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

Annotation of mandoc/roff.c, Revision 1.398

1.398   ! schwarze    1: /* $Id: roff.c,v 1.397 2023/10/21 17:10:18 schwarze Exp $ */
1.1       kristaps    2: /*
1.381     schwarze    3:  * Copyright (c) 2010-2015, 2017-2022 Ingo Schwarze <schwarze@openbsd.org>
1.267     schwarze    4:  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
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.
1.372     schwarze   17:  *
                     18:  * Implementation of the roff(7) parser for mandoc(1).
1.1       kristaps   19:  */
1.66      kristaps   20: #include "config.h"
1.225     schwarze   21:
                     22: #include <sys/types.h>
1.30      kristaps   23:
1.67      kristaps   24: #include <assert.h>
1.85      kristaps   25: #include <ctype.h>
1.245     schwarze   26: #include <limits.h>
1.295     schwarze   27: #include <stddef.h>
                     28: #include <stdint.h>
1.178     schwarze   29: #include <stdio.h>
1.1       kristaps   30: #include <stdlib.h>
1.67      kristaps   31: #include <string.h>
1.1       kristaps   32:
1.201     schwarze   33: #include "mandoc_aux.h"
1.295     schwarze   34: #include "mandoc_ohash.h"
1.345     schwarze   35: #include "mandoc.h"
1.265     schwarze   36: #include "roff.h"
1.349     schwarze   37: #include "mandoc_parse.h"
1.226     schwarze   38: #include "libmandoc.h"
1.266     schwarze   39: #include "roff_int.h"
1.346     schwarze   40: #include "tbl_parse.h"
1.347     schwarze   41: #include "eqn_parse.h"
1.82      kristaps   42:
1.170     schwarze   43: /* Maximum number of string expansions per line, to break infinite loops. */
                     44: #define        EXPAND_LIMIT    1000
                     45:
1.315     schwarze   46: /* Types of definitions of macros and strings. */
                     47: #define        ROFFDEF_USER    (1 << 1)  /* User-defined. */
                     48: #define        ROFFDEF_PRE     (1 << 2)  /* Predefined. */
                     49: #define        ROFFDEF_REN     (1 << 3)  /* Renamed standard macro. */
                     50: #define        ROFFDEF_STD     (1 << 4)  /* mdoc(7) or man(7) macro. */
                     51: #define        ROFFDEF_ANY     (ROFFDEF_USER | ROFFDEF_PRE | \
                     52:                         ROFFDEF_REN | ROFFDEF_STD)
1.325     schwarze   53: #define        ROFFDEF_UNDEF   (1 << 5)  /* Completely undefined. */
1.315     schwarze   54:
1.266     schwarze   55: /* --- data types --------------------------------------------------------- */
                     56:
1.147     kristaps   57: /*
1.167     kristaps   58:  * An incredibly-simple string buffer.
                     59:  */
1.94      kristaps   60: struct roffstr {
1.167     kristaps   61:        char            *p; /* nil-terminated buffer */
                     62:        size_t           sz; /* saved strlen(p) */
1.166     kristaps   63: };
                     64:
                     65: /*
1.167     kristaps   66:  * A key-value roffstr pair as part of a singly-linked list.
1.166     kristaps   67:  */
                     68: struct roffkv {
                     69:        struct roffstr   key;
                     70:        struct roffstr   val;
                     71:        struct roffkv   *next; /* next in list */
1.94      kristaps   72: };
                     73:
1.180     schwarze   74: /*
                     75:  * A single number register as part of a singly-linked list.
                     76:  */
                     77: struct roffreg {
                     78:        struct roffstr   key;
1.181     schwarze   79:        int              val;
1.327     schwarze   80:        int              step;
1.180     schwarze   81:        struct roffreg  *next;
                     82: };
                     83:
1.295     schwarze   84: /*
                     85:  * Association of request and macro names with token IDs.
                     86:  */
                     87: struct roffreq {
                     88:        enum roff_tok    tok;
                     89:        char             name[];
                     90: };
                     91:
1.339     schwarze   92: /*
                     93:  * A macro processing context.
                     94:  * More than one is needed when macro calls are nested.
                     95:  */
                     96: struct mctx {
                     97:        char            **argv;
                     98:        int              argc;
                     99:        int              argsz;
                    100: };
                    101:
1.67      kristaps  102: struct roff {
1.296     schwarze  103:        struct roff_man *man; /* mdoc or man parser */
1.67      kristaps  104:        struct roffnode *last; /* leaf of stack */
1.339     schwarze  105:        struct mctx     *mstack; /* stack of macro contexts */
1.223     schwarze  106:        int             *rstack; /* stack of inverted `ie' values */
1.295     schwarze  107:        struct ohash    *reqtab; /* request lookup table */
1.180     schwarze  108:        struct roffreg  *regtab; /* number registers */
1.166     kristaps  109:        struct roffkv   *strtab; /* user-defined strings & macros */
1.306     schwarze  110:        struct roffkv   *rentab; /* renamed strings & macros */
1.167     kristaps  111:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                    112:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.106     kristaps  113:        const char      *current_string; /* value of last called user macro */
1.118     kristaps  114:        struct tbl_node *first_tbl; /* first table parsed */
                    115:        struct tbl_node *last_tbl; /* last table parsed */
                    116:        struct tbl_node *tbl; /* current table being parsed */
1.319     schwarze  117:        struct eqn_node *last_eqn; /* equation parser */
                    118:        struct eqn_node *eqn; /* active equation parser */
1.230     schwarze  119:        int              eqn_inline; /* current equation is inline */
1.223     schwarze  120:        int              options; /* parse options */
1.339     schwarze  121:        int              mstacksz; /* current size of mstack */
                    122:        int              mstackpos; /* position in mstack */
1.223     schwarze  123:        int              rstacksz; /* current size limit of rstack */
                    124:        int              rstackpos; /* position in rstack */
1.227     schwarze  125:        int              format; /* current file in mdoc or man format */
1.223     schwarze  126:        char             control; /* control character */
1.303     schwarze  127:        char             escape; /* escape character */
1.79      kristaps  128: };
                    129:
1.364     schwarze  130: /*
                    131:  * A macro definition, condition, or ignored block.
