[BACK]Return to roff.c CVS log [TXT][DIR] Up to [cvsweb.bsd.lv] / mandoc

Annotation of mandoc/roff.c, Revision 1.296

1.296   ! schwarze    1: /*     $Id: roff.c,v 1.295 2017/04/29 12:45:42 schwarze Exp $ */
1.1       kristaps    2: /*
1.267     schwarze    3:  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.287     schwarze    4:  * Copyright (c) 2010-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
1.1       kristaps    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
1.66      kristaps    7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
1.1       kristaps    9:  *
1.106     kristaps   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66      kristaps   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106     kristaps   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66      kristaps   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.
1.1       kristaps   17:  */
1.66      kristaps   18: #include "config.h"
1.225     schwarze   19:
                     20: #include <sys/types.h>
1.30      kristaps   21:
1.67      kristaps   22: #include <assert.h>
1.85      kristaps   23: #include <ctype.h>
1.245     schwarze   24: #include <limits.h>
1.295     schwarze   25: #include <stddef.h>
                     26: #include <stdint.h>
1.178     schwarze   27: #include <stdio.h>
1.1       kristaps   28: #include <stdlib.h>
1.67      kristaps   29: #include <string.h>
1.1       kristaps   30:
1.67      kristaps   31: #include "mandoc.h"
1.201     schwarze   32: #include "mandoc_aux.h"
1.295     schwarze   33: #include "mandoc_ohash.h"
1.265     schwarze   34: #include "roff.h"
1.226     schwarze   35: #include "libmandoc.h"
1.266     schwarze   36: #include "roff_int.h"
1.109     kristaps   37: #include "libroff.h"
1.82      kristaps   38:
1.170     schwarze   39: /* Maximum number of string expansions per line, to break infinite loops. */
                     40: #define        EXPAND_LIMIT    1000
                     41:
1.266     schwarze   42: /* --- data types --------------------------------------------------------- */
                     43:
1.147     kristaps   44: /*
1.167     kristaps   45:  * An incredibly-simple string buffer.
                     46:  */
1.94      kristaps   47: struct roffstr {
1.167     kristaps   48:        char            *p; /* nil-terminated buffer */
                     49:        size_t           sz; /* saved strlen(p) */
1.166     kristaps   50: };
                     51:
                     52: /*
1.167     kristaps   53:  * A key-value roffstr pair as part of a singly-linked list.
1.166     kristaps   54:  */
                     55: struct roffkv {
                     56:        struct roffstr   key;
                     57:        struct roffstr   val;
                     58:        struct roffkv   *next; /* next in list */
1.94      kristaps   59: };
                     60:
1.180     schwarze   61: /*
                     62:  * A single number register as part of a singly-linked list.
                     63:  */
                     64: struct roffreg {
                     65:        struct roffstr   key;
1.181     schwarze   66:        int              val;
1.180     schwarze   67:        struct roffreg  *next;
                     68: };
                     69:
1.295     schwarze   70: /*
                     71:  * Association of request and macro names with token IDs.
                     72:  */
                     73: struct roffreq {
                     74:        enum roff_tok    tok;
                     75:        char             name[];
                     76: };
                     77:
1.67      kristaps   78: struct roff {
1.128     kristaps   79:        struct mparse   *parse; /* parse point */
1.296   ! schwarze   80:        struct roff_man *man; /* mdoc or man parser */
1.67      kristaps   81:        struct roffnode *last; /* leaf of stack */
1.223     schwarze   82:        int             *rstack; /* stack of inverted `ie' values */
1.295     schwarze   83:        struct ohash    *reqtab; /* request lookup table */
1.180     schwarze   84:        struct roffreg  *regtab; /* number registers */
1.166     kristaps   85:        struct roffkv   *strtab; /* user-defined strings & macros */
1.167     kristaps   86:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     87:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.106     kristaps   88:        const char      *current_string; /* value of last called user macro */
1.118     kristaps   89:        struct tbl_node *first_tbl; /* first table parsed */
                     90:        struct tbl_node *last_tbl; /* last table parsed */
                     91:        struct tbl_node *tbl; /* current table being parsed */
1.125     kristaps   92:        struct eqn_node *last_eqn; /* last equation parsed */
                     93:        struct eqn_node *first_eqn; /* first equation parsed */
                     94:        struct eqn_node *eqn; /* current equation being parsed */
1.230     schwarze   95:        int              eqn_inline; /* current equation is inline */
1.223     schwarze   96:        int              options; /* parse options */
                     97:        int              rstacksz; /* current size limit of rstack */
                     98:        int              rstackpos; /* position in rstack */
1.227     schwarze   99:        int              format; /* current file in mdoc or man format */
1.273     schwarze  100:        int              argc; /* number of args of the last macro */
1.223     schwarze  101:        char             control; /* control character */
1.79      kristaps  102: };
                    103:
1.67      kristaps  104: struct roffnode {
1.294     schwarze  105:        enum roff_tok    tok; /* type of node */
1.67      kristaps  106:        struct roffnode *parent; /* up one in stack */
                    107:        int              line; /* parse line */
                    108:        int              col; /* parse col */
1.106     kristaps  109:        char            *name; /* node name, e.g. macro name */
1.79      kristaps  110:        char            *end; /* end-rules: custom token */
                    111:        int              endspan; /* end-rules: next-line or infty */
1.198     schwarze  112:        int              rule; /* current evaluation rule */
1.67      kristaps  113: };
                    114:
                    115: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.294     schwarze  116:                         enum roff_tok tok, /* tok of macro */ \
1.238     schwarze  117:                         struct buf *buf, /* input buffer */ \
1.67      kristaps  118:                         int ln, /* parse line */ \
1.75      kristaps  119:                         int ppos, /* original pos in buffer */ \
                    120:                         int pos, /* current pos in buffer */ \
1.74      kristaps  121:                         int *offs /* reset offset of buffer data */
1.67      kristaps  122:
                    123: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    124:
                    125: struct roffmac {
1.79      kristaps  126:        roffproc         proc; /* process new macro */
                    127:        roffproc         text; /* process as child text of macro */
                    128:        roffproc         sub; /* process as child of macro */
                    129:        int              flags;
                    130: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.67      kristaps  131: };
                    132:
1.141     kristaps  133: struct predef {
                    134:        const char      *name; /* predefined input name */
                    135:        const char      *str; /* replacement symbol */
                    136: };
                    137:
                    138: #define        PREDEF(__name, __str) \
                    139:        { (__name), (__str) },
                    140:
1.266     schwarze  141: /* --- function prototypes ------------------------------------------------ */
                    142:
1.155     kristaps  143: static void             roffnode_cleanscope(struct roff *);
                    144: static void             roffnode_pop(struct roff *);
1.294     schwarze  145: static void             roffnode_push(struct roff *, enum roff_tok,
1.155     kristaps  146:                                const char *, int, int);
1.80      kristaps  147: static enum rofferr     roff_block(ROFF_ARGS);
                    148: static enum rofferr     roff_block_text(ROFF_ARGS);
                    149: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.296   ! schwarze  150: static enum rofferr     roff_br(ROFF_ARGS);
1.80      kristaps  151: static enum rofferr     roff_cblock(ROFF_ARGS);
1.174     kristaps  152: static enum rofferr     roff_cc(ROFF_ARGS);
1.195     schwarze  153: static void             roff_ccond(struct roff *, int, int);
1.82      kristaps  154: static enum rofferr     roff_cond(ROFF_ARGS);
                    155: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    156: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.92      schwarze  157: static enum rofferr     roff_ds(ROFF_ARGS);
1.238     schwarze  158: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.271     schwarze  159: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.234     kristaps  160: static int              roff_evalnum(struct roff *, int,
                    161:                                const char *, int *, int *, int);
                    162: static int              roff_evalpar(struct roff *, int,
1.261     schwarze  163:                                const char *, int *, int *, int);
1.198     schwarze  164: static int              roff_evalstrcond(const char *, int *);
1.155     kristaps  165: static void             roff_free1(struct roff *);
1.180     schwarze  166: static void             roff_freereg(struct roffreg *);
1.167     kristaps  167: static void             roff_freestr(struct roffkv *);
1.212     schwarze  168: static size_t           roff_getname(struct roff *, char **, int, int);
1.261     schwarze  169: static int              roff_getnum(const char *, int *, int *, int);
1.184     schwarze  170: static int              roff_getop(const char *, int *, char *);
1.181     schwarze  171: static int              roff_getregn(const struct roff *,
                    172:                                const char *, size_t);
1.273     schwarze  173: static int              roff_getregro(const struct roff *,
                    174:                                const char *name);
1.207     schwarze  175: static const char      *roff_getstrn(const struct roff *,
1.94      kristaps  176:                                const char *, size_t);
1.271     schwarze  177: static int              roff_hasregn(const struct roff *,
                    178:                                const char *, size_t);
1.251     schwarze  179: static enum rofferr     roff_insec(ROFF_ARGS);
1.178     schwarze  180: static enum rofferr     roff_it(ROFF_ARGS);
1.103     kristaps  181: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.265     schwarze  182: static void             roff_man_alloc1(struct roff_man *);
                    183: static void             roff_man_free1(struct roff_man *);
1.89      kristaps  184: static enum rofferr     roff_nr(ROFF_ARGS);
1.294     schwarze  185: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.214     schwarze  186:                                int, int);
1.238     schwarze  187: static enum rofferr     roff_parsetext(struct buf *, int, int *);
                    188: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.122     schwarze  189: static enum rofferr     roff_rm(ROFF_ARGS);
1.203     schwarze  190: static enum rofferr     roff_rr(ROFF_ARGS);
1.94      kristaps  191: static void             roff_setstr(struct roff *,
1.106     kristaps  192:                                const char *, const char *, int);
1.207     schwarze  193: static void             roff_setstrn(struct roffkv **, const char *,
1.164     kristaps  194:                                size_t, const char *, size_t, int);
1.105     kristaps  195: static enum rofferr     roff_so(ROFF_ARGS);
1.164     kristaps  196: static enum rofferr     roff_tr(ROFF_ARGS);
1.175     schwarze  197: static enum rofferr     roff_Dd(ROFF_ARGS);
                    198: static enum rofferr     roff_TH(ROFF_ARGS);
1.109     kristaps  199: static enum rofferr     roff_TE(ROFF_ARGS);
                    200: static enum rofferr     roff_TS(ROFF_ARGS);
1.125     kristaps  201: static enum rofferr     roff_EQ(ROFF_ARGS);
                    202: static enum rofferr     roff_EN(ROFF_ARGS);
1.112     kristaps  203: static enum rofferr     roff_T_(ROFF_ARGS);
1.251     schwarze  204: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.106     kristaps  205: static enum rofferr     roff_userdef(ROFF_ARGS);
1.67      kristaps  206:
1.266     schwarze  207: /* --- constant data ------------------------------------------------------ */
                    208:
1.261     schwarze  209: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    210: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    211:
1.294     schwarze  212: const char *__roff_name[MAN_MAX + 1] = {
1.296   ! schwarze  213:        "br",           NULL,
1.294     schwarze  214:        "ab",           "ad",           "af",           "aln",
                    215:        "als",          "am",           "am1",          "ami",
                    216:        "ami1",         "as",           "as1",          "asciify",
                    217:        "backtrace",    "bd",           "bleedat",      "blm",
                    218:         "box",         "boxa",         "bp",           "BP",
                    219:        "break",        "breakchar",    "brnl",         "brp",
                    220:        "brpnl",        "c2",           "cc",           "ce",
                    221:        "cf",           "cflags",       "ch",           "char",
                    222:        "chop",         "class",        "close",        "CL",
                    223:        "color",        "composite",    "continue",     "cp",
                    224:        "cropat",       "cs",           "cu",           "da",
                    225:        "dch",          "Dd",           "de",           "de1",
                    226:        "defcolor",     "dei",          "dei1",         "device",
                    227:        "devicem",      "di",           "do",           "ds",
                    228:        "ds1",          "dwh",          "dt",           "ec",
                    229:        "ecr",          "ecs",          "el",           "em",
                    230:        "EN",           "eo",           "EP",           "EQ",
                    231:        "errprint",     "ev",           "evc",          "ex",
                    232:        "fallback",     "fam",          "fc",           "fchar",
                    233:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    234:        "fl",           "flig",         "fp",           "fps",
                    235:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    236:        "fzoom",        "gcolor",       "hc",           "hcode",
                    237:        "hidechar",     "hla",          "hlm",          "hpf",
                    238:        "hpfa",         "hpfcode",      "hw",           "hy",
                    239:        "hylang",       "hylen",        "hym",          "hypp",
                    240:        "hys",          "ie",           "if",           "ig",
                    241:        "index",        "it",           "itc",          "IX",
                    242:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    243:        "lc",           "lc_ctype",     "lds",          "length",
                    244:        "letadj",       "lf",           "lg",           "lhang",
                    245:        "linetabs",     "lnr",          "lnrf",         "lpfx",
                    246:        "ls",           "lsm",          "lt",           "mc",
                    247:        "mediasize",    "minss",        "mk",           "mso",
                    248:        "na",           "ne",           "nh",           "nhychar",
                    249:        "nm",           "nn",           "nop",          "nr",
                    250:        "nrf",          "nroff",        "ns",           "nx",
                    251:        "open",         "opena",        "os",           "output",
                    252:        "padj",         "papersize",    "pc",           "pev",
                    253:        "pi",           "PI",           "pl",           "pm",
                    254:        "pn",           "pnr",          "po",           "ps",
                    255:        "psbb",         "pshape",       "pso",          "ptr",
                    256:        "pvs",          "rchar",        "rd",           "recursionlimit",
                    257:        "return",       "rfschar",      "rhang",        "rj",
                    258:        "rm",           "rn",           "rnn",          "rr",
                    259:        "rs",           "rt",           "schar",        "sentchar",
                    260:        "shc",          "shift",        "sizes",        "so",
                    261:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    262:        "sty",          "substring",    "sv",           "sy",
                    263:        "T&",           "ta",           "tc",           "TE",
                    264:        "TH",           "ti",           "tkf",          "tl",
                    265:        "tm",           "tm1",          "tmc",          "tr",
                    266:        "track",        "transchar",    "trf",          "trimat",
                    267:        "trin",         "trnt",         "troff",        "TS",
                    268:        "uf",           "ul",           "unformat",     "unwatch",
                    269:        "unwatchn",     "vpt",          "vs",           "warn",
                    270:        "warnscale",    "watch",        "watchlength",  "watchn",
                    271:        "wh",           "while",        "write",        "writec",
                    272:        "writem",       "xflag",        ".",            NULL,
                    273:        "text",
                    274:        "Dd",           "Dt",           "Os",           "Sh",
                    275:        "Ss",           "Pp",           "D1",           "Dl",
                    276:        "Bd",           "Ed",           "Bl",           "El",
                    277:        "It",           "Ad",           "An",           "Ap",
                    278:        "Ar",           "Cd",           "Cm",           "Dv",
                    279:        "Er",           "Ev",           "Ex",           "Fa",
                    280:        "Fd",           "Fl",           "Fn",           "Ft",
                    281:        "Ic",           "In",           "Li",           "Nd",
                    282:        "Nm",           "Op",           "Ot",           "Pa",
                    283:        "Rv",           "St",           "Va",           "Vt",
                    284:        "Xr",           "%A",           "%B",           "%D",
                    285:        "%I",           "%J",           "%N",           "%O",
                    286:        "%P",           "%R",           "%T",           "%V",
                    287:        "Ac",           "Ao",           "Aq",           "At",
                    288:        "Bc",           "Bf",           "Bo",           "Bq",
                    289:        "Bsx",          "Bx",           "Db",           "Dc",
                    290:        "Do",           "Dq",           "Ec",           "Ef",
                    291:        "Em",           "Eo",           "Fx",           "Ms",
                    292:        "No",           "Ns",           "Nx",           "Ox",
                    293:        "Pc",           "Pf",           "Po",           "Pq",
                    294:        "Qc",           "Ql",           "Qo",           "Qq",
                    295:        "Re",           "Rs",           "Sc",           "So",
                    296:        "Sq",           "Sm",           "Sx",           "Sy",
                    297:        "Tn",           "Ux",           "Xc",           "Xo",
                    298:        "Fo",           "Fc",           "Oo",           "Oc",
                    299:        "Bk",           "Ek",           "Bt",           "Hf",
                    300:        "Fr",           "Ud",           "Lb",           "Lp",
                    301:        "Lk",           "Mt",           "Brq",          "Bro",
                    302:        "Brc",          "%C",           "Es",           "En",
1.296   ! schwarze  303:        "Dx",           "%Q",           "sp",
1.294     schwarze  304:        "%U",           "Ta",           "ll",           NULL,
                    305:        "TH",           "SH",           "SS",           "TP",
                    306:        "LP",           "PP",           "P",            "IP",
                    307:        "HP",           "SM",           "SB",           "BI",
                    308:        "IB",           "BR",           "RB",           "R",
                    309:        "B",            "I",            "IR",           "RI",
1.296   ! schwarze  310:        "sp",           "nf",           "fi",
1.294     schwarze  311:        "RE",           "RS",           "DT",           "UC",
                    312:        "PD",           "AT",           "in",           "ft",
                    313:        "OP",           "EX",           "EE",           "UR",
                    314:        "UE",           "ll",           NULL
                    315: };
                    316: const  char *const *roff_name = __roff_name;
                    317:
                    318: static struct roffmac   roffs[TOKEN_NONE] = {
1.296   ! schwarze  319:        { roff_br, NULL, NULL, 0 },  /* br */
        !           320:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.295     schwarze  321:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    322:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    323:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    324:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
                    325:        { roff_unsupp, NULL, NULL, 0 },  /* als */
                    326:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    327:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    328:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    329:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    330:        { roff_ds, NULL, NULL, 0 },  /* as */
                    331:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    332:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    333:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    334:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    335:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    336:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    337:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    338:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    339:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    340:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    341:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    342:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    343:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.296   ! schwarze  344:        { roff_br, NULL, NULL, 0 },  /* brp */
1.295     schwarze  345:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    346:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    347:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    348:        { roff_line_ignore, NULL, NULL, 0 },  /* ce */
                    349:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    350:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    351:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    352:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    353:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    354:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    355:        { roff_insec, NULL, NULL, 0 },  /* close */
                    356:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    357:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    358:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    359:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    360:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    361:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    362:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    363:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    364:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    365:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    366:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    367:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    368:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    369:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    370:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    371:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    372:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    373:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    374:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    375:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    376:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    377:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    378:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    379:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
                    380:        { roff_unsupp, NULL, NULL, 0 },  /* ec */
                    381:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    382:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    383:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    384:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    385:        { roff_EN, NULL, NULL, 0 },  /* EN */
