        /* Expand some escape sequences. */
1.8       schwarze 1481:
1.110     schwarze 1482:        e = roff_res(r, buf, ln, pos);
1.175     schwarze 1483:        if (e == ROFF_IGN || e == ROFF_APPEND)
1.149     schwarze 1484:                return e;
1.110     schwarze 1485:        assert(e == ROFF_CONT);
1.8       schwarze 1486:
1.110     schwarze 1487:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1488:
1.8       schwarze 1489:        /*
1.2       schwarze 1490:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1491:         * text through the macro's filter.
                   1492:         * Equations process all content themselves.
                   1493:         * Tables process almost all content themselves, but we want
                   1494:         * to warn about macros before passing it there.
1.2       schwarze 1495:         */
                   1496:
1.124     schwarze 1497:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1498:                t = r->last->tok;
1.110     schwarze 1499:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1500:                if (e == ROFF_IGN)
1.149     schwarze 1501:                        return e;
1.166     schwarze 1502:                assert(e == ROFF_CONT);
1.54      schwarze 1503:        }
1.191     schwarze 1504:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1505:                eqn_read(r->eqn, buf->buf + ppos);
                   1506:                return ROFF_IGN;
                   1507:        }
1.193     schwarze 1508:        if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
                   1509:                tbl_read(r->tbl, ln, buf->buf, ppos);
                   1510:                roff_addtbl(r->man, r->tbl);
                   1511:                return ROFF_IGN;
                   1512:        }
1.124     schwarze 1513:        if ( ! ctl)
1.177     schwarze 1514:                return roff_parsetext(r, buf, pos, offs);
1.100     schwarze 1515:
                   1516:        /* Skip empty request lines. */
                   1517:
1.110     schwarze 1518:        if (buf->buf[pos] == '"') {
1.100     schwarze 1519:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1520:                    ln, pos, NULL);
1.149     schwarze 1521:                return ROFF_IGN;
1.110     schwarze 1522:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1523:                return ROFF_IGN;
1.2       schwarze 1524:
                   1525:        /*
                   1526:         * If a scope is open, go to the child handler for that macro,
                   1527:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1528:         * Don't do so if an equation is open.
1.2       schwarze 1529:         */
                   1530:
                   1531:        if (r->last) {
1.1       schwarze 1532:                t = r->last->tok;
1.149     schwarze 1533:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1534:        }
                   1535:
1.115     schwarze 1536:        /* No scope is open.  This is a new request or macro. */
                   1537:
                   1538:        spos = pos;
                   1539:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1540:
                   1541:        /* Tables ignore most macros. */
                   1542:
1.180     schwarze 1543:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181     schwarze 1544:            t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.115     schwarze 1545:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1546:                    ln, pos, buf->buf + spos);
1.180     schwarze 1547:                if (t != TOKEN_NONE)
1.149     schwarze 1548:                        return ROFF_IGN;
1.129     schwarze 1549:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1550:                        pos++;
1.163     schwarze 1551:                while (buf->buf[pos] == ' ')
1.129     schwarze 1552:                        pos++;
1.193     schwarze 1553:                tbl_read(r->tbl, ln, buf->buf, pos);
                   1554:                roff_addtbl(r->man, r->tbl);
                   1555:                return ROFF_IGN;
1.115     schwarze 1556:        }
                   1557:
1.177     schwarze 1558:        /* For now, let high level macros abort .ce mode. */
                   1559:
                   1560:        if (ctl && roffce_node != NULL &&
1.196     schwarze 1561:            (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
                   1562:             t == ROFF_TH || t == ROFF_TS)) {
1.177     schwarze 1563:                r->man->last = roffce_node;
                   1564:                r->man->next = ROFF_NEXT_SIBLING;
                   1565:                roffce_lines = 0;
                   1566:                roffce_node = NULL;
                   1567:        }
                   1568:
1.2       schwarze 1569:        /*
1.115     schwarze 1570:         * This is neither a roff request nor a user-defined macro.
                   1571:         * Let the standard macro set parsers handle it.
1.2       schwarze 1572:         */
                   1573:
1.166     schwarze 1574:        if (t == TOKEN_NONE)
1.149     schwarze 1575:                return ROFF_CONT;
1.115     schwarze 1576:
                   1577:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1578:
1.168     schwarze 1579:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2       schwarze 1580: }
                   1581:
1.27      schwarze 1582: void
1.2       schwarze 1583: roff_endparse(struct roff *r)
                   1584: {
1.193     schwarze 1585:        if (r->last != NULL)
1.94      schwarze 1586:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1587:                    r->last->line, r->last->col,
1.167     schwarze 1588:                    roff_name[r->last->tok]);
1.27      schwarze 1589:
1.193     schwarze 1590:        if (r->eqn != NULL) {
1.94      schwarze 1591:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.191     schwarze 1592:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1593:                eqn_parse(r->eqn);
                   1594:                r->eqn = NULL;
1.32      schwarze 1595:        }
                   1596:
1.193     schwarze 1597:        if (r->tbl != NULL) {
1.94      schwarze 1598:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1599:                    r->tbl->line, r->tbl->pos, "TS");
1.193     schwarze 1600:                tbl_end(r->tbl);
                   1601:                r->tbl = NULL;
1.27      schwarze 1602:        }
1.1       schwarze 1603: }
                   1604:
                   1605: /*
                   1606:  * Parse a roff node's type from the input buffer.  This must be in the
                   1607:  * form of ".foo xxx" in the usual way.
                   1608:  */
1.166     schwarze 1609: static enum roff_tok
1.87      schwarze 1610: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1611: {
1.87      schwarze 1612:        char            *cp;
1.16      schwarze 1613:        const char      *mac;
                   1614:        size_t           maclen;
1.187     schwarze 1615:        int              deftype;
1.166     schwarze 1616:        enum roff_tok    t;
1.1       schwarze 1617:
1.87      schwarze 1618:        cp = buf + *pos;
                   1619:
                   1620:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1621:                return TOKEN_NONE;
1.1       schwarze 1622:
1.87      schwarze 1623:        mac = cp;
                   1624:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1625:
1.187     schwarze 1626:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1627:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1628:        switch (deftype) {
                   1629:        case ROFFDEF_USER:
                   1630:                t = ROFF_USERDEF;
                   1631:                break;
                   1632:        case ROFFDEF_REN:
                   1633:                t = ROFF_RENAMED;
                   1634:                break;
                   1635:        default:
                   1636:                t = roffhash_find(r->reqtab, mac, maclen);
                   1637:                break;
                   1638:        }
1.166     schwarze 1639:        if (t != TOKEN_NONE)
1.87      schwarze 1640:                *pos = cp - buf;
1.197   ! schwarze 1641:        else if (deftype == ROFFDEF_UNDEF) {
        !          1642:                /* Using an undefined macro defines it to be empty. */
        !          1643:                roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
        !          1644:                roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
        !          1645:        }
1.149     schwarze 1646:        return t;
1.1       schwarze 1647: }
                   1648:
1.138     schwarze 1649: /* --- handling of request blocks ----------------------------------------- */
                   1650:
1.1       schwarze 1651: static enum rofferr
1.2       schwarze 1652: roff_cblock(ROFF_ARGS)
1.1       schwarze 1653: {
                   1654:
1.2       schwarze 1655:        /*
                   1656:         * A block-close `..' should only be invoked as a child of an
                   1657:         * ignore macro, otherwise raise a warning and just ignore it.
                   1658:         */
                   1659:
1.110     schwarze 1660:        if (r->last == NULL) {
1.94      schwarze 1661:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1662:                    ln, ppos, "..");
1.149     schwarze 1663:                return ROFF_IGN;
1.2       schwarze 1664:        }
1.1       schwarze 1665:
1.2       schwarze 1666:        switch (r->last->tok) {
1.80      schwarze 1667:        case ROFF_am:
1.93      schwarze 1668:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1669:        case ROFF_ami:
                   1670:        case ROFF_de:
1.23      schwarze 1671:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1672:        case ROFF_dei:
                   1673:        case ROFF_ig:
1.2       schwarze 1674:                break;
                   1675:        default:
1.94      schwarze 1676:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1677:                    ln, ppos, "..");
1.149     schwarze 1678:                return ROFF_IGN;
1.2       schwarze 1679:        }
                   1680:
1.110     schwarze 1681:        if (buf->buf[pos] != '\0')
1.90      schwarze 1682:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1683:                    ".. %s", buf->buf + pos);
1.2       schwarze 1684:
                   1685:        roffnode_pop(r);
                   1686:        roffnode_cleanscope(r);
1.149     schwarze 1687:        return ROFF_IGN;
1.2       schwarze 1688:
                   1689: }
1.1       schwarze 1690:
1.2       schwarze 1691: static void
                   1692: roffnode_cleanscope(struct roff *r)
                   1693: {
1.1       schwarze 1694:
1.2       schwarze 1695:        while (r->last) {
1.46      schwarze 1696:                if (--r->last->endspan != 0)
1.2       schwarze 1697:                        break;
                   1698:                roffnode_pop(r);
                   1699:        }
                   1700: }
1.1       schwarze 1701:
1.68      schwarze 1702: static void
                   1703: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1704: {
1.1       schwarze 1705:
1.2       schwarze 1706:        if (NULL == r->last) {
1.94      schwarze 1707:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1708:                    ln, ppos, "\\}");
1.68      schwarze 1709:                return;
1.2       schwarze 1710:        }
1.1       schwarze 1711:
1.2       schwarze 1712:        switch (r->last->tok) {
1.80      schwarze 1713:        case ROFF_el:
                   1714:        case ROFF_ie:
                   1715:        case ROFF_if:
1.2       schwarze 1716:                break;
                   1717:        default:
1.94      schwarze 1718:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1719:                    ln, ppos, "\\}");
1.68      schwarze 1720:                return;
1.2       schwarze 1721:        }
1.1       schwarze 1722:
1.2       schwarze 1723:        if (r->last->endspan > -1) {
1.94      schwarze 1724:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1725:                    ln, ppos, "\\}");
1.68      schwarze 1726:                return;
1.2       schwarze 1727:        }
                   1728:
                   1729:        roffnode_pop(r);
                   1730:        roffnode_cleanscope(r);
1.68      schwarze 1731:        return;
1.1       schwarze 1732: }
                   1733:
                   1734: static enum rofferr
1.2       schwarze 1735: roff_block(ROFF_ARGS)
1.1       schwarze 1736: {
1.187     schwarze 1737:        const char      *name, *value;
                   1738:        char            *call, *cp, *iname, *rname;
                   1739:        size_t           csz, namesz, rsz;
                   1740:        int              deftype;
1.16      schwarze 1741:
1.93      schwarze 1742:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1743:
1.110     schwarze 1744:        if (tok == ROFF_de1)
1.93      schwarze 1745:                tok = ROFF_de;
1.123     schwarze 1746:        else if (tok == ROFF_dei1)
                   1747:                tok = ROFF_dei;
1.110     schwarze 1748:        else if (tok == ROFF_am1)
1.93      schwarze 1749:                tok = ROFF_am;
1.123     schwarze 1750:        else if (tok == ROFF_ami1)
                   1751:                tok = ROFF_ami;
1.93      schwarze 1752:
                   1753:        /* Parse the macro name argument. */
                   1754:
1.110     schwarze 1755:        cp = buf->buf + pos;
                   1756:        if (tok == ROFF_ig) {
1.93      schwarze 1757:                iname = NULL;
                   1758:                namesz = 0;
                   1759:        } else {
                   1760:                iname = cp;
                   1761:                namesz = roff_getname(r, &cp, ln, ppos);
                   1762:                iname[namesz] = '\0';
                   1763:        }
1.22      schwarze 1764:
1.93      schwarze 1765:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1766:
1.110     schwarze 1767:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1768:                deftype = ROFFDEF_USER;
                   1769:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1770:                if (name == NULL) {
1.93      schwarze 1771:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1772:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1773:                            "%.*s", (int)namesz, iname);
                   1774:                        namesz = 0;