                    132:  */
1.67      kristaps  133: struct roffnode {
1.294     schwarze  134:        enum roff_tok    tok; /* type of node */
1.67      kristaps  135:        struct roffnode *parent; /* up one in stack */
                    136:        int              line; /* parse line */
                    137:        int              col; /* parse col */
1.106     kristaps  138:        char            *name; /* node name, e.g. macro name */
1.364     schwarze  139:        char            *end; /* custom end macro of the block */
                    140:        int              endspan; /* scope to: 1=eol 2=next line -1=\} */
                    141:        int              rule; /* content is: 1=evaluated 0=skipped */
1.67      kristaps  142: };
                    143:
                    144: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.294     schwarze  145:                         enum roff_tok tok, /* tok of macro */ \
1.238     schwarze  146:                         struct buf *buf, /* input buffer */ \
1.67      kristaps  147:                         int ln, /* parse line */ \
1.75      kristaps  148:                         int ppos, /* original pos in buffer */ \
                    149:                         int pos, /* current pos in buffer */ \
1.74      kristaps  150:                         int *offs /* reset offset of buffer data */
1.67      kristaps  151:
1.340     schwarze  152: typedef        int (*roffproc)(ROFF_ARGS);
1.67      kristaps  153:
                    154: struct roffmac {
1.79      kristaps  155:        roffproc         proc; /* process new macro */
                    156:        roffproc         text; /* process as child text of macro */
                    157:        roffproc         sub; /* process as child of macro */
                    158:        int              flags;
                    159: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.67      kristaps  160: };
                    161:
1.141     kristaps  162: struct predef {
                    163:        const char      *name; /* predefined input name */
                    164:        const char      *str; /* replacement symbol */
                    165: };
                    166:
                    167: #define        PREDEF(__name, __str) \
                    168:        { (__name), (__str) },
                    169:
1.266     schwarze  170: /* --- function prototypes ------------------------------------------------ */
                    171:
1.340     schwarze  172: static int              roffnode_cleanscope(struct roff *);
                    173: static int              roffnode_pop(struct roff *);
1.294     schwarze  174: static void             roffnode_push(struct roff *, enum roff_tok,
1.155     kristaps  175:                                const char *, int, int);
1.346     schwarze  176: static void             roff_addtbl(struct roff_man *, int, struct tbl_node *);
1.340     schwarze  177: static int              roff_als(ROFF_ARGS);
                    178: static int              roff_block(ROFF_ARGS);
                    179: static int              roff_block_text(ROFF_ARGS);
                    180: static int              roff_block_sub(ROFF_ARGS);
1.364     schwarze  181: static int              roff_break(ROFF_ARGS);
1.340     schwarze  182: static int              roff_cblock(ROFF_ARGS);
                    183: static int              roff_cc(ROFF_ARGS);
                    184: static int              roff_ccond(struct roff *, int, int);
1.341     schwarze  185: static int              roff_char(ROFF_ARGS);
1.340     schwarze  186: static int              roff_cond(ROFF_ARGS);
1.375     schwarze  187: static int              roff_cond_checkend(ROFF_ARGS);
1.340     schwarze  188: static int              roff_cond_text(ROFF_ARGS);
                    189: static int              roff_cond_sub(ROFF_ARGS);
                    190: static int              roff_ds(ROFF_ARGS);
                    191: static int              roff_ec(ROFF_ARGS);
                    192: static int              roff_eo(ROFF_ARGS);
                    193: static int              roff_eqndelim(struct roff *, struct buf *, int);
1.372     schwarze  194: static int              roff_evalcond(struct roff *, int, char *, int *);
1.234     kristaps  195: static int              roff_evalnum(struct roff *, int,
                    196:                                const char *, int *, int *, int);
                    197: static int              roff_evalpar(struct roff *, int,
1.261     schwarze  198:                                const char *, int *, int *, int);
1.198     schwarze  199: static int              roff_evalstrcond(const char *, int *);
1.355     schwarze  200: static int              roff_expand(struct roff *, struct buf *,
                    201:                                int, int, char);
1.388     schwarze  202: static void             roff_expand_patch(struct buf *, int,
                    203:                                const char *, int);
1.155     kristaps  204: static void             roff_free1(struct roff *);
1.180     schwarze  205: static void             roff_freereg(struct roffreg *);
1.167     kristaps  206: static void             roff_freestr(struct roffkv *);
1.212     schwarze  207: static size_t           roff_getname(struct roff *, char **, int, int);
1.261     schwarze  208: static int              roff_getnum(const char *, int *, int *, int);
1.184     schwarze  209: static int              roff_getop(const char *, int *, char *);
1.327     schwarze  210: static int              roff_getregn(struct roff *,
                    211:                                const char *, size_t, char);
1.273     schwarze  212: static int              roff_getregro(const struct roff *,
                    213:                                const char *name);
1.325     schwarze  214: static const char      *roff_getstrn(struct roff *,
1.315     schwarze  215:                                const char *, size_t, int *);
1.271     schwarze  216: static int              roff_hasregn(const struct roff *,
                    217:                                const char *, size_t);
1.340     schwarze  218: static int              roff_insec(ROFF_ARGS);
                    219: static int              roff_it(ROFF_ARGS);
                    220: static int              roff_line_ignore(ROFF_ARGS);
1.265     schwarze  221: static void             roff_man_alloc1(struct roff_man *);
                    222: static void             roff_man_free1(struct roff_man *);
1.340     schwarze  223: static int              roff_manyarg(ROFF_ARGS);
1.384     schwarze  224: static int              roff_mc(ROFF_ARGS);
1.358     schwarze  225: static int              roff_noarg(ROFF_ARGS);
1.340     schwarze  226: static int              roff_nop(ROFF_ARGS);
                    227: static int              roff_nr(ROFF_ARGS);
                    228: static int              roff_onearg(ROFF_ARGS);
1.294     schwarze  229: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.214     schwarze  230:                                int, int);
1.387     schwarze  231: static int              roff_parse_comment(struct roff *, struct buf *,
                    232:                                int, int, char);
1.340     schwarze  233: static int              roff_parsetext(struct roff *, struct buf *,
1.305     schwarze  234:                                int, int *);
1.340     schwarze  235: static int              roff_renamed(ROFF_ARGS);
1.386     schwarze  236: static int              roff_req_or_macro(ROFF_ARGS);
1.340     schwarze  237: static int              roff_return(ROFF_ARGS);
                    238: static int              roff_rm(ROFF_ARGS);
                    239: static int              roff_rn(ROFF_ARGS);
                    240: static int              roff_rr(ROFF_ARGS);
1.326     schwarze  241: static void             roff_setregn(struct roff *, const char *,
1.327     schwarze  242:                                size_t, int, char, int);
1.94      kristaps  243: static void             roff_setstr(struct roff *,
1.106     kristaps  244:                                const char *, const char *, int);
1.207     schwarze  245: static void             roff_setstrn(struct roffkv **, const char *,
1.164     kristaps  246:                                size_t, const char *, size_t, int);
1.340     schwarze  247: static int              roff_shift(ROFF_ARGS);
                    248: static int              roff_so(ROFF_ARGS);
                    249: static int              roff_tr(ROFF_ARGS);
                    250: static int              roff_Dd(ROFF_ARGS);
                    251: static int              roff_TE(ROFF_ARGS);
                    252: static int              roff_TS(ROFF_ARGS);
                    253: static int              roff_EQ(ROFF_ARGS);
                    254: static int              roff_EN(ROFF_ARGS);
                    255: static int              roff_T_(ROFF_ARGS);
                    256: static int              roff_unsupp(ROFF_ARGS);
                    257: static int              roff_userdef(ROFF_ARGS);
1.67      kristaps  258:
1.266     schwarze  259: /* --- constant data ------------------------------------------------------ */
                    260:
1.261     schwarze  261: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    262: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    263:
1.294     schwarze  264: const char *__roff_name[MAN_MAX + 1] = {
1.358     schwarze  265:        "br",           "ce",           "fi",           "ft",
                    266:        "ll",           "mc",           "nf",
                    267:        "po",           "rj",           "sp",
1.312     schwarze  268:        "ta",           "ti",           NULL,
1.294     schwarze  269:        "ab",           "ad",           "af",           "aln",
                    270:        "als",          "am",           "am1",          "ami",
                    271:        "ami1",         "as",           "as1",          "asciify",
                    272:        "backtrace",    "bd",           "bleedat",      "blm",
                    273:         "box",         "boxa",         "bp",           "BP",
                    274:        "break",        "breakchar",    "brnl",         "brp",
1.305     schwarze  275:        "brpnl",        "c2",           "cc",
1.294     schwarze  276:        "cf",           "cflags",       "ch",           "char",
                    277:        "chop",         "class",        "close",        "CL",
                    278:        "color",        "composite",    "continue",     "cp",
                    279:        "cropat",       "cs",           "cu",           "da",
                    280:        "dch",          "Dd",           "de",           "de1",
                    281:        "defcolor",     "dei",          "dei1",         "device",
                    282:        "devicem",      "di",           "do",           "ds",
                    283:        "ds1",          "dwh",          "dt",           "ec",
                    284:        "ecr",          "ecs",          "el",           "em",
                    285:        "EN",           "eo",           "EP",           "EQ",
                    286:        "errprint",     "ev",           "evc",          "ex",
                    287:        "fallback",     "fam",          "fc",           "fchar",
                    288:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    289:        "fl",           "flig",         "fp",           "fps",
                    290:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    291:        "fzoom",        "gcolor",       "hc",           "hcode",
