[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 1.200

1.200   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.199 2018/04/10 00:52:21 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;
1.199     schwarze   74:        int              step;
1.52      schwarze   75:        struct roffreg  *next;
                     76: };
                     77:
1.167     schwarze   78: /*
                     79:  * Association of request and macro names with token IDs.
                     80:  */
                     81: struct roffreq {
                     82:        enum roff_tok    tok;
                     83:        char             name[];
                     84: };
                     85:
1.1       schwarze   86: struct roff {
1.35      schwarze   87:        struct mparse   *parse; /* parse point */
1.168     schwarze   88:        struct roff_man *man; /* mdoc or man parser */
1.1       schwarze   89:        struct roffnode *last; /* leaf of stack */
1.96      schwarze   90:        int             *rstack; /* stack of inverted `ie' values */
1.167     schwarze   91:        struct ohash    *reqtab; /* request lookup table */
1.52      schwarze   92:        struct roffreg  *regtab; /* number registers */
1.42      schwarze   93:        struct roffkv   *strtab; /* user-defined strings & macros */
1.178     schwarze   94:        struct roffkv   *rentab; /* renamed strings & macros */
1.42      schwarze   95:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     96:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze   97:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   98:        struct tbl_node *first_tbl; /* first table parsed */
                     99:        struct tbl_node *last_tbl; /* last table parsed */
                    100:        struct tbl_node *tbl; /* current table being parsed */
1.191     schwarze  101:        struct eqn_node *last_eqn; /* equation parser */
                    102:        struct eqn_node *eqn; /* active equation parser */
1.102     schwarze  103:        int              eqn_inline; /* current equation is inline */
1.96      schwarze  104:        int              options; /* parse options */
                    105:        int              rstacksz; /* current size limit of rstack */
                    106:        int              rstackpos; /* position in rstack */
1.99      schwarze  107:        int              format; /* current file in mdoc or man format */
1.145     schwarze  108:        int              argc; /* number of args of the last macro */
1.96      schwarze  109:        char             control; /* control character */
1.175     schwarze  110:        char             escape; /* escape character */
1.1       schwarze  111: };
                    112:
                    113: struct roffnode {
1.166     schwarze  114:        enum roff_tok    tok; /* type of node */
1.1       schwarze  115:        struct roffnode *parent; /* up one in stack */
                    116:        int              line; /* parse line */
                    117:        int              col; /* parse col */
1.16      schwarze  118:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  119:        char            *end; /* end-rules: custom token */
                    120:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  121:        int              rule; /* current evaluation rule */
1.1       schwarze  122: };
                    123:
                    124: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.166     schwarze  125:                         enum roff_tok tok, /* tok of macro */ \
1.110     schwarze  126:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  127:                         int ln, /* parse line */ \
1.2       schwarze  128:                         int ppos, /* original pos in buffer */ \
                    129:                         int pos, /* current pos in buffer */ \
                    130:                         int *offs /* reset offset of buffer data */
1.1       schwarze  131:
                    132: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    133:
                    134: struct roffmac {
1.2       schwarze  135:        roffproc         proc; /* process new macro */
                    136:        roffproc         text; /* process as child text of macro */
                    137:        roffproc         sub; /* process as child of macro */
                    138:        int              flags;
                    139: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.1       schwarze  140: };
                    141:
1.37      schwarze  142: struct predef {
                    143:        const char      *name; /* predefined input name */
                    144:        const char      *str; /* replacement symbol */
                    145: };
                    146:
                    147: #define        PREDEF(__name, __str) \
                    148:        { (__name), (__str) },
                    149:
1.138     schwarze  150: /* --- function prototypes ------------------------------------------------ */
                    151:
1.42      schwarze  152: static void             roffnode_cleanscope(struct roff *);
                    153: static void             roffnode_pop(struct roff *);
1.166     schwarze  154: static void             roffnode_push(struct roff *, enum roff_tok,
1.42      schwarze  155:                                const char *, int, int);
1.193     schwarze  156: static void             roff_addtbl(struct roff_man *, struct tbl_node *);
1.183     schwarze  157: static enum rofferr     roff_als(ROFF_ARGS);
1.2       schwarze  158: static enum rofferr     roff_block(ROFF_ARGS);
                    159: static enum rofferr     roff_block_text(ROFF_ARGS);
                    160: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.168     schwarze  161: static enum rofferr     roff_br(ROFF_ARGS);
1.2       schwarze  162: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  163: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  164: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  165: static enum rofferr     roff_cond(ROFF_ARGS);
                    166: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    167: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  168: static enum rofferr     roff_ds(ROFF_ARGS);
1.175     schwarze  169: static enum rofferr     roff_ec(ROFF_ARGS);
                    170: static enum rofferr     roff_eo(ROFF_ARGS);
1.110     schwarze  171: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.143     schwarze  172: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.106     schwarze  173: static int              roff_evalnum(struct roff *, int,
                    174:                                const char *, int *, int *, int);
                    175: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  176:                                const char *, int *, int *, int);
1.71      schwarze  177: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  178: static void             roff_free1(struct roff *);
1.52      schwarze  179: static void             roff_freereg(struct roffreg *);
1.42      schwarze  180: static void             roff_freestr(struct roffkv *);
1.85      schwarze  181: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  182: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  183: static int              roff_getop(const char *, int *, char *);
1.199     schwarze  184: static int              roff_getregn(struct roff *,
                    185:                                const char *, size_t, char);
1.145     schwarze  186: static int              roff_getregro(const struct roff *,
                    187:                                const char *name);
1.197     schwarze  188: static const char      *roff_getstrn(struct roff *,
1.187     schwarze  189:                                const char *, size_t, int *);
1.143     schwarze  190: static int              roff_hasregn(const struct roff *,
                    191:                                const char *, size_t);
1.123     schwarze  192: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  193: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  194: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137     schwarze  195: static void             roff_man_alloc1(struct roff_man *);
                    196: static void             roff_man_free1(struct roff_man *);
1.172     schwarze  197: static enum rofferr     roff_manyarg(ROFF_ARGS);
1.6       schwarze  198: static enum rofferr     roff_nr(ROFF_ARGS);
1.169     schwarze  199: static enum rofferr     roff_onearg(ROFF_ARGS);
1.166     schwarze  200: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.87      schwarze  201:                                int, int);
1.177     schwarze  202: static enum rofferr     roff_parsetext(struct roff *, struct buf *,
                    203:                                int, int *);
1.178     schwarze  204: static enum rofferr     roff_renamed(ROFF_ARGS);
1.110     schwarze  205: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  206: static enum rofferr     roff_rm(ROFF_ARGS);
1.178     schwarze  207: static enum rofferr     roff_rn(ROFF_ARGS);
1.76      schwarze  208: static enum rofferr     roff_rr(ROFF_ARGS);
1.198     schwarze  209: static void             roff_setregn(struct roff *, const char *,
1.199     schwarze  210:                                size_t, int, char, int);
1.8       schwarze  211: static void             roff_setstr(struct roff *,
1.16      schwarze  212:                                const char *, const char *, int);
1.80      schwarze  213: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  214:                                size_t, const char *, size_t, int);
1.14      schwarze  215: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  216: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  217: static enum rofferr     roff_Dd(ROFF_ARGS);
1.27      schwarze  218: static enum rofferr     roff_TE(ROFF_ARGS);
                    219: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  220: static enum rofferr     roff_EQ(ROFF_ARGS);
                    221: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  222: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  223: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  224: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  225:
1.138     schwarze  226: /* --- constant data ------------------------------------------------------ */
                    227:
1.133     schwarze  228: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    229: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    230:
1.166     schwarze  231: const char *__roff_name[MAN_MAX + 1] = {
1.177     schwarze  232:        "br",           "ce",           "ft",           "ll",
1.184     schwarze  233:        "mc",           "po",           "rj",           "sp",
                    234:        "ta",           "ti",           NULL,
1.166     schwarze  235:        "ab",           "ad",           "af",           "aln",
                    236:        "als",          "am",           "am1",          "ami",
                    237:        "ami1",         "as",           "as1",          "asciify",
                    238:        "backtrace",    "bd",           "bleedat",      "blm",
                    239:         "box",         "boxa",         "bp",           "BP",
                    240:        "break",        "breakchar",    "brnl",         "brp",
1.177     schwarze  241:        "brpnl",        "c2",           "cc",
1.166     schwarze  242:        "cf",           "cflags",       "ch",           "char",
                    243:        "chop",         "class",        "close",        "CL",
                    244:        "color",        "composite",    "continue",     "cp",
                    245:        "cropat",       "cs",           "cu",           "da",
                    246:        "dch",          "Dd",           "de",           "de1",
                    247:        "defcolor",     "dei",          "dei1",         "device",
                    248:        "devicem",      "di",           "do",           "ds",
                    249:        "ds1",          "dwh",          "dt",           "ec",
                    250:        "ecr",          "ecs",          "el",           "em",
                    251:        "EN",           "eo",           "EP",           "EQ",
                    252:        "errprint",     "ev",           "evc",          "ex",
                    253:        "fallback",     "fam",          "fc",           "fchar",
                    254:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    255:        "fl",           "flig",         "fp",           "fps",
                    256:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    257:        "fzoom",        "gcolor",       "hc",           "hcode",
                    258:        "hidechar",     "hla",          "hlm",          "hpf",
                    259:        "hpfa",         "hpfcode",      "hw",           "hy",
                    260:        "hylang",       "hylen",        "hym",          "hypp",
                    261:        "hys",          "ie",           "if",           "ig",
                    262:        "index",        "it",           "itc",          "IX",
                    263:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    264:        "lc",           "lc_ctype",     "lds",          "length",
                    265:        "letadj",       "lf",           "lg",           "lhang",
                    266:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.176     schwarze  267:        "ls",           "lsm",          "lt",
1.166     schwarze  268:        "mediasize",    "minss",        "mk",           "mso",
                    269:        "na",           "ne",           "nh",           "nhychar",
                    270:        "nm",           "nn",           "nop",          "nr",
                    271:        "nrf",          "nroff",        "ns",           "nx",
                    272:        "open",         "opena",        "os",           "output",
                    273:        "padj",         "papersize",    "pc",           "pev",
                    274:        "pi",           "PI",           "pl",           "pm",
1.184     schwarze  275:        "pn",           "pnr",          "ps",
1.166     schwarze  276:        "psbb",         "pshape",       "pso",          "ptr",
                    277:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.181     schwarze  278:        "return",       "rfschar",      "rhang",
1.166     schwarze  279:        "rm",           "rn",           "rnn",          "rr",
                    280:        "rs",           "rt",           "schar",        "sentchar",
                    281:        "shc",          "shift",        "sizes",        "so",
                    282:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    283:        "sty",          "substring",    "sv",           "sy",
1.172     schwarze  284:        "T&",           "tc",           "TE",
1.173     schwarze  285:        "TH",           "tkf",          "tl",
1.166     schwarze  286:        "tm",           "tm1",          "tmc",          "tr",
                    287:        "track",        "transchar",    "trf",          "trimat",
                    288:        "trin",         "trnt",         "troff",        "TS",
                    289:        "uf",           "ul",           "unformat",     "unwatch",
                    290:        "unwatchn",     "vpt",          "vs",           "warn",
                    291:        "warnscale",    "watch",        "watchlength",  "watchn",
                    292:        "wh",           "while",        "write",        "writec",
                    293:        "writem",       "xflag",        ".",            NULL,
1.178     schwarze  294:        NULL,           "text",
1.166     schwarze  295:        "Dd",           "Dt",           "Os",           "Sh",
                    296:        "Ss",           "Pp",           "D1",           "Dl",
                    297:        "Bd",           "Ed",           "Bl",           "El",
                    298:        "It",           "Ad",           "An",           "Ap",
                    299:        "Ar",           "Cd",           "Cm",           "Dv",
                    300:        "Er",           "Ev",           "Ex",           "Fa",
                    301:        "Fd",           "Fl",           "Fn",           "Ft",
                    302:        "Ic",           "In",           "Li",           "Nd",
                    303:        "Nm",           "Op",           "Ot",           "Pa",
                    304:        "Rv",           "St",           "Va",           "Vt",
                    305:        "Xr",           "%A",           "%B",           "%D",
                    306:        "%I",           "%J",           "%N",           "%O",
                    307:        "%P",           "%R",           "%T",           "%V",
                    308:        "Ac",           "Ao",           "Aq",           "At",
                    309:        "Bc",           "Bf",           "Bo",           "Bq",
                    310:        "Bsx",          "Bx",           "Db",           "Dc",
                    311:        "Do",           "Dq",           "Ec",           "Ef",
                    312:        "Em",           "Eo",           "Fx",           "Ms",
                    313:        "No",           "Ns",           "Nx",           "Ox",
                    314:        "Pc",           "Pf",           "Po",           "Pq",
                    315:        "Qc",           "Ql",           "Qo",           "Qq",
                    316:        "Re",           "Rs",           "Sc",           "So",
                    317:        "Sq",           "Sm",           "Sx",           "Sy",
                    318:        "Tn",           "Ux",           "Xc",           "Xo",
                    319:        "Fo",           "Fc",           "Oo",           "Oc",
                    320:        "Bk",           "Ek",           "Bt",           "Hf",
                    321:        "Fr",           "Ud",           "Lb",           "Lp",
                    322:        "Lk",           "Mt",           "Brq",          "Bro",
                    323:        "Brc",          "%C",           "Es",           "En",
1.171     schwarze  324:        "Dx",           "%Q",           "%U",           "Ta",
                    325:        NULL,
1.166     schwarze  326:        "TH",           "SH",           "SS",           "TP",
                    327:        "LP",           "PP",           "P",            "IP",
                    328:        "HP",           "SM",           "SB",           "BI",
                    329:        "IB",           "BR",           "RB",           "R",
                    330:        "B",            "I",            "IR",           "RI",
1.171     schwarze  331:        "nf",           "fi",
1.166     schwarze  332:        "RE",           "RS",           "DT",           "UC",
1.169     schwarze  333:        "PD",           "AT",           "in",
1.166     schwarze  334:        "OP",           "EX",           "EE",           "UR",
1.189     bentley   335:        "UE",           "MT",           "ME",           NULL
1.166     schwarze  336: };
                    337: const  char *const *roff_name = __roff_name;
                    338:
                    339: static struct roffmac   roffs[TOKEN_NONE] = {
1.168     schwarze  340:        { roff_br, NULL, NULL, 0 },  /* br */
1.177     schwarze  341:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.169     schwarze  342:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.170     schwarze  343:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.176     schwarze  344:        { roff_onearg, NULL, NULL, 0 },  /* mc */
1.184     schwarze  345:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.181     schwarze  346:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.171     schwarze  347:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.172     schwarze  348:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.173     schwarze  349:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.168     schwarze  350:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.167     schwarze  351:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    352:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    353:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    354:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.183     schwarze  355:        { roff_als, NULL, NULL, 0 },  /* als */
1.167     schwarze  356:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    357:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    358:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    359:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    360:        { roff_ds, NULL, NULL, 0 },  /* as */
                    361:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    362:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    363:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    364:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    365:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    366:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    367:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    368:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    369:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    370:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    371:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    372:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    373:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.168     schwarze  374:        { roff_br, NULL, NULL, 0 },  /* brp */
1.167     schwarze  375:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    376:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    377:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    378:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    379:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    380:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    381:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    382:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    383:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    384:        { roff_insec, NULL, NULL, 0 },  /* close */
                    385:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    386:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    387:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    388:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    389:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    390:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    391:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    392:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    393:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    394:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    395:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    396:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    397:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    398:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    399:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    400:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    401:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    402:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    403:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    404:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    405:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    406:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    407:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    408:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.175     schwarze  409:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.167     schwarze  410:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    411:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    412:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    413:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    414:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.175     schwarze  415:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.167     schwarze  416:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    417:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    419:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    420:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    421:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    422:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    424:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    425:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    429:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    430:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    431:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    432:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    434:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    438:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    439:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    440:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    442:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    444:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    445:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    446:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    447:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    448:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    449:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    450:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    451:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    452:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    453:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    454:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    455:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    456:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    457:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    458:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    459:        { roff_it, NULL, NULL, 0 },  /* it */
                    460:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    461:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    462:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    463:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    464:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    465:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    466:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    467:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    468:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    469:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    470:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    471:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    472:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    474:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    475:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    476:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    477:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    479:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    484:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    487:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    488:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    489:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    490:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    491:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    492:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    493:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    494:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    495:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    496:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    497:        { roff_insec, NULL, NULL, 0 },  /* open */
                    498:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    500:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    501:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    503:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    504:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    505:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    506:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    508:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    511:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    512:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    513:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    514:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    516:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    517:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    518:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    519:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    520:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    521:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    522:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    523:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.178     schwarze  524:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.167     schwarze  525:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    526:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    529:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    530:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    532:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    533:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    534:        { roff_so, NULL, NULL, 0 },  /* so */
                    535:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    536:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    537:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    538:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    539:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    540:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    541:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    542:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    543:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    544:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    545:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.187     schwarze  546:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.167     schwarze  547:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    548:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    549:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    550:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    551:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    552:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    553:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    554:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    555:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    557:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    558:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    559:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    560:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    561:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    562:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    563:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    565:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    566:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    568:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    569:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    570:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    571:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    572:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    573:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    574:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    575:        { roff_insec, NULL, NULL, 0 },  /* write */
                    576:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    577:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    578:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    579:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.178     schwarze  580:        { roff_renamed, NULL, NULL, 0 },
1.167     schwarze  581:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  582: };
                    583:
1.37      schwarze  584: /* Array of injected predefined strings. */
                    585: #define        PREDEFS_MAX      38
                    586: static const struct predef predefs[PREDEFS_MAX] = {
                    587: #include "predefs.in"
                    588: };
                    589:
1.177     schwarze  590: static int      roffce_lines;  /* number of input lines to center */
                    591: static struct roff_node *roffce_node;  /* active request */
1.51      schwarze  592: static int      roffit_lines;  /* number of lines to delay */
                    593: static char    *roffit_macro;  /* nil-terminated macro line */
                    594:
1.80      schwarze  595:
1.138     schwarze  596: /* --- request table ------------------------------------------------------ */
                    597:
1.167     schwarze  598: struct ohash *
                    599: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  600: {
1.167     schwarze  601:        struct ohash    *htab;
                    602:        struct roffreq  *req;
                    603:        enum roff_tok    tok;
                    604:        size_t           sz;
                    605:        unsigned int     slot;
1.3       schwarze  606:
1.167     schwarze  607:        htab = mandoc_malloc(sizeof(*htab));
                    608:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  609:
1.167     schwarze  610:        for (tok = mintok; tok < maxtok; tok++) {
1.168     schwarze  611:                if (roff_name[tok] == NULL)
                    612:                        continue;
1.167     schwarze  613:                sz = strlen(roff_name[tok]);
                    614:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    615:                req->tok = tok;
                    616:                memcpy(req->name, roff_name[tok], sz + 1);
                    617:                slot = ohash_qlookup(htab, req->name);
                    618:                ohash_insert(htab, slot, req);
1.3       schwarze  619:        }
1.167     schwarze  620:        return htab;
1.3       schwarze  621: }
                    622:
1.167     schwarze  623: void
                    624: roffhash_free(struct ohash *htab)
1.1       schwarze  625: {
1.167     schwarze  626:        struct roffreq  *req;
                    627:        unsigned int     slot;
1.1       schwarze  628:
1.167     schwarze  629:        if (htab == NULL)
                    630:                return;
                    631:        for (req = ohash_first(htab, &slot); req != NULL;
                    632:             req = ohash_next(htab, &slot))
                    633:                free(req);
                    634:        ohash_delete(htab);
                    635:        free(htab);
                    636: }
                    637:
                    638: enum roff_tok
                    639: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    640: {
                    641:        struct roffreq  *req;
                    642:        const char      *end;
                    643:
                    644:        if (sz) {
                    645:                end = name + sz;
                    646:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    647:        } else
                    648:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    649:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  650: }
                    651:
1.138     schwarze  652: /* --- stack of request blocks -------------------------------------------- */
                    653:
1.1       schwarze  654: /*
                    655:  * Pop the current node off of the stack of roff instructions currently
                    656:  * pending.