                    386:        { roff_unsupp, NULL, NULL, 0 },  /* eo */
                    387:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    388:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    389:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    390:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    391:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    392:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    393:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    394:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    395:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    396:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    397:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    398:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    399:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    400:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    401:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    402:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    403:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    404:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    405:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    406:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    407:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    408:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    409:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    410:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    411:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    412:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    413:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    414:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    415:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    416:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    417:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    419:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    421:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    422:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    425:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    426:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    427:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    428:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    429:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    430:        { roff_it, NULL, NULL, 0 },  /* it */
                    431:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    432:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    437:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    438:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    439:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    440:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    442:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    444:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    445:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    446:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    447:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    448:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    449:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    450:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    451:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    452:        { roff_line_ignore, NULL, NULL, 0 },  /* mc */
                    453:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    454:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    455:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    456:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    457:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    458:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    461:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    462:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    463:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    464:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    465:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    466:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    467:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    468:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    469:        { roff_insec, NULL, NULL, 0 },  /* open */
                    470:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    471:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    472:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    474:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    475:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    477:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    478:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* po */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    485:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    486:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    487:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    488:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    489:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    490:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    491:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    492:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    493:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    494:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    495:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    496:        { roff_line_ignore, NULL, NULL, 0 },  /* rj */
                    497:        { roff_rm, NULL, NULL, 0 },  /* rm */
                    498:        { roff_unsupp, NULL, NULL, 0 },  /* rn */
                    499:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    500:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    501:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    503:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    504:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    505:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    506:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    508:        { roff_so, NULL, NULL, 0 },  /* so */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    511:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    512:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    513:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    514:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    516:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    517:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    518:        { roff_unsupp, NULL, NULL, 0 },  /* ta */
                    519:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    520:        { roff_TE, NULL, NULL, 0 },  /* TE */
                    521:        { roff_TH, NULL, NULL, 0 },  /* TH */
                    522:        { roff_unsupp, NULL, NULL, 0 },  /* ti */
                    523:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    524:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    526:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    528:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    529:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    530:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    531:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    533:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    534:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    535:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    536:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    537:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    538:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    539:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    540:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    541:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    542:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    543:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    549:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    550:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    551:        { roff_insec, NULL, NULL, 0 },  /* write */
                    552:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    553:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    554:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    555:        { roff_cblock, NULL, NULL, 0 },  /* . */
                    556:        { roff_userdef, NULL, NULL, 0 }
1.67      kristaps  557: };
                    558:
1.200     schwarze  559: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175     schwarze  560: const  char *const __mdoc_reserved[] = {
                    561:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    562:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    563:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    564:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200     schwarze  565:        "Dt", "Dv", "Dx", "D1",
                    566:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    567:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175     schwarze  568:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200     schwarze  569:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    570:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175     schwarze  571:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200     schwarze  572:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    573:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    574:        "Sc", "Sh", "Sm", "So", "Sq",
1.175     schwarze  575:        "Ss", "St", "Sx", "Sy",
                    576:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200     schwarze  577:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175     schwarze  578:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    579:        NULL
                    580: };
                    581:
1.200     schwarze  582: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175     schwarze  583: const  char *const __man_reserved[] = {
1.200     schwarze  584:        "AT", "B", "BI", "BR", "DT",
                    585:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    586:        "LP", "OP", "P", "PD", "PP",
                    587:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    588:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175     schwarze  589:        NULL
                    590: };
                    591:
1.141     kristaps  592: /* Array of injected predefined strings. */
                    593: #define        PREDEFS_MAX      38
                    594: static const struct predef predefs[PREDEFS_MAX] = {
                    595: #include "predefs.in"
                    596: };
                    597:
1.178     schwarze  598: static int      roffit_lines;  /* number of lines to delay */
                    599: static char    *roffit_macro;  /* nil-terminated macro line */
                    600:
1.207     schwarze  601:
1.266     schwarze  602: /* --- request table ------------------------------------------------------ */
                    603:
1.295     schwarze  604: struct ohash *
                    605: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85      kristaps  606: {
1.295     schwarze  607:        struct ohash    *htab;
                    608:        struct roffreq  *req;
                    609:        enum roff_tok    tok;
                    610:        size_t           sz;
                    611:        unsigned int     slot;
1.85      kristaps  612:
1.295     schwarze  613:        htab = mandoc_malloc(sizeof(*htab));
                    614:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85      kristaps  615:
1.295     schwarze  616:        for (tok = mintok; tok < maxtok; tok++) {
1.296   ! schwarze  617:                if (roff_name[tok] == NULL)
        !           618:                        continue;
1.295     schwarze  619:                sz = strlen(roff_name[tok]);
                    620:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    621:                req->tok = tok;
                    622:                memcpy(req->name, roff_name[tok], sz + 1);
                    623:                slot = ohash_qlookup(htab, req->name);
                    624:                ohash_insert(htab, slot, req);
1.85      kristaps  625:        }
1.295     schwarze  626:        return htab;
1.85      kristaps  627: }
                    628:
1.295     schwarze  629: void
                    630: roffhash_free(struct ohash *htab)
1.67      kristaps  631: {
1.295     schwarze  632:        struct roffreq  *req;
                    633:        unsigned int     slot;
1.67      kristaps  634:
1.295     schwarze  635:        if (htab == NULL)
                    636:                return;
                    637:        for (req = ohash_first(htab, &slot); req != NULL;
                    638:             req = ohash_next(htab, &slot))
                    639:                free(req);
                    640:        ohash_delete(htab);
                    641:        free(htab);
                    642: }
                    643:
                    644: enum roff_tok
                    645: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    646: {
                    647:        struct roffreq  *req;
                    648:        const char      *end;
                    649:
                    650:        if (sz) {
                    651:                end = name + sz;
                    652:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    653:        } else
                    654:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    655:        return req == NULL ? TOKEN_NONE : req->tok;
1.67      kristaps  656: }
                    657:
1.266     schwarze  658: /* --- stack of request blocks -------------------------------------------- */
                    659:
1.67      kristaps  660: /*
                    661:  * Pop the current node off of the stack of roff instructions currently
                    662:  * pending.
                    663:  */
                    664: static void
                    665: roffnode_pop(struct roff *r)
                    666: {
                    667:        struct roffnode *p;
                    668:
1.75      kristaps  669:        assert(r->last);
1.207     schwarze  670:        p = r->last;
1.82      kristaps  671:
1.75      kristaps  672:        r->last = r->last->parent;
1.106     kristaps  673:        free(p->name);
                    674:        free(p->end);
1.67      kristaps  675:        free(p);
                    676: }
                    677:
                    678: /*
                    679:  * Push a roff node onto the instruction stack.  This must later be
                    680:  * removed with roffnode_pop().
                    681:  */
1.98      schwarze  682: static void
1.294     schwarze  683: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106     kristaps  684:                int line, int col)
1.67      kristaps  685: {
                    686:        struct roffnode *p;
                    687:
1.98      schwarze  688:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.67      kristaps  689:        p->tok = tok;
1.106     kristaps  690:        if (name)
                    691:                p->name = mandoc_strdup(name);
1.67      kristaps  692:        p->parent = r->last;
                    693:        p->line = line;
                    694:        p->col = col;
1.198     schwarze  695:        p->rule = p->parent ? p->parent->rule : 0;
1.67      kristaps  696:
                    697:        r->last = p;
                    698: }
                    699:
1.266     schwarze  700: /* --- roff parser state data management ---------------------------------- */
                    701:
1.67      kristaps  702: static void
                    703: roff_free1(struct roff *r)
                    704: {
1.176     schwarze  705:        struct tbl_node *tbl;
1.125     kristaps  706:        struct eqn_node *e;
1.167     kristaps  707:        int              i;
1.67      kristaps  708:
1.176     schwarze  709:        while (NULL != (tbl = r->first_tbl)) {
                    710:                r->first_tbl = tbl->next;
                    711:                tbl_free(tbl);
1.109     kristaps  712:        }
1.113     kristaps  713:        r->first_tbl = r->last_tbl = r->tbl = NULL;
                    714:
1.125     kristaps  715:        while (NULL != (e = r->first_eqn)) {
                    716:                r->first_eqn = e->next;
                    717:                eqn_free(e);
                    718:        }
                    719:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    720:
1.67      kristaps  721:        while (r->last)
                    722:                roffnode_pop(r);
1.109     kristaps  723:
1.223     schwarze  724:        free (r->rstack);
                    725:        r->rstack = NULL;
                    726:        r->rstacksz = 0;
                    727:        r->rstackpos = -1;
                    728:
                    729:        roff_freereg(r->regtab);
                    730:        r->regtab = NULL;
                    731:
1.167     kristaps  732:        roff_freestr(r->strtab);
                    733:        roff_freestr(r->xmbtab);
                    734:        r->strtab = r->xmbtab = NULL;
                    735:
                    736:        if (r->xtab)
                    737:                for (i = 0; i < 128; i++)
                    738:                        free(r->xtab[i].p);
                    739:        free(r->xtab);
                    740:        r->xtab = NULL;
1.67      kristaps  741: }
                    742:
                    743: void
                    744: roff_reset(struct roff *r)
                    745: {
                    746:        roff_free1(r);
1.227     schwarze  747:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174     kristaps  748:        r->control = 0;
1.67      kristaps  749: }
                    750:
                    751: void
                    752: roff_free(struct roff *r)
                    753: {
                    754:        roff_free1(r);
1.295     schwarze  755:        roffhash_free(r->reqtab);
1.67      kristaps  756:        free(r);
                    757: }
                    758:
                    759: struct roff *
1.279     schwarze  760: roff_alloc(struct mparse *parse, int options)
1.67      kristaps  761: {
                    762:        struct roff     *r;
                    763:
1.98      schwarze  764:        r = mandoc_calloc(1, sizeof(struct roff));
1.128     kristaps  765:        r->parse = parse;
1.295     schwarze  766:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.199     schwarze  767:        r->options = options;
1.227     schwarze  768:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82      kristaps  769:        r->rstackpos = -1;
1.277     schwarze  770:        return r;
1.265     schwarze  771: }
                    772:
1.266     schwarze  773: /* --- syntax tree state data management ---------------------------------- */
                    774:
1.265     schwarze  775: static void
                    776: roff_man_free1(struct roff_man *man)
                    777: {
                    778:
1.266     schwarze  779:        if (man->first != NULL)
                    780:                roff_node_delete(man, man->first);
1.265     schwarze  781:        free(man->meta.msec);
                    782:        free(man->meta.vol);
                    783:        free(man->meta.os);
                    784:        free(man->meta.arch);
                    785:        free(man->meta.title);
                    786:        free(man->meta.name);
                    787:        free(man->meta.date);
                    788: }
                    789:
                    790: static void
                    791: roff_man_alloc1(struct roff_man *man)
                    792: {
                    793:
                    794:        memset(&man->meta, 0, sizeof(man->meta));
                    795:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    796:        man->first->type = ROFFT_ROOT;
                    797:        man->last = man->first;
                    798:        man->last_es = NULL;
                    799:        man->flags = 0;
                    800:        man->macroset = MACROSET_NONE;
                    801:        man->lastsec = man->lastnamed = SEC_NONE;
                    802:        man->next = ROFF_NEXT_CHILD;
                    803: }
                    804:
                    805: void
                    806: roff_man_reset(struct roff_man *man)
                    807: {
                    808:
                    809:        roff_man_free1(man);
                    810:        roff_man_alloc1(man);
                    811: }
                    812:
                    813: void
                    814: roff_man_free(struct roff_man *man)
                    815: {
                    816:
                    817:        roff_man_free1(man);
                    818:        free(man);
                    819: }
                    820:
                    821: struct roff_man *
                    822: roff_man_alloc(struct roff *roff, struct mparse *parse,
                    823:        const char *defos, int quick)
                    824: {
                    825:        struct roff_man *man;
                    826:
                    827:        man = mandoc_calloc(1, sizeof(*man));
                    828:        man->parse = parse;
                    829:        man->roff = roff;
                    830:        man->defos = defos;
                    831:        man->quick = quick;
                    832:        roff_man_alloc1(man);
1.296   ! schwarze  833:        roff->man = man;
1.277     schwarze  834:        return man;
1.67      kristaps  835: }
                    836:
1.266     schwarze  837: /* --- syntax tree handling ----------------------------------------------- */
                    838:
                    839: struct roff_node *
                    840: roff_node_alloc(struct roff_man *man, int line, int pos,
                    841:        enum roff_type type, int tok)
                    842: {
                    843:        struct roff_node        *n;
                    844:
                    845:        n = mandoc_calloc(1, sizeof(*n));
                    846:        n->line = line;
                    847:        n->pos = pos;
                    848:        n->tok = tok;
                    849:        n->type = type;
                    850:        n->sec = man->lastsec;
                    851:
                    852:        if (man->flags & MDOC_SYNOPSIS)
1.285     schwarze  853:                n->flags |= NODE_SYNPRETTY;
1.266     schwarze  854:        else
1.285     schwarze  855:                n->flags &= ~NODE_SYNPRETTY;
1.266     schwarze  856:        if (man->flags & MDOC_NEWLINE)
1.285     schwarze  857:                n->flags |= NODE_LINE;
1.266     schwarze  858:        man->flags &= ~MDOC_NEWLINE;
                    859:
1.277     schwarze  860:        return n;
1.266     schwarze  861: }
                    862:
                    863: void
                    864: roff_node_append(struct roff_man *man, struct roff_node *n)
                    865: {
                    866:
                    867:        switch (man->next) {
                    868:        case ROFF_NEXT_SIBLING:
1.281     schwarze  869:                if (man->last->next != NULL) {
                    870:                        n->next = man->last->next;
                    871:                        man->last->next->prev = n;
                    872:                } else
                    873:                        man->last->parent->last = n;
1.266     schwarze  874:                man->last->next = n;
                    875:                n->prev = man->last;
                    876:                n->parent = man->last->parent;
                    877:                break;
                    878:        case ROFF_NEXT_CHILD:
1.287     schwarze  879:                if (man->last->child != NULL) {
                    880:                        n->next = man->last->child;
                    881:                        man->last->child->prev = n;
                    882:                } else
                    883:                        man->last->last = n;
1.266     schwarze  884:                man->last->child = n;
                    885:                n->parent = man->last;
                    886:                break;
                    887:        default:
                    888:                abort();
                    889:        }
1.282     schwarze  890:        man->last = n;
1.266     schwarze  891:
                    892:        switch (n->type) {
                    893:        case ROFFT_HEAD:
                    894:                n->parent->head = n;
                    895:                break;
                    896:        case ROFFT_BODY:
1.282     schwarze  897:                if (n->end != ENDBODY_NOT)
                    898:                        return;
1.266     schwarze  899:                n->parent->body = n;
                    900:                break;
                    901:        case ROFFT_TAIL:
                    902:                n->parent->tail = n;
                    903:                break;
                    904:        default:
1.282     schwarze  905:                return;
1.266     schwarze  906:        }
1.282     schwarze  907:
                    908:        /*
                    909:         * Copy over the normalised-data pointer of our parent.  Not
                    910:         * everybody has one, but copying a null pointer is fine.
