[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 1.186

1.186   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.185 2017/06/17 22:40:27 schwarze Exp $ */
1.1       schwarze    2: /*
1.139     schwarze    3:  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.159     schwarze    4:  * Copyright (c) 2010-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
1.1       schwarze    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
                      7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
                      9:  *
1.16      schwarze   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1       schwarze   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16      schwarze   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1       schwarze   13:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     14:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     15:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     16:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
                     17:  */
1.99      schwarze   18: #include <sys/types.h>
                     19:
1.1       schwarze   20: #include <assert.h>
1.3       schwarze   21: #include <ctype.h>
1.117     schwarze   22: #include <limits.h>
1.167     schwarze   23: #include <stddef.h>
                     24: #include <stdint.h>
1.51      schwarze   25: #include <stdio.h>
1.1       schwarze   26: #include <stdlib.h>
                     27: #include <string.h>
                     28:
                     29: #include "mandoc.h"
1.74      schwarze   30: #include "mandoc_aux.h"
1.167     schwarze   31: #include "mandoc_ohash.h"
1.137     schwarze   32: #include "roff.h"
1.98      daniel     33: #include "libmandoc.h"
1.138     schwarze   34: #include "roff_int.h"
1.27      schwarze   35: #include "libroff.h"
1.2       schwarze   36:
1.43      schwarze   37: /* Maximum number of string expansions per line, to break infinite loops. */
                     38: #define        EXPAND_LIMIT    1000
                     39:
1.138     schwarze   40: /* --- data types --------------------------------------------------------- */
                     41:
1.41      schwarze   42: /*
1.42      schwarze   43:  * An incredibly-simple string buffer.
                     44:  */
1.8       schwarze   45: struct roffstr {
1.42      schwarze   46:        char            *p; /* nil-terminated buffer */
                     47:        size_t           sz; /* saved strlen(p) */
                     48: };
                     49:
                     50: /*
                     51:  * A key-value roffstr pair as part of a singly-linked list.
                     52:  */
                     53: struct roffkv {
                     54:        struct roffstr   key;
                     55:        struct roffstr   val;
                     56:        struct roffkv   *next; /* next in list */
1.8       schwarze   57: };
                     58:
1.52      schwarze   59: /*
                     60:  * A single number register as part of a singly-linked list.
                     61:  */
                     62: struct roffreg {
                     63:        struct roffstr   key;
1.53      schwarze   64:        int              val;
1.52      schwarze   65:        struct roffreg  *next;
                     66: };
                     67:
1.167     schwarze   68: /*
                     69:  * Association of request and macro names with token IDs.
                     70:  */
                     71: struct roffreq {
                     72:        enum roff_tok    tok;
                     73:        char             name[];
                     74: };
                     75:
1.1       schwarze   76: struct roff {
1.35      schwarze   77:        struct mparse   *parse; /* parse point */
1.168     schwarze   78:        struct roff_man *man; /* mdoc or man parser */
1.1       schwarze   79:        struct roffnode *last; /* leaf of stack */
1.96      schwarze   80:        int             *rstack; /* stack of inverted `ie' values */
1.167     schwarze   81:        struct ohash    *reqtab; /* request lookup table */
1.52      schwarze   82:        struct roffreg  *regtab; /* number registers */
1.42      schwarze   83:        struct roffkv   *strtab; /* user-defined strings & macros */
1.178     schwarze   84:        struct roffkv   *rentab; /* renamed strings & macros */
1.42      schwarze   85:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     86:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze   87:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   88:        struct tbl_node *first_tbl; /* first table parsed */
                     89:        struct tbl_node *last_tbl; /* last table parsed */
                     90:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze   91:        struct eqn_node *last_eqn; /* last equation parsed */
                     92:        struct eqn_node *first_eqn; /* first equation parsed */
                     93:        struct eqn_node *eqn; /* current equation being parsed */
1.102     schwarze   94:        int              eqn_inline; /* current equation is inline */
1.96      schwarze   95:        int              options; /* parse options */
                     96:        int              rstacksz; /* current size limit of rstack */
                     97:        int              rstackpos; /* position in rstack */
1.99      schwarze   98:        int              format; /* current file in mdoc or man format */
1.145     schwarze   99:        int              argc; /* number of args of the last macro */
1.96      schwarze  100:        char             control; /* control character */
1.175     schwarze  101:        char             escape; /* escape character */
1.1       schwarze  102: };
                    103:
                    104: struct roffnode {
1.166     schwarze  105:        enum roff_tok    tok; /* type of node */
1.1       schwarze  106:        struct roffnode *parent; /* up one in stack */
                    107:        int              line; /* parse line */
                    108:        int              col; /* parse col */
1.16      schwarze  109:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  110:        char            *end; /* end-rules: custom token */
                    111:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  112:        int              rule; /* current evaluation rule */
1.1       schwarze  113: };
                    114:
                    115: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.166     schwarze  116:                         enum roff_tok tok, /* tok of macro */ \
1.110     schwarze  117:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  118:                         int ln, /* parse line */ \
1.2       schwarze  119:                         int ppos, /* original pos in buffer */ \
                    120:                         int pos, /* current pos in buffer */ \
                    121:                         int *offs /* reset offset of buffer data */
1.1       schwarze  122:
                    123: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    124:
                    125: struct roffmac {
1.2       schwarze  126:        roffproc         proc; /* process new macro */
                    127:        roffproc         text; /* process as child text of macro */
                    128:        roffproc         sub; /* process as child of macro */
                    129:        int              flags;
                    130: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.1       schwarze  131: };
                    132:
1.37      schwarze  133: struct predef {
                    134:        const char      *name; /* predefined input name */
                    135:        const char      *str; /* replacement symbol */
                    136: };
                    137:
                    138: #define        PREDEF(__name, __str) \
                    139:        { (__name), (__str) },
                    140:
1.138     schwarze  141: /* --- function prototypes ------------------------------------------------ */
                    142:
1.42      schwarze  143: static void             roffnode_cleanscope(struct roff *);
                    144: static void             roffnode_pop(struct roff *);
1.166     schwarze  145: static void             roffnode_push(struct roff *, enum roff_tok,
1.42      schwarze  146:                                const char *, int, int);
1.183     schwarze  147: static enum rofferr     roff_als(ROFF_ARGS);
1.2       schwarze  148: static enum rofferr     roff_block(ROFF_ARGS);
                    149: static enum rofferr     roff_block_text(ROFF_ARGS);
                    150: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.168     schwarze  151: static enum rofferr     roff_br(ROFF_ARGS);
1.2       schwarze  152: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  153: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  154: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  155: static enum rofferr     roff_cond(ROFF_ARGS);
                    156: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    157: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  158: static enum rofferr     roff_ds(ROFF_ARGS);
1.175     schwarze  159: static enum rofferr     roff_ec(ROFF_ARGS);
                    160: static enum rofferr     roff_eo(ROFF_ARGS);
1.110     schwarze  161: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.143     schwarze  162: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.106     schwarze  163: static int              roff_evalnum(struct roff *, int,
                    164:                                const char *, int *, int *, int);
                    165: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  166:                                const char *, int *, int *, int);
1.71      schwarze  167: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  168: static void             roff_free1(struct roff *);
1.52      schwarze  169: static void             roff_freereg(struct roffreg *);
1.42      schwarze  170: static void             roff_freestr(struct roffkv *);
1.85      schwarze  171: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  172: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  173: static int              roff_getop(const char *, int *, char *);
1.53      schwarze  174: static int              roff_getregn(const struct roff *,
                    175:                                const char *, size_t);
1.145     schwarze  176: static int              roff_getregro(const struct roff *,
                    177:                                const char *name);
1.178     schwarze  178: static const char      *roff_getrenn(const struct roff *,
                    179:                                const char *, size_t);
1.80      schwarze  180: static const char      *roff_getstrn(const struct roff *,
1.8       schwarze  181:                                const char *, size_t);
1.143     schwarze  182: static int              roff_hasregn(const struct roff *,
                    183:                                const char *, size_t);
1.123     schwarze  184: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  185: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  186: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137     schwarze  187: static void             roff_man_alloc1(struct roff_man *);
                    188: static void             roff_man_free1(struct roff_man *);
1.172     schwarze  189: static enum rofferr     roff_manyarg(ROFF_ARGS);
1.6       schwarze  190: static enum rofferr     roff_nr(ROFF_ARGS);
1.169     schwarze  191: static enum rofferr     roff_onearg(ROFF_ARGS);
1.166     schwarze  192: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.87      schwarze  193:                                int, int);
1.177     schwarze  194: static enum rofferr     roff_parsetext(struct roff *, struct buf *,
                    195:                                int, int *);
1.178     schwarze  196: static enum rofferr     roff_renamed(ROFF_ARGS);
1.110     schwarze  197: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  198: static enum rofferr     roff_rm(ROFF_ARGS);
1.178     schwarze  199: static enum rofferr     roff_rn(ROFF_ARGS);
1.76      schwarze  200: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  201: static void             roff_setstr(struct roff *,
1.16      schwarze  202:                                const char *, const char *, int);
1.80      schwarze  203: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  204:                                size_t, const char *, size_t, int);
1.14      schwarze  205: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  206: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  207: static enum rofferr     roff_Dd(ROFF_ARGS);
                    208: static enum rofferr     roff_TH(ROFF_ARGS);
1.27      schwarze  209: static enum rofferr     roff_TE(ROFF_ARGS);
                    210: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  211: static enum rofferr     roff_EQ(ROFF_ARGS);
                    212: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  213: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  214: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  215: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  216:
1.138     schwarze  217: /* --- constant data ------------------------------------------------------ */
                    218:
1.133     schwarze  219: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    220: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    221:
1.166     schwarze  222: const char *__roff_name[MAN_MAX + 1] = {
1.177     schwarze  223:        "br",           "ce",           "ft",           "ll",
1.184     schwarze  224:        "mc",           "po",           "rj",           "sp",
                    225:        "ta",           "ti",           NULL,
1.166     schwarze  226:        "ab",           "ad",           "af",           "aln",
                    227:        "als",          "am",           "am1",          "ami",
                    228:        "ami1",         "as",           "as1",          "asciify",
                    229:        "backtrace",    "bd",           "bleedat",      "blm",
                    230:         "box",         "boxa",         "bp",           "BP",
                    231:        "break",        "breakchar",    "brnl",         "brp",
1.177     schwarze  232:        "brpnl",        "c2",           "cc",
1.166     schwarze  233:        "cf",           "cflags",       "ch",           "char",
                    234:        "chop",         "class",        "close",        "CL",
                    235:        "color",        "composite",    "continue",     "cp",
                    236:        "cropat",       "cs",           "cu",           "da",
                    237:        "dch",          "Dd",           "de",           "de1",
                    238:        "defcolor",     "dei",          "dei1",         "device",
                    239:        "devicem",      "di",           "do",           "ds",
                    240:        "ds1",          "dwh",          "dt",           "ec",
                    241:        "ecr",          "ecs",          "el",           "em",
                    242:        "EN",           "eo",           "EP",           "EQ",
                    243:        "errprint",     "ev",           "evc",          "ex",
                    244:        "fallback",     "fam",          "fc",           "fchar",
                    245:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    246:        "fl",           "flig",         "fp",           "fps",
                    247:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    248:        "fzoom",        "gcolor",       "hc",           "hcode",
                    249:        "hidechar",     "hla",          "hlm",          "hpf",
                    250:        "hpfa",         "hpfcode",      "hw",           "hy",
                    251:        "hylang",       "hylen",        "hym",          "hypp",
                    252:        "hys",          "ie",           "if",           "ig",
                    253:        "index",        "it",           "itc",          "IX",
                    254:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    255:        "lc",           "lc_ctype",     "lds",          "length",
                    256:        "letadj",       "lf",           "lg",           "lhang",
                    257:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.176     schwarze  258:        "ls",           "lsm",          "lt",
1.166     schwarze  259:        "mediasize",    "minss",        "mk",           "mso",
                    260:        "na",           "ne",           "nh",           "nhychar",
                    261:        "nm",           "nn",           "nop",          "nr",
                    262:        "nrf",          "nroff",        "ns",           "nx",
                    263:        "open",         "opena",        "os",           "output",
                    264:        "padj",         "papersize",    "pc",           "pev",
                    265:        "pi",           "PI",           "pl",           "pm",
1.184     schwarze  266:        "pn",           "pnr",          "ps",
1.166     schwarze  267:        "psbb",         "pshape",       "pso",          "ptr",
                    268:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.181     schwarze  269:        "return",       "rfschar",      "rhang",
1.166     schwarze  270:        "rm",           "rn",           "rnn",          "rr",
                    271:        "rs",           "rt",           "schar",        "sentchar",
                    272:        "shc",          "shift",        "sizes",        "so",
                    273:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    274:        "sty",          "substring",    "sv",           "sy",
1.172     schwarze  275:        "T&",           "tc",           "TE",
1.173     schwarze  276:        "TH",           "tkf",          "tl",
1.166     schwarze  277:        "tm",           "tm1",          "tmc",          "tr",
                    278:        "track",        "transchar",    "trf",          "trimat",
                    279:        "trin",         "trnt",         "troff",        "TS",
                    280:        "uf",           "ul",           "unformat",     "unwatch",
                    281:        "unwatchn",     "vpt",          "vs",           "warn",
                    282:        "warnscale",    "watch",        "watchlength",  "watchn",
                    283:        "wh",           "while",        "write",        "writec",
                    284:        "writem",       "xflag",        ".",            NULL,
1.178     schwarze  285:        NULL,           "text",
1.166     schwarze  286:        "Dd",           "Dt",           "Os",           "Sh",
                    287:        "Ss",           "Pp",           "D1",           "Dl",
                    288:        "Bd",           "Ed",           "Bl",           "El",
                    289:        "It",           "Ad",           "An",           "Ap",
                    290:        "Ar",           "Cd",           "Cm",           "Dv",
                    291:        "Er",           "Ev",           "Ex",           "Fa",
                    292:        "Fd",           "Fl",           "Fn",           "Ft",
                    293:        "Ic",           "In",           "Li",           "Nd",
                    294:        "Nm",           "Op",           "Ot",           "Pa",
                    295:        "Rv",           "St",           "Va",           "Vt",
                    296:        "Xr",           "%A",           "%B",           "%D",
                    297:        "%I",           "%J",           "%N",           "%O",
                    298:        "%P",           "%R",           "%T",           "%V",
                    299:        "Ac",           "Ao",           "Aq",           "At",
                    300:        "Bc",           "Bf",           "Bo",           "Bq",
                    301:        "Bsx",          "Bx",           "Db",           "Dc",
                    302:        "Do",           "Dq",           "Ec",           "Ef",
                    303:        "Em",           "Eo",           "Fx",           "Ms",
                    304:        "No",           "Ns",           "Nx",           "Ox",
                    305:        "Pc",           "Pf",           "Po",           "Pq",
                    306:        "Qc",           "Ql",           "Qo",           "Qq",
                    307:        "Re",           "Rs",           "Sc",           "So",
                    308:        "Sq",           "Sm",           "Sx",           "Sy",
                    309:        "Tn",           "Ux",           "Xc",           "Xo",
                    310:        "Fo",           "Fc",           "Oo",           "Oc",
                    311:        "Bk",           "Ek",           "Bt",           "Hf",
                    312:        "Fr",           "Ud",           "Lb",           "Lp",
                    313:        "Lk",           "Mt",           "Brq",          "Bro",
                    314:        "Brc",          "%C",           "Es",           "En",
1.171     schwarze  315:        "Dx",           "%Q",           "%U",           "Ta",
                    316:        NULL,
1.166     schwarze  317:        "TH",           "SH",           "SS",           "TP",
                    318:        "LP",           "PP",           "P",            "IP",
                    319:        "HP",           "SM",           "SB",           "BI",
                    320:        "IB",           "BR",           "RB",           "R",
                    321:        "B",            "I",            "IR",           "RI",
1.171     schwarze  322:        "nf",           "fi",
1.166     schwarze  323:        "RE",           "RS",           "DT",           "UC",
1.169     schwarze  324:        "PD",           "AT",           "in",
1.166     schwarze  325:        "OP",           "EX",           "EE",           "UR",
1.170     schwarze  326:        "UE",           NULL
1.166     schwarze  327: };
                    328: const  char *const *roff_name = __roff_name;
                    329:
                    330: static struct roffmac   roffs[TOKEN_NONE] = {
1.168     schwarze  331:        { roff_br, NULL, NULL, 0 },  /* br */
1.177     schwarze  332:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.169     schwarze  333:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.170     schwarze  334:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.176     schwarze  335:        { roff_onearg, NULL, NULL, 0 },  /* mc */
1.184     schwarze  336:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.181     schwarze  337:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.171     schwarze  338:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.172     schwarze  339:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.173     schwarze  340:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.168     schwarze  341:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.167     schwarze  342:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    343:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    344:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    345:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.183     schwarze  346:        { roff_als, NULL, NULL, 0 },  /* als */
1.167     schwarze  347:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    348:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    349:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    350:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    351:        { roff_ds, NULL, NULL, 0 },  /* as */
                    352:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    353:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    354:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    355:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    356:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    357:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    358:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    359:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    360:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    361:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    362:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    363:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    364:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.168     schwarze  365:        { roff_br, NULL, NULL, 0 },  /* brp */
1.167     schwarze  366:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    367:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    368:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    369:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    370:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    371:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    372:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    373:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    374:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    375:        { roff_insec, NULL, NULL, 0 },  /* close */