                    657:  */
                    658: static void
                    659: roffnode_pop(struct roff *r)
                    660: {
                    661:        struct roffnode *p;
                    662:
1.2       schwarze  663:        assert(r->last);
1.80      schwarze  664:        p = r->last;
1.2       schwarze  665:
                    666:        r->last = r->last->parent;
1.16      schwarze  667:        free(p->name);
                    668:        free(p->end);
1.1       schwarze  669:        free(p);
                    670: }
                    671:
                    672: /*
                    673:  * Push a roff node onto the instruction stack.  This must later be
                    674:  * removed with roffnode_pop().
                    675:  */
1.11      schwarze  676: static void
1.166     schwarze  677: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  678:                int line, int col)
1.1       schwarze  679: {
                    680:        struct roffnode *p;
                    681:
1.11      schwarze  682:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  683:        p->tok = tok;
1.16      schwarze  684:        if (name)
                    685:                p->name = mandoc_strdup(name);
1.1       schwarze  686:        p->parent = r->last;
                    687:        p->line = line;
                    688:        p->col = col;
1.71      schwarze  689:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  690:
                    691:        r->last = p;
                    692: }
                    693:
1.138     schwarze  694: /* --- roff parser state data management ---------------------------------- */
                    695:
1.1       schwarze  696: static void
                    697: roff_free1(struct roff *r)
                    698: {
1.49      schwarze  699:        struct tbl_node *tbl;
1.42      schwarze  700:        int              i;
1.27      schwarze  701:
1.49      schwarze  702:        while (NULL != (tbl = r->first_tbl)) {
                    703:                r->first_tbl = tbl->next;
                    704:                tbl_free(tbl);
1.27      schwarze  705:        }
                    706:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  707:
1.191     schwarze  708:        if (r->last_eqn != NULL)
                    709:                eqn_free(r->last_eqn);
                    710:        r->last_eqn = r->eqn = NULL;
1.32      schwarze  711:
1.1       schwarze  712:        while (r->last)
                    713:                roffnode_pop(r);
1.27      schwarze  714:
1.96      schwarze  715:        free (r->rstack);
                    716:        r->rstack = NULL;
                    717:        r->rstacksz = 0;
                    718:        r->rstackpos = -1;
                    719:
                    720:        roff_freereg(r->regtab);
                    721:        r->regtab = NULL;
                    722:
1.42      schwarze  723:        roff_freestr(r->strtab);
1.178     schwarze  724:        roff_freestr(r->rentab);
1.42      schwarze  725:        roff_freestr(r->xmbtab);
1.178     schwarze  726:        r->strtab = r->rentab = r->xmbtab = NULL;
1.42      schwarze  727:
                    728:        if (r->xtab)
                    729:                for (i = 0; i < 128; i++)
                    730:                        free(r->xtab[i].p);
                    731:        free(r->xtab);
                    732:        r->xtab = NULL;
1.1       schwarze  733: }
                    734:
                    735: void
                    736: roff_reset(struct roff *r)
                    737: {
                    738:        roff_free1(r);
1.99      schwarze  739:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175     schwarze  740:        r->control = '\0';
                    741:        r->escape = '\\';
1.179     schwarze  742:        roffce_lines = 0;
                    743:        roffce_node = NULL;
                    744:        roffit_lines = 0;
                    745:        roffit_macro = NULL;
1.1       schwarze  746: }
                    747:
                    748: void
                    749: roff_free(struct roff *r)
                    750: {
                    751:        roff_free1(r);
1.167     schwarze  752:        roffhash_free(r->reqtab);
1.1       schwarze  753:        free(r);
                    754: }
                    755:
                    756: struct roff *
1.151     schwarze  757: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  758: {
                    759:        struct roff     *r;
                    760:
1.11      schwarze  761:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  762:        r->parse = parse;
1.200   ! schwarze  763:        r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.72      schwarze  764:        r->options = options;
1.99      schwarze  765:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  766:        r->rstackpos = -1;
1.175     schwarze  767:        r->escape = '\\';
1.149     schwarze  768:        return r;
1.137     schwarze  769: }
                    770:
1.138     schwarze  771: /* --- syntax tree state data management ---------------------------------- */
                    772:
1.137     schwarze  773: static void
                    774: roff_man_free1(struct roff_man *man)
                    775: {
                    776:
1.138     schwarze  777:        if (man->first != NULL)
                    778:                roff_node_delete(man, man->first);
1.137     schwarze  779:        free(man->meta.msec);
                    780:        free(man->meta.vol);
                    781:        free(man->meta.os);
                    782:        free(man->meta.arch);
                    783:        free(man->meta.title);
                    784:        free(man->meta.name);
                    785:        free(man->meta.date);
                    786: }
                    787:
                    788: static void
                    789: roff_man_alloc1(struct roff_man *man)
                    790: {
                    791:
                    792:        memset(&man->meta, 0, sizeof(man->meta));
                    793:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    794:        man->first->type = ROFFT_ROOT;
                    795:        man->last = man->first;
                    796:        man->last_es = NULL;
                    797:        man->flags = 0;
                    798:        man->macroset = MACROSET_NONE;
                    799:        man->lastsec = man->lastnamed = SEC_NONE;
                    800:        man->next = ROFF_NEXT_CHILD;
                    801: }
                    802:
                    803: void
                    804: roff_man_reset(struct roff_man *man)
                    805: {
                    806:
                    807:        roff_man_free1(man);
                    808:        roff_man_alloc1(man);
                    809: }
                    810:
                    811: void
                    812: roff_man_free(struct roff_man *man)
                    813: {
                    814:
                    815:        roff_man_free1(man);
                    816:        free(man);
                    817: }
                    818:
                    819: struct roff_man *
                    820: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.188     schwarze  821:        const char *os_s, int quick)
1.137     schwarze  822: {
                    823:        struct roff_man *man;
                    824:
                    825:        man = mandoc_calloc(1, sizeof(*man));
                    826:        man->parse = parse;
                    827:        man->roff = roff;
1.188     schwarze  828:        man->os_s = os_s;
1.137     schwarze  829:        man->quick = quick;
                    830:        roff_man_alloc1(man);
1.168     schwarze  831:        roff->man = man;
1.149     schwarze  832:        return man;
1.1       schwarze  833: }
                    834:
1.138     schwarze  835: /* --- syntax tree handling ----------------------------------------------- */
                    836:
                    837: struct roff_node *
                    838: roff_node_alloc(struct roff_man *man, int line, int pos,
                    839:        enum roff_type type, int tok)
                    840: {
                    841:        struct roff_node        *n;
                    842:
                    843:        n = mandoc_calloc(1, sizeof(*n));
                    844:        n->line = line;
                    845:        n->pos = pos;
                    846:        n->tok = tok;
                    847:        n->type = type;
                    848:        n->sec = man->lastsec;
                    849:
                    850:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  851:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  852:        else
1.157     schwarze  853:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  854:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  855:                n->flags |= NODE_LINE;
1.138     schwarze  856:        man->flags &= ~MDOC_NEWLINE;
                    857:
1.149     schwarze  858:        return n;
1.138     schwarze  859: }
                    860:
                    861: void
                    862: roff_node_append(struct roff_man *man, struct roff_node *n)
                    863: {
                    864:
                    865:        switch (man->next) {
                    866:        case ROFF_NEXT_SIBLING:
1.153     schwarze  867:                if (man->last->next != NULL) {
                    868:                        n->next = man->last->next;
                    869:                        man->last->next->prev = n;
                    870:                } else
                    871:                        man->last->parent->last = n;
1.138     schwarze  872:                man->last->next = n;
                    873:                n->prev = man->last;
                    874:                n->parent = man->last->parent;
                    875:                break;
                    876:        case ROFF_NEXT_CHILD:
1.159     schwarze  877:                if (man->last->child != NULL) {
                    878:                        n->next = man->last->child;
                    879:                        man->last->child->prev = n;
                    880:                } else
                    881:                        man->last->last = n;
1.138     schwarze  882:                man->last->child = n;
                    883:                n->parent = man->last;
                    884:                break;
                    885:        default:
                    886:                abort();
                    887:        }
1.154     schwarze  888:        man->last = n;
1.138     schwarze  889:
                    890:        switch (n->type) {
                    891:        case ROFFT_HEAD:
                    892:                n->parent->head = n;
                    893:                break;
                    894:        case ROFFT_BODY:
1.154     schwarze  895:                if (n->end != ENDBODY_NOT)
                    896:                        return;
1.138     schwarze  897:                n->parent->body = n;
                    898:                break;
                    899:        case ROFFT_TAIL:
                    900:                n->parent->tail = n;
                    901:                break;
                    902:        default:
1.154     schwarze  903:                return;
1.138     schwarze  904:        }
1.154     schwarze  905:
                    906:        /*
                    907:         * Copy over the normalised-data pointer of our parent.  Not
                    908:         * everybody has one, but copying a null pointer is fine.
                    909:         */
                    910:
                    911:        n->norm = n->parent->norm;
                    912:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  913: }
                    914:
1.139     schwarze  915: void
                    916: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    917: {
                    918:        struct roff_node        *n;
                    919:
                    920:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    921:        n->string = roff_strdup(man->roff, word);
                    922:        roff_node_append(man, n);
1.158     schwarze  923:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  924:        man->next = ROFF_NEXT_SIBLING;
                    925: }
                    926:
                    927: void
                    928: roff_word_append(struct roff_man *man, const char *word)
                    929: {
                    930:        struct roff_node        *n;
                    931:        char                    *addstr, *newstr;
                    932:
                    933:        n = man->last;
                    934:        addstr = roff_strdup(man->roff, word);
                    935:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    936:        free(addstr);
                    937:        free(n->string);
                    938:        n->string = newstr;
                    939:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  940: }
                    941:
                    942: void
                    943: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    944: {
                    945:        struct roff_node        *n;
                    946:
                    947:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    948:        roff_node_append(man, n);
                    949:        man->next = ROFF_NEXT_CHILD;
                    950: }
                    951:
                    952: struct roff_node *
                    953: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    954: {
                    955:        struct roff_node        *n;
                    956:
                    957:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    958:        roff_node_append(man, n);
                    959:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  960:        return n;
1.139     schwarze  961: }
                    962:
1.138     schwarze  963: struct roff_node *
                    964: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    965: {
                    966:        struct roff_node        *n;
                    967:
                    968:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    969:        roff_node_append(man, n);
                    970:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  971:        return n;
1.138     schwarze  972: }
                    973:
                    974: struct roff_node *
                    975: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    976: {
                    977:        struct roff_node        *n;
                    978:
                    979:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    980:        roff_node_append(man, n);
                    981:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  982:        return n;
1.139     schwarze  983: }
                    984:
1.193     schwarze  985: static void
                    986: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.139     schwarze  987: {
                    988:        struct roff_node        *n;
1.193     schwarze  989:        const struct tbl_span   *span;
1.139     schwarze  990:
                    991:        if (man->macroset == MACROSET_MAN)
1.174     schwarze  992:                man_breakscope(man, ROFF_TS);
1.193     schwarze  993:        while ((span = tbl_span(tbl)) != NULL) {
                    994:                n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                    995:                n->span = span;
                    996:                roff_node_append(man, n);
                    997:                n->flags |= NODE_VALID | NODE_ENDED;
                    998:                man->next = ROFF_NEXT_SIBLING;
                    999:        }
1.138     schwarze 1000: }
                   1001:
                   1002: void
                   1003: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1004: {
                   1005:
                   1006:        /* Adjust siblings. */
                   1007:
                   1008:        if (n->prev)
                   1009:                n->prev->next = n->next;
                   1010:        if (n->next)
                   1011:                n->next->prev = n->prev;
                   1012:
                   1013:        /* Adjust parent. */
                   1014:
                   1015:        if (n->parent != NULL) {
                   1016:                if (n->parent->child == n)
                   1017:                        n->parent->child = n->next;
                   1018:                if (n->parent->last == n)
                   1019:                        n->parent->last = n->prev;
                   1020:        }
                   1021:
                   1022:        /* Adjust parse point. */
                   1023:
                   1024:        if (man == NULL)
                   1025:                return;
                   1026:        if (man->last == n) {
                   1027:                if (n->prev == NULL) {
                   1028:                        man->last = n->parent;
                   1029:                        man->next = ROFF_NEXT_CHILD;
                   1030:                } else {
                   1031:                        man->last = n->prev;
                   1032:                        man->next = ROFF_NEXT_SIBLING;
                   1033:                }
                   1034:        }
                   1035:        if (man->first == n)
                   1036:                man->first = NULL;
                   1037: }
                   1038:
                   1039: void
                   1040: roff_node_free(struct roff_node *n)
                   1041: {
                   1042:
                   1043:        if (n->args != NULL)
                   1044:                mdoc_argv_free(n->args);
                   1045:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1046:                free(n->norm);
1.191     schwarze 1047:        if (n->eqn != NULL)
                   1048:                eqn_box_free(n->eqn);
1.138     schwarze 1049:        free(n->string);
                   1050:        free(n);
                   1051: }
                   1052:
                   1053: void
                   1054: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1055: {
                   1056:
                   1057:        while (n->child != NULL)
                   1058:                roff_node_delete(man, n->child);
                   1059:        roff_node_unlink(man, n);
                   1060:        roff_node_free(n);
1.141     schwarze 1061: }
                   1062:
                   1063: void
                   1064: deroff(char **dest, const struct roff_node *n)
                   1065: {
                   1066:        char    *cp;
                   1067:        size_t   sz;
                   1068:
                   1069:        if (n->type != ROFFT_TEXT) {
                   1070:                for (n = n->child; n != NULL; n = n->next)
                   1071:                        deroff(dest, n);
                   1072:                return;
                   1073:        }
                   1074:
1.160     schwarze 1075:        /* Skip leading whitespace. */
1.141     schwarze 1076:
1.160     schwarze 1077:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1078:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1079:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1080:                        cp++;
1.160     schwarze 1081:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1082:                        break;
                   1083:        }
                   1084:
1.161     schwarze 1085:        /* Skip trailing backslash. */
                   1086:
                   1087:        sz = strlen(cp);
1.162     schwarze 1088:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1089:                sz--;
                   1090:
1.141     schwarze 1091:        /* Skip trailing whitespace. */
                   1092:
1.161     schwarze 1093:        for (; sz; sz--)
1.141     schwarze 1094:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1095:                        break;
                   1096:
                   1097:        /* Skip empty strings. */
                   1098:
                   1099:        if (sz == 0)
                   1100:                return;
                   1101:
                   1102:        if (*dest == NULL) {
                   1103:                *dest = mandoc_strndup(cp, sz);
                   1104:                return;
                   1105:        }
                   1106:
                   1107:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1108:        free(*dest);
                   1109:        *dest = cp;
1.138     schwarze 1110: }
                   1111:
                   1112: /* --- main functions of the roff parser ---------------------------------- */
                   1113:
1.8       schwarze 1114: /*
1.79      schwarze 1115:  * In the current line, expand escape sequences that tend to get
                   1116:  * used in numerical expressions and conditional requests.
