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

Annotation of mandoc/roff.c, Revision 1.297

1.297   ! schwarze    1: /*     $Id: roff.c,v 1.296 2017/05/04 17:48:29 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);
1.297   ! schwarze  185: static enum rofferr     roff_onearg(ROFF_ARGS);
1.294     schwarze  186: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.214     schwarze  187:                                int, int);
1.238     schwarze  188: static enum rofferr     roff_parsetext(struct buf *, int, int *);
                    189: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.122     schwarze  190: static enum rofferr     roff_rm(ROFF_ARGS);
1.203     schwarze  191: static enum rofferr     roff_rr(ROFF_ARGS);
1.94      kristaps  192: static void             roff_setstr(struct roff *,
1.106     kristaps  193:                                const char *, const char *, int);
1.207     schwarze  194: static void             roff_setstrn(struct roffkv **, const char *,
1.164     kristaps  195:                                size_t, const char *, size_t, int);
1.105     kristaps  196: static enum rofferr     roff_so(ROFF_ARGS);
1.164     kristaps  197: static enum rofferr     roff_tr(ROFF_ARGS);
1.175     schwarze  198: static enum rofferr     roff_Dd(ROFF_ARGS);
                    199: static enum rofferr     roff_TH(ROFF_ARGS);
1.109     kristaps  200: static enum rofferr     roff_TE(ROFF_ARGS);
                    201: static enum rofferr     roff_TS(ROFF_ARGS);
1.125     kristaps  202: static enum rofferr     roff_EQ(ROFF_ARGS);
                    203: static enum rofferr     roff_EN(ROFF_ARGS);
1.112     kristaps  204: static enum rofferr     roff_T_(ROFF_ARGS);
1.251     schwarze  205: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.106     kristaps  206: static enum rofferr     roff_userdef(ROFF_ARGS);
1.67      kristaps  207:
1.266     schwarze  208: /* --- constant data ------------------------------------------------------ */
                    209:
1.261     schwarze  210: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    211: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    212:
1.294     schwarze  213: const char *__roff_name[MAN_MAX + 1] = {
1.297   ! schwarze  214:        "br",           "ft",           NULL,
1.294     schwarze  215:        "ab",           "ad",           "af",           "aln",
                    216:        "als",          "am",           "am1",          "ami",
                    217:        "ami1",         "as",           "as1",          "asciify",
                    218:        "backtrace",    "bd",           "bleedat",      "blm",
                    219:         "box",         "boxa",         "bp",           "BP",
                    220:        "break",        "breakchar",    "brnl",         "brp",
                    221:        "brpnl",        "c2",           "cc",           "ce",
                    222:        "cf",           "cflags",       "ch",           "char",
                    223:        "chop",         "class",        "close",        "CL",
                    224:        "color",        "composite",    "continue",     "cp",
                    225:        "cropat",       "cs",           "cu",           "da",
                    226:        "dch",          "Dd",           "de",           "de1",
                    227:        "defcolor",     "dei",          "dei1",         "device",
                    228:        "devicem",      "di",           "do",           "ds",
                    229:        "ds1",          "dwh",          "dt",           "ec",
                    230:        "ecr",          "ecs",          "el",           "em",
                    231:        "EN",           "eo",           "EP",           "EQ",
                    232:        "errprint",     "ev",           "evc",          "ex",
                    233:        "fallback",     "fam",          "fc",           "fchar",
                    234:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    235:        "fl",           "flig",         "fp",           "fps",
                    236:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    237:        "fzoom",        "gcolor",       "hc",           "hcode",
                    238:        "hidechar",     "hla",          "hlm",          "hpf",
                    239:        "hpfa",         "hpfcode",      "hw",           "hy",
                    240:        "hylang",       "hylen",        "hym",          "hypp",
                    241:        "hys",          "ie",           "if",           "ig",
                    242:        "index",        "it",           "itc",          "IX",
                    243:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    244:        "lc",           "lc_ctype",     "lds",          "length",
                    245:        "letadj",       "lf",           "lg",           "lhang",
                    246:        "linetabs",     "lnr",          "lnrf",         "lpfx",
                    247:        "ls",           "lsm",          "lt",           "mc",
                    248:        "mediasize",    "minss",        "mk",           "mso",
                    249:        "na",           "ne",           "nh",           "nhychar",
                    250:        "nm",           "nn",           "nop",          "nr",
                    251:        "nrf",          "nroff",        "ns",           "nx",
                    252:        "open",         "opena",        "os",           "output",
                    253:        "padj",         "papersize",    "pc",           "pev",
                    254:        "pi",           "PI",           "pl",           "pm",
                    255:        "pn",           "pnr",          "po",           "ps",
                    256:        "psbb",         "pshape",       "pso",          "ptr",
                    257:        "pvs",          "rchar",        "rd",           "recursionlimit",
                    258:        "return",       "rfschar",      "rhang",        "rj",
                    259:        "rm",           "rn",           "rnn",          "rr",
                    260:        "rs",           "rt",           "schar",        "sentchar",
                    261:        "shc",          "shift",        "sizes",        "so",
                    262:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    263:        "sty",          "substring",    "sv",           "sy",
                    264:        "T&",           "ta",           "tc",           "TE",
                    265:        "TH",           "ti",           "tkf",          "tl",
                    266:        "tm",           "tm1",          "tmc",          "tr",
                    267:        "track",        "transchar",    "trf",          "trimat",
                    268:        "trin",         "trnt",         "troff",        "TS",
                    269:        "uf",           "ul",           "unformat",     "unwatch",
                    270:        "unwatchn",     "vpt",          "vs",           "warn",
                    271:        "warnscale",    "watch",        "watchlength",  "watchn",
                    272:        "wh",           "while",        "write",        "writec",
                    273:        "writem",       "xflag",        ".",            NULL,
                    274:        "text",
                    275:        "Dd",           "Dt",           "Os",           "Sh",
                    276:        "Ss",           "Pp",           "D1",           "Dl",
                    277:        "Bd",           "Ed",           "Bl",           "El",
                    278:        "It",           "Ad",           "An",           "Ap",
                    279:        "Ar",           "Cd",           "Cm",           "Dv",
                    280:        "Er",           "Ev",           "Ex",           "Fa",
                    281:        "Fd",           "Fl",           "Fn",           "Ft",
                    282:        "Ic",           "In",           "Li",           "Nd",
                    283:        "Nm",           "Op",           "Ot",           "Pa",
                    284:        "Rv",           "St",           "Va",           "Vt",
                    285:        "Xr",           "%A",           "%B",           "%D",
                    286:        "%I",           "%J",           "%N",           "%O",
                    287:        "%P",           "%R",           "%T",           "%V",
                    288:        "Ac",           "Ao",           "Aq",           "At",
                    289:        "Bc",           "Bf",           "Bo",           "Bq",
                    290:        "Bsx",          "Bx",           "Db",           "Dc",
                    291:        "Do",           "Dq",           "Ec",           "Ef",
                    292:        "Em",           "Eo",           "Fx",           "Ms",
                    293:        "No",           "Ns",           "Nx",           "Ox",
                    294:        "Pc",           "Pf",           "Po",           "Pq",
                    295:        "Qc",           "Ql",           "Qo",           "Qq",
                    296:        "Re",           "Rs",           "Sc",           "So",
                    297:        "Sq",           "Sm",           "Sx",           "Sy",
                    298:        "Tn",           "Ux",           "Xc",           "Xo",
                    299:        "Fo",           "Fc",           "Oo",           "Oc",
                    300:        "Bk",           "Ek",           "Bt",           "Hf",
                    301:        "Fr",           "Ud",           "Lb",           "Lp",
                    302:        "Lk",           "Mt",           "Brq",          "Bro",
                    303:        "Brc",          "%C",           "Es",           "En",
1.296     schwarze  304:        "Dx",           "%Q",           "sp",
1.294     schwarze  305:        "%U",           "Ta",           "ll",           NULL,
                    306:        "TH",           "SH",           "SS",           "TP",
                    307:        "LP",           "PP",           "P",            "IP",
                    308:        "HP",           "SM",           "SB",           "BI",
                    309:        "IB",           "BR",           "RB",           "R",
                    310:        "B",            "I",            "IR",           "RI",
1.296     schwarze  311:        "sp",           "nf",           "fi",
1.294     schwarze  312:        "RE",           "RS",           "DT",           "UC",
1.297   ! schwarze  313:        "PD",           "AT",           "in",
1.294     schwarze  314:        "OP",           "EX",           "EE",           "UR",
                    315:        "UE",           "ll",           NULL
                    316: };
                    317: const  char *const *roff_name = __roff_name;
                    318:
                    319: static struct roffmac   roffs[TOKEN_NONE] = {
1.296     schwarze  320:        { roff_br, NULL, NULL, 0 },  /* br */
1.297   ! schwarze  321:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.296     schwarze  322:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.295     schwarze  323:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    324:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    325:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    326:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
                    327:        { roff_unsupp, NULL, NULL, 0 },  /* als */
                    328:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    329:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    330:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    331:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    332:        { roff_ds, NULL, NULL, 0 },  /* as */
                    333:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    334:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    335:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    336:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    337:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    338:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    339:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    340:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    341:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    342:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    343:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    344:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    345:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.296     schwarze  346:        { roff_br, NULL, NULL, 0 },  /* brp */
1.295     schwarze  347:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    348:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    349:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    350:        { roff_line_ignore, NULL, NULL, 0 },  /* ce */
                    351:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    352:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    353:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    354:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    355:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    356:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    357:        { roff_insec, NULL, NULL, 0 },  /* close */
                    358:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    359:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    360:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    361:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    362:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    363:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    364:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    365:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    366:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    367:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    368:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    369:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    370:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    371:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    372:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    373:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    374:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    375:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    376:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    377:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    378:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    379:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    380:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    381:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
                    382:        { roff_unsupp, NULL, NULL, 0 },  /* ec */
                    383:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    384:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    385:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    386:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    387:        { roff_EN, NULL, NULL, 0 },  /* EN */
                    388:        { roff_unsupp, NULL, NULL, 0 },  /* eo */
                    389:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    390:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    391:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    392:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    393:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    394:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    395:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    396:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    397:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    398:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    399:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    400:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    401:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    402:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    403:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    404:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    405:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    406:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    407:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    408:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    409:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    410:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    411:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    412:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    413:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    414:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    415:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    416:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    417:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    419:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    421:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    422:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    425:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    428:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    429:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    430:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    431:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    432:        { roff_it, NULL, NULL, 0 },  /* it */
                    433:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    438:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    439:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    440:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    441:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    442:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    444:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    445:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    446:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    447:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    448:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    449:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    450:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    451:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    452:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    453:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    454:        { roff_line_ignore, NULL, NULL, 0 },  /* mc */
                    455:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    456:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    457:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    458:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    461:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    462:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    463:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    464:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    465:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    466:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    467:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    468:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    469:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    470:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    471:        { roff_insec, NULL, NULL, 0 },  /* open */
                    472:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    474:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    475:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    479:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    480:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* po */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    487:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    488:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    489:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    491:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    492:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    493:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    494:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    495:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    496:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    497:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* rj */
                    499:        { roff_rm, NULL, NULL, 0 },  /* rm */
                    500:        { roff_unsupp, NULL, NULL, 0 },  /* rn */
                    501:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    502:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    503:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    504:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    505:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    506:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    508:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    510:        { roff_so, NULL, NULL, 0 },  /* so */
                    511:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    512:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    513:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    514:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    516:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    517:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    518:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    519:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    520:        { roff_unsupp, NULL, NULL, 0 },  /* ta */
                    521:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    522:        { roff_TE, NULL, NULL, 0 },  /* TE */
                    523:        { roff_TH, NULL, NULL, 0 },  /* TH */
                    524:        { roff_unsupp, NULL, NULL, 0 },  /* ti */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    526:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    529:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    530:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    533:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    534:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    535:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    536:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    537:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    538:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    539:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    540:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    541:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    542:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    543:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    549:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    550:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    551:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    552:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    553:        { roff_insec, NULL, NULL, 0 },  /* write */
                    554:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    555:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    557:        { roff_cblock, NULL, NULL, 0 },  /* . */
                    558:        { roff_userdef, NULL, NULL, 0 }
1.67      kristaps  559: };
                    560:
1.200     schwarze  561: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175     schwarze  562: const  char *const __mdoc_reserved[] = {
                    563:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    564:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    565:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    566:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200     schwarze  567:        "Dt", "Dv", "Dx", "D1",
                    568:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    569:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175     schwarze  570:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200     schwarze  571:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    572:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175     schwarze  573:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200     schwarze  574:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    575:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    576:        "Sc", "Sh", "Sm", "So", "Sq",
1.175     schwarze  577:        "Ss", "St", "Sx", "Sy",
                    578:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200     schwarze  579:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175     schwarze  580:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    581:        NULL
                    582: };
                    583:
1.200     schwarze  584: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175     schwarze  585: const  char *const __man_reserved[] = {
1.200     schwarze  586:        "AT", "B", "BI", "BR", "DT",
                    587:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    588:        "LP", "OP", "P", "PD", "PP",
                    589:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    590:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175     schwarze  591:        NULL
                    592: };
                    593:
1.141     kristaps  594: /* Array of injected predefined strings. */
                    595: #define        PREDEFS_MAX      38
                    596: static const struct predef predefs[PREDEFS_MAX] = {
                    597: #include "predefs.in"
                    598: };
                    599:
1.178     schwarze  600: static int      roffit_lines;  /* number of lines to delay */
                    601: static char    *roffit_macro;  /* nil-terminated macro line */
                    602:
1.207     schwarze  603:
1.266     schwarze  604: /* --- request table ------------------------------------------------------ */
                    605:
1.295     schwarze  606: struct ohash *
                    607: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85      kristaps  608: {
1.295     schwarze  609:        struct ohash    *htab;
                    610:        struct roffreq  *req;
                    611:        enum roff_tok    tok;
                    612:        size_t           sz;
                    613:        unsigned int     slot;
1.85      kristaps  614:
1.295     schwarze  615:        htab = mandoc_malloc(sizeof(*htab));
                    616:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85      kristaps  617:
1.295     schwarze  618:        for (tok = mintok; tok < maxtok; tok++) {
1.296     schwarze  619:                if (roff_name[tok] == NULL)
                    620:                        continue;
1.295     schwarze  621:                sz = strlen(roff_name[tok]);
                    622:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    623:                req->tok = tok;
                    624:                memcpy(req->name, roff_name[tok], sz + 1);
                    625:                slot = ohash_qlookup(htab, req->name);
                    626:                ohash_insert(htab, slot, req);
1.85      kristaps  627:        }
1.295     schwarze  628:        return htab;
1.85      kristaps  629: }
                    630:
1.295     schwarze  631: void
                    632: roffhash_free(struct ohash *htab)
1.67      kristaps  633: {
1.295     schwarze  634:        struct roffreq  *req;
                    635:        unsigned int     slot;
1.67      kristaps  636:
1.295     schwarze  637:        if (htab == NULL)
                    638:                return;
                    639:        for (req = ohash_first(htab, &slot); req != NULL;
                    640:             req = ohash_next(htab, &slot))
                    641:                free(req);
                    642:        ohash_delete(htab);
                    643:        free(htab);
                    644: }
                    645:
                    646: enum roff_tok
                    647: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    648: {
                    649:        struct roffreq  *req;
                    650:        const char      *end;
                    651:
                    652:        if (sz) {
                    653:                end = name + sz;
                    654:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    655:        } else
                    656:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    657:        return req == NULL ? TOKEN_NONE : req->tok;
1.67      kristaps  658: }
                    659:
1.266     schwarze  660: /* --- stack of request blocks -------------------------------------------- */
                    661:
1.67      kristaps  662: /*
                    663:  * Pop the current node off of the stack of roff instructions currently
                    664:  * pending.