                    911:         */
                    912:
                    913:        n->norm = n->parent->norm;
                    914:        assert(n->parent->type == ROFFT_BLOCK);
1.266     schwarze  915: }
                    916:
1.267     schwarze  917: void
                    918: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    919: {
                    920:        struct roff_node        *n;
                    921:
                    922:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    923:        n->string = roff_strdup(man->roff, word);
                    924:        roff_node_append(man, n);
1.286     schwarze  925:        n->flags |= NODE_VALID | NODE_ENDED;
1.267     schwarze  926:        man->next = ROFF_NEXT_SIBLING;
                    927: }
                    928:
                    929: void
                    930: roff_word_append(struct roff_man *man, const char *word)
                    931: {
                    932:        struct roff_node        *n;
                    933:        char                    *addstr, *newstr;
                    934:
                    935:        n = man->last;
                    936:        addstr = roff_strdup(man->roff, word);
                    937:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    938:        free(addstr);
                    939:        free(n->string);
                    940:        n->string = newstr;
                    941:        man->next = ROFF_NEXT_SIBLING;
1.268     schwarze  942: }
                    943:
                    944: void
                    945: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    946: {
                    947:        struct roff_node        *n;
                    948:
                    949:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    950:        roff_node_append(man, n);
                    951:        man->next = ROFF_NEXT_CHILD;
                    952: }
                    953:
                    954: struct roff_node *
                    955: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    956: {
                    957:        struct roff_node        *n;
                    958:
                    959:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    960:        roff_node_append(man, n);
                    961:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze  962:        return n;
1.267     schwarze  963: }
                    964:
1.266     schwarze  965: struct roff_node *
                    966: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    967: {
                    968:        struct roff_node        *n;
                    969:
                    970:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    971:        roff_node_append(man, n);
                    972:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze  973:        return n;
1.266     schwarze  974: }
                    975:
                    976: struct roff_node *
                    977: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    978: {
                    979:        struct roff_node        *n;
                    980:
                    981:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    982:        roff_node_append(man, n);
                    983:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze  984:        return n;
1.267     schwarze  985: }
                    986:
                    987: void
                    988: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
                    989: {
                    990:        struct roff_node        *n;
                    991:
                    992:        n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
                    993:        n->eqn = eqn;
                    994:        if (eqn->ln > man->last->line)
1.285     schwarze  995:                n->flags |= NODE_LINE;
1.267     schwarze  996:        roff_node_append(man, n);
                    997:        man->next = ROFF_NEXT_SIBLING;
                    998: }
                    999:
                   1000: void
                   1001: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
                   1002: {
                   1003:        struct roff_node        *n;
                   1004:
                   1005:        if (man->macroset == MACROSET_MAN)
                   1006:                man_breakscope(man, TOKEN_NONE);
                   1007:        n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                   1008:        n->span = tbl;
                   1009:        roff_node_append(man, n);
1.286     schwarze 1010:        n->flags |= NODE_VALID | NODE_ENDED;
1.267     schwarze 1011:        man->next = ROFF_NEXT_SIBLING;
1.266     schwarze 1012: }
                   1013:
                   1014: void
                   1015: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1016: {
                   1017:
                   1018:        /* Adjust siblings. */
                   1019:
                   1020:        if (n->prev)
                   1021:                n->prev->next = n->next;
                   1022:        if (n->next)
                   1023:                n->next->prev = n->prev;
                   1024:
                   1025:        /* Adjust parent. */
                   1026:
                   1027:        if (n->parent != NULL) {
                   1028:                if (n->parent->child == n)
                   1029:                        n->parent->child = n->next;
                   1030:                if (n->parent->last == n)
                   1031:                        n->parent->last = n->prev;
                   1032:        }
                   1033:
                   1034:        /* Adjust parse point. */
                   1035:
                   1036:        if (man == NULL)
                   1037:                return;
                   1038:        if (man->last == n) {
                   1039:                if (n->prev == NULL) {
                   1040:                        man->last = n->parent;
                   1041:                        man->next = ROFF_NEXT_CHILD;
                   1042:                } else {
                   1043:                        man->last = n->prev;
                   1044:                        man->next = ROFF_NEXT_SIBLING;
                   1045:                }
                   1046:        }
                   1047:        if (man->first == n)
                   1048:                man->first = NULL;
                   1049: }
                   1050:
                   1051: void
                   1052: roff_node_free(struct roff_node *n)
                   1053: {
                   1054:
                   1055:        if (n->args != NULL)
                   1056:                mdoc_argv_free(n->args);
                   1057:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1058:                free(n->norm);
                   1059:        free(n->string);
                   1060:        free(n);
                   1061: }
                   1062:
                   1063: void
                   1064: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1065: {
                   1066:
                   1067:        while (n->child != NULL)
                   1068:                roff_node_delete(man, n->child);
                   1069:        roff_node_unlink(man, n);
                   1070:        roff_node_free(n);
1.269     schwarze 1071: }
                   1072:
                   1073: void
                   1074: deroff(char **dest, const struct roff_node *n)
                   1075: {
                   1076:        char    *cp;
                   1077:        size_t   sz;
                   1078:
                   1079:        if (n->type != ROFFT_TEXT) {
                   1080:                for (n = n->child; n != NULL; n = n->next)
                   1081:                        deroff(dest, n);
                   1082:                return;
                   1083:        }
                   1084:
1.288     schwarze 1085:        /* Skip leading whitespace. */
1.269     schwarze 1086:
1.288     schwarze 1087:        for (cp = n->string; *cp != '\0'; cp++) {
1.289     schwarze 1088:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1089:                    strchr(" %&0^|~", cp[1]) != NULL)
1.269     schwarze 1090:                        cp++;
1.288     schwarze 1091:                else if ( ! isspace((unsigned char)*cp))
1.269     schwarze 1092:                        break;
                   1093:        }
                   1094:
1.289     schwarze 1095:        /* Skip trailing backslash. */
                   1096:
                   1097:        sz = strlen(cp);
1.290     schwarze 1098:        if (sz > 0 && cp[sz - 1] == '\\')
1.289     schwarze 1099:                sz--;
                   1100:
1.269     schwarze 1101:        /* Skip trailing whitespace. */
                   1102:
1.289     schwarze 1103:        for (; sz; sz--)
1.269     schwarze 1104:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1105:                        break;
                   1106:
                   1107:        /* Skip empty strings. */
                   1108:
                   1109:        if (sz == 0)
                   1110:                return;
                   1111:
                   1112:        if (*dest == NULL) {
                   1113:                *dest = mandoc_strndup(cp, sz);
                   1114:                return;
                   1115:        }
                   1116:
                   1117:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1118:        free(*dest);
                   1119:        *dest = cp;
1.266     schwarze 1120: }
                   1121:
                   1122: /* --- main functions of the roff parser ---------------------------------- */
                   1123:
1.94      kristaps 1124: /*
1.206     schwarze 1125:  * In the current line, expand escape sequences that tend to get
                   1126:  * used in numerical expressions and conditional requests.
                   1127:  * Also check the syntax of the remaining escape sequences.
1.154     kristaps 1128:  */
1.172     schwarze 1129: static enum rofferr
1.238     schwarze 1130: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94      kristaps 1131: {
1.208     schwarze 1132:        char             ubuf[24]; /* buffer to print the number */
1.205     schwarze 1133:        const char      *start; /* start of the string to process */
1.209     schwarze 1134:        char            *stesc; /* start of an escape sequence ('\\') */
1.108     schwarze 1135:        const char      *stnam; /* start of the name, after "[(*" */
                   1136:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1137:        const char      *res;   /* the string to be substituted */
1.238     schwarze 1138:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.181     schwarze 1139:        size_t           maxl;  /* expected length of the escape name */
                   1140:        size_t           naml;  /* actual length of the escape name */
1.237     schwarze 1141:        enum mandoc_esc  esc;   /* type of the escape sequence */
                   1142:        int              inaml; /* length returned from mandoc_escape() */
1.181     schwarze 1143:        int              expand_count;  /* to avoid infinite loops */
1.206     schwarze 1144:        int              npos;  /* position in numeric expression */
1.218     schwarze 1145:        int              arg_complete; /* argument not interrupted by eol */
1.206     schwarze 1146:        char             term;  /* character terminating the escape */
1.94      kristaps 1147:
1.170     schwarze 1148:        expand_count = 0;
1.238     schwarze 1149:        start = buf->buf + pos;
1.205     schwarze 1150:        stesc = strchr(start, '\0') - 1;
                   1151:        while (stesc-- > start) {
1.170     schwarze 1152:
1.205     schwarze 1153:                /* Search backwards for the next backslash. */
                   1154:
1.238     schwarze 1155:                if (*stesc != '\\')
1.205     schwarze 1156:                        continue;
                   1157:
                   1158:                /* If it is escaped, skip it. */
                   1159:
                   1160:                for (cp = stesc - 1; cp >= start; cp--)
1.238     schwarze 1161:                        if (*cp != '\\')
1.205     schwarze 1162:                                break;
                   1163:
1.238     schwarze 1164:                if ((stesc - cp) % 2 == 0) {
1.209     schwarze 1165:                        stesc = (char *)cp;
1.205     schwarze 1166:                        continue;
                   1167:                }
1.108     schwarze 1168:
1.206     schwarze 1169:                /* Decide whether to expand or to check only. */
1.108     schwarze 1170:
1.206     schwarze 1171:                term = '\0';
1.205     schwarze 1172:                cp = stesc + 1;
1.181     schwarze 1173:                switch (*cp) {
1.207     schwarze 1174:                case '*':
1.181     schwarze 1175:                        res = NULL;
                   1176:                        break;
1.207     schwarze 1177:                case 'B':
                   1178:                case 'w':
1.206     schwarze 1179:                        term = cp[1];
                   1180:                        /* FALLTHROUGH */
1.207     schwarze 1181:                case 'n':
1.181     schwarze 1182:                        res = ubuf;
                   1183:                        break;
                   1184:                default:
1.237     schwarze 1185:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1186:                        if (esc == ESCAPE_ERROR ||
                   1187:                            (esc == ESCAPE_SPECIAL &&
1.279     schwarze 1188:                             mchars_spec2cp(stnam, inaml) < 0))
1.219     schwarze 1189:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238     schwarze 1190:                                    r->parse, ln, (int)(stesc - buf->buf),
1.219     schwarze 1191:                                    "%.*s", (int)(cp - stesc), stesc);
1.205     schwarze 1192:                        continue;
1.152     kristaps 1193:                }
                   1194:
1.205     schwarze 1195:                if (EXPAND_LIMIT < ++expand_count) {
                   1196:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238     schwarze 1197:                            ln, (int)(stesc - buf->buf), NULL);
1.277     schwarze 1198:                        return ROFF_IGN;
1.205     schwarze 1199:                }
1.108     schwarze 1200:
                   1201:                /*
                   1202:                 * The third character decides the length
1.181     schwarze 1203:                 * of the name of the string or register.
1.108     schwarze 1204:                 * Save a pointer to the name.
                   1205:                 */
                   1206:
1.238     schwarze 1207:                if (term == '\0') {
1.206     schwarze 1208:                        switch (*++cp) {
1.207     schwarze 1209:                        case '\0':
1.206     schwarze 1210:                                maxl = 0;
                   1211:                                break;
1.207     schwarze 1212:                        case '(':
1.206     schwarze 1213:                                cp++;
                   1214:                                maxl = 2;
                   1215:                                break;
1.207     schwarze 1216:                        case '[':
1.206     schwarze 1217:                                cp++;
                   1218:                                term = ']';
                   1219:                                maxl = 0;
                   1220:                                break;
                   1221:                        default:
                   1222:                                maxl = 1;
                   1223:                                break;
                   1224:                        }
                   1225:                } else {
                   1226:                        cp += 2;
1.94      kristaps 1227:                        maxl = 0;
                   1228:                }
1.108     schwarze 1229:                stnam = cp;
1.94      kristaps 1230:
1.108     schwarze 1231:                /* Advance to the end of the name. */
1.94      kristaps 1232:
1.253     schwarze 1233:                naml = 0;
1.218     schwarze 1234:                arg_complete = 1;
1.253     schwarze 1235:                while (maxl == 0 || naml < maxl) {
1.238     schwarze 1236:                        if (*cp == '\0') {
1.219     schwarze 1237:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238     schwarze 1238:                                    ln, (int)(stesc - buf->buf), stesc);
1.218     schwarze 1239:                                arg_complete = 0;
1.206     schwarze 1240:                                break;
1.153     kristaps 1241:                        }
1.238     schwarze 1242:                        if (maxl == 0 && *cp == term) {
1.206     schwarze 1243:                                cp++;
1.253     schwarze 1244:                                break;
                   1245:                        }
                   1246:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1247:                                naml++;
                   1248:                                continue;
                   1249:                        }
                   1250:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1251:                        case ESCAPE_SPECIAL:
                   1252:                        case ESCAPE_UNICODE:
                   1253:                        case ESCAPE_NUMBERED:
                   1254:                        case ESCAPE_OVERSTRIKE:
                   1255:                                naml++;
                   1256:                                break;
                   1257:                        default:
1.94      kristaps 1258:                                break;
1.206     schwarze 1259:                        }
1.94      kristaps 1260:                }
                   1261:
1.108     schwarze 1262:                /*
                   1263:                 * Retrieve the replacement string; if it is
                   1264:                 * undefined, resume searching for escapes.
                   1265:                 */
                   1266:
1.206     schwarze 1267:                switch (stesc[1]) {
1.207     schwarze 1268:                case '*':
1.218     schwarze 1269:                        if (arg_complete)
                   1270:                                res = roff_getstrn(r, stnam, naml);
1.206     schwarze 1271:                        break;
1.207     schwarze 1272:                case 'B':
1.206     schwarze 1273:                        npos = 0;
1.218     schwarze 1274:                        ubuf[0] = arg_complete &&
1.261     schwarze 1275:                            roff_evalnum(r, ln, stnam, &npos,
                   1276:                              NULL, ROFFNUM_SCALE) &&
1.218     schwarze 1277:                            stnam + npos + 1 == cp ? '1' : '0';
1.206     schwarze 1278:                        ubuf[1] = '\0';
                   1279:                        break;
1.207     schwarze 1280:                case 'n':
1.218     schwarze 1281:                        if (arg_complete)
                   1282:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1283:                                    roff_getregn(r, stnam, naml));
                   1284:                        else
                   1285:                                ubuf[0] = '\0';
1.206     schwarze 1286:                        break;
1.207     schwarze 1287:                case 'w':
1.218     schwarze 1288:                        /* use even incomplete args */
1.208     schwarze 1289:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206     schwarze 1290:                            24 * (int)naml);
                   1291:                        break;
                   1292:                }
1.94      kristaps 1293:
1.238     schwarze 1294:                if (res == NULL) {
1.219     schwarze 1295:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238     schwarze 1296:                            r->parse, ln, (int)(stesc - buf->buf),
1.219     schwarze 1297:                            "%.*s", (int)naml, stnam);
1.142     kristaps 1298:                        res = "";
1.246     schwarze 1299:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1300:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1301:                            ln, (int)(stesc - buf->buf), NULL);
1.277     schwarze 1302:                        return ROFF_IGN;
1.94      kristaps 1303:                }
                   1304:
1.108     schwarze 1305:                /* Replace the escape sequence by the string. */
                   1306:
1.209     schwarze 1307:                *stesc = '\0';
1.238     schwarze 1308:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1309:                    buf->buf, res, cp) + 1;
1.94      kristaps 1310:
1.205     schwarze 1311:                /* Prepare for the next replacement. */
1.94      kristaps 1312:
1.205     schwarze 1313:                start = nbuf + pos;
1.238     schwarze 1314:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1315:                free(buf->buf);
                   1316:                buf->buf = nbuf;
1.154     kristaps 1317:        }
1.277     schwarze 1318:        return ROFF_CONT;
1.154     kristaps 1319: }
                   1320:
                   1321: /*
1.275     schwarze 1322:  * Process text streams.