                   1117:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1118:  */
1.45      schwarze 1119: static enum rofferr
1.110     schwarze 1120: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1121: {
1.81      schwarze 1122:        char             ubuf[24]; /* buffer to print the number */
1.200   ! schwarze 1123:        struct roff_node *n;    /* used for header comments */
1.78      schwarze 1124:        const char      *start; /* start of the string to process */
1.82      schwarze 1125:        char            *stesc; /* start of an escape sequence ('\\') */
1.200   ! schwarze 1126:        char            *ep;    /* end of comment string */
1.23      schwarze 1127:        const char      *stnam; /* start of the name, after "[(*" */
                   1128:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1129:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1130:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1131:        size_t           maxl;  /* expected length of the escape name */
                   1132:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1133:        enum mandoc_esc  esc;   /* type of the escape sequence */
                   1134:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1135:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1136:        int              npos;  /* position in numeric expression */
1.91      schwarze 1137:        int              arg_complete; /* argument not interrupted by eol */
1.175     schwarze 1138:        int              done;  /* no more input available */
1.187     schwarze 1139:        int              deftype; /* type of definition to paste */
1.195     schwarze 1140:        int              rcsid; /* kind of RCS id seen */
1.199     schwarze 1141:        char             sign;  /* increment number register */
1.79      schwarze 1142:        char             term;  /* character terminating the escape */
1.8       schwarze 1143:
1.175     schwarze 1144:        /* Search forward for comments. */
                   1145:
                   1146:        done = 0;
                   1147:        start = buf->buf + pos;
                   1148:        for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
                   1149:                if (stesc[0] != r->escape || stesc[1] == '\0')
                   1150:                        continue;
                   1151:                stesc++;
                   1152:                if (*stesc != '"' && *stesc != '#')
                   1153:                        continue;
1.185     schwarze 1154:
                   1155:                /* Comment found, look for RCS id. */
                   1156:
1.195     schwarze 1157:                rcsid = 0;
1.185     schwarze 1158:                if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.195     schwarze 1159:                        rcsid = 1 << MANDOC_OS_OPENBSD;
1.185     schwarze 1160:                        cp += 8;
                   1161:                } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.195     schwarze 1162:                        rcsid = 1 << MANDOC_OS_NETBSD;
1.185     schwarze 1163:                        cp += 7;
                   1164:                }
                   1165:                if (cp != NULL &&
                   1166:                    isalnum((unsigned char)*cp) == 0 &&
1.186     schwarze 1167:                    strchr(cp, '$') != NULL) {
1.195     schwarze 1168:                        if (r->man->meta.rcsids & rcsid)
1.186     schwarze 1169:                                mandoc_msg(MANDOCERR_RCS_REP, r->parse,
                   1170:                                    ln, stesc + 1 - buf->buf, stesc + 1);
1.195     schwarze 1171:                        r->man->meta.rcsids |= rcsid;
1.186     schwarze 1172:                }
1.185     schwarze 1173:
                   1174:                /* Handle trailing whitespace. */
                   1175:
1.200   ! schwarze 1176:                ep = strchr(stesc--, '\0') - 1;
        !          1177:                if (*ep == '\n') {
1.175     schwarze 1178:                        done = 1;
1.200   ! schwarze 1179:                        ep--;
1.175     schwarze 1180:                }
1.200   ! schwarze 1181:                if (*ep == ' ' || *ep == '\t')
1.175     schwarze 1182:                        mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1.200   ! schwarze 1183:                            ln, ep - buf->buf, NULL);
        !          1184:
        !          1185:                /*
        !          1186:                 * Save comments preceding the title macro
        !          1187:                 * in the syntax tree.
        !          1188:                 */
        !          1189:
        !          1190:                if (r->format == 0) {
        !          1191:                        while (*ep == ' ' || *ep == '\t')
        !          1192:                                ep--;
        !          1193:                        ep[1] = '\0';
        !          1194:                        n = roff_node_alloc(r->man,
        !          1195:                            ln, stesc + 1 - buf->buf,
        !          1196:                            ROFFT_COMMENT, TOKEN_NONE);
        !          1197:                        n->string = mandoc_strdup(stesc + 2);
        !          1198:                        roff_node_append(r->man, n);
        !          1199:                        n->flags |= NODE_VALID | NODE_ENDED;
        !          1200:                        r->man->next = ROFF_NEXT_SIBLING;
        !          1201:                }
        !          1202:
        !          1203:                /* Discard comments. */
        !          1204:
1.175     schwarze 1205:                while (stesc > start && stesc[-1] == ' ')
                   1206:                        stesc--;
                   1207:                *stesc = '\0';
                   1208:                break;
                   1209:        }
                   1210:        if (stesc == start)
                   1211:                return ROFF_CONT;
                   1212:        stesc--;
                   1213:
                   1214:        /* Notice the end of the input. */
                   1215:
                   1216:        if (*stesc == '\n') {
                   1217:                *stesc-- = '\0';
                   1218:                done = 1;
                   1219:        }
                   1220:
1.43      schwarze 1221:        expand_count = 0;
1.175     schwarze 1222:        while (stesc >= start) {
1.43      schwarze 1223:
1.78      schwarze 1224:                /* Search backwards for the next backslash. */
                   1225:
1.175     schwarze 1226:                if (*stesc != r->escape) {
                   1227:                        if (*stesc == '\\') {
                   1228:                                *stesc = '\0';
                   1229:                                buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
                   1230:                                    buf->buf, stesc + 1) + 1;
                   1231:                                start = nbuf + pos;
                   1232:                                stesc = nbuf + (stesc - buf->buf);
                   1233:                                free(buf->buf);
                   1234:                                buf->buf = nbuf;
                   1235:                        }
                   1236:                        stesc--;
1.78      schwarze 1237:                        continue;
1.175     schwarze 1238:                }
1.78      schwarze 1239:
                   1240:                /* If it is escaped, skip it. */
                   1241:
                   1242:                for (cp = stesc - 1; cp >= start; cp--)
1.175     schwarze 1243:                        if (*cp != r->escape)
1.78      schwarze 1244:                                break;
                   1245:
1.110     schwarze 1246:                if ((stesc - cp) % 2 == 0) {
1.175     schwarze 1247:                        while (stesc > cp)
                   1248:                                *stesc-- = '\\';
1.78      schwarze 1249:                        continue;
1.175     schwarze 1250:                } else if (stesc[1] != '\0') {
                   1251:                        *stesc = '\\';
                   1252:                } else {
                   1253:                        *stesc-- = '\0';
                   1254:                        if (done)
                   1255:                                continue;
                   1256:                        else
                   1257:                                return ROFF_APPEND;
1.78      schwarze 1258:                }
1.23      schwarze 1259:
1.79      schwarze 1260:                /* Decide whether to expand or to check only. */
1.23      schwarze 1261:
1.79      schwarze 1262:                term = '\0';
1.78      schwarze 1263:                cp = stesc + 1;
1.53      schwarze 1264:                switch (*cp) {
1.80      schwarze 1265:                case '*':
1.53      schwarze 1266:                        res = NULL;
                   1267:                        break;
1.80      schwarze 1268:                case 'B':
                   1269:                case 'w':
1.79      schwarze 1270:                        term = cp[1];
                   1271:                        /* FALLTHROUGH */
1.80      schwarze 1272:                case 'n':
1.199     schwarze 1273:                        sign = cp[1];
                   1274:                        if (sign == '+' || sign == '-')
                   1275:                                cp++;
1.53      schwarze 1276:                        res = ubuf;
                   1277:                        break;
                   1278:                default:
1.109     schwarze 1279:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1280:                        if (esc == ESCAPE_ERROR ||
                   1281:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1282:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1283:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1284:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1285:                                    "%.*s", (int)(cp - stesc), stesc);
1.175     schwarze 1286:                        stesc--;
1.78      schwarze 1287:                        continue;
1.42      schwarze 1288:                }
                   1289:
1.78      schwarze 1290:                if (EXPAND_LIMIT < ++expand_count) {
                   1291:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1292:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1293:                        return ROFF_IGN;
1.78      schwarze 1294:                }
1.23      schwarze 1295:
                   1296:                /*
                   1297:                 * The third character decides the length
1.53      schwarze 1298:                 * of the name of the string or register.
1.23      schwarze 1299:                 * Save a pointer to the name.
                   1300:                 */
                   1301:
1.110     schwarze 1302:                if (term == '\0') {
1.79      schwarze 1303:                        switch (*++cp) {
1.80      schwarze 1304:                        case '\0':
1.79      schwarze 1305:                                maxl = 0;
                   1306:                                break;
1.80      schwarze 1307:                        case '(':
1.79      schwarze 1308:                                cp++;
                   1309:                                maxl = 2;
                   1310:                                break;
1.80      schwarze 1311:                        case '[':
1.79      schwarze 1312:                                cp++;
                   1313:                                term = ']';
                   1314:                                maxl = 0;
                   1315:                                break;
                   1316:                        default:
                   1317:                                maxl = 1;
                   1318:                                break;
                   1319:                        }
                   1320:                } else {
                   1321:                        cp += 2;
1.8       schwarze 1322:                        maxl = 0;
                   1323:                }
1.23      schwarze 1324:                stnam = cp;
1.8       schwarze 1325:
1.23      schwarze 1326:                /* Advance to the end of the name. */
1.8       schwarze 1327:
1.125     schwarze 1328:                naml = 0;
1.91      schwarze 1329:                arg_complete = 1;
1.125     schwarze 1330:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1331:                        if (*cp == '\0') {
1.92      schwarze 1332:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1333:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1334:                                arg_complete = 0;
1.79      schwarze 1335:                                break;
1.42      schwarze 1336:                        }
1.110     schwarze 1337:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1338:                                cp++;
1.125     schwarze 1339:                                break;
                   1340:                        }
                   1341:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1342:                                naml++;
                   1343:                                continue;
                   1344:                        }
                   1345:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1346:                        case ESCAPE_SPECIAL:
                   1347:                        case ESCAPE_UNICODE:
                   1348:                        case ESCAPE_NUMBERED:
                   1349:                        case ESCAPE_OVERSTRIKE:
                   1350:                                naml++;
                   1351:                                break;
                   1352:                        default:
1.8       schwarze 1353:                                break;
1.79      schwarze 1354:                        }
1.8       schwarze 1355:                }
                   1356:
1.23      schwarze 1357:                /*
                   1358:                 * Retrieve the replacement string; if it is
                   1359:                 * undefined, resume searching for escapes.
                   1360:                 */
                   1361:
1.79      schwarze 1362:                switch (stesc[1]) {
1.80      schwarze 1363:                case '*':
1.187     schwarze 1364:                        if (arg_complete) {
                   1365:                                deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1366:                                res = roff_getstrn(r, stnam, naml, &deftype);
                   1367:                        }
1.79      schwarze 1368:                        break;
1.80      schwarze 1369:                case 'B':
1.79      schwarze 1370:                        npos = 0;
1.91      schwarze 1371:                        ubuf[0] = arg_complete &&
1.133     schwarze 1372:                            roff_evalnum(r, ln, stnam, &npos,
                   1373:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1374:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1375:                        ubuf[1] = '\0';
                   1376:                        break;
1.80      schwarze 1377:                case 'n':
1.91      schwarze 1378:                        if (arg_complete)
                   1379:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.199     schwarze 1380:                                    roff_getregn(r, stnam, naml, sign));
1.91      schwarze 1381:                        else
                   1382:                                ubuf[0] = '\0';
1.79      schwarze 1383:                        break;
1.80      schwarze 1384:                case 'w':
1.91      schwarze 1385:                        /* use even incomplete args */
1.81      schwarze 1386:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1387:                            24 * (int)naml);
                   1388:                        break;
                   1389:                }
1.8       schwarze 1390:
1.110     schwarze 1391:                if (res == NULL) {
1.92      schwarze 1392:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1393:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1394:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1395:                        res = "";
1.118     schwarze 1396:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1397:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1398:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1399:                        return ROFF_IGN;
1.8       schwarze 1400:                }
                   1401:
1.23      schwarze 1402:                /* Replace the escape sequence by the string. */
                   1403:
1.82      schwarze 1404:                *stesc = '\0';
1.110     schwarze 1405:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1406:                    buf->buf, res, cp) + 1;
1.8       schwarze 1407:
1.78      schwarze 1408:                /* Prepare for the next replacement. */
1.8       schwarze 1409:
1.78      schwarze 1410:                start = nbuf + pos;
1.110     schwarze 1411:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1412:                free(buf->buf);
                   1413:                buf->buf = nbuf;
1.42      schwarze 1414:        }
1.149     schwarze 1415:        return ROFF_CONT;
1.42      schwarze 1416: }
                   1417:
                   1418: /*
1.147     schwarze 1419:  * Process text streams.
1.42      schwarze 1420:  */
                   1421: static enum rofferr
1.177     schwarze 1422: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42      schwarze 1423: {
                   1424:        size_t           sz;
                   1425:        const char      *start;
1.51      schwarze 1426:        char            *p;
                   1427:        int              isz;
1.42      schwarze 1428:        enum mandoc_esc  esc;
                   1429:
1.147     schwarze 1430:        /* Spring the input line trap. */
                   1431:
                   1432:        if (roffit_lines == 1) {
                   1433:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1434:                free(buf->buf);
                   1435:                buf->buf = p;
                   1436:                buf->sz = isz + 1;
                   1437:                *offs = 0;
                   1438:                free(roffit_macro);
                   1439:                roffit_lines = 0;
1.149     schwarze 1440:                return ROFF_REPARSE;
1.147     schwarze 1441:        } else if (roffit_lines > 1)
                   1442:                --roffit_lines;
                   1443:
1.177     schwarze 1444:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1445:                if (roffce_lines < 1) {
                   1446:                        r->man->last = roffce_node;
                   1447:                        r->man->next = ROFF_NEXT_SIBLING;
                   1448:                        roffce_lines = 0;
                   1449:                        roffce_node = NULL;
                   1450:                } else
                   1451:                        roffce_lines--;
                   1452:        }
                   1453:
1.147     schwarze 1454:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1455:
1.110     schwarze 1456:        start = p = buf->buf + pos;
1.42      schwarze 1457:
1.110     schwarze 1458:        while (*p != '\0') {
1.42      schwarze 1459:                sz = strcspn(p, "-\\");
                   1460:                p += sz;
                   1461:
1.110     schwarze 1462:                if (*p == '\0')
1.42      schwarze 1463:                        break;
                   1464:
1.110     schwarze 1465:                if (*p == '\\') {
1.42      schwarze 1466:                        /* Skip over escapes. */
                   1467:                        p++;
1.62      schwarze 1468:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1469:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1470:                                break;
1.136     schwarze 1471:                        while (*p == '-')
                   1472:                                p++;
1.42      schwarze 1473:                        continue;
                   1474:                } else if (p == start) {
                   1475:                        p++;
                   1476:                        continue;
                   1477:                }
                   1478:
1.44      schwarze 1479:                if (isalpha((unsigned char)p[-1]) &&
                   1480:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1481:                        *p = ASCII_HYPH;
                   1482:                p++;
1.8       schwarze 1483:        }
1.149     schwarze 1484:        return ROFF_CONT;
1.8       schwarze 1485: }
                   1486:
1.1       schwarze 1487: enum rofferr
1.110     schwarze 1488: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1489: {
1.166     schwarze 1490:        enum roff_tok    t;
1.27      schwarze 1491:        enum rofferr     e;
1.110     schwarze 1492:        int              pos;   /* parse point */
1.115     schwarze 1493:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1494:        int              ppos;  /* original offset in buf->buf */
                   1495:        int              ctl;   /* macro line (boolean) */
                   1496:
                   1497:        ppos = pos = *offs;
1.1       schwarze 1498:
1.102     schwarze 1499:        /* Handle in-line equation delimiters. */
                   1500:
1.108     schwarze 1501:        if (r->tbl == NULL &&
                   1502:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1503:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1504:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1505:                if (e == ROFF_REPARSE)
1.149     schwarze 1506:                        return e;
1.102     schwarze 1507:                assert(e == ROFF_CONT);
                   1508:        }
                   1509:
                   1510:        /* Expand some escape sequences. */
1.8       schwarze 1511:
1.110     schwarze 1512:        e = roff_res(r, buf, ln, pos);
1.175     schwarze 1513:        if (e == ROFF_IGN || e == ROFF_APPEND)
1.149     schwarze 1514:                return e;
1.110     schwarze 1515:        assert(e == ROFF_CONT);
1.8       schwarze 1516:
1.110     schwarze 1517:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1518:
1.8       schwarze 1519:        /*
1.2       schwarze 1520:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1521:         * text through the macro's filter.
                   1522:         * Equations process all content themselves.
                   1523:         * Tables process almost all content themselves, but we want
                   1524:         * to warn about macros before passing it there.