                    665:  */
                    666: static void
                    667: roffnode_pop(struct roff *r)
                    668: {
                    669:        struct roffnode *p;
                    670:
1.75      kristaps  671:        assert(r->last);
1.207     schwarze  672:        p = r->last;
1.82      kristaps  673:
1.75      kristaps  674:        r->last = r->last->parent;
1.106     kristaps  675:        free(p->name);
                    676:        free(p->end);
1.67      kristaps  677:        free(p);
                    678: }
                    679:
                    680: /*
                    681:  * Push a roff node onto the instruction stack.  This must later be
                    682:  * removed with roffnode_pop().
                    683:  */
1.98      schwarze  684: static void
1.294     schwarze  685: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106     kristaps  686:                int line, int col)
1.67      kristaps  687: {
                    688:        struct roffnode *p;
                    689:
1.98      schwarze  690:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.67      kristaps  691:        p->tok = tok;
1.106     kristaps  692:        if (name)
                    693:                p->name = mandoc_strdup(name);
1.67      kristaps  694:        p->parent = r->last;
                    695:        p->line = line;
                    696:        p->col = col;
1.198     schwarze  697:        p->rule = p->parent ? p->parent->rule : 0;
1.67      kristaps  698:
                    699:        r->last = p;
                    700: }
                    701:
1.266     schwarze  702: /* --- roff parser state data management ---------------------------------- */
                    703:
1.67      kristaps  704: static void
                    705: roff_free1(struct roff *r)
                    706: {
1.176     schwarze  707:        struct tbl_node *tbl;
1.125     kristaps  708:        struct eqn_node *e;
1.167     kristaps  709:        int              i;
1.67      kristaps  710:
1.176     schwarze  711:        while (NULL != (tbl = r->first_tbl)) {
                    712:                r->first_tbl = tbl->next;
                    713:                tbl_free(tbl);
1.109     kristaps  714:        }
1.113     kristaps  715:        r->first_tbl = r->last_tbl = r->tbl = NULL;
                    716:
1.125     kristaps  717:        while (NULL != (e = r->first_eqn)) {
                    718:                r->first_eqn = e->next;
                    719:                eqn_free(e);
                    720:        }
                    721:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    722:
1.67      kristaps  723:        while (r->last)
                    724:                roffnode_pop(r);
1.109     kristaps  725:
1.223     schwarze  726:        free (r->rstack);
                    727:        r->rstack = NULL;
                    728:        r->rstacksz = 0;
                    729:        r->rstackpos = -1;
                    730:
                    731:        roff_freereg(r->regtab);
                    732:        r->regtab = NULL;
                    733:
1.167     kristaps  734:        roff_freestr(r->strtab);
                    735:        roff_freestr(r->xmbtab);
                    736:        r->strtab = r->xmbtab = NULL;
                    737:
                    738:        if (r->xtab)
                    739:                for (i = 0; i < 128; i++)
                    740:                        free(r->xtab[i].p);
                    741:        free(r->xtab);
                    742:        r->xtab = NULL;
1.67      kristaps  743: }
                    744:
                    745: void
                    746: roff_reset(struct roff *r)
                    747: {
                    748:        roff_free1(r);
1.227     schwarze  749:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174     kristaps  750:        r->control = 0;
1.67      kristaps  751: }
                    752:
                    753: void
                    754: roff_free(struct roff *r)
                    755: {
                    756:        roff_free1(r);
1.295     schwarze  757:        roffhash_free(r->reqtab);
1.67      kristaps  758:        free(r);
                    759: }
                    760:
                    761: struct roff *
1.279     schwarze  762: roff_alloc(struct mparse *parse, int options)
1.67      kristaps  763: {
                    764:        struct roff     *r;
                    765:
1.98      schwarze  766:        r = mandoc_calloc(1, sizeof(struct roff));
1.128     kristaps  767:        r->parse = parse;
1.295     schwarze  768:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.199     schwarze  769:        r->options = options;
1.227     schwarze  770:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82      kristaps  771:        r->rstackpos = -1;
1.277     schwarze  772:        return r;
1.265     schwarze  773: }
                    774:
1.266     schwarze  775: /* --- syntax tree state data management ---------------------------------- */
                    776:
1.265     schwarze  777: static void
                    778: roff_man_free1(struct roff_man *man)
                    779: {
                    780:
1.266     schwarze  781:        if (man->first != NULL)
                    782:                roff_node_delete(man, man->first);
1.265     schwarze  783:        free(man->meta.msec);
                    784:        free(man->meta.vol);
                    785:        free(man->meta.os);
                    786:        free(man->meta.arch);
                    787:        free(man->meta.title);
                    788:        free(man->meta.name);
                    789:        free(man->meta.date);
                    790: }
                    791:
                    792: static void
                    793: roff_man_alloc1(struct roff_man *man)
                    794: {
                    795:
                    796:        memset(&man->meta, 0, sizeof(man->meta));
                    797:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    798:        man->first->type = ROFFT_ROOT;
                    799:        man->last = man->first;
                    800:        man->last_es = NULL;
                    801:        man->flags = 0;
                    802:        man->macroset = MACROSET_NONE;
                    803:        man->lastsec = man->lastnamed = SEC_NONE;
                    804:        man->next = ROFF_NEXT_CHILD;
                    805: }
                    806:
                    807: void
                    808: roff_man_reset(struct roff_man *man)
                    809: {
                    810:
                    811:        roff_man_free1(man);
                    812:        roff_man_alloc1(man);
                    813: }
                    814:
                    815: void
                    816: roff_man_free(struct roff_man *man)
                    817: {
                    818:
                    819:        roff_man_free1(man);
                    820:        free(man);
                    821: }
                    822:
                    823: struct roff_man *
                    824: roff_man_alloc(struct roff *roff, struct mparse *parse,
                    825:        const char *defos, int quick)
                    826: {
                    827:        struct roff_man *man;
                    828:
                    829:        man = mandoc_calloc(1, sizeof(*man));
                    830:        man->parse = parse;
                    831:        man->roff = roff;
                    832:        man->defos = defos;
                    833:        man->quick = quick;
                    834:        roff_man_alloc1(man);
1.296     schwarze  835:        roff->man = man;
1.277     schwarze  836:        return man;
1.67      kristaps  837: }
                    838:
1.266     schwarze  839: /* --- syntax tree handling ----------------------------------------------- */
                    840:
                    841: struct roff_node *
                    842: roff_node_alloc(struct roff_man *man, int line, int pos,
                    843:        enum roff_type type, int tok)
                    844: {
                    845:        struct roff_node        *n;
                    846:
                    847:        n = mandoc_calloc(1, sizeof(*n));
                    848:        n->line = line;
                    849:        n->pos = pos;
                    850:        n->tok = tok;
                    851:        n->type = type;
                    852:        n->sec = man->lastsec;
                    853:
                    854:        if (man->flags & MDOC_SYNOPSIS)
1.285     schwarze  855:                n->flags |= NODE_SYNPRETTY;
1.266     schwarze  856:        else
1.285     schwarze  857:                n->flags &= ~NODE_SYNPRETTY;
1.266     schwarze  858:        if (man->flags & MDOC_NEWLINE)
1.285     schwarze  859:                n->flags |= NODE_LINE;
1.266     schwarze  860:        man->flags &= ~MDOC_NEWLINE;
                    861:
1.277     schwarze  862:        return n;
1.266     schwarze  863: }
                    864:
                    865: void
                    866: roff_node_append(struct roff_man *man, struct roff_node *n)
                    867: {
                    868:
                    869:        switch (man->next) {
                    870:        case ROFF_NEXT_SIBLING:
1.281     schwarze  871:                if (man->last->next != NULL) {
                    872:                        n->next = man->last->next;
                    873:                        man->last->next->prev = n;
                    874:                } else
                    875:                        man->last->parent->last = n;
1.266     schwarze  876:                man->last->next = n;
                    877:                n->prev = man->last;
                    878:                n->parent = man->last->parent;
                    879:                break;
                    880:        case ROFF_NEXT_CHILD:
1.287     schwarze  881:                if (man->last->child != NULL) {
                    882:                        n->next = man->last->child;
                    883:                        man->last->child->prev = n;
                    884:                } else
                    885:                        man->last->last = n;
1.266     schwarze  886:                man->last->child = n;
                    887:                n->parent = man->last;
                    888:                break;
                    889:        default:
                    890:                abort();
                    891:        }
1.282     schwarze  892:        man->last = n;
1.266     schwarze  893:
                    894:        switch (n->type) {
                    895:        case ROFFT_HEAD:
                    896:                n->parent->head = n;
                    897:                break;
                    898:        case ROFFT_BODY:
1.282     schwarze  899:                if (n->end != ENDBODY_NOT)
                    900:                        return;
1.266     schwarze  901:                n->parent->body = n;
                    902:                break;
                    903:        case ROFFT_TAIL:
                    904:                n->parent->tail = n;
                    905:                break;
                    906:        default:
1.282     schwarze  907:                return;
1.266     schwarze  908:        }
1.282     schwarze  909:
                    910:        /*
                    911:         * Copy over the normalised-data pointer of our parent.  Not
                    912:         * everybody has one, but copying a null pointer is fine.
                    913:         */
                    914:
                    915:        n->norm = n->parent->norm;
                    916:        assert(n->parent->type == ROFFT_BLOCK);
1.266     schwarze  917: }
                    918:
1.267     schwarze  919: void
                    920: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    921: {
                    922:        struct roff_node        *n;
                    923:
                    924:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    925:        n->string = roff_strdup(man->roff, word);
                    926:        roff_node_append(man, n);
1.286     schwarze  927:        n->flags |= NODE_VALID | NODE_ENDED;
1.267     schwarze  928:        man->next = ROFF_NEXT_SIBLING;
                    929: }
                    930:
                    931: void
                    932: roff_word_append(struct roff_man *man, const char *word)
                    933: {
                    934:        struct roff_node        *n;
                    935:        char                    *addstr, *newstr;
                    936:
                    937:        n = man->last;
                    938:        addstr = roff_strdup(man->roff, word);
                    939:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    940:        free(addstr);
                    941:        free(n->string);
                    942:        n->string = newstr;
                    943:        man->next = ROFF_NEXT_SIBLING;
1.268     schwarze  944: }
                    945:
                    946: void
                    947: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    948: {
                    949:        struct roff_node        *n;
                    950:
                    951:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    952:        roff_node_append(man, n);
                    953:        man->next = ROFF_NEXT_CHILD;
                    954: }
                    955:
                    956: struct roff_node *
                    957: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    958: {
                    959:        struct roff_node        *n;
                    960:
                    961:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    962:        roff_node_append(man, n);
                    963:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze  964:        return n;
1.267     schwarze  965: }
                    966:
1.266     schwarze  967: struct roff_node *
                    968: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    969: {
                    970:        struct roff_node        *n;
                    971:
                    972:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    973:        roff_node_append(man, n);
                    974:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze  975:        return n;
1.266     schwarze  976: }
                    977:
                    978: struct roff_node *
                    979: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    980: {
                    981:        struct roff_node        *n;
                    982:
                    983:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    984:        roff_node_append(man, n);
                    985:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze  986:        return n;
1.267     schwarze  987: }
                    988:
                    989: void
                    990: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
                    991: {
                    992:        struct roff_node        *n;
                    993:
                    994:        n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
                    995:        n->eqn = eqn;
                    996:        if (eqn->ln > man->last->line)
1.285     schwarze  997:                n->flags |= NODE_LINE;
1.267     schwarze  998:        roff_node_append(man, n);
                    999:        man->next = ROFF_NEXT_SIBLING;
                   1000: }
                   1001:
                   1002: void
                   1003: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
                   1004: {
                   1005:        struct roff_node        *n;
                   1006:
                   1007:        if (man->macroset == MACROSET_MAN)
                   1008:                man_breakscope(man, TOKEN_NONE);
                   1009:        n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                   1010:        n->span = tbl;
                   1011:        roff_node_append(man, n);
1.286     schwarze 1012:        n->flags |= NODE_VALID | NODE_ENDED;
1.267     schwarze 1013:        man->next = ROFF_NEXT_SIBLING;
1.266     schwarze 1014: }
                   1015:
                   1016: void
                   1017: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1018: {
                   1019:
                   1020:        /* Adjust siblings. */
                   1021:
                   1022:        if (n->prev)
                   1023:                n->prev->next = n->next;
                   1024:        if (n->next)
                   1025:                n->next->prev = n->prev;
                   1026:
                   1027:        /* Adjust parent. */
                   1028:
                   1029:        if (n->parent != NULL) {
                   1030:                if (n->parent->child == n)
                   1031:                        n->parent->child = n->next;
                   1032:                if (n->parent->last == n)
                   1033:                        n->parent->last = n->prev;
                   1034:        }
                   1035:
                   1036:        /* Adjust parse point. */
                   1037:
                   1038:        if (man == NULL)
                   1039:                return;
                   1040:        if (man->last == n) {
                   1041:                if (n->prev == NULL) {
                   1042:                        man->last = n->parent;
                   1043:                        man->next = ROFF_NEXT_CHILD;
                   1044:                } else {
                   1045:                        man->last = n->prev;
                   1046:                        man->next = ROFF_NEXT_SIBLING;
                   1047:                }
                   1048:        }
                   1049:        if (man->first == n)
                   1050:                man->first = NULL;
                   1051: }
                   1052:
                   1053: void
                   1054: roff_node_free(struct roff_node *n)
                   1055: {
                   1056:
                   1057:        if (n->args != NULL)
                   1058:                mdoc_argv_free(n->args);
                   1059:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1060:                free(n->norm);
                   1061:        free(n->string);
                   1062:        free(n);
                   1063: }
                   1064:
                   1065: void
                   1066: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1067: {
                   1068:
                   1069:        while (n->child != NULL)
                   1070:                roff_node_delete(man, n->child);
                   1071:        roff_node_unlink(man, n);
                   1072:        roff_node_free(n);
1.269     schwarze 1073: }
                   1074:
                   1075: void
                   1076: deroff(char **dest, const struct roff_node *n)
                   1077: {
                   1078:        char    *cp;
                   1079:        size_t   sz;
                   1080:
                   1081:        if (n->type != ROFFT_TEXT) {
                   1082:                for (n = n->child; n != NULL; n = n->next)
                   1083:                        deroff(dest, n);
                   1084:                return;
                   1085:        }
                   1086:
1.288     schwarze 1087:        /* Skip leading whitespace. */
1.269     schwarze 1088:
1.288     schwarze 1089:        for (cp = n->string; *cp != '\0'; cp++) {
1.289     schwarze 1090:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1091:                    strchr(" %&0^|~", cp[1]) != NULL)
1.269     schwarze 1092:                        cp++;
1.288     schwarze 1093:                else if ( ! isspace((unsigned char)*cp))
1.269     schwarze 1094:                        break;
                   1095:        }
                   1096:
1.289     schwarze 1097:        /* Skip trailing backslash. */
                   1098:
                   1099:        sz = strlen(cp);
1.290     schwarze 1100:        if (sz > 0 && cp[sz - 1] == '\\')
1.289     schwarze 1101:                sz--;
                   1102:
1.269     schwarze 1103:        /* Skip trailing whitespace. */
                   1104:
1.289     schwarze 1105:        for (; sz; sz--)
1.269     schwarze 1106:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1107:                        break;
                   1108:
                   1109:        /* Skip empty strings. */
                   1110:
                   1111:        if (sz == 0)
                   1112:                return;
                   1113:
                   1114:        if (*dest == NULL) {
                   1115:                *dest = mandoc_strndup(cp, sz);
                   1116:                return;
                   1117:        }
                   1118:
                   1119:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1120:        free(*dest);
                   1121:        *dest = cp;
1.266     schwarze 1122: }
                   1123:
                   1124: /* --- main functions of the roff parser ---------------------------------- */
                   1125:
1.94      kristaps 1126: /*
1.206     schwarze 1127:  * In the current line, expand escape sequences that tend to get
                   1128:  * used in numerical expressions and conditional requests.