1.154     kristaps 1323:  */
                   1324: static enum rofferr
1.238     schwarze 1325: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154     kristaps 1326: {
                   1327:        size_t           sz;
                   1328:        const char      *start;
1.178     schwarze 1329:        char            *p;
                   1330:        int              isz;
1.154     kristaps 1331:        enum mandoc_esc  esc;
                   1332:
1.275     schwarze 1333:        /* Spring the input line trap. */
                   1334:
                   1335:        if (roffit_lines == 1) {
                   1336:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1337:                free(buf->buf);
                   1338:                buf->buf = p;
                   1339:                buf->sz = isz + 1;
                   1340:                *offs = 0;
                   1341:                free(roffit_macro);
                   1342:                roffit_lines = 0;
1.277     schwarze 1343:                return ROFF_REPARSE;
1.275     schwarze 1344:        } else if (roffit_lines > 1)
                   1345:                --roffit_lines;
                   1346:
                   1347:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1348:
1.238     schwarze 1349:        start = p = buf->buf + pos;
1.154     kristaps 1350:
1.238     schwarze 1351:        while (*p != '\0') {
1.154     kristaps 1352:                sz = strcspn(p, "-\\");
                   1353:                p += sz;
                   1354:
1.238     schwarze 1355:                if (*p == '\0')
1.159     kristaps 1356:                        break;
                   1357:
1.238     schwarze 1358:                if (*p == '\\') {
1.154     kristaps 1359:                        /* Skip over escapes. */
                   1360:                        p++;
1.189     schwarze 1361:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238     schwarze 1362:                        if (esc == ESCAPE_ERROR)
1.154     kristaps 1363:                                break;
1.264     schwarze 1364:                        while (*p == '-')
                   1365:                                p++;
1.155     kristaps 1366:                        continue;
1.159     kristaps 1367:                } else if (p == start) {
1.158     kristaps 1368:                        p++;
1.155     kristaps 1369:                        continue;
1.158     kristaps 1370:                }
1.155     kristaps 1371:
1.171     schwarze 1372:                if (isalpha((unsigned char)p[-1]) &&
                   1373:                    isalpha((unsigned char)p[1]))
1.155     kristaps 1374:                        *p = ASCII_HYPH;
                   1375:                p++;
1.94      kristaps 1376:        }
1.277     schwarze 1377:        return ROFF_CONT;
1.94      kristaps 1378: }
                   1379:
1.67      kristaps 1380: enum rofferr
1.238     schwarze 1381: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67      kristaps 1382: {
1.294     schwarze 1383:        enum roff_tok    t;
1.109     kristaps 1384:        enum rofferr     e;
1.238     schwarze 1385:        int              pos;   /* parse point */
1.243     schwarze 1386:        int              spos;  /* saved parse point for messages */
1.238     schwarze 1387:        int              ppos;  /* original offset in buf->buf */
                   1388:        int              ctl;   /* macro line (boolean) */
                   1389:
                   1390:        ppos = pos = *offs;
1.79      kristaps 1391:
1.230     schwarze 1392:        /* Handle in-line equation delimiters. */
                   1393:
1.236     schwarze 1394:        if (r->tbl == NULL &&
                   1395:            r->last_eqn != NULL && r->last_eqn->delim &&
1.230     schwarze 1396:            (r->eqn == NULL || r->eqn_inline)) {
1.238     schwarze 1397:                e = roff_eqndelim(r, buf, pos);
1.230     schwarze 1398:                if (e == ROFF_REPARSE)
1.277     schwarze 1399:                        return e;
1.230     schwarze 1400:                assert(e == ROFF_CONT);
                   1401:        }
                   1402:
                   1403:        /* Expand some escape sequences. */
1.94      kristaps 1404:
1.238     schwarze 1405:        e = roff_res(r, buf, ln, pos);
                   1406:        if (e == ROFF_IGN)
1.277     schwarze 1407:                return e;
1.238     schwarze 1408:        assert(e == ROFF_CONT);
1.94      kristaps 1409:
1.238     schwarze 1410:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.130     kristaps 1411:
1.94      kristaps 1412:        /*
1.79      kristaps 1413:         * First, if a scope is open and we're not a macro, pass the
1.252     schwarze 1414:         * text through the macro's filter.
                   1415:         * Equations process all content themselves.
                   1416:         * Tables process almost all content themselves, but we want
                   1417:         * to warn about macros before passing it there.
1.79      kristaps 1418:         */
1.74      kristaps 1419:
1.252     schwarze 1420:        if (r->last != NULL && ! ctl) {
1.78      kristaps 1421:                t = r->last->tok;
1.238     schwarze 1422:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.294     schwarze 1423:                if (e == ROFF_IGN)
1.277     schwarze 1424:                        return e;
1.294     schwarze 1425:                assert(e == ROFF_CONT);
1.182     schwarze 1426:        }
1.252     schwarze 1427:        if (r->eqn != NULL)
1.277     schwarze 1428:                return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.252     schwarze 1429:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.277     schwarze 1430:                return tbl_read(r->tbl, ln, buf->buf, ppos);
1.252     schwarze 1431:        if ( ! ctl)
1.277     schwarze 1432:                return roff_parsetext(buf, pos, offs);
1.228     schwarze 1433:
                   1434:        /* Skip empty request lines. */
                   1435:
1.238     schwarze 1436:        if (buf->buf[pos] == '"') {
1.228     schwarze 1437:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1438:                    ln, pos, NULL);
1.277     schwarze 1439:                return ROFF_IGN;
1.238     schwarze 1440:        } else if (buf->buf[pos] == '\0')
1.277     schwarze 1441:                return ROFF_IGN;
1.67      kristaps 1442:
1.79      kristaps 1443:        /*
                   1444:         * If a scope is open, go to the child handler for that macro,
                   1445:         * as it may want to preprocess before doing anything with it.
1.125     kristaps 1446:         * Don't do so if an equation is open.
1.79      kristaps 1447:         */
1.78      kristaps 1448:
1.79      kristaps 1449:        if (r->last) {
                   1450:                t = r->last->tok;
1.277     schwarze 1451:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1452:        }
1.78      kristaps 1453:
1.243     schwarze 1454:        /* No scope is open.  This is a new request or macro. */
                   1455:
                   1456:        spos = pos;
                   1457:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1458:
                   1459:        /* Tables ignore most macros. */
                   1460:
1.294     schwarze 1461:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS)) {
1.243     schwarze 1462:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1463:                    ln, pos, buf->buf + spos);
1.257     schwarze 1464:                if (t == ROFF_TS)
1.277     schwarze 1465:                        return ROFF_IGN;
1.257     schwarze 1466:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1467:                        pos++;
1.291     schwarze 1468:                while (buf->buf[pos] == ' ')
1.257     schwarze 1469:                        pos++;
1.277     schwarze 1470:                return tbl_read(r->tbl, ln, buf->buf, pos);
1.243     schwarze 1471:        }
                   1472:
1.79      kristaps 1473:        /*
1.243     schwarze 1474:         * This is neither a roff request nor a user-defined macro.
                   1475:         * Let the standard macro set parsers handle it.
1.79      kristaps 1476:         */
1.67      kristaps 1477:
1.294     schwarze 1478:        if (t == TOKEN_NONE)
1.277     schwarze 1479:                return ROFF_CONT;
1.243     schwarze 1480:
                   1481:        /* Execute a roff request or a user defined macro. */
1.67      kristaps 1482:
1.296   ! schwarze 1483:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74      kristaps 1484: }
                   1485:
1.117     kristaps 1486: void
1.74      kristaps 1487: roff_endparse(struct roff *r)
                   1488: {
                   1489:
1.110     kristaps 1490:        if (r->last)
1.221     schwarze 1491:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1492:                    r->last->line, r->last->col,
1.295     schwarze 1493:                    roff_name[r->last->tok]);
1.117     kristaps 1494:
1.125     kristaps 1495:        if (r->eqn) {
1.221     schwarze 1496:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1497:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151     kristaps 1498:                eqn_end(&r->eqn);
1.125     kristaps 1499:        }
                   1500:
1.117     kristaps 1501:        if (r->tbl) {
1.221     schwarze 1502:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1503:                    r->tbl->line, r->tbl->pos, "TS");
1.151     kristaps 1504:                tbl_end(&r->tbl);
1.117     kristaps 1505:        }
1.67      kristaps 1506: }
                   1507:
                   1508: /*
                   1509:  * Parse a roff node's type from the input buffer.  This must be in the
                   1510:  * form of ".foo xxx" in the usual way.
                   1511:  */
1.294     schwarze 1512: static enum roff_tok
1.214     schwarze 1513: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67      kristaps 1514: {
1.214     schwarze 1515:        char            *cp;
1.106     kristaps 1516:        const char      *mac;
                   1517:        size_t           maclen;
1.294     schwarze 1518:        enum roff_tok    t;
1.67      kristaps 1519:
1.214     schwarze 1520:        cp = buf + *pos;
                   1521:
                   1522:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294     schwarze 1523:                return TOKEN_NONE;
1.67      kristaps 1524:
1.214     schwarze 1525:        mac = cp;
                   1526:        maclen = roff_getname(r, &cp, ln, ppos);
1.67      kristaps 1527:
1.106     kristaps 1528:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.295     schwarze 1529:            ? ROFF_USERDEF : roffhash_find(r->reqtab, mac, maclen);
1.67      kristaps 1530:
1.294     schwarze 1531:        if (t != TOKEN_NONE)
1.214     schwarze 1532:                *pos = cp - buf;
1.67      kristaps 1533:
1.277     schwarze 1534:        return t;
1.67      kristaps 1535: }
                   1536:
1.266     schwarze 1537: /* --- handling of request blocks ----------------------------------------- */
                   1538:
1.67      kristaps 1539: static enum rofferr
1.76      kristaps 1540: roff_cblock(ROFF_ARGS)
1.67      kristaps 1541: {
                   1542:
1.79      kristaps 1543:        /*
                   1544:         * A block-close `..' should only be invoked as a child of an
                   1545:         * ignore macro, otherwise raise a warning and just ignore it.
                   1546:         */
                   1547:
1.238     schwarze 1548:        if (r->last == NULL) {
1.221     schwarze 1549:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1550:                    ln, ppos, "..");
1.277     schwarze 1551:                return ROFF_IGN;
1.76      kristaps 1552:        }
1.67      kristaps 1553:
1.81      kristaps 1554:        switch (r->last->tok) {
1.207     schwarze 1555:        case ROFF_am:
1.220     schwarze 1556:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207     schwarze 1557:        case ROFF_ami:
                   1558:        case ROFF_de:
1.108     schwarze 1559:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207     schwarze 1560:        case ROFF_dei:
                   1561:        case ROFF_ig:
1.81      kristaps 1562:                break;
                   1563:        default:
1.221     schwarze 1564:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1565:                    ln, ppos, "..");
1.277     schwarze 1566:                return ROFF_IGN;
1.76      kristaps 1567:        }
1.67      kristaps 1568:
1.238     schwarze 1569:        if (buf->buf[pos] != '\0')
1.217     schwarze 1570:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238     schwarze 1571:                    ".. %s", buf->buf + pos);
1.71      kristaps 1572:
                   1573:        roffnode_pop(r);
1.76      kristaps 1574:        roffnode_cleanscope(r);
1.277     schwarze 1575:        return ROFF_IGN;
1.71      kristaps 1576:
1.67      kristaps 1577: }
                   1578:
1.76      kristaps 1579: static void
                   1580: roffnode_cleanscope(struct roff *r)
1.67      kristaps 1581: {
                   1582:
1.76      kristaps 1583:        while (r->last) {
1.173     schwarze 1584:                if (--r->last->endspan != 0)
1.76      kristaps 1585:                        break;
                   1586:                roffnode_pop(r);
                   1587:        }
1.67      kristaps 1588: }
                   1589:
1.195     schwarze 1590: static void
                   1591: roff_ccond(struct roff *r, int ln, int ppos)
1.74      kristaps 1592: {
                   1593:
1.76      kristaps 1594:        if (NULL == r->last) {
1.221     schwarze 1595:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1596:                    ln, ppos, "\\}");
1.195     schwarze 1597:                return;
1.76      kristaps 1598:        }
                   1599:
1.82      kristaps 1600:        switch (r->last->tok) {
1.207     schwarze 1601:        case ROFF_el:
                   1602:        case ROFF_ie:
                   1603:        case ROFF_if:
1.82      kristaps 1604:                break;
                   1605:        default:
1.221     schwarze 1606:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1607:                    ln, ppos, "\\}");
1.195     schwarze 1608:                return;
1.75      kristaps 1609:        }
                   1610:
1.76      kristaps 1611:        if (r->last->endspan > -1) {
1.221     schwarze 1612:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1613:                    ln, ppos, "\\}");
1.195     schwarze 1614:                return;
1.76      kristaps 1615:        }
                   1616:
1.75      kristaps 1617:        roffnode_pop(r);
1.76      kristaps 1618:        roffnode_cleanscope(r);
1.195     schwarze 1619:        return;
1.76      kristaps 1620: }
                   1621:
                   1622: static enum rofferr
1.80      kristaps 1623: roff_block(ROFF_ARGS)
1.76      kristaps 1624: {
1.220     schwarze 1625:        const char      *name;
                   1626:        char            *iname, *cp;
1.213     schwarze 1627:        size_t           namesz;
1.106     kristaps 1628:
1.220     schwarze 1629:        /* Ignore groff compatibility mode for now. */
1.76      kristaps 1630:
1.238     schwarze 1631:        if (tok == ROFF_de1)
1.220     schwarze 1632:                tok = ROFF_de;
1.251     schwarze 1633:        else if (tok == ROFF_dei1)
                   1634:                tok = ROFF_dei;
1.238     schwarze 1635:        else if (tok == ROFF_am1)
1.220     schwarze 1636:                tok = ROFF_am;
1.251     schwarze 1637:        else if (tok == ROFF_ami1)
                   1638:                tok = ROFF_ami;
1.220     schwarze 1639:
                   1640:        /* Parse the macro name argument. */
                   1641:
1.238     schwarze 1642:        cp = buf->buf + pos;
                   1643:        if (tok == ROFF_ig) {
1.220     schwarze 1644:                iname = NULL;
                   1645:                namesz = 0;
                   1646:        } else {
                   1647:                iname = cp;
                   1648:                namesz = roff_getname(r, &cp, ln, ppos);
                   1649:                iname[namesz] = '\0';
                   1650:        }
1.107     kristaps 1651:
1.220     schwarze 1652:        /* Resolve the macro name argument if it is indirect. */
1.107     kristaps 1653:
1.238     schwarze 1654:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1655:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220     schwarze 1656:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238     schwarze 1657:                            r->parse, ln, (int)(iname - buf->buf),
1.220     schwarze 1658:                            "%.*s", (int)namesz, iname);
                   1659:                        namesz = 0;
                   1660:                } else
                   1661:                        namesz = strlen(name);
                   1662:        } else
                   1663:                name = iname;
1.107     kristaps 1664:
1.238     schwarze 1665:        if (namesz == 0 && tok != ROFF_ig) {
1.220     schwarze 1666:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295     schwarze 1667:                    ln, ppos, roff_name[tok]);
1.277     schwarze 1668:                return ROFF_IGN;
1.220     schwarze 1669:        }
1.80      kristaps 1670:
1.106     kristaps 1671:        roffnode_push(r, tok, name, ln, ppos);
                   1672:
                   1673:        /*
                   1674:         * At the beginning of a `de' macro, clear the existing string
                   1675:         * with the same name, if there is one.  New content will be
1.193     schwarze 1676:         * appended from roff_block_text() in multiline mode.
1.106     kristaps 1677:         */
1.107     kristaps 1678:
1.238     schwarze 1679:        if (tok == ROFF_de || tok == ROFF_dei)
1.213     schwarze 1680:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76      kristaps 1681:
1.238     schwarze 1682:        if (*cp == '\0')
1.277     schwarze 1683:                return ROFF_IGN;
1.78      kristaps 1684:
1.220     schwarze 1685:        /* Get the custom end marker. */
1.107     kristaps 1686:
1.220     schwarze 1687:        iname = cp;
1.213     schwarze 1688:        namesz = roff_getname(r, &cp, ln, ppos);
1.220     schwarze 1689:
                   1690:        /* Resolve the end marker if it is indirect. */
                   1691:
1.238     schwarze 1692:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1693:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220     schwarze 1694:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238     schwarze 1695:                            r->parse, ln, (int)(iname - buf->buf),
1.220     schwarze 1696:                            "%.*s", (int)namesz, iname);
                   1697:                        namesz = 0;
                   1698:                } else
                   1699:                        namesz = strlen(name);
                   1700:        } else
                   1701:                name = iname;
                   1702:
1.213     schwarze 1703:        if (namesz)
                   1704:                r->last->end = mandoc_strndup(name, namesz);
1.78      kristaps 1705:
1.238     schwarze 1706:        if (*cp != '\0')
1.217     schwarze 1707:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295     schwarze 1708:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74      kristaps 1709:
1.277     schwarze 1710:        return ROFF_IGN;
1.78      kristaps 1711: }
                   1712:
                   1713: static enum rofferr
1.80      kristaps 1714: roff_block_sub(ROFF_ARGS)
1.79      kristaps 1715: {
1.294     schwarze 1716:        enum roff_tok   t;
1.79      kristaps 1717:        int             i, j;
                   1718:
                   1719:        /*
                   1720:         * First check whether a custom macro exists at this level.  If
                   1721:         * it does, then check against it.  This is some of groff's
                   1722:         * stranger behaviours.  If we encountered a custom end-scope
                   1723:         * tag and that tag also happens to be a "real" macro, then we
                   1724:         * need to try interpreting it again as a real macro.  If it's
                   1725:         * not, then return ignore.  Else continue.