                    376:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    377:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    378:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    379:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    380:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    381:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    382:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    383:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    384:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    385:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    386:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    387:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    388:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    389:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    390:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    391:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    392:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    393:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    394:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    395:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    396:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    397:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    398:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.175     schwarze  400:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.167     schwarze  401:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    402:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    403:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    404:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    405:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.175     schwarze  406:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.167     schwarze  407:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    408:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    409:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    410:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    411:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    412:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    413:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    414:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    415:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    416:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    417:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    419:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    421:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    422:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    425:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    429:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    430:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    431:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    432:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    438:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    439:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    440:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    442:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    444:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    445:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    446:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    447:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    448:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    449:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    450:        { roff_it, NULL, NULL, 0 },  /* it */
                    451:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    452:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    453:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    454:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    455:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    456:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    457:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    458:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    459:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    460:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    461:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    462:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    463:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    464:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    465:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    466:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    467:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    468:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    469:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    470:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    471:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    472:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    474:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    475:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    480:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    481:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    482:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    483:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    484:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    487:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    488:        { roff_insec, NULL, NULL, 0 },  /* open */
                    489:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    491:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    492:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    493:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    494:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    495:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    496:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    497:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    500:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    501:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    503:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    504:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    505:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    506:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    508:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    511:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    512:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    513:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    514:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.178     schwarze  515:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.167     schwarze  516:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    517:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    518:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    519:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    520:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    521:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    522:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    523:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    524:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    525:        { roff_so, NULL, NULL, 0 },  /* so */
                    526:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    529:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    530:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    531:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    533:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    534:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    535:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    536:        { roff_TE, NULL, NULL, 0 },  /* TE */
                    537:        { roff_TH, NULL, NULL, 0 },  /* TH */
                    538:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    539:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    540:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    541:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    542:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    543:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    546:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    548:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    549:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    550:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    551:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    552:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    553:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    554:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    555:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    557:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    558:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    559:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    560:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    561:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    562:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    563:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    564:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    565:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    566:        { roff_insec, NULL, NULL, 0 },  /* write */
                    567:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    568:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    569:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    570:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.178     schwarze  571:        { roff_renamed, NULL, NULL, 0 },
1.167     schwarze  572:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  573: };
                    574:
1.73      schwarze  575: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47      schwarze  576: const  char *const __mdoc_reserved[] = {
                    577:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    578:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    579:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    580:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73      schwarze  581:        "Dt", "Dv", "Dx", "D1",
                    582:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    583:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47      schwarze  584:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73      schwarze  585:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    586:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47      schwarze  587:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73      schwarze  588:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    589:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    590:        "Sc", "Sh", "Sm", "So", "Sq",
1.47      schwarze  591:        "Ss", "St", "Sx", "Sy",
                    592:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73      schwarze  593:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47      schwarze  594:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    595:        NULL
                    596: };
                    597:
1.73      schwarze  598: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47      schwarze  599: const  char *const __man_reserved[] = {
1.73      schwarze  600:        "AT", "B", "BI", "BR", "DT",
                    601:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    602:        "LP", "OP", "P", "PD", "PP",
                    603:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    604:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47      schwarze  605:        NULL
                    606: };
                    607:
1.37      schwarze  608: /* Array of injected predefined strings. */
                    609: #define        PREDEFS_MAX      38
                    610: static const struct predef predefs[PREDEFS_MAX] = {
                    611: #include "predefs.in"
                    612: };
                    613:
1.177     schwarze  614: static int      roffce_lines;  /* number of input lines to center */
                    615: static struct roff_node *roffce_node;  /* active request */
1.51      schwarze  616: static int      roffit_lines;  /* number of lines to delay */
                    617: static char    *roffit_macro;  /* nil-terminated macro line */
                    618:
1.80      schwarze  619:
1.138     schwarze  620: /* --- request table ------------------------------------------------------ */
                    621:
1.167     schwarze  622: struct ohash *
                    623: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  624: {
1.167     schwarze  625:        struct ohash    *htab;
                    626:        struct roffreq  *req;
                    627:        enum roff_tok    tok;
                    628:        size_t           sz;
                    629:        unsigned int     slot;
1.3       schwarze  630:
1.167     schwarze  631:        htab = mandoc_malloc(sizeof(*htab));
                    632:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  633:
1.167     schwarze  634:        for (tok = mintok; tok < maxtok; tok++) {
1.168     schwarze  635:                if (roff_name[tok] == NULL)
                    636:                        continue;
1.167     schwarze  637:                sz = strlen(roff_name[tok]);
                    638:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    639:                req->tok = tok;
                    640:                memcpy(req->name, roff_name[tok], sz + 1);
                    641:                slot = ohash_qlookup(htab, req->name);
                    642:                ohash_insert(htab, slot, req);
1.3       schwarze  643:        }
1.167     schwarze  644:        return htab;
1.3       schwarze  645: }
                    646:
1.167     schwarze  647: void
                    648: roffhash_free(struct ohash *htab)
1.1       schwarze  649: {
1.167     schwarze  650:        struct roffreq  *req;
                    651:        unsigned int     slot;
1.1       schwarze  652:
1.167     schwarze  653:        if (htab == NULL)
                    654:                return;
                    655:        for (req = ohash_first(htab, &slot); req != NULL;
                    656:             req = ohash_next(htab, &slot))
                    657:                free(req);
                    658:        ohash_delete(htab);
                    659:        free(htab);
                    660: }
                    661:
                    662: enum roff_tok
                    663: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    664: {
                    665:        struct roffreq  *req;
                    666:        const char      *end;
                    667:
                    668:        if (sz) {
                    669:                end = name + sz;
                    670:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    671:        } else
                    672:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    673:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  674: }
                    675:
1.138     schwarze  676: /* --- stack of request blocks -------------------------------------------- */
                    677:
1.1       schwarze  678: /*
                    679:  * Pop the current node off of the stack of roff instructions currently
                    680:  * pending.
                    681:  */
                    682: static void
                    683: roffnode_pop(struct roff *r)
                    684: {
                    685:        struct roffnode *p;
                    686:
1.2       schwarze  687:        assert(r->last);
1.80      schwarze  688:        p = r->last;
1.2       schwarze  689:
                    690:        r->last = r->last->parent;
1.16      schwarze  691:        free(p->name);
                    692:        free(p->end);
1.1       schwarze  693:        free(p);
                    694: }
                    695:
                    696: /*
                    697:  * Push a roff node onto the instruction stack.  This must later be
                    698:  * removed with roffnode_pop().
                    699:  */
1.11      schwarze  700: static void
1.166     schwarze  701: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  702:                int line, int col)
1.1       schwarze  703: {
                    704:        struct roffnode *p;
                    705:
1.11      schwarze  706:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  707:        p->tok = tok;
1.16      schwarze  708:        if (name)
                    709:                p->name = mandoc_strdup(name);
1.1       schwarze  710:        p->parent = r->last;
                    711:        p->line = line;
                    712:        p->col = col;
1.71      schwarze  713:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  714:
                    715:        r->last = p;
                    716: }
                    717:
1.138     schwarze  718: /* --- roff parser state data management ---------------------------------- */
                    719:
1.1       schwarze  720: static void
                    721: roff_free1(struct roff *r)
                    722: {
1.49      schwarze  723:        struct tbl_node *tbl;
1.32      schwarze  724:        struct eqn_node *e;
1.42      schwarze  725:        int              i;
1.27      schwarze  726:
1.49      schwarze  727:        while (NULL != (tbl = r->first_tbl)) {
                    728:                r->first_tbl = tbl->next;
                    729:                tbl_free(tbl);
1.27      schwarze  730:        }
                    731:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  732:
1.32      schwarze  733:        while (NULL != (e = r->first_eqn)) {
                    734:                r->first_eqn = e->next;
                    735:                eqn_free(e);
                    736:        }
                    737:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    738:
1.1       schwarze  739:        while (r->last)
                    740:                roffnode_pop(r);
1.27      schwarze  741:
1.96      schwarze  742:        free (r->rstack);
                    743:        r->rstack = NULL;
                    744:        r->rstacksz = 0;
                    745:        r->rstackpos = -1;
                    746:
                    747:        roff_freereg(r->regtab);
                    748:        r->regtab = NULL;
                    749:
1.42      schwarze  750:        roff_freestr(r->strtab);
1.178     schwarze  751:        roff_freestr(r->rentab);
1.42      schwarze  752:        roff_freestr(r->xmbtab);
1.178     schwarze  753:        r->strtab = r->rentab = r->xmbtab = NULL;
1.42      schwarze  754:
                    755:        if (r->xtab)
                    756:                for (i = 0; i < 128; i++)
                    757:                        free(r->xtab[i].p);
                    758:        free(r->xtab);
                    759:        r->xtab = NULL;
1.1       schwarze  760: }
                    761:
                    762: void
                    763: roff_reset(struct roff *r)
                    764: {
                    765:        roff_free1(r);
1.99      schwarze  766:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175     schwarze  767:        r->control = '\0';
                    768:        r->escape = '\\';
1.179     schwarze  769:        roffce_lines = 0;
                    770:        roffce_node = NULL;
                    771:        roffit_lines = 0;
                    772:        roffit_macro = NULL;
1.1       schwarze  773: }
                    774:
                    775: void
                    776: roff_free(struct roff *r)
                    777: {
                    778:        roff_free1(r);
1.167     schwarze  779:        roffhash_free(r->reqtab);
1.1       schwarze  780:        free(r);
                    781: }
                    782:
                    783: struct roff *
1.151     schwarze  784: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  785: {
                    786:        struct roff     *r;
                    787:
1.11      schwarze  788:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  789:        r->parse = parse;
1.167     schwarze  790:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72      schwarze  791:        r->options = options;
1.99      schwarze  792:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  793:        r->rstackpos = -1;
1.175     schwarze  794:        r->escape = '\\';
1.149     schwarze  795:        return r;
1.137     schwarze  796: }
                    797:
1.138     schwarze  798: /* --- syntax tree state data management ---------------------------------- */
                    799:
1.137     schwarze  800: static void
                    801: roff_man_free1(struct roff_man *man)
                    802: {
                    803:
1.138     schwarze  804:        if (man->first != NULL)
                    805:                roff_node_delete(man, man->first);
1.137     schwarze  806:        free(man->meta.msec);
                    807:        free(man->meta.vol);
                    808:        free(man->meta.os);
                    809:        free(man->meta.arch);
                    810:        free(man->meta.title);
                    811:        free(man->meta.name);
                    812:        free(man->meta.date);
                    813: }
                    814:
                    815: static void
                    816: roff_man_alloc1(struct roff_man *man)
                    817: {
                    818:
                    819:        memset(&man->meta, 0, sizeof(man->meta));
                    820:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    821:        man->first->type = ROFFT_ROOT;
                    822:        man->last = man->first;
                    823:        man->last_es = NULL;
                    824:        man->flags = 0;
                    825:        man->macroset = MACROSET_NONE;
                    826:        man->lastsec = man->lastnamed = SEC_NONE;
                    827:        man->next = ROFF_NEXT_CHILD;
                    828: }
                    829:
                    830: void
                    831: roff_man_reset(struct roff_man *man)
                    832: {
                    833:
                    834:        roff_man_free1(man);
                    835:        roff_man_alloc1(man);
                    836: }
                    837:
                    838: void
                    839: roff_man_free(struct roff_man *man)
                    840: {
                    841:
                    842:        roff_man_free1(man);
                    843:        free(man);
                    844: }
                    845:
                    846: struct roff_man *
                    847: roff_man_alloc(struct roff *roff, struct mparse *parse,
                    848:        const char *defos, int quick)
                    849: {
                    850:        struct roff_man *man;
                    851:
                    852:        man = mandoc_calloc(1, sizeof(*man));
                    853:        man->parse = parse;
                    854:        man->roff = roff;
                    855:        man->defos = defos;
                    856:        man->quick = quick;
                    857:        roff_man_alloc1(man);
1.168     schwarze  858:        roff->man = man;
1.149     schwarze  859:        return man;
1.1       schwarze  860: }
                    861:
1.138     schwarze  862: /* --- syntax tree handling ----------------------------------------------- */
                    863:
                    864: struct roff_node *
                    865: roff_node_alloc(struct roff_man *man, int line, int pos,
                    866:        enum roff_type type, int tok)
                    867: {
                    868:        struct roff_node        *n;
                    869:
                    870:        n = mandoc_calloc(1, sizeof(*n));
                    871:        n->line = line;
                    872:        n->pos = pos;
                    873:        n->tok = tok;
                    874:        n->type = type;
                    875:        n->sec = man->lastsec;
                    876:
                    877:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  878:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  879:        else
1.157     schwarze  880:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  881:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  882:                n->flags |= NODE_LINE;
1.138     schwarze  883:        man->flags &= ~MDOC_NEWLINE;
                    884:
1.149     schwarze  885:        return n;
1.138     schwarze  886: }
                    887:
                    888: void
                    889: roff_node_append(struct roff_man *man, struct roff_node *n)
                    890: {
                    891:
                    892:        switch (man->next) {
                    893:        case ROFF_NEXT_SIBLING:
1.153     schwarze  894:                if (man->last->next != NULL) {
                    895:                        n->next = man->last->next;
                    896:                        man->last->next->prev = n;
                    897:                } else
                    898:                        man->last->parent->last = n;
1.138     schwarze  899:                man->last->next = n;
                    900:                n->prev = man->last;
                    901:                n->parent = man->last->parent;
                    902:                break;
                    903:        case ROFF_NEXT_CHILD:
1.159     schwarze  904:                if (man->last->child != NULL) {
                    905:                        n->next = man->last->child;
                    906:                        man->last->child->prev = n;
                    907:                } else
                    908:                        man->last->last = n;
1.138     schwarze  909:                man->last->child = n;
                    910:                n->parent = man->last;
                    911:                break;
                    912:        default:
                    913:                abort();
                    914:        }
1.154     schwarze  915:        man->last = n;
1.138     schwarze  916:
                    917:        switch (n->type) {
                    918:        case ROFFT_HEAD:
                    919:                n->parent->head = n;
                    920:                break;
                    921:        case ROFFT_BODY:
1.154     schwarze  922:                if (n->end != ENDBODY_NOT)
                    923:                        return;
1.138     schwarze  924:                n->parent->body = n;
                    925:                break;
                    926:        case ROFFT_TAIL:
                    927:                n->parent->tail = n;
                    928:                break;
                    929:        default:
1.154     schwarze  930:                return;
1.138     schwarze  931:        }
1.154     schwarze  932:
                    933:        /*
                    934:         * Copy over the normalised-data pointer of our parent.  Not
                    935:         * everybody has one, but copying a null pointer is fine.
