[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 1.192

1.192   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.191 2017/07/08 14:51:01 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:
1.187     schwarze   40: /* Types of definitions of macros and strings. */
                     41: #define        ROFFDEF_USER    (1 << 1)  /* User-defined. */
                     42: #define        ROFFDEF_PRE     (1 << 2)  /* Predefined. */
                     43: #define        ROFFDEF_REN     (1 << 3)  /* Renamed standard macro. */
                     44: #define        ROFFDEF_STD     (1 << 4)  /* mdoc(7) or man(7) macro. */
                     45: #define        ROFFDEF_ANY     (ROFFDEF_USER | ROFFDEF_PRE | \
                     46:                         ROFFDEF_REN | ROFFDEF_STD)
                     47:
1.138     schwarze   48: /* --- data types --------------------------------------------------------- */
                     49:
1.41      schwarze   50: /*
1.42      schwarze   51:  * An incredibly-simple string buffer.
                     52:  */
1.8       schwarze   53: struct roffstr {
1.42      schwarze   54:        char            *p; /* nil-terminated buffer */
                     55:        size_t           sz; /* saved strlen(p) */
                     56: };
                     57:
                     58: /*
                     59:  * A key-value roffstr pair as part of a singly-linked list.
                     60:  */
                     61: struct roffkv {
                     62:        struct roffstr   key;
                     63:        struct roffstr   val;
                     64:        struct roffkv   *next; /* next in list */
1.8       schwarze   65: };
                     66:
1.52      schwarze   67: /*
                     68:  * A single number register as part of a singly-linked list.
                     69:  */
                     70: struct roffreg {
                     71:        struct roffstr   key;
1.53      schwarze   72:        int              val;
1.52      schwarze   73:        struct roffreg  *next;
                     74: };
                     75:
1.167     schwarze   76: /*
                     77:  * Association of request and macro names with token IDs.
                     78:  */
                     79: struct roffreq {
                     80:        enum roff_tok    tok;
                     81:        char             name[];
                     82: };
                     83:
1.1       schwarze   84: struct roff {
1.35      schwarze   85:        struct mparse   *parse; /* parse point */
1.168     schwarze   86:        struct roff_man *man; /* mdoc or man parser */
1.1       schwarze   87:        struct roffnode *last; /* leaf of stack */
1.96      schwarze   88:        int             *rstack; /* stack of inverted `ie' values */
1.167     schwarze   89:        struct ohash    *reqtab; /* request lookup table */
1.52      schwarze   90:        struct roffreg  *regtab; /* number registers */
1.42      schwarze   91:        struct roffkv   *strtab; /* user-defined strings & macros */
1.178     schwarze   92:        struct roffkv   *rentab; /* renamed strings & macros */
1.42      schwarze   93:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     94:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze   95:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   96:        struct tbl_node *first_tbl; /* first table parsed */
                     97:        struct tbl_node *last_tbl; /* last table parsed */
                     98:        struct tbl_node *tbl; /* current table being parsed */
1.191     schwarze   99:        struct eqn_node *last_eqn; /* equation parser */
                    100:        struct eqn_node *eqn; /* active equation parser */
1.102     schwarze  101:        int              eqn_inline; /* current equation is inline */
1.96      schwarze  102:        int              options; /* parse options */
                    103:        int              rstacksz; /* current size limit of rstack */
                    104:        int              rstackpos; /* position in rstack */
1.99      schwarze  105:        int              format; /* current file in mdoc or man format */
1.145     schwarze  106:        int              argc; /* number of args of the last macro */
1.96      schwarze  107:        char             control; /* control character */
1.175     schwarze  108:        char             escape; /* escape character */
1.1       schwarze  109: };
                    110:
                    111: struct roffnode {
1.166     schwarze  112:        enum roff_tok    tok; /* type of node */
1.1       schwarze  113:        struct roffnode *parent; /* up one in stack */
                    114:        int              line; /* parse line */
                    115:        int              col; /* parse col */
1.16      schwarze  116:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  117:        char            *end; /* end-rules: custom token */
                    118:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  119:        int              rule; /* current evaluation rule */
1.1       schwarze  120: };
                    121:
                    122: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.166     schwarze  123:                         enum roff_tok tok, /* tok of macro */ \
1.110     schwarze  124:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  125:                         int ln, /* parse line */ \
1.2       schwarze  126:                         int ppos, /* original pos in buffer */ \
                    127:                         int pos, /* current pos in buffer */ \
                    128:                         int *offs /* reset offset of buffer data */
1.1       schwarze  129:
                    130: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    131:
                    132: struct roffmac {
1.2       schwarze  133:        roffproc         proc; /* process new macro */
                    134:        roffproc         text; /* process as child text of macro */
                    135:        roffproc         sub; /* process as child of macro */
                    136:        int              flags;
                    137: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.1       schwarze  138: };
                    139:
1.37      schwarze  140: struct predef {
                    141:        const char      *name; /* predefined input name */
                    142:        const char      *str; /* replacement symbol */
                    143: };
                    144:
                    145: #define        PREDEF(__name, __str) \
                    146:        { (__name), (__str) },
                    147:
1.138     schwarze  148: /* --- function prototypes ------------------------------------------------ */
                    149:
1.42      schwarze  150: static void             roffnode_cleanscope(struct roff *);
                    151: static void             roffnode_pop(struct roff *);
1.166     schwarze  152: static void             roffnode_push(struct roff *, enum roff_tok,
1.42      schwarze  153:                                const char *, int, int);
1.183     schwarze  154: static enum rofferr     roff_als(ROFF_ARGS);
1.2       schwarze  155: static enum rofferr     roff_block(ROFF_ARGS);
                    156: static enum rofferr     roff_block_text(ROFF_ARGS);
                    157: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.168     schwarze  158: static enum rofferr     roff_br(ROFF_ARGS);
1.2       schwarze  159: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  160: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  161: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  162: static enum rofferr     roff_cond(ROFF_ARGS);
                    163: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    164: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  165: static enum rofferr     roff_ds(ROFF_ARGS);
1.175     schwarze  166: static enum rofferr     roff_ec(ROFF_ARGS);
                    167: static enum rofferr     roff_eo(ROFF_ARGS);
1.110     schwarze  168: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.143     schwarze  169: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.106     schwarze  170: static int              roff_evalnum(struct roff *, int,
                    171:                                const char *, int *, int *, int);
                    172: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  173:                                const char *, int *, int *, int);
1.71      schwarze  174: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  175: static void             roff_free1(struct roff *);
1.52      schwarze  176: static void             roff_freereg(struct roffreg *);
1.42      schwarze  177: static void             roff_freestr(struct roffkv *);
1.85      schwarze  178: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  179: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  180: static int              roff_getop(const char *, int *, char *);
1.53      schwarze  181: static int              roff_getregn(const struct roff *,
                    182:                                const char *, size_t);
1.145     schwarze  183: static int              roff_getregro(const struct roff *,
                    184:                                const char *name);
1.80      schwarze  185: static const char      *roff_getstrn(const struct roff *,
1.187     schwarze  186:                                const char *, size_t, int *);
1.143     schwarze  187: static int              roff_hasregn(const struct roff *,
                    188:                                const char *, size_t);
1.123     schwarze  189: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  190: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  191: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137     schwarze  192: static void             roff_man_alloc1(struct roff_man *);
                    193: static void             roff_man_free1(struct roff_man *);
1.172     schwarze  194: static enum rofferr     roff_manyarg(ROFF_ARGS);
1.6       schwarze  195: static enum rofferr     roff_nr(ROFF_ARGS);
1.169     schwarze  196: static enum rofferr     roff_onearg(ROFF_ARGS);
1.166     schwarze  197: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.87      schwarze  198:                                int, int);
1.177     schwarze  199: static enum rofferr     roff_parsetext(struct roff *, struct buf *,
                    200:                                int, int *);
1.178     schwarze  201: static enum rofferr     roff_renamed(ROFF_ARGS);
1.110     schwarze  202: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  203: static enum rofferr     roff_rm(ROFF_ARGS);
1.178     schwarze  204: static enum rofferr     roff_rn(ROFF_ARGS);
1.76      schwarze  205: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  206: static void             roff_setstr(struct roff *,
1.16      schwarze  207:                                const char *, const char *, int);
1.80      schwarze  208: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  209:                                size_t, const char *, size_t, int);
1.14      schwarze  210: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  211: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  212: static enum rofferr     roff_Dd(ROFF_ARGS);
1.27      schwarze  213: static enum rofferr     roff_TE(ROFF_ARGS);
                    214: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  215: static enum rofferr     roff_EQ(ROFF_ARGS);
                    216: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  217: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  218: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  219: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  220:
1.138     schwarze  221: /* --- constant data ------------------------------------------------------ */
                    222:
1.133     schwarze  223: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    224: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    225:
1.166     schwarze  226: const char *__roff_name[MAN_MAX + 1] = {
1.177     schwarze  227:        "br",           "ce",           "ft",           "ll",
1.184     schwarze  228:        "mc",           "po",           "rj",           "sp",
                    229:        "ta",           "ti",           NULL,
1.166     schwarze  230:        "ab",           "ad",           "af",           "aln",
                    231:        "als",          "am",           "am1",          "ami",
                    232:        "ami1",         "as",           "as1",          "asciify",
                    233:        "backtrace",    "bd",           "bleedat",      "blm",
                    234:         "box",         "boxa",         "bp",           "BP",
                    235:        "break",        "breakchar",    "brnl",         "brp",
1.177     schwarze  236:        "brpnl",        "c2",           "cc",
1.166     schwarze  237:        "cf",           "cflags",       "ch",           "char",
                    238:        "chop",         "class",        "close",        "CL",
                    239:        "color",        "composite",    "continue",     "cp",
                    240:        "cropat",       "cs",           "cu",           "da",
                    241:        "dch",          "Dd",           "de",           "de1",
                    242:        "defcolor",     "dei",          "dei1",         "device",
                    243:        "devicem",      "di",           "do",           "ds",
                    244:        "ds1",          "dwh",          "dt",           "ec",
                    245:        "ecr",          "ecs",          "el",           "em",
                    246:        "EN",           "eo",           "EP",           "EQ",
                    247:        "errprint",     "ev",           "evc",          "ex",
                    248:        "fallback",     "fam",          "fc",           "fchar",
                    249:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    250:        "fl",           "flig",         "fp",           "fps",
                    251:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    252:        "fzoom",        "gcolor",       "hc",           "hcode",
                    253:        "hidechar",     "hla",          "hlm",          "hpf",
                    254:        "hpfa",         "hpfcode",      "hw",           "hy",
                    255:        "hylang",       "hylen",        "hym",          "hypp",
                    256:        "hys",          "ie",           "if",           "ig",
                    257:        "index",        "it",           "itc",          "IX",
                    258:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    259:        "lc",           "lc_ctype",     "lds",          "length",
                    260:        "letadj",       "lf",           "lg",           "lhang",
                    261:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.176     schwarze  262:        "ls",           "lsm",          "lt",
1.166     schwarze  263:        "mediasize",    "minss",        "mk",           "mso",
                    264:        "na",           "ne",           "nh",           "nhychar",
                    265:        "nm",           "nn",           "nop",          "nr",
                    266:        "nrf",          "nroff",        "ns",           "nx",
                    267:        "open",         "opena",        "os",           "output",
                    268:        "padj",         "papersize",    "pc",           "pev",
                    269:        "pi",           "PI",           "pl",           "pm",
1.184     schwarze  270:        "pn",           "pnr",          "ps",
1.166     schwarze  271:        "psbb",         "pshape",       "pso",          "ptr",
                    272:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.181     schwarze  273:        "return",       "rfschar",      "rhang",
1.166     schwarze  274:        "rm",           "rn",           "rnn",          "rr",
                    275:        "rs",           "rt",           "schar",        "sentchar",
                    276:        "shc",          "shift",        "sizes",        "so",
                    277:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    278:        "sty",          "substring",    "sv",           "sy",
1.172     schwarze  279:        "T&",           "tc",           "TE",
1.173     schwarze  280:        "TH",           "tkf",          "tl",
1.166     schwarze  281:        "tm",           "tm1",          "tmc",          "tr",
                    282:        "track",        "transchar",    "trf",          "trimat",
                    283:        "trin",         "trnt",         "troff",        "TS",
                    284:        "uf",           "ul",           "unformat",     "unwatch",
                    285:        "unwatchn",     "vpt",          "vs",           "warn",
                    286:        "warnscale",    "watch",        "watchlength",  "watchn",
                    287:        "wh",           "while",        "write",        "writec",
                    288:        "writem",       "xflag",        ".",            NULL,
1.178     schwarze  289:        NULL,           "text",
1.166     schwarze  290:        "Dd",           "Dt",           "Os",           "Sh",
                    291:        "Ss",           "Pp",           "D1",           "Dl",
                    292:        "Bd",           "Ed",           "Bl",           "El",
                    293:        "It",           "Ad",           "An",           "Ap",
                    294:        "Ar",           "Cd",           "Cm",           "Dv",
                    295:        "Er",           "Ev",           "Ex",           "Fa",
                    296:        "Fd",           "Fl",           "Fn",           "Ft",
                    297:        "Ic",           "In",           "Li",           "Nd",
                    298:        "Nm",           "Op",           "Ot",           "Pa",
                    299:        "Rv",           "St",           "Va",           "Vt",
                    300:        "Xr",           "%A",           "%B",           "%D",
                    301:        "%I",           "%J",           "%N",           "%O",
                    302:        "%P",           "%R",           "%T",           "%V",
                    303:        "Ac",           "Ao",           "Aq",           "At",
                    304:        "Bc",           "Bf",           "Bo",           "Bq",
                    305:        "Bsx",          "Bx",           "Db",           "Dc",
                    306:        "Do",           "Dq",           "Ec",           "Ef",
                    307:        "Em",           "Eo",           "Fx",           "Ms",
                    308:        "No",           "Ns",           "Nx",           "Ox",
                    309:        "Pc",           "Pf",           "Po",           "Pq",
                    310:        "Qc",           "Ql",           "Qo",           "Qq",
                    311:        "Re",           "Rs",           "Sc",           "So",
                    312:        "Sq",           "Sm",           "Sx",           "Sy",
                    313:        "Tn",           "Ux",           "Xc",           "Xo",
                    314:        "Fo",           "Fc",           "Oo",           "Oc",
                    315:        "Bk",           "Ek",           "Bt",           "Hf",
                    316:        "Fr",           "Ud",           "Lb",           "Lp",
                    317:        "Lk",           "Mt",           "Brq",          "Bro",
                    318:        "Brc",          "%C",           "Es",           "En",
1.171     schwarze  319:        "Dx",           "%Q",           "%U",           "Ta",
                    320:        NULL,
1.166     schwarze  321:        "TH",           "SH",           "SS",           "TP",
                    322:        "LP",           "PP",           "P",            "IP",
                    323:        "HP",           "SM",           "SB",           "BI",
                    324:        "IB",           "BR",           "RB",           "R",
                    325:        "B",            "I",            "IR",           "RI",
1.171     schwarze  326:        "nf",           "fi",
1.166     schwarze  327:        "RE",           "RS",           "DT",           "UC",
1.169     schwarze  328:        "PD",           "AT",           "in",
1.166     schwarze  329:        "OP",           "EX",           "EE",           "UR",
1.189     bentley   330:        "UE",           "MT",           "ME",           NULL
1.166     schwarze  331: };
                    332: const  char *const *roff_name = __roff_name;
                    333:
                    334: static struct roffmac   roffs[TOKEN_NONE] = {
1.168     schwarze  335:        { roff_br, NULL, NULL, 0 },  /* br */
1.177     schwarze  336:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.169     schwarze  337:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.170     schwarze  338:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.176     schwarze  339:        { roff_onearg, NULL, NULL, 0 },  /* mc */
1.184     schwarze  340:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.181     schwarze  341:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.171     schwarze  342:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.172     schwarze  343:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.173     schwarze  344:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.168     schwarze  345:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.167     schwarze  346:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    347:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    348:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    349:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.183     schwarze  350:        { roff_als, NULL, NULL, 0 },  /* als */
1.167     schwarze  351:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    352:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    353:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    354:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    355:        { roff_ds, NULL, NULL, 0 },  /* as */
                    356:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    357:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    358:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    359:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    360:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    361:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    362:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    363:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    364:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    365:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    366:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    367:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    368:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.168     schwarze  369:        { roff_br, NULL, NULL, 0 },  /* brp */
1.167     schwarze  370:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    371:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    372:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    373:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    374:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    375:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    376:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    377:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    378:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    379:        { roff_insec, NULL, NULL, 0 },  /* close */
                    380:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    381:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    382:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    383:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    384:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    385:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    386:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    387:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    388:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    389:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    390:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    391:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    392:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    393:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    394:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    395:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    396:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    397:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    398:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    400:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    401:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    402:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    403:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.175     schwarze  404:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.167     schwarze  405:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    406:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    407:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    408:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    409:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.175     schwarze  410:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.167     schwarze  411:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    412:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    413:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    414:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    415:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    416:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    417:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    419:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    420:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    421:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    422:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    425:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    429:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    430:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    431:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    432:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    438:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    439:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    440:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    442:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    444:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    445:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    446:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    447:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    448:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    449:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    450:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    451:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    452:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    453:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    454:        { roff_it, NULL, NULL, 0 },  /* it */
                    455:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    456:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    457:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    458:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    461:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    462:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    463:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    464:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    465:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    466:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    467:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    468:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    469:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    470:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    471:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    472:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    474:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    475:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    479:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    484:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    485:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    486:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    487:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    488:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    489:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    491:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    492:        { roff_insec, NULL, NULL, 0 },  /* open */
                    493:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    494:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    495:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    496:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    497:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    500:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    501:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    503:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    504:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    505:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    506:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    507:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    508:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    509:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    511:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    512:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    513:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    514:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    515:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    516:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    517:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    518:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.178     schwarze  519:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.167     schwarze  520:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    521:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    522:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    523:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    524:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    526:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    527:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    529:        { roff_so, NULL, NULL, 0 },  /* so */
                    530:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    533:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    534:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    535:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    536:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    537:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    538:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    539:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    540:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.187     schwarze  541:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.167     schwarze  542:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    543:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    547:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    549:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    550:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    551:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    552:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    553:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    554:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    555:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    557:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    558:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    559:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    560:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    561:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    562:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    563:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    565:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    566:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    568:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    569:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    570:        { roff_insec, NULL, NULL, 0 },  /* write */
                    571:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    572:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    573:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    574:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.178     schwarze  575:        { roff_renamed, NULL, NULL, 0 },
1.167     schwarze  576:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  577: };
                    578:
1.37      schwarze  579: /* Array of injected predefined strings. */
                    580: #define        PREDEFS_MAX      38
                    581: static const struct predef predefs[PREDEFS_MAX] = {
                    582: #include "predefs.in"
                    583: };
                    584:
1.177     schwarze  585: static int      roffce_lines;  /* number of input lines to center */
                    586: static struct roff_node *roffce_node;  /* active request */
1.51      schwarze  587: static int      roffit_lines;  /* number of lines to delay */
                    588: static char    *roffit_macro;  /* nil-terminated macro line */
                    589:
1.80      schwarze  590:
1.138     schwarze  591: /* --- request table ------------------------------------------------------ */
                    592:
1.167     schwarze  593: struct ohash *
                    594: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  595: {
1.167     schwarze  596:        struct ohash    *htab;
                    597:        struct roffreq  *req;
                    598:        enum roff_tok    tok;
                    599:        size_t           sz;
                    600:        unsigned int     slot;
1.3       schwarze  601:
1.167     schwarze  602:        htab = mandoc_malloc(sizeof(*htab));
                    603:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  604:
1.167     schwarze  605:        for (tok = mintok; tok < maxtok; tok++) {
1.168     schwarze  606:                if (roff_name[tok] == NULL)
                    607:                        continue;
1.167     schwarze  608:                sz = strlen(roff_name[tok]);
                    609:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    610:                req->tok = tok;
                    611:                memcpy(req->name, roff_name[tok], sz + 1);
                    612:                slot = ohash_qlookup(htab, req->name);
                    613:                ohash_insert(htab, slot, req);
1.3       schwarze  614:        }
1.167     schwarze  615:        return htab;
1.3       schwarze  616: }
                    617:
1.167     schwarze  618: void
                    619: roffhash_free(struct ohash *htab)
1.1       schwarze  620: {
1.167     schwarze  621:        struct roffreq  *req;
                    622:        unsigned int     slot;
1.1       schwarze  623:
1.167     schwarze  624:        if (htab == NULL)
                    625:                return;
                    626:        for (req = ohash_first(htab, &slot); req != NULL;
                    627:             req = ohash_next(htab, &slot))
                    628:                free(req);
                    629:        ohash_delete(htab);
                    630:        free(htab);
                    631: }
                    632:
                    633: enum roff_tok
                    634: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    635: {
                    636:        struct roffreq  *req;
                    637:        const char      *end;
                    638:
                    639:        if (sz) {
                    640:                end = name + sz;
                    641:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    642:        } else
                    643:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    644:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  645: }
                    646:
1.138     schwarze  647: /* --- stack of request blocks -------------------------------------------- */
                    648:
1.1       schwarze  649: /*
                    650:  * Pop the current node off of the stack of roff instructions currently
                    651:  * pending.