                   1129:  * Also check the syntax of the remaining escape sequences.
1.154     kristaps 1130:  */
1.172     schwarze 1131: static enum rofferr
1.238     schwarze 1132: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94      kristaps 1133: {
1.208     schwarze 1134:        char             ubuf[24]; /* buffer to print the number */
1.205     schwarze 1135:        const char      *start; /* start of the string to process */
1.209     schwarze 1136:        char            *stesc; /* start of an escape sequence ('\\') */
1.108     schwarze 1137:        const char      *stnam; /* start of the name, after "[(*" */
                   1138:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1139:        const char      *res;   /* the string to be substituted */
1.238     schwarze 1140:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.181     schwarze 1141:        size_t           maxl;  /* expected length of the escape name */
                   1142:        size_t           naml;  /* actual length of the escape name */
1.237     schwarze 1143:        enum mandoc_esc  esc;   /* type of the escape sequence */
                   1144:        int              inaml; /* length returned from mandoc_escape() */
1.181     schwarze 1145:        int              expand_count;  /* to avoid infinite loops */
1.206     schwarze 1146:        int              npos;  /* position in numeric expression */
1.218     schwarze 1147:        int              arg_complete; /* argument not interrupted by eol */
1.206     schwarze 1148:        char             term;  /* character terminating the escape */
1.94      kristaps 1149:
1.170     schwarze 1150:        expand_count = 0;
1.238     schwarze 1151:        start = buf->buf + pos;
1.205     schwarze 1152:        stesc = strchr(start, '\0') - 1;
                   1153:        while (stesc-- > start) {
1.170     schwarze 1154:
1.205     schwarze 1155:                /* Search backwards for the next backslash. */
                   1156:
1.238     schwarze 1157:                if (*stesc != '\\')
1.205     schwarze 1158:                        continue;
                   1159:
                   1160:                /* If it is escaped, skip it. */
                   1161:
                   1162:                for (cp = stesc - 1; cp >= start; cp--)
1.238     schwarze 1163:                        if (*cp != '\\')
1.205     schwarze 1164:                                break;
                   1165:
1.238     schwarze 1166:                if ((stesc - cp) % 2 == 0) {
1.209     schwarze 1167:                        stesc = (char *)cp;
1.205     schwarze 1168:                        continue;
                   1169:                }
1.108     schwarze 1170:
1.206     schwarze 1171:                /* Decide whether to expand or to check only. */
1.108     schwarze 1172:
1.206     schwarze 1173:                term = '\0';
1.205     schwarze 1174:                cp = stesc + 1;
1.181     schwarze 1175:                switch (*cp) {
1.207     schwarze 1176:                case '*':
1.181     schwarze 1177:                        res = NULL;
                   1178:                        break;
1.207     schwarze 1179:                case 'B':
                   1180:                case 'w':
1.206     schwarze 1181:                        term = cp[1];
                   1182:                        /* FALLTHROUGH */
1.207     schwarze 1183:                case 'n':
1.181     schwarze 1184:                        res = ubuf;
                   1185:                        break;
                   1186:                default:
1.237     schwarze 1187:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1188:                        if (esc == ESCAPE_ERROR ||
                   1189:                            (esc == ESCAPE_SPECIAL &&
1.279     schwarze 1190:                             mchars_spec2cp(stnam, inaml) < 0))
1.219     schwarze 1191:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238     schwarze 1192:                                    r->parse, ln, (int)(stesc - buf->buf),
1.219     schwarze 1193:                                    "%.*s", (int)(cp - stesc), stesc);
1.205     schwarze 1194:                        continue;
1.152     kristaps 1195:                }
                   1196:
1.205     schwarze 1197:                if (EXPAND_LIMIT < ++expand_count) {
                   1198:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238     schwarze 1199:                            ln, (int)(stesc - buf->buf), NULL);
1.277     schwarze 1200:                        return ROFF_IGN;
1.205     schwarze 1201:                }
1.108     schwarze 1202:
                   1203:                /*
                   1204:                 * The third character decides the length
1.181     schwarze 1205:                 * of the name of the string or register.
1.108     schwarze 1206:                 * Save a pointer to the name.
                   1207:                 */
                   1208:
1.238     schwarze 1209:                if (term == '\0') {
1.206     schwarze 1210:                        switch (*++cp) {
1.207     schwarze 1211:                        case '\0':
1.206     schwarze 1212:                                maxl = 0;
                   1213:                                break;
1.207     schwarze 1214:                        case '(':
1.206     schwarze 1215:                                cp++;
                   1216:                                maxl = 2;
                   1217:                                break;
1.207     schwarze 1218:                        case '[':
1.206     schwarze 1219:                                cp++;
                   1220:                                term = ']';
                   1221:                                maxl = 0;
                   1222:                                break;
                   1223:                        default:
                   1224:                                maxl = 1;
                   1225:                                break;
                   1226:                        }
                   1227:                } else {
                   1228:                        cp += 2;
1.94      kristaps 1229:                        maxl = 0;
                   1230:                }
1.108     schwarze 1231:                stnam = cp;
1.94      kristaps 1232:
1.108     schwarze 1233:                /* Advance to the end of the name. */
1.94      kristaps 1234:
1.253     schwarze 1235:                naml = 0;
1.218     schwarze 1236:                arg_complete = 1;
1.253     schwarze 1237:                while (maxl == 0 || naml < maxl) {
1.238     schwarze 1238:                        if (*cp == '\0') {
1.219     schwarze 1239:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238     schwarze 1240:                                    ln, (int)(stesc - buf->buf), stesc);
1.218     schwarze 1241:                                arg_complete = 0;
1.206     schwarze 1242:                                break;
1.153     kristaps 1243:                        }
1.238     schwarze 1244:                        if (maxl == 0 && *cp == term) {
1.206     schwarze 1245:                                cp++;
1.253     schwarze 1246:                                break;
                   1247:                        }
                   1248:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1249:                                naml++;
                   1250:                                continue;
                   1251:                        }
                   1252:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1253:                        case ESCAPE_SPECIAL:
                   1254:                        case ESCAPE_UNICODE:
                   1255:                        case ESCAPE_NUMBERED:
                   1256:                        case ESCAPE_OVERSTRIKE:
                   1257:                                naml++;
                   1258:                                break;
                   1259:                        default:
1.94      kristaps 1260:                                break;
1.206     schwarze 1261:                        }
1.94      kristaps 1262:                }
                   1263:
1.108     schwarze 1264:                /*
                   1265:                 * Retrieve the replacement string; if it is
                   1266:                 * undefined, resume searching for escapes.
                   1267:                 */
                   1268:
1.206     schwarze 1269:                switch (stesc[1]) {
1.207     schwarze 1270:                case '*':
1.218     schwarze 1271:                        if (arg_complete)
                   1272:                                res = roff_getstrn(r, stnam, naml);
1.206     schwarze 1273:                        break;
1.207     schwarze 1274:                case 'B':
1.206     schwarze 1275:                        npos = 0;
1.218     schwarze 1276:                        ubuf[0] = arg_complete &&
1.261     schwarze 1277:                            roff_evalnum(r, ln, stnam, &npos,
                   1278:                              NULL, ROFFNUM_SCALE) &&
1.218     schwarze 1279:                            stnam + npos + 1 == cp ? '1' : '0';
1.206     schwarze 1280:                        ubuf[1] = '\0';
                   1281:                        break;
1.207     schwarze 1282:                case 'n':
1.218     schwarze 1283:                        if (arg_complete)
                   1284:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1285:                                    roff_getregn(r, stnam, naml));
                   1286:                        else
                   1287:                                ubuf[0] = '\0';
1.206     schwarze 1288:                        break;
1.207     schwarze 1289:                case 'w':
1.218     schwarze 1290:                        /* use even incomplete args */
1.208     schwarze 1291:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206     schwarze 1292:                            24 * (int)naml);
                   1293:                        break;
                   1294:                }
1.94      kristaps 1295:
1.238     schwarze 1296:                if (res == NULL) {
1.219     schwarze 1297:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238     schwarze 1298:                            r->parse, ln, (int)(stesc - buf->buf),
1.219     schwarze 1299:                            "%.*s", (int)naml, stnam);
1.142     kristaps 1300:                        res = "";
1.246     schwarze 1301:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1302:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1303:                            ln, (int)(stesc - buf->buf), NULL);
1.277     schwarze 1304:                        return ROFF_IGN;
1.94      kristaps 1305:                }
                   1306:
1.108     schwarze 1307:                /* Replace the escape sequence by the string. */
                   1308:
1.209     schwarze 1309:                *stesc = '\0';
1.238     schwarze 1310:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1311:                    buf->buf, res, cp) + 1;
1.94      kristaps 1312:
1.205     schwarze 1313:                /* Prepare for the next replacement. */
1.94      kristaps 1314:
1.205     schwarze 1315:                start = nbuf + pos;
1.238     schwarze 1316:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1317:                free(buf->buf);
                   1318:                buf->buf = nbuf;
1.154     kristaps 1319:        }
1.277     schwarze 1320:        return ROFF_CONT;
1.154     kristaps 1321: }
                   1322:
                   1323: /*
1.275     schwarze 1324:  * Process text streams.
1.154     kristaps 1325:  */
                   1326: static enum rofferr
1.238     schwarze 1327: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154     kristaps 1328: {
                   1329:        size_t           sz;
                   1330:        const char      *start;
1.178     schwarze 1331:        char            *p;
                   1332:        int              isz;
1.154     kristaps 1333:        enum mandoc_esc  esc;
                   1334:
1.275     schwarze 1335:        /* Spring the input line trap. */
                   1336:
                   1337:        if (roffit_lines == 1) {
                   1338:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1339:                free(buf->buf);
                   1340:                buf->buf = p;
                   1341:                buf->sz = isz + 1;
                   1342:                *offs = 0;
                   1343:                free(roffit_macro);
                   1344:                roffit_lines = 0;
1.277     schwarze 1345:                return ROFF_REPARSE;
1.275     schwarze 1346:        } else if (roffit_lines > 1)
                   1347:                --roffit_lines;
                   1348:
                   1349:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1350:
1.238     schwarze 1351:        start = p = buf->buf + pos;
1.154     kristaps 1352:
1.238     schwarze 1353:        while (*p != '\0') {
1.154     kristaps 1354:                sz = strcspn(p, "-\\");
                   1355:                p += sz;
                   1356:
1.238     schwarze 1357:                if (*p == '\0')
1.159     kristaps 1358:                        break;
                   1359:
1.238     schwarze 1360:                if (*p == '\\') {
1.154     kristaps 1361:                        /* Skip over escapes. */
                   1362:                        p++;
1.189     schwarze 1363:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238     schwarze 1364:                        if (esc == ESCAPE_ERROR)
1.154     kristaps 1365:                                break;
1.264     schwarze 1366:                        while (*p == '-')
                   1367:                                p++;
1.155     kristaps 1368:                        continue;
1.159     kristaps 1369:                } else if (p == start) {
1.158     kristaps 1370:                        p++;
1.155     kristaps 1371:                        continue;
1.158     kristaps 1372:                }
1.155     kristaps 1373:
1.171     schwarze 1374:                if (isalpha((unsigned char)p[-1]) &&
                   1375:                    isalpha((unsigned char)p[1]))
1.155     kristaps 1376:                        *p = ASCII_HYPH;
                   1377:                p++;
1.94      kristaps 1378:        }
1.277     schwarze 1379:        return ROFF_CONT;
1.94      kristaps 1380: }
                   1381:
1.67      kristaps 1382: enum rofferr
1.238     schwarze 1383: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67      kristaps 1384: {
1.294     schwarze 1385:        enum roff_tok    t;
1.109     kristaps 1386:        enum rofferr     e;
1.238     schwarze 1387:        int              pos;   /* parse point */
1.243     schwarze 1388:        int              spos;  /* saved parse point for messages */
1.238     schwarze 1389:        int              ppos;  /* original offset in buf->buf */
                   1390:        int              ctl;   /* macro line (boolean) */
                   1391:
                   1392:        ppos = pos = *offs;
1.79      kristaps 1393:
1.230     schwarze 1394:        /* Handle in-line equation delimiters. */
                   1395:
1.236     schwarze 1396:        if (r->tbl == NULL &&
                   1397:            r->last_eqn != NULL && r->last_eqn->delim &&
1.230     schwarze 1398:            (r->eqn == NULL || r->eqn_inline)) {
1.238     schwarze 1399:                e = roff_eqndelim(r, buf, pos);
1.230     schwarze 1400:                if (e == ROFF_REPARSE)
1.277     schwarze 1401:                        return e;
1.230     schwarze 1402:                assert(e == ROFF_CONT);
                   1403:        }
                   1404:
                   1405:        /* Expand some escape sequences. */
1.94      kristaps 1406:
1.238     schwarze 1407:        e = roff_res(r, buf, ln, pos);
                   1408:        if (e == ROFF_IGN)
1.277     schwarze 1409:                return e;
1.238     schwarze 1410:        assert(e == ROFF_CONT);
1.94      kristaps 1411:
1.238     schwarze 1412:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.130     kristaps 1413:
1.94      kristaps 1414:        /*
1.79      kristaps 1415:         * First, if a scope is open and we're not a macro, pass the
1.252     schwarze 1416:         * text through the macro's filter.
                   1417:         * Equations process all content themselves.
                   1418:         * Tables process almost all content themselves, but we want
                   1419:         * to warn about macros before passing it there.
1.79      kristaps 1420:         */
1.74      kristaps 1421:
1.252     schwarze 1422:        if (r->last != NULL && ! ctl) {
1.78      kristaps 1423:                t = r->last->tok;
1.238     schwarze 1424:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.294     schwarze 1425:                if (e == ROFF_IGN)
1.277     schwarze 1426:                        return e;
1.294     schwarze 1427:                assert(e == ROFF_CONT);
1.182     schwarze 1428:        }
1.252     schwarze 1429:        if (r->eqn != NULL)
1.277     schwarze 1430:                return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.252     schwarze 1431:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.277     schwarze 1432:                return tbl_read(r->tbl, ln, buf->buf, ppos);
1.252     schwarze 1433:        if ( ! ctl)
1.277     schwarze 1434:                return roff_parsetext(buf, pos, offs);
1.228     schwarze 1435:
                   1436:        /* Skip empty request lines. */
                   1437:
1.238     schwarze 1438:        if (buf->buf[pos] == '"') {
1.228     schwarze 1439:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1440:                    ln, pos, NULL);
1.277     schwarze 1441:                return ROFF_IGN;
1.238     schwarze 1442:        } else if (buf->buf[pos] == '\0')
1.277     schwarze 1443:                return ROFF_IGN;
1.67      kristaps 1444:
1.79      kristaps 1445:        /*
                   1446:         * If a scope is open, go to the child handler for that macro,
                   1447:         * as it may want to preprocess before doing anything with it.
1.125     kristaps 1448:         * Don't do so if an equation is open.