                   1775:                } else
                   1776:                        namesz = strlen(name);
                   1777:        } else
                   1778:                name = iname;
1.22      schwarze 1779:
1.110     schwarze 1780:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1781:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167     schwarze 1782:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1783:                return ROFF_IGN;
1.93      schwarze 1784:        }
1.2       schwarze 1785:
1.16      schwarze 1786:        roffnode_push(r, tok, name, ln, ppos);
                   1787:
                   1788:        /*
                   1789:         * At the beginning of a `de' macro, clear the existing string
                   1790:         * with the same name, if there is one.  New content will be
1.66      schwarze 1791:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1792:         */
1.22      schwarze 1793:
1.183     schwarze 1794:        if (tok == ROFF_de || tok == ROFF_dei) {
1.86      schwarze 1795:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183     schwarze 1796:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.187     schwarze 1797:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   1798:                deftype = ROFFDEF_ANY;
                   1799:                value = roff_getstrn(r, iname, namesz, &deftype);
                   1800:                switch (deftype) {  /* Before appending, ... */
                   1801:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   1802:                        roff_setstrn(&r->strtab, name, namesz,
                   1803:                            value, strlen(value), 0);
                   1804:                        break;
                   1805:                case ROFFDEF_REN: /* call original standard macro. */
                   1806:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1807:                            (int)strlen(value), value);
                   1808:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1809:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   1810:                        free(call);
                   1811:                        break;
                   1812:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   1813:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   1814:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   1815:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1816:                            (int)rsz, rname);
                   1817:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1818:                        free(call);
                   1819:                        free(rname);
                   1820:                        break;
                   1821:                default:
                   1822:                        break;
                   1823:                }
1.183     schwarze 1824:        }
1.2       schwarze 1825:
1.110     schwarze 1826:        if (*cp == '\0')
1.149     schwarze 1827:                return ROFF_IGN;
1.1       schwarze 1828:
1.93      schwarze 1829:        /* Get the custom end marker. */
1.22      schwarze 1830:
1.93      schwarze 1831:        iname = cp;
1.86      schwarze 1832:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1833:
                   1834:        /* Resolve the end marker if it is indirect. */
                   1835:
1.110     schwarze 1836:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1837:                deftype = ROFFDEF_USER;
                   1838:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1839:                if (name == NULL) {
1.93      schwarze 1840:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1841:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1842:                            "%.*s", (int)namesz, iname);
                   1843:                        namesz = 0;
                   1844:                } else
                   1845:                        namesz = strlen(name);
                   1846:        } else
                   1847:                name = iname;
                   1848:
1.86      schwarze 1849:        if (namesz)
                   1850:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1851:
1.110     schwarze 1852:        if (*cp != '\0')
1.90      schwarze 1853:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167     schwarze 1854:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1855:
1.149     schwarze 1856:        return ROFF_IGN;
1.1       schwarze 1857: }
                   1858:
                   1859: static enum rofferr
1.2       schwarze 1860: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1861: {
1.166     schwarze 1862:        enum roff_tok   t;
1.2       schwarze 1863:        int             i, j;
                   1864:
                   1865:        /*
                   1866:         * First check whether a custom macro exists at this level.  If
                   1867:         * it does, then check against it.  This is some of groff's
                   1868:         * stranger behaviours.  If we encountered a custom end-scope
                   1869:         * tag and that tag also happens to be a "real" macro, then we
                   1870:         * need to try interpreting it again as a real macro.  If it's
                   1871:         * not, then return ignore.  Else continue.
                   1872:         */
                   1873:
                   1874:        if (r->last->end) {
1.35      schwarze 1875:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1876:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1877:                                break;
1.1       schwarze 1878:
1.110     schwarze 1879:                if (r->last->end[j] == '\0' &&
                   1880:                    (buf->buf[i] == '\0' ||
                   1881:                     buf->buf[i] == ' ' ||
                   1882:                     buf->buf[i] == '\t')) {
1.2       schwarze 1883:                        roffnode_pop(r);
                   1884:                        roffnode_cleanscope(r);
1.1       schwarze 1885:
1.110     schwarze 1886:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1887:                                i++;
                   1888:
                   1889:                        pos = i;
1.110     schwarze 1890:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1891:                            TOKEN_NONE)
1.149     schwarze 1892:                                return ROFF_RERUN;
                   1893:                        return ROFF_IGN;
1.2       schwarze 1894:                }
1.1       schwarze 1895:        }
                   1896:
1.2       schwarze 1897:        /*
                   1898:         * If we have no custom end-query or lookup failed, then try
                   1899:         * pulling it out of the hashtable.
                   1900:         */
1.1       schwarze 1901:
1.110     schwarze 1902:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1903:
1.110     schwarze 1904:        if (t != ROFF_cblock) {
                   1905:                if (tok != ROFF_ig)
                   1906:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1907:                return ROFF_IGN;
1.16      schwarze 1908:        }
1.1       schwarze 1909:
1.149     schwarze 1910:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1911: }
                   1912:
                   1913: static enum rofferr
                   1914: roff_block_text(ROFF_ARGS)
                   1915: {
                   1916:
1.110     schwarze 1917:        if (tok != ROFF_ig)
                   1918:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1919:
1.149     schwarze 1920:        return ROFF_IGN;
1.2       schwarze 1921: }
                   1922:
                   1923: static enum rofferr
                   1924: roff_cond_sub(ROFF_ARGS)
                   1925: {
1.166     schwarze 1926:        enum roff_tok    t;
1.37      schwarze 1927:        char            *ep;
1.71      schwarze 1928:        int              rr;
1.2       schwarze 1929:
                   1930:        rr = r->last->rule;
1.37      schwarze 1931:        roffnode_cleanscope(r);
1.39      schwarze 1932:
1.69      schwarze 1933:        /*
                   1934:         * If `\}' occurs on a macro line without a preceding macro,
                   1935:         * drop the line completely.
                   1936:         */
                   1937:
1.110     schwarze 1938:        ep = buf->buf + pos;
                   1939:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1940:                rr = 0;
1.69      schwarze 1941:
1.50      schwarze 1942:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1943:
1.110     schwarze 1944:        while ((ep = strchr(ep, '\\')) != NULL) {
1.190     schwarze 1945:                switch (ep[1]) {
                   1946:                case '}':
                   1947:                        memmove(ep, ep + 2, strlen(ep + 2) + 1);
                   1948:                        roff_ccond(r, ln, ep - buf->buf);
                   1949:                        break;
                   1950:                case '\0':
                   1951:                        ++ep;
                   1952:                        break;
                   1953:                default:
                   1954:                        ep += 2;
                   1955:                        break;
1.70      schwarze 1956:                }
1.50      schwarze 1957:        }
1.190     schwarze 1958:
                   1959:        /*
                   1960:         * Fully handle known macros when they are structurally
                   1961:         * required or when the conditional evaluated to true.
                   1962:         */
                   1963:
                   1964:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1965:        return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
                   1966:            ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
                   1967:            ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1968: }
                   1969:
                   1970: static enum rofferr
                   1971: roff_cond_text(ROFF_ARGS)
                   1972: {
1.37      schwarze 1973:        char            *ep;
1.71      schwarze 1974:        int              rr;
1.2       schwarze 1975:
                   1976:        rr = r->last->rule;
1.37      schwarze 1977:        roffnode_cleanscope(r);
1.1       schwarze 1978:
1.110     schwarze 1979:        ep = buf->buf + pos;
                   1980:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1981:                if (*(++ep) == '}') {
1.70      schwarze 1982:                        *ep = '&';
1.110     schwarze 1983:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1984:                }
1.119     schwarze 1985:                if (*ep != '\0')
                   1986:                        ++ep;
1.2       schwarze 1987:        }
1.149     schwarze 1988:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1989: }
                   1990:
1.138     schwarze 1991: /* --- handling of numeric and conditional expressions -------------------- */
                   1992:
1.77      schwarze 1993: /*
                   1994:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1995:  * If there is at least one digit, return success and advance the
                   1996:  * parse point, else return failure and let the parse point unchanged.
                   1997:  * Ignore overflows, treat them just like the C language.
                   1998:  */
1.56      schwarze 1999: static int
1.133     schwarze 2000: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 2001: {
1.133     schwarze 2002:        int      myres, scaled, n, p;
1.79      schwarze 2003:
                   2004:        if (NULL == res)
                   2005:                res = &myres;
1.56      schwarze 2006:
                   2007:        p = *pos;
                   2008:        n = v[p] == '-';
1.133     schwarze 2009:        if (n || v[p] == '+')
1.56      schwarze 2010:                p++;
                   2011:
1.133     schwarze 2012:        if (flags & ROFFNUM_WHITE)
                   2013:                while (isspace((unsigned char)v[p]))
                   2014:                        p++;
                   2015:
1.56      schwarze 2016:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 2017:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 2018:        if (p == *pos + n)
                   2019:                return 0;
                   2020:
                   2021:        if (n)
                   2022:                *res = -*res;
                   2023:
1.126     schwarze 2024:        /* Each number may be followed by one optional scaling unit. */
                   2025:
                   2026:        switch (v[p]) {
                   2027:        case 'f':
1.133     schwarze 2028:                scaled = *res * 65536;
1.126     schwarze 2029:                break;
                   2030:        case 'i':
1.133     schwarze 2031:                scaled = *res * 240;
1.126     schwarze 2032:                break;
                   2033:        case 'c':
1.133     schwarze 2034:                scaled = *res * 240 / 2.54;
1.126     schwarze 2035:                break;
                   2036:        case 'v':
                   2037:        case 'P':
1.133     schwarze 2038:                scaled = *res * 40;
1.126     schwarze 2039:                break;
                   2040:        case 'm':
                   2041:        case 'n':
1.133     schwarze 2042:                scaled = *res * 24;
1.126     schwarze 2043:                break;
                   2044:        case 'p':
1.133     schwarze 2045:                scaled = *res * 10 / 3;
1.126     schwarze 2046:                break;
                   2047:        case 'u':
1.133     schwarze 2048:                scaled = *res;
1.126     schwarze 2049:                break;
                   2050:        case 'M':
1.133     schwarze 2051:                scaled = *res * 6 / 25;
1.126     schwarze 2052:                break;
                   2053:        default:
1.133     schwarze 2054:                scaled = *res;
1.126     schwarze 2055:                p--;
                   2056:                break;
                   2057:        }
1.133     schwarze 2058:        if (flags & ROFFNUM_SCALE)
                   2059:                *res = scaled;
1.126     schwarze 2060:
                   2061:        *pos = p + 1;
1.149     schwarze 2062:        return 1;
1.56      schwarze 2063: }
                   2064:
1.71      schwarze 2065: /*
                   2066:  * Evaluate a string comparison condition.