                    652:  */
                    653: static void
                    654: roffnode_pop(struct roff *r)
                    655: {
                    656:        struct roffnode *p;
                    657:
1.2       schwarze  658:        assert(r->last);
1.80      schwarze  659:        p = r->last;
1.2       schwarze  660:
                    661:        r->last = r->last->parent;
1.16      schwarze  662:        free(p->name);
                    663:        free(p->end);
1.1       schwarze  664:        free(p);
                    665: }
                    666:
                    667: /*
                    668:  * Push a roff node onto the instruction stack.  This must later be
                    669:  * removed with roffnode_pop().
                    670:  */
1.11      schwarze  671: static void
1.166     schwarze  672: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  673:                int line, int col)
1.1       schwarze  674: {
                    675:        struct roffnode *p;
                    676:
1.11      schwarze  677:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  678:        p->tok = tok;
1.16      schwarze  679:        if (name)
                    680:                p->name = mandoc_strdup(name);
1.1       schwarze  681:        p->parent = r->last;
                    682:        p->line = line;
                    683:        p->col = col;
1.71      schwarze  684:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  685:
                    686:        r->last = p;
                    687: }
                    688:
1.138     schwarze  689: /* --- roff parser state data management ---------------------------------- */
                    690:
1.1       schwarze  691: static void
                    692: roff_free1(struct roff *r)
                    693: {
1.49      schwarze  694:        struct tbl_node *tbl;
1.42      schwarze  695:        int              i;
1.27      schwarze  696:
1.49      schwarze  697:        while (NULL != (tbl = r->first_tbl)) {
                    698:                r->first_tbl = tbl->next;
                    699:                tbl_free(tbl);
1.27      schwarze  700:        }
                    701:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  702:
1.191     schwarze  703:        if (r->last_eqn != NULL)
                    704:                eqn_free(r->last_eqn);
                    705:        r->last_eqn = r->eqn = NULL;
1.32      schwarze  706:
1.1       schwarze  707:        while (r->last)
                    708:                roffnode_pop(r);
1.27      schwarze  709:
1.96      schwarze  710:        free (r->rstack);
                    711:        r->rstack = NULL;
                    712:        r->rstacksz = 0;
                    713:        r->rstackpos = -1;
                    714:
                    715:        roff_freereg(r->regtab);
                    716:        r->regtab = NULL;
                    717:
1.42      schwarze  718:        roff_freestr(r->strtab);
1.178     schwarze  719:        roff_freestr(r->rentab);
1.42      schwarze  720:        roff_freestr(r->xmbtab);
1.178     schwarze  721:        r->strtab = r->rentab = r->xmbtab = NULL;
1.42      schwarze  722:
                    723:        if (r->xtab)
                    724:                for (i = 0; i < 128; i++)
                    725:                        free(r->xtab[i].p);
                    726:        free(r->xtab);
                    727:        r->xtab = NULL;
1.1       schwarze  728: }
                    729:
                    730: void
                    731: roff_reset(struct roff *r)
                    732: {
                    733:        roff_free1(r);
1.99      schwarze  734:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175     schwarze  735:        r->control = '\0';
                    736:        r->escape = '\\';
1.179     schwarze  737:        roffce_lines = 0;
                    738:        roffce_node = NULL;
                    739:        roffit_lines = 0;
                    740:        roffit_macro = NULL;
1.1       schwarze  741: }
                    742:
                    743: void
                    744: roff_free(struct roff *r)
                    745: {
                    746:        roff_free1(r);
1.167     schwarze  747:        roffhash_free(r->reqtab);
1.1       schwarze  748:        free(r);
                    749: }
                    750:
                    751: struct roff *
1.151     schwarze  752: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  753: {
                    754:        struct roff     *r;
                    755:
1.11      schwarze  756:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  757:        r->parse = parse;
1.167     schwarze  758:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72      schwarze  759:        r->options = options;
1.99      schwarze  760:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  761:        r->rstackpos = -1;
1.175     schwarze  762:        r->escape = '\\';
1.149     schwarze  763:        return r;
1.137     schwarze  764: }
                    765:
1.138     schwarze  766: /* --- syntax tree state data management ---------------------------------- */
                    767:
1.137     schwarze  768: static void
                    769: roff_man_free1(struct roff_man *man)
                    770: {
                    771:
1.138     schwarze  772:        if (man->first != NULL)
                    773:                roff_node_delete(man, man->first);
1.137     schwarze  774:        free(man->meta.msec);
                    775:        free(man->meta.vol);
                    776:        free(man->meta.os);
                    777:        free(man->meta.arch);
                    778:        free(man->meta.title);
                    779:        free(man->meta.name);
                    780:        free(man->meta.date);
                    781: }
                    782:
                    783: static void
                    784: roff_man_alloc1(struct roff_man *man)
                    785: {
                    786:
                    787:        memset(&man->meta, 0, sizeof(man->meta));
                    788:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    789:        man->first->type = ROFFT_ROOT;
                    790:        man->last = man->first;
                    791:        man->last_es = NULL;
                    792:        man->flags = 0;
                    793:        man->macroset = MACROSET_NONE;
                    794:        man->lastsec = man->lastnamed = SEC_NONE;
                    795:        man->next = ROFF_NEXT_CHILD;
                    796: }
                    797:
                    798: void
                    799: roff_man_reset(struct roff_man *man)
                    800: {
                    801:
                    802:        roff_man_free1(man);
                    803:        roff_man_alloc1(man);
                    804: }
                    805:
                    806: void
                    807: roff_man_free(struct roff_man *man)
                    808: {
                    809:
                    810:        roff_man_free1(man);
                    811:        free(man);
                    812: }
                    813:
                    814: struct roff_man *
                    815: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.188     schwarze  816:        const char *os_s, int quick)
1.137     schwarze  817: {
                    818:        struct roff_man *man;
                    819:
                    820:        man = mandoc_calloc(1, sizeof(*man));
                    821:        man->parse = parse;
                    822:        man->roff = roff;
1.188     schwarze  823:        man->os_s = os_s;
1.137     schwarze  824:        man->quick = quick;
                    825:        roff_man_alloc1(man);
1.168     schwarze  826:        roff->man = man;
1.149     schwarze  827:        return man;
1.1       schwarze  828: }
                    829:
1.138     schwarze  830: /* --- syntax tree handling ----------------------------------------------- */
                    831:
                    832: struct roff_node *
                    833: roff_node_alloc(struct roff_man *man, int line, int pos,
                    834:        enum roff_type type, int tok)
                    835: {
                    836:        struct roff_node        *n;
                    837:
                    838:        n = mandoc_calloc(1, sizeof(*n));
                    839:        n->line = line;
                    840:        n->pos = pos;
                    841:        n->tok = tok;
                    842:        n->type = type;
                    843:        n->sec = man->lastsec;
                    844:
                    845:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  846:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  847:        else
1.157     schwarze  848:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  849:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  850:                n->flags |= NODE_LINE;
1.138     schwarze  851:        man->flags &= ~MDOC_NEWLINE;
                    852:
1.149     schwarze  853:        return n;
1.138     schwarze  854: }
                    855:
                    856: void
                    857: roff_node_append(struct roff_man *man, struct roff_node *n)
                    858: {
                    859:
                    860:        switch (man->next) {
                    861:        case ROFF_NEXT_SIBLING:
1.153     schwarze  862:                if (man->last->next != NULL) {
                    863:                        n->next = man->last->next;
                    864:                        man->last->next->prev = n;
                    865:                } else
                    866:                        man->last->parent->last = n;
1.138     schwarze  867:                man->last->next = n;
                    868:                n->prev = man->last;
                    869:                n->parent = man->last->parent;
                    870:                break;
                    871:        case ROFF_NEXT_CHILD:
1.159     schwarze  872:                if (man->last->child != NULL) {
                    873:                        n->next = man->last->child;
                    874:                        man->last->child->prev = n;
                    875:                } else
                    876:                        man->last->last = n;
1.138     schwarze  877:                man->last->child = n;
                    878:                n->parent = man->last;
                    879:                break;
                    880:        default:
                    881:                abort();
                    882:        }
1.154     schwarze  883:        man->last = n;
1.138     schwarze  884:
                    885:        switch (n->type) {
                    886:        case ROFFT_HEAD:
                    887:                n->parent->head = n;
                    888:                break;
                    889:        case ROFFT_BODY:
1.154     schwarze  890:                if (n->end != ENDBODY_NOT)
                    891:                        return;
1.138     schwarze  892:                n->parent->body = n;
                    893:                break;
                    894:        case ROFFT_TAIL:
                    895:                n->parent->tail = n;
                    896:                break;
                    897:        default:
1.154     schwarze  898:                return;
1.138     schwarze  899:        }
1.154     schwarze  900:
                    901:        /*
                    902:         * Copy over the normalised-data pointer of our parent.  Not
                    903:         * everybody has one, but copying a null pointer is fine.
                    904:         */
                    905:
                    906:        n->norm = n->parent->norm;
                    907:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  908: }
                    909:
1.139     schwarze  910: void
                    911: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    912: {
                    913:        struct roff_node        *n;
                    914:
                    915:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    916:        n->string = roff_strdup(man->roff, word);
                    917:        roff_node_append(man, n);
1.158     schwarze  918:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  919:        man->next = ROFF_NEXT_SIBLING;
                    920: }
                    921:
                    922: void
                    923: roff_word_append(struct roff_man *man, const char *word)
                    924: {
                    925:        struct roff_node        *n;
                    926:        char                    *addstr, *newstr;
                    927:
                    928:        n = man->last;
                    929:        addstr = roff_strdup(man->roff, word);
                    930:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    931:        free(addstr);
                    932:        free(n->string);
                    933:        n->string = newstr;
                    934:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  935: }
                    936:
                    937: void
                    938: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    939: {
                    940:        struct roff_node        *n;
                    941:
                    942:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    943:        roff_node_append(man, n);
                    944:        man->next = ROFF_NEXT_CHILD;
                    945: }
                    946:
                    947: struct roff_node *
                    948: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    949: {
                    950:        struct roff_node        *n;
                    951:
                    952:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    953:        roff_node_append(man, n);
                    954:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  955:        return n;
1.139     schwarze  956: }
                    957:
1.138     schwarze  958: struct roff_node *
                    959: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    960: {
                    961:        struct roff_node        *n;
                    962:
                    963:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    964:        roff_node_append(man, n);
                    965:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  966:        return n;
1.138     schwarze  967: }
                    968:
                    969: struct roff_node *
                    970: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    971: {
                    972:        struct roff_node        *n;
                    973:
                    974:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    975:        roff_node_append(man, n);
                    976:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  977:        return n;
1.139     schwarze  978: }
                    979:
                    980: void
                    981: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
                    982: {
                    983:        struct roff_node        *n;
                    984:
                    985:        if (man->macroset == MACROSET_MAN)
1.174     schwarze  986:                man_breakscope(man, ROFF_TS);
1.139     schwarze  987:        n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                    988:        n->span = tbl;
                    989:        roff_node_append(man, n);
1.158     schwarze  990:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  991:        man->next = ROFF_NEXT_SIBLING;
1.138     schwarze  992: }
                    993:
                    994: void
                    995: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                    996: {
                    997:
                    998:        /* Adjust siblings. */
                    999:
                   1000:        if (n->prev)
                   1001:                n->prev->next = n->next;
                   1002:        if (n->next)
                   1003:                n->next->prev = n->prev;
                   1004:
                   1005:        /* Adjust parent. */
                   1006:
                   1007:        if (n->parent != NULL) {
                   1008:                if (n->parent->child == n)
                   1009:                        n->parent->child = n->next;
                   1010:                if (n->parent->last == n)
                   1011:                        n->parent->last = n->prev;
                   1012:        }
                   1013:
                   1014:        /* Adjust parse point. */
                   1015:
                   1016:        if (man == NULL)
                   1017:                return;
                   1018:        if (man->last == n) {
                   1019:                if (n->prev == NULL) {
                   1020:                        man->last = n->parent;
                   1021:                        man->next = ROFF_NEXT_CHILD;
                   1022:                } else {
                   1023:                        man->last = n->prev;
                   1024:                        man->next = ROFF_NEXT_SIBLING;
                   1025:                }
                   1026:        }
                   1027:        if (man->first == n)
                   1028:                man->first = NULL;
                   1029: }
                   1030:
                   1031: void
                   1032: roff_node_free(struct roff_node *n)
                   1033: {
                   1034:
                   1035:        if (n->args != NULL)
                   1036:                mdoc_argv_free(n->args);
                   1037:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1038:                free(n->norm);
1.191     schwarze 1039:        if (n->eqn != NULL)
                   1040:                eqn_box_free(n->eqn);
1.138     schwarze 1041:        free(n->string);
                   1042:        free(n);
                   1043: }
                   1044:
                   1045: void
                   1046: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1047: {
                   1048:
                   1049:        while (n->child != NULL)
                   1050:                roff_node_delete(man, n->child);
                   1051:        roff_node_unlink(man, n);
                   1052:        roff_node_free(n);
1.141     schwarze 1053: }
                   1054:
                   1055: void
                   1056: deroff(char **dest, const struct roff_node *n)
                   1057: {
                   1058:        char    *cp;
                   1059:        size_t   sz;
                   1060:
                   1061:        if (n->type != ROFFT_TEXT) {
                   1062:                for (n = n->child; n != NULL; n = n->next)
                   1063:                        deroff(dest, n);
                   1064:                return;
                   1065:        }
                   1066:
1.160     schwarze 1067:        /* Skip leading whitespace. */
1.141     schwarze 1068:
1.160     schwarze 1069:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1070:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1071:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1072:                        cp++;
1.160     schwarze 1073:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1074:                        break;
                   1075:        }
                   1076:
1.161     schwarze 1077:        /* Skip trailing backslash. */
                   1078:
                   1079:        sz = strlen(cp);
1.162     schwarze 1080:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1081:                sz--;
                   1082:
1.141     schwarze 1083:        /* Skip trailing whitespace. */
                   1084:
1.161     schwarze 1085:        for (; sz; sz--)
1.141     schwarze 1086:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1087:                        break;
                   1088:
                   1089:        /* Skip empty strings. */
                   1090:
                   1091:        if (sz == 0)
                   1092:                return;
                   1093:
                   1094:        if (*dest == NULL) {
                   1095:                *dest = mandoc_strndup(cp, sz);
                   1096:                return;
                   1097:        }
                   1098:
                   1099:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1100:        free(*dest);
                   1101:        *dest = cp;
1.138     schwarze 1102: }
                   1103:
                   1104: /* --- main functions of the roff parser ---------------------------------- */
                   1105:
1.8       schwarze 1106: /*
1.79      schwarze 1107:  * In the current line, expand escape sequences that tend to get
                   1108:  * used in numerical expressions and conditional requests.