1.2       schwarze 1525:         */
                   1526:
1.124     schwarze 1527:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1528:                t = r->last->tok;
1.110     schwarze 1529:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1530:                if (e == ROFF_IGN)
1.149     schwarze 1531:                        return e;
1.166     schwarze 1532:                assert(e == ROFF_CONT);
1.54      schwarze 1533:        }
1.191     schwarze 1534:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1535:                eqn_read(r->eqn, buf->buf + ppos);
                   1536:                return ROFF_IGN;
                   1537:        }
1.193     schwarze 1538:        if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
                   1539:                tbl_read(r->tbl, ln, buf->buf, ppos);
                   1540:                roff_addtbl(r->man, r->tbl);
                   1541:                return ROFF_IGN;
                   1542:        }
1.124     schwarze 1543:        if ( ! ctl)
1.177     schwarze 1544:                return roff_parsetext(r, buf, pos, offs);
1.100     schwarze 1545:
                   1546:        /* Skip empty request lines. */
                   1547:
1.110     schwarze 1548:        if (buf->buf[pos] == '"') {
1.100     schwarze 1549:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1550:                    ln, pos, NULL);
1.149     schwarze 1551:                return ROFF_IGN;
1.110     schwarze 1552:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1553:                return ROFF_IGN;
1.2       schwarze 1554:
                   1555:        /*
                   1556:         * If a scope is open, go to the child handler for that macro,
                   1557:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1558:         * Don't do so if an equation is open.
1.2       schwarze 1559:         */
                   1560:
                   1561:        if (r->last) {
1.1       schwarze 1562:                t = r->last->tok;
1.149     schwarze 1563:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1564:        }
                   1565:
1.115     schwarze 1566:        /* No scope is open.  This is a new request or macro. */
                   1567:
                   1568:        spos = pos;
                   1569:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1570:
                   1571:        /* Tables ignore most macros. */
                   1572:
1.180     schwarze 1573:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181     schwarze 1574:            t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.115     schwarze 1575:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1576:                    ln, pos, buf->buf + spos);
1.180     schwarze 1577:                if (t != TOKEN_NONE)
1.149     schwarze 1578:                        return ROFF_IGN;
1.129     schwarze 1579:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1580:                        pos++;
1.163     schwarze 1581:                while (buf->buf[pos] == ' ')
1.129     schwarze 1582:                        pos++;
1.193     schwarze 1583:                tbl_read(r->tbl, ln, buf->buf, pos);
                   1584:                roff_addtbl(r->man, r->tbl);
                   1585:                return ROFF_IGN;
1.115     schwarze 1586:        }
                   1587:
1.177     schwarze 1588:        /* For now, let high level macros abort .ce mode. */
                   1589:
                   1590:        if (ctl && roffce_node != NULL &&
1.196     schwarze 1591:            (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
                   1592:             t == ROFF_TH || t == ROFF_TS)) {
1.177     schwarze 1593:                r->man->last = roffce_node;
                   1594:                r->man->next = ROFF_NEXT_SIBLING;
                   1595:                roffce_lines = 0;
                   1596:                roffce_node = NULL;
                   1597:        }
                   1598:
1.2       schwarze 1599:        /*
1.115     schwarze 1600:         * This is neither a roff request nor a user-defined macro.
                   1601:         * Let the standard macro set parsers handle it.
1.2       schwarze 1602:         */
                   1603:
1.166     schwarze 1604:        if (t == TOKEN_NONE)
1.149     schwarze 1605:                return ROFF_CONT;
1.115     schwarze 1606:
                   1607:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1608:
1.168     schwarze 1609:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2       schwarze 1610: }
                   1611:
1.27      schwarze 1612: void
1.2       schwarze 1613: roff_endparse(struct roff *r)
                   1614: {
1.193     schwarze 1615:        if (r->last != NULL)
1.94      schwarze 1616:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1617:                    r->last->line, r->last->col,
1.167     schwarze 1618:                    roff_name[r->last->tok]);
1.27      schwarze 1619:
1.193     schwarze 1620:        if (r->eqn != NULL) {
1.94      schwarze 1621:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.191     schwarze 1622:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1623:                eqn_parse(r->eqn);
                   1624:                r->eqn = NULL;
1.32      schwarze 1625:        }
                   1626:
1.193     schwarze 1627:        if (r->tbl != NULL) {
1.94      schwarze 1628:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1629:                    r->tbl->line, r->tbl->pos, "TS");
1.193     schwarze 1630:                tbl_end(r->tbl);
                   1631:                r->tbl = NULL;
1.27      schwarze 1632:        }
1.1       schwarze 1633: }
                   1634:
                   1635: /*
                   1636:  * Parse a roff node's type from the input buffer.  This must be in the
                   1637:  * form of ".foo xxx" in the usual way.
                   1638:  */
1.166     schwarze 1639: static enum roff_tok
1.87      schwarze 1640: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1641: {
1.87      schwarze 1642:        char            *cp;
1.16      schwarze 1643:        const char      *mac;
                   1644:        size_t           maclen;
1.187     schwarze 1645:        int              deftype;
1.166     schwarze 1646:        enum roff_tok    t;
1.1       schwarze 1647:
1.87      schwarze 1648:        cp = buf + *pos;
                   1649:
                   1650:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1651:                return TOKEN_NONE;
1.1       schwarze 1652:
1.87      schwarze 1653:        mac = cp;
                   1654:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1655:
1.187     schwarze 1656:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1657:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1658:        switch (deftype) {
                   1659:        case ROFFDEF_USER:
                   1660:                t = ROFF_USERDEF;
                   1661:                break;
                   1662:        case ROFFDEF_REN:
                   1663:                t = ROFF_RENAMED;
                   1664:                break;
                   1665:        default:
                   1666:                t = roffhash_find(r->reqtab, mac, maclen);
                   1667:                break;
                   1668:        }
1.166     schwarze 1669:        if (t != TOKEN_NONE)
1.87      schwarze 1670:                *pos = cp - buf;
1.197     schwarze 1671:        else if (deftype == ROFFDEF_UNDEF) {
                   1672:                /* Using an undefined macro defines it to be empty. */
                   1673:                roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
                   1674:                roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
                   1675:        }
1.149     schwarze 1676:        return t;
1.1       schwarze 1677: }
                   1678:
1.138     schwarze 1679: /* --- handling of request blocks ----------------------------------------- */
                   1680:
1.1       schwarze 1681: static enum rofferr
1.2       schwarze 1682: roff_cblock(ROFF_ARGS)
1.1       schwarze 1683: {
                   1684:
1.2       schwarze 1685:        /*
                   1686:         * A block-close `..' should only be invoked as a child of an
                   1687:         * ignore macro, otherwise raise a warning and just ignore it.
                   1688:         */
                   1689:
1.110     schwarze 1690:        if (r->last == NULL) {
1.94      schwarze 1691:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1692:                    ln, ppos, "..");
1.149     schwarze 1693:                return ROFF_IGN;
1.2       schwarze 1694:        }
1.1       schwarze 1695:
1.2       schwarze 1696:        switch (r->last->tok) {
1.80      schwarze 1697:        case ROFF_am:
1.93      schwarze 1698:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1699:        case ROFF_ami:
                   1700:        case ROFF_de:
1.23      schwarze 1701:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1702:        case ROFF_dei:
                   1703:        case ROFF_ig:
1.2       schwarze 1704:                break;
                   1705:        default:
1.94      schwarze 1706:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1707:                    ln, ppos, "..");
1.149     schwarze 1708:                return ROFF_IGN;
1.2       schwarze 1709:        }
                   1710:
1.110     schwarze 1711:        if (buf->buf[pos] != '\0')
1.90      schwarze 1712:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1713:                    ".. %s", buf->buf + pos);
1.2       schwarze 1714:
                   1715:        roffnode_pop(r);
                   1716:        roffnode_cleanscope(r);
1.149     schwarze 1717:        return ROFF_IGN;
1.2       schwarze 1718:
                   1719: }
1.1       schwarze 1720:
1.2       schwarze 1721: static void
                   1722: roffnode_cleanscope(struct roff *r)
                   1723: {
1.1       schwarze 1724:
1.2       schwarze 1725:        while (r->last) {
1.46      schwarze 1726:                if (--r->last->endspan != 0)
1.2       schwarze 1727:                        break;
                   1728:                roffnode_pop(r);
                   1729:        }
                   1730: }
1.1       schwarze 1731:
1.68      schwarze 1732: static void
                   1733: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1734: {
1.1       schwarze 1735:
1.2       schwarze 1736:        if (NULL == r->last) {
1.94      schwarze 1737:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1738:                    ln, ppos, "\\}");
1.68      schwarze 1739:                return;
1.2       schwarze 1740:        }
1.1       schwarze 1741:
1.2       schwarze 1742:        switch (r->last->tok) {
1.80      schwarze 1743:        case ROFF_el:
                   1744:        case ROFF_ie:
                   1745:        case ROFF_if:
1.2       schwarze 1746:                break;
                   1747:        default:
1.94      schwarze 1748:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1749:                    ln, ppos, "\\}");
1.68      schwarze 1750:                return;
1.2       schwarze 1751:        }
1.1       schwarze 1752:
1.2       schwarze 1753:        if (r->last->endspan > -1) {
1.94      schwarze 1754:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1755:                    ln, ppos, "\\}");
1.68      schwarze 1756:                return;
1.2       schwarze 1757:        }
                   1758:
                   1759:        roffnode_pop(r);
                   1760:        roffnode_cleanscope(r);
1.68      schwarze 1761:        return;
1.1       schwarze 1762: }
                   1763:
                   1764: static enum rofferr
1.2       schwarze 1765: roff_block(ROFF_ARGS)
1.1       schwarze 1766: {
1.187     schwarze 1767:        const char      *name, *value;
                   1768:        char            *call, *cp, *iname, *rname;
                   1769:        size_t           csz, namesz, rsz;
                   1770:        int              deftype;
1.16      schwarze 1771:
1.93      schwarze 1772:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1773:
1.110     schwarze 1774:        if (tok == ROFF_de1)
1.93      schwarze 1775:                tok = ROFF_de;
1.123     schwarze 1776:        else if (tok == ROFF_dei1)
                   1777:                tok = ROFF_dei;
1.110     schwarze 1778:        else if (tok == ROFF_am1)
1.93      schwarze 1779:                tok = ROFF_am;
1.123     schwarze 1780:        else if (tok == ROFF_ami1)
                   1781:                tok = ROFF_ami;
1.93      schwarze 1782:
                   1783:        /* Parse the macro name argument. */
                   1784:
1.110     schwarze 1785:        cp = buf->buf + pos;
                   1786:        if (tok == ROFF_ig) {
1.93      schwarze 1787:                iname = NULL;
                   1788:                namesz = 0;
                   1789:        } else {
                   1790:                iname = cp;
                   1791:                namesz = roff_getname(r, &cp, ln, ppos);
                   1792:                iname[namesz] = '\0';
                   1793:        }
1.22      schwarze 1794:
1.93      schwarze 1795:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1796:
1.110     schwarze 1797:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1798:                deftype = ROFFDEF_USER;
                   1799:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1800:                if (name == NULL) {
1.93      schwarze 1801:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1802:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1803:                            "%.*s", (int)namesz, iname);
                   1804:                        namesz = 0;
                   1805:                } else
                   1806:                        namesz = strlen(name);
                   1807:        } else
                   1808:                name = iname;
1.22      schwarze 1809:
1.110     schwarze 1810:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1811:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167     schwarze 1812:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1813:                return ROFF_IGN;
1.93      schwarze 1814:        }
1.2       schwarze 1815:
1.16      schwarze 1816:        roffnode_push(r, tok, name, ln, ppos);
                   1817:
                   1818:        /*
                   1819:         * At the beginning of a `de' macro, clear the existing string
                   1820:         * with the same name, if there is one.  New content will be
1.66      schwarze 1821:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1822:         */
1.22      schwarze 1823:
1.183     schwarze 1824:        if (tok == ROFF_de || tok == ROFF_dei) {
1.86      schwarze 1825:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183     schwarze 1826:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.187     schwarze 1827:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   1828:                deftype = ROFFDEF_ANY;
                   1829:                value = roff_getstrn(r, iname, namesz, &deftype);
                   1830:                switch (deftype) {  /* Before appending, ... */
                   1831:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   1832:                        roff_setstrn(&r->strtab, name, namesz,
                   1833:                            value, strlen(value), 0);
                   1834:                        break;
                   1835:                case ROFFDEF_REN: /* call original standard macro. */
                   1836:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1837:                            (int)strlen(value), value);
                   1838:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1839:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   1840:                        free(call);
                   1841:                        break;
                   1842:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   1843:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   1844:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   1845:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1846:                            (int)rsz, rname);
                   1847:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1848:                        free(call);
                   1849:                        free(rname);
                   1850:                        break;
                   1851:                default:
                   1852:                        break;
                   1853:                }
1.183     schwarze 1854:        }
1.2       schwarze 1855:
1.110     schwarze 1856:        if (*cp == '\0')
1.149     schwarze 1857:                return ROFF_IGN;
1.1       schwarze 1858:
1.93      schwarze 1859:        /* Get the custom end marker. */
1.22      schwarze 1860:
1.93      schwarze 1861:        iname = cp;
1.86      schwarze 1862:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1863:
                   1864:        /* Resolve the end marker if it is indirect. */
                   1865:
1.110     schwarze 1866:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1867:                deftype = ROFFDEF_USER;
                   1868:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1869:                if (name == NULL) {
1.93      schwarze 1870:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1871:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1872:                            "%.*s", (int)namesz, iname);
                   1873:                        namesz = 0;
                   1874:                } else
                   1875:                        namesz = strlen(name);
                   1876:        } else
                   1877:                name = iname;
                   1878:
1.86      schwarze 1879:        if (namesz)
                   1880:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1881:
1.110     schwarze 1882:        if (*cp != '\0')
1.90      schwarze 1883:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167     schwarze 1884:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1885:
1.149     schwarze 1886:        return ROFF_IGN;
1.1       schwarze 1887: }
                   1888:
                   1889: static enum rofferr
1.2       schwarze 1890: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1891: {
1.166     schwarze 1892:        enum roff_tok   t;
1.2       schwarze 1893:        int             i, j;
                   1894:
                   1895:        /*
                   1896:         * First check whether a custom macro exists at this level.  If
                   1897:         * it does, then check against it.  This is some of groff's
                   1898:         * stranger behaviours.  If we encountered a custom end-scope
                   1899:         * tag and that tag also happens to be a "real" macro, then we
                   1900:         * need to try interpreting it again as a real macro.  If it's
                   1901:         * not, then return ignore.  Else continue.
                   1902:         */
                   1903:
                   1904:        if (r->last->end) {
1.35      schwarze 1905:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1906:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1907:                                break;
1.1       schwarze 1908:
1.110     schwarze 1909:                if (r->last->end[j] == '\0' &&
                   1910:                    (buf->buf[i] == '\0' ||
                   1911:                     buf->buf[i] == ' ' ||
                   1912:                     buf->buf[i] == '\t')) {
1.2       schwarze 1913:                        roffnode_pop(r);
                   1914:                        roffnode_cleanscope(r);
1.1       schwarze 1915:
1.110     schwarze 1916:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1917:                                i++;
                   1918:
                   1919:                        pos = i;
1.110     schwarze 1920:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1921:                            TOKEN_NONE)
1.149     schwarze 1922:                                return ROFF_RERUN;
                   1923:                        return ROFF_IGN;
1.2       schwarze 1924:                }
1.1       schwarze 1925:        }
                   1926:
1.2       schwarze 1927:        /*
                   1928:         * If we have no custom end-query or lookup failed, then try
                   1929:         * pulling it out of the hashtable.
                   1930:         */
1.1       schwarze 1931:
1.110     schwarze 1932:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1933:
1.110     schwarze 1934:        if (t != ROFF_cblock) {
                   1935:                if (tok != ROFF_ig)
                   1936:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1937:                return ROFF_IGN;
1.16      schwarze 1938:        }
1.1       schwarze 1939:
1.149     schwarze 1940:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1941: }
                   1942:
                   1943: static enum rofferr
                   1944: roff_block_text(ROFF_ARGS)
                   1945: {
                   1946:
1.110     schwarze 1947:        if (tok != ROFF_ig)
                   1948:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1949:
1.149     schwarze 1950:        return ROFF_IGN;
1.2       schwarze 1951: }
                   1952:
                   1953: static enum rofferr
                   1954: roff_cond_sub(ROFF_ARGS)
                   1955: {
1.166     schwarze 1956:        enum roff_tok    t;
1.37      schwarze 1957:        char            *ep;
1.71      schwarze 1958:        int              rr;
1.2       schwarze 1959:
                   1960:        rr = r->last->rule;
1.37      schwarze 1961:        roffnode_cleanscope(r);
1.39      schwarze 1962:
1.69      schwarze 1963:        /*
                   1964:         * If `\}' occurs on a macro line without a preceding macro,
                   1965:         * drop the line completely.
                   1966:         */
                   1967:
1.110     schwarze 1968:        ep = buf->buf + pos;
                   1969:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1970:                rr = 0;
1.69      schwarze 1971:
1.50      schwarze 1972:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1973:
1.110     schwarze 1974:        while ((ep = strchr(ep, '\\')) != NULL) {
1.190     schwarze 1975:                switch (ep[1]) {
                   1976:                case '}':
                   1977:                        memmove(ep, ep + 2, strlen(ep + 2) + 1);
                   1978:                        roff_ccond(r, ln, ep - buf->buf);
                   1979:                        break;
                   1980:                case '\0':
                   1981:                        ++ep;
                   1982:                        break;
                   1983:                default:
                   1984:                        ep += 2;
                   1985:                        break;
1.70      schwarze 1986:                }
1.50      schwarze 1987:        }
1.190     schwarze 1988:
                   1989:        /*
                   1990:         * Fully handle known macros when they are structurally
                   1991:         * required or when the conditional evaluated to true.