                   2067:  * The first character is the delimiter.
                   2068:  * Succeed if the string up to its second occurrence
                   2069:  * matches the string up to its third occurence.
                   2070:  * Advance the cursor after the third occurrence
                   2071:  * or lacking that, to the end of the line.
                   2072:  */
                   2073: static int
                   2074: roff_evalstrcond(const char *v, int *pos)
                   2075: {
                   2076:        const char      *s1, *s2, *s3;
                   2077:        int              match;
                   2078:
                   2079:        match = 0;
                   2080:        s1 = v + *pos;          /* initial delimiter */
                   2081:        s2 = s1 + 1;            /* for scanning the first string */
                   2082:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2083:
                   2084:        if (NULL == s3)         /* found no middle delimiter */
                   2085:                goto out;
                   2086:
                   2087:        while ('\0' != *++s3) {
                   2088:                if (*s2 != *s3) {  /* mismatch */
                   2089:                        s3 = strchr(s3, *s1);
                   2090:                        break;
                   2091:                }
                   2092:                if (*s3 == *s1) {  /* found the final delimiter */
                   2093:                        match = 1;
                   2094:                        break;
                   2095:                }
                   2096:                s2++;
                   2097:        }
                   2098:
                   2099: out:
                   2100:        if (NULL == s3)
                   2101:                s3 = strchr(s2, '\0');
1.114     schwarze 2102:        else if (*s3 != '\0')
1.71      schwarze 2103:                s3++;
                   2104:        *pos = s3 - v;
1.149     schwarze 2105:        return match;
1.71      schwarze 2106: }
                   2107:
1.77      schwarze 2108: /*
                   2109:  * Evaluate an optionally negated single character, numerical,
                   2110:  * or string condition.
                   2111:  */
1.71      schwarze 2112: static int
1.143     schwarze 2113: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 2114: {
1.143     schwarze 2115:        char    *cp, *name;
                   2116:        size_t   sz;
1.187     schwarze 2117:        int      deftype, number, savepos, istrue, wanttrue;
1.5       schwarze 2118:
1.71      schwarze 2119:        if ('!' == v[*pos]) {
                   2120:                wanttrue = 0;
                   2121:                (*pos)++;
                   2122:        } else
                   2123:                wanttrue = 1;
                   2124:
1.5       schwarze 2125:        switch (v[*pos]) {
1.112     schwarze 2126:        case '\0':
1.149     schwarze 2127:                return 0;
1.80      schwarze 2128:        case 'n':
                   2129:        case 'o':
1.5       schwarze 2130:                (*pos)++;
1.149     schwarze 2131:                return wanttrue;
1.80      schwarze 2132:        case 'c':
                   2133:        case 'e':
                   2134:        case 't':
1.111     schwarze 2135:        case 'v':
1.5       schwarze 2136:                (*pos)++;
1.149     schwarze 2137:                return !wanttrue;
1.182     schwarze 2138:        case 'd':
1.143     schwarze 2139:        case 'r':
1.182     schwarze 2140:                cp = v + *pos + 1;
                   2141:                while (*cp == ' ')
                   2142:                        cp++;
                   2143:                name = cp;
                   2144:                sz = roff_getname(r, &cp, ln, cp - v);
1.187     schwarze 2145:                if (sz == 0)
                   2146:                        istrue = 0;
                   2147:                else if (v[*pos] == 'r')
                   2148:                        istrue = roff_hasregn(r, name, sz);
                   2149:                else {
                   2150:                        deftype = ROFFDEF_ANY;
                   2151:                        roff_getstrn(r, name, sz, &deftype);
                   2152:                        istrue = !!deftype;
                   2153:                }
1.143     schwarze 2154:                *pos = cp - v;
1.182     schwarze 2155:                return istrue == wanttrue;
1.5       schwarze 2156:        default:
                   2157:                break;
                   2158:        }
                   2159:
1.113     schwarze 2160:        savepos = *pos;
1.133     schwarze 2161:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 2162:                return (number > 0) == wanttrue;
1.113     schwarze 2163:        else if (*pos == savepos)
1.149     schwarze 2164:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 2165:        else
1.149     schwarze 2166:                return 0;
1.5       schwarze 2167: }
                   2168:
1.2       schwarze 2169: static enum rofferr
1.21      schwarze 2170: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 2171: {
1.30      schwarze 2172:
1.149     schwarze 2173:        return ROFF_IGN;
1.21      schwarze 2174: }
                   2175:
                   2176: static enum rofferr
1.123     schwarze 2177: roff_insec(ROFF_ARGS)
                   2178: {
                   2179:
                   2180:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167     schwarze 2181:            ln, ppos, roff_name[tok]);
1.149     schwarze 2182:        return ROFF_IGN;
1.123     schwarze 2183: }
                   2184:
                   2185: static enum rofferr
                   2186: roff_unsupp(ROFF_ARGS)
                   2187: {
                   2188:
                   2189:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167     schwarze 2190:            ln, ppos, roff_name[tok]);
1.149     schwarze 2191:        return ROFF_IGN;
1.123     schwarze 2192: }
                   2193:
                   2194: static enum rofferr
1.2       schwarze 2195: roff_cond(ROFF_ARGS)
                   2196: {
1.46      schwarze 2197:
                   2198:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2199:
1.80      schwarze 2200:        /*
1.35      schwarze 2201:         * An `.el' has no conditional body: it will consume the value
                   2202:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2203:         * defaults to DENY.
1.35      schwarze 2204:         *
                   2205:         * If we're not an `el', however, then evaluate the conditional.
                   2206:         */
1.1       schwarze 2207:
1.110     schwarze 2208:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2209:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2210:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2211:
1.35      schwarze 2212:        /*
                   2213:         * An if-else will put the NEGATION of the current evaluated
                   2214:         * conditional into the stack of rules.
                   2215:         */
                   2216:
1.110     schwarze 2217:        if (tok == ROFF_ie) {
1.96      schwarze 2218:                if (r->rstackpos + 1 == r->rstacksz) {
                   2219:                        r->rstacksz += 16;
                   2220:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2221:                            r->rstacksz, sizeof(int));
1.35      schwarze 2222:                }
1.71      schwarze 2223:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2224:        }
1.5       schwarze 2225:
                   2226:        /* If the parent has false as its rule, then so do we. */
                   2227:
1.71      schwarze 2228:        if (r->last->parent && !r->last->parent->rule)
                   2229:                r->last->rule = 0;
1.5       schwarze 2230:
                   2231:        /*
1.46      schwarze 2232:         * Determine scope.
                   2233:         * If there is nothing on the line after the conditional,
                   2234:         * not even whitespace, use next-line scope.
1.5       schwarze 2235:         */
1.2       schwarze 2236:
1.110     schwarze 2237:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2238:                r->last->endspan = 2;
                   2239:                goto out;
                   2240:        }
                   2241:
1.110     schwarze 2242:        while (buf->buf[pos] == ' ')
1.46      schwarze 2243:                pos++;
                   2244:
                   2245:        /* An opening brace requests multiline scope. */
1.2       schwarze 2246:
1.110     schwarze 2247:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2248:                r->last->endspan = -1;
                   2249:                pos += 2;
1.144     schwarze 2250:                while (buf->buf[pos] == ' ')
                   2251:                        pos++;
1.46      schwarze 2252:                goto out;
1.80      schwarze 2253:        }
1.2       schwarze 2254:
                   2255:        /*
1.46      schwarze 2256:         * Anything else following the conditional causes
                   2257:         * single-line scope.  Warn if the scope contains
                   2258:         * nothing but trailing whitespace.
1.2       schwarze 2259:         */
                   2260:
1.110     schwarze 2261:        if (buf->buf[pos] == '\0')
1.89      schwarze 2262:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167     schwarze 2263:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2264:
1.46      schwarze 2265:        r->last->endspan = 1;
1.1       schwarze 2266:
1.46      schwarze 2267: out:
1.2       schwarze 2268:        *offs = pos;
1.149     schwarze 2269:        return ROFF_RERUN;
1.1       schwarze 2270: }
                   2271:
1.2       schwarze 2272: static enum rofferr
1.7       schwarze 2273: roff_ds(ROFF_ARGS)
                   2274: {
1.85      schwarze 2275:        char            *string;
                   2276:        const char      *name;
                   2277:        size_t           namesz;
1.10      schwarze 2278:
1.123     schwarze 2279:        /* Ignore groff compatibility mode for now. */
                   2280:
                   2281:        if (tok == ROFF_ds1)
                   2282:                tok = ROFF_ds;
                   2283:        else if (tok == ROFF_as1)
                   2284:                tok = ROFF_as;
                   2285:
1.10      schwarze 2286:        /*
1.85      schwarze 2287:         * The first word is the name of the string.
                   2288:         * If it is empty or terminated by an escape sequence,
                   2289:         * abort the `ds' request without defining anything.
1.10      schwarze 2290:         */
1.7       schwarze 2291:
1.110     schwarze 2292:        name = string = buf->buf + pos;
                   2293:        if (*name == '\0')
1.149     schwarze 2294:                return ROFF_IGN;
1.7       schwarze 2295:
1.85      schwarze 2296:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2297:        if (name[namesz] == '\\')
1.149     schwarze 2298:                return ROFF_IGN;
1.85      schwarze 2299:
                   2300:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2301:        if (*string == '"')
1.7       schwarze 2302:                string++;
                   2303:
1.10      schwarze 2304:        /* The rest is the value. */
1.85      schwarze 2305:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2306:            ROFF_as == tok);
1.183     schwarze 2307:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149     schwarze 2308:        return ROFF_IGN;
1.7       schwarze 2309: }
                   2310:
1.77      schwarze 2311: /*
                   2312:  * Parse a single operator, one or two characters long.
                   2313:  * If the operator is recognized, return success and advance the
                   2314:  * parse point, else return failure and let the parse point unchanged.