                   1109:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1110:  */
1.45      schwarze 1111: static enum rofferr
1.110     schwarze 1112: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1113: {
1.81      schwarze 1114:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze 1115:        const char      *start; /* start of the string to process */
1.82      schwarze 1116:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze 1117:        const char      *stnam; /* start of the name, after "[(*" */
                   1118:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1119:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1120:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1121:        size_t           maxl;  /* expected length of the escape name */
                   1122:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1123:        enum mandoc_esc  esc;   /* type of the escape sequence */
1.188     schwarze 1124:        enum mandoc_os   os_e;  /* kind of RCS id seen */
1.109     schwarze 1125:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1126:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1127:        int              npos;  /* position in numeric expression */
1.91      schwarze 1128:        int              arg_complete; /* argument not interrupted by eol */
1.175     schwarze 1129:        int              done;  /* no more input available */
1.187     schwarze 1130:        int              deftype; /* type of definition to paste */
1.79      schwarze 1131:        char             term;  /* character terminating the escape */
1.8       schwarze 1132:
1.175     schwarze 1133:        /* Search forward for comments. */
                   1134:
                   1135:        done = 0;
                   1136:        start = buf->buf + pos;
                   1137:        for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
                   1138:                if (stesc[0] != r->escape || stesc[1] == '\0')
                   1139:                        continue;
                   1140:                stesc++;
                   1141:                if (*stesc != '"' && *stesc != '#')
                   1142:                        continue;
1.185     schwarze 1143:
                   1144:                /* Comment found, look for RCS id. */
                   1145:
                   1146:                if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.188     schwarze 1147:                        os_e = MANDOC_OS_OPENBSD;
1.185     schwarze 1148:                        cp += 8;
                   1149:                } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.188     schwarze 1150:                        os_e = MANDOC_OS_NETBSD;
1.185     schwarze 1151:                        cp += 7;
                   1152:                }
                   1153:                if (cp != NULL &&
                   1154:                    isalnum((unsigned char)*cp) == 0 &&
1.186     schwarze 1155:                    strchr(cp, '$') != NULL) {
                   1156:                        if (r->man->meta.rcsids & (1 << os_e))
                   1157:                                mandoc_msg(MANDOCERR_RCS_REP, r->parse,
                   1158:                                    ln, stesc + 1 - buf->buf, stesc + 1);
1.185     schwarze 1159:                        r->man->meta.rcsids |= 1 << os_e;
1.186     schwarze 1160:                }
1.185     schwarze 1161:
                   1162:                /* Handle trailing whitespace. */
                   1163:
1.175     schwarze 1164:                cp = strchr(stesc--, '\0') - 1;
                   1165:                if (*cp == '\n') {
                   1166:                        done = 1;
                   1167:                        cp--;
                   1168:                }
                   1169:                if (*cp == ' ' || *cp == '\t')
                   1170:                        mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
                   1171:                            ln, cp - buf->buf, NULL);
                   1172:                while (stesc > start && stesc[-1] == ' ')
                   1173:                        stesc--;
                   1174:                *stesc = '\0';
                   1175:                break;
                   1176:        }
                   1177:        if (stesc == start)
                   1178:                return ROFF_CONT;
                   1179:        stesc--;
                   1180:
                   1181:        /* Notice the end of the input. */
                   1182:
                   1183:        if (*stesc == '\n') {
                   1184:                *stesc-- = '\0';
                   1185:                done = 1;
                   1186:        }
                   1187:
1.43      schwarze 1188:        expand_count = 0;
1.175     schwarze 1189:        while (stesc >= start) {
1.43      schwarze 1190:
1.78      schwarze 1191:                /* Search backwards for the next backslash. */
                   1192:
1.175     schwarze 1193:                if (*stesc != r->escape) {
                   1194:                        if (*stesc == '\\') {
                   1195:                                *stesc = '\0';
                   1196:                                buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
                   1197:                                    buf->buf, stesc + 1) + 1;
                   1198:                                start = nbuf + pos;
                   1199:                                stesc = nbuf + (stesc - buf->buf);
                   1200:                                free(buf->buf);
                   1201:                                buf->buf = nbuf;
                   1202:                        }
                   1203:                        stesc--;
1.78      schwarze 1204:                        continue;
1.175     schwarze 1205:                }
1.78      schwarze 1206:
                   1207:                /* If it is escaped, skip it. */
                   1208:
                   1209:                for (cp = stesc - 1; cp >= start; cp--)
1.175     schwarze 1210:                        if (*cp != r->escape)
1.78      schwarze 1211:                                break;
                   1212:
1.110     schwarze 1213:                if ((stesc - cp) % 2 == 0) {
1.175     schwarze 1214:                        while (stesc > cp)
                   1215:                                *stesc-- = '\\';
1.78      schwarze 1216:                        continue;
1.175     schwarze 1217:                } else if (stesc[1] != '\0') {
                   1218:                        *stesc = '\\';
                   1219:                } else {
                   1220:                        *stesc-- = '\0';
                   1221:                        if (done)
                   1222:                                continue;
                   1223:                        else
                   1224:                                return ROFF_APPEND;
1.78      schwarze 1225:                }
1.23      schwarze 1226:
1.79      schwarze 1227:                /* Decide whether to expand or to check only. */
1.23      schwarze 1228:
1.79      schwarze 1229:                term = '\0';
1.78      schwarze 1230:                cp = stesc + 1;
1.53      schwarze 1231:                switch (*cp) {
1.80      schwarze 1232:                case '*':
1.53      schwarze 1233:                        res = NULL;
                   1234:                        break;
1.80      schwarze 1235:                case 'B':
                   1236:                case 'w':
1.79      schwarze 1237:                        term = cp[1];
                   1238:                        /* FALLTHROUGH */
1.80      schwarze 1239:                case 'n':
1.53      schwarze 1240:                        res = ubuf;
                   1241:                        break;
                   1242:                default:
1.109     schwarze 1243:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1244:                        if (esc == ESCAPE_ERROR ||
                   1245:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1246:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1247:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1248:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1249:                                    "%.*s", (int)(cp - stesc), stesc);
1.175     schwarze 1250:                        stesc--;
1.78      schwarze 1251:                        continue;
1.42      schwarze 1252:                }
                   1253:
1.78      schwarze 1254:                if (EXPAND_LIMIT < ++expand_count) {
                   1255:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1256:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1257:                        return ROFF_IGN;
1.78      schwarze 1258:                }
1.23      schwarze 1259:
                   1260:                /*
                   1261:                 * The third character decides the length
1.53      schwarze 1262:                 * of the name of the string or register.
1.23      schwarze 1263:                 * Save a pointer to the name.
                   1264:                 */
                   1265:
1.110     schwarze 1266:                if (term == '\0') {
1.79      schwarze 1267:                        switch (*++cp) {
1.80      schwarze 1268:                        case '\0':
1.79      schwarze 1269:                                maxl = 0;
                   1270:                                break;
1.80      schwarze 1271:                        case '(':
1.79      schwarze 1272:                                cp++;
                   1273:                                maxl = 2;
                   1274:                                break;
1.80      schwarze 1275:                        case '[':
1.79      schwarze 1276:                                cp++;
                   1277:                                term = ']';
                   1278:                                maxl = 0;
                   1279:                                break;
                   1280:                        default:
                   1281:                                maxl = 1;
                   1282:                                break;
                   1283:                        }
                   1284:                } else {
                   1285:                        cp += 2;
1.8       schwarze 1286:                        maxl = 0;
                   1287:                }
1.23      schwarze 1288:                stnam = cp;
1.8       schwarze 1289:
1.23      schwarze 1290:                /* Advance to the end of the name. */
1.8       schwarze 1291:
1.125     schwarze 1292:                naml = 0;
1.91      schwarze 1293:                arg_complete = 1;
1.125     schwarze 1294:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1295:                        if (*cp == '\0') {
1.92      schwarze 1296:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1297:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1298:                                arg_complete = 0;
1.79      schwarze 1299:                                break;
1.42      schwarze 1300:                        }
1.110     schwarze 1301:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1302:                                cp++;
1.125     schwarze 1303:                                break;
                   1304:                        }
                   1305:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1306:                                naml++;
                   1307:                                continue;
                   1308:                        }
                   1309:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1310:                        case ESCAPE_SPECIAL:
                   1311:                        case ESCAPE_UNICODE:
                   1312:                        case ESCAPE_NUMBERED:
                   1313:                        case ESCAPE_OVERSTRIKE:
                   1314:                                naml++;
                   1315:                                break;
                   1316:                        default:
1.8       schwarze 1317:                                break;
1.79      schwarze 1318:                        }
1.8       schwarze 1319:                }
                   1320:
1.23      schwarze 1321:                /*
                   1322:                 * Retrieve the replacement string; if it is
                   1323:                 * undefined, resume searching for escapes.
                   1324:                 */
                   1325:
1.79      schwarze 1326:                switch (stesc[1]) {
1.80      schwarze 1327:                case '*':
1.187     schwarze 1328:                        if (arg_complete) {
                   1329:                                deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1330:                                res = roff_getstrn(r, stnam, naml, &deftype);
                   1331:                        }
1.79      schwarze 1332:                        break;
1.80      schwarze 1333:                case 'B':
1.79      schwarze 1334:                        npos = 0;
1.91      schwarze 1335:                        ubuf[0] = arg_complete &&
1.133     schwarze 1336:                            roff_evalnum(r, ln, stnam, &npos,
                   1337:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1338:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1339:                        ubuf[1] = '\0';
                   1340:                        break;
1.80      schwarze 1341:                case 'n':
1.91      schwarze 1342:                        if (arg_complete)
                   1343:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1344:                                    roff_getregn(r, stnam, naml));
                   1345:                        else
                   1346:                                ubuf[0] = '\0';
1.79      schwarze 1347:                        break;
1.80      schwarze 1348:                case 'w':
1.91      schwarze 1349:                        /* use even incomplete args */
1.81      schwarze 1350:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1351:                            24 * (int)naml);
                   1352:                        break;
                   1353:                }
1.8       schwarze 1354:
1.110     schwarze 1355:                if (res == NULL) {
1.92      schwarze 1356:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1357:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1358:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1359:                        res = "";
1.118     schwarze 1360:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1361:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1362:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1363:                        return ROFF_IGN;
1.8       schwarze 1364:                }
                   1365:
1.23      schwarze 1366:                /* Replace the escape sequence by the string. */
                   1367:
1.82      schwarze 1368:                *stesc = '\0';
1.110     schwarze 1369:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1370:                    buf->buf, res, cp) + 1;
1.8       schwarze 1371:
1.78      schwarze 1372:                /* Prepare for the next replacement. */
1.8       schwarze 1373:
1.78      schwarze 1374:                start = nbuf + pos;
1.110     schwarze 1375:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1376:                free(buf->buf);
                   1377:                buf->buf = nbuf;
1.42      schwarze 1378:        }
1.149     schwarze 1379:        return ROFF_CONT;
1.42      schwarze 1380: }
                   1381:
                   1382: /*
1.147     schwarze 1383:  * Process text streams.
1.42      schwarze 1384:  */
                   1385: static enum rofferr
1.177     schwarze 1386: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42      schwarze 1387: {
                   1388:        size_t           sz;
                   1389:        const char      *start;
1.51      schwarze 1390:        char            *p;
                   1391:        int              isz;
1.42      schwarze 1392:        enum mandoc_esc  esc;
                   1393:
1.147     schwarze 1394:        /* Spring the input line trap. */
                   1395:
                   1396:        if (roffit_lines == 1) {
                   1397:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1398:                free(buf->buf);
                   1399:                buf->buf = p;
                   1400:                buf->sz = isz + 1;
                   1401:                *offs = 0;
                   1402:                free(roffit_macro);
                   1403:                roffit_lines = 0;
1.149     schwarze 1404:                return ROFF_REPARSE;
1.147     schwarze 1405:        } else if (roffit_lines > 1)
                   1406:                --roffit_lines;
                   1407:
1.177     schwarze 1408:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1409:                if (roffce_lines < 1) {
                   1410:                        r->man->last = roffce_node;
                   1411:                        r->man->next = ROFF_NEXT_SIBLING;
                   1412:                        roffce_lines = 0;
                   1413:                        roffce_node = NULL;
                   1414:                } else
                   1415:                        roffce_lines--;
                   1416:        }
                   1417:
1.147     schwarze 1418:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1419:
1.110     schwarze 1420:        start = p = buf->buf + pos;
1.42      schwarze 1421:
1.110     schwarze 1422:        while (*p != '\0') {
1.42      schwarze 1423:                sz = strcspn(p, "-\\");
                   1424:                p += sz;
                   1425:
1.110     schwarze 1426:                if (*p == '\0')
1.42      schwarze 1427:                        break;
                   1428:
1.110     schwarze 1429:                if (*p == '\\') {
1.42      schwarze 1430:                        /* Skip over escapes. */
                   1431:                        p++;
1.62      schwarze 1432:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1433:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1434:                                break;
1.136     schwarze 1435:                        while (*p == '-')
                   1436:                                p++;
1.42      schwarze 1437:                        continue;
                   1438:                } else if (p == start) {
                   1439:                        p++;
                   1440:                        continue;
                   1441:                }
                   1442:
1.44      schwarze 1443:                if (isalpha((unsigned char)p[-1]) &&
                   1444:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1445:                        *p = ASCII_HYPH;
                   1446:                p++;
1.8       schwarze 1447:        }
1.149     schwarze 1448:        return ROFF_CONT;
1.8       schwarze 1449: }
                   1450:
1.1       schwarze 1451: enum rofferr
1.110     schwarze 1452: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1453: {
1.166     schwarze 1454:        enum roff_tok    t;
1.27      schwarze 1455:        enum rofferr     e;
1.110     schwarze 1456:        int              pos;   /* parse point */
1.115     schwarze 1457:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1458:        int              ppos;  /* original offset in buf->buf */
                   1459:        int              ctl;   /* macro line (boolean) */
                   1460:
                   1461:        ppos = pos = *offs;
1.1       schwarze 1462:
1.102     schwarze 1463:        /* Handle in-line equation delimiters. */
                   1464:
1.108     schwarze 1465:        if (r->tbl == NULL &&
                   1466:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1467:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1468:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1469:                if (e == ROFF_REPARSE)
1.149     schwarze 1470:                        return e;
1.102     schwarze 1471:                assert(e == ROFF_CONT);
                   1472:        }
                   1473:
                   1474:        /* Expand some escape sequences. */
1.8       schwarze 1475:
1.110     schwarze 1476:        e = roff_res(r, buf, ln, pos);
1.175     schwarze 1477:        if (e == ROFF_IGN || e == ROFF_APPEND)
1.149     schwarze 1478:                return e;
1.110     schwarze 1479:        assert(e == ROFF_CONT);
1.8       schwarze 1480:
1.110     schwarze 1481:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1482:
1.8       schwarze 1483:        /*
1.2       schwarze 1484:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1485:         * text through the macro's filter.
                   1486:         * Equations process all content themselves.
                   1487:         * Tables process almost all content themselves, but we want
                   1488:         * to warn about macros before passing it there.