                    292:        "hidechar",     "hla",          "hlm",          "hpf",
                    293:        "hpfa",         "hpfcode",      "hw",           "hy",
                    294:        "hylang",       "hylen",        "hym",          "hypp",
                    295:        "hys",          "ie",           "if",           "ig",
                    296:        "index",        "it",           "itc",          "IX",
                    297:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    298:        "lc",           "lc_ctype",     "lds",          "length",
                    299:        "letadj",       "lf",           "lg",           "lhang",
                    300:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.304     schwarze  301:        "ls",           "lsm",          "lt",
1.294     schwarze  302:        "mediasize",    "minss",        "mk",           "mso",
                    303:        "na",           "ne",           "nh",           "nhychar",
                    304:        "nm",           "nn",           "nop",          "nr",
                    305:        "nrf",          "nroff",        "ns",           "nx",
                    306:        "open",         "opena",        "os",           "output",
                    307:        "padj",         "papersize",    "pc",           "pev",
                    308:        "pi",           "PI",           "pl",           "pm",
1.312     schwarze  309:        "pn",           "pnr",          "ps",
1.294     schwarze  310:        "psbb",         "pshape",       "pso",          "ptr",
                    311:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.309     schwarze  312:        "return",       "rfschar",      "rhang",
1.294     schwarze  313:        "rm",           "rn",           "rnn",          "rr",
                    314:        "rs",           "rt",           "schar",        "sentchar",
                    315:        "shc",          "shift",        "sizes",        "so",
                    316:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    317:        "sty",          "substring",    "sv",           "sy",
1.300     schwarze  318:        "T&",           "tc",           "TE",
1.301     schwarze  319:        "TH",           "tkf",          "tl",
1.294     schwarze  320:        "tm",           "tm1",          "tmc",          "tr",
                    321:        "track",        "transchar",    "trf",          "trimat",
                    322:        "trin",         "trnt",         "troff",        "TS",
                    323:        "uf",           "ul",           "unformat",     "unwatch",
                    324:        "unwatchn",     "vpt",          "vs",           "warn",
                    325:        "warnscale",    "watch",        "watchlength",  "watchn",
                    326:        "wh",           "while",        "write",        "writec",
                    327:        "writem",       "xflag",        ".",            NULL,
1.306     schwarze  328:        NULL,           "text",
1.294     schwarze  329:        "Dd",           "Dt",           "Os",           "Sh",
                    330:        "Ss",           "Pp",           "D1",           "Dl",
                    331:        "Bd",           "Ed",           "Bl",           "El",
                    332:        "It",           "Ad",           "An",           "Ap",
                    333:        "Ar",           "Cd",           "Cm",           "Dv",
                    334:        "Er",           "Ev",           "Ex",           "Fa",
                    335:        "Fd",           "Fl",           "Fn",           "Ft",
                    336:        "Ic",           "In",           "Li",           "Nd",
                    337:        "Nm",           "Op",           "Ot",           "Pa",
                    338:        "Rv",           "St",           "Va",           "Vt",
                    339:        "Xr",           "%A",           "%B",           "%D",
                    340:        "%I",           "%J",           "%N",           "%O",
                    341:        "%P",           "%R",           "%T",           "%V",
                    342:        "Ac",           "Ao",           "Aq",           "At",
                    343:        "Bc",           "Bf",           "Bo",           "Bq",
                    344:        "Bsx",          "Bx",           "Db",           "Dc",
                    345:        "Do",           "Dq",           "Ec",           "Ef",
                    346:        "Em",           "Eo",           "Fx",           "Ms",
                    347:        "No",           "Ns",           "Nx",           "Ox",
                    348:        "Pc",           "Pf",           "Po",           "Pq",
                    349:        "Qc",           "Ql",           "Qo",           "Qq",
                    350:        "Re",           "Rs",           "Sc",           "So",
                    351:        "Sq",           "Sm",           "Sx",           "Sy",
                    352:        "Tn",           "Ux",           "Xc",           "Xo",
                    353:        "Fo",           "Fc",           "Oo",           "Oc",
                    354:        "Bk",           "Ek",           "Bt",           "Hf",
                    355:        "Fr",           "Ud",           "Lb",           "Lp",
                    356:        "Lk",           "Mt",           "Brq",          "Bro",
                    357:        "Brc",          "%C",           "Es",           "En",
1.299     schwarze  358:        "Dx",           "%Q",           "%U",           "Ta",
1.369     schwarze  359:        "Tg",           NULL,
1.294     schwarze  360:        "TH",           "SH",           "SS",           "TP",
1.332     schwarze  361:        "TQ",
1.294     schwarze  362:        "LP",           "PP",           "P",            "IP",
                    363:        "HP",           "SM",           "SB",           "BI",
                    364:        "IB",           "BR",           "RB",           "R",
                    365:        "B",            "I",            "IR",           "RI",
                    366:        "RE",           "RS",           "DT",           "UC",
1.297     schwarze  367:        "PD",           "AT",           "in",
1.333     schwarze  368:        "SY",           "YS",           "OP",
                    369:        "EX",           "EE",           "UR",
1.317     schwarze  370:        "UE",           "MT",           "ME",           NULL
1.294     schwarze  371: };
                    372: const  char *const *roff_name = __roff_name;
                    373:
                    374: static struct roffmac   roffs[TOKEN_NONE] = {
1.358     schwarze  375:        { roff_noarg, NULL, NULL, 0 },  /* br */
1.305     schwarze  376:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.358     schwarze  377:        { roff_noarg, NULL, NULL, 0 },  /* fi */
1.297     schwarze  378:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.298     schwarze  379:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.384     schwarze  380:        { roff_mc, NULL, NULL, 0 },  /* mc */
1.358     schwarze  381:        { roff_noarg, NULL, NULL, 0 },  /* nf */
1.312     schwarze  382:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.309     schwarze  383:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.299     schwarze  384:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.300     schwarze  385:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.301     schwarze  386:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.296     schwarze  387:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.295     schwarze  388:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    389:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    390:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    391:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.311     schwarze  392:        { roff_als, NULL, NULL, 0 },  /* als */
1.295     schwarze  393:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    394:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    395:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    396:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    397:        { roff_ds, NULL, NULL, 0 },  /* as */
                    398:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    400:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    401:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    402:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    403:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    404:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    405:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    406:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    407:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
1.364     schwarze  408:        { roff_break, NULL, NULL, 0 },  /* break */
1.295     schwarze  409:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    410:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.358     schwarze  411:        { roff_noarg, NULL, NULL, 0 },  /* brp */
1.295     schwarze  412:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    413:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    414:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    415:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    416:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    417:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
1.341     schwarze  418:        { roff_char, NULL, NULL, 0 },  /* char */
1.295     schwarze  419:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    421:        { roff_insec, NULL, NULL, 0 },  /* close */
                    422:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    424:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    425:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    429:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    430:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    431:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    432:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    433:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    434:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    436:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    437:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    438:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    439:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    440:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    441:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    442:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    443:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    444:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    445:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.303     schwarze  446:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.295     schwarze  447:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    448:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    449:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    450:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    451:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.303     schwarze  452:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.295     schwarze  453:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    454:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    455:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    456:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    457:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    458:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    461:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    462:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    463:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    464:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    465:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    466:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    467:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    468:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    469:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    470:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    471:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    472:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    474:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    475:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    487:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    488:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    489:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    491:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    492:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    493:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    494:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    495:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    496:        { roff_it, NULL, NULL, 0 },  /* it */
