[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 1.181

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