1.2       schwarze 1489:         */
                   1490:
1.124     schwarze 1491:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1492:                t = r->last->tok;
1.110     schwarze 1493:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1494:                if (e == ROFF_IGN)
1.149     schwarze 1495:                        return e;
1.166     schwarze 1496:                assert(e == ROFF_CONT);
1.54      schwarze 1497:        }
1.191     schwarze 1498:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1499:                eqn_read(r->eqn, buf->buf + ppos);
                   1500:                return ROFF_IGN;
                   1501:        }
1.124     schwarze 1502:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.149     schwarze 1503:                return tbl_read(r->tbl, ln, buf->buf, ppos);
1.124     schwarze 1504:        if ( ! ctl)
1.177     schwarze 1505:                return roff_parsetext(r, buf, pos, offs);
1.100     schwarze 1506:
                   1507:        /* Skip empty request lines. */
                   1508:
1.110     schwarze 1509:        if (buf->buf[pos] == '"') {
1.100     schwarze 1510:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1511:                    ln, pos, NULL);
1.149     schwarze 1512:                return ROFF_IGN;
1.110     schwarze 1513:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1514:                return ROFF_IGN;
1.2       schwarze 1515:
                   1516:        /*
                   1517:         * If a scope is open, go to the child handler for that macro,
                   1518:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1519:         * Don't do so if an equation is open.
1.2       schwarze 1520:         */
                   1521:
                   1522:        if (r->last) {
1.1       schwarze 1523:                t = r->last->tok;
1.149     schwarze 1524:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1525:        }
                   1526:
1.115     schwarze 1527:        /* No scope is open.  This is a new request or macro. */
                   1528:
                   1529:        spos = pos;
                   1530:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1531:
                   1532:        /* Tables ignore most macros. */
                   1533:
1.180     schwarze 1534:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181     schwarze 1535:            t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.115     schwarze 1536:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1537:                    ln, pos, buf->buf + spos);
1.180     schwarze 1538:                if (t != TOKEN_NONE)
1.149     schwarze 1539:                        return ROFF_IGN;
1.129     schwarze 1540:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1541:                        pos++;
1.163     schwarze 1542:                while (buf->buf[pos] == ' ')
1.129     schwarze 1543:                        pos++;
1.149     schwarze 1544:                return tbl_read(r->tbl, ln, buf->buf, pos);
1.115     schwarze 1545:        }
                   1546:
1.177     schwarze 1547:        /* For now, let high level macros abort .ce mode. */
                   1548:
                   1549:        if (ctl && roffce_node != NULL &&
                   1550:            (t == TOKEN_NONE || t == ROFF_EQ || t == ROFF_TS)) {
                   1551:                r->man->last = roffce_node;
                   1552:                r->man->next = ROFF_NEXT_SIBLING;
                   1553:                roffce_lines = 0;
                   1554:                roffce_node = NULL;
                   1555:        }
                   1556:
1.2       schwarze 1557:        /*
1.115     schwarze 1558:         * This is neither a roff request nor a user-defined macro.
                   1559:         * Let the standard macro set parsers handle it.
1.2       schwarze 1560:         */
                   1561:
1.166     schwarze 1562:        if (t == TOKEN_NONE)
1.149     schwarze 1563:                return ROFF_CONT;
1.115     schwarze 1564:
                   1565:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1566:
1.168     schwarze 1567:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2       schwarze 1568: }
                   1569:
1.27      schwarze 1570: void
1.2       schwarze 1571: roff_endparse(struct roff *r)
                   1572: {
1.1       schwarze 1573:
1.27      schwarze 1574:        if (r->last)
1.94      schwarze 1575:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1576:                    r->last->line, r->last->col,
1.167     schwarze 1577:                    roff_name[r->last->tok]);
1.27      schwarze 1578:
1.32      schwarze 1579:        if (r->eqn) {
1.94      schwarze 1580:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.191     schwarze 1581:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1582:                eqn_parse(r->eqn);
                   1583:                r->eqn = NULL;
1.32      schwarze 1584:        }
                   1585:
1.27      schwarze 1586:        if (r->tbl) {
1.94      schwarze 1587:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1588:                    r->tbl->line, r->tbl->pos, "TS");
1.41      schwarze 1589:                tbl_end(&r->tbl);
1.27      schwarze 1590:        }
1.1       schwarze 1591: }
                   1592:
                   1593: /*
                   1594:  * Parse a roff node's type from the input buffer.  This must be in the
                   1595:  * form of ".foo xxx" in the usual way.
                   1596:  */
1.166     schwarze 1597: static enum roff_tok
1.87      schwarze 1598: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1599: {
1.87      schwarze 1600:        char            *cp;
1.16      schwarze 1601:        const char      *mac;
                   1602:        size_t           maclen;
1.187     schwarze 1603:        int              deftype;
1.166     schwarze 1604:        enum roff_tok    t;
1.1       schwarze 1605:
1.87      schwarze 1606:        cp = buf + *pos;
                   1607:
                   1608:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1609:                return TOKEN_NONE;
1.1       schwarze 1610:
1.87      schwarze 1611:        mac = cp;
                   1612:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1613:
1.187     schwarze 1614:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1615:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1616:        switch (deftype) {
                   1617:        case ROFFDEF_USER:
                   1618:                t = ROFF_USERDEF;
                   1619:                break;
                   1620:        case ROFFDEF_REN:
                   1621:                t = ROFF_RENAMED;
                   1622:                break;
                   1623:        default:
                   1624:                t = roffhash_find(r->reqtab, mac, maclen);
                   1625:                break;
                   1626:        }
1.166     schwarze 1627:        if (t != TOKEN_NONE)
1.87      schwarze 1628:                *pos = cp - buf;
1.149     schwarze 1629:        return t;
1.1       schwarze 1630: }
                   1631:
1.138     schwarze 1632: /* --- handling of request blocks ----------------------------------------- */
                   1633:
1.1       schwarze 1634: static enum rofferr
1.2       schwarze 1635: roff_cblock(ROFF_ARGS)
1.1       schwarze 1636: {
                   1637:
1.2       schwarze 1638:        /*
                   1639:         * A block-close `..' should only be invoked as a child of an
                   1640:         * ignore macro, otherwise raise a warning and just ignore it.
                   1641:         */
                   1642:
1.110     schwarze 1643:        if (r->last == NULL) {
1.94      schwarze 1644:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1645:                    ln, ppos, "..");
1.149     schwarze 1646:                return ROFF_IGN;
1.2       schwarze 1647:        }
1.1       schwarze 1648:
1.2       schwarze 1649:        switch (r->last->tok) {
1.80      schwarze 1650:        case ROFF_am:
1.93      schwarze 1651:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1652:        case ROFF_ami:
                   1653:        case ROFF_de:
1.23      schwarze 1654:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1655:        case ROFF_dei:
                   1656:        case ROFF_ig:
1.2       schwarze 1657:                break;
                   1658:        default:
1.94      schwarze 1659:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1660:                    ln, ppos, "..");
1.149     schwarze 1661:                return ROFF_IGN;
1.2       schwarze 1662:        }
                   1663:
1.110     schwarze 1664:        if (buf->buf[pos] != '\0')
1.90      schwarze 1665:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1666:                    ".. %s", buf->buf + pos);
1.2       schwarze 1667:
                   1668:        roffnode_pop(r);
                   1669:        roffnode_cleanscope(r);
1.149     schwarze 1670:        return ROFF_IGN;
1.2       schwarze 1671:
                   1672: }
1.1       schwarze 1673:
1.2       schwarze 1674: static void
                   1675: roffnode_cleanscope(struct roff *r)
                   1676: {
1.1       schwarze 1677:
1.2       schwarze 1678:        while (r->last) {
1.46      schwarze 1679:                if (--r->last->endspan != 0)
1.2       schwarze 1680:                        break;
                   1681:                roffnode_pop(r);
                   1682:        }
                   1683: }
1.1       schwarze 1684:
1.68      schwarze 1685: static void
                   1686: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1687: {
1.1       schwarze 1688:
1.2       schwarze 1689:        if (NULL == r->last) {
1.94      schwarze 1690:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1691:                    ln, ppos, "\\}");
1.68      schwarze 1692:                return;
1.2       schwarze 1693:        }
1.1       schwarze 1694:
1.2       schwarze 1695:        switch (r->last->tok) {
1.80      schwarze 1696:        case ROFF_el:
                   1697:        case ROFF_ie:
                   1698:        case ROFF_if:
1.2       schwarze 1699:                break;
                   1700:        default:
1.94      schwarze 1701:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1702:                    ln, ppos, "\\}");
1.68      schwarze 1703:                return;
1.2       schwarze 1704:        }
1.1       schwarze 1705:
1.2       schwarze 1706:        if (r->last->endspan > -1) {
1.94      schwarze 1707:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1708:                    ln, ppos, "\\}");
1.68      schwarze 1709:                return;
1.2       schwarze 1710:        }
                   1711:
                   1712:        roffnode_pop(r);
                   1713:        roffnode_cleanscope(r);
1.68      schwarze 1714:        return;
1.1       schwarze 1715: }
                   1716:
                   1717: static enum rofferr
1.2       schwarze 1718: roff_block(ROFF_ARGS)
1.1       schwarze 1719: {
1.187     schwarze 1720:        const char      *name, *value;
                   1721:        char            *call, *cp, *iname, *rname;
                   1722:        size_t           csz, namesz, rsz;
                   1723:        int              deftype;
1.16      schwarze 1724:
1.93      schwarze 1725:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1726:
1.110     schwarze 1727:        if (tok == ROFF_de1)
1.93      schwarze 1728:                tok = ROFF_de;
1.123     schwarze 1729:        else if (tok == ROFF_dei1)
                   1730:                tok = ROFF_dei;
1.110     schwarze 1731:        else if (tok == ROFF_am1)
1.93      schwarze 1732:                tok = ROFF_am;
1.123     schwarze 1733:        else if (tok == ROFF_ami1)
                   1734:                tok = ROFF_ami;
1.93      schwarze 1735:
                   1736:        /* Parse the macro name argument. */
                   1737:
1.110     schwarze 1738:        cp = buf->buf + pos;
                   1739:        if (tok == ROFF_ig) {
1.93      schwarze 1740:                iname = NULL;
                   1741:                namesz = 0;
                   1742:        } else {
                   1743:                iname = cp;
                   1744:                namesz = roff_getname(r, &cp, ln, ppos);
                   1745:                iname[namesz] = '\0';
                   1746:        }
1.22      schwarze 1747:
1.93      schwarze 1748:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1749:
1.110     schwarze 1750:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1751:                deftype = ROFFDEF_USER;
                   1752:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1753:                if (name == NULL) {
1.93      schwarze 1754:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1755:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1756:                            "%.*s", (int)namesz, iname);
                   1757:                        namesz = 0;
                   1758:                } else
                   1759:                        namesz = strlen(name);
                   1760:        } else
                   1761:                name = iname;
1.22      schwarze 1762:
1.110     schwarze 1763:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1764:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167     schwarze 1765:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1766:                return ROFF_IGN;
1.93      schwarze 1767:        }
1.2       schwarze 1768:
1.16      schwarze 1769:        roffnode_push(r, tok, name, ln, ppos);
                   1770:
                   1771:        /*
                   1772:         * At the beginning of a `de' macro, clear the existing string
                   1773:         * with the same name, if there is one.  New content will be
1.66      schwarze 1774:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1775:         */
1.22      schwarze 1776:
1.183     schwarze 1777:        if (tok == ROFF_de || tok == ROFF_dei) {
1.86      schwarze 1778:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183     schwarze 1779:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.187     schwarze 1780:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   1781:                deftype = ROFFDEF_ANY;
                   1782:                value = roff_getstrn(r, iname, namesz, &deftype);
                   1783:                switch (deftype) {  /* Before appending, ... */
                   1784:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   1785:                        roff_setstrn(&r->strtab, name, namesz,
                   1786:                            value, strlen(value), 0);
                   1787:                        break;
                   1788:                case ROFFDEF_REN: /* call original standard macro. */
                   1789:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1790:                            (int)strlen(value), value);
                   1791:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1792:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   1793:                        free(call);
                   1794:                        break;
                   1795:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   1796:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   1797:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   1798:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   1799:                            (int)rsz, rname);
                   1800:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   1801:                        free(call);
                   1802:                        free(rname);
                   1803:                        break;
                   1804:                default:
                   1805:                        break;
                   1806:                }
1.183     schwarze 1807:        }
1.2       schwarze 1808:
1.110     schwarze 1809:        if (*cp == '\0')
1.149     schwarze 1810:                return ROFF_IGN;
1.1       schwarze 1811:
1.93      schwarze 1812:        /* Get the custom end marker. */
1.22      schwarze 1813:
1.93      schwarze 1814:        iname = cp;
1.86      schwarze 1815:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1816:
                   1817:        /* Resolve the end marker if it is indirect. */
                   1818:
1.110     schwarze 1819:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187     schwarze 1820:                deftype = ROFFDEF_USER;
                   1821:                name = roff_getstrn(r, iname, namesz, &deftype);
                   1822:                if (name == NULL) {
1.93      schwarze 1823:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1824:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1825:                            "%.*s", (int)namesz, iname);
                   1826:                        namesz = 0;
                   1827:                } else
                   1828:                        namesz = strlen(name);
                   1829:        } else
                   1830:                name = iname;
                   1831:
1.86      schwarze 1832:        if (namesz)
                   1833:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1834:
1.110     schwarze 1835:        if (*cp != '\0')
1.90      schwarze 1836:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167     schwarze 1837:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1838:
1.149     schwarze 1839:        return ROFF_IGN;
1.1       schwarze 1840: }
                   1841:
                   1842: static enum rofferr
1.2       schwarze 1843: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1844: {
1.166     schwarze 1845:        enum roff_tok   t;
1.2       schwarze 1846:        int             i, j;
                   1847:
                   1848:        /*
                   1849:         * First check whether a custom macro exists at this level.  If
                   1850:         * it does, then check against it.  This is some of groff's
                   1851:         * stranger behaviours.  If we encountered a custom end-scope
                   1852:         * tag and that tag also happens to be a "real" macro, then we
                   1853:         * need to try interpreting it again as a real macro.  If it's
                   1854:         * not, then return ignore.  Else continue.
                   1855:         */
                   1856:
                   1857:        if (r->last->end) {
1.35      schwarze 1858:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1859:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1860:                                break;
1.1       schwarze 1861:
1.110     schwarze 1862:                if (r->last->end[j] == '\0' &&
                   1863:                    (buf->buf[i] == '\0' ||
                   1864:                     buf->buf[i] == ' ' ||
                   1865:                     buf->buf[i] == '\t')) {
1.2       schwarze 1866:                        roffnode_pop(r);
                   1867:                        roffnode_cleanscope(r);
1.1       schwarze 1868:
1.110     schwarze 1869:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1870:                                i++;
                   1871:
                   1872:                        pos = i;
1.110     schwarze 1873:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1874:                            TOKEN_NONE)
1.149     schwarze 1875:                                return ROFF_RERUN;
                   1876:                        return ROFF_IGN;
1.2       schwarze 1877:                }
1.1       schwarze 1878:        }
                   1879:
1.2       schwarze 1880:        /*
                   1881:         * If we have no custom end-query or lookup failed, then try
                   1882:         * pulling it out of the hashtable.
                   1883:         */
1.1       schwarze 1884:
1.110     schwarze 1885:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1886:
1.110     schwarze 1887:        if (t != ROFF_cblock) {
                   1888:                if (tok != ROFF_ig)
                   1889:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1890:                return ROFF_IGN;
1.16      schwarze 1891:        }
1.1       schwarze 1892:
1.149     schwarze 1893:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1894: }
                   1895:
                   1896: static enum rofferr
                   1897: roff_block_text(ROFF_ARGS)
                   1898: {
                   1899:
1.110     schwarze 1900:        if (tok != ROFF_ig)
                   1901:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1902:
1.149     schwarze 1903:        return ROFF_IGN;
1.2       schwarze 1904: }
                   1905:
                   1906: static enum rofferr
                   1907: roff_cond_sub(ROFF_ARGS)
                   1908: {
1.166     schwarze 1909:        enum roff_tok    t;
1.37      schwarze 1910:        char            *ep;
1.71      schwarze 1911:        int              rr;
1.2       schwarze 1912:
                   1913:        rr = r->last->rule;
1.37      schwarze 1914:        roffnode_cleanscope(r);
1.39      schwarze 1915:
1.69      schwarze 1916:        /*
                   1917:         * If `\}' occurs on a macro line without a preceding macro,
                   1918:         * drop the line completely.