                    497:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    500:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    501:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    503:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    504:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    505:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    506:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    508:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    511:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    512:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    513:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    514:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    516:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    517:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    518:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    519:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    520:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    521:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    522:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    523:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    524:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    526:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    527:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
1.330     schwarze  528:        { roff_nop, NULL, NULL, 0 },  /* nop */
1.295     schwarze  529:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    530:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    533:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    534:        { roff_insec, NULL, NULL, 0 },  /* open */
                    535:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    536:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    537:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    538:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    539:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    540:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    541:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    542:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    543:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    549:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    550:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    551:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    552:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    553:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    554:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    555:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
1.339     schwarze  557:        { roff_return, NULL, NULL, 0 },  /* return */
1.295     schwarze  558:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    559:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    560:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.306     schwarze  561:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.295     schwarze  562:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    563:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    565:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    566:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    568:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
1.339     schwarze  569:        { roff_shift, NULL, NULL, 0 },  /* shift */
1.295     schwarze  570:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    571:        { roff_so, NULL, NULL, 0 },  /* so */
                    572:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    573:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    574:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    575:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    576:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    577:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    578:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    579:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    580:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    581:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    582:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.315     schwarze  583:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.295     schwarze  584:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    585:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    586:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    587:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    588:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    589:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    590:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    591:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    592:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    593:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    594:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    595:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    596:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    597:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    598:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    599:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    600:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    601:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    602:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    603:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    604:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    605:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    606:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    607:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    608:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    609:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    610:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
1.340     schwarze  611:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /*while*/
1.295     schwarze  612:        { roff_insec, NULL, NULL, 0 },  /* write */
                    613:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    614:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    615:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    616:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.306     schwarze  617:        { roff_renamed, NULL, NULL, 0 },
1.295     schwarze  618:        { roff_userdef, NULL, NULL, 0 }
1.67      kristaps  619: };
                    620:
1.141     kristaps  621: /* Array of injected predefined strings. */
                    622: #define        PREDEFS_MAX      38
                    623: static const struct predef predefs[PREDEFS_MAX] = {
                    624: #include "predefs.in"
                    625: };
                    626:
1.305     schwarze  627: static int      roffce_lines;  /* number of input lines to center */
                    628: static struct roff_node *roffce_node;  /* active request */
1.178     schwarze  629: static int      roffit_lines;  /* number of lines to delay */
                    630: static char    *roffit_macro;  /* nil-terminated macro line */
                    631:
1.207     schwarze  632:
1.266     schwarze  633: /* --- request table ------------------------------------------------------ */
                    634:
1.295     schwarze  635: struct ohash *
                    636: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85      kristaps  637: {
1.295     schwarze  638:        struct ohash    *htab;
                    639:        struct roffreq  *req;
                    640:        enum roff_tok    tok;
                    641:        size_t           sz;
                    642:        unsigned int     slot;
1.85      kristaps  643:
1.295     schwarze  644:        htab = mandoc_malloc(sizeof(*htab));
                    645:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85      kristaps  646:
1.295     schwarze  647:        for (tok = mintok; tok < maxtok; tok++) {
1.296     schwarze  648:                if (roff_name[tok] == NULL)
                    649:                        continue;
1.295     schwarze  650:                sz = strlen(roff_name[tok]);
                    651:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    652:                req->tok = tok;
                    653:                memcpy(req->name, roff_name[tok], sz + 1);
                    654:                slot = ohash_qlookup(htab, req->name);
                    655:                ohash_insert(htab, slot, req);
1.85      kristaps  656:        }
1.295     schwarze  657:        return htab;
1.85      kristaps  658: }
                    659:
1.295     schwarze  660: void
                    661: roffhash_free(struct ohash *htab)
1.67      kristaps  662: {
1.295     schwarze  663:        struct roffreq  *req;
                    664:        unsigned int     slot;
1.67      kristaps  665:
1.295     schwarze  666:        if (htab == NULL)
                    667:                return;
                    668:        for (req = ohash_first(htab, &slot); req != NULL;
                    669:             req = ohash_next(htab, &slot))
                    670:                free(req);
                    671:        ohash_delete(htab);
                    672:        free(htab);
                    673: }
                    674:
                    675: enum roff_tok
                    676: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    677: {
                    678:        struct roffreq  *req;
                    679:        const char      *end;
                    680:
                    681:        if (sz) {
                    682:                end = name + sz;
                    683:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    684:        } else
                    685:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    686:        return req == NULL ? TOKEN_NONE : req->tok;
1.67      kristaps  687: }
                    688:
1.266     schwarze  689: /* --- stack of request blocks -------------------------------------------- */
                    690:
1.67      kristaps  691: /*
                    692:  * Pop the current node off of the stack of roff instructions currently
1.364     schwarze  693:  * pending.  Return 1 if it is a loop or 0 otherwise.