1.79      kristaps 1449:         */
1.78      kristaps 1450:
1.79      kristaps 1451:        if (r->last) {
                   1452:                t = r->last->tok;
1.277     schwarze 1453:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1454:        }
1.78      kristaps 1455:
1.243     schwarze 1456:        /* No scope is open.  This is a new request or macro. */
                   1457:
                   1458:        spos = pos;
                   1459:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1460:
                   1461:        /* Tables ignore most macros. */
                   1462:
1.294     schwarze 1463:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS)) {
1.243     schwarze 1464:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1465:                    ln, pos, buf->buf + spos);
1.257     schwarze 1466:                if (t == ROFF_TS)
1.277     schwarze 1467:                        return ROFF_IGN;
1.257     schwarze 1468:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1469:                        pos++;
1.291     schwarze 1470:                while (buf->buf[pos] == ' ')
1.257     schwarze 1471:                        pos++;
1.277     schwarze 1472:                return tbl_read(r->tbl, ln, buf->buf, pos);
1.243     schwarze 1473:        }
                   1474:
1.79      kristaps 1475:        /*
1.243     schwarze 1476:         * This is neither a roff request nor a user-defined macro.
                   1477:         * Let the standard macro set parsers handle it.
1.79      kristaps 1478:         */
1.67      kristaps 1479:
1.294     schwarze 1480:        if (t == TOKEN_NONE)
1.277     schwarze 1481:                return ROFF_CONT;
1.243     schwarze 1482:
                   1483:        /* Execute a roff request or a user defined macro. */
1.67      kristaps 1484:
1.296     schwarze 1485:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74      kristaps 1486: }
                   1487:
1.117     kristaps 1488: void
1.74      kristaps 1489: roff_endparse(struct roff *r)
                   1490: {
                   1491:
1.110     kristaps 1492:        if (r->last)
1.221     schwarze 1493:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1494:                    r->last->line, r->last->col,
1.295     schwarze 1495:                    roff_name[r->last->tok]);
1.117     kristaps 1496:
1.125     kristaps 1497:        if (r->eqn) {
1.221     schwarze 1498:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1499:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151     kristaps 1500:                eqn_end(&r->eqn);
1.125     kristaps 1501:        }
                   1502:
1.117     kristaps 1503:        if (r->tbl) {
1.221     schwarze 1504:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1505:                    r->tbl->line, r->tbl->pos, "TS");
1.151     kristaps 1506:                tbl_end(&r->tbl);
1.117     kristaps 1507:        }
1.67      kristaps 1508: }
                   1509:
                   1510: /*
                   1511:  * Parse a roff node's type from the input buffer.  This must be in the
                   1512:  * form of ".foo xxx" in the usual way.
                   1513:  */
1.294     schwarze 1514: static enum roff_tok
1.214     schwarze 1515: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67      kristaps 1516: {
1.214     schwarze 1517:        char            *cp;
1.106     kristaps 1518:        const char      *mac;
                   1519:        size_t           maclen;
1.294     schwarze 1520:        enum roff_tok    t;
1.67      kristaps 1521:
1.214     schwarze 1522:        cp = buf + *pos;
                   1523:
                   1524:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294     schwarze 1525:                return TOKEN_NONE;
1.67      kristaps 1526:
1.214     schwarze 1527:        mac = cp;
                   1528:        maclen = roff_getname(r, &cp, ln, ppos);
1.67      kristaps 1529:
1.106     kristaps 1530:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.295     schwarze 1531:            ? ROFF_USERDEF : roffhash_find(r->reqtab, mac, maclen);
1.67      kristaps 1532:
1.294     schwarze 1533:        if (t != TOKEN_NONE)
1.214     schwarze 1534:                *pos = cp - buf;
1.67      kristaps 1535:
1.277     schwarze 1536:        return t;
1.67      kristaps 1537: }
                   1538:
1.266     schwarze 1539: /* --- handling of request blocks ----------------------------------------- */
                   1540:
1.67      kristaps 1541: static enum rofferr
1.76      kristaps 1542: roff_cblock(ROFF_ARGS)
1.67      kristaps 1543: {
                   1544:
1.79      kristaps 1545:        /*
                   1546:         * A block-close `..' should only be invoked as a child of an
                   1547:         * ignore macro, otherwise raise a warning and just ignore it.
                   1548:         */
                   1549:
1.238     schwarze 1550:        if (r->last == NULL) {
1.221     schwarze 1551:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1552:                    ln, ppos, "..");
1.277     schwarze 1553:                return ROFF_IGN;
1.76      kristaps 1554:        }
1.67      kristaps 1555:
1.81      kristaps 1556:        switch (r->last->tok) {
1.207     schwarze 1557:        case ROFF_am:
1.220     schwarze 1558:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207     schwarze 1559:        case ROFF_ami:
                   1560:        case ROFF_de:
1.108     schwarze 1561:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207     schwarze 1562:        case ROFF_dei:
                   1563:        case ROFF_ig:
1.81      kristaps 1564:                break;
                   1565:        default:
1.221     schwarze 1566:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1567:                    ln, ppos, "..");
1.277     schwarze 1568:                return ROFF_IGN;
1.76      kristaps 1569:        }
1.67      kristaps 1570:
1.238     schwarze 1571:        if (buf->buf[pos] != '\0')
1.217     schwarze 1572:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238     schwarze 1573:                    ".. %s", buf->buf + pos);
1.71      kristaps 1574:
                   1575:        roffnode_pop(r);
1.76      kristaps 1576:        roffnode_cleanscope(r);
1.277     schwarze 1577:        return ROFF_IGN;
1.71      kristaps 1578:
1.67      kristaps 1579: }
                   1580:
1.76      kristaps 1581: static void
                   1582: roffnode_cleanscope(struct roff *r)
1.67      kristaps 1583: {
                   1584:
1.76      kristaps 1585:        while (r->last) {
1.173     schwarze 1586:                if (--r->last->endspan != 0)
1.76      kristaps 1587:                        break;
                   1588:                roffnode_pop(r);
                   1589:        }
1.67      kristaps 1590: }
                   1591:
1.195     schwarze 1592: static void
                   1593: roff_ccond(struct roff *r, int ln, int ppos)
1.74      kristaps 1594: {
                   1595:
1.76      kristaps 1596:        if (NULL == r->last) {
1.221     schwarze 1597:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1598:                    ln, ppos, "\\}");
1.195     schwarze 1599:                return;
1.76      kristaps 1600:        }
                   1601:
1.82      kristaps 1602:        switch (r->last->tok) {
1.207     schwarze 1603:        case ROFF_el:
                   1604:        case ROFF_ie:
                   1605:        case ROFF_if:
1.82      kristaps 1606:                break;
                   1607:        default:
1.221     schwarze 1608:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1609:                    ln, ppos, "\\}");
1.195     schwarze 1610:                return;
1.75      kristaps 1611:        }
                   1612:
1.76      kristaps 1613:        if (r->last->endspan > -1) {
1.221     schwarze 1614:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1615:                    ln, ppos, "\\}");
1.195     schwarze 1616:                return;
1.76      kristaps 1617:        }
                   1618:
1.75      kristaps 1619:        roffnode_pop(r);
1.76      kristaps 1620:        roffnode_cleanscope(r);
1.195     schwarze 1621:        return;
1.76      kristaps 1622: }
                   1623:
                   1624: static enum rofferr
1.80      kristaps 1625: roff_block(ROFF_ARGS)
1.76      kristaps 1626: {
1.220     schwarze 1627:        const char      *name;
                   1628:        char            *iname, *cp;
1.213     schwarze 1629:        size_t           namesz;
1.106     kristaps 1630:
1.220     schwarze 1631:        /* Ignore groff compatibility mode for now. */
1.76      kristaps 1632:
1.238     schwarze 1633:        if (tok == ROFF_de1)
1.220     schwarze 1634:                tok = ROFF_de;
1.251     schwarze 1635:        else if (tok == ROFF_dei1)
                   1636:                tok = ROFF_dei;
1.238     schwarze 1637:        else if (tok == ROFF_am1)
1.220     schwarze 1638:                tok = ROFF_am;
1.251     schwarze 1639:        else if (tok == ROFF_ami1)
                   1640:                tok = ROFF_ami;
1.220     schwarze 1641:
                   1642:        /* Parse the macro name argument. */
                   1643:
1.238     schwarze 1644:        cp = buf->buf + pos;
                   1645:        if (tok == ROFF_ig) {
1.220     schwarze 1646:                iname = NULL;
                   1647:                namesz = 0;
                   1648:        } else {
                   1649:                iname = cp;
                   1650:                namesz = roff_getname(r, &cp, ln, ppos);
                   1651:                iname[namesz] = '\0';
                   1652:        }
1.107     kristaps 1653:
1.220     schwarze 1654:        /* Resolve the macro name argument if it is indirect. */
1.107     kristaps 1655:
1.238     schwarze 1656:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1657:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220     schwarze 1658:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238     schwarze 1659:                            r->parse, ln, (int)(iname - buf->buf),
1.220     schwarze 1660:                            "%.*s", (int)namesz, iname);
                   1661:                        namesz = 0;
                   1662:                } else
                   1663:                        namesz = strlen(name);
                   1664:        } else
                   1665:                name = iname;
1.107     kristaps 1666:
1.238     schwarze 1667:        if (namesz == 0 && tok != ROFF_ig) {
1.220     schwarze 1668:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295     schwarze 1669:                    ln, ppos, roff_name[tok]);
1.277     schwarze 1670:                return ROFF_IGN;
1.220     schwarze 1671:        }
1.80      kristaps 1672:
1.106     kristaps 1673:        roffnode_push(r, tok, name, ln, ppos);
                   1674:
                   1675:        /*
                   1676:         * At the beginning of a `de' macro, clear the existing string
                   1677:         * with the same name, if there is one.  New content will be
1.193     schwarze 1678:         * appended from roff_block_text() in multiline mode.
1.106     kristaps 1679:         */
1.107     kristaps 1680:
1.238     schwarze 1681:        if (tok == ROFF_de || tok == ROFF_dei)
1.213     schwarze 1682:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76      kristaps 1683:
1.238     schwarze 1684:        if (*cp == '\0')
1.277     schwarze 1685:                return ROFF_IGN;
1.78      kristaps 1686:
1.220     schwarze 1687:        /* Get the custom end marker. */
1.107     kristaps 1688:
1.220     schwarze 1689:        iname = cp;
1.213     schwarze 1690:        namesz = roff_getname(r, &cp, ln, ppos);
1.220     schwarze 1691:
                   1692:        /* Resolve the end marker if it is indirect. */
                   1693:
1.238     schwarze 1694:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1695:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220     schwarze 1696:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238     schwarze 1697:                            r->parse, ln, (int)(iname - buf->buf),
1.220     schwarze 1698:                            "%.*s", (int)namesz, iname);
                   1699:                        namesz = 0;
                   1700:                } else
                   1701:                        namesz = strlen(name);
                   1702:        } else
                   1703:                name = iname;
                   1704:
1.213     schwarze 1705:        if (namesz)
                   1706:                r->last->end = mandoc_strndup(name, namesz);
1.78      kristaps 1707:
1.238     schwarze 1708:        if (*cp != '\0')
1.217     schwarze 1709:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295     schwarze 1710:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74      kristaps 1711:
1.277     schwarze 1712:        return ROFF_IGN;
1.78      kristaps 1713: }
                   1714:
                   1715: static enum rofferr
1.80      kristaps 1716: roff_block_sub(ROFF_ARGS)
1.79      kristaps 1717: {
1.294     schwarze 1718:        enum roff_tok   t;
1.79      kristaps 1719:        int             i, j;
                   1720:
                   1721:        /*
                   1722:         * First check whether a custom macro exists at this level.  If
                   1723:         * it does, then check against it.  This is some of groff's
                   1724:         * stranger behaviours.  If we encountered a custom end-scope
                   1725:         * tag and that tag also happens to be a "real" macro, then we
                   1726:         * need to try interpreting it again as a real macro.  If it's
                   1727:         * not, then return ignore.  Else continue.
                   1728:         */
                   1729:
                   1730:        if (r->last->end) {
1.130     kristaps 1731:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238     schwarze 1732:                        if (buf->buf[i] != r->last->end[j])
1.79      kristaps 1733:                                break;
                   1734:
1.238     schwarze 1735:                if (r->last->end[j] == '\0' &&
                   1736:                    (buf->buf[i] == '\0' ||
                   1737:                     buf->buf[i] == ' ' ||
                   1738:                     buf->buf[i] == '\t')) {
1.79      kristaps 1739:                        roffnode_pop(r);
                   1740:                        roffnode_cleanscope(r);
                   1741:
1.238     schwarze 1742:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130     kristaps 1743:                                i++;
                   1744:
                   1745:                        pos = i;
1.238     schwarze 1746:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294     schwarze 1747:                            TOKEN_NONE)
1.277     schwarze 1748:                                return ROFF_RERUN;
                   1749:                        return ROFF_IGN;
1.79      kristaps 1750:                }
                   1751:        }
                   1752:
                   1753:        /*
                   1754:         * If we have no custom end-query or lookup failed, then try
                   1755:         * pulling it out of the hashtable.
                   1756:         */
                   1757:
1.238     schwarze 1758:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79      kristaps 1759:
1.238     schwarze 1760:        if (t != ROFF_cblock) {
                   1761:                if (tok != ROFF_ig)
                   1762:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277     schwarze 1763:                return ROFF_IGN;
1.106     kristaps 1764:        }
1.79      kristaps 1765:
1.277     schwarze 1766:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1767: }
                   1768:
                   1769: static enum rofferr
1.80      kristaps 1770: roff_block_text(ROFF_ARGS)
1.78      kristaps 1771: {
                   1772:
1.238     schwarze 1773:        if (tok != ROFF_ig)
                   1774:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106     kristaps 1775:
1.277     schwarze 1776:        return ROFF_IGN;
1.78      kristaps 1777: }
                   1778:
                   1779: static enum rofferr
1.82      kristaps 1780: roff_cond_sub(ROFF_ARGS)
                   1781: {
1.294     schwarze 1782:        enum roff_tok    t;
1.139     kristaps 1783:        char            *ep;
1.198     schwarze 1784:        int              rr;
1.82      kristaps 1785:
                   1786:        rr = r->last->rule;
1.139     kristaps 1787:        roffnode_cleanscope(r);
1.238     schwarze 1788:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82      kristaps 1789:
1.139     kristaps 1790:        /*
1.177     schwarze 1791:         * Fully handle known macros when they are structurally
                   1792:         * required or when the conditional evaluated to true.
1.87      kristaps 1793:         */
                   1794:
1.294     schwarze 1795:        if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.277     schwarze 1796:                return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.144     kristaps 1797:
1.196     schwarze 1798:        /*
                   1799:         * If `\}' occurs on a macro line without a preceding macro,
                   1800:         * drop the line completely.
                   1801:         */
                   1802:
1.238     schwarze 1803:        ep = buf->buf + pos;
                   1804:        if (ep[0] == '\\' && ep[1] == '}')
1.198     schwarze 1805:                rr = 0;
1.196     schwarze 1806:
1.177     schwarze 1807:        /* Always check for the closing delimiter `\}'. */
1.144     kristaps 1808:
1.238     schwarze 1809:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1810:                if (*(++ep) == '}') {
1.197     schwarze 1811:                        *ep = '&';
1.238     schwarze 1812:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.197     schwarze 1813:                }
1.247     schwarze 1814:                if (*ep != '\0')
                   1815:                        ++ep;
1.177     schwarze 1816:        }
1.277     schwarze 1817:        return rr ? ROFF_CONT : ROFF_IGN;
1.82      kristaps 1818: }
                   1819:
                   1820: static enum rofferr
                   1821: roff_cond_text(ROFF_ARGS)
1.78      kristaps 1822: {
1.140     kristaps 1823:        char            *ep;
1.198     schwarze 1824:        int              rr;
1.82      kristaps 1825:
                   1826:        rr = r->last->rule;
1.140     kristaps 1827:        roffnode_cleanscope(r);
1.82      kristaps 1828:
1.238     schwarze 1829:        ep = buf->buf + pos;
                   1830:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1831:                if (*(++ep) == '}') {
1.197     schwarze 1832:                        *ep = '&';
1.238     schwarze 1833:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.197     schwarze 1834:                }
1.247     schwarze 1835:                if (*ep != '\0')
                   1836:                        ++ep;
1.78      kristaps 1837:        }
1.277     schwarze 1838:        return rr ? ROFF_CONT : ROFF_IGN;
1.74      kristaps 1839: }
                   1840:
1.266     schwarze 1841: /* --- handling of numeric and conditional expressions -------------------- */
                   1842:
1.204     schwarze 1843: /*
                   1844:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1845:  * If there is at least one digit, return success and advance the
                   1846:  * parse point, else return failure and let the parse point unchanged.