                   1919:         */
                   1920:
1.110     schwarze 1921:        ep = buf->buf + pos;
                   1922:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1923:                rr = 0;
1.69      schwarze 1924:
1.50      schwarze 1925:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1926:
1.110     schwarze 1927:        while ((ep = strchr(ep, '\\')) != NULL) {
1.190     schwarze 1928:                switch (ep[1]) {
                   1929:                case '}':
                   1930:                        memmove(ep, ep + 2, strlen(ep + 2) + 1);
                   1931:                        roff_ccond(r, ln, ep - buf->buf);
                   1932:                        break;
                   1933:                case '\0':
                   1934:                        ++ep;
                   1935:                        break;
                   1936:                default:
                   1937:                        ep += 2;
                   1938:                        break;
1.70      schwarze 1939:                }
1.50      schwarze 1940:        }
1.190     schwarze 1941:
                   1942:        /*
                   1943:         * Fully handle known macros when they are structurally
                   1944:         * required or when the conditional evaluated to true.
                   1945:         */
                   1946:
                   1947:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1948:        return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
                   1949:            ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
                   1950:            ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1951: }
                   1952:
                   1953: static enum rofferr
                   1954: roff_cond_text(ROFF_ARGS)
                   1955: {
1.37      schwarze 1956:        char            *ep;
1.71      schwarze 1957:        int              rr;
1.2       schwarze 1958:
                   1959:        rr = r->last->rule;
1.37      schwarze 1960:        roffnode_cleanscope(r);
1.1       schwarze 1961:
1.110     schwarze 1962:        ep = buf->buf + pos;
                   1963:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1964:                if (*(++ep) == '}') {
1.70      schwarze 1965:                        *ep = '&';
1.110     schwarze 1966:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1967:                }
1.119     schwarze 1968:                if (*ep != '\0')
                   1969:                        ++ep;
1.2       schwarze 1970:        }
1.149     schwarze 1971:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1972: }
                   1973:
1.138     schwarze 1974: /* --- handling of numeric and conditional expressions -------------------- */
                   1975:
1.77      schwarze 1976: /*
                   1977:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1978:  * If there is at least one digit, return success and advance the
                   1979:  * parse point, else return failure and let the parse point unchanged.
                   1980:  * Ignore overflows, treat them just like the C language.
                   1981:  */
1.56      schwarze 1982: static int
1.133     schwarze 1983: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 1984: {
1.133     schwarze 1985:        int      myres, scaled, n, p;
1.79      schwarze 1986:
                   1987:        if (NULL == res)
                   1988:                res = &myres;
1.56      schwarze 1989:
                   1990:        p = *pos;
                   1991:        n = v[p] == '-';
1.133     schwarze 1992:        if (n || v[p] == '+')
1.56      schwarze 1993:                p++;
                   1994:
1.133     schwarze 1995:        if (flags & ROFFNUM_WHITE)
                   1996:                while (isspace((unsigned char)v[p]))
                   1997:                        p++;
                   1998:
1.56      schwarze 1999:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 2000:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 2001:        if (p == *pos + n)
                   2002:                return 0;
                   2003:
                   2004:        if (n)
                   2005:                *res = -*res;
                   2006:
1.126     schwarze 2007:        /* Each number may be followed by one optional scaling unit. */
                   2008:
                   2009:        switch (v[p]) {
                   2010:        case 'f':
1.133     schwarze 2011:                scaled = *res * 65536;
1.126     schwarze 2012:                break;
                   2013:        case 'i':
1.133     schwarze 2014:                scaled = *res * 240;
1.126     schwarze 2015:                break;
                   2016:        case 'c':
1.133     schwarze 2017:                scaled = *res * 240 / 2.54;
1.126     schwarze 2018:                break;
                   2019:        case 'v':
                   2020:        case 'P':
1.133     schwarze 2021:                scaled = *res * 40;
1.126     schwarze 2022:                break;
                   2023:        case 'm':
                   2024:        case 'n':
1.133     schwarze 2025:                scaled = *res * 24;
1.126     schwarze 2026:                break;
                   2027:        case 'p':
1.133     schwarze 2028:                scaled = *res * 10 / 3;
1.126     schwarze 2029:                break;
                   2030:        case 'u':
1.133     schwarze 2031:                scaled = *res;
1.126     schwarze 2032:                break;
                   2033:        case 'M':
1.133     schwarze 2034:                scaled = *res * 6 / 25;
1.126     schwarze 2035:                break;
                   2036:        default:
1.133     schwarze 2037:                scaled = *res;
1.126     schwarze 2038:                p--;
                   2039:                break;
                   2040:        }
1.133     schwarze 2041:        if (flags & ROFFNUM_SCALE)
                   2042:                *res = scaled;
1.126     schwarze 2043:
                   2044:        *pos = p + 1;
1.149     schwarze 2045:        return 1;
1.56      schwarze 2046: }
                   2047:
1.71      schwarze 2048: /*
                   2049:  * Evaluate a string comparison condition.
                   2050:  * The first character is the delimiter.
                   2051:  * Succeed if the string up to its second occurrence
                   2052:  * matches the string up to its third occurence.
                   2053:  * Advance the cursor after the third occurrence
                   2054:  * or lacking that, to the end of the line.
                   2055:  */
                   2056: static int
                   2057: roff_evalstrcond(const char *v, int *pos)
                   2058: {
                   2059:        const char      *s1, *s2, *s3;
                   2060:        int              match;
                   2061:
                   2062:        match = 0;
                   2063:        s1 = v + *pos;          /* initial delimiter */
                   2064:        s2 = s1 + 1;            /* for scanning the first string */
                   2065:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2066:
                   2067:        if (NULL == s3)         /* found no middle delimiter */
                   2068:                goto out;
                   2069:
                   2070:        while ('\0' != *++s3) {
                   2071:                if (*s2 != *s3) {  /* mismatch */
                   2072:                        s3 = strchr(s3, *s1);
                   2073:                        break;
                   2074:                }
                   2075:                if (*s3 == *s1) {  /* found the final delimiter */
                   2076:                        match = 1;
                   2077:                        break;
                   2078:                }
                   2079:                s2++;
                   2080:        }
                   2081:
                   2082: out:
                   2083:        if (NULL == s3)
                   2084:                s3 = strchr(s2, '\0');
1.114     schwarze 2085:        else if (*s3 != '\0')
1.71      schwarze 2086:                s3++;
                   2087:        *pos = s3 - v;
1.149     schwarze 2088:        return match;
1.71      schwarze 2089: }
                   2090:
1.77      schwarze 2091: /*
                   2092:  * Evaluate an optionally negated single character, numerical,
                   2093:  * or string condition.
                   2094:  */
1.71      schwarze 2095: static int
1.143     schwarze 2096: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 2097: {
1.143     schwarze 2098:        char    *cp, *name;
                   2099:        size_t   sz;
1.187     schwarze 2100:        int      deftype, number, savepos, istrue, wanttrue;
1.5       schwarze 2101:
1.71      schwarze 2102:        if ('!' == v[*pos]) {
                   2103:                wanttrue = 0;
                   2104:                (*pos)++;
                   2105:        } else
                   2106:                wanttrue = 1;
                   2107:
1.5       schwarze 2108:        switch (v[*pos]) {
1.112     schwarze 2109:        case '\0':
1.149     schwarze 2110:                return 0;
1.80      schwarze 2111:        case 'n':
                   2112:        case 'o':
1.5       schwarze 2113:                (*pos)++;
1.149     schwarze 2114:                return wanttrue;
1.80      schwarze 2115:        case 'c':
                   2116:        case 'e':
                   2117:        case 't':
1.111     schwarze 2118:        case 'v':
1.5       schwarze 2119:                (*pos)++;
1.149     schwarze 2120:                return !wanttrue;
1.182     schwarze 2121:        case 'd':
1.143     schwarze 2122:        case 'r':
1.182     schwarze 2123:                cp = v + *pos + 1;
                   2124:                while (*cp == ' ')
                   2125:                        cp++;
                   2126:                name = cp;
                   2127:                sz = roff_getname(r, &cp, ln, cp - v);
1.187     schwarze 2128:                if (sz == 0)
                   2129:                        istrue = 0;
                   2130:                else if (v[*pos] == 'r')
                   2131:                        istrue = roff_hasregn(r, name, sz);
                   2132:                else {
                   2133:                        deftype = ROFFDEF_ANY;
                   2134:                        roff_getstrn(r, name, sz, &deftype);
                   2135:                        istrue = !!deftype;
                   2136:                }
1.143     schwarze 2137:                *pos = cp - v;
1.182     schwarze 2138:                return istrue == wanttrue;
1.5       schwarze 2139:        default:
                   2140:                break;
                   2141:        }
                   2142:
1.113     schwarze 2143:        savepos = *pos;
1.133     schwarze 2144:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 2145:                return (number > 0) == wanttrue;
1.113     schwarze 2146:        else if (*pos == savepos)
1.149     schwarze 2147:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 2148:        else
1.149     schwarze 2149:                return 0;
1.5       schwarze 2150: }
                   2151:
1.2       schwarze 2152: static enum rofferr
1.21      schwarze 2153: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 2154: {
1.30      schwarze 2155:
1.149     schwarze 2156:        return ROFF_IGN;
1.21      schwarze 2157: }
                   2158:
                   2159: static enum rofferr
1.123     schwarze 2160: roff_insec(ROFF_ARGS)
                   2161: {
                   2162:
                   2163:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167     schwarze 2164:            ln, ppos, roff_name[tok]);
1.149     schwarze 2165:        return ROFF_IGN;
1.123     schwarze 2166: }
                   2167:
                   2168: static enum rofferr
                   2169: roff_unsupp(ROFF_ARGS)
                   2170: {
                   2171:
                   2172:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167     schwarze 2173:            ln, ppos, roff_name[tok]);
1.149     schwarze 2174:        return ROFF_IGN;
1.123     schwarze 2175: }
                   2176:
                   2177: static enum rofferr
1.2       schwarze 2178: roff_cond(ROFF_ARGS)
                   2179: {
1.46      schwarze 2180:
                   2181:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2182:
1.80      schwarze 2183:        /*
1.35      schwarze 2184:         * An `.el' has no conditional body: it will consume the value
                   2185:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2186:         * defaults to DENY.
1.35      schwarze 2187:         *
                   2188:         * If we're not an `el', however, then evaluate the conditional.
                   2189:         */
1.1       schwarze 2190:
1.110     schwarze 2191:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2192:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2193:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2194:
1.35      schwarze 2195:        /*
                   2196:         * An if-else will put the NEGATION of the current evaluated
                   2197:         * conditional into the stack of rules.
                   2198:         */
                   2199:
1.110     schwarze 2200:        if (tok == ROFF_ie) {
1.96      schwarze 2201:                if (r->rstackpos + 1 == r->rstacksz) {
                   2202:                        r->rstacksz += 16;
                   2203:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2204:                            r->rstacksz, sizeof(int));
1.35      schwarze 2205:                }
1.71      schwarze 2206:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2207:        }
1.5       schwarze 2208:
                   2209:        /* If the parent has false as its rule, then so do we. */
                   2210:
1.71      schwarze 2211:        if (r->last->parent && !r->last->parent->rule)
                   2212:                r->last->rule = 0;
1.5       schwarze 2213:
                   2214:        /*
1.46      schwarze 2215:         * Determine scope.
                   2216:         * If there is nothing on the line after the conditional,
                   2217:         * not even whitespace, use next-line scope.
1.5       schwarze 2218:         */
1.2       schwarze 2219:
1.110     schwarze 2220:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2221:                r->last->endspan = 2;
                   2222:                goto out;
                   2223:        }
                   2224:
1.110     schwarze 2225:        while (buf->buf[pos] == ' ')
1.46      schwarze 2226:                pos++;
                   2227:
                   2228:        /* An opening brace requests multiline scope. */
1.2       schwarze 2229:
1.110     schwarze 2230:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2231:                r->last->endspan = -1;
                   2232:                pos += 2;
1.144     schwarze 2233:                while (buf->buf[pos] == ' ')
                   2234:                        pos++;
1.46      schwarze 2235:                goto out;
1.80      schwarze 2236:        }
1.2       schwarze 2237:
                   2238:        /*
1.46      schwarze 2239:         * Anything else following the conditional causes
                   2240:         * single-line scope.  Warn if the scope contains
                   2241:         * nothing but trailing whitespace.
1.2       schwarze 2242:         */
                   2243:
1.110     schwarze 2244:        if (buf->buf[pos] == '\0')
1.89      schwarze 2245:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167     schwarze 2246:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2247:
1.46      schwarze 2248:        r->last->endspan = 1;
1.1       schwarze 2249:
1.46      schwarze 2250: out:
1.2       schwarze 2251:        *offs = pos;
1.149     schwarze 2252:        return ROFF_RERUN;
1.1       schwarze 2253: }
                   2254:
1.2       schwarze 2255: static enum rofferr
1.7       schwarze 2256: roff_ds(ROFF_ARGS)
                   2257: {
1.85      schwarze 2258:        char            *string;
                   2259:        const char      *name;
                   2260:        size_t           namesz;
1.10      schwarze 2261:
1.123     schwarze 2262:        /* Ignore groff compatibility mode for now. */
                   2263:
                   2264:        if (tok == ROFF_ds1)
                   2265:                tok = ROFF_ds;
                   2266:        else if (tok == ROFF_as1)
                   2267:                tok = ROFF_as;
                   2268:
1.10      schwarze 2269:        /*
1.85      schwarze 2270:         * The first word is the name of the string.
                   2271:         * If it is empty or terminated by an escape sequence,
                   2272:         * abort the `ds' request without defining anything.
1.10      schwarze 2273:         */
1.7       schwarze 2274:
1.110     schwarze 2275:        name = string = buf->buf + pos;
                   2276:        if (*name == '\0')
1.149     schwarze 2277:                return ROFF_IGN;
1.7       schwarze 2278:
1.85      schwarze 2279:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2280:        if (name[namesz] == '\\')
1.149     schwarze 2281:                return ROFF_IGN;
1.85      schwarze 2282:
                   2283:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2284:        if (*string == '"')
1.7       schwarze 2285:                string++;
                   2286:
1.10      schwarze 2287:        /* The rest is the value. */
1.85      schwarze 2288:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2289:            ROFF_as == tok);
1.183     schwarze 2290:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149     schwarze 2291:        return ROFF_IGN;
1.7       schwarze 2292: }
                   2293:
1.77      schwarze 2294: /*
                   2295:  * Parse a single operator, one or two characters long.
                   2296:  * If the operator is recognized, return success and advance the
                   2297:  * parse point, else return failure and let the parse point unchanged.
                   2298:  */
                   2299: static int
                   2300: roff_getop(const char *v, int *pos, char *res)
                   2301: {
                   2302:
                   2303:        *res = v[*pos];
                   2304:
                   2305:        switch (*res) {
1.80      schwarze 2306:        case '+':
                   2307:        case '-':
                   2308:        case '*':
                   2309:        case '/':
                   2310:        case '%':
                   2311:        case '&':
                   2312:        case ':':
1.77      schwarze 2313:                break;
                   2314:        case '<':
                   2315:                switch (v[*pos + 1]) {
1.80      schwarze 2316:                case '=':
1.77      schwarze 2317:                        *res = 'l';
                   2318:                        (*pos)++;
                   2319:                        break;
1.80      schwarze 2320:                case '>':
1.77      schwarze 2321:                        *res = '!';
                   2322:                        (*pos)++;
                   2323:                        break;
1.80      schwarze 2324:                case '?':
1.77      schwarze 2325:                        *res = 'i';
                   2326:                        (*pos)++;
                   2327:                        break;
                   2328:                default:
                   2329:                        break;
                   2330:                }
                   2331:                break;
                   2332:        case '>':
                   2333:                switch (v[*pos + 1]) {
1.80      schwarze 2334:                case '=':
1.77      schwarze 2335:                        *res = 'g';
                   2336:                        (*pos)++;
                   2337:                        break;
1.80      schwarze 2338:                case '?':
1.77      schwarze 2339:                        *res = 'a';
                   2340:                        (*pos)++;
                   2341:                        break;
                   2342:                default:
                   2343:                        break;
                   2344:                }
                   2345:                break;
                   2346:        case '=':
                   2347:                if ('=' == v[*pos + 1])
                   2348:                        (*pos)++;
                   2349:                break;
                   2350:        default:
1.149     schwarze 2351:                return 0;
1.77      schwarze 2352:        }
                   2353:        (*pos)++;
                   2354:
1.149     schwarze 2355:        return *res;
1.77      schwarze 2356: }
                   2357:
                   2358: /*
                   2359:  * Evaluate either a parenthesized numeric expression
                   2360:  * or a single signed integer number.
                   2361:  */
                   2362: static int
1.106     schwarze 2363: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2364:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2365: {
                   2366:
                   2367:        if ('(' != v[*pos])
1.149     schwarze 2368:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2369:
                   2370:        (*pos)++;
1.133     schwarze 2371:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2372:                return 0;
1.77      schwarze 2373:
1.79      schwarze 2374:        /*
                   2375:         * Omission of the closing parenthesis
                   2376:         * is an error in validation mode,
                   2377:         * but ignored in evaluation mode.
                   2378:         */
                   2379:
1.77      schwarze 2380:        if (')' == v[*pos])
                   2381:                (*pos)++;
1.79      schwarze 2382:        else if (NULL == res)
1.149     schwarze 2383:                return 0;
1.77      schwarze 2384:
1.149     schwarze 2385:        return 1;
1.77      schwarze 2386: }
                   2387:
                   2388: /*
                   2389:  * Evaluate a complete numeric expression.