1.67      kristaps  694:  */
1.340     schwarze  695: static int
1.67      kristaps  696: roffnode_pop(struct roff *r)
                    697: {
                    698:        struct roffnode *p;
1.340     schwarze  699:        int              inloop;
1.67      kristaps  700:
1.207     schwarze  701:        p = r->last;
1.340     schwarze  702:        inloop = p->tok == ROFF_while;
                    703:        r->last = p->parent;
1.106     kristaps  704:        free(p->name);
                    705:        free(p->end);
1.67      kristaps  706:        free(p);
1.340     schwarze  707:        return inloop;
1.67      kristaps  708: }
                    709:
                    710: /*
                    711:  * Push a roff node onto the instruction stack.  This must later be
                    712:  * removed with roffnode_pop().
                    713:  */
1.98      schwarze  714: static void
1.294     schwarze  715: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106     kristaps  716:                int line, int col)
1.67      kristaps  717: {
                    718:        struct roffnode *p;
                    719:
1.98      schwarze  720:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.67      kristaps  721:        p->tok = tok;
1.106     kristaps  722:        if (name)
                    723:                p->name = mandoc_strdup(name);
1.67      kristaps  724:        p->parent = r->last;
                    725:        p->line = line;
                    726:        p->col = col;
1.198     schwarze  727:        p->rule = p->parent ? p->parent->rule : 0;
1.67      kristaps  728:
                    729:        r->last = p;
                    730: }
                    731:
1.266     schwarze  732: /* --- roff parser state data management ---------------------------------- */
                    733:
1.67      kristaps  734: static void
                    735: roff_free1(struct roff *r)
                    736: {
1.167     kristaps  737:        int              i;
1.67      kristaps  738:
1.346     schwarze  739:        tbl_free(r->first_tbl);
1.113     kristaps  740:        r->first_tbl = r->last_tbl = r->tbl = NULL;
                    741:
1.347     schwarze  742:        eqn_free(r->last_eqn);
1.319     schwarze  743:        r->last_eqn = r->eqn = NULL;
1.125     kristaps  744:
1.339     schwarze  745:        while (r->mstackpos >= 0)
                    746:                roff_userret(r);
                    747:
1.67      kristaps  748:        while (r->last)
                    749:                roffnode_pop(r);
1.109     kristaps  750:
1.223     schwarze  751:        free (r->rstack);
                    752:        r->rstack = NULL;
                    753:        r->rstacksz = 0;
                    754:        r->rstackpos = -1;
                    755:
                    756:        roff_freereg(r->regtab);
                    757:        r->regtab = NULL;
                    758:
1.167     kristaps  759:        roff_freestr(r->strtab);
1.306     schwarze  760:        roff_freestr(r->rentab);
1.167     kristaps  761:        roff_freestr(r->xmbtab);
1.306     schwarze  762:        r->strtab = r->rentab = r->xmbtab = NULL;
1.167     kristaps  763:
                    764:        if (r->xtab)
                    765:                for (i = 0; i < 128; i++)
                    766:                        free(r->xtab[i].p);
                    767:        free(r->xtab);
                    768:        r->xtab = NULL;
1.67      kristaps  769: }
                    770:
                    771: void
                    772: roff_reset(struct roff *r)
                    773: {
                    774:        roff_free1(r);
1.367     schwarze  775:        r->options |= MPARSE_COMMENT;
1.227     schwarze  776:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303     schwarze  777:        r->control = '\0';
                    778:        r->escape = '\\';
1.307     schwarze  779:        roffce_lines = 0;
                    780:        roffce_node = NULL;
                    781:        roffit_lines = 0;
                    782:        roffit_macro = NULL;
1.67      kristaps  783: }
                    784:
                    785: void
                    786: roff_free(struct roff *r)
                    787: {
1.366     schwarze  788:        int              i;
1.339     schwarze  789:
1.67      kristaps  790:        roff_free1(r);
1.339     schwarze  791:        for (i = 0; i < r->mstacksz; i++)
                    792:                free(r->mstack[i].argv);
                    793:        free(r->mstack);
1.295     schwarze  794:        roffhash_free(r->reqtab);
1.67      kristaps  795:        free(r);
                    796: }
                    797:
                    798: struct roff *
1.351     schwarze  799: roff_alloc(int options)
1.67      kristaps  800: {
                    801:        struct roff     *r;
                    802:
1.98      schwarze  803:        r = mandoc_calloc(1, sizeof(struct roff));
1.328     schwarze  804:        r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.367     schwarze  805:        r->options = options | MPARSE_COMMENT;
1.227     schwarze  806:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.339     schwarze  807:        r->mstackpos = -1;
1.82      kristaps  808:        r->rstackpos = -1;
1.303     schwarze  809:        r->escape = '\\';
1.277     schwarze  810:        return r;
1.265     schwarze  811: }
                    812:
1.266     schwarze  813: /* --- syntax tree state data management ---------------------------------- */
                    814:
1.265     schwarze  815: static void
                    816: roff_man_free1(struct roff_man *man)
                    817: {
1.356     schwarze  818:        if (man->meta.first != NULL)
                    819:                roff_node_delete(man, man->meta.first);
1.265     schwarze  820:        free(man->meta.msec);
                    821:        free(man->meta.vol);
                    822:        free(man->meta.os);
                    823:        free(man->meta.arch);
                    824:        free(man->meta.title);
                    825:        free(man->meta.name);
                    826:        free(man->meta.date);
1.356     schwarze  827:        free(man->meta.sodest);
1.265     schwarze  828: }
                    829:
1.357     schwarze  830: void