                   1847:  * Ignore overflows, treat them just like the C language.
                   1848:  */
1.184     schwarze 1849: static int
1.261     schwarze 1850: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184     schwarze 1851: {
1.261     schwarze 1852:        int      myres, scaled, n, p;
1.206     schwarze 1853:
                   1854:        if (NULL == res)
                   1855:                res = &myres;
1.184     schwarze 1856:
                   1857:        p = *pos;
                   1858:        n = v[p] == '-';
1.261     schwarze 1859:        if (n || v[p] == '+')
1.184     schwarze 1860:                p++;
                   1861:
1.261     schwarze 1862:        if (flags & ROFFNUM_WHITE)
                   1863:                while (isspace((unsigned char)v[p]))
                   1864:                        p++;
                   1865:
1.184     schwarze 1866:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204     schwarze 1867:                *res = 10 * *res + v[p] - '0';
1.184     schwarze 1868:        if (p == *pos + n)
                   1869:                return 0;
                   1870:
                   1871:        if (n)
                   1872:                *res = -*res;
                   1873:
1.254     schwarze 1874:        /* Each number may be followed by one optional scaling unit. */
                   1875:
                   1876:        switch (v[p]) {
                   1877:        case 'f':
1.261     schwarze 1878:                scaled = *res * 65536;
1.254     schwarze 1879:                break;
                   1880:        case 'i':
1.261     schwarze 1881:                scaled = *res * 240;
1.254     schwarze 1882:                break;
                   1883:        case 'c':
1.261     schwarze 1884:                scaled = *res * 240 / 2.54;
1.254     schwarze 1885:                break;
                   1886:        case 'v':
                   1887:        case 'P':
1.261     schwarze 1888:                scaled = *res * 40;
1.254     schwarze 1889:                break;
                   1890:        case 'm':
                   1891:        case 'n':
1.261     schwarze 1892:                scaled = *res * 24;
1.254     schwarze 1893:                break;
                   1894:        case 'p':
1.261     schwarze 1895:                scaled = *res * 10 / 3;
1.254     schwarze 1896:                break;
                   1897:        case 'u':
1.261     schwarze 1898:                scaled = *res;
1.254     schwarze 1899:                break;
                   1900:        case 'M':
1.261     schwarze 1901:                scaled = *res * 6 / 25;
1.254     schwarze 1902:                break;
                   1903:        default:
1.261     schwarze 1904:                scaled = *res;
1.254     schwarze 1905:                p--;
                   1906:                break;
                   1907:        }
1.261     schwarze 1908:        if (flags & ROFFNUM_SCALE)
                   1909:                *res = scaled;
1.254     schwarze 1910:
                   1911:        *pos = p + 1;
1.277     schwarze 1912:        return 1;
1.184     schwarze 1913: }
                   1914:
1.198     schwarze 1915: /*
                   1916:  * Evaluate a string comparison condition.
                   1917:  * The first character is the delimiter.
                   1918:  * Succeed if the string up to its second occurrence
                   1919:  * matches the string up to its third occurence.
                   1920:  * Advance the cursor after the third occurrence
                   1921:  * or lacking that, to the end of the line.
                   1922:  */
                   1923: static int
                   1924: roff_evalstrcond(const char *v, int *pos)
                   1925: {
                   1926:        const char      *s1, *s2, *s3;
                   1927:        int              match;
                   1928:
                   1929:        match = 0;
                   1930:        s1 = v + *pos;          /* initial delimiter */
                   1931:        s2 = s1 + 1;            /* for scanning the first string */
                   1932:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   1933:
                   1934:        if (NULL == s3)         /* found no middle delimiter */
                   1935:                goto out;
                   1936:
                   1937:        while ('\0' != *++s3) {
                   1938:                if (*s2 != *s3) {  /* mismatch */
                   1939:                        s3 = strchr(s3, *s1);
                   1940:                        break;
                   1941:                }
                   1942:                if (*s3 == *s1) {  /* found the final delimiter */
                   1943:                        match = 1;
                   1944:                        break;
                   1945:                }
                   1946:                s2++;
                   1947:        }
                   1948:
                   1949: out:
                   1950:        if (NULL == s3)
                   1951:                s3 = strchr(s2, '\0');
1.242     schwarze 1952:        else if (*s3 != '\0')
1.198     schwarze 1953:                s3++;
                   1954:        *pos = s3 - v;
1.277     schwarze 1955:        return match;
1.198     schwarze 1956: }
                   1957:
1.204     schwarze 1958: /*
                   1959:  * Evaluate an optionally negated single character, numerical,
                   1960:  * or string condition.
                   1961:  */
1.198     schwarze 1962: static int
1.271     schwarze 1963: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88      kristaps 1964: {
1.271     schwarze 1965:        char    *cp, *name;
                   1966:        size_t   sz;
1.241     schwarze 1967:        int      number, savepos, wanttrue;
1.88      kristaps 1968:
1.198     schwarze 1969:        if ('!' == v[*pos]) {
                   1970:                wanttrue = 0;
                   1971:                (*pos)++;
                   1972:        } else
                   1973:                wanttrue = 1;
                   1974:
1.88      kristaps 1975:        switch (v[*pos]) {
1.240     schwarze 1976:        case '\0':
1.277     schwarze 1977:                return 0;
1.207     schwarze 1978:        case 'n':
                   1979:        case 'o':
1.88      kristaps 1980:                (*pos)++;
1.277     schwarze 1981:                return wanttrue;
1.207     schwarze 1982:        case 'c':
                   1983:        case 'd':
                   1984:        case 'e':
                   1985:        case 't':
1.239     schwarze 1986:        case 'v':
1.88      kristaps 1987:                (*pos)++;
1.277     schwarze 1988:                return !wanttrue;
1.271     schwarze 1989:        case 'r':
                   1990:                cp = name = v + ++*pos;
                   1991:                sz = roff_getname(r, &cp, ln, *pos);
                   1992:                *pos = cp - v;
1.277     schwarze 1993:                return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.88      kristaps 1994:        default:
                   1995:                break;
                   1996:        }
                   1997:
1.241     schwarze 1998:        savepos = *pos;
1.261     schwarze 1999:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277     schwarze 2000:                return (number > 0) == wanttrue;
1.241     schwarze 2001:        else if (*pos == savepos)
1.277     schwarze 2002:                return roff_evalstrcond(v, pos) == wanttrue;
1.204     schwarze 2003:        else
1.277     schwarze 2004:                return 0;
1.88      kristaps 2005: }
                   2006:
1.74      kristaps 2007: static enum rofferr
1.103     kristaps 2008: roff_line_ignore(ROFF_ARGS)
1.89      kristaps 2009: {
1.123     schwarze 2010:
1.277     schwarze 2011:        return ROFF_IGN;
1.89      kristaps 2012: }
                   2013:
1.104     kristaps 2014: static enum rofferr
1.251     schwarze 2015: roff_insec(ROFF_ARGS)
                   2016: {
                   2017:
                   2018:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.295     schwarze 2019:            ln, ppos, roff_name[tok]);
1.277     schwarze 2020:        return ROFF_IGN;
1.251     schwarze 2021: }
                   2022:
                   2023: static enum rofferr
                   2024: roff_unsupp(ROFF_ARGS)
                   2025: {
                   2026:
                   2027:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295     schwarze 2028:            ln, ppos, roff_name[tok]);
1.277     schwarze 2029:        return ROFF_IGN;
1.251     schwarze 2030: }
                   2031:
                   2032: static enum rofferr
1.82      kristaps 2033: roff_cond(ROFF_ARGS)
1.74      kristaps 2034: {
1.173     schwarze 2035:
                   2036:        roffnode_push(r, tok, NULL, ln, ppos);
1.74      kristaps 2037:
1.207     schwarze 2038:        /*
1.134     kristaps 2039:         * An `.el' has no conditional body: it will consume the value
                   2040:         * of the current rstack entry set in prior `ie' calls or
1.207     schwarze 2041:         * defaults to DENY.
1.134     kristaps 2042:         *
                   2043:         * If we're not an `el', however, then evaluate the conditional.
                   2044:         */
1.133     kristaps 2045:
1.238     schwarze 2046:        r->last->rule = tok == ROFF_el ?
1.207     schwarze 2047:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238     schwarze 2048:            roff_evalcond(r, ln, buf->buf, &pos);
1.77      kristaps 2049:
1.134     kristaps 2050:        /*
                   2051:         * An if-else will put the NEGATION of the current evaluated
                   2052:         * conditional into the stack of rules.
                   2053:         */
                   2054:
1.238     schwarze 2055:        if (tok == ROFF_ie) {
1.223     schwarze 2056:                if (r->rstackpos + 1 == r->rstacksz) {
                   2057:                        r->rstacksz += 16;
                   2058:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2059:                            r->rstacksz, sizeof(int));
1.134     kristaps 2060:                }
1.198     schwarze 2061:                r->rstack[++r->rstackpos] = !r->last->rule;
1.82      kristaps 2062:        }
1.88      kristaps 2063:
                   2064:        /* If the parent has false as its rule, then so do we. */
                   2065:
1.198     schwarze 2066:        if (r->last->parent && !r->last->parent->rule)
                   2067:                r->last->rule = 0;
1.88      kristaps 2068:
                   2069:        /*
1.173     schwarze 2070:         * Determine scope.
                   2071:         * If there is nothing on the line after the conditional,
                   2072:         * not even whitespace, use next-line scope.
1.88      kristaps 2073:         */
1.74      kristaps 2074:
1.238     schwarze 2075:        if (buf->buf[pos] == '\0') {
1.173     schwarze 2076:                r->last->endspan = 2;
                   2077:                goto out;
                   2078:        }
                   2079:
1.238     schwarze 2080:        while (buf->buf[pos] == ' ')
1.173     schwarze 2081:                pos++;
                   2082:
                   2083:        /* An opening brace requests multiline scope. */
1.75      kristaps 2084:
1.238     schwarze 2085:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75      kristaps 2086:                r->last->endspan = -1;
                   2087:                pos += 2;
1.272     schwarze 2088:                while (buf->buf[pos] == ' ')
                   2089:                        pos++;
1.173     schwarze 2090:                goto out;
1.207     schwarze 2091:        }
1.74      kristaps 2092:
1.77      kristaps 2093:        /*
1.173     schwarze 2094:         * Anything else following the conditional causes
                   2095:         * single-line scope.  Warn if the scope contains
                   2096:         * nothing but trailing whitespace.
1.77      kristaps 2097:         */
                   2098:
1.238     schwarze 2099:        if (buf->buf[pos] == '\0')
1.216     schwarze 2100:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295     schwarze 2101:                    ln, ppos, roff_name[tok]);
1.77      kristaps 2102:
1.173     schwarze 2103:        r->last->endspan = 1;
1.74      kristaps 2104:
1.173     schwarze 2105: out:
1.75      kristaps 2106:        *offs = pos;
1.277     schwarze 2107:        return ROFF_RERUN;
1.83      schwarze 2108: }
                   2109:
                   2110: static enum rofferr
1.92      schwarze 2111: roff_ds(ROFF_ARGS)
                   2112: {
1.212     schwarze 2113:        char            *string;
                   2114:        const char      *name;
                   2115:        size_t           namesz;
1.96      kristaps 2116:
1.251     schwarze 2117:        /* Ignore groff compatibility mode for now. */
                   2118:
                   2119:        if (tok == ROFF_ds1)
                   2120:                tok = ROFF_ds;
                   2121:        else if (tok == ROFF_as1)
                   2122:                tok = ROFF_as;
                   2123:
1.96      kristaps 2124:        /*
1.212     schwarze 2125:         * The first word is the name of the string.
                   2126:         * If it is empty or terminated by an escape sequence,
                   2127:         * abort the `ds' request without defining anything.
1.96      kristaps 2128:         */
1.92      schwarze 2129:
1.238     schwarze 2130:        name = string = buf->buf + pos;
                   2131:        if (*name == '\0')
1.277     schwarze 2132:                return ROFF_IGN;
1.92      schwarze 2133:
1.212     schwarze 2134:        namesz = roff_getname(r, &string, ln, pos);
1.238     schwarze 2135:        if (name[namesz] == '\\')
1.277     schwarze 2136:                return ROFF_IGN;
1.212     schwarze 2137:
                   2138:        /* Read past the initial double-quote, if any. */
1.238     schwarze 2139:        if (*string == '"')
1.92      schwarze 2140:                string++;
                   2141:
1.96      kristaps 2142:        /* The rest is the value. */
1.212     schwarze 2143:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2144:            ROFF_as == tok);
1.277     schwarze 2145:        return ROFF_IGN;
1.92      schwarze 2146: }
                   2147:
1.204     schwarze 2148: /*
                   2149:  * Parse a single operator, one or two characters long.
                   2150:  * If the operator is recognized, return success and advance the
                   2151:  * parse point, else return failure and let the parse point unchanged.
                   2152:  */
                   2153: static int
                   2154: roff_getop(const char *v, int *pos, char *res)
                   2155: {
                   2156:
                   2157:        *res = v[*pos];
                   2158:
                   2159:        switch (*res) {
1.207     schwarze 2160:        case '+':
                   2161:        case '-':
                   2162:        case '*':
                   2163:        case '/':
                   2164:        case '%':
                   2165:        case '&':
                   2166:        case ':':
1.204     schwarze 2167:                break;
                   2168:        case '<':
                   2169:                switch (v[*pos + 1]) {
1.207     schwarze 2170:                case '=':
1.204     schwarze 2171:                        *res = 'l';
                   2172:                        (*pos)++;
                   2173:                        break;
1.207     schwarze 2174:                case '>':
1.204     schwarze 2175:                        *res = '!';
                   2176:                        (*pos)++;
                   2177:                        break;
1.207     schwarze 2178:                case '?':
1.204     schwarze 2179:                        *res = 'i';
                   2180:                        (*pos)++;
                   2181:                        break;
                   2182:                default:
                   2183:                        break;
                   2184:                }
                   2185:                break;
                   2186:        case '>':
                   2187:                switch (v[*pos + 1]) {
1.207     schwarze 2188:                case '=':
1.204     schwarze 2189:                        *res = 'g';
                   2190:                        (*pos)++;
                   2191:                        break;
1.207     schwarze 2192:                case '?':
1.204     schwarze 2193:                        *res = 'a';
                   2194:                        (*pos)++;
                   2195:                        break;
                   2196:                default:
                   2197:                        break;
                   2198:                }
                   2199:                break;
                   2200:        case '=':
                   2201:                if ('=' == v[*pos + 1])
                   2202:                        (*pos)++;
                   2203:                break;
                   2204:        default:
1.277     schwarze 2205:                return 0;
1.204     schwarze 2206:        }
                   2207:        (*pos)++;
                   2208:
1.277     schwarze 2209:        return *res;
1.204     schwarze 2210: }
                   2211:
                   2212: /*
                   2213:  * Evaluate either a parenthesized numeric expression
                   2214:  * or a single signed integer number.