                    936:         */
                    937:
                    938:        n->norm = n->parent->norm;
                    939:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  940: }
                    941:
1.139     schwarze  942: void
                    943: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    944: {
                    945:        struct roff_node        *n;
                    946:
                    947:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    948:        n->string = roff_strdup(man->roff, word);
                    949:        roff_node_append(man, n);
1.158     schwarze  950:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  951:        man->next = ROFF_NEXT_SIBLING;
                    952: }
                    953:
                    954: void
                    955: roff_word_append(struct roff_man *man, const char *word)
                    956: {
                    957:        struct roff_node        *n;
                    958:        char                    *addstr, *newstr;
                    959:
                    960:        n = man->last;
                    961:        addstr = roff_strdup(man->roff, word);
                    962:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    963:        free(addstr);
                    964:        free(n->string);
                    965:        n->string = newstr;
                    966:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  967: }
                    968:
                    969: void
                    970: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    971: {
                    972:        struct roff_node        *n;
                    973:
                    974:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    975:        roff_node_append(man, n);
                    976:        man->next = ROFF_NEXT_CHILD;
                    977: }
                    978:
                    979: struct roff_node *
                    980: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    981: {
                    982:        struct roff_node        *n;
                    983:
                    984:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    985:        roff_node_append(man, n);
                    986:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  987:        return n;
1.139     schwarze  988: }
                    989:
1.138     schwarze  990: struct roff_node *
                    991: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    992: {
                    993:        struct roff_node        *n;
                    994:
                    995:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    996:        roff_node_append(man, n);
                    997:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  998:        return n;
1.138     schwarze  999: }
                   1000:
                   1001: struct roff_node *
                   1002: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                   1003: {
                   1004:        struct roff_node        *n;
                   1005:
                   1006:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                   1007:        roff_node_append(man, n);
                   1008:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze 1009:        return n;
1.139     schwarze 1010: }
                   1011:
                   1012: void
                   1013: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
                   1014: {
                   1015:        struct roff_node        *n;
                   1016:
                   1017:        n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
                   1018:        n->eqn = eqn;
                   1019:        if (eqn->ln > man->last->line)
1.157     schwarze 1020:                n->flags |= NODE_LINE;
1.139     schwarze 1021:        roff_node_append(man, n);
                   1022:        man->next = ROFF_NEXT_SIBLING;
                   1023: }
                   1024:
                   1025: void
                   1026: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
                   1027: {
                   1028:        struct roff_node        *n;
                   1029:
                   1030:        if (man->macroset == MACROSET_MAN)
1.174     schwarze 1031:                man_breakscope(man, ROFF_TS);
1.139     schwarze 1032:        n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                   1033:        n->span = tbl;
                   1034:        roff_node_append(man, n);
1.158     schwarze 1035:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze 1036:        man->next = ROFF_NEXT_SIBLING;
1.138     schwarze 1037: }
                   1038:
                   1039: void
                   1040: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1041: {
                   1042:
                   1043:        /* Adjust siblings. */
                   1044:
                   1045:        if (n->prev)
                   1046:                n->prev->next = n->next;
                   1047:        if (n->next)
                   1048:                n->next->prev = n->prev;
                   1049:
                   1050:        /* Adjust parent. */
                   1051:
                   1052:        if (n->parent != NULL) {
                   1053:                if (n->parent->child == n)
                   1054:                        n->parent->child = n->next;
                   1055:                if (n->parent->last == n)
                   1056:                        n->parent->last = n->prev;
                   1057:        }
                   1058:
                   1059:        /* Adjust parse point. */
                   1060:
                   1061:        if (man == NULL)
                   1062:                return;
                   1063:        if (man->last == n) {
                   1064:                if (n->prev == NULL) {
                   1065:                        man->last = n->parent;
                   1066:                        man->next = ROFF_NEXT_CHILD;
                   1067:                } else {
                   1068:                        man->last = n->prev;
                   1069:                        man->next = ROFF_NEXT_SIBLING;
                   1070:                }
                   1071:        }
                   1072:        if (man->first == n)
                   1073:                man->first = NULL;
                   1074: }
                   1075:
                   1076: void
                   1077: roff_node_free(struct roff_node *n)
                   1078: {
                   1079:
                   1080:        if (n->args != NULL)
                   1081:                mdoc_argv_free(n->args);
                   1082:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1083:                free(n->norm);
                   1084:        free(n->string);
                   1085:        free(n);
                   1086: }
                   1087:
                   1088: void
                   1089: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1090: {
                   1091:
                   1092:        while (n->child != NULL)
                   1093:                roff_node_delete(man, n->child);
                   1094:        roff_node_unlink(man, n);
                   1095:        roff_node_free(n);
1.141     schwarze 1096: }
                   1097:
                   1098: void
                   1099: deroff(char **dest, const struct roff_node *n)
                   1100: {
                   1101:        char    *cp;
                   1102:        size_t   sz;
                   1103:
                   1104:        if (n->type != ROFFT_TEXT) {
                   1105:                for (n = n->child; n != NULL; n = n->next)
                   1106:                        deroff(dest, n);
                   1107:                return;
                   1108:        }
                   1109:
1.160     schwarze 1110:        /* Skip leading whitespace. */
1.141     schwarze 1111:
1.160     schwarze 1112:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1113:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1114:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1115:                        cp++;
1.160     schwarze 1116:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1117:                        break;
                   1118:        }
                   1119:
1.161     schwarze 1120:        /* Skip trailing backslash. */
                   1121:
                   1122:        sz = strlen(cp);
1.162     schwarze 1123:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1124:                sz--;
                   1125:
1.141     schwarze 1126:        /* Skip trailing whitespace. */
                   1127:
1.161     schwarze 1128:        for (; sz; sz--)
1.141     schwarze 1129:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1130:                        break;
                   1131:
                   1132:        /* Skip empty strings. */
                   1133:
                   1134:        if (sz == 0)
                   1135:                return;
                   1136:
                   1137:        if (*dest == NULL) {
                   1138:                *dest = mandoc_strndup(cp, sz);
                   1139:                return;
                   1140:        }
                   1141:
                   1142:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1143:        free(*dest);
                   1144:        *dest = cp;
1.138     schwarze 1145: }
                   1146:
                   1147: /* --- main functions of the roff parser ---------------------------------- */
                   1148:
1.8       schwarze 1149: /*
1.79      schwarze 1150:  * In the current line, expand escape sequences that tend to get
                   1151:  * used in numerical expressions and conditional requests.
                   1152:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1153:  */
1.45      schwarze 1154: static enum rofferr
1.110     schwarze 1155: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1156: {
1.81      schwarze 1157:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze 1158:        const char      *start; /* start of the string to process */
1.82      schwarze 1159:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze 1160:        const char      *stnam; /* start of the name, after "[(*" */
                   1161:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1162:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1163:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1164:        size_t           maxl;  /* expected length of the escape name */
                   1165:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1166:        enum mandoc_esc  esc;   /* type of the escape sequence */
1.185     schwarze 1167:        enum mdoc_os     os_e;  /* kind of RCS id seen */
1.109     schwarze 1168:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1169:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1170:        int              npos;  /* position in numeric expression */
1.91      schwarze 1171:        int              arg_complete; /* argument not interrupted by eol */
1.175     schwarze 1172:        int              done;  /* no more input available */
1.79      schwarze 1173:        char             term;  /* character terminating the escape */
1.8       schwarze 1174:
1.175     schwarze 1175:        /* Search forward for comments. */
                   1176:
                   1177:        done = 0;
                   1178:        start = buf->buf + pos;
                   1179:        for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
                   1180:                if (stesc[0] != r->escape || stesc[1] == '\0')
                   1181:                        continue;
                   1182:                stesc++;
                   1183:                if (*stesc != '"' && *stesc != '#')
                   1184:                        continue;
1.185     schwarze 1185:
                   1186:                /* Comment found, look for RCS id. */
                   1187:
                   1188:                if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
                   1189:                        os_e = MDOC_OS_OPENBSD;
                   1190:                        cp += 8;
                   1191:                } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
                   1192:                        os_e = MDOC_OS_NETBSD;
                   1193:                        cp += 7;
                   1194:                }
                   1195:                if (cp != NULL &&
                   1196:                    isalnum((unsigned char)*cp) == 0 &&
1.186   ! schwarze 1197:                    strchr(cp, '$') != NULL) {
        !          1198:                        if (r->man->meta.rcsids & (1 << os_e))
        !          1199:                                mandoc_msg(MANDOCERR_RCS_REP, r->parse,
        !          1200:                                    ln, stesc + 1 - buf->buf, stesc + 1);
1.185     schwarze 1201:                        r->man->meta.rcsids |= 1 << os_e;
1.186   ! schwarze 1202:                }
1.185     schwarze 1203:
                   1204:                /* Handle trailing whitespace. */
                   1205:
1.175     schwarze 1206:                cp = strchr(stesc--, '\0') - 1;
                   1207:                if (*cp == '\n') {
                   1208:                        done = 1;
                   1209:                        cp--;
                   1210:                }
                   1211:                if (*cp == ' ' || *cp == '\t')
                   1212:                        mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
                   1213:                            ln, cp - buf->buf, NULL);
                   1214:                while (stesc > start && stesc[-1] == ' ')
                   1215:                        stesc--;
                   1216:                *stesc = '\0';
                   1217:                break;
                   1218:        }
                   1219:        if (stesc == start)
                   1220:                return ROFF_CONT;
                   1221:        stesc--;
                   1222:
                   1223:        /* Notice the end of the input. */
                   1224:
                   1225:        if (*stesc == '\n') {
                   1226:                *stesc-- = '\0';
                   1227:                done = 1;
                   1228:        }
                   1229:
1.43      schwarze 1230:        expand_count = 0;
1.175     schwarze 1231:        while (stesc >= start) {
1.43      schwarze 1232:
1.78      schwarze 1233:                /* Search backwards for the next backslash. */
                   1234:
1.175     schwarze 1235:                if (*stesc != r->escape) {
                   1236:                        if (*stesc == '\\') {
                   1237:                                *stesc = '\0';
                   1238:                                buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
                   1239:                                    buf->buf, stesc + 1) + 1;
                   1240:                                start = nbuf + pos;
                   1241:                                stesc = nbuf + (stesc - buf->buf);
                   1242:                                free(buf->buf);
                   1243:                                buf->buf = nbuf;
                   1244:                        }
                   1245:                        stesc--;
1.78      schwarze 1246:                        continue;
1.175     schwarze 1247:                }
1.78      schwarze 1248:
                   1249:                /* If it is escaped, skip it. */
                   1250:
                   1251:                for (cp = stesc - 1; cp >= start; cp--)
1.175     schwarze 1252:                        if (*cp != r->escape)
1.78      schwarze 1253:                                break;
                   1254:
1.110     schwarze 1255:                if ((stesc - cp) % 2 == 0) {
1.175     schwarze 1256:                        while (stesc > cp)
                   1257:                                *stesc-- = '\\';
1.78      schwarze 1258:                        continue;
1.175     schwarze 1259:                } else if (stesc[1] != '\0') {
                   1260:                        *stesc = '\\';
                   1261:                } else {
                   1262:                        *stesc-- = '\0';
                   1263:                        if (done)
                   1264:                                continue;
                   1265:                        else
                   1266:                                return ROFF_APPEND;
1.78      schwarze 1267:                }
1.23      schwarze 1268:
1.79      schwarze 1269:                /* Decide whether to expand or to check only. */
1.23      schwarze 1270:
1.79      schwarze 1271:                term = '\0';
1.78      schwarze 1272:                cp = stesc + 1;
1.53      schwarze 1273:                switch (*cp) {
1.80      schwarze 1274:                case '*':
1.53      schwarze 1275:                        res = NULL;
                   1276:                        break;
1.80      schwarze 1277:                case 'B':
                   1278:                case 'w':
1.79      schwarze 1279:                        term = cp[1];
                   1280:                        /* FALLTHROUGH */
1.80      schwarze 1281:                case 'n':
1.53      schwarze 1282:                        res = ubuf;
                   1283:                        break;
                   1284:                default:
1.109     schwarze 1285:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1286:                        if (esc == ESCAPE_ERROR ||
                   1287:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1288:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1289:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1290:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1291:                                    "%.*s", (int)(cp - stesc), stesc);
1.175     schwarze 1292:                        stesc--;
1.78      schwarze 1293:                        continue;
1.42      schwarze 1294:                }
                   1295:
1.78      schwarze 1296:                if (EXPAND_LIMIT < ++expand_count) {
                   1297:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1298:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1299:                        return ROFF_IGN;
1.78      schwarze 1300:                }
1.23      schwarze 1301:
                   1302:                /*
                   1303:                 * The third character decides the length
1.53      schwarze 1304:                 * of the name of the string or register.
1.23      schwarze 1305:                 * Save a pointer to the name.
                   1306:                 */
                   1307:
1.110     schwarze 1308:                if (term == '\0') {
1.79      schwarze 1309:                        switch (*++cp) {
1.80      schwarze 1310:                        case '\0':
1.79      schwarze 1311:                                maxl = 0;
                   1312:                                break;
1.80      schwarze 1313:                        case '(':
1.79      schwarze 1314:                                cp++;
                   1315:                                maxl = 2;
                   1316:                                break;
1.80      schwarze 1317:                        case '[':
1.79      schwarze 1318:                                cp++;
                   1319:                                term = ']';
                   1320:                                maxl = 0;
                   1321:                                break;
                   1322:                        default:
                   1323:                                maxl = 1;
                   1324:                                break;
                   1325:                        }
                   1326:                } else {
                   1327:                        cp += 2;
1.8       schwarze 1328:                        maxl = 0;
                   1329:                }
1.23      schwarze 1330:                stnam = cp;
1.8       schwarze 1331:
1.23      schwarze 1332:                /* Advance to the end of the name. */
1.8       schwarze 1333:
1.125     schwarze 1334:                naml = 0;
1.91      schwarze 1335:                arg_complete = 1;
1.125     schwarze 1336:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1337:                        if (*cp == '\0') {
1.92      schwarze 1338:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1339:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1340:                                arg_complete = 0;
1.79      schwarze 1341:                                break;
1.42      schwarze 1342:                        }
1.110     schwarze 1343:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1344:                                cp++;
1.125     schwarze 1345:                                break;
                   1346:                        }
                   1347:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1348:                                naml++;
                   1349:                                continue;
                   1350:                        }
                   1351:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1352:                        case ESCAPE_SPECIAL:
                   1353:                        case ESCAPE_UNICODE:
                   1354:                        case ESCAPE_NUMBERED:
                   1355:                        case ESCAPE_OVERSTRIKE:
                   1356:                                naml++;
                   1357:                                break;
                   1358:                        default:
1.8       schwarze 1359:                                break;
1.79      schwarze 1360:                        }
1.8       schwarze 1361:                }
                   1362:
1.23      schwarze 1363:                /*
                   1364:                 * Retrieve the replacement string; if it is
                   1365:                 * undefined, resume searching for escapes.
                   1366:                 */
                   1367:
1.79      schwarze 1368:                switch (stesc[1]) {
1.80      schwarze 1369:                case '*':
1.91      schwarze 1370:                        if (arg_complete)
                   1371:                                res = roff_getstrn(r, stnam, naml);
1.79      schwarze 1372:                        break;
1.80      schwarze 1373:                case 'B':
1.79      schwarze 1374:                        npos = 0;
1.91      schwarze 1375:                        ubuf[0] = arg_complete &&
1.133     schwarze 1376:                            roff_evalnum(r, ln, stnam, &npos,
                   1377:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1378:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1379:                        ubuf[1] = '\0';
                   1380:                        break;
1.80      schwarze 1381:                case 'n':
1.91      schwarze 1382:                        if (arg_complete)
                   1383:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1384:                                    roff_getregn(r, stnam, naml));
                   1385:                        else
                   1386:                                ubuf[0] = '\0';
1.79      schwarze 1387:                        break;
1.80      schwarze 1388:                case 'w':
1.91      schwarze 1389:                        /* use even incomplete args */
1.81      schwarze 1390:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1391:                            24 * (int)naml);
                   1392:                        break;
                   1393:                }
1.8       schwarze 1394:
1.110     schwarze 1395:                if (res == NULL) {
1.92      schwarze 1396:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1397:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1398:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1399:                        res = "";
1.118     schwarze 1400:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1401:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1402:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1403:                        return ROFF_IGN;
1.8       schwarze 1404:                }
                   1405:
1.23      schwarze 1406:                /* Replace the escape sequence by the string. */
                   1407:
1.82      schwarze 1408:                *stesc = '\0';
1.110     schwarze 1409:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1410:                    buf->buf, res, cp) + 1;
1.8       schwarze 1411:
1.78      schwarze 1412:                /* Prepare for the next replacement. */
1.8       schwarze 1413:
1.78      schwarze 1414:                start = nbuf + pos;
1.110     schwarze 1415:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1416:                free(buf->buf);
                   1417:                buf->buf = nbuf;
1.42      schwarze 1418:        }
1.149     schwarze 1419:        return ROFF_CONT;
1.42      schwarze 1420: }
                   1421:
                   1422: /*
1.147     schwarze 1423:  * Process text streams.