                   1992:         */
                   1993:
                   1994:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1995:        return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
                   1996:            ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
                   1997:            ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1998: }
                   1999:
                   2000: static enum rofferr
                   2001: roff_cond_text(ROFF_ARGS)
                   2002: {
1.37      schwarze 2003:        char            *ep;
1.71      schwarze 2004:        int              rr;
1.2       schwarze 2005:
                   2006:        rr = r->last->rule;
1.37      schwarze 2007:        roffnode_cleanscope(r);
1.1       schwarze 2008:
1.110     schwarze 2009:        ep = buf->buf + pos;
                   2010:        while ((ep = strchr(ep, '\\')) != NULL) {
                   2011:                if (*(++ep) == '}') {
1.70      schwarze 2012:                        *ep = '&';
1.110     schwarze 2013:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 2014:                }
1.119     schwarze 2015:                if (*ep != '\0')
                   2016:                        ++ep;
1.2       schwarze 2017:        }
1.149     schwarze 2018:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 2019: }
                   2020:
1.138     schwarze 2021: /* --- handling of numeric and conditional expressions -------------------- */
                   2022:
1.77      schwarze 2023: /*
                   2024:  * Parse a single signed integer number.  Stop at the first non-digit.
                   2025:  * If there is at least one digit, return success and advance the
                   2026:  * parse point, else return failure and let the parse point unchanged.
                   2027:  * Ignore overflows, treat them just like the C language.
                   2028:  */
1.56      schwarze 2029: static int
1.133     schwarze 2030: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 2031: {
1.133     schwarze 2032:        int      myres, scaled, n, p;
1.79      schwarze 2033:
                   2034:        if (NULL == res)
                   2035:                res = &myres;
1.56      schwarze 2036:
                   2037:        p = *pos;
                   2038:        n = v[p] == '-';
1.133     schwarze 2039:        if (n || v[p] == '+')
1.56      schwarze 2040:                p++;
                   2041:
1.133     schwarze 2042:        if (flags & ROFFNUM_WHITE)
                   2043:                while (isspace((unsigned char)v[p]))
                   2044:                        p++;
                   2045:
1.56      schwarze 2046:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 2047:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 2048:        if (p == *pos + n)
                   2049:                return 0;
                   2050:
                   2051:        if (n)
                   2052:                *res = -*res;
                   2053:
1.126     schwarze 2054:        /* Each number may be followed by one optional scaling unit. */
                   2055:
                   2056:        switch (v[p]) {
                   2057:        case 'f':
1.133     schwarze 2058:                scaled = *res * 65536;
1.126     schwarze 2059:                break;
                   2060:        case 'i':
1.133     schwarze 2061:                scaled = *res * 240;
1.126     schwarze 2062:                break;
                   2063:        case 'c':
1.133     schwarze 2064:                scaled = *res * 240 / 2.54;
1.126     schwarze 2065:                break;
                   2066:        case 'v':
                   2067:        case 'P':
1.133     schwarze 2068:                scaled = *res * 40;
1.126     schwarze 2069:                break;
                   2070:        case 'm':
                   2071:        case 'n':
1.133     schwarze 2072:                scaled = *res * 24;
1.126     schwarze 2073:                break;
                   2074:        case 'p':
1.133     schwarze 2075:                scaled = *res * 10 / 3;
1.126     schwarze 2076:                break;
                   2077:        case 'u':
1.133     schwarze 2078:                scaled = *res;
1.126     schwarze 2079:                break;
                   2080:        case 'M':
1.133     schwarze 2081:                scaled = *res * 6 / 25;
1.126     schwarze 2082:                break;
                   2083:        default:
1.133     schwarze 2084:                scaled = *res;
1.126     schwarze 2085:                p--;
                   2086:                break;
                   2087:        }
1.133     schwarze 2088:        if (flags & ROFFNUM_SCALE)
                   2089:                *res = scaled;
1.126     schwarze 2090:
                   2091:        *pos = p + 1;
1.149     schwarze 2092:        return 1;
1.56      schwarze 2093: }
                   2094:
1.71      schwarze 2095: /*
                   2096:  * Evaluate a string comparison condition.
                   2097:  * The first character is the delimiter.
                   2098:  * Succeed if the string up to its second occurrence
                   2099:  * matches the string up to its third occurence.
                   2100:  * Advance the cursor after the third occurrence
                   2101:  * or lacking that, to the end of the line.
                   2102:  */
                   2103: static int
                   2104: roff_evalstrcond(const char *v, int *pos)
                   2105: {
                   2106:        const char      *s1, *s2, *s3;
                   2107:        int              match;
                   2108:
                   2109:        match = 0;
                   2110:        s1 = v + *pos;          /* initial delimiter */
                   2111:        s2 = s1 + 1;            /* for scanning the first string */
                   2112:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2113:
                   2114:        if (NULL == s3)         /* found no middle delimiter */
                   2115:                goto out;
                   2116:
                   2117:        while ('\0' != *++s3) {
                   2118:                if (*s2 != *s3) {  /* mismatch */
                   2119:                        s3 = strchr(s3, *s1);
                   2120:                        break;
                   2121:                }
                   2122:                if (*s3 == *s1) {  /* found the final delimiter */
                   2123:                        match = 1;
                   2124:                        break;
                   2125:                }
                   2126:                s2++;
                   2127:        }
                   2128:
                   2129: out:
                   2130:        if (NULL == s3)
                   2131:                s3 = strchr(s2, '\0');
1.114     schwarze 2132:        else if (*s3 != '\0')
1.71      schwarze 2133:                s3++;
                   2134:        *pos = s3 - v;
1.149     schwarze 2135:        return match;
1.71      schwarze 2136: }
                   2137:
1.77      schwarze 2138: /*
                   2139:  * Evaluate an optionally negated single character, numerical,
                   2140:  * or string condition.
                   2141:  */
1.71      schwarze 2142: static int
1.143     schwarze 2143: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 2144: {
1.143     schwarze 2145:        char    *cp, *name;
                   2146:        size_t   sz;
1.187     schwarze 2147:        int      deftype, number, savepos, istrue, wanttrue;
1.5       schwarze 2148:
1.71      schwarze 2149:        if ('!' == v[*pos]) {
                   2150:                wanttrue = 0;
                   2151:                (*pos)++;
                   2152:        } else
                   2153:                wanttrue = 1;
                   2154:
1.5       schwarze 2155:        switch (v[*pos]) {
1.112     schwarze 2156:        case '\0':
1.149     schwarze 2157:                return 0;
1.80      schwarze 2158:        case 'n':
                   2159:        case 'o':
1.5       schwarze 2160:                (*pos)++;
1.149     schwarze 2161:                return wanttrue;
1.80      schwarze 2162:        case 'c':
                   2163:        case 'e':
                   2164:        case 't':
1.111     schwarze 2165:        case 'v':
1.5       schwarze 2166:                (*pos)++;
1.149     schwarze 2167:                return !wanttrue;
1.182     schwarze 2168:        case 'd':
1.143     schwarze 2169:        case 'r':
1.182     schwarze 2170:                cp = v + *pos + 1;
                   2171:                while (*cp == ' ')
                   2172:                        cp++;
                   2173:                name = cp;
                   2174:                sz = roff_getname(r, &cp, ln, cp - v);
1.187     schwarze 2175:                if (sz == 0)
                   2176:                        istrue = 0;
                   2177:                else if (v[*pos] == 'r')
                   2178:                        istrue = roff_hasregn(r, name, sz);
                   2179:                else {
                   2180:                        deftype = ROFFDEF_ANY;
                   2181:                        roff_getstrn(r, name, sz, &deftype);
                   2182:                        istrue = !!deftype;
                   2183:                }
1.143     schwarze 2184:                *pos = cp - v;
1.182     schwarze 2185:                return istrue == wanttrue;
1.5       schwarze 2186:        default:
                   2187:                break;
                   2188:        }
                   2189:
1.113     schwarze 2190:        savepos = *pos;
1.133     schwarze 2191:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 2192:                return (number > 0) == wanttrue;
1.113     schwarze 2193:        else if (*pos == savepos)
1.149     schwarze 2194:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 2195:        else
1.149     schwarze 2196:                return 0;
1.5       schwarze 2197: }
                   2198:
1.2       schwarze 2199: static enum rofferr
1.21      schwarze 2200: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 2201: {
1.30      schwarze 2202:
1.149     schwarze 2203:        return ROFF_IGN;
1.21      schwarze 2204: }
                   2205:
                   2206: static enum rofferr
1.123     schwarze 2207: roff_insec(ROFF_ARGS)
                   2208: {
                   2209:
                   2210:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167     schwarze 2211:            ln, ppos, roff_name[tok]);
1.149     schwarze 2212:        return ROFF_IGN;
1.123     schwarze 2213: }
                   2214:
                   2215: static enum rofferr
                   2216: roff_unsupp(ROFF_ARGS)
                   2217: {
                   2218:
                   2219:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167     schwarze 2220:            ln, ppos, roff_name[tok]);
1.149     schwarze 2221:        return ROFF_IGN;
1.123     schwarze 2222: }
                   2223:
                   2224: static enum rofferr
1.2       schwarze 2225: roff_cond(ROFF_ARGS)
                   2226: {
1.46      schwarze 2227:
                   2228:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2229:
1.80      schwarze 2230:        /*
1.35      schwarze 2231:         * An `.el' has no conditional body: it will consume the value
                   2232:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2233:         * defaults to DENY.
1.35      schwarze 2234:         *
                   2235:         * If we're not an `el', however, then evaluate the conditional.
                   2236:         */
1.1       schwarze 2237:
1.110     schwarze 2238:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2239:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2240:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2241:
1.35      schwarze 2242:        /*
                   2243:         * An if-else will put the NEGATION of the current evaluated
                   2244:         * conditional into the stack of rules.
                   2245:         */
                   2246:
1.110     schwarze 2247:        if (tok == ROFF_ie) {
1.96      schwarze 2248:                if (r->rstackpos + 1 == r->rstacksz) {
                   2249:                        r->rstacksz += 16;
                   2250:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2251:                            r->rstacksz, sizeof(int));
1.35      schwarze 2252:                }
1.71      schwarze 2253:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2254:        }
1.5       schwarze 2255:
                   2256:        /* If the parent has false as its rule, then so do we. */
                   2257:
1.71      schwarze 2258:        if (r->last->parent && !r->last->parent->rule)
                   2259:                r->last->rule = 0;
1.5       schwarze 2260:
                   2261:        /*
1.46      schwarze 2262:         * Determine scope.
                   2263:         * If there is nothing on the line after the conditional,
                   2264:         * not even whitespace, use next-line scope.
1.5       schwarze 2265:         */
1.2       schwarze 2266:
1.110     schwarze 2267:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2268:                r->last->endspan = 2;
                   2269:                goto out;
                   2270:        }
                   2271:
1.110     schwarze 2272:        while (buf->buf[pos] == ' ')
1.46      schwarze 2273:                pos++;
                   2274:
                   2275:        /* An opening brace requests multiline scope. */
1.2       schwarze 2276:
1.110     schwarze 2277:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2278:                r->last->endspan = -1;
                   2279:                pos += 2;
1.144     schwarze 2280:                while (buf->buf[pos] == ' ')
                   2281:                        pos++;
1.46      schwarze 2282:                goto out;
1.80      schwarze 2283:        }
1.2       schwarze 2284:
                   2285:        /*
1.46      schwarze 2286:         * Anything else following the conditional causes
                   2287:         * single-line scope.  Warn if the scope contains
                   2288:         * nothing but trailing whitespace.
1.2       schwarze 2289:         */
                   2290:
1.110     schwarze 2291:        if (buf->buf[pos] == '\0')
1.89      schwarze 2292:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167     schwarze 2293:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2294:
1.46      schwarze 2295:        r->last->endspan = 1;
1.1       schwarze 2296:
1.46      schwarze 2297: out:
1.2       schwarze 2298:        *offs = pos;
1.149     schwarze 2299:        return ROFF_RERUN;
1.1       schwarze 2300: }
                   2301:
1.2       schwarze 2302: static enum rofferr
1.7       schwarze 2303: roff_ds(ROFF_ARGS)
                   2304: {
1.85      schwarze 2305:        char            *string;
                   2306:        const char      *name;
                   2307:        size_t           namesz;
1.10      schwarze 2308:
1.123     schwarze 2309:        /* Ignore groff compatibility mode for now. */
                   2310:
                   2311:        if (tok == ROFF_ds1)
                   2312:                tok = ROFF_ds;
                   2313:        else if (tok == ROFF_as1)
                   2314:                tok = ROFF_as;
                   2315:
1.10      schwarze 2316:        /*
1.85      schwarze 2317:         * The first word is the name of the string.
                   2318:         * If it is empty or terminated by an escape sequence,
                   2319:         * abort the `ds' request without defining anything.
1.10      schwarze 2320:         */
1.7       schwarze 2321:
1.110     schwarze 2322:        name = string = buf->buf + pos;
                   2323:        if (*name == '\0')
1.149     schwarze 2324:                return ROFF_IGN;
1.7       schwarze 2325:
1.85      schwarze 2326:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2327:        if (name[namesz] == '\\')
1.149     schwarze 2328:                return ROFF_IGN;
1.85      schwarze 2329:
                   2330:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2331:        if (*string == '"')
1.7       schwarze 2332:                string++;
                   2333:
1.10      schwarze 2334:        /* The rest is the value. */
1.85      schwarze 2335:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2336:            ROFF_as == tok);
1.183     schwarze 2337:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149     schwarze 2338:        return ROFF_IGN;
1.7       schwarze 2339: }
                   2340:
1.77      schwarze 2341: /*
                   2342:  * Parse a single operator, one or two characters long.
                   2343:  * If the operator is recognized, return success and advance the
                   2344:  * parse point, else return failure and let the parse point unchanged.
                   2345:  */
                   2346: static int
                   2347: roff_getop(const char *v, int *pos, char *res)
                   2348: {
                   2349:
                   2350:        *res = v[*pos];
                   2351:
                   2352:        switch (*res) {
1.80      schwarze 2353:        case '+':
                   2354:        case '-':
                   2355:        case '*':
                   2356:        case '/':
                   2357:        case '%':
                   2358:        case '&':
                   2359:        case ':':
1.77      schwarze 2360:                break;
                   2361:        case '<':
                   2362:                switch (v[*pos + 1]) {
1.80      schwarze 2363:                case '=':
1.77      schwarze 2364:                        *res = 'l';
                   2365:                        (*pos)++;
                   2366:                        break;
1.80      schwarze 2367:                case '>':
1.77      schwarze 2368:                        *res = '!';
                   2369:                        (*pos)++;
                   2370:                        break;
1.80      schwarze 2371:                case '?':
1.77      schwarze 2372:                        *res = 'i';
                   2373:                        (*pos)++;
                   2374:                        break;
                   2375:                default:
                   2376:                        break;
                   2377:                }
                   2378:                break;
                   2379:        case '>':
                   2380:                switch (v[*pos + 1]) {
1.80      schwarze 2381:                case '=':
1.77      schwarze 2382:                        *res = 'g';
                   2383:                        (*pos)++;
                   2384:                        break;
1.80      schwarze 2385:                case '?':
1.77      schwarze 2386:                        *res = 'a';
                   2387:                        (*pos)++;
                   2388:                        break;
                   2389:                default:
                   2390:                        break;
                   2391:                }
                   2392:                break;
                   2393:        case '=':
                   2394:                if ('=' == v[*pos + 1])
                   2395:                        (*pos)++;
                   2396:                break;
                   2397:        default:
1.149     schwarze 2398:                return 0;
1.77      schwarze 2399:        }
                   2400:        (*pos)++;
                   2401:
1.149     schwarze 2402:        return *res;
1.77      schwarze 2403: }
                   2404:
                   2405: /*
                   2406:  * Evaluate either a parenthesized numeric expression
                   2407:  * or a single signed integer number.
                   2408:  */
                   2409: static int
1.106     schwarze 2410: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2411:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2412: {
                   2413:
                   2414:        if ('(' != v[*pos])
1.149     schwarze 2415:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2416:
                   2417:        (*pos)++;
1.133     schwarze 2418:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2419:                return 0;
1.77      schwarze 2420:
1.79      schwarze 2421:        /*
                   2422:         * Omission of the closing parenthesis
                   2423:         * is an error in validation mode,
                   2424:         * but ignored in evaluation mode.
                   2425:         */
                   2426:
1.77      schwarze 2427:        if (')' == v[*pos])
                   2428:                (*pos)++;
1.79      schwarze 2429:        else if (NULL == res)
1.149     schwarze 2430:                return 0;
1.77      schwarze 2431:
1.149     schwarze 2432:        return 1;
1.77      schwarze 2433: }
                   2434:
                   2435: /*
                   2436:  * Evaluate a complete numeric expression.
                   2437:  * Proceed left to right, there is no concept of precedence.