                    831: roff_state_reset(struct roff_man *man)
                    832: {
                    833:        man->last = man->meta.first;
                    834:        man->last_es = NULL;
                    835:        man->flags = 0;
                    836:        man->lastsec = man->lastnamed = SEC_NONE;
                    837:        man->next = ROFF_NEXT_CHILD;
                    838:        roff_setreg(man->roff, "nS", 0, '=');
                    839: }
                    840:
1.265     schwarze  841: static void
                    842: roff_man_alloc1(struct roff_man *man)
                    843: {
                    844:        memset(&man->meta, 0, sizeof(man->meta));
1.356     schwarze  845:        man->meta.first = mandoc_calloc(1, sizeof(*man->meta.first));
                    846:        man->meta.first->type = ROFFT_ROOT;
                    847:        man->meta.macroset = MACROSET_NONE;
1.357     schwarze  848:        roff_state_reset(man);
1.265     schwarze  849: }
                    850:
                    851: void
                    852: roff_man_reset(struct roff_man *man)
                    853: {
                    854:        roff_man_free1(man);
                    855:        roff_man_alloc1(man);
                    856: }
                    857:
                    858: void
                    859: roff_man_free(struct roff_man *man)
                    860: {
                    861:        roff_man_free1(man);
1.380     schwarze  862:        free(man->os_r);
1.265     schwarze  863:        free(man);
                    864: }
                    865:
                    866: struct roff_man *
1.351     schwarze  867: roff_man_alloc(struct roff *roff, const char *os_s, int quick)
1.265     schwarze  868: {
                    869:        struct roff_man *man;
                    870:
                    871:        man = mandoc_calloc(1, sizeof(*man));
                    872:        man->roff = roff;
1.316     schwarze  873:        man->os_s = os_s;
1.265     schwarze  874:        man->quick = quick;
                    875:        roff_man_alloc1(man);
1.296     schwarze  876:        roff->man = man;
1.277     schwarze  877:        return man;
1.67      kristaps  878: }
                    879:
1.266     schwarze  880: /* --- syntax tree handling ----------------------------------------------- */
                    881:
                    882: struct roff_node *
                    883: roff_node_alloc(struct roff_man *man, int line, int pos,
                    884:        enum roff_type type, int tok)
                    885: {
                    886:        struct roff_node        *n;
                    887:
                    888:        n = mandoc_calloc(1, sizeof(*n));
                    889:        n->line = line;
                    890:        n->pos = pos;
                    891:        n->tok = tok;
                    892:        n->type = type;
                    893:        n->sec = man->lastsec;
                    894:
                    895:        if (man->flags & MDOC_SYNOPSIS)
1.285     schwarze  896:                n->flags |= NODE_SYNPRETTY;
1.266     schwarze  897:        else
1.285     schwarze  898:                n->flags &= ~NODE_SYNPRETTY;
1.360     schwarze  899:        if ((man->flags & (ROFF_NOFILL | ROFF_NONOFILL)) == ROFF_NOFILL)
1.359     schwarze  900:                n->flags |= NODE_NOFILL;
                    901:        else
                    902:                n->flags &= ~NODE_NOFILL;
1.266     schwarze  903:        if (man->flags & MDOC_NEWLINE)
1.285     schwarze  904:                n->flags |= NODE_LINE;
1.266     schwarze  905:        man->flags &= ~MDOC_NEWLINE;
                    906:
1.277     schwarze  907:        return n;
1.266     schwarze  908: }
                    909:
                    910: void
                    911: roff_node_append(struct roff_man *man, struct roff_node *n)
                    912: {
                    913:
                    914:        switch (man->next) {
                    915:        case ROFF_NEXT_SIBLING:
1.281     schwarze  916:                if (man->last->next != NULL) {
                    917:                        n->next = man->last->next;
                    918:                        man->last->next->prev = n;
                    919:                } else
                    920:                        man->last->parent->last = n;
1.266     schwarze  921:                man->last->next = n;
                    922:                n->prev = man->last;
                    923:                n->parent = man->last->parent;
                    924:                break;
                    925:        case ROFF_NEXT_CHILD:
1.287     schwarze  926:                if (man->last->child != NULL) {
                    927:                        n->next = man->last->child;
                    928:                        man->last->child->prev = n;
                    929:                } else
                    930:                        man->last->last = n;
1.266     schwarze  931:                man->last->child = n;
                    932:                n->parent = man->last;
                    933:                break;
                    934:        default:
                    935:                abort();
                    936:        }
1.282     schwarze  937:        man->last = n;
1.266     schwarze  938:
                    939:        switch (n->type) {
                    940:        case ROFFT_HEAD:
                    941:                n->parent->head = n;
                    942:                break;
                    943:        case ROFFT_BODY:
1.282     schwarze  944:                if (n->end != ENDBODY_NOT)
                    945:                        return;
1.266     schwarze  946:                n->parent->body = n;
                    947:                break;
                    948:        case ROFFT_TAIL:
                    949:                n->parent->tail = n;
                    950:                break;
                    951:        default:
1.282     schwarze  952:                return;
1.266     schwarze  953:        }
1.282     schwarze  954:
                    955:        /*
                    956:         * Copy over the normalised-data pointer of our parent.  Not
                    957:         * everybody has one, but copying a null pointer is fine.