                   2315:  */
                   2316: static int
                   2317: roff_getop(const char *v, int *pos, char *res)
                   2318: {
                   2319:
                   2320:        *res = v[*pos];
                   2321:
                   2322:        switch (*res) {
1.80      schwarze 2323:        case '+':
                   2324:        case '-':
                   2325:        case '*':
                   2326:        case '/':
                   2327:        case '%':
                   2328:        case '&':
                   2329:        case ':':
1.77      schwarze 2330:                break;
                   2331:        case '<':
                   2332:                switch (v[*pos + 1]) {
1.80      schwarze 2333:                case '=':
1.77      schwarze 2334:                        *res = 'l';
                   2335:                        (*pos)++;
                   2336:                        break;
1.80      schwarze 2337:                case '>':
1.77      schwarze 2338:                        *res = '!';
                   2339:                        (*pos)++;
                   2340:                        break;
1.80      schwarze 2341:                case '?':
1.77      schwarze 2342:                        *res = 'i';
                   2343:                        (*pos)++;
                   2344:                        break;
                   2345:                default:
                   2346:                        break;
                   2347:                }
                   2348:                break;
                   2349:        case '>':
                   2350:                switch (v[*pos + 1]) {
1.80      schwarze 2351:                case '=':
1.77      schwarze 2352:                        *res = 'g';
                   2353:                        (*pos)++;
                   2354:                        break;
1.80      schwarze 2355:                case '?':
1.77      schwarze 2356:                        *res = 'a';
                   2357:                        (*pos)++;
                   2358:                        break;
                   2359:                default:
                   2360:                        break;
                   2361:                }
                   2362:                break;
                   2363:        case '=':
                   2364:                if ('=' == v[*pos + 1])
                   2365:                        (*pos)++;
                   2366:                break;
                   2367:        default:
1.149     schwarze 2368:                return 0;
1.77      schwarze 2369:        }
                   2370:        (*pos)++;
                   2371:
1.149     schwarze 2372:        return *res;
1.77      schwarze 2373: }
                   2374:
                   2375: /*
                   2376:  * Evaluate either a parenthesized numeric expression
                   2377:  * or a single signed integer number.
                   2378:  */
                   2379: static int
1.106     schwarze 2380: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2381:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2382: {
                   2383:
                   2384:        if ('(' != v[*pos])
1.149     schwarze 2385:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2386:
                   2387:        (*pos)++;
1.133     schwarze 2388:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2389:                return 0;
1.77      schwarze 2390:
1.79      schwarze 2391:        /*
                   2392:         * Omission of the closing parenthesis
                   2393:         * is an error in validation mode,
                   2394:         * but ignored in evaluation mode.
                   2395:         */
                   2396:
1.77      schwarze 2397:        if (')' == v[*pos])
                   2398:                (*pos)++;
1.79      schwarze 2399:        else if (NULL == res)
1.149     schwarze 2400:                return 0;
1.77      schwarze 2401:
1.149     schwarze 2402:        return 1;
1.77      schwarze 2403: }
                   2404:
                   2405: /*
                   2406:  * Evaluate a complete numeric expression.
                   2407:  * Proceed left to right, there is no concept of precedence.
                   2408:  */
                   2409: static int
1.106     schwarze 2410: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2411:        int *pos, int *res, int flags)
1.77      schwarze 2412: {
                   2413:        int              mypos, operand2;
                   2414:        char             operator;
                   2415:
                   2416:        if (NULL == pos) {
                   2417:                mypos = 0;
                   2418:                pos = &mypos;
                   2419:        }
                   2420:
1.133     schwarze 2421:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2422:                while (isspace((unsigned char)v[*pos]))
                   2423:                        (*pos)++;
                   2424:
1.133     schwarze 2425:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2426:                return 0;
1.77      schwarze 2427:
                   2428:        while (1) {
1.133     schwarze 2429:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2430:                        while (isspace((unsigned char)v[*pos]))
                   2431:                                (*pos)++;
                   2432:
                   2433:                if ( ! roff_getop(v, pos, &operator))
                   2434:                        break;
                   2435:
1.133     schwarze 2436:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2437:                        while (isspace((unsigned char)v[*pos]))
                   2438:                                (*pos)++;
                   2439:
1.133     schwarze 2440:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2441:                        return 0;
1.77      schwarze 2442:
1.133     schwarze 2443:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2444:                        while (isspace((unsigned char)v[*pos]))
                   2445:                                (*pos)++;
1.79      schwarze 2446:
                   2447:                if (NULL == res)
                   2448:                        continue;
1.77      schwarze 2449:
                   2450:                switch (operator) {
1.80      schwarze 2451:                case '+':
1.77      schwarze 2452:                        *res += operand2;
                   2453:                        break;
1.80      schwarze 2454:                case '-':
1.77      schwarze 2455:                        *res -= operand2;
                   2456:                        break;
1.80      schwarze 2457:                case '*':
1.77      schwarze 2458:                        *res *= operand2;
                   2459:                        break;
1.80      schwarze 2460:                case '/':
1.116     schwarze 2461:                        if (operand2 == 0) {
1.106     schwarze 2462:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2463:                                        r->parse, ln, *pos, v);
                   2464:                                *res = 0;
                   2465:                                break;
                   2466:                        }
1.77      schwarze 2467:                        *res /= operand2;
                   2468:                        break;
1.80      schwarze 2469:                case '%':
1.116     schwarze 2470:                        if (operand2 == 0) {
                   2471:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2472:                                        r->parse, ln, *pos, v);
                   2473:                                *res = 0;
                   2474:                                break;
                   2475:                        }
1.77      schwarze 2476:                        *res %= operand2;
                   2477:                        break;
1.80      schwarze 2478:                case '<':
1.77      schwarze 2479:                        *res = *res < operand2;
                   2480:                        break;
1.80      schwarze 2481:                case '>':
1.77      schwarze 2482:                        *res = *res > operand2;
                   2483:                        break;
1.80      schwarze 2484:                case 'l':
1.77      schwarze 2485:                        *res = *res <= operand2;
                   2486:                        break;
1.80      schwarze 2487:                case 'g':
1.77      schwarze 2488:                        *res = *res >= operand2;
                   2489:                        break;
1.80      schwarze 2490:                case '=':
1.77      schwarze 2491:                        *res = *res == operand2;
                   2492:                        break;
1.80      schwarze 2493:                case '!':
1.77      schwarze 2494:                        *res = *res != operand2;
                   2495:                        break;
1.80      schwarze 2496:                case '&':
1.77      schwarze 2497:                        *res = *res && operand2;
                   2498:                        break;
1.80      schwarze 2499:                case ':':
1.77      schwarze 2500:                        *res = *res || operand2;
                   2501:                        break;
1.80      schwarze 2502:                case 'i':
1.77      schwarze 2503:                        if (operand2 < *res)
                   2504:                                *res = operand2;
                   2505:                        break;
1.80      schwarze 2506:                case 'a':
1.77      schwarze 2507:                        if (operand2 > *res)
                   2508:                                *res = operand2;
                   2509:                        break;
                   2510:                default:
                   2511:                        abort();
                   2512:                }
                   2513:        }
1.149     schwarze 2514:        return 1;
1.77      schwarze 2515: }
                   2516:
1.138     schwarze 2517: /* --- register management ------------------------------------------------ */
                   2518:
1.52      schwarze 2519: void
1.60      schwarze 2520: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2521: {
1.52      schwarze 2522:        struct roffreg  *reg;
                   2523:
                   2524:        /* Search for an existing register with the same name. */
                   2525:        reg = r->regtab;
                   2526:
                   2527:        while (reg && strcmp(name, reg->key.p))
                   2528:                reg = reg->next;
1.41      schwarze 2529:
1.52      schwarze 2530:        if (NULL == reg) {
                   2531:                /* Create a new register. */
                   2532:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2533:                reg->key.p = mandoc_strdup(name);
                   2534:                reg->key.sz = strlen(name);
1.60      schwarze 2535:                reg->val = 0;
1.52      schwarze 2536:                reg->next = r->regtab;
                   2537:                r->regtab = reg;
                   2538:        }
                   2539:
1.60      schwarze 2540:        if ('+' == sign)
                   2541:                reg->val += val;
                   2542:        else if ('-' == sign)
                   2543:                reg->val -= val;
                   2544:        else
                   2545:                reg->val = val;
1.41      schwarze 2546: }
                   2547:
1.65      schwarze 2548: /*
                   2549:  * Handle some predefined read-only number registers.
                   2550:  * For now, return -1 if the requested register is not predefined;
                   2551:  * in case a predefined read-only register having the value -1
                   2552:  * were to turn up, another special value would have to be chosen.