1.42      schwarze 1424:  */
                   1425: static enum rofferr
1.177     schwarze 1426: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42      schwarze 1427: {
                   1428:        size_t           sz;
                   1429:        const char      *start;
1.51      schwarze 1430:        char            *p;
                   1431:        int              isz;
1.42      schwarze 1432:        enum mandoc_esc  esc;
                   1433:
1.147     schwarze 1434:        /* Spring the input line trap. */
                   1435:
                   1436:        if (roffit_lines == 1) {
                   1437:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1438:                free(buf->buf);
                   1439:                buf->buf = p;
                   1440:                buf->sz = isz + 1;
                   1441:                *offs = 0;
                   1442:                free(roffit_macro);
                   1443:                roffit_lines = 0;
1.149     schwarze 1444:                return ROFF_REPARSE;
1.147     schwarze 1445:        } else if (roffit_lines > 1)
                   1446:                --roffit_lines;
                   1447:
1.177     schwarze 1448:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1449:                if (roffce_lines < 1) {
                   1450:                        r->man->last = roffce_node;
                   1451:                        r->man->next = ROFF_NEXT_SIBLING;
                   1452:                        roffce_lines = 0;
                   1453:                        roffce_node = NULL;
                   1454:                } else
                   1455:                        roffce_lines--;
                   1456:        }
                   1457:
1.147     schwarze 1458:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1459:
1.110     schwarze 1460:        start = p = buf->buf + pos;
1.42      schwarze 1461:
1.110     schwarze 1462:        while (*p != '\0') {
1.42      schwarze 1463:                sz = strcspn(p, "-\\");
                   1464:                p += sz;
                   1465:
1.110     schwarze 1466:                if (*p == '\0')
1.42      schwarze 1467:                        break;
                   1468:
1.110     schwarze 1469:                if (*p == '\\') {
1.42      schwarze 1470:                        /* Skip over escapes. */
                   1471:                        p++;
1.62      schwarze 1472:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1473:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1474:                                break;
1.136     schwarze 1475:                        while (*p == '-')
                   1476:                                p++;
1.42      schwarze 1477:                        continue;
                   1478:                } else if (p == start) {
                   1479:                        p++;
                   1480:                        continue;
                   1481:                }
                   1482:
1.44      schwarze 1483:                if (isalpha((unsigned char)p[-1]) &&
                   1484:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1485:                        *p = ASCII_HYPH;
                   1486:                p++;
1.8       schwarze 1487:        }
1.149     schwarze 1488:        return ROFF_CONT;
1.8       schwarze 1489: }
                   1490:
1.1       schwarze 1491: enum rofferr
1.110     schwarze 1492: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1493: {
1.166     schwarze 1494:        enum roff_tok    t;
1.27      schwarze 1495:        enum rofferr     e;
1.110     schwarze 1496:        int              pos;   /* parse point */
1.115     schwarze 1497:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1498:        int              ppos;  /* original offset in buf->buf */
                   1499:        int              ctl;   /* macro line (boolean) */
                   1500:
                   1501:        ppos = pos = *offs;
1.1       schwarze 1502:
1.102     schwarze 1503:        /* Handle in-line equation delimiters. */
                   1504:
1.108     schwarze 1505:        if (r->tbl == NULL &&
                   1506:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1507:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1508:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1509:                if (e == ROFF_REPARSE)
1.149     schwarze 1510:                        return e;
1.102     schwarze 1511:                assert(e == ROFF_CONT);
                   1512:        }
                   1513:
                   1514:        /* Expand some escape sequences. */
1.8       schwarze 1515:
1.110     schwarze 1516:        e = roff_res(r, buf, ln, pos);
1.175     schwarze 1517:        if (e == ROFF_IGN || e == ROFF_APPEND)
1.149     schwarze 1518:                return e;
1.110     schwarze 1519:        assert(e == ROFF_CONT);
1.8       schwarze 1520:
1.110     schwarze 1521:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1522:
1.8       schwarze 1523:        /*
1.2       schwarze 1524:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1525:         * text through the macro's filter.
                   1526:         * Equations process all content themselves.
                   1527:         * Tables process almost all content themselves, but we want
                   1528:         * to warn about macros before passing it there.
1.2       schwarze 1529:         */
                   1530:
1.124     schwarze 1531:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1532:                t = r->last->tok;
1.110     schwarze 1533:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1534:                if (e == ROFF_IGN)
1.149     schwarze 1535:                        return e;
1.166     schwarze 1536:                assert(e == ROFF_CONT);
1.54      schwarze 1537:        }
1.124     schwarze 1538:        if (r->eqn != NULL)
1.149     schwarze 1539:                return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.124     schwarze 1540:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.149     schwarze 1541:                return tbl_read(r->tbl, ln, buf->buf, ppos);
1.124     schwarze 1542:        if ( ! ctl)
1.177     schwarze 1543:                return roff_parsetext(r, buf, pos, offs);
1.100     schwarze 1544:
                   1545:        /* Skip empty request lines. */
                   1546:
1.110     schwarze 1547:        if (buf->buf[pos] == '"') {
1.100     schwarze 1548:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1549:                    ln, pos, NULL);
1.149     schwarze 1550:                return ROFF_IGN;
1.110     schwarze 1551:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1552:                return ROFF_IGN;
1.2       schwarze 1553:
                   1554:        /*
                   1555:         * If a scope is open, go to the child handler for that macro,
                   1556:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1557:         * Don't do so if an equation is open.
1.2       schwarze 1558:         */
                   1559:
                   1560:        if (r->last) {
1.1       schwarze 1561:                t = r->last->tok;
1.149     schwarze 1562:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1563:        }
                   1564:
1.115     schwarze 1565:        /* No scope is open.  This is a new request or macro. */
                   1566:
                   1567:        spos = pos;
                   1568:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1569:
                   1570:        /* Tables ignore most macros. */
                   1571:
1.180     schwarze 1572:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181     schwarze 1573:            t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.115     schwarze 1574:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1575:                    ln, pos, buf->buf + spos);
1.180     schwarze 1576:                if (t != TOKEN_NONE)
1.149     schwarze 1577:                        return ROFF_IGN;
1.129     schwarze 1578:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1579:                        pos++;
1.163     schwarze 1580:                while (buf->buf[pos] == ' ')
1.129     schwarze 1581:                        pos++;
1.149     schwarze 1582:                return tbl_read(r->tbl, ln, buf->buf, pos);
1.115     schwarze 1583:        }
                   1584:
1.177     schwarze 1585:        /* For now, let high level macros abort .ce mode. */
                   1586:
                   1587:        if (ctl && roffce_node != NULL &&
                   1588:            (t == TOKEN_NONE || t == ROFF_EQ || t == ROFF_TS)) {
                   1589:                r->man->last = roffce_node;
                   1590:                r->man->next = ROFF_NEXT_SIBLING;
                   1591:                roffce_lines = 0;
                   1592:                roffce_node = NULL;
                   1593:        }
                   1594:
1.2       schwarze 1595:        /*
1.115     schwarze 1596:         * This is neither a roff request nor a user-defined macro.
                   1597:         * Let the standard macro set parsers handle it.
1.2       schwarze 1598:         */
                   1599:
1.166     schwarze 1600:        if (t == TOKEN_NONE)
1.149     schwarze 1601:                return ROFF_CONT;
1.115     schwarze 1602:
                   1603:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1604:
1.168     schwarze 1605:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2       schwarze 1606: }
                   1607:
1.27      schwarze 1608: void
1.2       schwarze 1609: roff_endparse(struct roff *r)
                   1610: {
1.1       schwarze 1611:
1.27      schwarze 1612:        if (r->last)
1.94      schwarze 1613:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1614:                    r->last->line, r->last->col,
1.167     schwarze 1615:                    roff_name[r->last->tok]);
1.27      schwarze 1616:
1.32      schwarze 1617:        if (r->eqn) {
1.94      schwarze 1618:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1619:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41      schwarze 1620:                eqn_end(&r->eqn);
1.32      schwarze 1621:        }
                   1622:
1.27      schwarze 1623:        if (r->tbl) {
1.94      schwarze 1624:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1625:                    r->tbl->line, r->tbl->pos, "TS");
1.41      schwarze 1626:                tbl_end(&r->tbl);
1.27      schwarze 1627:        }
1.1       schwarze 1628: }
                   1629:
                   1630: /*
                   1631:  * Parse a roff node's type from the input buffer.  This must be in the
                   1632:  * form of ".foo xxx" in the usual way.
                   1633:  */
1.166     schwarze 1634: static enum roff_tok
1.87      schwarze 1635: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1636: {
1.87      schwarze 1637:        char            *cp;
1.16      schwarze 1638:        const char      *mac;
                   1639:        size_t           maclen;
1.166     schwarze 1640:        enum roff_tok    t;
1.1       schwarze 1641:
1.87      schwarze 1642:        cp = buf + *pos;
                   1643:
                   1644:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1645:                return TOKEN_NONE;
1.1       schwarze 1646:
1.87      schwarze 1647:        mac = cp;
                   1648:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1649:
1.178     schwarze 1650:        t = (r->current_string = roff_getstrn(r, mac, maclen)) ?
                   1651:            ROFF_USERDEF :
                   1652:            (r->current_string = roff_getrenn(r, mac, maclen)) ?
                   1653:            ROFF_RENAMED : roffhash_find(r->reqtab, mac, maclen);
1.1       schwarze 1654:
1.166     schwarze 1655:        if (t != TOKEN_NONE)
1.87      schwarze 1656:                *pos = cp - buf;
1.1       schwarze 1657:
1.149     schwarze 1658:        return t;
1.1       schwarze 1659: }
                   1660:
1.138     schwarze 1661: /* --- handling of request blocks ----------------------------------------- */
                   1662:
1.1       schwarze 1663: static enum rofferr
1.2       schwarze 1664: roff_cblock(ROFF_ARGS)
1.1       schwarze 1665: {
                   1666:
1.2       schwarze 1667:        /*
                   1668:         * A block-close `..' should only be invoked as a child of an
                   1669:         * ignore macro, otherwise raise a warning and just ignore it.
                   1670:         */
                   1671:
1.110     schwarze 1672:        if (r->last == NULL) {
1.94      schwarze 1673:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1674:                    ln, ppos, "..");
1.149     schwarze 1675:                return ROFF_IGN;
1.2       schwarze 1676:        }
1.1       schwarze 1677:
1.2       schwarze 1678:        switch (r->last->tok) {
1.80      schwarze 1679:        case ROFF_am:
1.93      schwarze 1680:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1681:        case ROFF_ami:
                   1682:        case ROFF_de:
1.23      schwarze 1683:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1684:        case ROFF_dei:
                   1685:        case ROFF_ig:
1.2       schwarze 1686:                break;
                   1687:        default:
1.94      schwarze 1688:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1689:                    ln, ppos, "..");
1.149     schwarze 1690:                return ROFF_IGN;
1.2       schwarze 1691:        }
                   1692:
1.110     schwarze 1693:        if (buf->buf[pos] != '\0')
1.90      schwarze 1694:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1695:                    ".. %s", buf->buf + pos);
1.2       schwarze 1696:
                   1697:        roffnode_pop(r);
                   1698:        roffnode_cleanscope(r);
1.149     schwarze 1699:        return ROFF_IGN;
1.2       schwarze 1700:
                   1701: }
1.1       schwarze 1702:
1.2       schwarze 1703: static void
                   1704: roffnode_cleanscope(struct roff *r)
                   1705: {
1.1       schwarze 1706:
1.2       schwarze 1707:        while (r->last) {
1.46      schwarze 1708:                if (--r->last->endspan != 0)
1.2       schwarze 1709:                        break;
                   1710:                roffnode_pop(r);
                   1711:        }
                   1712: }
1.1       schwarze 1713:
1.68      schwarze 1714: static void
                   1715: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1716: {
1.1       schwarze 1717:
1.2       schwarze 1718:        if (NULL == r->last) {
1.94      schwarze 1719:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1720:                    ln, ppos, "\\}");
1.68      schwarze 1721:                return;
1.2       schwarze 1722:        }
1.1       schwarze 1723:
1.2       schwarze 1724:        switch (r->last->tok) {
1.80      schwarze 1725:        case ROFF_el:
                   1726:        case ROFF_ie:
                   1727:        case ROFF_if:
1.2       schwarze 1728:                break;
                   1729:        default:
1.94      schwarze 1730:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1731:                    ln, ppos, "\\}");
1.68      schwarze 1732:                return;
1.2       schwarze 1733:        }
1.1       schwarze 1734:
1.2       schwarze 1735:        if (r->last->endspan > -1) {
1.94      schwarze 1736:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1737:                    ln, ppos, "\\}");
1.68      schwarze 1738:                return;
1.2       schwarze 1739:        }
                   1740:
                   1741:        roffnode_pop(r);
                   1742:        roffnode_cleanscope(r);
1.68      schwarze 1743:        return;
1.1       schwarze 1744: }
                   1745:
                   1746: static enum rofferr
1.2       schwarze 1747: roff_block(ROFF_ARGS)
1.1       schwarze 1748: {
1.93      schwarze 1749:        const char      *name;
                   1750:        char            *iname, *cp;
1.86      schwarze 1751:        size_t           namesz;
1.16      schwarze 1752:
1.93      schwarze 1753:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1754:
1.110     schwarze 1755:        if (tok == ROFF_de1)
1.93      schwarze 1756:                tok = ROFF_de;
1.123     schwarze 1757:        else if (tok == ROFF_dei1)
                   1758:                tok = ROFF_dei;
1.110     schwarze 1759:        else if (tok == ROFF_am1)
1.93      schwarze 1760:                tok = ROFF_am;
1.123     schwarze 1761:        else if (tok == ROFF_ami1)
                   1762:                tok = ROFF_ami;
1.93      schwarze 1763:
                   1764:        /* Parse the macro name argument. */
                   1765:
1.110     schwarze 1766:        cp = buf->buf + pos;
                   1767:        if (tok == ROFF_ig) {
1.93      schwarze 1768:                iname = NULL;
                   1769:                namesz = 0;
                   1770:        } else {
                   1771:                iname = cp;
                   1772:                namesz = roff_getname(r, &cp, ln, ppos);
                   1773:                iname[namesz] = '\0';
                   1774:        }
1.22      schwarze 1775:
1.93      schwarze 1776:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1777:
1.110     schwarze 1778:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1779:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1780:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1781:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1782:                            "%.*s", (int)namesz, iname);
                   1783:                        namesz = 0;
                   1784:                } else
                   1785:                        namesz = strlen(name);
                   1786:        } else
                   1787:                name = iname;
1.22      schwarze 1788:
1.110     schwarze 1789:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1790:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167     schwarze 1791:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1792:                return ROFF_IGN;
1.93      schwarze 1793:        }
1.2       schwarze 1794:
1.16      schwarze 1795:        roffnode_push(r, tok, name, ln, ppos);
                   1796:
                   1797:        /*
                   1798:         * At the beginning of a `de' macro, clear the existing string
                   1799:         * with the same name, if there is one.  New content will be
1.66      schwarze 1800:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1801:         */
1.22      schwarze 1802:
1.183     schwarze 1803:        if (tok == ROFF_de || tok == ROFF_dei) {
1.86      schwarze 1804:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183     schwarze 1805:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   1806:        }
1.2       schwarze 1807:
1.110     schwarze 1808:        if (*cp == '\0')
1.149     schwarze 1809:                return ROFF_IGN;
1.1       schwarze 1810:
1.93      schwarze 1811:        /* Get the custom end marker. */
1.22      schwarze 1812:
1.93      schwarze 1813:        iname = cp;
1.86      schwarze 1814:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1815:
                   1816:        /* Resolve the end marker if it is indirect. */
                   1817:
1.110     schwarze 1818:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1819:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1820:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1821:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1822:                            "%.*s", (int)namesz, iname);
                   1823:                        namesz = 0;
                   1824:                } else
                   1825:                        namesz = strlen(name);
                   1826:        } else
                   1827:                name = iname;
                   1828:
1.86      schwarze 1829:        if (namesz)
                   1830:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1831:
1.110     schwarze 1832:        if (*cp != '\0')
1.90      schwarze 1833:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167     schwarze 1834:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1835:
1.149     schwarze 1836:        return ROFF_IGN;
1.1       schwarze 1837: }
                   1838:
                   1839: static enum rofferr
1.2       schwarze 1840: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1841: {
1.166     schwarze 1842:        enum roff_tok   t;
1.2       schwarze 1843:        int             i, j;
                   1844:
                   1845:        /*
                   1846:         * First check whether a custom macro exists at this level.  If
                   1847:         * it does, then check against it.  This is some of groff's
                   1848:         * stranger behaviours.  If we encountered a custom end-scope
                   1849:         * tag and that tag also happens to be a "real" macro, then we
                   1850:         * need to try interpreting it again as a real macro.  If it's
                   1851:         * not, then return ignore.  Else continue.
                   1852:         */
                   1853:
                   1854:        if (r->last->end) {
1.35      schwarze 1855:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1856:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1857:                                break;
1.1       schwarze 1858:
1.110     schwarze 1859:                if (r->last->end[j] == '\0' &&
                   1860:                    (buf->buf[i] == '\0' ||
                   1861:                     buf->buf[i] == ' ' ||
                   1862:                     buf->buf[i] == '\t')) {
1.2       schwarze 1863:                        roffnode_pop(r);
                   1864:                        roffnode_cleanscope(r);
1.1       schwarze 1865:
1.110     schwarze 1866:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1867:                                i++;
                   1868:
                   1869:                        pos = i;
1.110     schwarze 1870:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1871:                            TOKEN_NONE)
1.149     schwarze 1872:                                return ROFF_RERUN;
                   1873:                        return ROFF_IGN;
1.2       schwarze 1874:                }
1.1       schwarze 1875:        }
                   1876:
1.2       schwarze 1877:        /*
                   1878:         * If we have no custom end-query or lookup failed, then try
                   1879:         * pulling it out of the hashtable.