                   2215:  */
                   2216: static int
1.235     schwarze 2217: roff_evalpar(struct roff *r, int ln,
1.261     schwarze 2218:        const char *v, int *pos, int *res, int flags)
1.204     schwarze 2219: {
                   2220:
                   2221:        if ('(' != v[*pos])
1.277     schwarze 2222:                return roff_getnum(v, pos, res, flags);
1.204     schwarze 2223:
                   2224:        (*pos)++;
1.261     schwarze 2225:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277     schwarze 2226:                return 0;
1.204     schwarze 2227:
1.206     schwarze 2228:        /*
                   2229:         * Omission of the closing parenthesis
                   2230:         * is an error in validation mode,
                   2231:         * but ignored in evaluation mode.
                   2232:         */
                   2233:
1.204     schwarze 2234:        if (')' == v[*pos])
                   2235:                (*pos)++;
1.206     schwarze 2236:        else if (NULL == res)
1.277     schwarze 2237:                return 0;
1.204     schwarze 2238:
1.277     schwarze 2239:        return 1;
1.204     schwarze 2240: }
                   2241:
                   2242: /*
                   2243:  * Evaluate a complete numeric expression.
                   2244:  * Proceed left to right, there is no concept of precedence.
                   2245:  */
                   2246: static int
1.235     schwarze 2247: roff_evalnum(struct roff *r, int ln, const char *v,
1.261     schwarze 2248:        int *pos, int *res, int flags)
1.204     schwarze 2249: {
                   2250:        int              mypos, operand2;
                   2251:        char             operator;
                   2252:
                   2253:        if (NULL == pos) {
                   2254:                mypos = 0;
                   2255:                pos = &mypos;
                   2256:        }
                   2257:
1.261     schwarze 2258:        if (flags & ROFFNUM_WHITE)
1.204     schwarze 2259:                while (isspace((unsigned char)v[*pos]))
                   2260:                        (*pos)++;
                   2261:
1.261     schwarze 2262:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277     schwarze 2263:                return 0;
1.204     schwarze 2264:
                   2265:        while (1) {
1.261     schwarze 2266:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2267:                        while (isspace((unsigned char)v[*pos]))
                   2268:                                (*pos)++;
                   2269:
                   2270:                if ( ! roff_getop(v, pos, &operator))
                   2271:                        break;
                   2272:
1.261     schwarze 2273:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2274:                        while (isspace((unsigned char)v[*pos]))
                   2275:                                (*pos)++;
                   2276:
1.261     schwarze 2277:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277     schwarze 2278:                        return 0;
1.204     schwarze 2279:
1.261     schwarze 2280:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2281:                        while (isspace((unsigned char)v[*pos]))
                   2282:                                (*pos)++;
1.206     schwarze 2283:
                   2284:                if (NULL == res)
                   2285:                        continue;
1.204     schwarze 2286:
                   2287:                switch (operator) {
1.207     schwarze 2288:                case '+':
1.204     schwarze 2289:                        *res += operand2;
                   2290:                        break;
1.207     schwarze 2291:                case '-':
1.204     schwarze 2292:                        *res -= operand2;
                   2293:                        break;
1.207     schwarze 2294:                case '*':
1.204     schwarze 2295:                        *res *= operand2;
                   2296:                        break;
1.207     schwarze 2297:                case '/':
1.244     schwarze 2298:                        if (operand2 == 0) {
1.235     schwarze 2299:                                mandoc_msg(MANDOCERR_DIVZERO,
1.234     kristaps 2300:                                        r->parse, ln, *pos, v);
                   2301:                                *res = 0;
                   2302:                                break;
                   2303:                        }
1.204     schwarze 2304:                        *res /= operand2;
                   2305:                        break;
1.207     schwarze 2306:                case '%':
1.244     schwarze 2307:                        if (operand2 == 0) {
                   2308:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2309:                                        r->parse, ln, *pos, v);
                   2310:                                *res = 0;
                   2311:                                break;
                   2312:                        }
1.204     schwarze 2313:                        *res %= operand2;
                   2314:                        break;
1.207     schwarze 2315:                case '<':
1.204     schwarze 2316:                        *res = *res < operand2;
                   2317:                        break;
1.207     schwarze 2318:                case '>':
1.204     schwarze 2319:                        *res = *res > operand2;
                   2320:                        break;
1.207     schwarze 2321:                case 'l':
1.204     schwarze 2322:                        *res = *res <= operand2;
                   2323:                        break;
1.207     schwarze 2324:                case 'g':
1.204     schwarze 2325:                        *res = *res >= operand2;
                   2326:                        break;
1.207     schwarze 2327:                case '=':
1.204     schwarze 2328:                        *res = *res == operand2;
                   2329:                        break;
1.207     schwarze 2330:                case '!':
1.204     schwarze 2331:                        *res = *res != operand2;
                   2332:                        break;
1.207     schwarze 2333:                case '&':
1.204     schwarze 2334:                        *res = *res && operand2;
                   2335:                        break;
1.207     schwarze 2336:                case ':':
1.204     schwarze 2337:                        *res = *res || operand2;
                   2338:                        break;
1.207     schwarze 2339:                case 'i':
1.204     schwarze 2340:                        if (operand2 < *res)
                   2341:                                *res = operand2;
                   2342:                        break;
1.207     schwarze 2343:                case 'a':
1.204     schwarze 2344:                        if (operand2 > *res)
                   2345:                                *res = operand2;
                   2346:                        break;
                   2347:                default:
                   2348:                        abort();
                   2349:                }
                   2350:        }
1.277     schwarze 2351:        return 1;
1.204     schwarze 2352: }
                   2353:
1.266     schwarze 2354: /* --- register management ------------------------------------------------ */
                   2355:
1.180     schwarze 2356: void
1.187     schwarze 2357: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147     kristaps 2358: {
1.180     schwarze 2359:        struct roffreg  *reg;
                   2360:
                   2361:        /* Search for an existing register with the same name. */
                   2362:        reg = r->regtab;
                   2363:
                   2364:        while (reg && strcmp(name, reg->key.p))
                   2365:                reg = reg->next;
1.147     kristaps 2366:
1.180     schwarze 2367:        if (NULL == reg) {
                   2368:                /* Create a new register. */
                   2369:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2370:                reg->key.p = mandoc_strdup(name);
                   2371:                reg->key.sz = strlen(name);
1.187     schwarze 2372:                reg->val = 0;
1.180     schwarze 2373:                reg->next = r->regtab;
                   2374:                r->regtab = reg;
                   2375:        }
                   2376:
1.187     schwarze 2377:        if ('+' == sign)
                   2378:                reg->val += val;
                   2379:        else if ('-' == sign)
                   2380:                reg->val -= val;
                   2381:        else
                   2382:                reg->val = val;
1.147     kristaps 2383: }
                   2384:
1.192     schwarze 2385: /*
                   2386:  * Handle some predefined read-only number registers.
                   2387:  * For now, return -1 if the requested register is not predefined;
                   2388:  * in case a predefined read-only register having the value -1
                   2389:  * were to turn up, another special value would have to be chosen.
                   2390:  */
                   2391: static int
1.273     schwarze 2392: roff_getregro(const struct roff *r, const char *name)
1.192     schwarze 2393: {
                   2394:
                   2395:        switch (*name) {
1.273     schwarze 2396:        case '$':  /* Number of arguments of the last macro evaluated. */
1.277     schwarze 2397:                return r->argc;
1.207     schwarze 2398:        case 'A':  /* ASCII approximation mode is always off. */
1.277     schwarze 2399:                return 0;
1.207     schwarze 2400:        case 'g':  /* Groff compatibility mode is always on. */
1.277     schwarze 2401:                return 1;
1.207     schwarze 2402:        case 'H':  /* Fixed horizontal resolution. */
1.277     schwarze 2403:                return 24;
1.207     schwarze 2404:        case 'j':  /* Always adjust left margin only. */
1.277     schwarze 2405:                return 0;
1.207     schwarze 2406:        case 'T':  /* Some output device is always defined. */
1.277     schwarze 2407:                return 1;
1.207     schwarze 2408:        case 'V':  /* Fixed vertical resolution. */
1.277     schwarze 2409:                return 40;
1.192     schwarze 2410:        default:
1.277     schwarze 2411:                return -1;
1.192     schwarze 2412:        }
                   2413: }
                   2414:
1.181     schwarze 2415: int
1.180     schwarze 2416: roff_getreg(const struct roff *r, const char *name)
1.147     kristaps 2417: {
1.180     schwarze 2418:        struct roffreg  *reg;
1.192     schwarze 2419:        int              val;
                   2420:
                   2421:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.273     schwarze 2422:                val = roff_getregro(r, name + 1);
1.192     schwarze 2423:                if (-1 != val)
1.277     schwarze 2424:                        return val;
1.192     schwarze 2425:        }
1.180     schwarze 2426:
                   2427:        for (reg = r->regtab; reg; reg = reg->next)
                   2428:                if (0 == strcmp(name, reg->key.p))
1.277     schwarze 2429:                        return reg->val;
1.181     schwarze 2430:
1.277     schwarze 2431:        return 0;
1.181     schwarze 2432: }
                   2433:
                   2434: static int
                   2435: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2436: {
                   2437:        struct roffreg  *reg;
1.192     schwarze 2438:        int              val;
                   2439:
                   2440:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 2441:                val = roff_getregro(r, name + 1);
1.192     schwarze 2442:                if (-1 != val)
1.277     schwarze 2443:                        return val;
1.192     schwarze 2444:        }
1.181     schwarze 2445:
                   2446:        for (reg = r->regtab; reg; reg = reg->next)
                   2447:                if (len == reg->key.sz &&
                   2448:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 2449:                        return reg->val;
1.271     schwarze 2450:
1.277     schwarze 2451:        return 0;
1.271     schwarze 2452: }
                   2453:
                   2454: static int
                   2455: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2456: {
                   2457:        struct roffreg  *reg;
                   2458:        int              val;
                   2459:
                   2460:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 2461:                val = roff_getregro(r, name + 1);
1.271     schwarze 2462:                if (-1 != val)
1.277     schwarze 2463:                        return 1;
1.271     schwarze 2464:        }
                   2465:
                   2466:        for (reg = r->regtab; reg; reg = reg->next)
                   2467:                if (len == reg->key.sz &&
                   2468:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 2469:                        return 1;
1.147     kristaps 2470:
1.277     schwarze 2471:        return 0;
1.147     kristaps 2472: }
                   2473:
1.180     schwarze 2474: static void
                   2475: roff_freereg(struct roffreg *reg)
1.147     kristaps 2476: {
1.180     schwarze 2477:        struct roffreg  *old_reg;
1.147     kristaps 2478:
1.180     schwarze 2479:        while (NULL != reg) {
                   2480:                free(reg->key.p);
                   2481:                old_reg = reg;
                   2482:                reg = reg->next;
                   2483:                free(old_reg);
                   2484:        }
1.147     kristaps 2485: }
1.92      schwarze 2486:
                   2487: static enum rofferr
1.89      kristaps 2488: roff_nr(ROFF_ARGS)
1.83      schwarze 2489: {
1.212     schwarze 2490:        char            *key, *val;
                   2491:        size_t           keysz;
1.138     kristaps 2492:        int              iv;
1.187     schwarze 2493:        char             sign;
1.89      kristaps 2494:
1.238     schwarze 2495:        key = val = buf->buf + pos;
                   2496:        if (*key == '\0')
1.277     schwarze 2497:                return ROFF_IGN;
1.212     schwarze 2498:
                   2499:        keysz = roff_getname(r, &val, ln, pos);
1.238     schwarze 2500:        if (key[keysz] == '\\')
1.277     schwarze 2501:                return ROFF_IGN;
1.212     schwarze 2502:        key[keysz] = '\0';
1.89      kristaps 2503:
1.187     schwarze 2504:        sign = *val;
1.238     schwarze 2505:        if (sign == '+' || sign == '-')
1.187     schwarze 2506:                val++;
                   2507:
1.261     schwarze 2508:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204     schwarze 2509:                roff_setreg(r, key, iv, sign);
1.109     kristaps 2510:
1.277     schwarze 2511:        return ROFF_IGN;
1.203     schwarze 2512: }
                   2513:
                   2514: static enum rofferr
                   2515: roff_rr(ROFF_ARGS)
                   2516: {
                   2517:        struct roffreg  *reg, **prev;
1.212     schwarze 2518:        char            *name, *cp;
                   2519:        size_t           namesz;
1.203     schwarze 2520:
1.238     schwarze 2521:        name = cp = buf->buf + pos;
                   2522:        if (*name == '\0')
1.277     schwarze 2523:                return ROFF_IGN;
1.212     schwarze 2524:        namesz = roff_getname(r, &cp, ln, pos);
                   2525:        name[namesz] = '\0';
1.203     schwarze 2526:
                   2527:        prev = &r->regtab;
                   2528:        while (1) {
                   2529:                reg = *prev;
1.238     schwarze 2530:                if (reg == NULL || !strcmp(name, reg->key.p))
1.203     schwarze 2531:                        break;
                   2532:                prev = &reg->next;
                   2533:        }
1.238     schwarze 2534:        if (reg != NULL) {
1.203     schwarze 2535:                *prev = reg->next;
                   2536:                free(reg->key.p);
                   2537:                free(reg);
                   2538:        }
1.277     schwarze 2539:        return ROFF_IGN;
1.122     schwarze 2540: }
                   2541:
1.266     schwarze 2542: /* --- handler functions for roff requests -------------------------------- */
                   2543:
1.122     schwarze 2544: static enum rofferr
                   2545: roff_rm(ROFF_ARGS)
                   2546: {
                   2547:        const char       *name;
                   2548:        char             *cp;
1.212     schwarze 2549:        size_t            namesz;
1.122     schwarze 2550:
1.238     schwarze 2551:        cp = buf->buf + pos;
                   2552:        while (*cp != '\0') {
1.212     schwarze 2553:                name = cp;
1.238     schwarze 2554:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212     schwarze 2555:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238     schwarze 2556:                if (name[namesz] == '\\')
1.212     schwarze 2557:                        break;
1.122     schwarze 2558:        }
1.277     schwarze 2559:        return ROFF_IGN;
1.178     schwarze 2560: }
                   2561:
                   2562: static enum rofferr
                   2563: roff_it(ROFF_ARGS)
                   2564: {
                   2565:        int              iv;
                   2566:
                   2567:        /* Parse the number of lines. */
1.261     schwarze 2568:
                   2569:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222     schwarze 2570:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238     schwarze 2571:                    ln, ppos, buf->buf + 1);
1.277     schwarze 2572:                return ROFF_IGN;
1.178     schwarze 2573:        }
                   2574:
1.262     schwarze 2575:        while (isspace((unsigned char)buf->buf[pos]))
                   2576:                pos++;
                   2577:
                   2578:        /*
                   2579:         * Arm the input line trap.
                   2580:         * Special-casing "an-trap" is an ugly workaround to cope
                   2581:         * with DocBook stupidly fiddling with man(7) internals.
                   2582:         */
1.261     schwarze 2583:
1.178     schwarze 2584:        roffit_lines = iv;
1.262     schwarze 2585:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2586:            strcmp(buf->buf + pos, "an-trap") ?