                   2553:  */
                   2554: static int
1.145     schwarze 2555: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2556: {
                   2557:
                   2558:        switch (*name) {
1.145     schwarze 2559:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2560:                return r->argc;
1.80      schwarze 2561:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2562:                return 0;
1.80      schwarze 2563:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2564:                return 1;
1.80      schwarze 2565:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2566:                return 24;
1.80      schwarze 2567:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2568:                return 0;
1.80      schwarze 2569:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2570:                return 1;
1.80      schwarze 2571:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2572:                return 40;
1.65      schwarze 2573:        default:
1.149     schwarze 2574:                return -1;
1.65      schwarze 2575:        }
                   2576: }
                   2577:
1.53      schwarze 2578: int
1.52      schwarze 2579: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2580: {
1.52      schwarze 2581:        struct roffreg  *reg;
1.65      schwarze 2582:        int              val;
                   2583:
                   2584:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145     schwarze 2585:                val = roff_getregro(r, name + 1);
1.65      schwarze 2586:                if (-1 != val)
1.149     schwarze 2587:                        return val;
1.65      schwarze 2588:        }
1.52      schwarze 2589:
                   2590:        for (reg = r->regtab; reg; reg = reg->next)
                   2591:                if (0 == strcmp(name, reg->key.p))
1.149     schwarze 2592:                        return reg->val;
1.53      schwarze 2593:
1.149     schwarze 2594:        return 0;
1.53      schwarze 2595: }
                   2596:
                   2597: static int
                   2598: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2599: {
                   2600:        struct roffreg  *reg;
1.65      schwarze 2601:        int              val;
                   2602:
                   2603:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2604:                val = roff_getregro(r, name + 1);
1.65      schwarze 2605:                if (-1 != val)
1.149     schwarze 2606:                        return val;
1.65      schwarze 2607:        }
1.53      schwarze 2608:
                   2609:        for (reg = r->regtab; reg; reg = reg->next)
                   2610:                if (len == reg->key.sz &&
                   2611:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2612:                        return reg->val;
1.143     schwarze 2613:
1.149     schwarze 2614:        return 0;
1.143     schwarze 2615: }
                   2616:
                   2617: static int
                   2618: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2619: {
                   2620:        struct roffreg  *reg;
                   2621:        int              val;
                   2622:
                   2623:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2624:                val = roff_getregro(r, name + 1);
1.143     schwarze 2625:                if (-1 != val)
1.149     schwarze 2626:                        return 1;
1.143     schwarze 2627:        }
                   2628:
                   2629:        for (reg = r->regtab; reg; reg = reg->next)
                   2630:                if (len == reg->key.sz &&
                   2631:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2632:                        return 1;
1.41      schwarze 2633:
1.149     schwarze 2634:        return 0;
1.41      schwarze 2635: }
                   2636:
1.52      schwarze 2637: static void
                   2638: roff_freereg(struct roffreg *reg)
1.41      schwarze 2639: {
1.52      schwarze 2640:        struct roffreg  *old_reg;
1.41      schwarze 2641:
1.52      schwarze 2642:        while (NULL != reg) {
                   2643:                free(reg->key.p);
                   2644:                old_reg = reg;
                   2645:                reg = reg->next;
                   2646:                free(old_reg);
                   2647:        }
1.41      schwarze 2648: }
1.7       schwarze 2649:
                   2650: static enum rofferr
1.6       schwarze 2651: roff_nr(ROFF_ARGS)
1.1       schwarze 2652: {
1.85      schwarze 2653:        char            *key, *val;
                   2654:        size_t           keysz;
1.37      schwarze 2655:        int              iv;
1.60      schwarze 2656:        char             sign;
1.6       schwarze 2657:
1.110     schwarze 2658:        key = val = buf->buf + pos;
                   2659:        if (*key == '\0')
1.149     schwarze 2660:                return ROFF_IGN;
1.85      schwarze 2661:
                   2662:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2663:        if (key[keysz] == '\\')
1.149     schwarze 2664:                return ROFF_IGN;
1.85      schwarze 2665:        key[keysz] = '\0';
1.6       schwarze 2666:
1.60      schwarze 2667:        sign = *val;
1.110     schwarze 2668:        if (sign == '+' || sign == '-')
1.60      schwarze 2669:                val++;
                   2670:
1.133     schwarze 2671:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77      schwarze 2672:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2673:
1.149     schwarze 2674:        return ROFF_IGN;
1.76      schwarze 2675: }
                   2676:
                   2677: static enum rofferr
                   2678: roff_rr(ROFF_ARGS)
                   2679: {
                   2680:        struct roffreg  *reg, **prev;
1.85      schwarze 2681:        char            *name, *cp;
                   2682:        size_t           namesz;
1.76      schwarze 2683:
1.110     schwarze 2684:        name = cp = buf->buf + pos;
                   2685:        if (*name == '\0')
1.149     schwarze 2686:                return ROFF_IGN;
1.85      schwarze 2687:        namesz = roff_getname(r, &cp, ln, pos);
                   2688:        name[namesz] = '\0';
1.76      schwarze 2689:
                   2690:        prev = &r->regtab;
                   2691:        while (1) {
                   2692:                reg = *prev;
1.110     schwarze 2693:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2694:                        break;
                   2695:                prev = &reg->next;
                   2696:        }
1.110     schwarze 2697:        if (reg != NULL) {
1.76      schwarze 2698:                *prev = reg->next;
                   2699:                free(reg->key.p);
                   2700:                free(reg);
                   2701:        }
1.149     schwarze 2702:        return ROFF_IGN;
1.29      schwarze 2703: }
                   2704:
1.138     schwarze 2705: /* --- handler functions for roff requests -------------------------------- */
                   2706:
1.29      schwarze 2707: static enum rofferr
                   2708: roff_rm(ROFF_ARGS)
                   2709: {
                   2710:        const char       *name;
                   2711:        char             *cp;
1.85      schwarze 2712:        size_t            namesz;
1.29      schwarze 2713:
1.110     schwarze 2714:        cp = buf->buf + pos;
                   2715:        while (*cp != '\0') {
1.85      schwarze 2716:                name = cp;
1.110     schwarze 2717:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2718:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183     schwarze 2719:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.110     schwarze 2720:                if (name[namesz] == '\\')
1.85      schwarze 2721:                        break;
1.29      schwarze 2722:        }
1.149     schwarze 2723:        return ROFF_IGN;
1.51      schwarze 2724: }
                   2725:
                   2726: static enum rofferr
                   2727: roff_it(ROFF_ARGS)
                   2728: {
                   2729:        int              iv;
                   2730:
                   2731:        /* Parse the number of lines. */
1.133     schwarze 2732:
                   2733:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2734:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2735:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2736:                return ROFF_IGN;
1.51      schwarze 2737:        }
                   2738:
1.134     schwarze 2739:        while (isspace((unsigned char)buf->buf[pos]))
                   2740:                pos++;
                   2741:
                   2742:        /*
                   2743:         * Arm the input line trap.
                   2744:         * Special-casing "an-trap" is an ugly workaround to cope
                   2745:         * with DocBook stupidly fiddling with man(7) internals.
                   2746:         */
1.133     schwarze 2747:
1.51      schwarze 2748:        roffit_lines = iv;
1.134     schwarze 2749:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2750:            strcmp(buf->buf + pos, "an-trap") ?
                   2751:            buf->buf + pos : "br");
1.149     schwarze 2752:        return ROFF_IGN;
1.47      schwarze 2753: }
                   2754:
                   2755: static enum rofferr
                   2756: roff_Dd(ROFF_ARGS)
                   2757: {
1.187     schwarze 2758:        int              mask;
                   2759:        enum roff_tok    t, te;
1.99      schwarze 2760:
1.187     schwarze 2761:        switch (tok) {
                   2762:        case ROFF_Dd:
                   2763:                tok = MDOC_Dd;
                   2764:                te = MDOC_MAX;
                   2765:                if (r->format == 0)
                   2766:                        r->format = MPARSE_MDOC;
                   2767:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   2768:                break;
                   2769:        case ROFF_TH:
                   2770:                tok = MAN_TH;
                   2771:                te = MAN_MAX;
                   2772:                if (r->format == 0)
                   2773:                        r->format = MPARSE_MAN;
                   2774:                mask = MPARSE_QUICK;
                   2775:                break;
                   2776:        default:
                   2777:                abort();
                   2778:        }
                   2779:        if ((r->options & mask) == 0)
                   2780:                for (t = tok; t < te; t++)
                   2781:                        roff_setstr(r, roff_name[t], NULL, 0);
1.149     schwarze 2782:        return ROFF_CONT;
1.14      schwarze 2783: }
                   2784:
                   2785: static enum rofferr
1.27      schwarze 2786: roff_TE(ROFF_ARGS)
                   2787: {
1.193     schwarze 2788:        if (r->tbl == NULL) {
1.94      schwarze 2789:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2790:                    ln, ppos, "TE");
1.193     schwarze 2791:                return ROFF_IGN;
                   2792:        }
                   2793:        if (tbl_end(r->tbl) == 0) {
                   2794:                r->tbl = NULL;
1.130     schwarze 2795:                free(buf->buf);
                   2796:                buf->buf = mandoc_strdup(".sp");
                   2797:                buf->sz = 4;
1.149     schwarze 2798:                return ROFF_REPARSE;
1.130     schwarze 2799:        }
1.193     schwarze 2800:        r->tbl = NULL;
1.149     schwarze 2801:        return ROFF_IGN;
1.27      schwarze 2802: }
                   2803:
                   2804: static enum rofferr
                   2805: roff_T_(ROFF_ARGS)
                   2806: {
                   2807:
                   2808:        if (NULL == r->tbl)
1.94      schwarze 2809:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2810:                    ln, ppos, "T&");
1.27      schwarze 2811:        else
1.168     schwarze 2812:                tbl_restart(ln, ppos, r->tbl);
1.27      schwarze 2813:
1.149     schwarze 2814:        return ROFF_IGN;
1.27      schwarze 2815: }
                   2816:
1.102     schwarze 2817: /*
                   2818:  * Handle in-line equation delimiters.
                   2819:  */
                   2820: static enum rofferr
1.110     schwarze 2821: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2822: {
1.105     schwarze 2823:        char            *cp1, *cp2;
                   2824:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2825:
1.102     schwarze 2826:        /*
                   2827:         * Outside equations, look for an opening delimiter.
                   2828:         * If we are inside an equation, we already know it is
                   2829:         * in-line, or this function wouldn't have been called;
                   2830:         * so look for a closing delimiter.
                   2831:         */
                   2832:
1.110     schwarze 2833:        cp1 = buf->buf + pos;
1.102     schwarze 2834:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2835:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2836:        if (cp2 == NULL)
1.149     schwarze 2837:                return ROFF_CONT;
1.102     schwarze 2838:
1.105     schwarze 2839:        *cp2++ = '\0';
                   2840:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2841:
                   2842:        /* Handle preceding text, protecting whitespace. */
                   2843:
1.110     schwarze 2844:        if (*buf->buf != '\0') {
1.105     schwarze 2845:                if (r->eqn == NULL)
                   2846:                        bef_pr = "\\&";
                   2847:                bef_nl = "\n";
                   2848:        }
                   2849:
                   2850:        /*
                   2851:         * Prepare replacing the delimiter with an equation macro
                   2852:         * and drop leading white space from the equation.