                   1880:         */
1.1       schwarze 1881:
1.110     schwarze 1882:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1883:
1.110     schwarze 1884:        if (t != ROFF_cblock) {
                   1885:                if (tok != ROFF_ig)
                   1886:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1887:                return ROFF_IGN;
1.16      schwarze 1888:        }
1.1       schwarze 1889:
1.149     schwarze 1890:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1891: }
                   1892:
                   1893: static enum rofferr
                   1894: roff_block_text(ROFF_ARGS)
                   1895: {
                   1896:
1.110     schwarze 1897:        if (tok != ROFF_ig)
                   1898:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1899:
1.149     schwarze 1900:        return ROFF_IGN;
1.2       schwarze 1901: }
                   1902:
                   1903: static enum rofferr
                   1904: roff_cond_sub(ROFF_ARGS)
                   1905: {
1.166     schwarze 1906:        enum roff_tok    t;
1.37      schwarze 1907:        char            *ep;
1.71      schwarze 1908:        int              rr;
1.2       schwarze 1909:
                   1910:        rr = r->last->rule;
1.37      schwarze 1911:        roffnode_cleanscope(r);
1.110     schwarze 1912:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2       schwarze 1913:
1.37      schwarze 1914:        /*
1.50      schwarze 1915:         * Fully handle known macros when they are structurally
                   1916:         * required or when the conditional evaluated to true.
1.5       schwarze 1917:         */
                   1918:
1.166     schwarze 1919:        if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.149     schwarze 1920:                return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.39      schwarze 1921:
1.69      schwarze 1922:        /*
                   1923:         * If `\}' occurs on a macro line without a preceding macro,
                   1924:         * drop the line completely.
                   1925:         */
                   1926:
1.110     schwarze 1927:        ep = buf->buf + pos;
                   1928:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1929:                rr = 0;
1.69      schwarze 1930:
1.50      schwarze 1931:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1932:
1.110     schwarze 1933:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1934:                if (*(++ep) == '}') {
1.70      schwarze 1935:                        *ep = '&';
1.110     schwarze 1936:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1937:                }
1.119     schwarze 1938:                if (*ep != '\0')
                   1939:                        ++ep;
1.50      schwarze 1940:        }
1.149     schwarze 1941:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1942: }
                   1943:
                   1944: static enum rofferr
                   1945: roff_cond_text(ROFF_ARGS)
                   1946: {
1.37      schwarze 1947:        char            *ep;
1.71      schwarze 1948:        int              rr;
1.2       schwarze 1949:
                   1950:        rr = r->last->rule;
1.37      schwarze 1951:        roffnode_cleanscope(r);
1.1       schwarze 1952:
1.110     schwarze 1953:        ep = buf->buf + pos;
                   1954:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1955:                if (*(++ep) == '}') {
1.70      schwarze 1956:                        *ep = '&';
1.110     schwarze 1957:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1958:                }
1.119     schwarze 1959:                if (*ep != '\0')
                   1960:                        ++ep;
1.2       schwarze 1961:        }
1.149     schwarze 1962:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1963: }
                   1964:
1.138     schwarze 1965: /* --- handling of numeric and conditional expressions -------------------- */
                   1966:
1.77      schwarze 1967: /*
                   1968:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1969:  * If there is at least one digit, return success and advance the
                   1970:  * parse point, else return failure and let the parse point unchanged.
                   1971:  * Ignore overflows, treat them just like the C language.
                   1972:  */
1.56      schwarze 1973: static int
1.133     schwarze 1974: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 1975: {
1.133     schwarze 1976:        int      myres, scaled, n, p;
1.79      schwarze 1977:
                   1978:        if (NULL == res)
                   1979:                res = &myres;
1.56      schwarze 1980:
                   1981:        p = *pos;
                   1982:        n = v[p] == '-';
1.133     schwarze 1983:        if (n || v[p] == '+')
1.56      schwarze 1984:                p++;
                   1985:
1.133     schwarze 1986:        if (flags & ROFFNUM_WHITE)
                   1987:                while (isspace((unsigned char)v[p]))
                   1988:                        p++;
                   1989:
1.56      schwarze 1990:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 1991:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 1992:        if (p == *pos + n)
                   1993:                return 0;
                   1994:
                   1995:        if (n)
                   1996:                *res = -*res;
                   1997:
1.126     schwarze 1998:        /* Each number may be followed by one optional scaling unit. */
                   1999:
                   2000:        switch (v[p]) {
                   2001:        case 'f':
1.133     schwarze 2002:                scaled = *res * 65536;
1.126     schwarze 2003:                break;
                   2004:        case 'i':
1.133     schwarze 2005:                scaled = *res * 240;
1.126     schwarze 2006:                break;
                   2007:        case 'c':
1.133     schwarze 2008:                scaled = *res * 240 / 2.54;
1.126     schwarze 2009:                break;
                   2010:        case 'v':
                   2011:        case 'P':
1.133     schwarze 2012:                scaled = *res * 40;
1.126     schwarze 2013:                break;
                   2014:        case 'm':
                   2015:        case 'n':
1.133     schwarze 2016:                scaled = *res * 24;
1.126     schwarze 2017:                break;
                   2018:        case 'p':
1.133     schwarze 2019:                scaled = *res * 10 / 3;
1.126     schwarze 2020:                break;
                   2021:        case 'u':
1.133     schwarze 2022:                scaled = *res;
1.126     schwarze 2023:                break;
                   2024:        case 'M':
1.133     schwarze 2025:                scaled = *res * 6 / 25;
1.126     schwarze 2026:                break;
                   2027:        default:
1.133     schwarze 2028:                scaled = *res;
1.126     schwarze 2029:                p--;
                   2030:                break;
                   2031:        }
1.133     schwarze 2032:        if (flags & ROFFNUM_SCALE)
                   2033:                *res = scaled;
1.126     schwarze 2034:
                   2035:        *pos = p + 1;
1.149     schwarze 2036:        return 1;
1.56      schwarze 2037: }
                   2038:
1.71      schwarze 2039: /*
                   2040:  * Evaluate a string comparison condition.
                   2041:  * The first character is the delimiter.
                   2042:  * Succeed if the string up to its second occurrence
                   2043:  * matches the string up to its third occurence.
                   2044:  * Advance the cursor after the third occurrence
                   2045:  * or lacking that, to the end of the line.
                   2046:  */
                   2047: static int
                   2048: roff_evalstrcond(const char *v, int *pos)
                   2049: {
                   2050:        const char      *s1, *s2, *s3;
                   2051:        int              match;
                   2052:
                   2053:        match = 0;
                   2054:        s1 = v + *pos;          /* initial delimiter */
                   2055:        s2 = s1 + 1;            /* for scanning the first string */
                   2056:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2057:
                   2058:        if (NULL == s3)         /* found no middle delimiter */
                   2059:                goto out;
                   2060:
                   2061:        while ('\0' != *++s3) {
                   2062:                if (*s2 != *s3) {  /* mismatch */
                   2063:                        s3 = strchr(s3, *s1);
                   2064:                        break;
                   2065:                }
                   2066:                if (*s3 == *s1) {  /* found the final delimiter */
                   2067:                        match = 1;
                   2068:                        break;
                   2069:                }
                   2070:                s2++;
                   2071:        }
                   2072:
                   2073: out:
                   2074:        if (NULL == s3)
                   2075:                s3 = strchr(s2, '\0');
1.114     schwarze 2076:        else if (*s3 != '\0')
1.71      schwarze 2077:                s3++;
                   2078:        *pos = s3 - v;
1.149     schwarze 2079:        return match;
1.71      schwarze 2080: }
                   2081:
1.77      schwarze 2082: /*
                   2083:  * Evaluate an optionally negated single character, numerical,
                   2084:  * or string condition.
                   2085:  */
1.71      schwarze 2086: static int
1.143     schwarze 2087: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 2088: {
1.143     schwarze 2089:        char    *cp, *name;
                   2090:        size_t   sz;
1.182     schwarze 2091:        int      number, savepos, istrue, wanttrue;
1.5       schwarze 2092:
1.71      schwarze 2093:        if ('!' == v[*pos]) {
                   2094:                wanttrue = 0;
                   2095:                (*pos)++;
                   2096:        } else
                   2097:                wanttrue = 1;
                   2098:
1.5       schwarze 2099:        switch (v[*pos]) {
1.112     schwarze 2100:        case '\0':
1.149     schwarze 2101:                return 0;
1.80      schwarze 2102:        case 'n':
                   2103:        case 'o':
1.5       schwarze 2104:                (*pos)++;
1.149     schwarze 2105:                return wanttrue;
1.80      schwarze 2106:        case 'c':
                   2107:        case 'e':
                   2108:        case 't':
1.111     schwarze 2109:        case 'v':
1.5       schwarze 2110:                (*pos)++;
1.149     schwarze 2111:                return !wanttrue;
1.182     schwarze 2112:        case 'd':
1.143     schwarze 2113:        case 'r':
1.182     schwarze 2114:                cp = v + *pos + 1;
                   2115:                while (*cp == ' ')
                   2116:                        cp++;
                   2117:                name = cp;
                   2118:                sz = roff_getname(r, &cp, ln, cp - v);
                   2119:                istrue = sz && (v[*pos] == 'r' ? roff_hasregn(r, name, sz) :
                   2120:                    (roff_getstrn(r, name, sz) != NULL ||
                   2121:                     roff_getrenn(r, name, sz) != NULL));
1.143     schwarze 2122:                *pos = cp - v;
1.182     schwarze 2123:                return istrue == wanttrue;
1.5       schwarze 2124:        default:
                   2125:                break;
                   2126:        }
                   2127:
1.113     schwarze 2128:        savepos = *pos;
1.133     schwarze 2129:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 2130:                return (number > 0) == wanttrue;
1.113     schwarze 2131:        else if (*pos == savepos)
1.149     schwarze 2132:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 2133:        else
1.149     schwarze 2134:                return 0;
1.5       schwarze 2135: }
                   2136:
1.2       schwarze 2137: static enum rofferr
1.21      schwarze 2138: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 2139: {
1.30      schwarze 2140:
1.149     schwarze 2141:        return ROFF_IGN;
1.21      schwarze 2142: }
                   2143:
                   2144: static enum rofferr
1.123     schwarze 2145: roff_insec(ROFF_ARGS)
                   2146: {
                   2147:
                   2148:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167     schwarze 2149:            ln, ppos, roff_name[tok]);
1.149     schwarze 2150:        return ROFF_IGN;
1.123     schwarze 2151: }
                   2152:
                   2153: static enum rofferr
                   2154: roff_unsupp(ROFF_ARGS)
                   2155: {
                   2156:
                   2157:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167     schwarze 2158:            ln, ppos, roff_name[tok]);
1.149     schwarze 2159:        return ROFF_IGN;
1.123     schwarze 2160: }
                   2161:
                   2162: static enum rofferr
1.2       schwarze 2163: roff_cond(ROFF_ARGS)
                   2164: {
1.46      schwarze 2165:
                   2166:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2167:
1.80      schwarze 2168:        /*
1.35      schwarze 2169:         * An `.el' has no conditional body: it will consume the value
                   2170:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2171:         * defaults to DENY.
1.35      schwarze 2172:         *
                   2173:         * If we're not an `el', however, then evaluate the conditional.
                   2174:         */
1.1       schwarze 2175:
1.110     schwarze 2176:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2177:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2178:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2179:
1.35      schwarze 2180:        /*
                   2181:         * An if-else will put the NEGATION of the current evaluated
                   2182:         * conditional into the stack of rules.
                   2183:         */
                   2184:
1.110     schwarze 2185:        if (tok == ROFF_ie) {
1.96      schwarze 2186:                if (r->rstackpos + 1 == r->rstacksz) {
                   2187:                        r->rstacksz += 16;
                   2188:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2189:                            r->rstacksz, sizeof(int));
1.35      schwarze 2190:                }
1.71      schwarze 2191:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2192:        }
1.5       schwarze 2193:
                   2194:        /* If the parent has false as its rule, then so do we. */
                   2195:
1.71      schwarze 2196:        if (r->last->parent && !r->last->parent->rule)
                   2197:                r->last->rule = 0;
1.5       schwarze 2198:
                   2199:        /*
1.46      schwarze 2200:         * Determine scope.
                   2201:         * If there is nothing on the line after the conditional,
                   2202:         * not even whitespace, use next-line scope.
1.5       schwarze 2203:         */
1.2       schwarze 2204:
1.110     schwarze 2205:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2206:                r->last->endspan = 2;
                   2207:                goto out;
                   2208:        }
                   2209:
1.110     schwarze 2210:        while (buf->buf[pos] == ' ')
1.46      schwarze 2211:                pos++;
                   2212:
                   2213:        /* An opening brace requests multiline scope. */
1.2       schwarze 2214:
1.110     schwarze 2215:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2216:                r->last->endspan = -1;
                   2217:                pos += 2;
1.144     schwarze 2218:                while (buf->buf[pos] == ' ')
                   2219:                        pos++;
1.46      schwarze 2220:                goto out;
1.80      schwarze 2221:        }
1.2       schwarze 2222:
                   2223:        /*
1.46      schwarze 2224:         * Anything else following the conditional causes
                   2225:         * single-line scope.  Warn if the scope contains
                   2226:         * nothing but trailing whitespace.
1.2       schwarze 2227:         */
                   2228:
1.110     schwarze 2229:        if (buf->buf[pos] == '\0')
1.89      schwarze 2230:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167     schwarze 2231:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2232:
1.46      schwarze 2233:        r->last->endspan = 1;
1.1       schwarze 2234:
1.46      schwarze 2235: out:
1.2       schwarze 2236:        *offs = pos;
1.149     schwarze 2237:        return ROFF_RERUN;
1.1       schwarze 2238: }
                   2239:
1.2       schwarze 2240: static enum rofferr
1.7       schwarze 2241: roff_ds(ROFF_ARGS)
                   2242: {
1.85      schwarze 2243:        char            *string;
                   2244:        const char      *name;
                   2245:        size_t           namesz;
1.10      schwarze 2246:
1.123     schwarze 2247:        /* Ignore groff compatibility mode for now. */
                   2248:
                   2249:        if (tok == ROFF_ds1)
                   2250:                tok = ROFF_ds;
                   2251:        else if (tok == ROFF_as1)
                   2252:                tok = ROFF_as;
                   2253:
1.10      schwarze 2254:        /*
1.85      schwarze 2255:         * The first word is the name of the string.
                   2256:         * If it is empty or terminated by an escape sequence,
                   2257:         * abort the `ds' request without defining anything.
1.10      schwarze 2258:         */
1.7       schwarze 2259:
1.110     schwarze 2260:        name = string = buf->buf + pos;
                   2261:        if (*name == '\0')
1.149     schwarze 2262:                return ROFF_IGN;
1.7       schwarze 2263:
1.85      schwarze 2264:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2265:        if (name[namesz] == '\\')
1.149     schwarze 2266:                return ROFF_IGN;
1.85      schwarze 2267:
                   2268:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2269:        if (*string == '"')
1.7       schwarze 2270:                string++;
                   2271:
1.10      schwarze 2272:        /* The rest is the value. */
1.85      schwarze 2273:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2274:            ROFF_as == tok);
1.183     schwarze 2275:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149     schwarze 2276:        return ROFF_IGN;
1.7       schwarze 2277: }
                   2278:
1.77      schwarze 2279: /*
                   2280:  * Parse a single operator, one or two characters long.
                   2281:  * If the operator is recognized, return success and advance the
                   2282:  * parse point, else return failure and let the parse point unchanged.
                   2283:  */
                   2284: static int
                   2285: roff_getop(const char *v, int *pos, char *res)
                   2286: {
                   2287:
                   2288:        *res = v[*pos];
                   2289:
                   2290:        switch (*res) {
1.80      schwarze 2291:        case '+':
                   2292:        case '-':
                   2293:        case '*':
                   2294:        case '/':
                   2295:        case '%':
                   2296:        case '&':
                   2297:        case ':':
1.77      schwarze 2298:                break;
                   2299:        case '<':
                   2300:                switch (v[*pos + 1]) {
1.80      schwarze 2301:                case '=':
1.77      schwarze 2302:                        *res = 'l';
                   2303:                        (*pos)++;
                   2304:                        break;
1.80      schwarze 2305:                case '>':
1.77      schwarze 2306:                        *res = '!';
                   2307:                        (*pos)++;
                   2308:                        break;
1.80      schwarze 2309:                case '?':
1.77      schwarze 2310:                        *res = 'i';
                   2311:                        (*pos)++;
                   2312:                        break;
                   2313:                default:
                   2314:                        break;
                   2315:                }
                   2316:                break;
                   2317:        case '>':
                   2318:                switch (v[*pos + 1]) {
1.80      schwarze 2319:                case '=':
1.77      schwarze 2320:                        *res = 'g';
                   2321:                        (*pos)++;
                   2322:                        break;
1.80      schwarze 2323:                case '?':
1.77      schwarze 2324:                        *res = 'a';
                   2325:                        (*pos)++;
                   2326:                        break;
                   2327:                default:
                   2328:                        break;
                   2329:                }
                   2330:                break;
                   2331:        case '=':
                   2332:                if ('=' == v[*pos + 1])
                   2333:                        (*pos)++;
                   2334:                break;
                   2335:        default:
1.149     schwarze 2336:                return 0;
1.77      schwarze 2337:        }
                   2338:        (*pos)++;
                   2339:
1.149     schwarze 2340:        return *res;
1.77      schwarze 2341: }
                   2342:
                   2343: /*
                   2344:  * Evaluate either a parenthesized numeric expression
                   2345:  * or a single signed integer number.