                   2587:            buf->buf + pos : "br");
1.277     schwarze 2588:        return ROFF_IGN;
1.175     schwarze 2589: }
                   2590:
                   2591: static enum rofferr
                   2592: roff_Dd(ROFF_ARGS)
                   2593: {
                   2594:        const char *const       *cp;
                   2595:
1.227     schwarze 2596:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175     schwarze 2597:                for (cp = __mdoc_reserved; *cp; cp++)
                   2598:                        roff_setstr(r, *cp, NULL, 0);
                   2599:
1.227     schwarze 2600:        if (r->format == 0)
                   2601:                r->format = MPARSE_MDOC;
                   2602:
1.277     schwarze 2603:        return ROFF_CONT;
1.175     schwarze 2604: }
                   2605:
                   2606: static enum rofferr
                   2607: roff_TH(ROFF_ARGS)
                   2608: {
                   2609:        const char *const       *cp;
                   2610:
1.227     schwarze 2611:        if ((r->options & MPARSE_QUICK) == 0)
1.175     schwarze 2612:                for (cp = __man_reserved; *cp; cp++)
                   2613:                        roff_setstr(r, *cp, NULL, 0);
                   2614:
1.227     schwarze 2615:        if (r->format == 0)
                   2616:                r->format = MPARSE_MAN;
                   2617:
1.277     schwarze 2618:        return ROFF_CONT;
1.109     kristaps 2619: }
                   2620:
                   2621: static enum rofferr
                   2622: roff_TE(ROFF_ARGS)
                   2623: {
                   2624:
                   2625:        if (NULL == r->tbl)
1.221     schwarze 2626:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2627:                    ln, ppos, "TE");
1.258     schwarze 2628:        else if ( ! tbl_end(&r->tbl)) {
                   2629:                free(buf->buf);
                   2630:                buf->buf = mandoc_strdup(".sp");
                   2631:                buf->sz = 4;
1.277     schwarze 2632:                return ROFF_REPARSE;
1.258     schwarze 2633:        }
1.277     schwarze 2634:        return ROFF_IGN;
1.112     kristaps 2635: }
                   2636:
                   2637: static enum rofferr
                   2638: roff_T_(ROFF_ARGS)
                   2639: {
                   2640:
                   2641:        if (NULL == r->tbl)
1.221     schwarze 2642:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2643:                    ln, ppos, "T&");
1.112     kristaps 2644:        else
1.296     schwarze 2645:                tbl_restart(ln, ppos, r->tbl);
1.112     kristaps 2646:
1.277     schwarze 2647:        return ROFF_IGN;
1.109     kristaps 2648: }
                   2649:
1.230     schwarze 2650: /*
                   2651:  * Handle in-line equation delimiters.
                   2652:  */
                   2653: static enum rofferr
1.238     schwarze 2654: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151     kristaps 2655: {
1.233     schwarze 2656:        char            *cp1, *cp2;
                   2657:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151     kristaps 2658:
1.230     schwarze 2659:        /*
                   2660:         * Outside equations, look for an opening delimiter.
                   2661:         * If we are inside an equation, we already know it is
                   2662:         * in-line, or this function wouldn't have been called;
                   2663:         * so look for a closing delimiter.
                   2664:         */
                   2665:
1.238     schwarze 2666:        cp1 = buf->buf + pos;
1.230     schwarze 2667:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2668:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2669:        if (cp2 == NULL)
1.277     schwarze 2670:                return ROFF_CONT;
1.230     schwarze 2671:
1.233     schwarze 2672:        *cp2++ = '\0';
                   2673:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2674:
                   2675:        /* Handle preceding text, protecting whitespace. */
                   2676:
1.238     schwarze 2677:        if (*buf->buf != '\0') {
1.233     schwarze 2678:                if (r->eqn == NULL)
                   2679:                        bef_pr = "\\&";
                   2680:                bef_nl = "\n";
                   2681:        }
                   2682:
                   2683:        /*
                   2684:         * Prepare replacing the delimiter with an equation macro
                   2685:         * and drop leading white space from the equation.
                   2686:         */
1.230     schwarze 2687:
1.233     schwarze 2688:        if (r->eqn == NULL) {
                   2689:                while (*cp2 == ' ')
                   2690:                        cp2++;
                   2691:                mac = ".EQ";
                   2692:        } else
                   2693:                mac = ".EN";
                   2694:
                   2695:        /* Handle following text, protecting whitespace. */
                   2696:
                   2697:        if (*cp2 != '\0') {
                   2698:                aft_nl = "\n";
                   2699:                if (r->eqn != NULL)
                   2700:                        aft_pr = "\\&";
                   2701:        }
                   2702:
                   2703:        /* Do the actual replacement. */
                   2704:
1.238     schwarze 2705:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233     schwarze 2706:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238     schwarze 2707:        free(buf->buf);
                   2708:        buf->buf = cp1;
1.230     schwarze 2709:
                   2710:        /* Toggle the in-line state of the eqn subsystem. */
                   2711:
                   2712:        r->eqn_inline = r->eqn == NULL;
1.277     schwarze 2713:        return ROFF_REPARSE;
1.151     kristaps 2714: }
                   2715:
1.235     schwarze 2716: static enum rofferr
                   2717: roff_EQ(ROFF_ARGS)
1.125     kristaps 2718: {
1.151     kristaps 2719:        struct eqn_node *e;
1.125     kristaps 2720:
1.238     schwarze 2721:        assert(r->eqn == NULL);
1.235     schwarze 2722:        e = eqn_alloc(ppos, ln, r->parse);
1.125     kristaps 2723:
1.230     schwarze 2724:        if (r->last_eqn) {
1.125     kristaps 2725:                r->last_eqn->next = e;
1.230     schwarze 2726:                e->delim = r->last_eqn->delim;
                   2727:                e->odelim = r->last_eqn->odelim;
                   2728:                e->cdelim = r->last_eqn->cdelim;
                   2729:        } else
1.125     kristaps 2730:                r->first_eqn = r->last_eqn = e;
                   2731:
                   2732:        r->eqn = r->last_eqn = e;
1.151     kristaps 2733:
1.238     schwarze 2734:        if (buf->buf[pos] != '\0')
1.235     schwarze 2735:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238     schwarze 2736:                    ".EQ %s", buf->buf + pos);
1.151     kristaps 2737:
1.277     schwarze 2738:        return ROFF_IGN;
1.125     kristaps 2739: }
                   2740:
                   2741: static enum rofferr
                   2742: roff_EN(ROFF_ARGS)
                   2743: {
                   2744:
1.221     schwarze 2745:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.277     schwarze 2746:        return ROFF_IGN;
1.125     kristaps 2747: }
                   2748:
                   2749: static enum rofferr
1.109     kristaps 2750: roff_TS(ROFF_ARGS)
                   2751: {
1.176     schwarze 2752:        struct tbl_node *tbl;
1.89      kristaps 2753:
1.115     kristaps 2754:        if (r->tbl) {
1.221     schwarze 2755:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2756:                    ln, ppos, "TS breaks TS");
1.151     kristaps 2757:                tbl_end(&r->tbl);
1.115     kristaps 2758:        }
1.83      schwarze 2759:
1.176     schwarze 2760:        tbl = tbl_alloc(ppos, ln, r->parse);
1.113     kristaps 2761:
                   2762:        if (r->last_tbl)
1.176     schwarze 2763:                r->last_tbl->next = tbl;
1.113     kristaps 2764:        else
1.176     schwarze 2765:                r->first_tbl = r->last_tbl = tbl;
1.113     kristaps 2766:
1.176     schwarze 2767:        r->tbl = r->last_tbl = tbl;
1.297   ! schwarze 2768:        return ROFF_IGN;
        !          2769: }
        !          2770:
        !          2771: static enum rofferr
        !          2772: roff_onearg(ROFF_ARGS)
        !          2773: {
        !          2774:        struct roff_node        *n;
        !          2775:        char                    *cp;
        !          2776:
        !          2777:        roff_elem_alloc(r->man, ln, ppos, tok);
        !          2778:        n = r->man->last;
        !          2779:
        !          2780:        cp = buf->buf + pos;
        !          2781:        if (*cp != '\0') {
        !          2782:                while (*cp != '\0' && *cp != ' ')
        !          2783:                        cp++;
        !          2784:                while (*cp == ' ')
        !          2785:                        *cp++ = '\0';
        !          2786:                if (*cp != '\0')
        !          2787:                        mandoc_vmsg(MANDOCERR_ARG_EXCESS,
        !          2788:                            r->parse, ln, cp - buf->buf,
        !          2789:                            "%s ... %s", roff_name[tok], cp);
        !          2790:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
        !          2791:        }
        !          2792:
        !          2793:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
        !          2794:        r->man->last = n;
        !          2795:        r->man->next = ROFF_NEXT_SIBLING;
1.277     schwarze 2796:        return ROFF_IGN;
1.251     schwarze 2797: }
                   2798:
                   2799: static enum rofferr
1.296     schwarze 2800: roff_br(ROFF_ARGS)
1.251     schwarze 2801: {
1.296     schwarze 2802:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
                   2803:        if (buf->buf[pos] != '\0')
                   2804:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
                   2805:                    "%s %s", roff_name[tok], buf->buf + pos);
                   2806:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   2807:        r->man->next = ROFF_NEXT_SIBLING;
                   2808:        return ROFF_IGN;
1.92      schwarze 2809: }
                   2810:
1.105     kristaps 2811: static enum rofferr
1.174     kristaps 2812: roff_cc(ROFF_ARGS)
                   2813: {
                   2814:        const char      *p;
                   2815:
1.238     schwarze 2816:        p = buf->buf + pos;
1.174     kristaps 2817:
1.238     schwarze 2818:        if (*p == '\0' || (r->control = *p++) == '.')
1.174     kristaps 2819:                r->control = 0;
                   2820:
1.238     schwarze 2821:        if (*p != '\0')
1.260     schwarze 2822:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   2823:                    ln, p - buf->buf, "cc ... %s", p);
1.174     kristaps 2824:
1.277     schwarze 2825:        return ROFF_IGN;
1.174     kristaps 2826: }
                   2827:
                   2828: static enum rofferr
1.164     kristaps 2829: roff_tr(ROFF_ARGS)
                   2830: {
                   2831:        const char      *p, *first, *second;
                   2832:        size_t           fsz, ssz;
                   2833:        enum mandoc_esc  esc;
                   2834:
1.238     schwarze 2835:        p = buf->buf + pos;
1.164     kristaps 2836:
1.238     schwarze 2837:        if (*p == '\0') {
1.260     schwarze 2838:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277     schwarze 2839:                return ROFF_IGN;
1.164     kristaps 2840:        }
                   2841:
1.238     schwarze 2842:        while (*p != '\0') {
1.164     kristaps 2843:                fsz = ssz = 1;
                   2844:
                   2845:                first = p++;
1.238     schwarze 2846:                if (*first == '\\') {
1.164     kristaps 2847:                        esc = mandoc_escape(&p, NULL, NULL);
1.238     schwarze 2848:                        if (esc == ESCAPE_ERROR) {
1.219     schwarze 2849:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238     schwarze 2850:                                    ln, (int)(p - buf->buf), first);
1.277     schwarze 2851:                                return ROFF_IGN;
1.164     kristaps 2852:                        }
                   2853:                        fsz = (size_t)(p - first);
                   2854:                }
                   2855:
                   2856:                second = p++;
1.238     schwarze 2857:                if (*second == '\\') {
1.164     kristaps 2858:                        esc = mandoc_escape(&p, NULL, NULL);
1.238     schwarze 2859:                        if (esc == ESCAPE_ERROR) {
1.219     schwarze 2860:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238     schwarze 2861:                                    ln, (int)(p - buf->buf), second);
1.277     schwarze 2862:                                return ROFF_IGN;
1.164     kristaps 2863:                        }
                   2864:                        ssz = (size_t)(p - second);
1.238     schwarze 2865:                } else if (*second == '\0') {
1.260     schwarze 2866:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   2867:                            ln, first - buf->buf, "tr %s", first);
1.164     kristaps 2868:                        second = " ";
1.165     kristaps 2869:                        p--;
1.164     kristaps 2870:                }
                   2871:
1.167     kristaps 2872:                if (fsz > 1) {
1.207     schwarze 2873:                        roff_setstrn(&r->xmbtab, first, fsz,
                   2874:                            second, ssz, 0);
1.167     kristaps 2875:                        continue;
                   2876:                }
                   2877:
1.238     schwarze 2878:                if (r->xtab == NULL)
1.207     schwarze 2879:                        r->xtab = mandoc_calloc(128,
                   2880:                            sizeof(struct roffstr));
1.167     kristaps 2881:
                   2882:                free(r->xtab[(int)*first].p);
                   2883:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   2884:                r->xtab[(int)*first].sz = ssz;
1.164     kristaps 2885:        }
                   2886:
1.277     schwarze 2887:        return ROFF_IGN;
1.164     kristaps 2888: }
                   2889:
                   2890: static enum rofferr
1.105     kristaps 2891: roff_so(ROFF_ARGS)
                   2892: {
1.249     schwarze 2893:        char *name, *cp;
1.105     kristaps 2894:
1.238     schwarze 2895:        name = buf->buf + pos;
1.224     schwarze 2896:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105     kristaps 2897:
                   2898:        /*
                   2899:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   2900:         * opening anything that's not in our cwd or anything beneath
                   2901:         * it.  Thus, explicitly disallow traversing up the file-system
                   2902:         * or using absolute paths.
                   2903:         */
                   2904:
1.238     schwarze 2905:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210     schwarze 2906:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   2907:                    ".so %s", name);
1.249     schwarze 2908:                buf->sz = mandoc_asprintf(&cp,
                   2909:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   2910:                free(buf->buf);
                   2911:                buf->buf = cp;
                   2912:                *offs = 0;
1.277     schwarze 2913:                return ROFF_REPARSE;
1.105     kristaps 2914:        }
                   2915:
                   2916:        *offs = pos;
1.277     schwarze 2917:        return ROFF_SO;
1.105     kristaps 2918: }
1.92      schwarze 2919:
1.266     schwarze 2920: /* --- user defined strings and macros ------------------------------------ */
                   2921:
1.106     kristaps 2922: static enum rofferr
                   2923: roff_userdef(ROFF_ARGS)
1.99      kristaps 2924: {
1.263     schwarze 2925:        const char       *arg[9], *ap;
1.106     kristaps 2926:        char             *cp, *n1, *n2;
1.292     schwarze 2927:        int               expand_count, i, ib, ie;
1.263     schwarze 2928:        size_t            asz, rsz;
1.106     kristaps 2929:
                   2930:        /*
                   2931:         * Collect pointers to macro argument strings
1.188     schwarze 2932:         * and NUL-terminate them.
1.106     kristaps 2933:         */
1.263     schwarze 2934:
1.273     schwarze 2935:        r->argc = 0;
1.238     schwarze 2936:        cp = buf->buf + pos;
1.273     schwarze 2937:        for (i = 0; i < 9; i++) {
                   2938:                if (*cp == '\0')
                   2939:                        arg[i] = "";
                   2940:                else {
                   2941:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   2942:                        r->argc = i + 1;
                   2943:                }
                   2944:        }
1.99      kristaps 2945:
1.106     kristaps 2946:        /*
                   2947:         * Expand macro arguments.
1.99      kristaps 2948:         */
1.263     schwarze 2949:
                   2950:        buf->sz = strlen(r->current_string) + 1;
1.292     schwarze 2951:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.263     schwarze 2952:        memcpy(n1, r->current_string, buf->sz);
1.292     schwarze 2953:        expand_count = 0;
1.263     schwarze 2954:        while (*cp != '\0') {
                   2955:
                   2956:                /* Scan ahead for the next argument invocation. */
                   2957:
                   2958:                if (*cp++ != '\\')
                   2959:                        continue;
                   2960:                if (*cp++ != '$')
                   2961:                        continue;
1.274     schwarze 2962:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   2963:                        ib = 0;
                   2964:                        ie = r->argc - 1;
                   2965:                } else {  /* \\$1 .. \\$9 insert one argument */
                   2966:                        ib = ie = *cp - '1';
                   2967:                        if (ib < 0 || ib > 8)
                   2968:                                continue;
                   2969:                }
1.263     schwarze 2970:                cp -= 2;
1.292     schwarze 2971:
                   2972:                /*
                   2973:                 * Prevent infinite recursion.