                   2853:         */
1.102     schwarze 2854:
1.105     schwarze 2855:        if (r->eqn == NULL) {
                   2856:                while (*cp2 == ' ')
                   2857:                        cp2++;
                   2858:                mac = ".EQ";
                   2859:        } else
                   2860:                mac = ".EN";
                   2861:
                   2862:        /* Handle following text, protecting whitespace. */
                   2863:
                   2864:        if (*cp2 != '\0') {
                   2865:                aft_nl = "\n";
                   2866:                if (r->eqn != NULL)
                   2867:                        aft_pr = "\\&";
                   2868:        }
                   2869:
                   2870:        /* Do the actual replacement. */
                   2871:
1.110     schwarze 2872:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2873:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2874:        free(buf->buf);
                   2875:        buf->buf = cp1;
1.102     schwarze 2876:
                   2877:        /* Toggle the in-line state of the eqn subsystem. */
                   2878:
                   2879:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2880:        return ROFF_REPARSE;
1.41      schwarze 2881: }
                   2882:
1.107     schwarze 2883: static enum rofferr
                   2884: roff_EQ(ROFF_ARGS)
1.32      schwarze 2885: {
1.191     schwarze 2886:        struct roff_node        *n;
                   2887:
1.194     schwarze 2888:        if (r->man->macroset == MACROSET_MAN)
                   2889:                man_breakscope(r->man, ROFF_EQ);
1.191     schwarze 2890:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   2891:        if (ln > r->man->last->line)
                   2892:                n->flags |= NODE_LINE;
                   2893:        n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
                   2894:        n->eqn->expectargs = UINT_MAX;
                   2895:        roff_node_append(r->man, n);
                   2896:        r->man->next = ROFF_NEXT_SIBLING;
1.32      schwarze 2897:
1.110     schwarze 2898:        assert(r->eqn == NULL);
1.191     schwarze 2899:        if (r->last_eqn == NULL)
                   2900:                r->last_eqn = eqn_alloc(r->parse);
                   2901:        else
                   2902:                eqn_reset(r->last_eqn);
                   2903:        r->eqn = r->last_eqn;
                   2904:        r->eqn->node = n;
1.41      schwarze 2905:
1.110     schwarze 2906:        if (buf->buf[pos] != '\0')
1.107     schwarze 2907:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2908:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2909:
1.149     schwarze 2910:        return ROFF_IGN;
1.32      schwarze 2911: }
                   2912:
                   2913: static enum rofferr
                   2914: roff_EN(ROFF_ARGS)
                   2915: {
1.191     schwarze 2916:        if (r->eqn != NULL) {
                   2917:                eqn_parse(r->eqn);
                   2918:                r->eqn = NULL;
                   2919:        } else
                   2920:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
                   2921:        if (buf->buf[pos] != '\0')
                   2922:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   2923:                    "EN %s", buf->buf + pos);
1.149     schwarze 2924:        return ROFF_IGN;
1.32      schwarze 2925: }
                   2926:
                   2927: static enum rofferr
1.27      schwarze 2928: roff_TS(ROFF_ARGS)
                   2929: {
1.193     schwarze 2930:        if (r->tbl != NULL) {
1.94      schwarze 2931:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2932:                    ln, ppos, "TS breaks TS");
1.193     schwarze 2933:                tbl_end(r->tbl);
1.27      schwarze 2934:        }
1.193     schwarze 2935:        r->tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2936:        if (r->last_tbl)
1.193     schwarze 2937:                r->last_tbl->next = r->tbl;
1.27      schwarze 2938:        else
1.193     schwarze 2939:                r->first_tbl = r->tbl;
                   2940:        r->last_tbl = r->tbl;
1.169     schwarze 2941:        return ROFF_IGN;
                   2942: }
                   2943:
                   2944: static enum rofferr
                   2945: roff_onearg(ROFF_ARGS)
                   2946: {
                   2947:        struct roff_node        *n;
                   2948:        char                    *cp;
1.177     schwarze 2949:        int                      npos;
1.169     schwarze 2950:
1.174     schwarze 2951:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.192     schwarze 2952:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
                   2953:             tok == ROFF_ti))
1.174     schwarze 2954:                man_breakscope(r->man, tok);
                   2955:
1.181     schwarze 2956:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177     schwarze 2957:                r->man->last = roffce_node;
                   2958:                r->man->next = ROFF_NEXT_SIBLING;
                   2959:        }
                   2960:
1.169     schwarze 2961:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2962:        n = r->man->last;
                   2963:
                   2964:        cp = buf->buf + pos;
                   2965:        if (*cp != '\0') {
                   2966:                while (*cp != '\0' && *cp != ' ')
                   2967:                        cp++;
                   2968:                while (*cp == ' ')
                   2969:                        *cp++ = '\0';
                   2970:                if (*cp != '\0')
                   2971:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS,
                   2972:                            r->parse, ln, cp - buf->buf,
                   2973:                            "%s ... %s", roff_name[tok], cp);
                   2974:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172     schwarze 2975:        }
                   2976:
1.181     schwarze 2977:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   2978:                if (r->man->last->type == ROFFT_ELEM) {
1.177     schwarze 2979:                        roff_word_alloc(r->man, ln, pos, "1");
                   2980:                        r->man->last->flags |= NODE_NOSRC;
                   2981:                }
                   2982:                npos = 0;
                   2983:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   2984:                    &roffce_lines, 0) == 0) {
                   2985:                        mandoc_vmsg(MANDOCERR_CE_NONUM,
                   2986:                            r->parse, ln, pos, "ce %s", buf->buf + pos);
                   2987:                        roffce_lines = 1;
                   2988:                }
                   2989:                if (roffce_lines < 1) {
                   2990:                        r->man->last = r->man->last->parent;
                   2991:                        roffce_node = NULL;
                   2992:                        roffce_lines = 0;
                   2993:                } else
                   2994:                        roffce_node = r->man->last->parent;
                   2995:        } else {
                   2996:                n->flags |= NODE_VALID | NODE_ENDED;
                   2997:                r->man->last = n;
                   2998:        }
                   2999:        n->flags |= NODE_LINE;
1.172     schwarze 3000:        r->man->next = ROFF_NEXT_SIBLING;
                   3001:        return ROFF_IGN;
                   3002: }
                   3003:
                   3004: static enum rofferr
                   3005: roff_manyarg(ROFF_ARGS)
                   3006: {
                   3007:        struct roff_node        *n;
                   3008:        char                    *sp, *ep;
                   3009:
                   3010:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3011:        n = r->man->last;
                   3012:
                   3013:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   3014:                while (*ep != '\0' && *ep != ' ')
                   3015:                        ep++;
                   3016:                while (*ep == ' ')
                   3017:                        *ep++ = '\0';
                   3018:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169     schwarze 3019:        }
                   3020:
                   3021:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3022:        r->man->last = n;
                   3023:        r->man->next = ROFF_NEXT_SIBLING;
1.149     schwarze 3024:        return ROFF_IGN;
1.123     schwarze 3025: }
                   3026:
                   3027: static enum rofferr
1.183     schwarze 3028: roff_als(ROFF_ARGS)
                   3029: {
                   3030:        char            *oldn, *newn, *end, *value;
                   3031:        size_t           oldsz, newsz, valsz;
                   3032:
                   3033:        newn = oldn = buf->buf + pos;
                   3034:        if (*newn == '\0')
                   3035:                return ROFF_IGN;
                   3036:
                   3037:        newsz = roff_getname(r, &oldn, ln, pos);
                   3038:        if (newn[newsz] == '\\' || *oldn == '\0')
                   3039:                return ROFF_IGN;
                   3040:
                   3041:        end = oldn;
                   3042:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3043:        if (oldsz == 0)
                   3044:                return ROFF_IGN;
                   3045:
1.187     schwarze 3046:        valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
                   3047:            (int)oldsz, oldn);
1.183     schwarze 3048:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3049:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3050:        free(value);
                   3051:        return ROFF_IGN;
                   3052: }
                   3053:
                   3054: static enum rofferr
1.168     schwarze 3055: roff_br(ROFF_ARGS)
1.123     schwarze 3056: {
1.174     schwarze 3057:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3058:                man_breakscope(r->man, ROFF_br);
1.168     schwarze 3059:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
                   3060:        if (buf->buf[pos] != '\0')
                   3061:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   3062:                    "%s %s", roff_name[tok], buf->buf + pos);
                   3063:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3064:        r->man->next = ROFF_NEXT_SIBLING;
                   3065:        return ROFF_IGN;
1.27      schwarze 3066: }
                   3067:
                   3068: static enum rofferr
1.48      schwarze 3069: roff_cc(ROFF_ARGS)
                   3070: {
                   3071:        const char      *p;
                   3072:
1.110     schwarze 3073:        p = buf->buf + pos;
1.48      schwarze 3074:
1.110     schwarze 3075:        if (*p == '\0' || (r->control = *p++) == '.')
1.175     schwarze 3076:                r->control = '\0';
1.48      schwarze 3077:
1.110     schwarze 3078:        if (*p != '\0')
1.132     schwarze 3079:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3080:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 3081:
1.149     schwarze 3082:        return ROFF_IGN;
1.48      schwarze 3083: }
                   3084:
                   3085: static enum rofferr
1.175     schwarze 3086: roff_ec(ROFF_ARGS)
                   3087: {
                   3088:        const char      *p;
                   3089:
                   3090:        p = buf->buf + pos;
                   3091:        if (*p == '\0')
                   3092:                r->escape = '\\';
                   3093:        else {
                   3094:                r->escape = *p;
                   3095:                if (*++p != '\0')
                   3096:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3097:                            ln, p - buf->buf, "ec ... %s", p);
                   3098:        }
                   3099:        return ROFF_IGN;
                   3100: }
                   3101:
                   3102: static enum rofferr
                   3103: roff_eo(ROFF_ARGS)
                   3104: {
                   3105:        r->escape = '\0';
                   3106:        if (buf->buf[pos] != '\0')
                   3107:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
                   3108:                    ln, pos, "eo %s", buf->buf + pos);
                   3109:        return ROFF_IGN;
                   3110: }
                   3111:
                   3112: static enum rofferr
1.42      schwarze 3113: roff_tr(ROFF_ARGS)
                   3114: {
                   3115:        const char      *p, *first, *second;
                   3116:        size_t           fsz, ssz;
                   3117:        enum mandoc_esc  esc;
                   3118:
1.110     schwarze 3119:        p = buf->buf + pos;
1.42      schwarze 3120:
1.110     schwarze 3121:        if (*p == '\0') {
1.132     schwarze 3122:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 3123:                return ROFF_IGN;
1.42      schwarze 3124:        }
                   3125:
1.110     schwarze 3126:        while (*p != '\0') {
1.42      schwarze 3127:                fsz = ssz = 1;
                   3128:
                   3129:                first = p++;
1.110     schwarze 3130:                if (*first == '\\') {
1.42      schwarze 3131:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3132:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3133:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3134:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 3135:                                return ROFF_IGN;
1.42      schwarze 3136:                        }
                   3137:                        fsz = (size_t)(p - first);
                   3138:                }
                   3139:
                   3140:                second = p++;
1.110     schwarze 3141:                if (*second == '\\') {
1.42      schwarze 3142:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3143:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3144:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3145:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 3146:                                return ROFF_IGN;
1.42      schwarze 3147:                        }
                   3148:                        ssz = (size_t)(p - second);
1.110     schwarze 3149:                } else if (*second == '\0') {
1.132     schwarze 3150:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   3151:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 3152:                        second = " ";
                   3153:                        p--;
                   3154:                }
                   3155:
                   3156:                if (fsz > 1) {
1.80      schwarze 3157:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3158:                            second, ssz, 0);
1.42      schwarze 3159:                        continue;
                   3160:                }
                   3161:
1.110     schwarze 3162:                if (r->xtab == NULL)
1.80      schwarze 3163:                        r->xtab = mandoc_calloc(128,
                   3164:                            sizeof(struct roffstr));
1.42      schwarze 3165:
                   3166:                free(r->xtab[(int)*first].p);
                   3167:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3168:                r->xtab[(int)*first].sz = ssz;
                   3169:        }
                   3170:
1.149     schwarze 3171:        return ROFF_IGN;
1.42      schwarze 3172: }
                   3173:
                   3174: static enum rofferr
1.178     schwarze 3175: roff_rn(ROFF_ARGS)
                   3176: {
                   3177:        const char      *value;
                   3178:        char            *oldn, *newn, *end;
                   3179:        size_t           oldsz, newsz;
1.187     schwarze 3180:        int              deftype;
1.178     schwarze 3181:
                   3182:        oldn = newn = buf->buf + pos;
                   3183:        if (*oldn == '\0')
                   3184:                return ROFF_IGN;
                   3185:
                   3186:        oldsz = roff_getname(r, &newn, ln, pos);
                   3187:        if (oldn[oldsz] == '\\' || *newn == '\0')
                   3188:                return ROFF_IGN;
                   3189:
                   3190:        end = newn;
                   3191:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3192:        if (newsz == 0)
                   3193:                return ROFF_IGN;
                   3194:
1.187     schwarze 3195:        deftype = ROFFDEF_ANY;
                   3196:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3197:        switch (deftype) {
                   3198:        case ROFFDEF_USER:
1.178     schwarze 3199:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3200:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3201:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.187     schwarze 3202:                break;
                   3203:        case ROFFDEF_PRE:
                   3204:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3205:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3206:                break;
                   3207:        case ROFFDEF_REN:
1.178     schwarze 3208:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3209:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.187     schwarze 3210:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3211:                break;
                   3212:        case ROFFDEF_STD:
1.178     schwarze 3213:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.187     schwarze 3214:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3215:                break;
                   3216:        default:
                   3217:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3218:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3219:                break;
                   3220:        }
1.178     schwarze 3221:        return ROFF_IGN;
                   3222: }
                   3223:
                   3224: static enum rofferr
1.14      schwarze 3225: roff_so(ROFF_ARGS)
                   3226: {
1.121     schwarze 3227:        char *name, *cp;
1.15      schwarze 3228:
1.110     schwarze 3229:        name = buf->buf + pos;
1.97      schwarze 3230:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 3231:
1.22      schwarze 3232:        /*
                   3233:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3234:         * opening anything that's not in our cwd or anything beneath
                   3235:         * it.  Thus, explicitly disallow traversing up the file-system
                   3236:         * or using absolute paths.