                   2438:  */
                   2439: static int
1.106     schwarze 2440: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2441:        int *pos, int *res, int flags)
1.77      schwarze 2442: {
                   2443:        int              mypos, operand2;
                   2444:        char             operator;
                   2445:
                   2446:        if (NULL == pos) {
                   2447:                mypos = 0;
                   2448:                pos = &mypos;
                   2449:        }
                   2450:
1.133     schwarze 2451:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2452:                while (isspace((unsigned char)v[*pos]))
                   2453:                        (*pos)++;
                   2454:
1.133     schwarze 2455:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2456:                return 0;
1.77      schwarze 2457:
                   2458:        while (1) {
1.133     schwarze 2459:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2460:                        while (isspace((unsigned char)v[*pos]))
                   2461:                                (*pos)++;
                   2462:
                   2463:                if ( ! roff_getop(v, pos, &operator))
                   2464:                        break;
                   2465:
1.133     schwarze 2466:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2467:                        while (isspace((unsigned char)v[*pos]))
                   2468:                                (*pos)++;
                   2469:
1.133     schwarze 2470:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2471:                        return 0;
1.77      schwarze 2472:
1.133     schwarze 2473:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2474:                        while (isspace((unsigned char)v[*pos]))
                   2475:                                (*pos)++;
1.79      schwarze 2476:
                   2477:                if (NULL == res)
                   2478:                        continue;
1.77      schwarze 2479:
                   2480:                switch (operator) {
1.80      schwarze 2481:                case '+':
1.77      schwarze 2482:                        *res += operand2;
                   2483:                        break;
1.80      schwarze 2484:                case '-':
1.77      schwarze 2485:                        *res -= operand2;
                   2486:                        break;
1.80      schwarze 2487:                case '*':
1.77      schwarze 2488:                        *res *= operand2;
                   2489:                        break;
1.80      schwarze 2490:                case '/':
1.116     schwarze 2491:                        if (operand2 == 0) {
1.106     schwarze 2492:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2493:                                        r->parse, ln, *pos, v);
                   2494:                                *res = 0;
                   2495:                                break;
                   2496:                        }
1.77      schwarze 2497:                        *res /= operand2;
                   2498:                        break;
1.80      schwarze 2499:                case '%':
1.116     schwarze 2500:                        if (operand2 == 0) {
                   2501:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2502:                                        r->parse, ln, *pos, v);
                   2503:                                *res = 0;
                   2504:                                break;
                   2505:                        }
1.77      schwarze 2506:                        *res %= operand2;
                   2507:                        break;
1.80      schwarze 2508:                case '<':
1.77      schwarze 2509:                        *res = *res < operand2;
                   2510:                        break;
1.80      schwarze 2511:                case '>':
1.77      schwarze 2512:                        *res = *res > operand2;
                   2513:                        break;
1.80      schwarze 2514:                case 'l':
1.77      schwarze 2515:                        *res = *res <= operand2;
                   2516:                        break;
1.80      schwarze 2517:                case 'g':
1.77      schwarze 2518:                        *res = *res >= operand2;
                   2519:                        break;
1.80      schwarze 2520:                case '=':
1.77      schwarze 2521:                        *res = *res == operand2;
                   2522:                        break;
1.80      schwarze 2523:                case '!':
1.77      schwarze 2524:                        *res = *res != operand2;
                   2525:                        break;
1.80      schwarze 2526:                case '&':
1.77      schwarze 2527:                        *res = *res && operand2;
                   2528:                        break;
1.80      schwarze 2529:                case ':':
1.77      schwarze 2530:                        *res = *res || operand2;
                   2531:                        break;
1.80      schwarze 2532:                case 'i':
1.77      schwarze 2533:                        if (operand2 < *res)
                   2534:                                *res = operand2;
                   2535:                        break;
1.80      schwarze 2536:                case 'a':
1.77      schwarze 2537:                        if (operand2 > *res)
                   2538:                                *res = operand2;
                   2539:                        break;
                   2540:                default:
                   2541:                        abort();
                   2542:                }
                   2543:        }
1.149     schwarze 2544:        return 1;
1.77      schwarze 2545: }
                   2546:
1.138     schwarze 2547: /* --- register management ------------------------------------------------ */
                   2548:
1.52      schwarze 2549: void
1.60      schwarze 2550: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2551: {
1.199     schwarze 2552:        roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.198     schwarze 2553: }
                   2554:
                   2555: static void
                   2556: roff_setregn(struct roff *r, const char *name, size_t len,
1.199     schwarze 2557:     int val, char sign, int step)
1.198     schwarze 2558: {
1.52      schwarze 2559:        struct roffreg  *reg;
                   2560:
                   2561:        /* Search for an existing register with the same name. */
                   2562:        reg = r->regtab;
                   2563:
1.198     schwarze 2564:        while (reg != NULL && (reg->key.sz != len ||
                   2565:            strncmp(reg->key.p, name, len) != 0))
1.52      schwarze 2566:                reg = reg->next;
1.41      schwarze 2567:
1.52      schwarze 2568:        if (NULL == reg) {
                   2569:                /* Create a new register. */
                   2570:                reg = mandoc_malloc(sizeof(struct roffreg));
1.198     schwarze 2571:                reg->key.p = mandoc_strndup(name, len);
                   2572:                reg->key.sz = len;
1.60      schwarze 2573:                reg->val = 0;
1.199     schwarze 2574:                reg->step = 0;
1.52      schwarze 2575:                reg->next = r->regtab;
                   2576:                r->regtab = reg;
                   2577:        }
                   2578:
1.60      schwarze 2579:        if ('+' == sign)
                   2580:                reg->val += val;
                   2581:        else if ('-' == sign)
                   2582:                reg->val -= val;
                   2583:        else
                   2584:                reg->val = val;
1.199     schwarze 2585:        if (step != INT_MIN)
                   2586:                reg->step = step;
1.41      schwarze 2587: }
                   2588:
1.65      schwarze 2589: /*
                   2590:  * Handle some predefined read-only number registers.
                   2591:  * For now, return -1 if the requested register is not predefined;
                   2592:  * in case a predefined read-only register having the value -1
                   2593:  * were to turn up, another special value would have to be chosen.
                   2594:  */
                   2595: static int
1.145     schwarze 2596: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2597: {
                   2598:
                   2599:        switch (*name) {
1.145     schwarze 2600:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2601:                return r->argc;
1.80      schwarze 2602:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2603:                return 0;
1.80      schwarze 2604:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2605:                return 1;
1.80      schwarze 2606:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2607:                return 24;
1.80      schwarze 2608:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2609:                return 0;
1.80      schwarze 2610:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2611:                return 1;
1.80      schwarze 2612:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2613:                return 40;
1.65      schwarze 2614:        default:
1.149     schwarze 2615:                return -1;
1.65      schwarze 2616:        }
                   2617: }
                   2618:
1.53      schwarze 2619: int
1.198     schwarze 2620: roff_getreg(struct roff *r, const char *name)
1.41      schwarze 2621: {
1.199     schwarze 2622:        return roff_getregn(r, name, strlen(name), '\0');
1.53      schwarze 2623: }
                   2624:
                   2625: static int
1.199     schwarze 2626: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.53      schwarze 2627: {
                   2628:        struct roffreg  *reg;
1.65      schwarze 2629:        int              val;
                   2630:
                   2631:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2632:                val = roff_getregro(r, name + 1);
1.65      schwarze 2633:                if (-1 != val)
1.149     schwarze 2634:                        return val;
1.65      schwarze 2635:        }
1.53      schwarze 2636:
1.199     schwarze 2637:        for (reg = r->regtab; reg; reg = reg->next) {
1.53      schwarze 2638:                if (len == reg->key.sz &&
1.199     schwarze 2639:                    0 == strncmp(name, reg->key.p, len)) {
                   2640:                        switch (sign) {
                   2641:                        case '+':
                   2642:                                reg->val += reg->step;
                   2643:                                break;
                   2644:                        case '-':
                   2645:                                reg->val -= reg->step;
                   2646:                                break;
                   2647:                        default:
                   2648:                                break;
                   2649:                        }
1.149     schwarze 2650:                        return reg->val;
1.199     schwarze 2651:                }
                   2652:        }
1.143     schwarze 2653:
1.199     schwarze 2654:        roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.149     schwarze 2655:        return 0;
1.143     schwarze 2656: }
                   2657:
                   2658: static int
                   2659: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2660: {
                   2661:        struct roffreg  *reg;
                   2662:        int              val;
                   2663:
                   2664:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2665:                val = roff_getregro(r, name + 1);
1.143     schwarze 2666:                if (-1 != val)
1.149     schwarze 2667:                        return 1;
1.143     schwarze 2668:        }
                   2669:
                   2670:        for (reg = r->regtab; reg; reg = reg->next)
                   2671:                if (len == reg->key.sz &&
                   2672:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2673:                        return 1;
1.41      schwarze 2674:
1.149     schwarze 2675:        return 0;
1.41      schwarze 2676: }
                   2677:
1.52      schwarze 2678: static void
                   2679: roff_freereg(struct roffreg *reg)
1.41      schwarze 2680: {
1.52      schwarze 2681:        struct roffreg  *old_reg;
1.41      schwarze 2682:
1.52      schwarze 2683:        while (NULL != reg) {
                   2684:                free(reg->key.p);
                   2685:                old_reg = reg;
                   2686:                reg = reg->next;
                   2687:                free(old_reg);
                   2688:        }
1.41      schwarze 2689: }
1.7       schwarze 2690:
                   2691: static enum rofferr
1.6       schwarze 2692: roff_nr(ROFF_ARGS)
1.1       schwarze 2693: {
1.199     schwarze 2694:        char            *key, *val, *step;
1.85      schwarze 2695:        size_t           keysz;
1.199     schwarze 2696:        int              iv, is, len;
1.60      schwarze 2697:        char             sign;
1.6       schwarze 2698:
1.110     schwarze 2699:        key = val = buf->buf + pos;
                   2700:        if (*key == '\0')
1.149     schwarze 2701:                return ROFF_IGN;
1.85      schwarze 2702:
                   2703:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2704:        if (key[keysz] == '\\')
1.149     schwarze 2705:                return ROFF_IGN;
1.6       schwarze 2706:
1.60      schwarze 2707:        sign = *val;
1.110     schwarze 2708:        if (sign == '+' || sign == '-')
1.60      schwarze 2709:                val++;
                   2710:
1.199     schwarze 2711:        len = 0;
                   2712:        if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
                   2713:                return ROFF_IGN;
                   2714:
                   2715:        step = val + len;
                   2716:        while (isspace((unsigned char)*step))
                   2717:                step++;
                   2718:        if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
                   2719:                is = INT_MIN;
1.1       schwarze 2720:
1.199     schwarze 2721:        roff_setregn(r, key, keysz, iv, sign, is);
1.149     schwarze 2722:        return ROFF_IGN;
1.76      schwarze 2723: }
                   2724:
                   2725: static enum rofferr
                   2726: roff_rr(ROFF_ARGS)
                   2727: {
                   2728:        struct roffreg  *reg, **prev;
1.85      schwarze 2729:        char            *name, *cp;
                   2730:        size_t           namesz;
1.76      schwarze 2731:
1.110     schwarze 2732:        name = cp = buf->buf + pos;
                   2733:        if (*name == '\0')
1.149     schwarze 2734:                return ROFF_IGN;
1.85      schwarze 2735:        namesz = roff_getname(r, &cp, ln, pos);
                   2736:        name[namesz] = '\0';
1.76      schwarze 2737:
                   2738:        prev = &r->regtab;
                   2739:        while (1) {
                   2740:                reg = *prev;
1.110     schwarze 2741:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2742:                        break;
                   2743:                prev = &reg->next;
                   2744:        }
1.110     schwarze 2745:        if (reg != NULL) {
1.76      schwarze 2746:                *prev = reg->next;
                   2747:                free(reg->key.p);
                   2748:                free(reg);
                   2749:        }
1.149     schwarze 2750:        return ROFF_IGN;
1.29      schwarze 2751: }
                   2752:
1.138     schwarze 2753: /* --- handler functions for roff requests -------------------------------- */
                   2754:
1.29      schwarze 2755: static enum rofferr
                   2756: roff_rm(ROFF_ARGS)
                   2757: {
                   2758:        const char       *name;
                   2759:        char             *cp;
1.85      schwarze 2760:        size_t            namesz;
1.29      schwarze 2761:
1.110     schwarze 2762:        cp = buf->buf + pos;
                   2763:        while (*cp != '\0') {
1.85      schwarze 2764:                name = cp;
1.110     schwarze 2765:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2766:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183     schwarze 2767:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.110     schwarze 2768:                if (name[namesz] == '\\')
1.85      schwarze 2769:                        break;
1.29      schwarze 2770:        }
1.149     schwarze 2771:        return ROFF_IGN;
1.51      schwarze 2772: }
                   2773:
                   2774: static enum rofferr
                   2775: roff_it(ROFF_ARGS)
                   2776: {
                   2777:        int              iv;
                   2778:
                   2779:        /* Parse the number of lines. */
1.133     schwarze 2780:
                   2781:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2782:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2783:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2784:                return ROFF_IGN;
1.51      schwarze 2785:        }
                   2786:
1.134     schwarze 2787:        while (isspace((unsigned char)buf->buf[pos]))
                   2788:                pos++;
                   2789:
                   2790:        /*
                   2791:         * Arm the input line trap.
                   2792:         * Special-casing "an-trap" is an ugly workaround to cope
                   2793:         * with DocBook stupidly fiddling with man(7) internals.
                   2794:         */
1.133     schwarze 2795:
1.51      schwarze 2796:        roffit_lines = iv;
1.134     schwarze 2797:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2798:            strcmp(buf->buf + pos, "an-trap") ?
                   2799:            buf->buf + pos : "br");
1.149     schwarze 2800:        return ROFF_IGN;
1.47      schwarze 2801: }
                   2802:
                   2803: static enum rofferr
                   2804: roff_Dd(ROFF_ARGS)
                   2805: {
1.187     schwarze 2806:        int              mask;
                   2807:        enum roff_tok    t, te;
1.99      schwarze 2808:
1.187     schwarze 2809:        switch (tok) {
                   2810:        case ROFF_Dd:
                   2811:                tok = MDOC_Dd;
                   2812:                te = MDOC_MAX;
                   2813:                if (r->format == 0)
                   2814:                        r->format = MPARSE_MDOC;
                   2815:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   2816:                break;
                   2817:        case ROFF_TH:
                   2818:                tok = MAN_TH;
                   2819:                te = MAN_MAX;
                   2820:                if (r->format == 0)
                   2821:                        r->format = MPARSE_MAN;
                   2822:                mask = MPARSE_QUICK;
                   2823:                break;
                   2824:        default:
                   2825:                abort();
                   2826:        }
                   2827:        if ((r->options & mask) == 0)
                   2828:                for (t = tok; t < te; t++)
                   2829:                        roff_setstr(r, roff_name[t], NULL, 0);
1.149     schwarze 2830:        return ROFF_CONT;
1.14      schwarze 2831: }
                   2832:
                   2833: static enum rofferr
1.27      schwarze 2834: roff_TE(ROFF_ARGS)
                   2835: {
1.193     schwarze 2836:        if (r->tbl == NULL) {
1.94      schwarze 2837:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2838:                    ln, ppos, "TE");
1.193     schwarze 2839:                return ROFF_IGN;
                   2840:        }
                   2841:        if (tbl_end(r->tbl) == 0) {
                   2842:                r->tbl = NULL;
1.130     schwarze 2843:                free(buf->buf);
                   2844:                buf->buf = mandoc_strdup(".sp");
                   2845:                buf->sz = 4;
1.149     schwarze 2846:                return ROFF_REPARSE;
1.130     schwarze 2847:        }
1.193     schwarze 2848:        r->tbl = NULL;
1.149     schwarze 2849:        return ROFF_IGN;
1.27      schwarze 2850: }
                   2851:
                   2852: static enum rofferr
                   2853: roff_T_(ROFF_ARGS)
                   2854: {
                   2855:
                   2856:        if (NULL == r->tbl)
1.94      schwarze 2857:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2858:                    ln, ppos, "T&");
1.27      schwarze 2859:        else
1.168     schwarze 2860:                tbl_restart(ln, ppos, r->tbl);
1.27      schwarze 2861:
1.149     schwarze 2862:        return ROFF_IGN;
1.27      schwarze 2863: }
                   2864:
1.102     schwarze 2865: /*
                   2866:  * Handle in-line equation delimiters.
                   2867:  */
                   2868: static enum rofferr
1.110     schwarze 2869: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2870: {
1.105     schwarze 2871:        char            *cp1, *cp2;
                   2872:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2873:
1.102     schwarze 2874:        /*
                   2875:         * Outside equations, look for an opening delimiter.
                   2876:         * If we are inside an equation, we already know it is
                   2877:         * in-line, or this function wouldn't have been called;
                   2878:         * so look for a closing delimiter.
                   2879:         */
                   2880:
1.110     schwarze 2881:        cp1 = buf->buf + pos;
1.102     schwarze 2882:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2883:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2884:        if (cp2 == NULL)
1.149     schwarze 2885:                return ROFF_CONT;
1.102     schwarze 2886:
1.105     schwarze 2887:        *cp2++ = '\0';
                   2888:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2889:
                   2890:        /* Handle preceding text, protecting whitespace. */
                   2891:
1.110     schwarze 2892:        if (*buf->buf != '\0') {
1.105     schwarze 2893:                if (r->eqn == NULL)
                   2894:                        bef_pr = "\\&";
                   2895:                bef_nl = "\n";
                   2896:        }
                   2897:
                   2898:        /*
                   2899:         * Prepare replacing the delimiter with an equation macro
                   2900:         * and drop leading white space from the equation.