                   2974:                 */
                   2975:
                   2976:                if (cp >= n2)
                   2977:                        expand_count = 1;
                   2978:                else if (++expand_count > EXPAND_LIMIT) {
                   2979:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   2980:                            ln, (int)(cp - n1), NULL);
1.293     schwarze 2981:                        free(buf->buf);
                   2982:                        buf->buf = n1;
1.292     schwarze 2983:                        return ROFF_IGN;
                   2984:                }
1.263     schwarze 2985:
                   2986:                /*
                   2987:                 * Determine the size of the expanded argument,
                   2988:                 * taking escaping of quotes into account.
                   2989:                 */
                   2990:
1.274     schwarze 2991:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   2992:                for (i = ib; i <= ie; i++) {
                   2993:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   2994:                                asz++;
                   2995:                                if (*ap == '"')
                   2996:                                        asz += 3;
                   2997:                        }
1.263     schwarze 2998:                }
                   2999:                if (asz != 3) {
                   3000:
                   3001:                        /*
                   3002:                         * Determine the size of the rest of the
                   3003:                         * unexpanded macro, including the NUL.
                   3004:                         */
                   3005:
                   3006:                        rsz = buf->sz - (cp - n1) - 3;
                   3007:
                   3008:                        /*
                   3009:                         * When shrinking, move before
                   3010:                         * releasing the storage.
                   3011:                         */
                   3012:
                   3013:                        if (asz < 3)
                   3014:                                memmove(cp + asz, cp + 3, rsz);
                   3015:
                   3016:                        /*
                   3017:                         * Resize the storage for the macro
                   3018:                         * and readjust the parse pointer.
                   3019:                         */
                   3020:
                   3021:                        buf->sz += asz - 3;
                   3022:                        n2 = mandoc_realloc(n1, buf->sz);
                   3023:                        cp = n2 + (cp - n1);
                   3024:                        n1 = n2;
                   3025:
                   3026:                        /*
                   3027:                         * When growing, make room
                   3028:                         * for the expanded argument.
                   3029:                         */
                   3030:
                   3031:                        if (asz > 3)
                   3032:                                memmove(cp + asz, cp + 3, rsz);
                   3033:                }
                   3034:
                   3035:                /* Copy the expanded argument, escaping quotes. */
                   3036:
                   3037:                n2 = cp;
1.274     schwarze 3038:                for (i = ib; i <= ie; i++) {
                   3039:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3040:                                if (*ap == '"') {
                   3041:                                        memcpy(n2, "\\(dq", 4);
                   3042:                                        n2 += 4;
                   3043:                                } else
                   3044:                                        *n2++ = *ap;
                   3045:                        }
                   3046:                        if (i < ie)
                   3047:                                *n2++ = ' ';
1.106     kristaps 3048:                }
1.99      kristaps 3049:        }
                   3050:
1.106     kristaps 3051:        /*
                   3052:         * Replace the macro invocation
                   3053:         * by the expanded macro.
                   3054:         */
1.263     schwarze 3055:
1.238     schwarze 3056:        free(buf->buf);
                   3057:        buf->buf = n1;
1.248     schwarze 3058:        *offs = 0;
1.106     kristaps 3059:
1.277     schwarze 3060:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3061:           ROFF_REPARSE : ROFF_APPEND;
1.99      kristaps 3062: }
1.121     schwarze 3063:
1.212     schwarze 3064: static size_t
1.121     schwarze 3065: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3066: {
                   3067:        char     *name, *cp;
1.212     schwarze 3068:        size_t    namesz;
1.121     schwarze 3069:
                   3070:        name = *cpp;
                   3071:        if ('\0' == *name)
1.277     schwarze 3072:                return 0;
1.121     schwarze 3073:
1.212     schwarze 3074:        /* Read until end of name and terminate it with NUL. */
                   3075:        for (cp = name; 1; cp++) {
                   3076:                if ('\0' == *cp || ' ' == *cp) {
                   3077:                        namesz = cp - name;
                   3078:                        break;
                   3079:                }
1.121     schwarze 3080:                if ('\\' != *cp)
                   3081:                        continue;
1.215     schwarze 3082:                namesz = cp - name;
                   3083:                if ('{' == cp[1] || '}' == cp[1])
                   3084:                        break;
1.121     schwarze 3085:                cp++;
                   3086:                if ('\\' == *cp)
                   3087:                        continue;
1.224     schwarze 3088:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3089:                    "%.*s", (int)(cp - name + 1), name);
1.212     schwarze 3090:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3091:                break;
1.121     schwarze 3092:        }
                   3093:
                   3094:        /* Read past spaces. */
                   3095:        while (' ' == *cp)
                   3096:                cp++;
                   3097:
                   3098:        *cpp = cp;
1.277     schwarze 3099:        return namesz;
1.121     schwarze 3100: }
                   3101:
1.106     kristaps 3102: /*
                   3103:  * Store *string into the user-defined string called *name.
                   3104:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193     schwarze 3105:  * append == 0: replace mode
                   3106:  * append == 1: single-line append mode
                   3107:  * append == 2: multiline append mode, append '\n' after each call
1.106     kristaps 3108:  */
1.94      kristaps 3109: static void
1.106     kristaps 3110: roff_setstr(struct roff *r, const char *name, const char *string,
1.193     schwarze 3111:        int append)
1.92      schwarze 3112: {
1.164     kristaps 3113:
                   3114:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.207     schwarze 3115:            string ? strlen(string) : 0, append);
1.164     kristaps 3116: }
                   3117:
                   3118: static void
1.166     kristaps 3119: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193     schwarze 3120:                const char *string, size_t stringsz, int append)
1.164     kristaps 3121: {
1.166     kristaps 3122:        struct roffkv   *n;
1.164     kristaps 3123:        char            *c;
                   3124:        int              i;
                   3125:        size_t           oldch, newch;
1.92      schwarze 3126:
1.106     kristaps 3127:        /* Search for an existing string with the same name. */
1.164     kristaps 3128:        n = *r;
                   3129:
1.211     schwarze 3130:        while (n && (namesz != n->key.sz ||
                   3131:                        strncmp(n->key.p, name, namesz)))
1.92      schwarze 3132:                n = n->next;
1.94      kristaps 3133:
                   3134:        if (NULL == n) {
1.106     kristaps 3135:                /* Create a new string table entry. */
1.166     kristaps 3136:                n = mandoc_malloc(sizeof(struct roffkv));
                   3137:                n->key.p = mandoc_strndup(name, namesz);
                   3138:                n->key.sz = namesz;
                   3139:                n->val.p = NULL;
                   3140:                n->val.sz = 0;
1.164     kristaps 3141:                n->next = *r;
                   3142:                *r = n;
1.193     schwarze 3143:        } else if (0 == append) {
1.166     kristaps 3144:                free(n->val.p);
                   3145:                n->val.p = NULL;
                   3146:                n->val.sz = 0;
1.106     kristaps 3147:        }
                   3148:
                   3149:        if (NULL == string)
                   3150:                return;
                   3151:
                   3152:        /*
                   3153:         * One additional byte for the '\n' in multiline mode,
                   3154:         * and one for the terminating '\0'.
                   3155:         */
1.193     schwarze 3156:        newch = stringsz + (1 < append ? 2u : 1u);
1.164     kristaps 3157:
1.166     kristaps 3158:        if (NULL == n->val.p) {
                   3159:                n->val.p = mandoc_malloc(newch);
                   3160:                *n->val.p = '\0';
1.106     kristaps 3161:                oldch = 0;
                   3162:        } else {
1.166     kristaps 3163:                oldch = n->val.sz;
                   3164:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106     kristaps 3165:        }
                   3166:
                   3167:        /* Skip existing content in the destination buffer. */
1.166     kristaps 3168:        c = n->val.p + (int)oldch;
1.106     kristaps 3169:
                   3170:        /* Append new content to the destination buffer. */
1.164     kristaps 3171:        i = 0;
                   3172:        while (i < (int)stringsz) {
1.106     kristaps 3173:                /*
                   3174:                 * Rudimentary roff copy mode:
                   3175:                 * Handle escaped backslashes.
                   3176:                 */
1.164     kristaps 3177:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3178:                        i++;
                   3179:                *c++ = string[i++];
1.106     kristaps 3180:        }
1.94      kristaps 3181:
1.106     kristaps 3182:        /* Append terminating bytes. */
1.193     schwarze 3183:        if (1 < append)
1.106     kristaps 3184:                *c++ = '\n';
1.163     kristaps 3185:
1.106     kristaps 3186:        *c = '\0';
1.166     kristaps 3187:        n->val.sz = (int)(c - n->val.p);
1.92      schwarze 3188: }
                   3189:
1.94      kristaps 3190: static const char *
                   3191: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92      schwarze 3192: {
1.166     kristaps 3193:        const struct roffkv *n;
1.191     schwarze 3194:        int i;
1.92      schwarze 3195:
1.164     kristaps 3196:        for (n = r->strtab; n; n = n->next)
1.207     schwarze 3197:                if (0 == strncmp(name, n->key.p, len) &&
                   3198:                    '\0' == n->key.p[(int)len])
1.277     schwarze 3199:                        return n->val.p;
1.191     schwarze 3200:
                   3201:        for (i = 0; i < PREDEFS_MAX; i++)
                   3202:                if (0 == strncmp(name, predefs[i].name, len) &&
                   3203:                                '\0' == predefs[i].name[(int)len])
1.277     schwarze 3204:                        return predefs[i].str;
1.94      kristaps 3205:
1.277     schwarze 3206:        return NULL;
1.92      schwarze 3207: }
                   3208:
1.94      kristaps 3209: static void
1.167     kristaps 3210: roff_freestr(struct roffkv *r)
1.92      schwarze 3211: {
1.166     kristaps 3212:        struct roffkv    *n, *nn;
1.92      schwarze 3213:
1.167     kristaps 3214:        for (n = r; n; n = nn) {
1.166     kristaps 3215:                free(n->key.p);
                   3216:                free(n->val.p);
1.92      schwarze 3217:                nn = n->next;
                   3218:                free(n);
                   3219:        }
1.114     kristaps 3220: }
1.266     schwarze 3221:
                   3222: /* --- accessors and utility functions ------------------------------------ */
1.114     kristaps 3223:
                   3224: const struct tbl_span *
                   3225: roff_span(const struct roff *r)
                   3226: {
1.207     schwarze 3227:
1.277     schwarze 3228:        return r->tbl ? tbl_span(r->tbl) : NULL;
1.125     kristaps 3229: }
                   3230:
                   3231: const struct eqn *
                   3232: roff_eqn(const struct roff *r)
                   3233: {
1.207     schwarze 3234:
1.277     schwarze 3235:        return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.164     kristaps 3236: }
                   3237:
                   3238: /*
                   3239:  * Duplicate an input string, making the appropriate character
                   3240:  * conversations (as stipulated by `tr') along the way.
                   3241:  * Returns a heap-allocated string with all the replacements made.
                   3242:  */
                   3243: char *
                   3244: roff_strdup(const struct roff *r, const char *p)
                   3245: {
1.166     kristaps 3246:        const struct roffkv *cp;
1.164     kristaps 3247:        char            *res;
                   3248:        const char      *pp;
                   3249:        size_t           ssz, sz;
                   3250:        enum mandoc_esc  esc;
                   3251:
1.167     kristaps 3252:        if (NULL == r->xmbtab && NULL == r->xtab)
1.277     schwarze 3253:                return mandoc_strdup(p);
1.164     kristaps 3254:        else if ('\0' == *p)
1.277     schwarze 3255:                return mandoc_strdup("");
1.164     kristaps 3256:
                   3257:        /*
                   3258:         * Step through each character looking for term matches
                   3259:         * (remember that a `tr' can be invoked with an escape, which is
                   3260:         * a glyph but the escape is multi-character).
                   3261:         * We only do this if the character hash has been initialised
                   3262:         * and the string is >0 length.
                   3263:         */
                   3264:
                   3265:        res = NULL;
                   3266:        ssz = 0;
                   3267:
                   3268:        while ('\0' != *p) {
1.289     schwarze 3269:                assert((unsigned int)*p < 128);
                   3270:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167     kristaps 3271:                        sz = r->xtab[(int)*p].sz;
                   3272:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3273:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3274:                        ssz += sz;
                   3275:                        p++;
                   3276:                        continue;
                   3277:                } else if ('\\' != *p) {
                   3278:                        res = mandoc_realloc(res, ssz + 2);
                   3279:                        res[ssz++] = *p++;
                   3280:                        continue;
                   3281:                }
                   3282:
1.164     kristaps 3283:                /* Search for term matches. */
1.167     kristaps 3284:                for (cp = r->xmbtab; cp; cp = cp->next)
1.166     kristaps 3285:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164     kristaps 3286:                                break;
                   3287:
                   3288:                if (NULL != cp) {
                   3289:                        /*
                   3290:                         * A match has been found.
                   3291:                         * Append the match to the array and move
                   3292:                         * forward by its keysize.
                   3293:                         */
1.207     schwarze 3294:                        res = mandoc_realloc(res,
                   3295:                            ssz + cp->val.sz + 1);
1.166     kristaps 3296:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3297:                        ssz += cp->val.sz;
                   3298:                        p += (int)cp->key.sz;
1.164     kristaps 3299:                        continue;
                   3300:                }
                   3301:
1.167     kristaps 3302:                /*
                   3303:                 * Handle escapes carefully: we need to copy
                   3304:                 * over just the escape itself, or else we might
                   3305:                 * do replacements within the escape itself.
                   3306:                 * Make sure to pass along the bogus string.
                   3307:                 */
                   3308:                pp = p++;
                   3309:                esc = mandoc_escape(&p, NULL, NULL);
                   3310:                if (ESCAPE_ERROR == esc) {
                   3311:                        sz = strlen(pp);
1.164     kristaps 3312:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3313:                        memcpy(res + ssz, pp, sz);
1.167     kristaps 3314:                        break;
1.164     kristaps 3315:                }
1.207     schwarze 3316:                /*
                   3317:                 * We bail out on bad escapes.
1.167     kristaps 3318:                 * No need to warn: we already did so when
                   3319:                 * roff_res() was called.
                   3320:                 */
                   3321:                sz = (int)(p - pp);
                   3322:                res = mandoc_realloc(res, ssz + sz + 1);
                   3323:                memcpy(res + ssz, pp, sz);
                   3324:                ssz += sz;
1.164     kristaps 3325:        }
                   3326:
                   3327:        res[(int)ssz] = '\0';
1.277     schwarze 3328:        return res;
1.227     schwarze 3329: }
                   3330:
                   3331: int
                   3332: roff_getformat(const struct roff *r)
                   3333: {
                   3334:
1.277     schwarze 3335:        return r->format;
1.174     kristaps 3336: }
                   3337:
                   3338: /*
1.207     schwarze 3339:  * Find out whether a line is a macro line or not.
1.174     kristaps 3340:  * If it is, adjust the current position and return one; if it isn't,
                   3341:  * return zero and don't change the current position.
                   3342:  * If the control character has been set with `.cc', then let that grain
                   3343:  * precedence.
                   3344:  * This is slighly contrary to groff, where using the non-breaking
                   3345:  * control character when `cc' has been invoked will cause the
                   3346:  * non-breaking macro contents to be printed verbatim.
                   3347:  */
                   3348: int
                   3349: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3350: {
                   3351:        int             pos;
                   3352:
                   3353:        pos = *ppos;
                   3354:
                   3355:        if (0 != r->control && cp[pos] == r->control)
                   3356:                pos++;
                   3357:        else if (0 != r->control)
1.277     schwarze 3358:                return 0;
1.174     kristaps 3359:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3360:                pos += 2;
                   3361:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3362:                pos++;
                   3363:        else
1.277     schwarze 3364:                return 0;
1.174     kristaps 3365:
                   3366:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3367:                pos++;
                   3368:
                   3369:        *ppos = pos;
1.277     schwarze 3370:        return 1;
1.74      kristaps 3371: }

CVSweb