                   2390:  * Proceed left to right, there is no concept of precedence.
                   2391:  */
                   2392: static int
1.106     schwarze 2393: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2394:        int *pos, int *res, int flags)
1.77      schwarze 2395: {
                   2396:        int              mypos, operand2;
                   2397:        char             operator;
                   2398:
                   2399:        if (NULL == pos) {
                   2400:                mypos = 0;
                   2401:                pos = &mypos;
                   2402:        }
                   2403:
1.133     schwarze 2404:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2405:                while (isspace((unsigned char)v[*pos]))
                   2406:                        (*pos)++;
                   2407:
1.133     schwarze 2408:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2409:                return 0;
1.77      schwarze 2410:
                   2411:        while (1) {
1.133     schwarze 2412:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2413:                        while (isspace((unsigned char)v[*pos]))
                   2414:                                (*pos)++;
                   2415:
                   2416:                if ( ! roff_getop(v, pos, &operator))
                   2417:                        break;
                   2418:
1.133     schwarze 2419:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2420:                        while (isspace((unsigned char)v[*pos]))
                   2421:                                (*pos)++;
                   2422:
1.133     schwarze 2423:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2424:                        return 0;
1.77      schwarze 2425:
1.133     schwarze 2426:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2427:                        while (isspace((unsigned char)v[*pos]))
                   2428:                                (*pos)++;
1.79      schwarze 2429:
                   2430:                if (NULL == res)
                   2431:                        continue;
1.77      schwarze 2432:
                   2433:                switch (operator) {
1.80      schwarze 2434:                case '+':
1.77      schwarze 2435:                        *res += operand2;
                   2436:                        break;
1.80      schwarze 2437:                case '-':
1.77      schwarze 2438:                        *res -= operand2;
                   2439:                        break;
1.80      schwarze 2440:                case '*':
1.77      schwarze 2441:                        *res *= operand2;
                   2442:                        break;
1.80      schwarze 2443:                case '/':
1.116     schwarze 2444:                        if (operand2 == 0) {
1.106     schwarze 2445:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2446:                                        r->parse, ln, *pos, v);
                   2447:                                *res = 0;
                   2448:                                break;
                   2449:                        }
1.77      schwarze 2450:                        *res /= operand2;
                   2451:                        break;
1.80      schwarze 2452:                case '%':
1.116     schwarze 2453:                        if (operand2 == 0) {
                   2454:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2455:                                        r->parse, ln, *pos, v);
                   2456:                                *res = 0;
                   2457:                                break;
                   2458:                        }
1.77      schwarze 2459:                        *res %= operand2;
                   2460:                        break;
1.80      schwarze 2461:                case '<':
1.77      schwarze 2462:                        *res = *res < operand2;
                   2463:                        break;
1.80      schwarze 2464:                case '>':
1.77      schwarze 2465:                        *res = *res > operand2;
                   2466:                        break;
1.80      schwarze 2467:                case 'l':
1.77      schwarze 2468:                        *res = *res <= operand2;
                   2469:                        break;
1.80      schwarze 2470:                case 'g':
1.77      schwarze 2471:                        *res = *res >= operand2;
                   2472:                        break;
1.80      schwarze 2473:                case '=':
1.77      schwarze 2474:                        *res = *res == operand2;
                   2475:                        break;
1.80      schwarze 2476:                case '!':
1.77      schwarze 2477:                        *res = *res != operand2;
                   2478:                        break;
1.80      schwarze 2479:                case '&':
1.77      schwarze 2480:                        *res = *res && operand2;
                   2481:                        break;
1.80      schwarze 2482:                case ':':
1.77      schwarze 2483:                        *res = *res || operand2;
                   2484:                        break;
1.80      schwarze 2485:                case 'i':
1.77      schwarze 2486:                        if (operand2 < *res)
                   2487:                                *res = operand2;
                   2488:                        break;
1.80      schwarze 2489:                case 'a':
1.77      schwarze 2490:                        if (operand2 > *res)
                   2491:                                *res = operand2;
                   2492:                        break;
                   2493:                default:
                   2494:                        abort();
                   2495:                }
                   2496:        }
1.149     schwarze 2497:        return 1;
1.77      schwarze 2498: }
                   2499:
1.138     schwarze 2500: /* --- register management ------------------------------------------------ */
                   2501:
1.52      schwarze 2502: void
1.60      schwarze 2503: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2504: {
1.52      schwarze 2505:        struct roffreg  *reg;
                   2506:
                   2507:        /* Search for an existing register with the same name. */
                   2508:        reg = r->regtab;
                   2509:
                   2510:        while (reg && strcmp(name, reg->key.p))
                   2511:                reg = reg->next;
1.41      schwarze 2512:
1.52      schwarze 2513:        if (NULL == reg) {
                   2514:                /* Create a new register. */
                   2515:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2516:                reg->key.p = mandoc_strdup(name);
                   2517:                reg->key.sz = strlen(name);
1.60      schwarze 2518:                reg->val = 0;
1.52      schwarze 2519:                reg->next = r->regtab;
                   2520:                r->regtab = reg;
                   2521:        }
                   2522:
1.60      schwarze 2523:        if ('+' == sign)
                   2524:                reg->val += val;
                   2525:        else if ('-' == sign)
                   2526:                reg->val -= val;
                   2527:        else
                   2528:                reg->val = val;
1.41      schwarze 2529: }
                   2530:
1.65      schwarze 2531: /*
                   2532:  * Handle some predefined read-only number registers.
                   2533:  * For now, return -1 if the requested register is not predefined;
                   2534:  * in case a predefined read-only register having the value -1
                   2535:  * were to turn up, another special value would have to be chosen.
                   2536:  */
                   2537: static int
1.145     schwarze 2538: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2539: {
                   2540:
                   2541:        switch (*name) {
1.145     schwarze 2542:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2543:                return r->argc;
1.80      schwarze 2544:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2545:                return 0;
1.80      schwarze 2546:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2547:                return 1;
1.80      schwarze 2548:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2549:                return 24;
1.80      schwarze 2550:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2551:                return 0;
1.80      schwarze 2552:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2553:                return 1;
1.80      schwarze 2554:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2555:                return 40;
1.65      schwarze 2556:        default:
1.149     schwarze 2557:                return -1;
1.65      schwarze 2558:        }
                   2559: }
                   2560:
1.53      schwarze 2561: int
1.52      schwarze 2562: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2563: {
1.52      schwarze 2564:        struct roffreg  *reg;
1.65      schwarze 2565:        int              val;
                   2566:
                   2567:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145     schwarze 2568:                val = roff_getregro(r, name + 1);
1.65      schwarze 2569:                if (-1 != val)
1.149     schwarze 2570:                        return val;
1.65      schwarze 2571:        }
1.52      schwarze 2572:
                   2573:        for (reg = r->regtab; reg; reg = reg->next)
                   2574:                if (0 == strcmp(name, reg->key.p))
1.149     schwarze 2575:                        return reg->val;
1.53      schwarze 2576:
1.149     schwarze 2577:        return 0;
1.53      schwarze 2578: }
                   2579:
                   2580: static int
                   2581: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2582: {
                   2583:        struct roffreg  *reg;
1.65      schwarze 2584:        int              val;
                   2585:
                   2586:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2587:                val = roff_getregro(r, name + 1);
1.65      schwarze 2588:                if (-1 != val)
1.149     schwarze 2589:                        return val;
1.65      schwarze 2590:        }
1.53      schwarze 2591:
                   2592:        for (reg = r->regtab; reg; reg = reg->next)
                   2593:                if (len == reg->key.sz &&
                   2594:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2595:                        return reg->val;
1.143     schwarze 2596:
1.149     schwarze 2597:        return 0;
1.143     schwarze 2598: }
                   2599:
                   2600: static int
                   2601: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2602: {
                   2603:        struct roffreg  *reg;
                   2604:        int              val;
                   2605:
                   2606:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2607:                val = roff_getregro(r, name + 1);
1.143     schwarze 2608:                if (-1 != val)
1.149     schwarze 2609:                        return 1;
1.143     schwarze 2610:        }
                   2611:
                   2612:        for (reg = r->regtab; reg; reg = reg->next)
                   2613:                if (len == reg->key.sz &&
                   2614:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2615:                        return 1;
1.41      schwarze 2616:
1.149     schwarze 2617:        return 0;
1.41      schwarze 2618: }
                   2619:
1.52      schwarze 2620: static void
                   2621: roff_freereg(struct roffreg *reg)
1.41      schwarze 2622: {
1.52      schwarze 2623:        struct roffreg  *old_reg;
1.41      schwarze 2624:
1.52      schwarze 2625:        while (NULL != reg) {
                   2626:                free(reg->key.p);
                   2627:                old_reg = reg;
                   2628:                reg = reg->next;
                   2629:                free(old_reg);
                   2630:        }
1.41      schwarze 2631: }
1.7       schwarze 2632:
                   2633: static enum rofferr
1.6       schwarze 2634: roff_nr(ROFF_ARGS)
1.1       schwarze 2635: {
1.85      schwarze 2636:        char            *key, *val;
                   2637:        size_t           keysz;
1.37      schwarze 2638:        int              iv;
1.60      schwarze 2639:        char             sign;
1.6       schwarze 2640:
1.110     schwarze 2641:        key = val = buf->buf + pos;
                   2642:        if (*key == '\0')
1.149     schwarze 2643:                return ROFF_IGN;
1.85      schwarze 2644:
                   2645:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2646:        if (key[keysz] == '\\')
1.149     schwarze 2647:                return ROFF_IGN;
1.85      schwarze 2648:        key[keysz] = '\0';
1.6       schwarze 2649:
1.60      schwarze 2650:        sign = *val;
1.110     schwarze 2651:        if (sign == '+' || sign == '-')
1.60      schwarze 2652:                val++;
                   2653:
1.133     schwarze 2654:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77      schwarze 2655:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2656:
1.149     schwarze 2657:        return ROFF_IGN;
1.76      schwarze 2658: }
                   2659:
                   2660: static enum rofferr
                   2661: roff_rr(ROFF_ARGS)
                   2662: {
                   2663:        struct roffreg  *reg, **prev;
1.85      schwarze 2664:        char            *name, *cp;
                   2665:        size_t           namesz;
1.76      schwarze 2666:
1.110     schwarze 2667:        name = cp = buf->buf + pos;
                   2668:        if (*name == '\0')
1.149     schwarze 2669:                return ROFF_IGN;
1.85      schwarze 2670:        namesz = roff_getname(r, &cp, ln, pos);
                   2671:        name[namesz] = '\0';
1.76      schwarze 2672:
                   2673:        prev = &r->regtab;
                   2674:        while (1) {
                   2675:                reg = *prev;
1.110     schwarze 2676:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2677:                        break;
                   2678:                prev = &reg->next;
                   2679:        }
1.110     schwarze 2680:        if (reg != NULL) {
1.76      schwarze 2681:                *prev = reg->next;
                   2682:                free(reg->key.p);
                   2683:                free(reg);
                   2684:        }
1.149     schwarze 2685:        return ROFF_IGN;
1.29      schwarze 2686: }
                   2687:
1.138     schwarze 2688: /* --- handler functions for roff requests -------------------------------- */
                   2689:
1.29      schwarze 2690: static enum rofferr
                   2691: roff_rm(ROFF_ARGS)
                   2692: {
                   2693:        const char       *name;
                   2694:        char             *cp;
1.85      schwarze 2695:        size_t            namesz;
1.29      schwarze 2696:
1.110     schwarze 2697:        cp = buf->buf + pos;
                   2698:        while (*cp != '\0') {
1.85      schwarze 2699:                name = cp;
1.110     schwarze 2700:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2701:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183     schwarze 2702:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.110     schwarze 2703:                if (name[namesz] == '\\')
1.85      schwarze 2704:                        break;
1.29      schwarze 2705:        }
1.149     schwarze 2706:        return ROFF_IGN;
1.51      schwarze 2707: }
                   2708:
                   2709: static enum rofferr
                   2710: roff_it(ROFF_ARGS)
                   2711: {
                   2712:        int              iv;
                   2713:
                   2714:        /* Parse the number of lines. */
1.133     schwarze 2715:
                   2716:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2717:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2718:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2719:                return ROFF_IGN;
1.51      schwarze 2720:        }
                   2721:
1.134     schwarze 2722:        while (isspace((unsigned char)buf->buf[pos]))
                   2723:                pos++;
                   2724:
                   2725:        /*
                   2726:         * Arm the input line trap.
                   2727:         * Special-casing "an-trap" is an ugly workaround to cope
                   2728:         * with DocBook stupidly fiddling with man(7) internals.
                   2729:         */
1.133     schwarze 2730:
1.51      schwarze 2731:        roffit_lines = iv;
1.134     schwarze 2732:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2733:            strcmp(buf->buf + pos, "an-trap") ?
                   2734:            buf->buf + pos : "br");
1.149     schwarze 2735:        return ROFF_IGN;
1.47      schwarze 2736: }
                   2737:
                   2738: static enum rofferr
                   2739: roff_Dd(ROFF_ARGS)
                   2740: {
1.187     schwarze 2741:        int              mask;
                   2742:        enum roff_tok    t, te;
1.99      schwarze 2743:
1.187     schwarze 2744:        switch (tok) {
                   2745:        case ROFF_Dd:
                   2746:                tok = MDOC_Dd;
                   2747:                te = MDOC_MAX;
                   2748:                if (r->format == 0)
                   2749:                        r->format = MPARSE_MDOC;
                   2750:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   2751:                break;
                   2752:        case ROFF_TH:
                   2753:                tok = MAN_TH;
                   2754:                te = MAN_MAX;
                   2755:                if (r->format == 0)
                   2756:                        r->format = MPARSE_MAN;
                   2757:                mask = MPARSE_QUICK;
                   2758:                break;
                   2759:        default:
                   2760:                abort();
                   2761:        }
                   2762:        if ((r->options & mask) == 0)
                   2763:                for (t = tok; t < te; t++)
                   2764:                        roff_setstr(r, roff_name[t], NULL, 0);
1.149     schwarze 2765:        return ROFF_CONT;
1.14      schwarze 2766: }
                   2767:
                   2768: static enum rofferr
1.27      schwarze 2769: roff_TE(ROFF_ARGS)
                   2770: {
                   2771:
                   2772:        if (NULL == r->tbl)
1.94      schwarze 2773:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2774:                    ln, ppos, "TE");
1.130     schwarze 2775:        else if ( ! tbl_end(&r->tbl)) {
                   2776:                free(buf->buf);
                   2777:                buf->buf = mandoc_strdup(".sp");
                   2778:                buf->sz = 4;
1.149     schwarze 2779:                return ROFF_REPARSE;
1.130     schwarze 2780:        }
1.149     schwarze 2781:        return ROFF_IGN;
1.27      schwarze 2782: }
                   2783:
                   2784: static enum rofferr
                   2785: roff_T_(ROFF_ARGS)
                   2786: {
                   2787:
                   2788:        if (NULL == r->tbl)
1.94      schwarze 2789:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2790:                    ln, ppos, "T&");
1.27      schwarze 2791:        else
1.168     schwarze 2792:                tbl_restart(ln, ppos, r->tbl);
1.27      schwarze 2793:
1.149     schwarze 2794:        return ROFF_IGN;
1.27      schwarze 2795: }
                   2796:
1.102     schwarze 2797: /*
                   2798:  * Handle in-line equation delimiters.
                   2799:  */
                   2800: static enum rofferr
1.110     schwarze 2801: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2802: {
1.105     schwarze 2803:        char            *cp1, *cp2;
                   2804:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2805:
1.102     schwarze 2806:        /*
                   2807:         * Outside equations, look for an opening delimiter.
                   2808:         * If we are inside an equation, we already know it is
                   2809:         * in-line, or this function wouldn't have been called;
                   2810:         * so look for a closing delimiter.
                   2811:         */
                   2812:
1.110     schwarze 2813:        cp1 = buf->buf + pos;
1.102     schwarze 2814:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2815:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2816:        if (cp2 == NULL)
1.149     schwarze 2817:                return ROFF_CONT;
1.102     schwarze 2818:
1.105     schwarze 2819:        *cp2++ = '\0';
                   2820:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2821:
                   2822:        /* Handle preceding text, protecting whitespace. */
                   2823:
1.110     schwarze 2824:        if (*buf->buf != '\0') {
1.105     schwarze 2825:                if (r->eqn == NULL)
                   2826:                        bef_pr = "\\&";
                   2827:                bef_nl = "\n";
                   2828:        }
                   2829:
                   2830:        /*
                   2831:         * Prepare replacing the delimiter with an equation macro
                   2832:         * and drop leading white space from the equation.