                   1726:         */
                   1727:
                   1728:        if (r->last->end) {
1.130     kristaps 1729:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238     schwarze 1730:                        if (buf->buf[i] != r->last->end[j])
1.79      kristaps 1731:                                break;
                   1732:
1.238     schwarze 1733:                if (r->last->end[j] == '\0' &&
                   1734:                    (buf->buf[i] == '\0' ||
                   1735:                     buf->buf[i] == ' ' ||
                   1736:                     buf->buf[i] == '\t')) {
1.79      kristaps 1737:                        roffnode_pop(r);
                   1738:                        roffnode_cleanscope(r);
                   1739:
1.238     schwarze 1740:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130     kristaps 1741:                                i++;
                   1742:
                   1743:                        pos = i;
1.238     schwarze 1744:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294     schwarze 1745:                            TOKEN_NONE)
1.277     schwarze 1746:                                return ROFF_RERUN;
                   1747:                        return ROFF_IGN;
1.79      kristaps 1748:                }
                   1749:        }
                   1750:
                   1751:        /*
                   1752:         * If we have no custom end-query or lookup failed, then try
                   1753:         * pulling it out of the hashtable.
                   1754:         */
                   1755:
1.238     schwarze 1756:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79      kristaps 1757:
1.238     schwarze 1758:        if (t != ROFF_cblock) {
                   1759:                if (tok != ROFF_ig)
                   1760:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277     schwarze 1761:                return ROFF_IGN;
1.106     kristaps 1762:        }
1.79      kristaps 1763:
1.277     schwarze 1764:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1765: }
                   1766:
                   1767: static enum rofferr
1.80      kristaps 1768: roff_block_text(ROFF_ARGS)
1.78      kristaps 1769: {
                   1770:
1.238     schwarze 1771:        if (tok != ROFF_ig)
                   1772:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106     kristaps 1773:
1.277     schwarze 1774:        return ROFF_IGN;
1.78      kristaps 1775: }
                   1776:
                   1777: static enum rofferr
1.82      kristaps 1778: roff_cond_sub(ROFF_ARGS)
                   1779: {
1.294     schwarze 1780:        enum roff_tok    t;
1.139     kristaps 1781:        char            *ep;
1.198     schwarze 1782:        int              rr;
1.82      kristaps 1783:
                   1784:        rr = r->last->rule;
1.139     kristaps 1785:        roffnode_cleanscope(r);
1.238     schwarze 1786:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82      kristaps 1787:
1.139     kristaps 1788:        /*
1.177     schwarze 1789:         * Fully handle known macros when they are structurally
                   1790:         * required or when the conditional evaluated to true.
1.87      kristaps 1791:         */
                   1792:
1.294     schwarze 1793:        if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.277     schwarze 1794:                return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.144     kristaps 1795:
1.196     schwarze 1796:        /*
                   1797:         * If `\}' occurs on a macro line without a preceding macro,
                   1798:         * drop the line completely.
                   1799:         */
                   1800:
1.238     schwarze 1801:        ep = buf->buf + pos;
                   1802:        if (ep[0] == '\\' && ep[1] == '}')
1.198     schwarze 1803:                rr = 0;
1.196     schwarze 1804:
1.177     schwarze 1805:        /* Always check for the closing delimiter `\}'. */
1.144     kristaps 1806:
1.238     schwarze 1807:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1808:                if (*(++ep) == '}') {
1.197     schwarze 1809:                        *ep = '&';
1.238     schwarze 1810:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.197     schwarze 1811:                }
1.247     schwarze 1812:                if (*ep != '\0')
                   1813:                        ++ep;
1.177     schwarze 1814:        }
1.277     schwarze 1815:        return rr ? ROFF_CONT : ROFF_IGN;
1.82      kristaps 1816: }
                   1817:
                   1818: static enum rofferr
                   1819: roff_cond_text(ROFF_ARGS)
1.78      kristaps 1820: {
1.140     kristaps 1821:        char            *ep;
1.198     schwarze 1822:        int              rr;
1.82      kristaps 1823:
                   1824:        rr = r->last->rule;
1.140     kristaps 1825:        roffnode_cleanscope(r);
1.82      kristaps 1826:
1.238     schwarze 1827:        ep = buf->buf + pos;
                   1828:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1829:                if (*(++ep) == '}') {
1.197     schwarze 1830:                        *ep = '&';
1.238     schwarze 1831:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.197     schwarze 1832:                }
1.247     schwarze 1833:                if (*ep != '\0')
                   1834:                        ++ep;
1.78      kristaps 1835:        }
1.277     schwarze 1836:        return rr ? ROFF_CONT : ROFF_IGN;
1.74      kristaps 1837: }
                   1838:
1.266     schwarze 1839: /* --- handling of numeric and conditional expressions -------------------- */
                   1840:
1.204     schwarze 1841: /*
                   1842:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1843:  * If there is at least one digit, return success and advance the
                   1844:  * parse point, else return failure and let the parse point unchanged.
                   1845:  * Ignore overflows, treat them just like the C language.
                   1846:  */
1.184     schwarze 1847: static int
1.261     schwarze 1848: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184     schwarze 1849: {
1.261     schwarze 1850:        int      myres, scaled, n, p;
1.206     schwarze 1851:
                   1852:        if (NULL == res)
                   1853:                res = &myres;
1.184     schwarze 1854:
                   1855:        p = *pos;
                   1856:        n = v[p] == '-';
1.261     schwarze 1857:        if (n || v[p] == '+')
1.184     schwarze 1858:                p++;
                   1859:
1.261     schwarze 1860:        if (flags & ROFFNUM_WHITE)
                   1861:                while (isspace((unsigned char)v[p]))
                   1862:                        p++;
                   1863:
1.184     schwarze 1864:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204     schwarze 1865:                *res = 10 * *res + v[p] - '0';
1.184     schwarze 1866:        if (p == *pos + n)
                   1867:                return 0;
                   1868:
                   1869:        if (n)
                   1870:                *res = -*res;
                   1871:
1.254     schwarze 1872:        /* Each number may be followed by one optional scaling unit. */
                   1873:
                   1874:        switch (v[p]) {
                   1875:        case 'f':
1.261     schwarze 1876:                scaled = *res * 65536;
1.254     schwarze 1877:                break;
                   1878:        case 'i':
1.261     schwarze 1879:                scaled = *res * 240;
1.254     schwarze 1880:                break;
                   1881:        case 'c':
1.261     schwarze 1882:                scaled = *res * 240 / 2.54;
1.254     schwarze 1883:                break;
                   1884:        case 'v':
                   1885:        case 'P':
1.261     schwarze 1886:                scaled = *res * 40;
1.254     schwarze 1887:                break;
                   1888:        case 'm':
                   1889:        case 'n':
1.261     schwarze 1890:                scaled = *res * 24;
1.254     schwarze 1891:                break;
                   1892:        case 'p':
1.261     schwarze 1893:                scaled = *res * 10 / 3;
1.254     schwarze 1894:                break;
                   1895:        case 'u':
1.261     schwarze 1896:                scaled = *res;
1.254     schwarze 1897:                break;
                   1898:        case 'M':
1.261     schwarze 1899:                scaled = *res * 6 / 25;
1.254     schwarze 1900:                break;
                   1901:        default:
1.261     schwarze 1902:                scaled = *res;
1.254     schwarze 1903:                p--;
                   1904:                break;
                   1905:        }
1.261     schwarze 1906:        if (flags & ROFFNUM_SCALE)
                   1907:                *res = scaled;
1.254     schwarze 1908:
                   1909:        *pos = p + 1;
1.277     schwarze 1910:        return 1;
1.184     schwarze 1911: }
                   1912:
1.198     schwarze 1913: /*
                   1914:  * Evaluate a string comparison condition.
                   1915:  * The first character is the delimiter.
                   1916:  * Succeed if the string up to its second occurrence
                   1917:  * matches the string up to its third occurence.
                   1918:  * Advance the cursor after the third occurrence
                   1919:  * or lacking that, to the end of the line.
                   1920:  */
                   1921: static int
                   1922: roff_evalstrcond(const char *v, int *pos)
                   1923: {
                   1924:        const char      *s1, *s2, *s3;
                   1925:        int              match;
                   1926:
                   1927:        match = 0;
                   1928:        s1 = v + *pos;          /* initial delimiter */
                   1929:        s2 = s1 + 1;            /* for scanning the first string */
                   1930:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   1931:
                   1932:        if (NULL == s3)         /* found no middle delimiter */
                   1933:                goto out;
                   1934:
                   1935:        while ('\0' != *++s3) {
                   1936:                if (*s2 != *s3) {  /* mismatch */
                   1937:                        s3 = strchr(s3, *s1);
                   1938:                        break;
                   1939:                }
                   1940:                if (*s3 == *s1) {  /* found the final delimiter */
                   1941:                        match = 1;
                   1942:                        break;
                   1943:                }
                   1944:                s2++;
                   1945:        }
                   1946:
                   1947: out:
                   1948:        if (NULL == s3)
                   1949:                s3 = strchr(s2, '\0');
1.242     schwarze 1950:        else if (*s3 != '\0')
1.198     schwarze 1951:                s3++;
                   1952:        *pos = s3 - v;
1.277     schwarze 1953:        return match;
1.198     schwarze 1954: }
                   1955:
1.204     schwarze 1956: /*
                   1957:  * Evaluate an optionally negated single character, numerical,
                   1958:  * or string condition.
                   1959:  */
1.198     schwarze 1960: static int
1.271     schwarze 1961: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88      kristaps 1962: {
1.271     schwarze 1963:        char    *cp, *name;
                   1964:        size_t   sz;
1.241     schwarze 1965:        int      number, savepos, wanttrue;
1.88      kristaps 1966:
1.198     schwarze 1967:        if ('!' == v[*pos]) {
                   1968:                wanttrue = 0;
                   1969:                (*pos)++;
                   1970:        } else
                   1971:                wanttrue = 1;
                   1972:
1.88      kristaps 1973:        switch (v[*pos]) {
1.240     schwarze 1974:        case '\0':
1.277     schwarze 1975:                return 0;
1.207     schwarze 1976:        case 'n':
                   1977:        case 'o':
1.88      kristaps 1978:                (*pos)++;
1.277     schwarze 1979:                return wanttrue;
1.207     schwarze 1980:        case 'c':
                   1981:        case 'd':
                   1982:        case 'e':
                   1983:        case 't':
1.239     schwarze 1984:        case 'v':
1.88      kristaps 1985:                (*pos)++;
1.277     schwarze 1986:                return !wanttrue;
1.271     schwarze 1987:        case 'r':
                   1988:                cp = name = v + ++*pos;
                   1989:                sz = roff_getname(r, &cp, ln, *pos);
                   1990:                *pos = cp - v;
1.277     schwarze 1991:                return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.88      kristaps 1992:        default:
                   1993:                break;
                   1994:        }
                   1995:
1.241     schwarze 1996:        savepos = *pos;
1.261     schwarze 1997:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277     schwarze 1998:                return (number > 0) == wanttrue;
1.241     schwarze 1999:        else if (*pos == savepos)
1.277     schwarze 2000:                return roff_evalstrcond(v, pos) == wanttrue;
1.204     schwarze 2001:        else
1.277     schwarze 2002:                return 0;
1.88      kristaps 2003: }
                   2004:
1.74      kristaps 2005: static enum rofferr
1.103     kristaps 2006: roff_line_ignore(ROFF_ARGS)
1.89      kristaps 2007: {
1.123     schwarze 2008:
1.277     schwarze 2009:        return ROFF_IGN;
1.89      kristaps 2010: }
                   2011:
1.104     kristaps 2012: static enum rofferr
1.251     schwarze 2013: roff_insec(ROFF_ARGS)
                   2014: {
                   2015:
                   2016:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.295     schwarze 2017:            ln, ppos, roff_name[tok]);
1.277     schwarze 2018:        return ROFF_IGN;
1.251     schwarze 2019: }
                   2020:
                   2021: static enum rofferr
                   2022: roff_unsupp(ROFF_ARGS)
                   2023: {
                   2024:
                   2025:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295     schwarze 2026:            ln, ppos, roff_name[tok]);
1.277     schwarze 2027:        return ROFF_IGN;
1.251     schwarze 2028: }
                   2029:
                   2030: static enum rofferr
1.82      kristaps 2031: roff_cond(ROFF_ARGS)
1.74      kristaps 2032: {
1.173     schwarze 2033:
                   2034:        roffnode_push(r, tok, NULL, ln, ppos);
1.74      kristaps 2035:
1.207     schwarze 2036:        /*
1.134     kristaps 2037:         * An `.el' has no conditional body: it will consume the value
                   2038:         * of the current rstack entry set in prior `ie' calls or
1.207     schwarze 2039:         * defaults to DENY.
1.134     kristaps 2040:         *
                   2041:         * If we're not an `el', however, then evaluate the conditional.
                   2042:         */
1.133     kristaps 2043:
1.238     schwarze 2044:        r->last->rule = tok == ROFF_el ?
1.207     schwarze 2045:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238     schwarze 2046:            roff_evalcond(r, ln, buf->buf, &pos);
1.77      kristaps 2047:
1.134     kristaps 2048:        /*
                   2049:         * An if-else will put the NEGATION of the current evaluated
                   2050:         * conditional into the stack of rules.
                   2051:         */
                   2052:
1.238     schwarze 2053:        if (tok == ROFF_ie) {
1.223     schwarze 2054:                if (r->rstackpos + 1 == r->rstacksz) {
                   2055:                        r->rstacksz += 16;
                   2056:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2057:                            r->rstacksz, sizeof(int));
1.134     kristaps 2058:                }
1.198     schwarze 2059:                r->rstack[++r->rstackpos] = !r->last->rule;
1.82      kristaps 2060:        }
1.88      kristaps 2061:
                   2062:        /* If the parent has false as its rule, then so do we. */
                   2063:
1.198     schwarze 2064:        if (r->last->parent && !r->last->parent->rule)
                   2065:                r->last->rule = 0;
1.88      kristaps 2066:
                   2067:        /*
1.173     schwarze 2068:         * Determine scope.
                   2069:         * If there is nothing on the line after the conditional,
                   2070:         * not even whitespace, use next-line scope.
1.88      kristaps 2071:         */
1.74      kristaps 2072:
1.238     schwarze 2073:        if (buf->buf[pos] == '\0') {
1.173     schwarze 2074:                r->last->endspan = 2;
                   2075:                goto out;
                   2076:        }
                   2077:
1.238     schwarze 2078:        while (buf->buf[pos] == ' ')
1.173     schwarze 2079:                pos++;
                   2080:
                   2081:        /* An opening brace requests multiline scope. */
1.75      kristaps 2082:
1.238     schwarze 2083:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75      kristaps 2084:                r->last->endspan = -1;
                   2085:                pos += 2;
1.272     schwarze 2086:                while (buf->buf[pos] == ' ')
                   2087:                        pos++;
1.173     schwarze 2088:                goto out;
1.207     schwarze 2089:        }
1.74      kristaps 2090:
1.77      kristaps 2091:        /*
1.173     schwarze 2092:         * Anything else following the conditional causes
                   2093:         * single-line scope.  Warn if the scope contains
                   2094:         * nothing but trailing whitespace.
1.77      kristaps 2095:         */
                   2096:
1.238     schwarze 2097:        if (buf->buf[pos] == '\0')
1.216     schwarze 2098:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295     schwarze 2099:                    ln, ppos, roff_name[tok]);
1.77      kristaps 2100:
1.173     schwarze 2101:        r->last->endspan = 1;
1.74      kristaps 2102:
1.173     schwarze 2103: out:
1.75      kristaps 2104:        *offs = pos;
1.277     schwarze 2105:        return ROFF_RERUN;
1.83      schwarze 2106: }
                   2107:
                   2108: static enum rofferr
1.92      schwarze 2109: roff_ds(ROFF_ARGS)
                   2110: {
1.212     schwarze 2111:        char            *string;
                   2112:        const char      *name;
                   2113:        size_t           namesz;
1.96      kristaps 2114:
1.251     schwarze 2115:        /* Ignore groff compatibility mode for now. */
                   2116:
                   2117:        if (tok == ROFF_ds1)
                   2118:                tok = ROFF_ds;
                   2119:        else if (tok == ROFF_as1)
                   2120:                tok = ROFF_as;
                   2121:
1.96      kristaps 2122:        /*
1.212     schwarze 2123:         * The first word is the name of the string.