                   2346:  */
                   2347: static int
1.106     schwarze 2348: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2349:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2350: {
                   2351:
                   2352:        if ('(' != v[*pos])
1.149     schwarze 2353:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2354:
                   2355:        (*pos)++;
1.133     schwarze 2356:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2357:                return 0;
1.77      schwarze 2358:
1.79      schwarze 2359:        /*
                   2360:         * Omission of the closing parenthesis
                   2361:         * is an error in validation mode,
                   2362:         * but ignored in evaluation mode.
                   2363:         */
                   2364:
1.77      schwarze 2365:        if (')' == v[*pos])
                   2366:                (*pos)++;
1.79      schwarze 2367:        else if (NULL == res)
1.149     schwarze 2368:                return 0;
1.77      schwarze 2369:
1.149     schwarze 2370:        return 1;
1.77      schwarze 2371: }
                   2372:
                   2373: /*
                   2374:  * Evaluate a complete numeric expression.
                   2375:  * Proceed left to right, there is no concept of precedence.
                   2376:  */
                   2377: static int
1.106     schwarze 2378: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2379:        int *pos, int *res, int flags)
1.77      schwarze 2380: {
                   2381:        int              mypos, operand2;
                   2382:        char             operator;
                   2383:
                   2384:        if (NULL == pos) {
                   2385:                mypos = 0;
                   2386:                pos = &mypos;
                   2387:        }
                   2388:
1.133     schwarze 2389:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2390:                while (isspace((unsigned char)v[*pos]))
                   2391:                        (*pos)++;
                   2392:
1.133     schwarze 2393:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2394:                return 0;
1.77      schwarze 2395:
                   2396:        while (1) {
1.133     schwarze 2397:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2398:                        while (isspace((unsigned char)v[*pos]))
                   2399:                                (*pos)++;
                   2400:
                   2401:                if ( ! roff_getop(v, pos, &operator))
                   2402:                        break;
                   2403:
1.133     schwarze 2404:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2405:                        while (isspace((unsigned char)v[*pos]))
                   2406:                                (*pos)++;
                   2407:
1.133     schwarze 2408:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2409:                        return 0;
1.77      schwarze 2410:
1.133     schwarze 2411:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2412:                        while (isspace((unsigned char)v[*pos]))
                   2413:                                (*pos)++;
1.79      schwarze 2414:
                   2415:                if (NULL == res)
                   2416:                        continue;
1.77      schwarze 2417:
                   2418:                switch (operator) {
1.80      schwarze 2419:                case '+':
1.77      schwarze 2420:                        *res += operand2;
                   2421:                        break;
1.80      schwarze 2422:                case '-':
1.77      schwarze 2423:                        *res -= operand2;
                   2424:                        break;
1.80      schwarze 2425:                case '*':
1.77      schwarze 2426:                        *res *= operand2;
                   2427:                        break;
1.80      schwarze 2428:                case '/':
1.116     schwarze 2429:                        if (operand2 == 0) {
1.106     schwarze 2430:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2431:                                        r->parse, ln, *pos, v);
                   2432:                                *res = 0;
                   2433:                                break;
                   2434:                        }
1.77      schwarze 2435:                        *res /= operand2;
                   2436:                        break;
1.80      schwarze 2437:                case '%':
1.116     schwarze 2438:                        if (operand2 == 0) {
                   2439:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2440:                                        r->parse, ln, *pos, v);
                   2441:                                *res = 0;
                   2442:                                break;
                   2443:                        }
1.77      schwarze 2444:                        *res %= operand2;
                   2445:                        break;
1.80      schwarze 2446:                case '<':
1.77      schwarze 2447:                        *res = *res < operand2;
                   2448:                        break;
1.80      schwarze 2449:                case '>':
1.77      schwarze 2450:                        *res = *res > operand2;
                   2451:                        break;
1.80      schwarze 2452:                case 'l':
1.77      schwarze 2453:                        *res = *res <= operand2;
                   2454:                        break;
1.80      schwarze 2455:                case 'g':
1.77      schwarze 2456:                        *res = *res >= operand2;
                   2457:                        break;
1.80      schwarze 2458:                case '=':
1.77      schwarze 2459:                        *res = *res == operand2;
                   2460:                        break;
1.80      schwarze 2461:                case '!':
1.77      schwarze 2462:                        *res = *res != operand2;
                   2463:                        break;
1.80      schwarze 2464:                case '&':
1.77      schwarze 2465:                        *res = *res && operand2;
                   2466:                        break;
1.80      schwarze 2467:                case ':':
1.77      schwarze 2468:                        *res = *res || operand2;
                   2469:                        break;
1.80      schwarze 2470:                case 'i':
1.77      schwarze 2471:                        if (operand2 < *res)
                   2472:                                *res = operand2;
                   2473:                        break;
1.80      schwarze 2474:                case 'a':
1.77      schwarze 2475:                        if (operand2 > *res)
                   2476:                                *res = operand2;
                   2477:                        break;
                   2478:                default:
                   2479:                        abort();
                   2480:                }
                   2481:        }
1.149     schwarze 2482:        return 1;
1.77      schwarze 2483: }
                   2484:
1.138     schwarze 2485: /* --- register management ------------------------------------------------ */
                   2486:
1.52      schwarze 2487: void
1.60      schwarze 2488: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2489: {
1.52      schwarze 2490:        struct roffreg  *reg;
                   2491:
                   2492:        /* Search for an existing register with the same name. */
                   2493:        reg = r->regtab;
                   2494:
                   2495:        while (reg && strcmp(name, reg->key.p))
                   2496:                reg = reg->next;
1.41      schwarze 2497:
1.52      schwarze 2498:        if (NULL == reg) {
                   2499:                /* Create a new register. */
                   2500:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2501:                reg->key.p = mandoc_strdup(name);
                   2502:                reg->key.sz = strlen(name);
1.60      schwarze 2503:                reg->val = 0;
1.52      schwarze 2504:                reg->next = r->regtab;
                   2505:                r->regtab = reg;
                   2506:        }
                   2507:
1.60      schwarze 2508:        if ('+' == sign)
                   2509:                reg->val += val;
                   2510:        else if ('-' == sign)
                   2511:                reg->val -= val;
                   2512:        else
                   2513:                reg->val = val;
1.41      schwarze 2514: }
                   2515:
1.65      schwarze 2516: /*
                   2517:  * Handle some predefined read-only number registers.
                   2518:  * For now, return -1 if the requested register is not predefined;
                   2519:  * in case a predefined read-only register having the value -1
                   2520:  * were to turn up, another special value would have to be chosen.
                   2521:  */
                   2522: static int
1.145     schwarze 2523: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2524: {
                   2525:
                   2526:        switch (*name) {
1.145     schwarze 2527:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2528:                return r->argc;
1.80      schwarze 2529:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2530:                return 0;
1.80      schwarze 2531:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2532:                return 1;
1.80      schwarze 2533:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2534:                return 24;
1.80      schwarze 2535:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2536:                return 0;
1.80      schwarze 2537:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2538:                return 1;
1.80      schwarze 2539:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2540:                return 40;
1.65      schwarze 2541:        default:
1.149     schwarze 2542:                return -1;
1.65      schwarze 2543:        }
                   2544: }
                   2545:
1.53      schwarze 2546: int
1.52      schwarze 2547: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2548: {
1.52      schwarze 2549:        struct roffreg  *reg;
1.65      schwarze 2550:        int              val;
                   2551:
                   2552:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145     schwarze 2553:                val = roff_getregro(r, name + 1);
1.65      schwarze 2554:                if (-1 != val)
1.149     schwarze 2555:                        return val;
1.65      schwarze 2556:        }
1.52      schwarze 2557:
                   2558:        for (reg = r->regtab; reg; reg = reg->next)
                   2559:                if (0 == strcmp(name, reg->key.p))
1.149     schwarze 2560:                        return reg->val;
1.53      schwarze 2561:
1.149     schwarze 2562:        return 0;
1.53      schwarze 2563: }
                   2564:
                   2565: static int
                   2566: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2567: {
                   2568:        struct roffreg  *reg;
1.65      schwarze 2569:        int              val;
                   2570:
                   2571:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2572:                val = roff_getregro(r, name + 1);
1.65      schwarze 2573:                if (-1 != val)
1.149     schwarze 2574:                        return val;
1.65      schwarze 2575:        }
1.53      schwarze 2576:
                   2577:        for (reg = r->regtab; reg; reg = reg->next)
                   2578:                if (len == reg->key.sz &&
                   2579:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2580:                        return reg->val;
1.143     schwarze 2581:
1.149     schwarze 2582:        return 0;
1.143     schwarze 2583: }
                   2584:
                   2585: static int
                   2586: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2587: {
                   2588:        struct roffreg  *reg;
                   2589:        int              val;
                   2590:
                   2591:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2592:                val = roff_getregro(r, name + 1);
1.143     schwarze 2593:                if (-1 != val)
1.149     schwarze 2594:                        return 1;
1.143     schwarze 2595:        }
                   2596:
                   2597:        for (reg = r->regtab; reg; reg = reg->next)
                   2598:                if (len == reg->key.sz &&
                   2599:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2600:                        return 1;
1.41      schwarze 2601:
1.149     schwarze 2602:        return 0;
1.41      schwarze 2603: }
                   2604:
1.52      schwarze 2605: static void
                   2606: roff_freereg(struct roffreg *reg)
1.41      schwarze 2607: {
1.52      schwarze 2608:        struct roffreg  *old_reg;
1.41      schwarze 2609:
1.52      schwarze 2610:        while (NULL != reg) {
                   2611:                free(reg->key.p);
                   2612:                old_reg = reg;
                   2613:                reg = reg->next;
                   2614:                free(old_reg);
                   2615:        }
1.41      schwarze 2616: }
1.7       schwarze 2617:
                   2618: static enum rofferr
1.6       schwarze 2619: roff_nr(ROFF_ARGS)
1.1       schwarze 2620: {
1.85      schwarze 2621:        char            *key, *val;
                   2622:        size_t           keysz;
1.37      schwarze 2623:        int              iv;
1.60      schwarze 2624:        char             sign;
1.6       schwarze 2625:
1.110     schwarze 2626:        key = val = buf->buf + pos;
                   2627:        if (*key == '\0')
1.149     schwarze 2628:                return ROFF_IGN;
1.85      schwarze 2629:
                   2630:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2631:        if (key[keysz] == '\\')
1.149     schwarze 2632:                return ROFF_IGN;
1.85      schwarze 2633:        key[keysz] = '\0';
1.6       schwarze 2634:
1.60      schwarze 2635:        sign = *val;
1.110     schwarze 2636:        if (sign == '+' || sign == '-')
1.60      schwarze 2637:                val++;
                   2638:
1.133     schwarze 2639:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77      schwarze 2640:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2641:
1.149     schwarze 2642:        return ROFF_IGN;
1.76      schwarze 2643: }
                   2644:
                   2645: static enum rofferr
                   2646: roff_rr(ROFF_ARGS)
                   2647: {
                   2648:        struct roffreg  *reg, **prev;
1.85      schwarze 2649:        char            *name, *cp;
                   2650:        size_t           namesz;
1.76      schwarze 2651:
1.110     schwarze 2652:        name = cp = buf->buf + pos;
                   2653:        if (*name == '\0')
1.149     schwarze 2654:                return ROFF_IGN;
1.85      schwarze 2655:        namesz = roff_getname(r, &cp, ln, pos);
                   2656:        name[namesz] = '\0';
1.76      schwarze 2657:
                   2658:        prev = &r->regtab;
                   2659:        while (1) {
                   2660:                reg = *prev;
1.110     schwarze 2661:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2662:                        break;
                   2663:                prev = &reg->next;
                   2664:        }
1.110     schwarze 2665:        if (reg != NULL) {
1.76      schwarze 2666:                *prev = reg->next;
                   2667:                free(reg->key.p);
                   2668:                free(reg);
                   2669:        }
1.149     schwarze 2670:        return ROFF_IGN;
1.29      schwarze 2671: }
                   2672:
1.138     schwarze 2673: /* --- handler functions for roff requests -------------------------------- */
                   2674:
1.29      schwarze 2675: static enum rofferr
                   2676: roff_rm(ROFF_ARGS)
                   2677: {
                   2678:        const char       *name;
                   2679:        char             *cp;
1.85      schwarze 2680:        size_t            namesz;
1.29      schwarze 2681:
1.110     schwarze 2682:        cp = buf->buf + pos;
                   2683:        while (*cp != '\0') {
1.85      schwarze 2684:                name = cp;
1.110     schwarze 2685:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2686:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183     schwarze 2687:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.110     schwarze 2688:                if (name[namesz] == '\\')
1.85      schwarze 2689:                        break;
1.29      schwarze 2690:        }
1.149     schwarze 2691:        return ROFF_IGN;
1.51      schwarze 2692: }
                   2693:
                   2694: static enum rofferr
                   2695: roff_it(ROFF_ARGS)
                   2696: {
                   2697:        int              iv;
                   2698:
                   2699:        /* Parse the number of lines. */
1.133     schwarze 2700:
                   2701:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2702:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2703:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2704:                return ROFF_IGN;
1.51      schwarze 2705:        }
                   2706:
1.134     schwarze 2707:        while (isspace((unsigned char)buf->buf[pos]))
                   2708:                pos++;
                   2709:
                   2710:        /*
                   2711:         * Arm the input line trap.
                   2712:         * Special-casing "an-trap" is an ugly workaround to cope
                   2713:         * with DocBook stupidly fiddling with man(7) internals.
                   2714:         */
1.133     schwarze 2715:
1.51      schwarze 2716:        roffit_lines = iv;
1.134     schwarze 2717:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2718:            strcmp(buf->buf + pos, "an-trap") ?
                   2719:            buf->buf + pos : "br");
1.149     schwarze 2720:        return ROFF_IGN;
1.47      schwarze 2721: }
                   2722:
                   2723: static enum rofferr
                   2724: roff_Dd(ROFF_ARGS)
                   2725: {
                   2726:        const char *const       *cp;
                   2727:
1.99      schwarze 2728:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47      schwarze 2729:                for (cp = __mdoc_reserved; *cp; cp++)
                   2730:                        roff_setstr(r, *cp, NULL, 0);
                   2731:
1.99      schwarze 2732:        if (r->format == 0)
                   2733:                r->format = MPARSE_MDOC;
                   2734:
1.149     schwarze 2735:        return ROFF_CONT;
1.47      schwarze 2736: }
                   2737:
                   2738: static enum rofferr
                   2739: roff_TH(ROFF_ARGS)
                   2740: {
                   2741:        const char *const       *cp;
                   2742:
1.99      schwarze 2743:        if ((r->options & MPARSE_QUICK) == 0)
1.47      schwarze 2744:                for (cp = __man_reserved; *cp; cp++)
                   2745:                        roff_setstr(r, *cp, NULL, 0);
                   2746:
1.99      schwarze 2747:        if (r->format == 0)
                   2748:                r->format = MPARSE_MAN;
                   2749:
1.149     schwarze 2750:        return ROFF_CONT;
1.14      schwarze 2751: }
                   2752:
                   2753: static enum rofferr
1.27      schwarze 2754: roff_TE(ROFF_ARGS)
                   2755: {
                   2756:
                   2757:        if (NULL == r->tbl)
1.94      schwarze 2758:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2759:                    ln, ppos, "TE");
1.130     schwarze 2760:        else if ( ! tbl_end(&r->tbl)) {
                   2761:                free(buf->buf);
                   2762:                buf->buf = mandoc_strdup(".sp");
                   2763:                buf->sz = 4;
1.149     schwarze 2764:                return ROFF_REPARSE;
1.130     schwarze 2765:        }
1.149     schwarze 2766:        return ROFF_IGN;
1.27      schwarze 2767: }
                   2768:
                   2769: static enum rofferr
                   2770: roff_T_(ROFF_ARGS)
                   2771: {
                   2772:
                   2773:        if (NULL == r->tbl)
1.94      schwarze 2774:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2775:                    ln, ppos, "T&");
1.27      schwarze 2776:        else
1.168     schwarze 2777:                tbl_restart(ln, ppos, r->tbl);
1.27      schwarze 2778:
1.149     schwarze 2779:        return ROFF_IGN;
1.27      schwarze 2780: }
                   2781:
1.102     schwarze 2782: /*
                   2783:  * Handle in-line equation delimiters.
                   2784:  */
                   2785: static enum rofferr
1.110     schwarze 2786: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2787: {
1.105     schwarze 2788:        char            *cp1, *cp2;
                   2789:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2790:
1.102     schwarze 2791:        /*
                   2792:         * Outside equations, look for an opening delimiter.
                   2793:         * If we are inside an equation, we already know it is
                   2794:         * in-line, or this function wouldn't have been called;
                   2795:         * so look for a closing delimiter.