                   2833:         */
1.102     schwarze 2834:
1.105     schwarze 2835:        if (r->eqn == NULL) {
                   2836:                while (*cp2 == ' ')
                   2837:                        cp2++;
                   2838:                mac = ".EQ";
                   2839:        } else
                   2840:                mac = ".EN";
                   2841:
                   2842:        /* Handle following text, protecting whitespace. */
                   2843:
                   2844:        if (*cp2 != '\0') {
                   2845:                aft_nl = "\n";
                   2846:                if (r->eqn != NULL)
                   2847:                        aft_pr = "\\&";
                   2848:        }
                   2849:
                   2850:        /* Do the actual replacement. */
                   2851:
1.110     schwarze 2852:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2853:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2854:        free(buf->buf);
                   2855:        buf->buf = cp1;
1.102     schwarze 2856:
                   2857:        /* Toggle the in-line state of the eqn subsystem. */
                   2858:
                   2859:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2860:        return ROFF_REPARSE;
1.41      schwarze 2861: }
                   2862:
1.107     schwarze 2863: static enum rofferr
                   2864: roff_EQ(ROFF_ARGS)
1.32      schwarze 2865: {
1.191     schwarze 2866:        struct roff_node        *n;
                   2867:
                   2868:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   2869:        if (ln > r->man->last->line)
                   2870:                n->flags |= NODE_LINE;
                   2871:        n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
                   2872:        n->eqn->expectargs = UINT_MAX;
                   2873:        roff_node_append(r->man, n);
                   2874:        r->man->next = ROFF_NEXT_SIBLING;
1.32      schwarze 2875:
1.110     schwarze 2876:        assert(r->eqn == NULL);
1.191     schwarze 2877:        if (r->last_eqn == NULL)
                   2878:                r->last_eqn = eqn_alloc(r->parse);
                   2879:        else
                   2880:                eqn_reset(r->last_eqn);
                   2881:        r->eqn = r->last_eqn;
                   2882:        r->eqn->node = n;
1.41      schwarze 2883:
1.110     schwarze 2884:        if (buf->buf[pos] != '\0')
1.107     schwarze 2885:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2886:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2887:
1.149     schwarze 2888:        return ROFF_IGN;
1.32      schwarze 2889: }
                   2890:
                   2891: static enum rofferr
                   2892: roff_EN(ROFF_ARGS)
                   2893: {
1.191     schwarze 2894:        if (r->eqn != NULL) {
                   2895:                eqn_parse(r->eqn);
                   2896:                r->eqn = NULL;
                   2897:        } else
                   2898:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
                   2899:        if (buf->buf[pos] != '\0')
                   2900:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   2901:                    "EN %s", buf->buf + pos);
1.149     schwarze 2902:        return ROFF_IGN;
1.32      schwarze 2903: }
                   2904:
                   2905: static enum rofferr
1.27      schwarze 2906: roff_TS(ROFF_ARGS)
                   2907: {
1.49      schwarze 2908:        struct tbl_node *tbl;
1.27      schwarze 2909:
                   2910:        if (r->tbl) {
1.94      schwarze 2911:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2912:                    ln, ppos, "TS breaks TS");
1.41      schwarze 2913:                tbl_end(&r->tbl);
1.27      schwarze 2914:        }
                   2915:
1.49      schwarze 2916:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2917:
                   2918:        if (r->last_tbl)
1.49      schwarze 2919:                r->last_tbl->next = tbl;
1.27      schwarze 2920:        else
1.49      schwarze 2921:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 2922:
1.49      schwarze 2923:        r->tbl = r->last_tbl = tbl;
1.169     schwarze 2924:        return ROFF_IGN;
                   2925: }
                   2926:
                   2927: static enum rofferr
                   2928: roff_onearg(ROFF_ARGS)
                   2929: {
                   2930:        struct roff_node        *n;
                   2931:        char                    *cp;
1.177     schwarze 2932:        int                      npos;
1.169     schwarze 2933:
1.174     schwarze 2934:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.192   ! schwarze 2935:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
        !          2936:             tok == ROFF_ti))
1.174     schwarze 2937:                man_breakscope(r->man, tok);
                   2938:
1.181     schwarze 2939:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177     schwarze 2940:                r->man->last = roffce_node;
                   2941:                r->man->next = ROFF_NEXT_SIBLING;
                   2942:        }
                   2943:
1.169     schwarze 2944:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2945:        n = r->man->last;
                   2946:
                   2947:        cp = buf->buf + pos;
                   2948:        if (*cp != '\0') {
                   2949:                while (*cp != '\0' && *cp != ' ')
                   2950:                        cp++;
                   2951:                while (*cp == ' ')
                   2952:                        *cp++ = '\0';
                   2953:                if (*cp != '\0')
                   2954:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS,
                   2955:                            r->parse, ln, cp - buf->buf,
                   2956:                            "%s ... %s", roff_name[tok], cp);
                   2957:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172     schwarze 2958:        }
                   2959:
1.181     schwarze 2960:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   2961:                if (r->man->last->type == ROFFT_ELEM) {
1.177     schwarze 2962:                        roff_word_alloc(r->man, ln, pos, "1");
                   2963:                        r->man->last->flags |= NODE_NOSRC;
                   2964:                }
                   2965:                npos = 0;
                   2966:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   2967:                    &roffce_lines, 0) == 0) {
                   2968:                        mandoc_vmsg(MANDOCERR_CE_NONUM,
                   2969:                            r->parse, ln, pos, "ce %s", buf->buf + pos);
                   2970:                        roffce_lines = 1;
                   2971:                }
                   2972:                if (roffce_lines < 1) {
                   2973:                        r->man->last = r->man->last->parent;
                   2974:                        roffce_node = NULL;
                   2975:                        roffce_lines = 0;
                   2976:                } else
                   2977:                        roffce_node = r->man->last->parent;
                   2978:        } else {
                   2979:                n->flags |= NODE_VALID | NODE_ENDED;
                   2980:                r->man->last = n;
                   2981:        }
                   2982:        n->flags |= NODE_LINE;
1.172     schwarze 2983:        r->man->next = ROFF_NEXT_SIBLING;
                   2984:        return ROFF_IGN;
                   2985: }
                   2986:
                   2987: static enum rofferr
                   2988: roff_manyarg(ROFF_ARGS)
                   2989: {
                   2990:        struct roff_node        *n;
                   2991:        char                    *sp, *ep;
                   2992:
                   2993:        roff_elem_alloc(r->man, ln, ppos, tok);
                   2994:        n = r->man->last;
                   2995:
                   2996:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   2997:                while (*ep != '\0' && *ep != ' ')
                   2998:                        ep++;
                   2999:                while (*ep == ' ')
                   3000:                        *ep++ = '\0';
                   3001:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169     schwarze 3002:        }
                   3003:
                   3004:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3005:        r->man->last = n;
                   3006:        r->man->next = ROFF_NEXT_SIBLING;
1.149     schwarze 3007:        return ROFF_IGN;
1.123     schwarze 3008: }
                   3009:
                   3010: static enum rofferr
1.183     schwarze 3011: roff_als(ROFF_ARGS)
                   3012: {
                   3013:        char            *oldn, *newn, *end, *value;
                   3014:        size_t           oldsz, newsz, valsz;
                   3015:
                   3016:        newn = oldn = buf->buf + pos;
                   3017:        if (*newn == '\0')
                   3018:                return ROFF_IGN;
                   3019:
                   3020:        newsz = roff_getname(r, &oldn, ln, pos);
                   3021:        if (newn[newsz] == '\\' || *oldn == '\0')
                   3022:                return ROFF_IGN;
                   3023:
                   3024:        end = oldn;
                   3025:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3026:        if (oldsz == 0)
                   3027:                return ROFF_IGN;
                   3028:
1.187     schwarze 3029:        valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
                   3030:            (int)oldsz, oldn);
1.183     schwarze 3031:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3032:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3033:        free(value);
                   3034:        return ROFF_IGN;
                   3035: }
                   3036:
                   3037: static enum rofferr
1.168     schwarze 3038: roff_br(ROFF_ARGS)
1.123     schwarze 3039: {
1.174     schwarze 3040:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3041:                man_breakscope(r->man, ROFF_br);
1.168     schwarze 3042:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
                   3043:        if (buf->buf[pos] != '\0')
                   3044:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   3045:                    "%s %s", roff_name[tok], buf->buf + pos);
                   3046:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3047:        r->man->next = ROFF_NEXT_SIBLING;
                   3048:        return ROFF_IGN;
1.27      schwarze 3049: }
                   3050:
                   3051: static enum rofferr
1.48      schwarze 3052: roff_cc(ROFF_ARGS)
                   3053: {
                   3054:        const char      *p;
                   3055:
1.110     schwarze 3056:        p = buf->buf + pos;
1.48      schwarze 3057:
1.110     schwarze 3058:        if (*p == '\0' || (r->control = *p++) == '.')
1.175     schwarze 3059:                r->control = '\0';
1.48      schwarze 3060:
1.110     schwarze 3061:        if (*p != '\0')
1.132     schwarze 3062:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3063:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 3064:
1.149     schwarze 3065:        return ROFF_IGN;
1.48      schwarze 3066: }
                   3067:
                   3068: static enum rofferr
1.175     schwarze 3069: roff_ec(ROFF_ARGS)
                   3070: {
                   3071:        const char      *p;
                   3072:
                   3073:        p = buf->buf + pos;
                   3074:        if (*p == '\0')
                   3075:                r->escape = '\\';
                   3076:        else {
                   3077:                r->escape = *p;
                   3078:                if (*++p != '\0')
                   3079:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   3080:                            ln, p - buf->buf, "ec ... %s", p);
                   3081:        }
                   3082:        return ROFF_IGN;
                   3083: }
                   3084:
                   3085: static enum rofferr
                   3086: roff_eo(ROFF_ARGS)
                   3087: {
                   3088:        r->escape = '\0';
                   3089:        if (buf->buf[pos] != '\0')
                   3090:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
                   3091:                    ln, pos, "eo %s", buf->buf + pos);
                   3092:        return ROFF_IGN;
                   3093: }
                   3094:
                   3095: static enum rofferr
1.42      schwarze 3096: roff_tr(ROFF_ARGS)
                   3097: {
                   3098:        const char      *p, *first, *second;
                   3099:        size_t           fsz, ssz;
                   3100:        enum mandoc_esc  esc;
                   3101:
1.110     schwarze 3102:        p = buf->buf + pos;
1.42      schwarze 3103:
1.110     schwarze 3104:        if (*p == '\0') {
1.132     schwarze 3105:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 3106:                return ROFF_IGN;
1.42      schwarze 3107:        }
                   3108:
1.110     schwarze 3109:        while (*p != '\0') {
1.42      schwarze 3110:                fsz = ssz = 1;
                   3111:
                   3112:                first = p++;
1.110     schwarze 3113:                if (*first == '\\') {
1.42      schwarze 3114:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3115:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3116:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3117:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 3118:                                return ROFF_IGN;
1.42      schwarze 3119:                        }
                   3120:                        fsz = (size_t)(p - first);
                   3121:                }
                   3122:
                   3123:                second = p++;
1.110     schwarze 3124:                if (*second == '\\') {
1.42      schwarze 3125:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 3126:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 3127:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 3128:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 3129:                                return ROFF_IGN;
1.42      schwarze 3130:                        }
                   3131:                        ssz = (size_t)(p - second);
1.110     schwarze 3132:                } else if (*second == '\0') {
1.132     schwarze 3133:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   3134:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 3135:                        second = " ";
                   3136:                        p--;
                   3137:                }
                   3138:
                   3139:                if (fsz > 1) {
1.80      schwarze 3140:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3141:                            second, ssz, 0);
1.42      schwarze 3142:                        continue;
                   3143:                }
                   3144:
1.110     schwarze 3145:                if (r->xtab == NULL)
1.80      schwarze 3146:                        r->xtab = mandoc_calloc(128,
                   3147:                            sizeof(struct roffstr));
1.42      schwarze 3148:
                   3149:                free(r->xtab[(int)*first].p);
                   3150:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3151:                r->xtab[(int)*first].sz = ssz;
                   3152:        }
                   3153:
1.149     schwarze 3154:        return ROFF_IGN;
1.42      schwarze 3155: }
                   3156:
                   3157: static enum rofferr
1.178     schwarze 3158: roff_rn(ROFF_ARGS)
                   3159: {
                   3160:        const char      *value;
                   3161:        char            *oldn, *newn, *end;
                   3162:        size_t           oldsz, newsz;
1.187     schwarze 3163:        int              deftype;
1.178     schwarze 3164:
                   3165:        oldn = newn = buf->buf + pos;
                   3166:        if (*oldn == '\0')
                   3167:                return ROFF_IGN;
                   3168:
                   3169:        oldsz = roff_getname(r, &newn, ln, pos);
                   3170:        if (oldn[oldsz] == '\\' || *newn == '\0')
                   3171:                return ROFF_IGN;
                   3172:
                   3173:        end = newn;
                   3174:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3175:        if (newsz == 0)
                   3176:                return ROFF_IGN;
                   3177:
1.187     schwarze 3178:        deftype = ROFFDEF_ANY;
                   3179:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3180:        switch (deftype) {
                   3181:        case ROFFDEF_USER:
1.178     schwarze 3182:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3183:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3184:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.187     schwarze 3185:                break;
                   3186:        case ROFFDEF_PRE:
                   3187:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3188:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3189:                break;
                   3190:        case ROFFDEF_REN:
1.178     schwarze 3191:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3192:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.187     schwarze 3193:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3194:                break;
                   3195:        case ROFFDEF_STD:
1.178     schwarze 3196:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.187     schwarze 3197:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3198:                break;
                   3199:        default:
                   3200:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3201:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3202:                break;
                   3203:        }
1.178     schwarze 3204:        return ROFF_IGN;
                   3205: }
                   3206:
                   3207: static enum rofferr
1.14      schwarze 3208: roff_so(ROFF_ARGS)
                   3209: {
1.121     schwarze 3210:        char *name, *cp;
1.15      schwarze 3211:
1.110     schwarze 3212:        name = buf->buf + pos;
1.97      schwarze 3213:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 3214:
1.22      schwarze 3215:        /*
                   3216:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3217:         * opening anything that's not in our cwd or anything beneath
                   3218:         * it.  Thus, explicitly disallow traversing up the file-system
                   3219:         * or using absolute paths.
                   3220:         */
                   3221:
1.110     schwarze 3222:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 3223:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   3224:                    ".so %s", name);
1.121     schwarze 3225:                buf->sz = mandoc_asprintf(&cp,
                   3226:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3227:                free(buf->buf);
                   3228:                buf->buf = cp;
                   3229:                *offs = 0;
1.149     schwarze 3230:                return ROFF_REPARSE;
1.14      schwarze 3231:        }
                   3232:
                   3233:        *offs = pos;
1.149     schwarze 3234:        return ROFF_SO;
1.7       schwarze 3235: }
                   3236:
1.138     schwarze 3237: /* --- user defined strings and macros ------------------------------------ */
                   3238:
1.16      schwarze 3239: static enum rofferr
                   3240: roff_userdef(ROFF_ARGS)
1.12      schwarze 3241: {
1.183     schwarze 3242:        const char       *arg[16], *ap;
1.16      schwarze 3243:        char             *cp, *n1, *n2;
1.164     schwarze 3244:        int               expand_count, i, ib, ie;
1.135     schwarze 3245:        size_t            asz, rsz;
1.12      schwarze 3246:
1.16      schwarze 3247:        /*
                   3248:         * Collect pointers to macro argument strings
1.61      schwarze 3249:         * and NUL-terminate them.
1.16      schwarze 3250:         */
1.135     schwarze 3251:
1.145     schwarze 3252:        r->argc = 0;
1.110     schwarze 3253:        cp = buf->buf + pos;
1.183     schwarze 3254:        for (i = 0; i < 16; i++) {
1.145     schwarze 3255:                if (*cp == '\0')
                   3256:                        arg[i] = "";
                   3257:                else {
                   3258:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   3259:                        r->argc = i + 1;
                   3260:                }
                   3261:        }
1.16      schwarze 3262:
                   3263:        /*
                   3264:         * Expand macro arguments.
1.12      schwarze 3265:         */
1.135     schwarze 3266:
                   3267:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 3268:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 3269:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 3270:        expand_count = 0;
1.135     schwarze 3271:        while (*cp != '\0') {
                   3272:
                   3273:                /* Scan ahead for the next argument invocation. */
                   3274:
                   3275:                if (*cp++ != '\\')
                   3276:                        continue;
                   3277:                if (*cp++ != '$')
                   3278:                        continue;
1.146     schwarze 3279:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   3280:                        ib = 0;
                   3281:                        ie = r->argc - 1;
                   3282:                } else {  /* \\$1 .. \\$9 insert one argument */
                   3283:                        ib = ie = *cp - '1';
                   3284:                        if (ib < 0 || ib > 8)
                   3285:                                continue;
                   3286:                }
1.135     schwarze 3287:                cp -= 2;
1.164     schwarze 3288:
                   3289:                /*
                   3290:                 * Prevent infinite recursion.