                   2901:         */
1.102     schwarze 2902:
1.105     schwarze 2903:        if (r->eqn == NULL) {
                   2904:                while (*cp2 == ' ')
                   2905:                        cp2++;
                   2906:                mac = ".EQ";
                   2907:        } else
                   2908:                mac = ".EN";
                   2909:
                   2910:        /* Handle following text, protecting whitespace. */
                   2911:
                   2912:        if (*cp2 != '\0') {
                   2913:                aft_nl = "\n";
                   2914:                if (r->eqn != NULL)
                   2915:                        aft_pr = "\\&";
                   2916:        }
                   2917:
                   2918:        /* Do the actual replacement. */
                   2919:
1.110     schwarze 2920:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2921:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2922:        free(buf->buf);
                   2923:        buf->buf = cp1;
1.102     schwarze 2924:
                   2925:        /* Toggle the in-line state of the eqn subsystem. */
                   2926:
                   2927:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2928:        return ROFF_REPARSE;
1.41      schwarze 2929: }
                   2930:
1.107     schwarze 2931: static enum rofferr
                   2932: roff_EQ(ROFF_ARGS)
1.32      schwarze 2933: {
1.191     schwarze 2934:        struct roff_node        *n;
                   2935:
1.194     schwarze 2936:        if (r->man->macroset == MACROSET_MAN)
                   2937:                man_breakscope(r->man, ROFF_EQ);
1.191     schwarze 2938:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   2939:        if (ln > r->man->last->line)
                   2940:                n->flags |= NODE_LINE;
                   2941:        n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
                   2942:        n->eqn->expectargs = UINT_MAX;
                   2943:        roff_node_append(r->man, n);
                   2944:        r->man->next = ROFF_NEXT_SIBLING;
1.32      schwarze 2945:
1.110     schwarze 2946:        assert(r->eqn == NULL);
1.191     schwarze 2947:        if (r->last_eqn == NULL)
                   2948:                r->last_eqn = eqn_alloc(r->parse);
                   2949:        else
                   2950:                eqn_reset(r->last_eqn);
                   2951:        r->eqn = r->last_eqn;
                   2952:        r->eqn->node = n;
1.41      schwarze 2953:
1.110     schwarze 2954:        if (buf->buf[pos] != '\0')
1.107     schwarze 2955:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2956:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2957:
1.149     schwarze 2958:        return ROFF_IGN;
1.32      schwarze 2959: }
                   2960:
                   2961: static enum rofferr
                   2962: roff_EN(ROFF_ARGS)
                   2963: {
1.191     schwarze 2964:        if (r->eqn != NULL) {
                   2965:                eqn_parse(r->eqn);
                   2966:                r->eqn = NULL;
                   2967:        } else
                   2968:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
                   2969:        if (buf->buf[pos] != '\0')
                   2970:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   2971:                    "EN %s", buf->buf + pos);
1.149     schwarze 2972:        return ROFF_IGN;
1.32      schwarze 2973: }
                   2974:
                   2975: static enum rofferr
1.27      schwarze 2976: roff_TS(ROFF_ARGS)
                   2977: {
1.193     schwarze 2978:        if (r->tbl != NULL) {
1.94      schwarze 2979:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2980:                    ln, ppos, "TS breaks TS");
1.193     schwarze 2981:                tbl_end(r->tbl);
1.27      schwarze 2982:        }
1.193     schwarze 2983:        r->tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2984:        if (r->last_tbl)
1.193     schwarze 2985:                r->last_tbl->next = r->tbl;
1.27      schwarze 2986:        else
1.193     schwarze 2987:                r->first_tbl = r->tbl;
                   2988:        r->last_tbl = r->tbl;
1.169     schwarze 2989:        return ROFF_IGN;
                   2990: }
                   2991:
                   2992: static enum rofferr
                   2993: roff_onearg(ROFF_ARGS)
                   2994: {
                   2995:        struct roff_node        *n;
                   2996:        char                    *cp;
1.177     schwarze 2997:        int                      npos;
1.169     schwarze 2998:
1.174     schwarze 2999:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.192     schwarze 3000:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
                   3001:             tok == ROFF_ti))
1.174     schwarze 3002:                man_breakscope(r->man, tok);
                   3003:
1.181     schwarze 3004:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177     schwarze 3005:                r->man->last = roffce_node;
                   3006:                r->man->next = ROFF_NEXT_SIBLING;
                   3007:        }
                   3008:
1.169     schwarze 3009:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3010:        n = r->man->last;
                   3011:
                   3012:        cp = buf->buf + pos;
                   3013:        if (*cp != '\0') {
                   3014:                while (*cp != '\0' && *cp != ' ')
                   3015:                        cp++;
                   3016:                while (*cp == ' ')
                   3017:                        *cp++ = '\0';
                   3018:                if (*cp != '\0')
                   3019:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS,
                   3020:                            r->parse, ln, cp - buf->buf,
                   3021:                            "%s ... %s", roff_name[tok], cp);
                   3022:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172     schwarze 3023:        }
                   3024:
1.181     schwarze 3025:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   3026:                if (r->man->last->type == ROFFT_ELEM) {
1.177     schwarze 3027:                        roff_word_alloc(r->man, ln, pos, "1");
                   3028:                        r->man->last->flags |= NODE_NOSRC;
                   3029:                }
                   3030:                npos = 0;
                   3031:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   3032:                    &roffce_lines, 0) == 0) {
                   3033:                        mandoc_vmsg(MANDOCERR_CE_NONUM,
                   3034:                            r->parse, ln, pos, "ce %s", buf->buf + pos);
                   3035:                        roffce_lines = 1;
                   3036:                }
                   3037:                if (roffce_lines < 1) {
                   3038:                        r->man->last = r->man->last->parent;
                   3039:                        roffce_node = NULL;
                   3040:                        roffce_lines = 0;
                   3041:                } else
                   3042:                        roffce_node = r->man->last->parent;
                   3043:        } else {
                   3044:                n->flags |= NODE_VALID | NODE_ENDED;
                   3045:                r->man->last = n;
                   3046:        }
                   3047:        n->flags |= NODE_LINE;
1.172     schwarze 3048:        r->man->next = ROFF_NEXT_SIBLING;
                   3049:        return ROFF_IGN;
                   3050: }
                   3051:
                   3052: static enum rofferr
                   3053: roff_manyarg(ROFF_ARGS)
                   3054: {
                   3055:        struct roff_node        *n;
                   3056:        char                    *sp, *ep;
                   3057:
                   3058:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3059:        n = r->man->last;
                   3060:
                   3061:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   3062:                while (*ep != '\0' && *ep != ' ')
                   3063:                        ep++;
                   3064:                while (*ep == ' ')
                   3065:                        *ep++ = '\0';
                   3066:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169     schwarze 3067:        }
                   3068:
                   3069:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3070:        r->man->last = n;
                   3071:        r->man->next = ROFF_NEXT_SIBLING;
1.149     schwarze 3072:        return ROFF_IGN;
1.123     schwarze 3073: }
                   3074:
                   3075: static enum rofferr
1.183     schwarze 3076: roff_als(ROFF_ARGS)
                   3077: {
                   3078:        char            *oldn, *newn, *end, *value;
                   3079:        size_t           oldsz, newsz, valsz;
                   3080:
                   3081:        newn = oldn = buf->buf + pos;
                   3082:        if (*newn == '\0')
                   3083:                return ROFF_IGN;
                   3084:
                   3085:        newsz = roff_getname(r, &oldn, ln, pos);
                   3086:        if (newn[newsz] == '\\' || *oldn == '\0')
                   3087:                return ROFF_IGN;
                   3088:
                   3089:        end = oldn;
                   3090:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3091:        if (oldsz == 0)
                   3092:                return ROFF_IGN;
                   3093:
1.187     schwarze 3094:        valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
                   3095:            (int)oldsz, oldn);
1.183     schwarze 3096:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3097:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3098:        free(value);
                   3099:        return ROFF_IGN;
                   3100: }
                   3101:
                   3102: static enum rofferr
1.168     schwarze 3103: roff_br(ROFF_ARGS)
1.123     schwarze 3104: {
1.174     schwarze 3105:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3106:                man_breakscope(r->man, ROFF_br);
1.168     schwarze 3107:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
                   3108:        if (buf->buf[pos] != '\0')
                   3109:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   3110:                    "%s %s", roff_name[tok], buf->buf + pos);
                   3111:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3112:        r->man->next = ROFF_NEXT_SIBLING;
                   3113:        return ROFF_IGN;
1.27      schwarze 3114: }
                   3115:
                   3116: static enum rofferr
1.48      schwarze 3117: roff_cc(ROFF_ARGS)
                   3118: {
                   3119:        const char      *p;
                   3120:
1.110     schwarze 3121:        p = buf->buf + pos;
1.48      schwarze 3122:
1.110     schwarze 3123:        if (*p == '\0' || (r->control = *p++) == '.')
1.175     schwarze 3124:                r->control = '\0';
1.48      schwarze 3125:
1.110     schwarze 3126:        if (*p != '\0')
1.132     schwarze 3127:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3128:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 3129:
1.149     schwarze 3130:        return ROFF_IGN;
1.48      schwarze 3131: }
                   3132:
                   3133: static enum rofferr
1.175     schwarze 3134: roff_ec(ROFF_ARGS)
                   3135: {
                   3136:        const char      *p;
                   3137:
                   3138:        p = buf->buf + pos;
                   3139:        if (*p == '\0')
                   3140:                r->escape = '\\';
                   3141:        else {
                   3142:                r->escape = *p;
                   3143:                if (*++p != '\0')
                   3144:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3145:                            ln, p - buf->buf, "ec ... %s", p);
                   3146:        }
                   3147:        return ROFF_IGN;
                   3148: }
                   3149:
                   3150: static enum rofferr
                   3151: roff_eo(ROFF_ARGS)
                   3152: {
                   3153:        r->escape = '\0';
                   3154:        if (buf->buf[pos] != '\0')
                   3155:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
                   3156:                    ln, pos, "eo %s", buf->buf + pos);
                   3157:        return ROFF_IGN;
                   3158: }
                   3159:
                   3160: static enum rofferr
1.42      schwarze 3161: roff_tr(ROFF_ARGS)
                   3162: {
                   3163:        const char      *p, *first, *second;
                   3164:        size_t           fsz, ssz;
                   3165:        enum mandoc_esc  esc;
                   3166:
1.110     schwarze 3167:        p = buf->buf + pos;
1.42      schwarze 3168:
1.110     schwarze 3169:        if (*p == '\0') {
1.132     schwarze 3170:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 3171:                return ROFF_IGN;
1.42      schwarze 3172:        }
                   3173:
1.110     schwarze 3174:        while (*p != '\0') {
1.42      schwarze 3175:                fsz = ssz = 1;
                   3176:
                   3177:                first = p++;
1.110     schwarze 3178:                if (*first == '\\') {
1.42      schwarze 3179:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3180:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3181:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3182:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 3183:                                return ROFF_IGN;
1.42      schwarze 3184:                        }
                   3185:                        fsz = (size_t)(p - first);
                   3186:                }
                   3187:
                   3188:                second = p++;
1.110     schwarze 3189:                if (*second == '\\') {
1.42      schwarze 3190:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3191:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3192:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3193:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 3194:                                return ROFF_IGN;
1.42      schwarze 3195:                        }
                   3196:                        ssz = (size_t)(p - second);
1.110     schwarze 3197:                } else if (*second == '\0') {
1.132     schwarze 3198:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   3199:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 3200:                        second = " ";
                   3201:                        p--;
                   3202:                }
                   3203:
                   3204:                if (fsz > 1) {
1.80      schwarze 3205:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3206:                            second, ssz, 0);
1.42      schwarze 3207:                        continue;
                   3208:                }
                   3209:
1.110     schwarze 3210:                if (r->xtab == NULL)
1.80      schwarze 3211:                        r->xtab = mandoc_calloc(128,
                   3212:                            sizeof(struct roffstr));
1.42      schwarze 3213:
                   3214:                free(r->xtab[(int)*first].p);
                   3215:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3216:                r->xtab[(int)*first].sz = ssz;
                   3217:        }
                   3218:
1.149     schwarze 3219:        return ROFF_IGN;
1.42      schwarze 3220: }
                   3221:
                   3222: static enum rofferr
1.178     schwarze 3223: roff_rn(ROFF_ARGS)
                   3224: {
                   3225:        const char      *value;
                   3226:        char            *oldn, *newn, *end;
                   3227:        size_t           oldsz, newsz;
1.187     schwarze 3228:        int              deftype;
1.178     schwarze 3229:
                   3230:        oldn = newn = buf->buf + pos;
                   3231:        if (*oldn == '\0')
                   3232:                return ROFF_IGN;
                   3233:
                   3234:        oldsz = roff_getname(r, &newn, ln, pos);
                   3235:        if (oldn[oldsz] == '\\' || *newn == '\0')
                   3236:                return ROFF_IGN;
                   3237:
                   3238:        end = newn;
                   3239:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3240:        if (newsz == 0)
                   3241:                return ROFF_IGN;
                   3242:
1.187     schwarze 3243:        deftype = ROFFDEF_ANY;
                   3244:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3245:        switch (deftype) {
                   3246:        case ROFFDEF_USER:
1.178     schwarze 3247:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3248:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3249:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.187     schwarze 3250:                break;
                   3251:        case ROFFDEF_PRE:
                   3252:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3253:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3254:                break;
                   3255:        case ROFFDEF_REN:
1.178     schwarze 3256:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3257:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.187     schwarze 3258:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3259:                break;
                   3260:        case ROFFDEF_STD:
1.178     schwarze 3261:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.187     schwarze 3262:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3263:                break;
                   3264:        default:
                   3265:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3266:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3267:                break;
                   3268:        }
1.178     schwarze 3269:        return ROFF_IGN;
                   3270: }
                   3271:
                   3272: static enum rofferr
1.14      schwarze 3273: roff_so(ROFF_ARGS)
                   3274: {
1.121     schwarze 3275:        char *name, *cp;
1.15      schwarze 3276:
1.110     schwarze 3277:        name = buf->buf + pos;
1.97      schwarze 3278:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 3279:
1.22      schwarze 3280:        /*
                   3281:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3282:         * opening anything that's not in our cwd or anything beneath
                   3283:         * it.  Thus, explicitly disallow traversing up the file-system
                   3284:         * or using absolute paths.
                   3285:         */
                   3286:
1.110     schwarze 3287:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 3288:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   3289:                    ".so %s", name);
1.121     schwarze 3290:                buf->sz = mandoc_asprintf(&cp,
                   3291:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3292:                free(buf->buf);
                   3293:                buf->buf = cp;
                   3294:                *offs = 0;
1.149     schwarze 3295:                return ROFF_REPARSE;
1.14      schwarze 3296:        }
                   3297:
                   3298:        *offs = pos;
1.149     schwarze 3299:        return ROFF_SO;
1.7       schwarze 3300: }
                   3301:
1.138     schwarze 3302: /* --- user defined strings and macros ------------------------------------ */
                   3303:
1.16      schwarze 3304: static enum rofferr
                   3305: roff_userdef(ROFF_ARGS)
1.12      schwarze 3306: {
1.183     schwarze 3307:        const char       *arg[16], *ap;
1.16      schwarze 3308:        char             *cp, *n1, *n2;
1.164     schwarze 3309:        int               expand_count, i, ib, ie;
1.135     schwarze 3310:        size_t            asz, rsz;
1.12      schwarze 3311:
1.16      schwarze 3312:        /*
                   3313:         * Collect pointers to macro argument strings
1.61      schwarze 3314:         * and NUL-terminate them.
1.16      schwarze 3315:         */
1.135     schwarze 3316:
1.145     schwarze 3317:        r->argc = 0;
1.110     schwarze 3318:        cp = buf->buf + pos;
1.183     schwarze 3319:        for (i = 0; i < 16; i++) {
1.145     schwarze 3320:                if (*cp == '\0')
                   3321:                        arg[i] = "";
                   3322:                else {
                   3323:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   3324:                        r->argc = i + 1;
                   3325:                }
                   3326:        }
1.16      schwarze 3327:
                   3328:        /*
                   3329:         * Expand macro arguments.
1.12      schwarze 3330:         */
1.135     schwarze 3331:
                   3332:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 3333:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 3334:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 3335:        expand_count = 0;
1.135     schwarze 3336:        while (*cp != '\0') {
                   3337:
                   3338:                /* Scan ahead for the next argument invocation. */
                   3339:
                   3340:                if (*cp++ != '\\')
                   3341:                        continue;
                   3342:                if (*cp++ != '$')
                   3343:                        continue;
1.146     schwarze 3344:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   3345:                        ib = 0;
                   3346:                        ie = r->argc - 1;
                   3347:                } else {  /* \\$1 .. \\$9 insert one argument */
                   3348:                        ib = ie = *cp - '1';
                   3349:                        if (ib < 0 || ib > 8)
                   3350:                                continue;
                   3351:                }
1.135     schwarze 3352:                cp -= 2;
1.164     schwarze 3353:
                   3354:                /*
                   3355:                 * Prevent infinite recursion.
                   3356:                 */
                   3357:
                   3358:                if (cp >= n2)
                   3359:                        expand_count = 1;
                   3360:                else if (++expand_count > EXPAND_LIMIT) {
                   3361:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   3362:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 3363:                        free(buf->buf);
                   3364:                        buf->buf = n1;
1.164     schwarze 3365:                        return ROFF_IGN;
                   3366:                }
1.135     schwarze 3367:
                   3368:                /*
                   3369:                 * Determine the size of the expanded argument,
                   3370:                 * taking escaping of quotes into account.