                   2124:         * If it is empty or terminated by an escape sequence,
                   2125:         * abort the `ds' request without defining anything.
1.96      kristaps 2126:         */
1.92      schwarze 2127:
1.238     schwarze 2128:        name = string = buf->buf + pos;
                   2129:        if (*name == '\0')
1.277     schwarze 2130:                return ROFF_IGN;
1.92      schwarze 2131:
1.212     schwarze 2132:        namesz = roff_getname(r, &string, ln, pos);
1.238     schwarze 2133:        if (name[namesz] == '\\')
1.277     schwarze 2134:                return ROFF_IGN;
1.212     schwarze 2135:
                   2136:        /* Read past the initial double-quote, if any. */
1.238     schwarze 2137:        if (*string == '"')
1.92      schwarze 2138:                string++;
                   2139:
1.96      kristaps 2140:        /* The rest is the value. */
1.212     schwarze 2141:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2142:            ROFF_as == tok);
1.277     schwarze 2143:        return ROFF_IGN;
1.92      schwarze 2144: }
                   2145:
1.204     schwarze 2146: /*
                   2147:  * Parse a single operator, one or two characters long.
                   2148:  * If the operator is recognized, return success and advance the
                   2149:  * parse point, else return failure and let the parse point unchanged.
                   2150:  */
                   2151: static int
                   2152: roff_getop(const char *v, int *pos, char *res)
                   2153: {
                   2154:
                   2155:        *res = v[*pos];
                   2156:
                   2157:        switch (*res) {
1.207     schwarze 2158:        case '+':
                   2159:        case '-':
                   2160:        case '*':
                   2161:        case '/':
                   2162:        case '%':
                   2163:        case '&':
                   2164:        case ':':
1.204     schwarze 2165:                break;
                   2166:        case '<':
                   2167:                switch (v[*pos + 1]) {
1.207     schwarze 2168:                case '=':
1.204     schwarze 2169:                        *res = 'l';
                   2170:                        (*pos)++;
                   2171:                        break;
1.207     schwarze 2172:                case '>':
1.204     schwarze 2173:                        *res = '!';
                   2174:                        (*pos)++;
                   2175:                        break;
1.207     schwarze 2176:                case '?':
1.204     schwarze 2177:                        *res = 'i';
                   2178:                        (*pos)++;
                   2179:                        break;
                   2180:                default:
                   2181:                        break;
                   2182:                }
                   2183:                break;
                   2184:        case '>':
                   2185:                switch (v[*pos + 1]) {
1.207     schwarze 2186:                case '=':
1.204     schwarze 2187:                        *res = 'g';
                   2188:                        (*pos)++;
                   2189:                        break;
1.207     schwarze 2190:                case '?':
1.204     schwarze 2191:                        *res = 'a';
                   2192:                        (*pos)++;
                   2193:                        break;
                   2194:                default:
                   2195:                        break;
                   2196:                }
                   2197:                break;
                   2198:        case '=':
                   2199:                if ('=' == v[*pos + 1])
                   2200:                        (*pos)++;
                   2201:                break;
                   2202:        default:
1.277     schwarze 2203:                return 0;
1.204     schwarze 2204:        }
                   2205:        (*pos)++;
                   2206:
1.277     schwarze 2207:        return *res;
1.204     schwarze 2208: }
                   2209:
                   2210: /*
                   2211:  * Evaluate either a parenthesized numeric expression
                   2212:  * or a single signed integer number.
                   2213:  */
                   2214: static int
1.235     schwarze 2215: roff_evalpar(struct roff *r, int ln,
1.261     schwarze 2216:        const char *v, int *pos, int *res, int flags)
1.204     schwarze 2217: {
                   2218:
                   2219:        if ('(' != v[*pos])
1.277     schwarze 2220:                return roff_getnum(v, pos, res, flags);
1.204     schwarze 2221:
                   2222:        (*pos)++;
1.261     schwarze 2223:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277     schwarze 2224:                return 0;
1.204     schwarze 2225:
1.206     schwarze 2226:        /*
                   2227:         * Omission of the closing parenthesis
                   2228:         * is an error in validation mode,
                   2229:         * but ignored in evaluation mode.
                   2230:         */
                   2231:
1.204     schwarze 2232:        if (')' == v[*pos])
                   2233:                (*pos)++;
1.206     schwarze 2234:        else if (NULL == res)
1.277     schwarze 2235:                return 0;
1.204     schwarze 2236:
1.277     schwarze 2237:        return 1;
1.204     schwarze 2238: }
                   2239:
                   2240: /*
                   2241:  * Evaluate a complete numeric expression.
                   2242:  * Proceed left to right, there is no concept of precedence.
                   2243:  */
                   2244: static int
1.235     schwarze 2245: roff_evalnum(struct roff *r, int ln, const char *v,
1.261     schwarze 2246:        int *pos, int *res, int flags)
1.204     schwarze 2247: {
                   2248:        int              mypos, operand2;
                   2249:        char             operator;
                   2250:
                   2251:        if (NULL == pos) {
                   2252:                mypos = 0;
                   2253:                pos = &mypos;
                   2254:        }
                   2255:
1.261     schwarze 2256:        if (flags & ROFFNUM_WHITE)
1.204     schwarze 2257:                while (isspace((unsigned char)v[*pos]))
                   2258:                        (*pos)++;
                   2259:
1.261     schwarze 2260:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277     schwarze 2261:                return 0;
1.204     schwarze 2262:
                   2263:        while (1) {
1.261     schwarze 2264:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2265:                        while (isspace((unsigned char)v[*pos]))
                   2266:                                (*pos)++;
                   2267:
                   2268:                if ( ! roff_getop(v, pos, &operator))
                   2269:                        break;
                   2270:
1.261     schwarze 2271:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2272:                        while (isspace((unsigned char)v[*pos]))
                   2273:                                (*pos)++;
                   2274:
1.261     schwarze 2275:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277     schwarze 2276:                        return 0;
1.204     schwarze 2277:
1.261     schwarze 2278:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2279:                        while (isspace((unsigned char)v[*pos]))
                   2280:                                (*pos)++;
1.206     schwarze 2281:
                   2282:                if (NULL == res)
                   2283:                        continue;
1.204     schwarze 2284:
                   2285:                switch (operator) {
1.207     schwarze 2286:                case '+':
1.204     schwarze 2287:                        *res += operand2;
                   2288:                        break;
1.207     schwarze 2289:                case '-':
1.204     schwarze 2290:                        *res -= operand2;
                   2291:                        break;
1.207     schwarze 2292:                case '*':
1.204     schwarze 2293:                        *res *= operand2;
                   2294:                        break;
1.207     schwarze 2295:                case '/':
1.244     schwarze 2296:                        if (operand2 == 0) {
1.235     schwarze 2297:                                mandoc_msg(MANDOCERR_DIVZERO,
1.234     kristaps 2298:                                        r->parse, ln, *pos, v);
                   2299:                                *res = 0;
                   2300:                                break;
                   2301:                        }
1.204     schwarze 2302:                        *res /= operand2;
                   2303:                        break;
1.207     schwarze 2304:                case '%':
1.244     schwarze 2305:                        if (operand2 == 0) {
                   2306:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2307:                                        r->parse, ln, *pos, v);
                   2308:                                *res = 0;
                   2309:                                break;
                   2310:                        }
1.204     schwarze 2311:                        *res %= operand2;
                   2312:                        break;
1.207     schwarze 2313:                case '<':
1.204     schwarze 2314:                        *res = *res < operand2;
                   2315:                        break;
1.207     schwarze 2316:                case '>':
1.204     schwarze 2317:                        *res = *res > operand2;
                   2318:                        break;
1.207     schwarze 2319:                case 'l':
1.204     schwarze 2320:                        *res = *res <= operand2;
                   2321:                        break;
1.207     schwarze 2322:                case 'g':
1.204     schwarze 2323:                        *res = *res >= operand2;
                   2324:                        break;
1.207     schwarze 2325:                case '=':
1.204     schwarze 2326:                        *res = *res == operand2;
                   2327:                        break;
1.207     schwarze 2328:                case '!':
1.204     schwarze 2329:                        *res = *res != operand2;
                   2330:                        break;
1.207     schwarze 2331:                case '&':
1.204     schwarze 2332:                        *res = *res && operand2;
                   2333:                        break;
1.207     schwarze 2334:                case ':':
1.204     schwarze 2335:                        *res = *res || operand2;
                   2336:                        break;
1.207     schwarze 2337:                case 'i':
1.204     schwarze 2338:                        if (operand2 < *res)
                   2339:                                *res = operand2;
                   2340:                        break;
1.207     schwarze 2341:                case 'a':
1.204     schwarze 2342:                        if (operand2 > *res)
                   2343:                                *res = operand2;
                   2344:                        break;
                   2345:                default:
                   2346:                        abort();
                   2347:                }
                   2348:        }
1.277     schwarze 2349:        return 1;
1.204     schwarze 2350: }
                   2351:
1.266     schwarze 2352: /* --- register management ------------------------------------------------ */
                   2353:
1.180     schwarze 2354: void
1.187     schwarze 2355: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147     kristaps 2356: {
1.180     schwarze 2357:        struct roffreg  *reg;
                   2358:
                   2359:        /* Search for an existing register with the same name. */
                   2360:        reg = r->regtab;
                   2361:
                   2362:        while (reg && strcmp(name, reg->key.p))
                   2363:                reg = reg->next;
1.147     kristaps 2364:
1.180     schwarze 2365:        if (NULL == reg) {
                   2366:                /* Create a new register. */
                   2367:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2368:                reg->key.p = mandoc_strdup(name);
                   2369:                reg->key.sz = strlen(name);
1.187     schwarze 2370:                reg->val = 0;
1.180     schwarze 2371:                reg->next = r->regtab;
                   2372:                r->regtab = reg;
                   2373:        }
                   2374:
1.187     schwarze 2375:        if ('+' == sign)
                   2376:                reg->val += val;
                   2377:        else if ('-' == sign)
                   2378:                reg->val -= val;
                   2379:        else
                   2380:                reg->val = val;
1.147     kristaps 2381: }
                   2382:
1.192     schwarze 2383: /*
                   2384:  * Handle some predefined read-only number registers.
                   2385:  * For now, return -1 if the requested register is not predefined;
                   2386:  * in case a predefined read-only register having the value -1
                   2387:  * were to turn up, another special value would have to be chosen.
                   2388:  */
                   2389: static int
1.273     schwarze 2390: roff_getregro(const struct roff *r, const char *name)
1.192     schwarze 2391: {
                   2392:
                   2393:        switch (*name) {
1.273     schwarze 2394:        case '$':  /* Number of arguments of the last macro evaluated. */
1.277     schwarze 2395:                return r->argc;
1.207     schwarze 2396:        case 'A':  /* ASCII approximation mode is always off. */
1.277     schwarze 2397:                return 0;
1.207     schwarze 2398:        case 'g':  /* Groff compatibility mode is always on. */
1.277     schwarze 2399:                return 1;
1.207     schwarze 2400:        case 'H':  /* Fixed horizontal resolution. */
1.277     schwarze 2401:                return 24;
1.207     schwarze 2402:        case 'j':  /* Always adjust left margin only. */
1.277     schwarze 2403:                return 0;
1.207     schwarze 2404:        case 'T':  /* Some output device is always defined. */
1.277     schwarze 2405:                return 1;
1.207     schwarze 2406:        case 'V':  /* Fixed vertical resolution. */
1.277     schwarze 2407:                return 40;
1.192     schwarze 2408:        default:
1.277     schwarze 2409:                return -1;
1.192     schwarze 2410:        }
                   2411: }
                   2412:
1.181     schwarze 2413: int
1.180     schwarze 2414: roff_getreg(const struct roff *r, const char *name)
1.147     kristaps 2415: {
1.180     schwarze 2416:        struct roffreg  *reg;
1.192     schwarze 2417:        int              val;
                   2418:
                   2419:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.273     schwarze 2420:                val = roff_getregro(r, name + 1);
1.192     schwarze 2421:                if (-1 != val)
1.277     schwarze 2422:                        return val;
1.192     schwarze 2423:        }
1.180     schwarze 2424:
                   2425:        for (reg = r->regtab; reg; reg = reg->next)
                   2426:                if (0 == strcmp(name, reg->key.p))
1.277     schwarze 2427:                        return reg->val;
1.181     schwarze 2428:
1.277     schwarze 2429:        return 0;
1.181     schwarze 2430: }
                   2431:
                   2432: static int
                   2433: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2434: {
                   2435:        struct roffreg  *reg;
1.192     schwarze 2436:        int              val;
                   2437:
                   2438:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 2439:                val = roff_getregro(r, name + 1);
1.192     schwarze 2440:                if (-1 != val)
1.277     schwarze 2441:                        return val;
1.192     schwarze 2442:        }
1.181     schwarze 2443:
                   2444:        for (reg = r->regtab; reg; reg = reg->next)
                   2445:                if (len == reg->key.sz &&
                   2446:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 2447:                        return reg->val;
1.271     schwarze 2448:
1.277     schwarze 2449:        return 0;
1.271     schwarze 2450: }
                   2451:
                   2452: static int
                   2453: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2454: {
                   2455:        struct roffreg  *reg;
                   2456:        int              val;
                   2457:
                   2458:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 2459:                val = roff_getregro(r, name + 1);
1.271     schwarze 2460:                if (-1 != val)
1.277     schwarze 2461:                        return 1;
1.271     schwarze 2462:        }
                   2463:
                   2464:        for (reg = r->regtab; reg; reg = reg->next)
                   2465:                if (len == reg->key.sz &&
                   2466:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 2467:                        return 1;
1.147     kristaps 2468:
1.277     schwarze 2469:        return 0;
1.147     kristaps 2470: }
                   2471:
1.180     schwarze 2472: static void
                   2473: roff_freereg(struct roffreg *reg)
1.147     kristaps 2474: {
1.180     schwarze 2475:        struct roffreg  *old_reg;
1.147     kristaps 2476:
1.180     schwarze 2477:        while (NULL != reg) {
                   2478:                free(reg->key.p);
                   2479:                old_reg = reg;
                   2480:                reg = reg->next;
                   2481:                free(old_reg);
                   2482:        }
1.147     kristaps 2483: }
1.92      schwarze 2484:
                   2485: static enum rofferr
1.89      kristaps 2486: roff_nr(ROFF_ARGS)
1.83      schwarze 2487: {
1.212     schwarze 2488:        char            *key, *val;
                   2489:        size_t           keysz;
1.138     kristaps 2490:        int              iv;
1.187     schwarze 2491:        char             sign;
1.89      kristaps 2492:
1.238     schwarze 2493:        key = val = buf->buf + pos;
                   2494:        if (*key == '\0')
1.277     schwarze 2495:                return ROFF_IGN;
1.212     schwarze 2496:
                   2497:        keysz = roff_getname(r, &val, ln, pos);
1.238     schwarze 2498:        if (key[keysz] == '\\')
1.277     schwarze 2499:                return ROFF_IGN;
1.212     schwarze 2500:        key[keysz] = '\0';
1.89      kristaps 2501:
1.187     schwarze 2502:        sign = *val;
1.238     schwarze 2503:        if (sign == '+' || sign == '-')
1.187     schwarze 2504:                val++;
                   2505:
1.261     schwarze 2506:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204     schwarze 2507:                roff_setreg(r, key, iv, sign);
1.109     kristaps 2508:
1.277     schwarze 2509:        return ROFF_IGN;
1.203     schwarze 2510: }
                   2511:
                   2512: static enum rofferr
                   2513: roff_rr(ROFF_ARGS)
                   2514: {
                   2515:        struct roffreg  *reg, **prev;
1.212     schwarze 2516:        char            *name, *cp;
                   2517:        size_t           namesz;
1.203     schwarze 2518:
1.238     schwarze 2519:        name = cp = buf->buf + pos;
                   2520:        if (*name == '\0')
1.277     schwarze 2521:                return ROFF_IGN;
1.212     schwarze 2522:        namesz = roff_getname(r, &cp, ln, pos);
                   2523:        name[namesz] = '\0';
1.203     schwarze 2524:
                   2525:        prev = &r->regtab;
                   2526:        while (1) {
                   2527:                reg = *prev;
1.238     schwarze 2528:                if (reg == NULL || !strcmp(name, reg->key.p))
1.203     schwarze 2529:                        break;
                   2530:                prev = &reg->next;
                   2531:        }
1.238     schwarze 2532:        if (reg != NULL) {
1.203     schwarze 2533:                *prev = reg->next;
                   2534:                free(reg->key.p);
                   2535:                free(reg);
                   2536:        }
1.277     schwarze 2537:        return ROFF_IGN;
1.122     schwarze 2538: }
                   2539:
1.266     schwarze 2540: /* --- handler functions for roff requests -------------------------------- */
                   2541:
1.122     schwarze 2542: static enum rofferr
                   2543: roff_rm(ROFF_ARGS)
                   2544: {
                   2545:        const char       *name;
                   2546:        char             *cp;
1.212     schwarze 2547:        size_t            namesz;
1.122     schwarze 2548:
1.238     schwarze 2549:        cp = buf->buf + pos;
                   2550:        while (*cp != '\0') {
1.212     schwarze 2551:                name = cp;
1.238     schwarze 2552:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212     schwarze 2553:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238     schwarze 2554:                if (name[namesz] == '\\')
1.212     schwarze 2555:                        break;
1.122     schwarze 2556:        }
1.277     schwarze 2557:        return ROFF_IGN;
1.178     schwarze 2558: }
                   2559:
                   2560: static enum rofferr
                   2561: roff_it(ROFF_ARGS)
                   2562: {
                   2563:        int              iv;
                   2564:
                   2565:        /* Parse the number of lines. */
1.261     schwarze 2566:
                   2567:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222     schwarze 2568:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238     schwarze 2569:                    ln, ppos, buf->buf + 1);
1.277     schwarze 2570:                return ROFF_IGN;
1.178     schwarze 2571:        }
                   2572:
1.262     schwarze 2573:        while (isspace((unsigned char)buf->buf[pos]))
                   2574:                pos++;
                   2575:
                   2576:        /*
                   2577:         * Arm the input line trap.