                   3237:         */
                   3238:
1.110     schwarze 3239:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 3240:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   3241:                    ".so %s", name);
1.121     schwarze 3242:                buf->sz = mandoc_asprintf(&cp,
                   3243:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3244:                free(buf->buf);
                   3245:                buf->buf = cp;
                   3246:                *offs = 0;
1.149     schwarze 3247:                return ROFF_REPARSE;
1.14      schwarze 3248:        }
                   3249:
                   3250:        *offs = pos;
1.149     schwarze 3251:        return ROFF_SO;
1.7       schwarze 3252: }
                   3253:
1.138     schwarze 3254: /* --- user defined strings and macros ------------------------------------ */
                   3255:
1.16      schwarze 3256: static enum rofferr
                   3257: roff_userdef(ROFF_ARGS)
1.12      schwarze 3258: {
1.183     schwarze 3259:        const char       *arg[16], *ap;
1.16      schwarze 3260:        char             *cp, *n1, *n2;
1.164     schwarze 3261:        int               expand_count, i, ib, ie;
1.135     schwarze 3262:        size_t            asz, rsz;
1.12      schwarze 3263:
1.16      schwarze 3264:        /*
                   3265:         * Collect pointers to macro argument strings
1.61      schwarze 3266:         * and NUL-terminate them.
1.16      schwarze 3267:         */
1.135     schwarze 3268:
1.145     schwarze 3269:        r->argc = 0;
1.110     schwarze 3270:        cp = buf->buf + pos;
1.183     schwarze 3271:        for (i = 0; i < 16; i++) {
1.145     schwarze 3272:                if (*cp == '\0')
                   3273:                        arg[i] = "";
                   3274:                else {
                   3275:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   3276:                        r->argc = i + 1;
                   3277:                }
                   3278:        }
1.16      schwarze 3279:
                   3280:        /*
                   3281:         * Expand macro arguments.
1.12      schwarze 3282:         */
1.135     schwarze 3283:
                   3284:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 3285:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 3286:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 3287:        expand_count = 0;
1.135     schwarze 3288:        while (*cp != '\0') {
                   3289:
                   3290:                /* Scan ahead for the next argument invocation. */
                   3291:
                   3292:                if (*cp++ != '\\')
                   3293:                        continue;
                   3294:                if (*cp++ != '$')
                   3295:                        continue;
1.146     schwarze 3296:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   3297:                        ib = 0;
                   3298:                        ie = r->argc - 1;
                   3299:                } else {  /* \\$1 .. \\$9 insert one argument */
                   3300:                        ib = ie = *cp - '1';
                   3301:                        if (ib < 0 || ib > 8)
                   3302:                                continue;
                   3303:                }
1.135     schwarze 3304:                cp -= 2;
1.164     schwarze 3305:
                   3306:                /*
                   3307:                 * Prevent infinite recursion.
                   3308:                 */
                   3309:
                   3310:                if (cp >= n2)
                   3311:                        expand_count = 1;
                   3312:                else if (++expand_count > EXPAND_LIMIT) {
                   3313:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   3314:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 3315:                        free(buf->buf);
                   3316:                        buf->buf = n1;
1.164     schwarze 3317:                        return ROFF_IGN;
                   3318:                }
1.135     schwarze 3319:
                   3320:                /*
                   3321:                 * Determine the size of the expanded argument,
                   3322:                 * taking escaping of quotes into account.
                   3323:                 */
                   3324:
1.146     schwarze 3325:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   3326:                for (i = ib; i <= ie; i++) {
                   3327:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3328:                                asz++;
                   3329:                                if (*ap == '"')
                   3330:                                        asz += 3;
                   3331:                        }
1.135     schwarze 3332:                }
                   3333:                if (asz != 3) {
                   3334:
                   3335:                        /*
                   3336:                         * Determine the size of the rest of the
                   3337:                         * unexpanded macro, including the NUL.
                   3338:                         */
                   3339:
                   3340:                        rsz = buf->sz - (cp - n1) - 3;
                   3341:
                   3342:                        /*
                   3343:                         * When shrinking, move before
                   3344:                         * releasing the storage.
                   3345:                         */
                   3346:
                   3347:                        if (asz < 3)
                   3348:                                memmove(cp + asz, cp + 3, rsz);
                   3349:
                   3350:                        /*
                   3351:                         * Resize the storage for the macro
                   3352:                         * and readjust the parse pointer.
                   3353:                         */
                   3354:
                   3355:                        buf->sz += asz - 3;
                   3356:                        n2 = mandoc_realloc(n1, buf->sz);
                   3357:                        cp = n2 + (cp - n1);
                   3358:                        n1 = n2;
                   3359:
                   3360:                        /*
                   3361:                         * When growing, make room
                   3362:                         * for the expanded argument.
                   3363:                         */
                   3364:
                   3365:                        if (asz > 3)
                   3366:                                memmove(cp + asz, cp + 3, rsz);
                   3367:                }
                   3368:
                   3369:                /* Copy the expanded argument, escaping quotes. */
                   3370:
                   3371:                n2 = cp;
1.146     schwarze 3372:                for (i = ib; i <= ie; i++) {
                   3373:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3374:                                if (*ap == '"') {
                   3375:                                        memcpy(n2, "\\(dq", 4);
                   3376:                                        n2 += 4;
                   3377:                                } else
                   3378:                                        *n2++ = *ap;
                   3379:                        }
                   3380:                        if (i < ie)
                   3381:                                *n2++ = ' ';
1.16      schwarze 3382:                }
1.12      schwarze 3383:        }
                   3384:
1.16      schwarze 3385:        /*
                   3386:         * Replace the macro invocation
                   3387:         * by the expanded macro.
                   3388:         */
1.135     schwarze 3389:
1.110     schwarze 3390:        free(buf->buf);
                   3391:        buf->buf = n1;
1.120     schwarze 3392:        *offs = 0;
1.16      schwarze 3393:
1.149     schwarze 3394:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3395:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3396: }
1.28      schwarze 3397:
1.178     schwarze 3398: /*
                   3399:  * Calling a high-level macro that was renamed with .rn.
                   3400:  * r->current_string has already been set up by roff_parse().
                   3401:  */
                   3402: static enum rofferr
                   3403: roff_renamed(ROFF_ARGS)
                   3404: {
                   3405:        char    *nbuf;
                   3406:
1.187     schwarze 3407:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3408:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.178     schwarze 3409:        free(buf->buf);
                   3410:        buf->buf = nbuf;
                   3411:        return ROFF_CONT;
                   3412: }
                   3413:
1.85      schwarze 3414: static size_t
1.28      schwarze 3415: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3416: {
                   3417:        char     *name, *cp;
1.85      schwarze 3418:        size_t    namesz;
1.28      schwarze 3419:
                   3420:        name = *cpp;
                   3421:        if ('\0' == *name)
1.149     schwarze 3422:                return 0;
1.28      schwarze 3423:
1.85      schwarze 3424:        /* Read until end of name and terminate it with NUL. */
                   3425:        for (cp = name; 1; cp++) {
                   3426:                if ('\0' == *cp || ' ' == *cp) {
                   3427:                        namesz = cp - name;
                   3428:                        break;
                   3429:                }
1.28      schwarze 3430:                if ('\\' != *cp)
                   3431:                        continue;
1.88      schwarze 3432:                namesz = cp - name;
                   3433:                if ('{' == cp[1] || '}' == cp[1])
                   3434:                        break;
1.28      schwarze 3435:                cp++;
                   3436:                if ('\\' == *cp)
                   3437:                        continue;
1.97      schwarze 3438:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3439:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3440:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3441:                break;
1.28      schwarze 3442:        }
                   3443:
                   3444:        /* Read past spaces. */
                   3445:        while (' ' == *cp)
                   3446:                cp++;
                   3447:
                   3448:        *cpp = cp;
1.149     schwarze 3449:        return namesz;
1.28      schwarze 3450: }
                   3451:
1.16      schwarze 3452: /*
                   3453:  * Store *string into the user-defined string called *name.
                   3454:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3455:  * append == 0: replace mode
                   3456:  * append == 1: single-line append mode
                   3457:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3458:  */
1.8       schwarze 3459: static void
1.16      schwarze 3460: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3461:        int append)
1.7       schwarze 3462: {
1.187     schwarze 3463:        size_t   namesz;
1.42      schwarze 3464:
1.187     schwarze 3465:        namesz = strlen(name);
                   3466:        roff_setstrn(&r->strtab, name, namesz, string,
1.80      schwarze 3467:            string ? strlen(string) : 0, append);
1.187     schwarze 3468:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.42      schwarze 3469: }
                   3470:
                   3471: static void
                   3472: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3473:                const char *string, size_t stringsz, int append)
1.42      schwarze 3474: {
                   3475:        struct roffkv   *n;
                   3476:        char            *c;
                   3477:        int              i;
                   3478:        size_t           oldch, newch;
1.7       schwarze 3479:
1.16      schwarze 3480:        /* Search for an existing string with the same name. */
1.42      schwarze 3481:        n = *r;
                   3482:
1.84      schwarze 3483:        while (n && (namesz != n->key.sz ||
                   3484:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3485:                n = n->next;
1.8       schwarze 3486:
                   3487:        if (NULL == n) {
1.16      schwarze 3488:                /* Create a new string table entry. */
1.42      schwarze 3489:                n = mandoc_malloc(sizeof(struct roffkv));
                   3490:                n->key.p = mandoc_strndup(name, namesz);
                   3491:                n->key.sz = namesz;
                   3492:                n->val.p = NULL;
                   3493:                n->val.sz = 0;
                   3494:                n->next = *r;
                   3495:                *r = n;
1.66      schwarze 3496:        } else if (0 == append) {
1.42      schwarze 3497:                free(n->val.p);
                   3498:                n->val.p = NULL;
                   3499:                n->val.sz = 0;
1.16      schwarze 3500:        }
                   3501:
                   3502:        if (NULL == string)
                   3503:                return;
                   3504:
                   3505:        /*
                   3506:         * One additional byte for the '\n' in multiline mode,
                   3507:         * and one for the terminating '\0'.