                   2796:         */
                   2797:
1.110     schwarze 2798:        cp1 = buf->buf + pos;
1.102     schwarze 2799:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2800:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2801:        if (cp2 == NULL)
1.149     schwarze 2802:                return ROFF_CONT;
1.102     schwarze 2803:
1.105     schwarze 2804:        *cp2++ = '\0';
                   2805:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2806:
                   2807:        /* Handle preceding text, protecting whitespace. */
                   2808:
1.110     schwarze 2809:        if (*buf->buf != '\0') {
1.105     schwarze 2810:                if (r->eqn == NULL)
                   2811:                        bef_pr = "\\&";
                   2812:                bef_nl = "\n";
                   2813:        }
                   2814:
                   2815:        /*
                   2816:         * Prepare replacing the delimiter with an equation macro
                   2817:         * and drop leading white space from the equation.
                   2818:         */
1.102     schwarze 2819:
1.105     schwarze 2820:        if (r->eqn == NULL) {
                   2821:                while (*cp2 == ' ')
                   2822:                        cp2++;
                   2823:                mac = ".EQ";
                   2824:        } else
                   2825:                mac = ".EN";
                   2826:
                   2827:        /* Handle following text, protecting whitespace. */
                   2828:
                   2829:        if (*cp2 != '\0') {
                   2830:                aft_nl = "\n";
                   2831:                if (r->eqn != NULL)
                   2832:                        aft_pr = "\\&";
                   2833:        }
                   2834:
                   2835:        /* Do the actual replacement. */
                   2836:
1.110     schwarze 2837:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2838:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2839:        free(buf->buf);
                   2840:        buf->buf = cp1;
1.102     schwarze 2841:
                   2842:        /* Toggle the in-line state of the eqn subsystem. */
                   2843:
                   2844:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2845:        return ROFF_REPARSE;
1.41      schwarze 2846: }
                   2847:
1.107     schwarze 2848: static enum rofferr
                   2849: roff_EQ(ROFF_ARGS)
1.32      schwarze 2850: {
1.41      schwarze 2851:        struct eqn_node *e;
1.32      schwarze 2852:
1.110     schwarze 2853:        assert(r->eqn == NULL);
1.107     schwarze 2854:        e = eqn_alloc(ppos, ln, r->parse);
1.32      schwarze 2855:
1.102     schwarze 2856:        if (r->last_eqn) {
1.32      schwarze 2857:                r->last_eqn->next = e;
1.102     schwarze 2858:                e->delim = r->last_eqn->delim;
                   2859:                e->odelim = r->last_eqn->odelim;
                   2860:                e->cdelim = r->last_eqn->cdelim;
                   2861:        } else
1.32      schwarze 2862:                r->first_eqn = r->last_eqn = e;
                   2863:
                   2864:        r->eqn = r->last_eqn = e;
1.41      schwarze 2865:
1.110     schwarze 2866:        if (buf->buf[pos] != '\0')
1.107     schwarze 2867:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2868:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2869:
1.149     schwarze 2870:        return ROFF_IGN;
1.32      schwarze 2871: }
                   2872:
                   2873: static enum rofferr
                   2874: roff_EN(ROFF_ARGS)
                   2875: {
                   2876:
1.94      schwarze 2877:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.149     schwarze 2878:        return ROFF_IGN;
1.32      schwarze 2879: }
                   2880:
                   2881: static enum rofferr
1.27      schwarze 2882: roff_TS(ROFF_ARGS)
                   2883: {
1.49      schwarze 2884:        struct tbl_node *tbl;
1.27      schwarze 2885:
                   2886:        if (r->tbl) {
1.94      schwarze 2887:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2888:                    ln, ppos, "TS breaks TS");
1.41      schwarze 2889:                tbl_end(&r->tbl);
1.27      schwarze 2890:        }
                   2891:
1.49      schwarze 2892:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2893:
                   2894:        if (r->last_tbl)
1.49      schwarze 2895:                r->last_tbl->next = tbl;
1.27      schwarze 2896:        else
1.49      schwarze 2897:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 2898:
1.49      schwarze 2899:        r->tbl = r->last_tbl = tbl;
1.169     schwarze 2900:        return ROFF_IGN;
                   2901: }
                   2902:
                   2903: static enum rofferr
                   2904: roff_onearg(ROFF_ARGS)
                   2905: {
                   2906:        struct roff_node        *n;
                   2907:        char                    *cp;
1.177     schwarze 2908:        int                      npos;
1.169     schwarze 2909:
1.174     schwarze 2910:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
                   2911:            (tok == ROFF_sp || tok == ROFF_ti))
                   2912:                man_breakscope(r->man, tok);
                   2913:
1.181     schwarze 2914:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177     schwarze 2915:                r->man->last = roffce_node;
                   2916:                r->man->next = ROFF_NEXT_SIBLING;
                   2917:        }
                   2918:
1.169     schwarze 2919:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2920:        n = r->man->last;
                   2921:
                   2922:        cp = buf->buf + pos;
                   2923:        if (*cp != '\0') {
                   2924:                while (*cp != '\0' && *cp != ' ')
                   2925:                        cp++;
                   2926:                while (*cp == ' ')
                   2927:                        *cp++ = '\0';
                   2928:                if (*cp != '\0')
                   2929:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS,
                   2930:                            r->parse, ln, cp - buf->buf,
                   2931:                            "%s ... %s", roff_name[tok], cp);
                   2932:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172     schwarze 2933:        }
                   2934:
1.181     schwarze 2935:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   2936:                if (r->man->last->type == ROFFT_ELEM) {
1.177     schwarze 2937:                        roff_word_alloc(r->man, ln, pos, "1");
                   2938:                        r->man->last->flags |= NODE_NOSRC;
                   2939:                }
                   2940:                npos = 0;
                   2941:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   2942:                    &roffce_lines, 0) == 0) {
                   2943:                        mandoc_vmsg(MANDOCERR_CE_NONUM,
                   2944:                            r->parse, ln, pos, "ce %s", buf->buf + pos);
                   2945:                        roffce_lines = 1;
                   2946:                }
                   2947:                if (roffce_lines < 1) {
                   2948:                        r->man->last = r->man->last->parent;
                   2949:                        roffce_node = NULL;
                   2950:                        roffce_lines = 0;
                   2951:                } else
                   2952:                        roffce_node = r->man->last->parent;
                   2953:        } else {
                   2954:                n->flags |= NODE_VALID | NODE_ENDED;
                   2955:                r->man->last = n;
                   2956:        }
                   2957:        n->flags |= NODE_LINE;
1.172     schwarze 2958:        r->man->next = ROFF_NEXT_SIBLING;
                   2959:        return ROFF_IGN;
                   2960: }
                   2961:
                   2962: static enum rofferr
                   2963: roff_manyarg(ROFF_ARGS)
                   2964: {
                   2965:        struct roff_node        *n;
                   2966:        char                    *sp, *ep;
                   2967:
                   2968:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2969:        n = r->man->last;
                   2970:
                   2971:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   2972:                while (*ep != '\0' && *ep != ' ')
                   2973:                        ep++;
                   2974:                while (*ep == ' ')
                   2975:                        *ep++ = '\0';
                   2976:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169     schwarze 2977:        }
                   2978:
                   2979:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   2980:        r->man->last = n;
                   2981:        r->man->next = ROFF_NEXT_SIBLING;
1.149     schwarze 2982:        return ROFF_IGN;
1.123     schwarze 2983: }
                   2984:
                   2985: static enum rofferr
1.183     schwarze 2986: roff_als(ROFF_ARGS)
                   2987: {
                   2988:        char            *oldn, *newn, *end, *value;
                   2989:        size_t           oldsz, newsz, valsz;
                   2990:
                   2991:        newn = oldn = buf->buf + pos;
                   2992:        if (*newn == '\0')
                   2993:                return ROFF_IGN;
                   2994:
                   2995:        newsz = roff_getname(r, &oldn, ln, pos);
                   2996:        if (newn[newsz] == '\\' || *oldn == '\0')
                   2997:                return ROFF_IGN;
                   2998:
                   2999:        end = oldn;
                   3000:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3001:        if (oldsz == 0)
                   3002:                return ROFF_IGN;
                   3003:
                   3004:        valsz = mandoc_asprintf(&value, ".%.*s \\$*\n", (int)oldsz, oldn);
                   3005:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3006:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3007:        free(value);
                   3008:        return ROFF_IGN;
                   3009: }
                   3010:
                   3011: static enum rofferr
1.168     schwarze 3012: roff_br(ROFF_ARGS)
1.123     schwarze 3013: {
1.174     schwarze 3014:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3015:                man_breakscope(r->man, ROFF_br);
1.168     schwarze 3016:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
                   3017:        if (buf->buf[pos] != '\0')
                   3018:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   3019:                    "%s %s", roff_name[tok], buf->buf + pos);
                   3020:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3021:        r->man->next = ROFF_NEXT_SIBLING;
                   3022:        return ROFF_IGN;
1.27      schwarze 3023: }
                   3024:
                   3025: static enum rofferr
1.48      schwarze 3026: roff_cc(ROFF_ARGS)
                   3027: {
                   3028:        const char      *p;
                   3029:
1.110     schwarze 3030:        p = buf->buf + pos;
1.48      schwarze 3031:
1.110     schwarze 3032:        if (*p == '\0' || (r->control = *p++) == '.')
1.175     schwarze 3033:                r->control = '\0';
1.48      schwarze 3034:
1.110     schwarze 3035:        if (*p != '\0')
1.132     schwarze 3036:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3037:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 3038:
1.149     schwarze 3039:        return ROFF_IGN;
1.48      schwarze 3040: }
                   3041:
                   3042: static enum rofferr
1.175     schwarze 3043: roff_ec(ROFF_ARGS)
                   3044: {
                   3045:        const char      *p;
                   3046:
                   3047:        p = buf->buf + pos;
                   3048:        if (*p == '\0')
                   3049:                r->escape = '\\';
                   3050:        else {
                   3051:                r->escape = *p;
                   3052:                if (*++p != '\0')
                   3053:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3054:                            ln, p - buf->buf, "ec ... %s", p);
                   3055:        }
                   3056:        return ROFF_IGN;
                   3057: }
                   3058:
                   3059: static enum rofferr
                   3060: roff_eo(ROFF_ARGS)
                   3061: {
                   3062:        r->escape = '\0';
                   3063:        if (buf->buf[pos] != '\0')
                   3064:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
                   3065:                    ln, pos, "eo %s", buf->buf + pos);
                   3066:        return ROFF_IGN;
                   3067: }
                   3068:
                   3069: static enum rofferr
1.42      schwarze 3070: roff_tr(ROFF_ARGS)
                   3071: {
                   3072:        const char      *p, *first, *second;
                   3073:        size_t           fsz, ssz;
                   3074:        enum mandoc_esc  esc;
                   3075:
1.110     schwarze 3076:        p = buf->buf + pos;
1.42      schwarze 3077:
1.110     schwarze 3078:        if (*p == '\0') {
1.132     schwarze 3079:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 3080:                return ROFF_IGN;
1.42      schwarze 3081:        }
                   3082:
1.110     schwarze 3083:        while (*p != '\0') {
1.42      schwarze 3084:                fsz = ssz = 1;
                   3085:
                   3086:                first = p++;
1.110     schwarze 3087:                if (*first == '\\') {
1.42      schwarze 3088:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3089:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3090:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3091:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 3092:                                return ROFF_IGN;
1.42      schwarze 3093:                        }
                   3094:                        fsz = (size_t)(p - first);
                   3095:                }
                   3096:
                   3097:                second = p++;
1.110     schwarze 3098:                if (*second == '\\') {
1.42      schwarze 3099:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3100:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3101:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3102:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 3103:                                return ROFF_IGN;
1.42      schwarze 3104:                        }
                   3105:                        ssz = (size_t)(p - second);
1.110     schwarze 3106:                } else if (*second == '\0') {
1.132     schwarze 3107:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   3108:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 3109:                        second = " ";
                   3110:                        p--;
                   3111:                }
                   3112:
                   3113:                if (fsz > 1) {
1.80      schwarze 3114:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3115:                            second, ssz, 0);
1.42      schwarze 3116:                        continue;
                   3117:                }
                   3118:
1.110     schwarze 3119:                if (r->xtab == NULL)
1.80      schwarze 3120:                        r->xtab = mandoc_calloc(128,
                   3121:                            sizeof(struct roffstr));
1.42      schwarze 3122:
                   3123:                free(r->xtab[(int)*first].p);
                   3124:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3125:                r->xtab[(int)*first].sz = ssz;
                   3126:        }
                   3127:
1.149     schwarze 3128:        return ROFF_IGN;
1.42      schwarze 3129: }
                   3130:
                   3131: static enum rofferr
1.178     schwarze 3132: roff_rn(ROFF_ARGS)
                   3133: {
                   3134:        const char      *value;
                   3135:        char            *oldn, *newn, *end;
                   3136:        size_t           oldsz, newsz;
                   3137:
                   3138:        oldn = newn = buf->buf + pos;
                   3139:        if (*oldn == '\0')
                   3140:                return ROFF_IGN;
                   3141:
                   3142:        oldsz = roff_getname(r, &newn, ln, pos);
                   3143:        if (oldn[oldsz] == '\\' || *newn == '\0')
                   3144:                return ROFF_IGN;
                   3145:
                   3146:        end = newn;
                   3147:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3148:        if (newsz == 0)
                   3149:                return ROFF_IGN;
                   3150:
                   3151:        /*
                   3152:         * Rename a user-defined macro bearing the old name,
                   3153:         * overriding an existing renamed high-level macro
                   3154:         * bearing the new name, if that exists.
                   3155:         */
                   3156:
                   3157:        if ((value = roff_getstrn(r, oldn, oldsz)) != NULL) {
                   3158:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3159:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3160:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3161:                return ROFF_IGN;
                   3162:        }
                   3163:
                   3164:        /*
                   3165:         * Rename a high-level macro bearing the old name,
                   3166:         * either renaming it a second time if it was already
                   3167:         * renamed before, or renaming it for the first time.
                   3168:         * In both cases, override an existing user-defined
                   3169:         * macro bearing the new name, if that exists.
                   3170:         */
                   3171:
                   3172:        if ((value = roff_getrenn(r, oldn, oldsz)) != NULL) {
                   3173:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3174:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
                   3175:        } else
                   3176:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
                   3177:        roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3178:        return ROFF_IGN;
                   3179: }
                   3180:
                   3181: static enum rofferr
1.14      schwarze 3182: roff_so(ROFF_ARGS)
                   3183: {
1.121     schwarze 3184:        char *name, *cp;
1.15      schwarze 3185:
1.110     schwarze 3186:        name = buf->buf + pos;
1.97      schwarze 3187:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 3188:
1.22      schwarze 3189:        /*
                   3190:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3191:         * opening anything that's not in our cwd or anything beneath
                   3192:         * it.  Thus, explicitly disallow traversing up the file-system
                   3193:         * or using absolute paths.
                   3194:         */
                   3195:
1.110     schwarze 3196:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 3197:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   3198:                    ".so %s", name);
1.121     schwarze 3199:                buf->sz = mandoc_asprintf(&cp,
                   3200:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3201:                free(buf->buf);
                   3202:                buf->buf = cp;
                   3203:                *offs = 0;
1.149     schwarze 3204:                return ROFF_REPARSE;
1.14      schwarze 3205:        }
                   3206:
                   3207:        *offs = pos;
1.149     schwarze 3208:        return ROFF_SO;
1.7       schwarze 3209: }
                   3210:
1.138     schwarze 3211: /* --- user defined strings and macros ------------------------------------ */
                   3212:
1.16      schwarze 3213: static enum rofferr
                   3214: roff_userdef(ROFF_ARGS)
1.12      schwarze 3215: {
1.183     schwarze 3216:        const char       *arg[16], *ap;
1.16      schwarze 3217:        char             *cp, *n1, *n2;
1.164     schwarze 3218:        int               expand_count, i, ib, ie;
1.135     schwarze 3219:        size_t            asz, rsz;
1.12      schwarze 3220:
1.16      schwarze 3221:        /*
                   3222:         * Collect pointers to macro argument strings
1.61      schwarze 3223:         * and NUL-terminate them.
1.16      schwarze 3224:         */
1.135     schwarze 3225:
1.145     schwarze 3226:        r->argc = 0;
1.110     schwarze 3227:        cp = buf->buf + pos;
1.183     schwarze 3228:        for (i = 0; i < 16; i++) {
1.145     schwarze 3229:                if (*cp == '\0')
                   3230:                        arg[i] = "";
                   3231:                else {
                   3232:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   3233:                        r->argc = i + 1;
                   3234:                }
                   3235:        }
1.16      schwarze 3236:
                   3237:        /*
                   3238:         * Expand macro arguments.