                   2578:         * Special-casing "an-trap" is an ugly workaround to cope
                   2579:         * with DocBook stupidly fiddling with man(7) internals.
                   2580:         */
1.261     schwarze 2581:
1.178     schwarze 2582:        roffit_lines = iv;
1.262     schwarze 2583:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2584:            strcmp(buf->buf + pos, "an-trap") ?
                   2585:            buf->buf + pos : "br");
1.277     schwarze 2586:        return ROFF_IGN;
1.175     schwarze 2587: }
                   2588:
                   2589: static enum rofferr
                   2590: roff_Dd(ROFF_ARGS)
                   2591: {
                   2592:        const char *const       *cp;
                   2593:
1.227     schwarze 2594:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175     schwarze 2595:                for (cp = __mdoc_reserved; *cp; cp++)
                   2596:                        roff_setstr(r, *cp, NULL, 0);
                   2597:
1.227     schwarze 2598:        if (r->format == 0)
                   2599:                r->format = MPARSE_MDOC;
                   2600:
1.277     schwarze 2601:        return ROFF_CONT;
1.175     schwarze 2602: }
                   2603:
                   2604: static enum rofferr
                   2605: roff_TH(ROFF_ARGS)
                   2606: {
                   2607:        const char *const       *cp;
                   2608:
1.227     schwarze 2609:        if ((r->options & MPARSE_QUICK) == 0)
1.175     schwarze 2610:                for (cp = __man_reserved; *cp; cp++)
                   2611:                        roff_setstr(r, *cp, NULL, 0);
                   2612:
1.227     schwarze 2613:        if (r->format == 0)
                   2614:                r->format = MPARSE_MAN;
                   2615:
1.277     schwarze 2616:        return ROFF_CONT;
1.109     kristaps 2617: }
                   2618:
                   2619: static enum rofferr
                   2620: roff_TE(ROFF_ARGS)
                   2621: {
                   2622:
                   2623:        if (NULL == r->tbl)
1.221     schwarze 2624:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2625:                    ln, ppos, "TE");
1.258     schwarze 2626:        else if ( ! tbl_end(&r->tbl)) {
                   2627:                free(buf->buf);
                   2628:                buf->buf = mandoc_strdup(".sp");
                   2629:                buf->sz = 4;
1.277     schwarze 2630:                return ROFF_REPARSE;
1.258     schwarze 2631:        }
1.277     schwarze 2632:        return ROFF_IGN;
1.112     kristaps 2633: }
                   2634:
                   2635: static enum rofferr
                   2636: roff_T_(ROFF_ARGS)
                   2637: {
                   2638:
                   2639:        if (NULL == r->tbl)
1.221     schwarze 2640:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2641:                    ln, ppos, "T&");
1.112     kristaps 2642:        else
1.296   ! schwarze 2643:                tbl_restart(ln, ppos, r->tbl);
1.112     kristaps 2644:
1.277     schwarze 2645:        return ROFF_IGN;
1.109     kristaps 2646: }
                   2647:
1.230     schwarze 2648: /*
                   2649:  * Handle in-line equation delimiters.
                   2650:  */
                   2651: static enum rofferr
1.238     schwarze 2652: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151     kristaps 2653: {
1.233     schwarze 2654:        char            *cp1, *cp2;
                   2655:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151     kristaps 2656:
1.230     schwarze 2657:        /*
                   2658:         * Outside equations, look for an opening delimiter.
                   2659:         * If we are inside an equation, we already know it is
                   2660:         * in-line, or this function wouldn't have been called;
                   2661:         * so look for a closing delimiter.
                   2662:         */
                   2663:
1.238     schwarze 2664:        cp1 = buf->buf + pos;
1.230     schwarze 2665:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2666:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2667:        if (cp2 == NULL)
1.277     schwarze 2668:                return ROFF_CONT;
1.230     schwarze 2669:
1.233     schwarze 2670:        *cp2++ = '\0';
                   2671:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2672:
                   2673:        /* Handle preceding text, protecting whitespace. */
                   2674:
1.238     schwarze 2675:        if (*buf->buf != '\0') {
1.233     schwarze 2676:                if (r->eqn == NULL)
                   2677:                        bef_pr = "\\&";
                   2678:                bef_nl = "\n";
                   2679:        }
                   2680:
                   2681:        /*
                   2682:         * Prepare replacing the delimiter with an equation macro
                   2683:         * and drop leading white space from the equation.
                   2684:         */
1.230     schwarze 2685:
1.233     schwarze 2686:        if (r->eqn == NULL) {
                   2687:                while (*cp2 == ' ')
                   2688:                        cp2++;
                   2689:                mac = ".EQ";
                   2690:        } else
                   2691:                mac = ".EN";
                   2692:
                   2693:        /* Handle following text, protecting whitespace. */
                   2694:
                   2695:        if (*cp2 != '\0') {
                   2696:                aft_nl = "\n";
                   2697:                if (r->eqn != NULL)
                   2698:                        aft_pr = "\\&";
                   2699:        }
                   2700:
                   2701:        /* Do the actual replacement. */
                   2702:
1.238     schwarze 2703:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233     schwarze 2704:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238     schwarze 2705:        free(buf->buf);
                   2706:        buf->buf = cp1;
1.230     schwarze 2707:
                   2708:        /* Toggle the in-line state of the eqn subsystem. */
                   2709:
                   2710:        r->eqn_inline = r->eqn == NULL;
1.277     schwarze 2711:        return ROFF_REPARSE;
1.151     kristaps 2712: }
                   2713:
1.235     schwarze 2714: static enum rofferr
                   2715: roff_EQ(ROFF_ARGS)
1.125     kristaps 2716: {
1.151     kristaps 2717:        struct eqn_node *e;
1.125     kristaps 2718:
1.238     schwarze 2719:        assert(r->eqn == NULL);
1.235     schwarze 2720:        e = eqn_alloc(ppos, ln, r->parse);
1.125     kristaps 2721:
1.230     schwarze 2722:        if (r->last_eqn) {
1.125     kristaps 2723:                r->last_eqn->next = e;
1.230     schwarze 2724:                e->delim = r->last_eqn->delim;
                   2725:                e->odelim = r->last_eqn->odelim;
                   2726:                e->cdelim = r->last_eqn->cdelim;
                   2727:        } else
1.125     kristaps 2728:                r->first_eqn = r->last_eqn = e;
                   2729:
                   2730:        r->eqn = r->last_eqn = e;
1.151     kristaps 2731:
1.238     schwarze 2732:        if (buf->buf[pos] != '\0')
1.235     schwarze 2733:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238     schwarze 2734:                    ".EQ %s", buf->buf + pos);
1.151     kristaps 2735:
1.277     schwarze 2736:        return ROFF_IGN;
1.125     kristaps 2737: }
                   2738:
                   2739: static enum rofferr
                   2740: roff_EN(ROFF_ARGS)
                   2741: {
                   2742:
1.221     schwarze 2743:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.277     schwarze 2744:        return ROFF_IGN;
1.125     kristaps 2745: }
                   2746:
                   2747: static enum rofferr
1.109     kristaps 2748: roff_TS(ROFF_ARGS)
                   2749: {
1.176     schwarze 2750:        struct tbl_node *tbl;
1.89      kristaps 2751:
1.115     kristaps 2752:        if (r->tbl) {
1.221     schwarze 2753:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2754:                    ln, ppos, "TS breaks TS");
1.151     kristaps 2755:                tbl_end(&r->tbl);
1.115     kristaps 2756:        }
1.83      schwarze 2757:
1.176     schwarze 2758:        tbl = tbl_alloc(ppos, ln, r->parse);
1.113     kristaps 2759:
                   2760:        if (r->last_tbl)
1.176     schwarze 2761:                r->last_tbl->next = tbl;
1.113     kristaps 2762:        else
1.176     schwarze 2763:                r->first_tbl = r->last_tbl = tbl;
1.113     kristaps 2764:
1.176     schwarze 2765:        r->tbl = r->last_tbl = tbl;
1.277     schwarze 2766:        return ROFF_IGN;
1.251     schwarze 2767: }
                   2768:
                   2769: static enum rofferr
1.296   ! schwarze 2770: roff_br(ROFF_ARGS)
1.251     schwarze 2771: {
1.296   ! schwarze 2772:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
        !          2773:        if (buf->buf[pos] != '\0')
        !          2774:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
        !          2775:                    "%s %s", roff_name[tok], buf->buf + pos);
        !          2776:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
        !          2777:        r->man->next = ROFF_NEXT_SIBLING;
        !          2778:        return ROFF_IGN;
1.92      schwarze 2779: }
                   2780:
1.105     kristaps 2781: static enum rofferr
1.174     kristaps 2782: roff_cc(ROFF_ARGS)
                   2783: {
                   2784:        const char      *p;
                   2785:
1.238     schwarze 2786:        p = buf->buf + pos;
1.174     kristaps 2787:
1.238     schwarze 2788:        if (*p == '\0' || (r->control = *p++) == '.')
1.174     kristaps 2789:                r->control = 0;
                   2790:
1.238     schwarze 2791:        if (*p != '\0')
1.260     schwarze 2792:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   2793:                    ln, p - buf->buf, "cc ... %s", p);
1.174     kristaps 2794:
1.277     schwarze 2795:        return ROFF_IGN;
1.174     kristaps 2796: }
                   2797:
                   2798: static enum rofferr
1.164     kristaps 2799: roff_tr(ROFF_ARGS)
                   2800: {
                   2801:        const char      *p, *first, *second;
                   2802:        size_t           fsz, ssz;
                   2803:        enum mandoc_esc  esc;
                   2804:
1.238     schwarze 2805:        p = buf->buf + pos;
1.164     kristaps 2806:
1.238     schwarze 2807:        if (*p == '\0') {
1.260     schwarze 2808:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277     schwarze 2809:                return ROFF_IGN;
1.164     kristaps 2810:        }
                   2811:
1.238     schwarze 2812:        while (*p != '\0') {
1.164     kristaps 2813:                fsz = ssz = 1;
                   2814:
                   2815:                first = p++;
1.238     schwarze 2816:                if (*first == '\\') {
1.164     kristaps 2817:                        esc = mandoc_escape(&p, NULL, NULL);
1.238     schwarze 2818:                        if (esc == ESCAPE_ERROR) {
1.219     schwarze 2819:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238     schwarze 2820:                                    ln, (int)(p - buf->buf), first);
1.277     schwarze 2821:                                return ROFF_IGN;
1.164     kristaps 2822:                        }
                   2823:                        fsz = (size_t)(p - first);
                   2824:                }
                   2825:
                   2826:                second = p++;
1.238     schwarze 2827:                if (*second == '\\') {
1.164     kristaps 2828:                        esc = mandoc_escape(&p, NULL, NULL);
1.238     schwarze 2829:                        if (esc == ESCAPE_ERROR) {
1.219     schwarze 2830:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238     schwarze 2831:                                    ln, (int)(p - buf->buf), second);
1.277     schwarze 2832:                                return ROFF_IGN;
1.164     kristaps 2833:                        }
                   2834:                        ssz = (size_t)(p - second);
1.238     schwarze 2835:                } else if (*second == '\0') {
1.260     schwarze 2836:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   2837:                            ln, first - buf->buf, "tr %s", first);
1.164     kristaps 2838:                        second = " ";
1.165     kristaps 2839:                        p--;
1.164     kristaps 2840:                }
                   2841:
1.167     kristaps 2842:                if (fsz > 1) {
1.207     schwarze 2843:                        roff_setstrn(&r->xmbtab, first, fsz,
                   2844:                            second, ssz, 0);
1.167     kristaps 2845:                        continue;
                   2846:                }
                   2847:
1.238     schwarze 2848:                if (r->xtab == NULL)
1.207     schwarze 2849:                        r->xtab = mandoc_calloc(128,
                   2850:                            sizeof(struct roffstr));
1.167     kristaps 2851:
                   2852:                free(r->xtab[(int)*first].p);
                   2853:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   2854:                r->xtab[(int)*first].sz = ssz;
1.164     kristaps 2855:        }
                   2856:
1.277     schwarze 2857:        return ROFF_IGN;
1.164     kristaps 2858: }
                   2859:
                   2860: static enum rofferr
1.105     kristaps 2861: roff_so(ROFF_ARGS)
                   2862: {
1.249     schwarze 2863:        char *name, *cp;
1.105     kristaps 2864:
1.238     schwarze 2865:        name = buf->buf + pos;
1.224     schwarze 2866:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105     kristaps 2867:
                   2868:        /*
                   2869:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   2870:         * opening anything that's not in our cwd or anything beneath
                   2871:         * it.  Thus, explicitly disallow traversing up the file-system
                   2872:         * or using absolute paths.
                   2873:         */
                   2874:
1.238     schwarze 2875:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210     schwarze 2876:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   2877:                    ".so %s", name);
1.249     schwarze 2878:                buf->sz = mandoc_asprintf(&cp,
                   2879:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   2880:                free(buf->buf);
                   2881:                buf->buf = cp;
                   2882:                *offs = 0;
1.277     schwarze 2883:                return ROFF_REPARSE;
1.105     kristaps 2884:        }
                   2885:
                   2886:        *offs = pos;
1.277     schwarze 2887:        return ROFF_SO;
1.105     kristaps 2888: }
1.92      schwarze 2889:
1.266     schwarze 2890: /* --- user defined strings and macros ------------------------------------ */
                   2891:
1.106     kristaps 2892: static enum rofferr
                   2893: roff_userdef(ROFF_ARGS)
1.99      kristaps 2894: {
1.263     schwarze 2895:        const char       *arg[9], *ap;
1.106     kristaps 2896:        char             *cp, *n1, *n2;
1.292     schwarze 2897:        int               expand_count, i, ib, ie;
1.263     schwarze 2898:        size_t            asz, rsz;
1.106     kristaps 2899:
                   2900:        /*
                   2901:         * Collect pointers to macro argument strings
1.188     schwarze 2902:         * and NUL-terminate them.
1.106     kristaps 2903:         */
1.263     schwarze 2904:
1.273     schwarze 2905:        r->argc = 0;
1.238     schwarze 2906:        cp = buf->buf + pos;
1.273     schwarze 2907:        for (i = 0; i < 9; i++) {
                   2908:                if (*cp == '\0')
                   2909:                        arg[i] = "";
                   2910:                else {
                   2911:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   2912:                        r->argc = i + 1;
                   2913:                }
                   2914:        }
1.99      kristaps 2915:
1.106     kristaps 2916:        /*
                   2917:         * Expand macro arguments.