                   3508:         */
1.66      schwarze 3509:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3510:
                   3511:        if (NULL == n->val.p) {
                   3512:                n->val.p = mandoc_malloc(newch);
                   3513:                *n->val.p = '\0';
1.16      schwarze 3514:                oldch = 0;
                   3515:        } else {
1.42      schwarze 3516:                oldch = n->val.sz;
                   3517:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3518:        }
                   3519:
                   3520:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3521:        c = n->val.p + (int)oldch;
1.16      schwarze 3522:
                   3523:        /* Append new content to the destination buffer. */
1.42      schwarze 3524:        i = 0;
                   3525:        while (i < (int)stringsz) {
1.16      schwarze 3526:                /*
                   3527:                 * Rudimentary roff copy mode:
                   3528:                 * Handle escaped backslashes.
                   3529:                 */
1.42      schwarze 3530:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3531:                        i++;
                   3532:                *c++ = string[i++];
1.16      schwarze 3533:        }
1.8       schwarze 3534:
1.16      schwarze 3535:        /* Append terminating bytes. */
1.66      schwarze 3536:        if (1 < append)
1.16      schwarze 3537:                *c++ = '\n';
1.42      schwarze 3538:
1.16      schwarze 3539:        *c = '\0';
1.42      schwarze 3540:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3541: }
                   3542:
1.8       schwarze 3543: static const char *
1.197   ! schwarze 3544: roff_getstrn(struct roff *r, const char *name, size_t len,
1.187     schwarze 3545:     int *deftype)
1.7       schwarze 3546: {
1.187     schwarze 3547:        const struct roffkv     *n;
1.197   ! schwarze 3548:        int                      found, i;
1.187     schwarze 3549:        enum roff_tok            tok;
                   3550:
1.197   ! schwarze 3551:        found = 0;
        !          3552:        for (n = r->strtab; n != NULL; n = n->next) {
        !          3553:                if (strncmp(name, n->key.p, len) != 0 ||
        !          3554:                    n->key.p[len] != '\0' || n->val.p == NULL)
        !          3555:                        continue;
        !          3556:                if (*deftype & ROFFDEF_USER) {
        !          3557:                        *deftype = ROFFDEF_USER;
        !          3558:                        return n->val.p;
        !          3559:                } else {
        !          3560:                        found = 1;
        !          3561:                        break;
        !          3562:                }
        !          3563:        }
        !          3564:        for (n = r->rentab; n != NULL; n = n->next) {
        !          3565:                if (strncmp(name, n->key.p, len) != 0 ||
        !          3566:                    n->key.p[len] != '\0' || n->val.p == NULL)
        !          3567:                        continue;
        !          3568:                if (*deftype & ROFFDEF_REN) {
        !          3569:                        *deftype = ROFFDEF_REN;
        !          3570:                        return n->val.p;
        !          3571:                } else {
        !          3572:                        found = 1;
        !          3573:                        break;
1.187     schwarze 3574:                }
                   3575:        }
1.197   ! schwarze 3576:        for (i = 0; i < PREDEFS_MAX; i++) {
        !          3577:                if (strncmp(name, predefs[i].name, len) != 0 ||
        !          3578:                    predefs[i].name[len] != '\0')
        !          3579:                        continue;
        !          3580:                if (*deftype & ROFFDEF_PRE) {
        !          3581:                        *deftype = ROFFDEF_PRE;
        !          3582:                        return predefs[i].str;
        !          3583:                } else {
        !          3584:                        found = 1;
        !          3585:                        break;
1.187     schwarze 3586:                }
                   3587:        }
1.197   ! schwarze 3588:        if (r->man->macroset != MACROSET_MAN) {
        !          3589:                for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
        !          3590:                        if (strncmp(name, roff_name[tok], len) != 0 ||
        !          3591:                            roff_name[tok][len] != '\0')
        !          3592:                                continue;
        !          3593:                        if (*deftype & ROFFDEF_STD) {
        !          3594:                                *deftype = ROFFDEF_STD;
        !          3595:                                return NULL;
        !          3596:                        } else {
        !          3597:                                found = 1;
        !          3598:                                break;
1.187     schwarze 3599:                        }
                   3600:                }
                   3601:        }
1.197   ! schwarze 3602:        if (r->man->macroset != MACROSET_MDOC) {
        !          3603:                for (tok = MAN_TH; tok < MAN_MAX; tok++) {
        !          3604:                        if (strncmp(name, roff_name[tok], len) != 0 ||
        !          3605:                            roff_name[tok][len] != '\0')
        !          3606:                                continue;
        !          3607:                        if (*deftype & ROFFDEF_STD) {
        !          3608:                                *deftype = ROFFDEF_STD;
        !          3609:                                return NULL;
        !          3610:                        } else {
        !          3611:                                found = 1;
        !          3612:                                break;
1.187     schwarze 3613:                        }
                   3614:                }
1.197   ! schwarze 3615:        }
        !          3616:
        !          3617:        if (found == 0 && *deftype != ROFFDEF_ANY) {
        !          3618:                if (*deftype & ROFFDEF_REN) {
        !          3619:                        /*
        !          3620:                         * This might still be a request,
        !          3621:                         * so do not treat it as undefined yet.
        !          3622:                         */
        !          3623:                        *deftype = ROFFDEF_UNDEF;
        !          3624:                        return NULL;
1.187     schwarze 3625:                }
1.197   ! schwarze 3626:
        !          3627:                /* Using an undefined string defines it to be empty. */
        !          3628:
        !          3629:                roff_setstrn(&r->strtab, name, len, "", 0, 0);
        !          3630:                roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.187     schwarze 3631:        }
1.197   ! schwarze 3632:
1.187     schwarze 3633:        *deftype = 0;
1.149     schwarze 3634:        return NULL;
1.7       schwarze 3635: }
                   3636:
1.8       schwarze 3637: static void
1.42      schwarze 3638: roff_freestr(struct roffkv *r)
1.7       schwarze 3639: {
1.42      schwarze 3640:        struct roffkv    *n, *nn;
1.7       schwarze 3641:
1.42      schwarze 3642:        for (n = r; n; n = nn) {
                   3643:                free(n->key.p);
                   3644:                free(n->val.p);
1.7       schwarze 3645:                nn = n->next;
                   3646:                free(n);
                   3647:        }
1.27      schwarze 3648: }
1.138     schwarze 3649:
                   3650: /* --- accessors and utility functions ------------------------------------ */
1.42      schwarze 3651:
                   3652: /*
                   3653:  * Duplicate an input string, making the appropriate character
                   3654:  * conversations (as stipulated by `tr') along the way.
                   3655:  * Returns a heap-allocated string with all the replacements made.
                   3656:  */
                   3657: char *
                   3658: roff_strdup(const struct roff *r, const char *p)
                   3659: {
                   3660:        const struct roffkv *cp;
                   3661:        char            *res;
                   3662:        const char      *pp;
                   3663:        size_t           ssz, sz;
                   3664:        enum mandoc_esc  esc;
                   3665:
                   3666:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3667:                return mandoc_strdup(p);
1.42      schwarze 3668:        else if ('\0' == *p)
1.149     schwarze 3669:                return mandoc_strdup("");
1.42      schwarze 3670:
                   3671:        /*
                   3672:         * Step through each character looking for term matches
                   3673:         * (remember that a `tr' can be invoked with an escape, which is
                   3674:         * a glyph but the escape is multi-character).
                   3675:         * We only do this if the character hash has been initialised
                   3676:         * and the string is >0 length.
                   3677:         */
                   3678:
                   3679:        res = NULL;
                   3680:        ssz = 0;
                   3681:
                   3682:        while ('\0' != *p) {
1.161     schwarze 3683:                assert((unsigned int)*p < 128);
                   3684:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3685:                        sz = r->xtab[(int)*p].sz;
                   3686:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3687:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3688:                        ssz += sz;
                   3689:                        p++;
                   3690:                        continue;
                   3691:                } else if ('\\' != *p) {
                   3692:                        res = mandoc_realloc(res, ssz + 2);
                   3693:                        res[ssz++] = *p++;
                   3694:                        continue;
                   3695:                }
                   3696:
                   3697:                /* Search for term matches. */
                   3698:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3699:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3700:                                break;
                   3701:
                   3702:                if (NULL != cp) {
                   3703:                        /*
                   3704:                         * A match has been found.
                   3705:                         * Append the match to the array and move
                   3706:                         * forward by its keysize.
                   3707:                         */
1.80      schwarze 3708:                        res = mandoc_realloc(res,
                   3709:                            ssz + cp->val.sz + 1);
1.42      schwarze 3710:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3711:                        ssz += cp->val.sz;
                   3712:                        p += (int)cp->key.sz;
                   3713:                        continue;
                   3714:                }
                   3715:
                   3716:                /*
                   3717:                 * Handle escapes carefully: we need to copy
                   3718:                 * over just the escape itself, or else we might
                   3719:                 * do replacements within the escape itself.
                   3720:                 * Make sure to pass along the bogus string.
                   3721:                 */
                   3722:                pp = p++;
                   3723:                esc = mandoc_escape(&p, NULL, NULL);
                   3724:                if (ESCAPE_ERROR == esc) {
                   3725:                        sz = strlen(pp);
                   3726:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3727:                        memcpy(res + ssz, pp, sz);
                   3728:                        break;
                   3729:                }
1.80      schwarze 3730:                /*
                   3731:                 * We bail out on bad escapes.
1.42      schwarze 3732:                 * No need to warn: we already did so when
                   3733:                 * roff_res() was called.
                   3734:                 */
                   3735:                sz = (int)(p - pp);
                   3736:                res = mandoc_realloc(res, ssz + sz + 1);
                   3737:                memcpy(res + ssz, pp, sz);
                   3738:                ssz += sz;
                   3739:        }
                   3740:
                   3741:        res[(int)ssz] = '\0';
1.149     schwarze 3742:        return res;
1.99      schwarze 3743: }
                   3744:
                   3745: int
                   3746: roff_getformat(const struct roff *r)
                   3747: {
                   3748:
1.149     schwarze 3749:        return r->format;
1.48      schwarze 3750: }
                   3751:
                   3752: /*
1.80      schwarze 3753:  * Find out whether a line is a macro line or not.
1.48      schwarze 3754:  * If it is, adjust the current position and return one; if it isn't,
                   3755:  * return zero and don't change the current position.
                   3756:  * If the control character has been set with `.cc', then let that grain
                   3757:  * precedence.
                   3758:  * This is slighly contrary to groff, where using the non-breaking
                   3759:  * control character when `cc' has been invoked will cause the
                   3760:  * non-breaking macro contents to be printed verbatim.
                   3761:  */
                   3762: int
                   3763: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3764: {
                   3765:        int             pos;
                   3766:
                   3767:        pos = *ppos;
                   3768:
1.175     schwarze 3769:        if (r->control != '\0' && cp[pos] == r->control)
1.48      schwarze 3770:                pos++;
1.175     schwarze 3771:        else if (r->control != '\0')
1.149     schwarze 3772:                return 0;
1.48      schwarze 3773:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3774:                pos += 2;
                   3775:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3776:                pos++;
                   3777:        else
1.149     schwarze 3778:                return 0;
1.48      schwarze 3779:
                   3780:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3781:                pos++;
                   3782:
                   3783:        *ppos = pos;
1.149     schwarze 3784:        return 1;
1.1       schwarze 3785: }