1.12      schwarze 3239:         */
1.135     schwarze 3240:
                   3241:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 3242:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 3243:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 3244:        expand_count = 0;
1.135     schwarze 3245:        while (*cp != '\0') {
                   3246:
                   3247:                /* Scan ahead for the next argument invocation. */
                   3248:
                   3249:                if (*cp++ != '\\')
                   3250:                        continue;
                   3251:                if (*cp++ != '$')
                   3252:                        continue;
1.146     schwarze 3253:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   3254:                        ib = 0;
                   3255:                        ie = r->argc - 1;
                   3256:                } else {  /* \\$1 .. \\$9 insert one argument */
                   3257:                        ib = ie = *cp - '1';
                   3258:                        if (ib < 0 || ib > 8)
                   3259:                                continue;
                   3260:                }
1.135     schwarze 3261:                cp -= 2;
1.164     schwarze 3262:
                   3263:                /*
                   3264:                 * Prevent infinite recursion.
                   3265:                 */
                   3266:
                   3267:                if (cp >= n2)
                   3268:                        expand_count = 1;
                   3269:                else if (++expand_count > EXPAND_LIMIT) {
                   3270:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   3271:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 3272:                        free(buf->buf);
                   3273:                        buf->buf = n1;
1.164     schwarze 3274:                        return ROFF_IGN;
                   3275:                }
1.135     schwarze 3276:
                   3277:                /*
                   3278:                 * Determine the size of the expanded argument,
                   3279:                 * taking escaping of quotes into account.
                   3280:                 */
                   3281:
1.146     schwarze 3282:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   3283:                for (i = ib; i <= ie; i++) {
                   3284:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3285:                                asz++;
                   3286:                                if (*ap == '"')
                   3287:                                        asz += 3;
                   3288:                        }
1.135     schwarze 3289:                }
                   3290:                if (asz != 3) {
                   3291:
                   3292:                        /*
                   3293:                         * Determine the size of the rest of the
                   3294:                         * unexpanded macro, including the NUL.
                   3295:                         */
                   3296:
                   3297:                        rsz = buf->sz - (cp - n1) - 3;
                   3298:
                   3299:                        /*
                   3300:                         * When shrinking, move before
                   3301:                         * releasing the storage.
                   3302:                         */
                   3303:
                   3304:                        if (asz < 3)
                   3305:                                memmove(cp + asz, cp + 3, rsz);
                   3306:
                   3307:                        /*
                   3308:                         * Resize the storage for the macro
                   3309:                         * and readjust the parse pointer.
                   3310:                         */
                   3311:
                   3312:                        buf->sz += asz - 3;
                   3313:                        n2 = mandoc_realloc(n1, buf->sz);
                   3314:                        cp = n2 + (cp - n1);
                   3315:                        n1 = n2;
                   3316:
                   3317:                        /*
                   3318:                         * When growing, make room
                   3319:                         * for the expanded argument.
                   3320:                         */
                   3321:
                   3322:                        if (asz > 3)
                   3323:                                memmove(cp + asz, cp + 3, rsz);
                   3324:                }
                   3325:
                   3326:                /* Copy the expanded argument, escaping quotes. */
                   3327:
                   3328:                n2 = cp;
1.146     schwarze 3329:                for (i = ib; i <= ie; i++) {
                   3330:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3331:                                if (*ap == '"') {
                   3332:                                        memcpy(n2, "\\(dq", 4);
                   3333:                                        n2 += 4;
                   3334:                                } else
                   3335:                                        *n2++ = *ap;
                   3336:                        }
                   3337:                        if (i < ie)
                   3338:                                *n2++ = ' ';
1.16      schwarze 3339:                }
1.12      schwarze 3340:        }
                   3341:
1.16      schwarze 3342:        /*
                   3343:         * Replace the macro invocation
                   3344:         * by the expanded macro.
                   3345:         */
1.135     schwarze 3346:
1.110     schwarze 3347:        free(buf->buf);
                   3348:        buf->buf = n1;
1.120     schwarze 3349:        *offs = 0;
1.16      schwarze 3350:
1.149     schwarze 3351:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3352:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3353: }
1.28      schwarze 3354:
1.178     schwarze 3355: /*
                   3356:  * Calling a high-level macro that was renamed with .rn.
                   3357:  * r->current_string has already been set up by roff_parse().
                   3358:  */
                   3359: static enum rofferr
                   3360: roff_renamed(ROFF_ARGS)
                   3361: {
                   3362:        char    *nbuf;
                   3363:
                   3364:        buf->sz = mandoc_asprintf(&nbuf, ".%s %s", r->current_string,
                   3365:            buf->buf + pos) + 1;
                   3366:        free(buf->buf);
                   3367:        buf->buf = nbuf;
                   3368:        return ROFF_CONT;
                   3369: }
                   3370:
1.85      schwarze 3371: static size_t
1.28      schwarze 3372: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3373: {
                   3374:        char     *name, *cp;
1.85      schwarze 3375:        size_t    namesz;
1.28      schwarze 3376:
                   3377:        name = *cpp;
                   3378:        if ('\0' == *name)
1.149     schwarze 3379:                return 0;
1.28      schwarze 3380:
1.85      schwarze 3381:        /* Read until end of name and terminate it with NUL. */
                   3382:        for (cp = name; 1; cp++) {
                   3383:                if ('\0' == *cp || ' ' == *cp) {
                   3384:                        namesz = cp - name;
                   3385:                        break;
                   3386:                }
1.28      schwarze 3387:                if ('\\' != *cp)
                   3388:                        continue;
1.88      schwarze 3389:                namesz = cp - name;
                   3390:                if ('{' == cp[1] || '}' == cp[1])
                   3391:                        break;
1.28      schwarze 3392:                cp++;
                   3393:                if ('\\' == *cp)
                   3394:                        continue;
1.97      schwarze 3395:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3396:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3397:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3398:                break;
1.28      schwarze 3399:        }
                   3400:
                   3401:        /* Read past spaces. */
                   3402:        while (' ' == *cp)
                   3403:                cp++;
                   3404:
                   3405:        *cpp = cp;
1.149     schwarze 3406:        return namesz;
1.28      schwarze 3407: }
                   3408:
1.16      schwarze 3409: /*
                   3410:  * Store *string into the user-defined string called *name.
                   3411:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3412:  * append == 0: replace mode
                   3413:  * append == 1: single-line append mode
                   3414:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3415:  */
1.8       schwarze 3416: static void
1.16      schwarze 3417: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3418:        int append)
1.7       schwarze 3419: {
1.42      schwarze 3420:
                   3421:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.80      schwarze 3422:            string ? strlen(string) : 0, append);
1.42      schwarze 3423: }
                   3424:
                   3425: static void
                   3426: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3427:                const char *string, size_t stringsz, int append)
1.42      schwarze 3428: {
                   3429:        struct roffkv   *n;
                   3430:        char            *c;
                   3431:        int              i;
                   3432:        size_t           oldch, newch;
1.7       schwarze 3433:
1.16      schwarze 3434:        /* Search for an existing string with the same name. */
1.42      schwarze 3435:        n = *r;
                   3436:
1.84      schwarze 3437:        while (n && (namesz != n->key.sz ||
                   3438:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3439:                n = n->next;
1.8       schwarze 3440:
                   3441:        if (NULL == n) {
1.16      schwarze 3442:                /* Create a new string table entry. */
1.42      schwarze 3443:                n = mandoc_malloc(sizeof(struct roffkv));
                   3444:                n->key.p = mandoc_strndup(name, namesz);
                   3445:                n->key.sz = namesz;
                   3446:                n->val.p = NULL;
                   3447:                n->val.sz = 0;
                   3448:                n->next = *r;
                   3449:                *r = n;
1.66      schwarze 3450:        } else if (0 == append) {
1.42      schwarze 3451:                free(n->val.p);
                   3452:                n->val.p = NULL;
                   3453:                n->val.sz = 0;
1.16      schwarze 3454:        }
                   3455:
                   3456:        if (NULL == string)
                   3457:                return;
                   3458:
                   3459:        /*
                   3460:         * One additional byte for the '\n' in multiline mode,
                   3461:         * and one for the terminating '\0'.
                   3462:         */
1.66      schwarze 3463:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3464:
                   3465:        if (NULL == n->val.p) {
                   3466:                n->val.p = mandoc_malloc(newch);
                   3467:                *n->val.p = '\0';
1.16      schwarze 3468:                oldch = 0;
                   3469:        } else {
1.42      schwarze 3470:                oldch = n->val.sz;
                   3471:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3472:        }
                   3473:
                   3474:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3475:        c = n->val.p + (int)oldch;
1.16      schwarze 3476:
                   3477:        /* Append new content to the destination buffer. */
1.42      schwarze 3478:        i = 0;
                   3479:        while (i < (int)stringsz) {
1.16      schwarze 3480:                /*
                   3481:                 * Rudimentary roff copy mode:
                   3482:                 * Handle escaped backslashes.
                   3483:                 */
1.42      schwarze 3484:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3485:                        i++;
                   3486:                *c++ = string[i++];
1.16      schwarze 3487:        }
1.8       schwarze 3488:
1.16      schwarze 3489:        /* Append terminating bytes. */
1.66      schwarze 3490:        if (1 < append)
1.16      schwarze 3491:                *c++ = '\n';
1.42      schwarze 3492:
1.16      schwarze 3493:        *c = '\0';
1.42      schwarze 3494:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3495: }
                   3496:
1.8       schwarze 3497: static const char *
                   3498: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7       schwarze 3499: {
1.42      schwarze 3500:        const struct roffkv *n;
1.64      schwarze 3501:        int i;
1.7       schwarze 3502:
1.42      schwarze 3503:        for (n = r->strtab; n; n = n->next)
1.80      schwarze 3504:                if (0 == strncmp(name, n->key.p, len) &&
                   3505:                    '\0' == n->key.p[(int)len])
1.149     schwarze 3506:                        return n->val.p;
1.64      schwarze 3507:
                   3508:        for (i = 0; i < PREDEFS_MAX; i++)
                   3509:                if (0 == strncmp(name, predefs[i].name, len) &&
                   3510:                                '\0' == predefs[i].name[(int)len])
1.149     schwarze 3511:                        return predefs[i].str;
1.178     schwarze 3512:
                   3513:        return NULL;
                   3514: }
                   3515:
                   3516: /*
                   3517:  * Check whether *name is the renamed name of a high-level macro.
                   3518:  * Return the standard name, or NULL if it is not.
                   3519:  */
                   3520: static const char *
                   3521: roff_getrenn(const struct roff *r, const char *name, size_t len)
                   3522: {
                   3523:        const struct roffkv *n;
                   3524:
                   3525:        for (n = r->rentab; n; n = n->next)
                   3526:                if (0 == strncmp(name, n->key.p, len) &&
                   3527:                    '\0' == n->key.p[(int)len])
                   3528:                        return n->val.p;
1.8       schwarze 3529:
1.149     schwarze 3530:        return NULL;
1.7       schwarze 3531: }
                   3532:
1.8       schwarze 3533: static void
1.42      schwarze 3534: roff_freestr(struct roffkv *r)
1.7       schwarze 3535: {
1.42      schwarze 3536:        struct roffkv    *n, *nn;
1.7       schwarze 3537:
1.42      schwarze 3538:        for (n = r; n; n = nn) {
                   3539:                free(n->key.p);
                   3540:                free(n->val.p);
1.7       schwarze 3541:                nn = n->next;
                   3542:                free(n);
                   3543:        }
1.27      schwarze 3544: }
1.138     schwarze 3545:
                   3546: /* --- accessors and utility functions ------------------------------------ */
1.27      schwarze 3547:
                   3548: const struct tbl_span *
                   3549: roff_span(const struct roff *r)
                   3550: {
1.80      schwarze 3551:
1.149     schwarze 3552:        return r->tbl ? tbl_span(r->tbl) : NULL;
1.32      schwarze 3553: }
                   3554:
                   3555: const struct eqn *
                   3556: roff_eqn(const struct roff *r)
                   3557: {
1.80      schwarze 3558:
1.149     schwarze 3559:        return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.42      schwarze 3560: }
                   3561:
                   3562: /*
                   3563:  * Duplicate an input string, making the appropriate character
                   3564:  * conversations (as stipulated by `tr') along the way.
                   3565:  * Returns a heap-allocated string with all the replacements made.
                   3566:  */
                   3567: char *
                   3568: roff_strdup(const struct roff *r, const char *p)
                   3569: {
                   3570:        const struct roffkv *cp;
                   3571:        char            *res;
                   3572:        const char      *pp;
                   3573:        size_t           ssz, sz;
                   3574:        enum mandoc_esc  esc;
                   3575:
                   3576:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3577:                return mandoc_strdup(p);
1.42      schwarze 3578:        else if ('\0' == *p)
1.149     schwarze 3579:                return mandoc_strdup("");
1.42      schwarze 3580:
                   3581:        /*
                   3582:         * Step through each character looking for term matches
                   3583:         * (remember that a `tr' can be invoked with an escape, which is
                   3584:         * a glyph but the escape is multi-character).
                   3585:         * We only do this if the character hash has been initialised
                   3586:         * and the string is >0 length.
                   3587:         */
                   3588:
                   3589:        res = NULL;
                   3590:        ssz = 0;
                   3591:
                   3592:        while ('\0' != *p) {
1.161     schwarze 3593:                assert((unsigned int)*p < 128);
                   3594:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3595:                        sz = r->xtab[(int)*p].sz;
                   3596:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3597:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3598:                        ssz += sz;
                   3599:                        p++;
                   3600:                        continue;
                   3601:                } else if ('\\' != *p) {
                   3602:                        res = mandoc_realloc(res, ssz + 2);
                   3603:                        res[ssz++] = *p++;
                   3604:                        continue;
                   3605:                }
                   3606:
                   3607:                /* Search for term matches. */
                   3608:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3609:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3610:                                break;
                   3611:
                   3612:                if (NULL != cp) {
                   3613:                        /*
                   3614:                         * A match has been found.
                   3615:                         * Append the match to the array and move
                   3616:                         * forward by its keysize.
                   3617:                         */
1.80      schwarze 3618:                        res = mandoc_realloc(res,
                   3619:                            ssz + cp->val.sz + 1);
1.42      schwarze 3620:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3621:                        ssz += cp->val.sz;
                   3622:                        p += (int)cp->key.sz;
                   3623:                        continue;
                   3624:                }
                   3625:
                   3626:                /*
                   3627:                 * Handle escapes carefully: we need to copy
                   3628:                 * over just the escape itself, or else we might
                   3629:                 * do replacements within the escape itself.
                   3630:                 * Make sure to pass along the bogus string.
                   3631:                 */
                   3632:                pp = p++;
                   3633:                esc = mandoc_escape(&p, NULL, NULL);
                   3634:                if (ESCAPE_ERROR == esc) {
                   3635:                        sz = strlen(pp);
                   3636:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3637:                        memcpy(res + ssz, pp, sz);
                   3638:                        break;
                   3639:                }
1.80      schwarze 3640:                /*
                   3641:                 * We bail out on bad escapes.
1.42      schwarze 3642:                 * No need to warn: we already did so when
                   3643:                 * roff_res() was called.
                   3644:                 */
                   3645:                sz = (int)(p - pp);
                   3646:                res = mandoc_realloc(res, ssz + sz + 1);
                   3647:                memcpy(res + ssz, pp, sz);
                   3648:                ssz += sz;
                   3649:        }
                   3650:
                   3651:        res[(int)ssz] = '\0';
1.149     schwarze 3652:        return res;
1.99      schwarze 3653: }
                   3654:
                   3655: int
                   3656: roff_getformat(const struct roff *r)
                   3657: {
                   3658:
1.149     schwarze 3659:        return r->format;
1.48      schwarze 3660: }
                   3661:
                   3662: /*
1.80      schwarze 3663:  * Find out whether a line is a macro line or not.
1.48      schwarze 3664:  * If it is, adjust the current position and return one; if it isn't,
                   3665:  * return zero and don't change the current position.
                   3666:  * If the control character has been set with `.cc', then let that grain
                   3667:  * precedence.
                   3668:  * This is slighly contrary to groff, where using the non-breaking
                   3669:  * control character when `cc' has been invoked will cause the
                   3670:  * non-breaking macro contents to be printed verbatim.
                   3671:  */
                   3672: int
                   3673: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3674: {
                   3675:        int             pos;
                   3676:
                   3677:        pos = *ppos;
                   3678:
1.175     schwarze 3679:        if (r->control != '\0' && cp[pos] == r->control)
1.48      schwarze 3680:                pos++;
1.175     schwarze 3681:        else if (r->control != '\0')
1.149     schwarze 3682:                return 0;
1.48      schwarze 3683:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3684:                pos += 2;
                   3685:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3686:                pos++;
                   3687:        else
1.149     schwarze 3688:                return 0;
1.48      schwarze 3689:
                   3690:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3691:                pos++;
                   3692:
                   3693:        *ppos = pos;
1.149     schwarze 3694:        return 1;
1.1       schwarze 3695: }