                   3371:                 */
                   3372:
1.146     schwarze 3373:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   3374:                for (i = ib; i <= ie; i++) {
                   3375:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3376:                                asz++;
                   3377:                                if (*ap == '"')
                   3378:                                        asz += 3;
                   3379:                        }
1.135     schwarze 3380:                }
                   3381:                if (asz != 3) {
                   3382:
                   3383:                        /*
                   3384:                         * Determine the size of the rest of the
                   3385:                         * unexpanded macro, including the NUL.
                   3386:                         */
                   3387:
                   3388:                        rsz = buf->sz - (cp - n1) - 3;
                   3389:
                   3390:                        /*
                   3391:                         * When shrinking, move before
                   3392:                         * releasing the storage.
                   3393:                         */
                   3394:
                   3395:                        if (asz < 3)
                   3396:                                memmove(cp + asz, cp + 3, rsz);
                   3397:
                   3398:                        /*
                   3399:                         * Resize the storage for the macro
                   3400:                         * and readjust the parse pointer.
                   3401:                         */
                   3402:
                   3403:                        buf->sz += asz - 3;
                   3404:                        n2 = mandoc_realloc(n1, buf->sz);
                   3405:                        cp = n2 + (cp - n1);
                   3406:                        n1 = n2;
                   3407:
                   3408:                        /*
                   3409:                         * When growing, make room
                   3410:                         * for the expanded argument.
                   3411:                         */
                   3412:
                   3413:                        if (asz > 3)
                   3414:                                memmove(cp + asz, cp + 3, rsz);
                   3415:                }
                   3416:
                   3417:                /* Copy the expanded argument, escaping quotes. */
                   3418:
                   3419:                n2 = cp;
1.146     schwarze 3420:                for (i = ib; i <= ie; i++) {
                   3421:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3422:                                if (*ap == '"') {
                   3423:                                        memcpy(n2, "\\(dq", 4);
                   3424:                                        n2 += 4;
                   3425:                                } else
                   3426:                                        *n2++ = *ap;
                   3427:                        }
                   3428:                        if (i < ie)
                   3429:                                *n2++ = ' ';
1.16      schwarze 3430:                }
1.12      schwarze 3431:        }
                   3432:
1.16      schwarze 3433:        /*
                   3434:         * Replace the macro invocation
                   3435:         * by the expanded macro.
                   3436:         */
1.135     schwarze 3437:
1.110     schwarze 3438:        free(buf->buf);
                   3439:        buf->buf = n1;
1.120     schwarze 3440:        *offs = 0;
1.16      schwarze 3441:
1.149     schwarze 3442:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3443:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3444: }
1.28      schwarze 3445:
1.178     schwarze 3446: /*
                   3447:  * Calling a high-level macro that was renamed with .rn.
                   3448:  * r->current_string has already been set up by roff_parse().
                   3449:  */
                   3450: static enum rofferr
                   3451: roff_renamed(ROFF_ARGS)
                   3452: {
                   3453:        char    *nbuf;
                   3454:
1.187     schwarze 3455:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3456:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.178     schwarze 3457:        free(buf->buf);
                   3458:        buf->buf = nbuf;
                   3459:        return ROFF_CONT;
                   3460: }
                   3461:
1.85      schwarze 3462: static size_t
1.28      schwarze 3463: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3464: {
                   3465:        char     *name, *cp;
1.85      schwarze 3466:        size_t    namesz;
1.28      schwarze 3467:
                   3468:        name = *cpp;
                   3469:        if ('\0' == *name)
1.149     schwarze 3470:                return 0;
1.28      schwarze 3471:
1.85      schwarze 3472:        /* Read until end of name and terminate it with NUL. */
                   3473:        for (cp = name; 1; cp++) {
                   3474:                if ('\0' == *cp || ' ' == *cp) {
                   3475:                        namesz = cp - name;
                   3476:                        break;
                   3477:                }
1.28      schwarze 3478:                if ('\\' != *cp)
                   3479:                        continue;
1.88      schwarze 3480:                namesz = cp - name;
                   3481:                if ('{' == cp[1] || '}' == cp[1])
                   3482:                        break;
1.28      schwarze 3483:                cp++;
                   3484:                if ('\\' == *cp)
                   3485:                        continue;
1.97      schwarze 3486:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3487:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3488:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3489:                break;
1.28      schwarze 3490:        }
                   3491:
                   3492:        /* Read past spaces. */
                   3493:        while (' ' == *cp)
                   3494:                cp++;
                   3495:
                   3496:        *cpp = cp;
1.149     schwarze 3497:        return namesz;
1.28      schwarze 3498: }
                   3499:
1.16      schwarze 3500: /*
                   3501:  * Store *string into the user-defined string called *name.
                   3502:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3503:  * append == 0: replace mode
                   3504:  * append == 1: single-line append mode
                   3505:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3506:  */
1.8       schwarze 3507: static void
1.16      schwarze 3508: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3509:        int append)
1.7       schwarze 3510: {
1.187     schwarze 3511:        size_t   namesz;
1.42      schwarze 3512:
1.187     schwarze 3513:        namesz = strlen(name);
                   3514:        roff_setstrn(&r->strtab, name, namesz, string,
1.80      schwarze 3515:            string ? strlen(string) : 0, append);
1.187     schwarze 3516:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.42      schwarze 3517: }
                   3518:
                   3519: static void
                   3520: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3521:                const char *string, size_t stringsz, int append)
1.42      schwarze 3522: {
                   3523:        struct roffkv   *n;
                   3524:        char            *c;
                   3525:        int              i;
                   3526:        size_t           oldch, newch;
1.7       schwarze 3527:
1.16      schwarze 3528:        /* Search for an existing string with the same name. */
1.42      schwarze 3529:        n = *r;
                   3530:
1.84      schwarze 3531:        while (n && (namesz != n->key.sz ||
                   3532:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3533:                n = n->next;
1.8       schwarze 3534:
                   3535:        if (NULL == n) {
1.16      schwarze 3536:                /* Create a new string table entry. */
1.42      schwarze 3537:                n = mandoc_malloc(sizeof(struct roffkv));
                   3538:                n->key.p = mandoc_strndup(name, namesz);
                   3539:                n->key.sz = namesz;
                   3540:                n->val.p = NULL;
                   3541:                n->val.sz = 0;
                   3542:                n->next = *r;
                   3543:                *r = n;
1.66      schwarze 3544:        } else if (0 == append) {
1.42      schwarze 3545:                free(n->val.p);
                   3546:                n->val.p = NULL;
                   3547:                n->val.sz = 0;
1.16      schwarze 3548:        }
                   3549:
                   3550:        if (NULL == string)
                   3551:                return;
                   3552:
                   3553:        /*
                   3554:         * One additional byte for the '\n' in multiline mode,
                   3555:         * and one for the terminating '\0'.
                   3556:         */
1.66      schwarze 3557:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3558:
                   3559:        if (NULL == n->val.p) {
                   3560:                n->val.p = mandoc_malloc(newch);
                   3561:                *n->val.p = '\0';
1.16      schwarze 3562:                oldch = 0;
                   3563:        } else {
1.42      schwarze 3564:                oldch = n->val.sz;
                   3565:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3566:        }
                   3567:
                   3568:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3569:        c = n->val.p + (int)oldch;
1.16      schwarze 3570:
                   3571:        /* Append new content to the destination buffer. */
1.42      schwarze 3572:        i = 0;
                   3573:        while (i < (int)stringsz) {
1.16      schwarze 3574:                /*
                   3575:                 * Rudimentary roff copy mode:
                   3576:                 * Handle escaped backslashes.
                   3577:                 */
1.42      schwarze 3578:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3579:                        i++;
                   3580:                *c++ = string[i++];
1.16      schwarze 3581:        }
1.8       schwarze 3582:
1.16      schwarze 3583:        /* Append terminating bytes. */
1.66      schwarze 3584:        if (1 < append)
1.16      schwarze 3585:                *c++ = '\n';
1.42      schwarze 3586:
1.16      schwarze 3587:        *c = '\0';
1.42      schwarze 3588:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3589: }
                   3590:
1.8       schwarze 3591: static const char *
1.197     schwarze 3592: roff_getstrn(struct roff *r, const char *name, size_t len,
1.187     schwarze 3593:     int *deftype)
1.7       schwarze 3594: {
1.187     schwarze 3595:        const struct roffkv     *n;
1.197     schwarze 3596:        int                      found, i;
1.187     schwarze 3597:        enum roff_tok            tok;
                   3598:
1.197     schwarze 3599:        found = 0;
                   3600:        for (n = r->strtab; n != NULL; n = n->next) {
                   3601:                if (strncmp(name, n->key.p, len) != 0 ||
                   3602:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   3603:                        continue;
                   3604:                if (*deftype & ROFFDEF_USER) {
                   3605:                        *deftype = ROFFDEF_USER;
                   3606:                        return n->val.p;
                   3607:                } else {
                   3608:                        found = 1;
                   3609:                        break;
                   3610:                }
                   3611:        }
                   3612:        for (n = r->rentab; n != NULL; n = n->next) {
                   3613:                if (strncmp(name, n->key.p, len) != 0 ||
                   3614:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   3615:                        continue;
                   3616:                if (*deftype & ROFFDEF_REN) {
                   3617:                        *deftype = ROFFDEF_REN;
                   3618:                        return n->val.p;
                   3619:                } else {
                   3620:                        found = 1;
                   3621:                        break;
1.187     schwarze 3622:                }
                   3623:        }
1.197     schwarze 3624:        for (i = 0; i < PREDEFS_MAX; i++) {
                   3625:                if (strncmp(name, predefs[i].name, len) != 0 ||
                   3626:                    predefs[i].name[len] != '\0')
                   3627:                        continue;
                   3628:                if (*deftype & ROFFDEF_PRE) {
                   3629:                        *deftype = ROFFDEF_PRE;
                   3630:                        return predefs[i].str;
                   3631:                } else {
                   3632:                        found = 1;
                   3633:                        break;
1.187     schwarze 3634:                }
                   3635:        }
1.197     schwarze 3636:        if (r->man->macroset != MACROSET_MAN) {
                   3637:                for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
                   3638:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   3639:                            roff_name[tok][len] != '\0')
                   3640:                                continue;
                   3641:                        if (*deftype & ROFFDEF_STD) {
                   3642:                                *deftype = ROFFDEF_STD;
                   3643:                                return NULL;
                   3644:                        } else {
                   3645:                                found = 1;
                   3646:                                break;
1.187     schwarze 3647:                        }
                   3648:                }
                   3649:        }
1.197     schwarze 3650:        if (r->man->macroset != MACROSET_MDOC) {
                   3651:                for (tok = MAN_TH; tok < MAN_MAX; tok++) {
                   3652:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   3653:                            roff_name[tok][len] != '\0')
                   3654:                                continue;
                   3655:                        if (*deftype & ROFFDEF_STD) {
                   3656:                                *deftype = ROFFDEF_STD;
                   3657:                                return NULL;
                   3658:                        } else {
                   3659:                                found = 1;
                   3660:                                break;
1.187     schwarze 3661:                        }
                   3662:                }
1.197     schwarze 3663:        }
                   3664:
                   3665:        if (found == 0 && *deftype != ROFFDEF_ANY) {
                   3666:                if (*deftype & ROFFDEF_REN) {
                   3667:                        /*
                   3668:                         * This might still be a request,
                   3669:                         * so do not treat it as undefined yet.
                   3670:                         */
                   3671:                        *deftype = ROFFDEF_UNDEF;
                   3672:                        return NULL;
1.187     schwarze 3673:                }
1.197     schwarze 3674:
                   3675:                /* Using an undefined string defines it to be empty. */
                   3676:
                   3677:                roff_setstrn(&r->strtab, name, len, "", 0, 0);
                   3678:                roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.187     schwarze 3679:        }
1.197     schwarze 3680:
1.187     schwarze 3681:        *deftype = 0;
1.149     schwarze 3682:        return NULL;
1.7       schwarze 3683: }
                   3684:
1.8       schwarze 3685: static void
1.42      schwarze 3686: roff_freestr(struct roffkv *r)
1.7       schwarze 3687: {
1.42      schwarze 3688:        struct roffkv    *n, *nn;
1.7       schwarze 3689:
1.42      schwarze 3690:        for (n = r; n; n = nn) {
                   3691:                free(n->key.p);
                   3692:                free(n->val.p);
1.7       schwarze 3693:                nn = n->next;
                   3694:                free(n);
                   3695:        }
1.27      schwarze 3696: }
1.138     schwarze 3697:
                   3698: /* --- accessors and utility functions ------------------------------------ */
1.42      schwarze 3699:
                   3700: /*
                   3701:  * Duplicate an input string, making the appropriate character
                   3702:  * conversations (as stipulated by `tr') along the way.
                   3703:  * Returns a heap-allocated string with all the replacements made.
                   3704:  */
                   3705: char *
                   3706: roff_strdup(const struct roff *r, const char *p)
                   3707: {
                   3708:        const struct roffkv *cp;
                   3709:        char            *res;
                   3710:        const char      *pp;
                   3711:        size_t           ssz, sz;
                   3712:        enum mandoc_esc  esc;
                   3713:
                   3714:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3715:                return mandoc_strdup(p);
1.42      schwarze 3716:        else if ('\0' == *p)
1.149     schwarze 3717:                return mandoc_strdup("");
1.42      schwarze 3718:
                   3719:        /*
                   3720:         * Step through each character looking for term matches
                   3721:         * (remember that a `tr' can be invoked with an escape, which is
                   3722:         * a glyph but the escape is multi-character).
                   3723:         * We only do this if the character hash has been initialised
                   3724:         * and the string is >0 length.
                   3725:         */
                   3726:
                   3727:        res = NULL;
                   3728:        ssz = 0;
                   3729:
                   3730:        while ('\0' != *p) {
1.161     schwarze 3731:                assert((unsigned int)*p < 128);
                   3732:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3733:                        sz = r->xtab[(int)*p].sz;
                   3734:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3735:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3736:                        ssz += sz;
                   3737:                        p++;
                   3738:                        continue;
                   3739:                } else if ('\\' != *p) {
                   3740:                        res = mandoc_realloc(res, ssz + 2);
                   3741:                        res[ssz++] = *p++;
                   3742:                        continue;
                   3743:                }
                   3744:
                   3745:                /* Search for term matches. */
                   3746:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3747:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3748:                                break;
                   3749:
                   3750:                if (NULL != cp) {
                   3751:                        /*
                   3752:                         * A match has been found.
                   3753:                         * Append the match to the array and move
                   3754:                         * forward by its keysize.
                   3755:                         */
1.80      schwarze 3756:                        res = mandoc_realloc(res,
                   3757:                            ssz + cp->val.sz + 1);
1.42      schwarze 3758:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3759:                        ssz += cp->val.sz;
                   3760:                        p += (int)cp->key.sz;
                   3761:                        continue;
                   3762:                }
                   3763:
                   3764:                /*
                   3765:                 * Handle escapes carefully: we need to copy
                   3766:                 * over just the escape itself, or else we might
                   3767:                 * do replacements within the escape itself.
                   3768:                 * Make sure to pass along the bogus string.
                   3769:                 */
                   3770:                pp = p++;
                   3771:                esc = mandoc_escape(&p, NULL, NULL);
                   3772:                if (ESCAPE_ERROR == esc) {
                   3773:                        sz = strlen(pp);
                   3774:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3775:                        memcpy(res + ssz, pp, sz);
                   3776:                        break;
                   3777:                }
1.80      schwarze 3778:                /*
                   3779:                 * We bail out on bad escapes.
1.42      schwarze 3780:                 * No need to warn: we already did so when
                   3781:                 * roff_res() was called.
                   3782:                 */
                   3783:                sz = (int)(p - pp);
                   3784:                res = mandoc_realloc(res, ssz + sz + 1);
                   3785:                memcpy(res + ssz, pp, sz);
                   3786:                ssz += sz;
                   3787:        }
                   3788:
                   3789:        res[(int)ssz] = '\0';
1.149     schwarze 3790:        return res;
1.99      schwarze 3791: }
                   3792:
                   3793: int
                   3794: roff_getformat(const struct roff *r)
                   3795: {
                   3796:
1.149     schwarze 3797:        return r->format;
1.48      schwarze 3798: }
                   3799:
                   3800: /*
1.80      schwarze 3801:  * Find out whether a line is a macro line or not.
1.48      schwarze 3802:  * If it is, adjust the current position and return one; if it isn't,
                   3803:  * return zero and don't change the current position.
                   3804:  * If the control character has been set with `.cc', then let that grain
                   3805:  * precedence.
                   3806:  * This is slighly contrary to groff, where using the non-breaking
                   3807:  * control character when `cc' has been invoked will cause the
                   3808:  * non-breaking macro contents to be printed verbatim.
                   3809:  */
                   3810: int
                   3811: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3812: {
                   3813:        int             pos;
                   3814:
                   3815:        pos = *ppos;
                   3816:
1.175     schwarze 3817:        if (r->control != '\0' && cp[pos] == r->control)
1.48      schwarze 3818:                pos++;
1.175     schwarze 3819:        else if (r->control != '\0')
1.149     schwarze 3820:                return 0;
1.48      schwarze 3821:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3822:                pos += 2;
                   3823:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3824:                pos++;
                   3825:        else
1.149     schwarze 3826:                return 0;
1.48      schwarze 3827:
                   3828:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3829:                pos++;
                   3830:
                   3831:        *ppos = pos;
1.149     schwarze 3832:        return 1;
1.1       schwarze 3833: }