                   3291:                 */
                   3292:
                   3293:                if (cp >= n2)
                   3294:                        expand_count = 1;
                   3295:                else if (++expand_count > EXPAND_LIMIT) {
                   3296:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   3297:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 3298:                        free(buf->buf);
                   3299:                        buf->buf = n1;
1.164     schwarze 3300:                        return ROFF_IGN;
                   3301:                }
1.135     schwarze 3302:
                   3303:                /*
                   3304:                 * Determine the size of the expanded argument,
                   3305:                 * taking escaping of quotes into account.
                   3306:                 */
                   3307:
1.146     schwarze 3308:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   3309:                for (i = ib; i <= ie; i++) {
                   3310:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3311:                                asz++;
                   3312:                                if (*ap == '"')
                   3313:                                        asz += 3;
                   3314:                        }
1.135     schwarze 3315:                }
                   3316:                if (asz != 3) {
                   3317:
                   3318:                        /*
                   3319:                         * Determine the size of the rest of the
                   3320:                         * unexpanded macro, including the NUL.
                   3321:                         */
                   3322:
                   3323:                        rsz = buf->sz - (cp - n1) - 3;
                   3324:
                   3325:                        /*
                   3326:                         * When shrinking, move before
                   3327:                         * releasing the storage.
                   3328:                         */
                   3329:
                   3330:                        if (asz < 3)
                   3331:                                memmove(cp + asz, cp + 3, rsz);
                   3332:
                   3333:                        /*
                   3334:                         * Resize the storage for the macro
                   3335:                         * and readjust the parse pointer.
                   3336:                         */
                   3337:
                   3338:                        buf->sz += asz - 3;
                   3339:                        n2 = mandoc_realloc(n1, buf->sz);
                   3340:                        cp = n2 + (cp - n1);
                   3341:                        n1 = n2;
                   3342:
                   3343:                        /*
                   3344:                         * When growing, make room
                   3345:                         * for the expanded argument.
                   3346:                         */
                   3347:
                   3348:                        if (asz > 3)
                   3349:                                memmove(cp + asz, cp + 3, rsz);
                   3350:                }
                   3351:
                   3352:                /* Copy the expanded argument, escaping quotes. */
                   3353:
                   3354:                n2 = cp;
1.146     schwarze 3355:                for (i = ib; i <= ie; i++) {
                   3356:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3357:                                if (*ap == '"') {
                   3358:                                        memcpy(n2, "\\(dq", 4);
                   3359:                                        n2 += 4;
                   3360:                                } else
                   3361:                                        *n2++ = *ap;
                   3362:                        }
                   3363:                        if (i < ie)
                   3364:                                *n2++ = ' ';
1.16      schwarze 3365:                }
1.12      schwarze 3366:        }
                   3367:
1.16      schwarze 3368:        /*
                   3369:         * Replace the macro invocation
                   3370:         * by the expanded macro.
                   3371:         */
1.135     schwarze 3372:
1.110     schwarze 3373:        free(buf->buf);
                   3374:        buf->buf = n1;
1.120     schwarze 3375:        *offs = 0;
1.16      schwarze 3376:
1.149     schwarze 3377:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3378:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3379: }
1.28      schwarze 3380:
1.178     schwarze 3381: /*
                   3382:  * Calling a high-level macro that was renamed with .rn.
                   3383:  * r->current_string has already been set up by roff_parse().
                   3384:  */
                   3385: static enum rofferr
                   3386: roff_renamed(ROFF_ARGS)
                   3387: {
                   3388:        char    *nbuf;
                   3389:
1.187     schwarze 3390:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3391:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.178     schwarze 3392:        free(buf->buf);
                   3393:        buf->buf = nbuf;
                   3394:        return ROFF_CONT;
                   3395: }
                   3396:
1.85      schwarze 3397: static size_t
1.28      schwarze 3398: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3399: {
                   3400:        char     *name, *cp;
1.85      schwarze 3401:        size_t    namesz;
1.28      schwarze 3402:
                   3403:        name = *cpp;
                   3404:        if ('\0' == *name)
1.149     schwarze 3405:                return 0;
1.28      schwarze 3406:
1.85      schwarze 3407:        /* Read until end of name and terminate it with NUL. */
                   3408:        for (cp = name; 1; cp++) {
                   3409:                if ('\0' == *cp || ' ' == *cp) {
                   3410:                        namesz = cp - name;
                   3411:                        break;
                   3412:                }
1.28      schwarze 3413:                if ('\\' != *cp)
                   3414:                        continue;
1.88      schwarze 3415:                namesz = cp - name;
                   3416:                if ('{' == cp[1] || '}' == cp[1])
                   3417:                        break;
1.28      schwarze 3418:                cp++;
                   3419:                if ('\\' == *cp)
                   3420:                        continue;
1.97      schwarze 3421:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3422:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3423:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3424:                break;
1.28      schwarze 3425:        }
                   3426:
                   3427:        /* Read past spaces. */
                   3428:        while (' ' == *cp)
                   3429:                cp++;
                   3430:
                   3431:        *cpp = cp;
1.149     schwarze 3432:        return namesz;
1.28      schwarze 3433: }
                   3434:
1.16      schwarze 3435: /*
                   3436:  * Store *string into the user-defined string called *name.
                   3437:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3438:  * append == 0: replace mode
                   3439:  * append == 1: single-line append mode
                   3440:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3441:  */
1.8       schwarze 3442: static void
1.16      schwarze 3443: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3444:        int append)
1.7       schwarze 3445: {
1.187     schwarze 3446:        size_t   namesz;
1.42      schwarze 3447:
1.187     schwarze 3448:        namesz = strlen(name);
                   3449:        roff_setstrn(&r->strtab, name, namesz, string,
1.80      schwarze 3450:            string ? strlen(string) : 0, append);
1.187     schwarze 3451:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.42      schwarze 3452: }
                   3453:
                   3454: static void
                   3455: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3456:                const char *string, size_t stringsz, int append)
1.42      schwarze 3457: {
                   3458:        struct roffkv   *n;
                   3459:        char            *c;
                   3460:        int              i;
                   3461:        size_t           oldch, newch;
1.7       schwarze 3462:
1.16      schwarze 3463:        /* Search for an existing string with the same name. */
1.42      schwarze 3464:        n = *r;
                   3465:
1.84      schwarze 3466:        while (n && (namesz != n->key.sz ||
                   3467:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3468:                n = n->next;
1.8       schwarze 3469:
                   3470:        if (NULL == n) {
1.16      schwarze 3471:                /* Create a new string table entry. */
1.42      schwarze 3472:                n = mandoc_malloc(sizeof(struct roffkv));
                   3473:                n->key.p = mandoc_strndup(name, namesz);
                   3474:                n->key.sz = namesz;
                   3475:                n->val.p = NULL;
                   3476:                n->val.sz = 0;
                   3477:                n->next = *r;
                   3478:                *r = n;
1.66      schwarze 3479:        } else if (0 == append) {
1.42      schwarze 3480:                free(n->val.p);
                   3481:                n->val.p = NULL;
                   3482:                n->val.sz = 0;
1.16      schwarze 3483:        }
                   3484:
                   3485:        if (NULL == string)
                   3486:                return;
                   3487:
                   3488:        /*
                   3489:         * One additional byte for the '\n' in multiline mode,
                   3490:         * and one for the terminating '\0'.
                   3491:         */
1.66      schwarze 3492:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3493:
                   3494:        if (NULL == n->val.p) {
                   3495:                n->val.p = mandoc_malloc(newch);
                   3496:                *n->val.p = '\0';
1.16      schwarze 3497:                oldch = 0;
                   3498:        } else {
1.42      schwarze 3499:                oldch = n->val.sz;
                   3500:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3501:        }
                   3502:
                   3503:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3504:        c = n->val.p + (int)oldch;
1.16      schwarze 3505:
                   3506:        /* Append new content to the destination buffer. */
1.42      schwarze 3507:        i = 0;
                   3508:        while (i < (int)stringsz) {
1.16      schwarze 3509:                /*
                   3510:                 * Rudimentary roff copy mode:
                   3511:                 * Handle escaped backslashes.
                   3512:                 */
1.42      schwarze 3513:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3514:                        i++;
                   3515:                *c++ = string[i++];
1.16      schwarze 3516:        }
1.8       schwarze 3517:
1.16      schwarze 3518:        /* Append terminating bytes. */
1.66      schwarze 3519:        if (1 < append)
1.16      schwarze 3520:                *c++ = '\n';
1.42      schwarze 3521:
1.16      schwarze 3522:        *c = '\0';
1.42      schwarze 3523:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3524: }
                   3525:
1.8       schwarze 3526: static const char *
1.187     schwarze 3527: roff_getstrn(const struct roff *r, const char *name, size_t len,
                   3528:     int *deftype)
1.7       schwarze 3529: {
1.187     schwarze 3530:        const struct roffkv     *n;
                   3531:        int                      i;
                   3532:        enum roff_tok            tok;
                   3533:
                   3534:        if (*deftype & ROFFDEF_USER) {
                   3535:                for (n = r->strtab; n != NULL; n = n->next) {
                   3536:                        if (strncmp(name, n->key.p, len) == 0 &&
                   3537:                            n->key.p[len] == '\0' &&
                   3538:                            n->val.p != NULL) {
                   3539:                                *deftype = ROFFDEF_USER;
                   3540:                                return n->val.p;
                   3541:                        }
                   3542:                }
                   3543:        }
                   3544:        if (*deftype & ROFFDEF_PRE) {
                   3545:                for (i = 0; i < PREDEFS_MAX; i++) {
                   3546:                        if (strncmp(name, predefs[i].name, len) == 0 &&
                   3547:                            predefs[i].name[len] == '\0') {
                   3548:                                *deftype = ROFFDEF_PRE;
                   3549:                                return predefs[i].str;
                   3550:                        }
                   3551:                }
                   3552:        }
                   3553:        if (*deftype & ROFFDEF_REN) {
                   3554:                for (n = r->rentab; n != NULL; n = n->next) {
                   3555:                        if (strncmp(name, n->key.p, len) == 0 &&
                   3556:                            n->key.p[len] == '\0' &&
                   3557:                            n->val.p != NULL) {
                   3558:                                *deftype = ROFFDEF_REN;
                   3559:                                return n->val.p;
                   3560:                        }
                   3561:                }
                   3562:        }
                   3563:        if (*deftype & ROFFDEF_STD) {
                   3564:                if (r->man->macroset != MACROSET_MAN) {
                   3565:                        for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
                   3566:                                if (strncmp(name, roff_name[tok], len) == 0 &&
                   3567:                                    roff_name[tok][len] == '\0') {
                   3568:                                        *deftype = ROFFDEF_STD;
                   3569:                                        return NULL;
                   3570:                                }
                   3571:                        }
                   3572:                }
                   3573:                if (r->man->macroset != MACROSET_MDOC) {
                   3574:                        for (tok = MAN_TH; tok < MAN_MAX; tok++) {
                   3575:                                if (strncmp(name, roff_name[tok], len) == 0 &&
                   3576:                                    roff_name[tok][len] == '\0') {
                   3577:                                        *deftype = ROFFDEF_STD;
                   3578:                                        return NULL;
                   3579:                                }
                   3580:                        }
                   3581:                }
                   3582:        }
                   3583:        *deftype = 0;
1.149     schwarze 3584:        return NULL;
1.7       schwarze 3585: }
                   3586:
1.8       schwarze 3587: static void
1.42      schwarze 3588: roff_freestr(struct roffkv *r)
1.7       schwarze 3589: {
1.42      schwarze 3590:        struct roffkv    *n, *nn;
1.7       schwarze 3591:
1.42      schwarze 3592:        for (n = r; n; n = nn) {
                   3593:                free(n->key.p);
                   3594:                free(n->val.p);
1.7       schwarze 3595:                nn = n->next;
                   3596:                free(n);
                   3597:        }
1.27      schwarze 3598: }
1.138     schwarze 3599:
                   3600: /* --- accessors and utility functions ------------------------------------ */
1.27      schwarze 3601:
                   3602: const struct tbl_span *
                   3603: roff_span(const struct roff *r)
                   3604: {
1.80      schwarze 3605:
1.149     schwarze 3606:        return r->tbl ? tbl_span(r->tbl) : NULL;
1.42      schwarze 3607: }
                   3608:
                   3609: /*
                   3610:  * Duplicate an input string, making the appropriate character
                   3611:  * conversations (as stipulated by `tr') along the way.
                   3612:  * Returns a heap-allocated string with all the replacements made.
                   3613:  */
                   3614: char *
                   3615: roff_strdup(const struct roff *r, const char *p)
                   3616: {
                   3617:        const struct roffkv *cp;
                   3618:        char            *res;
                   3619:        const char      *pp;
                   3620:        size_t           ssz, sz;
                   3621:        enum mandoc_esc  esc;
                   3622:
                   3623:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3624:                return mandoc_strdup(p);
1.42      schwarze 3625:        else if ('\0' == *p)
1.149     schwarze 3626:                return mandoc_strdup("");
1.42      schwarze 3627:
                   3628:        /*
                   3629:         * Step through each character looking for term matches
                   3630:         * (remember that a `tr' can be invoked with an escape, which is
                   3631:         * a glyph but the escape is multi-character).
                   3632:         * We only do this if the character hash has been initialised
                   3633:         * and the string is >0 length.
                   3634:         */
                   3635:
                   3636:        res = NULL;
                   3637:        ssz = 0;
                   3638:
                   3639:        while ('\0' != *p) {
1.161     schwarze 3640:                assert((unsigned int)*p < 128);
                   3641:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3642:                        sz = r->xtab[(int)*p].sz;
                   3643:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3644:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3645:                        ssz += sz;
                   3646:                        p++;
                   3647:                        continue;
                   3648:                } else if ('\\' != *p) {
                   3649:                        res = mandoc_realloc(res, ssz + 2);
                   3650:                        res[ssz++] = *p++;
                   3651:                        continue;
                   3652:                }
                   3653:
                   3654:                /* Search for term matches. */
                   3655:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3656:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3657:                                break;
                   3658:
                   3659:                if (NULL != cp) {
                   3660:                        /*
                   3661:                         * A match has been found.
                   3662:                         * Append the match to the array and move
                   3663:                         * forward by its keysize.
                   3664:                         */
1.80      schwarze 3665:                        res = mandoc_realloc(res,
                   3666:                            ssz + cp->val.sz + 1);
1.42      schwarze 3667:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3668:                        ssz += cp->val.sz;
                   3669:                        p += (int)cp->key.sz;
                   3670:                        continue;
                   3671:                }
                   3672:
                   3673:                /*
                   3674:                 * Handle escapes carefully: we need to copy
                   3675:                 * over just the escape itself, or else we might
                   3676:                 * do replacements within the escape itself.
                   3677:                 * Make sure to pass along the bogus string.
                   3678:                 */
                   3679:                pp = p++;
                   3680:                esc = mandoc_escape(&p, NULL, NULL);
                   3681:                if (ESCAPE_ERROR == esc) {
                   3682:                        sz = strlen(pp);
                   3683:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3684:                        memcpy(res + ssz, pp, sz);
                   3685:                        break;
                   3686:                }
1.80      schwarze 3687:                /*
                   3688:                 * We bail out on bad escapes.
1.42      schwarze 3689:                 * No need to warn: we already did so when
                   3690:                 * roff_res() was called.
                   3691:                 */
                   3692:                sz = (int)(p - pp);
                   3693:                res = mandoc_realloc(res, ssz + sz + 1);
                   3694:                memcpy(res + ssz, pp, sz);
                   3695:                ssz += sz;
                   3696:        }
                   3697:
                   3698:        res[(int)ssz] = '\0';
1.149     schwarze 3699:        return res;
1.99      schwarze 3700: }
                   3701:
                   3702: int
                   3703: roff_getformat(const struct roff *r)
                   3704: {
                   3705:
1.149     schwarze 3706:        return r->format;
1.48      schwarze 3707: }
                   3708:
                   3709: /*
1.80      schwarze 3710:  * Find out whether a line is a macro line or not.
1.48      schwarze 3711:  * If it is, adjust the current position and return one; if it isn't,
                   3712:  * return zero and don't change the current position.
                   3713:  * If the control character has been set with `.cc', then let that grain
                   3714:  * precedence.
                   3715:  * This is slighly contrary to groff, where using the non-breaking
                   3716:  * control character when `cc' has been invoked will cause the
                   3717:  * non-breaking macro contents to be printed verbatim.
                   3718:  */
                   3719: int
                   3720: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3721: {
                   3722:        int             pos;
                   3723:
                   3724:        pos = *ppos;
                   3725:
1.175     schwarze 3726:        if (r->control != '\0' && cp[pos] == r->control)
1.48      schwarze 3727:                pos++;
1.175     schwarze 3728:        else if (r->control != '\0')
1.149     schwarze 3729:                return 0;
1.48      schwarze 3730:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3731:                pos += 2;
                   3732:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3733:                pos++;
                   3734:        else
1.149     schwarze 3735:                return 0;
1.48      schwarze 3736:
                   3737:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3738:                pos++;
                   3739:
                   3740:        *ppos = pos;
1.149     schwarze 3741:        return 1;
1.1       schwarze 3742: }