1.99      kristaps 2918:         */
1.263     schwarze 2919:
                   2920:        buf->sz = strlen(r->current_string) + 1;
1.292     schwarze 2921:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.263     schwarze 2922:        memcpy(n1, r->current_string, buf->sz);
1.292     schwarze 2923:        expand_count = 0;
1.263     schwarze 2924:        while (*cp != '\0') {
                   2925:
                   2926:                /* Scan ahead for the next argument invocation. */
                   2927:
                   2928:                if (*cp++ != '\\')
                   2929:                        continue;
                   2930:                if (*cp++ != '$')
                   2931:                        continue;
1.274     schwarze 2932:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   2933:                        ib = 0;
                   2934:                        ie = r->argc - 1;
                   2935:                } else {  /* \\$1 .. \\$9 insert one argument */
                   2936:                        ib = ie = *cp - '1';
                   2937:                        if (ib < 0 || ib > 8)
                   2938:                                continue;
                   2939:                }
1.263     schwarze 2940:                cp -= 2;
1.292     schwarze 2941:
                   2942:                /*
                   2943:                 * Prevent infinite recursion.
                   2944:                 */
                   2945:
                   2946:                if (cp >= n2)
                   2947:                        expand_count = 1;
                   2948:                else if (++expand_count > EXPAND_LIMIT) {
                   2949:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   2950:                            ln, (int)(cp - n1), NULL);
1.293     schwarze 2951:                        free(buf->buf);
                   2952:                        buf->buf = n1;
1.292     schwarze 2953:                        return ROFF_IGN;
                   2954:                }
1.263     schwarze 2955:
                   2956:                /*
                   2957:                 * Determine the size of the expanded argument,
                   2958:                 * taking escaping of quotes into account.
                   2959:                 */
                   2960:
1.274     schwarze 2961:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   2962:                for (i = ib; i <= ie; i++) {
                   2963:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   2964:                                asz++;
                   2965:                                if (*ap == '"')
                   2966:                                        asz += 3;
                   2967:                        }
1.263     schwarze 2968:                }
                   2969:                if (asz != 3) {
                   2970:
                   2971:                        /*
                   2972:                         * Determine the size of the rest of the
                   2973:                         * unexpanded macro, including the NUL.
                   2974:                         */
                   2975:
                   2976:                        rsz = buf->sz - (cp - n1) - 3;
                   2977:
                   2978:                        /*
                   2979:                         * When shrinking, move before
                   2980:                         * releasing the storage.
                   2981:                         */
                   2982:
                   2983:                        if (asz < 3)
                   2984:                                memmove(cp + asz, cp + 3, rsz);
                   2985:
                   2986:                        /*
                   2987:                         * Resize the storage for the macro
                   2988:                         * and readjust the parse pointer.
                   2989:                         */
                   2990:
                   2991:                        buf->sz += asz - 3;
                   2992:                        n2 = mandoc_realloc(n1, buf->sz);
                   2993:                        cp = n2 + (cp - n1);
                   2994:                        n1 = n2;
                   2995:
                   2996:                        /*
                   2997:                         * When growing, make room
                   2998:                         * for the expanded argument.
                   2999:                         */
                   3000:
                   3001:                        if (asz > 3)
                   3002:                                memmove(cp + asz, cp + 3, rsz);
                   3003:                }
                   3004:
                   3005:                /* Copy the expanded argument, escaping quotes. */
                   3006:
                   3007:                n2 = cp;
1.274     schwarze 3008:                for (i = ib; i <= ie; i++) {
                   3009:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3010:                                if (*ap == '"') {
                   3011:                                        memcpy(n2, "\\(dq", 4);
                   3012:                                        n2 += 4;
                   3013:                                } else
                   3014:                                        *n2++ = *ap;
                   3015:                        }
                   3016:                        if (i < ie)
                   3017:                                *n2++ = ' ';
1.106     kristaps 3018:                }
1.99      kristaps 3019:        }
                   3020:
1.106     kristaps 3021:        /*
                   3022:         * Replace the macro invocation
                   3023:         * by the expanded macro.
                   3024:         */
1.263     schwarze 3025:
1.238     schwarze 3026:        free(buf->buf);
                   3027:        buf->buf = n1;
1.248     schwarze 3028:        *offs = 0;
1.106     kristaps 3029:
1.277     schwarze 3030:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3031:           ROFF_REPARSE : ROFF_APPEND;
1.99      kristaps 3032: }
1.121     schwarze 3033:
1.212     schwarze 3034: static size_t
1.121     schwarze 3035: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3036: {
                   3037:        char     *name, *cp;
1.212     schwarze 3038:        size_t    namesz;
1.121     schwarze 3039:
                   3040:        name = *cpp;
                   3041:        if ('\0' == *name)
1.277     schwarze 3042:                return 0;
1.121     schwarze 3043:
1.212     schwarze 3044:        /* Read until end of name and terminate it with NUL. */
                   3045:        for (cp = name; 1; cp++) {
                   3046:                if ('\0' == *cp || ' ' == *cp) {
                   3047:                        namesz = cp - name;
                   3048:                        break;
                   3049:                }
1.121     schwarze 3050:                if ('\\' != *cp)
                   3051:                        continue;
1.215     schwarze 3052:                namesz = cp - name;
                   3053:                if ('{' == cp[1] || '}' == cp[1])
                   3054:                        break;
1.121     schwarze 3055:                cp++;
                   3056:                if ('\\' == *cp)
                   3057:                        continue;
1.224     schwarze 3058:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3059:                    "%.*s", (int)(cp - name + 1), name);
1.212     schwarze 3060:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3061:                break;
1.121     schwarze 3062:        }
                   3063:
                   3064:        /* Read past spaces. */
                   3065:        while (' ' == *cp)
                   3066:                cp++;
                   3067:
                   3068:        *cpp = cp;
1.277     schwarze 3069:        return namesz;
1.121     schwarze 3070: }
                   3071:
1.106     kristaps 3072: /*
                   3073:  * Store *string into the user-defined string called *name.
                   3074:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193     schwarze 3075:  * append == 0: replace mode
                   3076:  * append == 1: single-line append mode
                   3077:  * append == 2: multiline append mode, append '\n' after each call
1.106     kristaps 3078:  */
1.94      kristaps 3079: static void
1.106     kristaps 3080: roff_setstr(struct roff *r, const char *name, const char *string,
1.193     schwarze 3081:        int append)
1.92      schwarze 3082: {
1.164     kristaps 3083:
                   3084:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.207     schwarze 3085:            string ? strlen(string) : 0, append);
1.164     kristaps 3086: }
                   3087:
                   3088: static void
1.166     kristaps 3089: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193     schwarze 3090:                const char *string, size_t stringsz, int append)
1.164     kristaps 3091: {
1.166     kristaps 3092:        struct roffkv   *n;
1.164     kristaps 3093:        char            *c;
                   3094:        int              i;
                   3095:        size_t           oldch, newch;
1.92      schwarze 3096:
1.106     kristaps 3097:        /* Search for an existing string with the same name. */
1.164     kristaps 3098:        n = *r;
                   3099:
1.211     schwarze 3100:        while (n && (namesz != n->key.sz ||
                   3101:                        strncmp(n->key.p, name, namesz)))
1.92      schwarze 3102:                n = n->next;
1.94      kristaps 3103:
                   3104:        if (NULL == n) {
1.106     kristaps 3105:                /* Create a new string table entry. */
1.166     kristaps 3106:                n = mandoc_malloc(sizeof(struct roffkv));
                   3107:                n->key.p = mandoc_strndup(name, namesz);
                   3108:                n->key.sz = namesz;
                   3109:                n->val.p = NULL;
                   3110:                n->val.sz = 0;
1.164     kristaps 3111:                n->next = *r;
                   3112:                *r = n;
1.193     schwarze 3113:        } else if (0 == append) {
1.166     kristaps 3114:                free(n->val.p);
                   3115:                n->val.p = NULL;
                   3116:                n->val.sz = 0;
1.106     kristaps 3117:        }
                   3118:
                   3119:        if (NULL == string)
                   3120:                return;
                   3121:
                   3122:        /*
                   3123:         * One additional byte for the '\n' in multiline mode,
                   3124:         * and one for the terminating '\0'.
                   3125:         */
1.193     schwarze 3126:        newch = stringsz + (1 < append ? 2u : 1u);
1.164     kristaps 3127:
1.166     kristaps 3128:        if (NULL == n->val.p) {
                   3129:                n->val.p = mandoc_malloc(newch);
                   3130:                *n->val.p = '\0';
1.106     kristaps 3131:                oldch = 0;
                   3132:        } else {
1.166     kristaps 3133:                oldch = n->val.sz;
                   3134:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106     kristaps 3135:        }
                   3136:
                   3137:        /* Skip existing content in the destination buffer. */
1.166     kristaps 3138:        c = n->val.p + (int)oldch;
1.106     kristaps 3139:
                   3140:        /* Append new content to the destination buffer. */
1.164     kristaps 3141:        i = 0;
                   3142:        while (i < (int)stringsz) {
1.106     kristaps 3143:                /*
                   3144:                 * Rudimentary roff copy mode:
                   3145:                 * Handle escaped backslashes.
                   3146:                 */
1.164     kristaps 3147:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3148:                        i++;
                   3149:                *c++ = string[i++];
1.106     kristaps 3150:        }
1.94      kristaps 3151:
1.106     kristaps 3152:        /* Append terminating bytes. */
1.193     schwarze 3153:        if (1 < append)
1.106     kristaps 3154:                *c++ = '\n';
1.163     kristaps 3155:
1.106     kristaps 3156:        *c = '\0';
1.166     kristaps 3157:        n->val.sz = (int)(c - n->val.p);
1.92      schwarze 3158: }
                   3159:
1.94      kristaps 3160: static const char *
                   3161: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92      schwarze 3162: {
1.166     kristaps 3163:        const struct roffkv *n;
1.191     schwarze 3164:        int i;
1.92      schwarze 3165:
1.164     kristaps 3166:        for (n = r->strtab; n; n = n->next)
1.207     schwarze 3167:                if (0 == strncmp(name, n->key.p, len) &&
                   3168:                    '\0' == n->key.p[(int)len])
1.277     schwarze 3169:                        return n->val.p;
1.191     schwarze 3170:
                   3171:        for (i = 0; i < PREDEFS_MAX; i++)
                   3172:                if (0 == strncmp(name, predefs[i].name, len) &&
                   3173:                                '\0' == predefs[i].name[(int)len])
1.277     schwarze 3174:                        return predefs[i].str;
1.94      kristaps 3175:
1.277     schwarze 3176:        return NULL;
1.92      schwarze 3177: }
                   3178:
1.94      kristaps 3179: static void
1.167     kristaps 3180: roff_freestr(struct roffkv *r)
1.92      schwarze 3181: {
1.166     kristaps 3182:        struct roffkv    *n, *nn;
1.92      schwarze 3183:
1.167     kristaps 3184:        for (n = r; n; n = nn) {
1.166     kristaps 3185:                free(n->key.p);
                   3186:                free(n->val.p);
1.92      schwarze 3187:                nn = n->next;
                   3188:                free(n);
                   3189:        }
1.114     kristaps 3190: }
1.266     schwarze 3191:
                   3192: /* --- accessors and utility functions ------------------------------------ */
1.114     kristaps 3193:
                   3194: const struct tbl_span *
                   3195: roff_span(const struct roff *r)
                   3196: {
1.207     schwarze 3197:
1.277     schwarze 3198:        return r->tbl ? tbl_span(r->tbl) : NULL;
1.125     kristaps 3199: }
                   3200:
                   3201: const struct eqn *
                   3202: roff_eqn(const struct roff *r)
                   3203: {
1.207     schwarze 3204:
1.277     schwarze 3205:        return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.164     kristaps 3206: }
                   3207:
                   3208: /*
                   3209:  * Duplicate an input string, making the appropriate character
                   3210:  * conversations (as stipulated by `tr') along the way.
                   3211:  * Returns a heap-allocated string with all the replacements made.
                   3212:  */
                   3213: char *
                   3214: roff_strdup(const struct roff *r, const char *p)
                   3215: {
1.166     kristaps 3216:        const struct roffkv *cp;
1.164     kristaps 3217:        char            *res;
                   3218:        const char      *pp;
                   3219:        size_t           ssz, sz;
                   3220:        enum mandoc_esc  esc;
                   3221:
1.167     kristaps 3222:        if (NULL == r->xmbtab && NULL == r->xtab)
1.277     schwarze 3223:                return mandoc_strdup(p);
1.164     kristaps 3224:        else if ('\0' == *p)
1.277     schwarze 3225:                return mandoc_strdup("");
1.164     kristaps 3226:
                   3227:        /*
                   3228:         * Step through each character looking for term matches
                   3229:         * (remember that a `tr' can be invoked with an escape, which is
                   3230:         * a glyph but the escape is multi-character).
                   3231:         * We only do this if the character hash has been initialised
                   3232:         * and the string is >0 length.
                   3233:         */
                   3234:
                   3235:        res = NULL;
                   3236:        ssz = 0;
                   3237:
                   3238:        while ('\0' != *p) {
1.289     schwarze 3239:                assert((unsigned int)*p < 128);
                   3240:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167     kristaps 3241:                        sz = r->xtab[(int)*p].sz;
                   3242:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3243:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3244:                        ssz += sz;
                   3245:                        p++;
                   3246:                        continue;
                   3247:                } else if ('\\' != *p) {
                   3248:                        res = mandoc_realloc(res, ssz + 2);
                   3249:                        res[ssz++] = *p++;
                   3250:                        continue;
                   3251:                }
                   3252:
1.164     kristaps 3253:                /* Search for term matches. */
1.167     kristaps 3254:                for (cp = r->xmbtab; cp; cp = cp->next)
1.166     kristaps 3255:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164     kristaps 3256:                                break;
                   3257:
                   3258:                if (NULL != cp) {
                   3259:                        /*
                   3260:                         * A match has been found.
                   3261:                         * Append the match to the array and move
                   3262:                         * forward by its keysize.
                   3263:                         */
1.207     schwarze 3264:                        res = mandoc_realloc(res,
                   3265:                            ssz + cp->val.sz + 1);
1.166     kristaps 3266:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3267:                        ssz += cp->val.sz;
                   3268:                        p += (int)cp->key.sz;
1.164     kristaps 3269:                        continue;
                   3270:                }
                   3271:
1.167     kristaps 3272:                /*
                   3273:                 * Handle escapes carefully: we need to copy
                   3274:                 * over just the escape itself, or else we might
                   3275:                 * do replacements within the escape itself.
                   3276:                 * Make sure to pass along the bogus string.
                   3277:                 */
                   3278:                pp = p++;
                   3279:                esc = mandoc_escape(&p, NULL, NULL);
                   3280:                if (ESCAPE_ERROR == esc) {
                   3281:                        sz = strlen(pp);
1.164     kristaps 3282:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3283:                        memcpy(res + ssz, pp, sz);
1.167     kristaps 3284:                        break;
1.164     kristaps 3285:                }
1.207     schwarze 3286:                /*
                   3287:                 * We bail out on bad escapes.
1.167     kristaps 3288:                 * No need to warn: we already did so when
                   3289:                 * roff_res() was called.
                   3290:                 */
                   3291:                sz = (int)(p - pp);
                   3292:                res = mandoc_realloc(res, ssz + sz + 1);
                   3293:                memcpy(res + ssz, pp, sz);
                   3294:                ssz += sz;
1.164     kristaps 3295:        }
                   3296:
                   3297:        res[(int)ssz] = '\0';
1.277     schwarze 3298:        return res;
1.227     schwarze 3299: }
                   3300:
                   3301: int
                   3302: roff_getformat(const struct roff *r)
                   3303: {
                   3304:
1.277     schwarze 3305:        return r->format;
1.174     kristaps 3306: }
                   3307:
                   3308: /*
1.207     schwarze 3309:  * Find out whether a line is a macro line or not.
1.174     kristaps 3310:  * If it is, adjust the current position and return one; if it isn't,
                   3311:  * return zero and don't change the current position.
                   3312:  * If the control character has been set with `.cc', then let that grain
                   3313:  * precedence.
                   3314:  * This is slighly contrary to groff, where using the non-breaking
                   3315:  * control character when `cc' has been invoked will cause the
                   3316:  * non-breaking macro contents to be printed verbatim.
                   3317:  */
                   3318: int
                   3319: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3320: {
                   3321:        int             pos;
                   3322:
                   3323:        pos = *ppos;
                   3324:
                   3325:        if (0 != r->control && cp[pos] == r->control)
                   3326:                pos++;
                   3327:        else if (0 != r->control)
1.277     schwarze 3328:                return 0;
1.174     kristaps 3329:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3330:                pos += 2;
                   3331:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3332:                pos++;
                   3333:        else
1.277     schwarze 3334:                return 0;
1.174     kristaps 3335:
                   3336:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3337:                pos++;
                   3338:
                   3339:        *ppos = pos;
1.277     schwarze 3340:        return 1;
1.74      kristaps 3341: }

CVSweb