                    958:         */
                    959:
                    960:        n->norm = n->parent->norm;
                    961:        assert(n->parent->type == ROFFT_BLOCK);
1.266     schwarze  962: }
                    963:
1.267     schwarze  964: void
                    965: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    966: {
                    967:        struct roff_node        *n;
                    968:
                    969:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    970:        n->string = roff_strdup(man->roff, word);
                    971:        roff_node_append(man, n);
1.286     schwarze  972:        n->flags |= NODE_VALID | NODE_ENDED;
1.267     schwarze  973:        man->next = ROFF_NEXT_SIBLING;
                    974: }
                    975:
                    976: void
                    977: roff_word_append(struct roff_man *man, const char *word)
                    978: {
                    979:        struct roff_node        *n;
                    980:        char                    *addstr, *newstr;
                    981:
                    982:        n = man->last;
                    983:        addstr = roff_strdup(man->roff, word);
                    984:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    985:        free(addstr);
                    986:        free(n->string);
                    987:        n->string = newstr;
                    988:        man->next = ROFF_NEXT_SIBLING;
1.268     schwarze  989: }
                    990:
                    991: void
                    992: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    993: {
                    994:        struct roff_node        *n;
                    995:
                    996:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    997:        roff_node_append(man, n);
                    998:        man->next = ROFF_NEXT_CHILD;
                    999: }
                   1000:
                   1001: struct roff_node *
                   1002: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                   1003: {
                   1004:        struct roff_node        *n;
                   1005:
                   1006:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                   1007:        roff_node_append(man, n);
                   1008:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1009:        return n;
1.267     schwarze 1010: }
                   1011:
1.266     schwarze 1012: struct roff_node *
                   1013: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                   1014: {
                   1015:        struct roff_node        *n;
                   1016:
                   1017:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                   1018:        roff_node_append(man, n);
                   1019:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1020:        return n;
1.266     schwarze 1021: }
                   1022:
                   1023: struct roff_node *
                   1024: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                   1025: {
                   1026:        struct roff_node        *n;
                   1027:
                   1028:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                   1029:        roff_node_append(man, n);
                   1030:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1031:        return n;
1.267     schwarze 1032: }
                   1033:
1.321     schwarze 1034: static void
1.346     schwarze 1035: roff_addtbl(struct roff_man *man, int line, struct tbl_node *tbl)
1.267     schwarze 1036: {
                   1037:        struct roff_node        *n;
1.346     schwarze 1038:        struct tbl_span         *span;
1.267     schwarze 1039:
1.356     schwarze 1040:        if (man->meta.macroset == MACROSET_MAN)
1.302     schwarze 1041:                man_breakscope(man, ROFF_TS);
1.321     schwarze 1042:        while ((span = tbl_span(tbl)) != NULL) {
1.346     schwarze 1043:                n = roff_node_alloc(man, line, 0, ROFFT_TBL, TOKEN_NONE);
1.321     schwarze 1044:                n->span = span;
                   1045:                roff_node_append(man, n);
                   1046:                n->flags |= NODE_VALID | NODE_ENDED;
                   1047:                man->next = ROFF_NEXT_SIBLING;
                   1048:        }
1.266     schwarze 1049: }
                   1050:
                   1051: void
                   1052: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1053: {
                   1054:
                   1055:        /* Adjust siblings. */
                   1056:
                   1057:        if (n->prev)
                   1058:                n->prev->next = n->next;
                   1059:        if (n->next)
                   1060:                n->next->prev = n->prev;
                   1061:
                   1062:        /* Adjust parent. */
                   1063:
                   1064:        if (n->parent != NULL) {
                   1065:                if (n->parent->child == n)
                   1066:                        n->parent->child = n->next;
                   1067:                if (n->parent->last == n)
                   1068:                        n->parent->last = n->prev;
                   1069:        }
                   1070:
                   1071:        /* Adjust parse point. */
                   1072:
                   1073:        if (man == NULL)
                   1074:                return;
                   1075:        if (man->last == n) {
                   1076:                if (n->prev == NULL) {
                   1077:                        man->last = n->parent;
                   1078:                        man->next = ROFF_NEXT_CHILD;
                   1079:                } else {
                   1080:                        man->last = n->prev;
                   1081:                        man->next = ROFF_NEXT_SIBLING;
                   1082:                }
                   1083:        }
1.356     schwarze 1084:        if (man->meta.first == n)
                   1085:                man->meta.first = NULL;
1.344     schwarze 1086: }
                   1087:
                   1088: void
                   1089: roff_node_relink(struct roff_man *man, struct roff_node *n)
                   1090: {
                   1091:        roff_node_unlink(man, n);
                   1092:        n->prev = n->next = NULL;
                   1093:        roff_node_append(man, n);
1.266     schwarze 1094: }
                   1095:
                   1096: void
                   1097: roff_node_free(struct roff_node *n)
                   1098: {
                   1099:
                   1100:        if (n->args != NULL)
                   1101:                mdoc_argv_free(n->args);
                   1102:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1103:                free(n->norm);
1.347     schwarze 1104:        eqn_box_free(n->eqn);
1.266     schwarze 1105:        free(n->string);
1.374     schwarze 1106:        free(n->tag);
1.266     schwarze 1107:        free(n);
                   1108: }
                   1109:
                   1110: void
                   1111: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1112: {
                   1113:
                   1114:        while (n->child != NULL)
                   1115:                roff_node_delete(man, n->child);
                   1116:        roff_node_unlink(man, n);
                   1117:        roff_node_free(n);
1.370     schwarze 1118: }
                   1119:
                   1120: int
                   1121: roff_node_transparent(struct roff_node *n)
                   1122: {
                   1123:        if (n == NULL)
                   1124:                return 0;
                   1125:        if (n->type == ROFFT_COMMENT || n->flags & NODE_NOPRT)
                   1126:                return 1;
1.373     schwarze 1127:        return roff_tok_transparent(n->tok);
                   1128: }
                   1129:
                   1130: int
                   1131: roff_tok_transparent(enum roff_tok tok)
                   1132: {
                   1133:        switch (tok) {
1.370     schwarze 1134:        case ROFF_ft:
                   1135:        case ROFF_ll:
                   1136:        case ROFF_mc:
                   1137:        case ROFF_po:
                   1138:        case ROFF_ta:
                   1139:        case MDOC_Db:
                   1140:        case MDOC_Es:
                   1141:        case MDOC_Sm:
                   1142:        case MDOC_Tg:
                   1143:        case MAN_DT:
                   1144:        case MAN_UC:
                   1145:        case MAN_PD:
                   1146:        case MAN_AT:
                   1147:                return 1;
                   1148:        default:
                   1149:                return 0;
                   1150:        }
                   1151: }
                   1152:
                   1153: struct roff_node *
                   1154: roff_node_child(struct roff_node *n)
                   1155: {
                   1156:        for (n = n->child; roff_node_transparent(n); n = n->next)
                   1157:                continue;
                   1158:        return n;
                   1159: }
                   1160:
                   1161: struct roff_node *
                   1162: roff_node_prev(struct roff_node *n)
                   1163: {
                   1164:        do {
                   1165:                n = n->prev;
                   1166:        } while (roff_node_transparent(n));
                   1167:        return n;
                   1168: }
                   1169:
                   1170: struct roff_node *
                   1171: roff_node_next(struct roff_node *n)
                   1172: {
                   1173:        do {
                   1174:                n = n->next;
                   1175:        } while (roff_node_transparent(n));
                   1176:        return n;
1.269     schwarze 1177: }
                   1178:
                   1179: void
                   1180: deroff(char **dest, const struct roff_node *n)
                   1181: {
                   1182:        char    *cp;
                   1183:        size_t   sz;
                   1184:
1.371     schwarze 1185:        if (n->string == NULL) {
1.269     schwarze 1186:                for (n = n->child; n != NULL; n = n->next)
                   1187:                        deroff(dest, n);
                   1188:                return;
                   1189:        }
                   1190:
1.288     schwarze 1191:        /* Skip leading whitespace. */
1.269     schwarze 1192:
1.288     schwarze 1193:        for (cp = n->string; *cp != '\0'; cp++) {
1.289     schwarze 1194:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1195:                    strchr(" %&0^|~", cp[1]) != NULL)
1.269     schwarze 1196:                        cp++;
1.288     schwarze 1197:                else if ( ! isspace((unsigned char)*cp))
1.269     schwarze 1198:                        break;
                   1199:        }
                   1200:
1.289     schwarze 1201:        /* Skip trailing backslash. */
                   1202:
                   1203:        sz = strlen(cp);
1.290     schwarze 1204:        if (sz > 0 && cp[sz - 1] == '\\')
1.289     schwarze 1205:                sz--;
                   1206:
1.269     schwarze 1207:        /* Skip trailing whitespace. */
                   1208:
1.289     schwarze 1209:        for (; sz; sz--)
1.269     schwarze 1210:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1211:                        break;
                   1212:
                   1213:        /* Skip empty strings. */
                   1214:
                   1215:        if (sz == 0)
                   1216:                return;
                   1217:
                   1218:        if (*dest == NULL) {
                   1219:                *dest = mandoc_strndup(cp, sz);
                   1220:                return;
                   1221:        }
                   1222:
                   1223:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1224:        free(*dest);
                   1225:        *dest = cp;
1.266     schwarze 1226: }
                   1227:
                   1228: /* --- main functions of the roff parser ---------------------------------- */
                   1229:
1.388     schwarze 1230: /*
                   1231:  * Save comments preceding the title macro, for example in order to
                   1232:  * preserve Copyright and license headers in HTML output,
                   1233:  * provide diagnostics about RCS ids and trailing whitespace in comments,
                   1234:  * then discard comments including preceding whitespace.
                   1235:  * This function also handles input line continuation.
                   1236:  */
1.387     schwarze 1237: static int
1.388     schwarze 1238: roff_parse_comment(struct roff *r, struct buf *buf, int ln, int pos, char ec)
1.387     schwarze 1239: {
                   1240:        struct roff_node *n;    /* used for header comments */
                   1241:        const char      *start; /* start of the string to process */
                   1242:        const char      *cp;    /* for RCS id parsing */
                   1243:        char            *stesc; /* start of an escape sequence ('\\') */
                   1244:        char            *ep;    /* end of comment string */
                   1245:        int              rcsid; /* kind of RCS id seen */
                   1246:
                   1247:        for (start = stesc = buf->buf + pos;; stesc++) {
1.388     schwarze 1248:                /*
                   1249:                 * XXX Ugly hack: Remove the newline character that
                   1250:                 * mparse_buf_r() appended to mark the end of input
                   1251:                 * if it is not preceded by an escape character.
                   1252:                 */
                   1253:                if (stesc[0] == '\n') {
                   1254:                        assert(stesc[1] == '\0');
                   1255:                        stesc[0] = '\0';
                   1256:                }
                   1257:
1.387     schwarze 1258:                /* The line ends without continuation or comment. */
                   1259:                if (stesc[0] == '\0')
                   1260:                        return ROFF_CONT;
                   1261:
                   1262:                /* Unescaped byte: skip it. */
1.388     schwarze 1263:                if (stesc[0] != ec)
1.387     schwarze 1264:                        continue;
                   1265:
1.388     schwarze 1266:                /*
                   1267:                 * XXX Ugly hack: Do not attempt to append another line
                   1268:                 * if the function mparse_buf_r() appended a newline
                   1269:                 * character to indicate the end of input.
                   1270:                 */
                   1271:                if (stesc[1] == '\n') {
                   1272:                        assert(stesc[2] == '\0');
                   1273:                        stesc[0] = '\0';
                   1274:                        return ROFF_CONT;
                   1275:                }
                   1276:
                   1277:                /*
                   1278:                 * An escape character at the end of an input line
                   1279:                 * requests line continuation.
                   1280:                 */
1.387     schwarze 1281:                if (stesc[1] == '\0') {
                   1282:                        stesc[0] = '\0';
                   1283:                        return ROFF_IGN | ROFF_APPEND;
                   1284:                }
                   1285:
                   1286:                /* Found a comment: process it. */
                   1287:                if (stesc[1] == '"' || stesc[1] == '#')
                   1288:                        break;
                   1289:
                   1290:                /* Escaped escape character: skip them both. */
1.388     schwarze 1291:                if (stesc[1] == ec)
1.387     schwarze 1292:                        stesc++;
                   1293:        }
                   1294:
                   1295:        /* Look for an RCS id in the comment. */
                   1296:
                   1297:        rcsid = 0;
                   1298:        if ((cp = strstr(stesc + 2, "$" "OpenBSD")) != NULL) {
                   1299:                rcsid = 1 << MANDOC_OS_OPENBSD;
                   1300:                cp += 8;
                   1301:        } else if ((cp = strstr(stesc + 2, "$" "NetBSD")) != NULL) {
                   1302:                rcsid = 1 << MANDOC_OS_NETBSD;
                   1303:                cp += 7;
                   1304:        }
                   1305:        if (cp != NULL && isalnum((unsigned char)*cp) == 0 &&
                   1306:            strchr(cp, '$') != NULL) {
                   1307:                if (r->man->meta.rcsids & rcsid)
                   1308:                        mandoc_msg(MANDOCERR_RCS_REP, ln,
                   1309:                            (int)(stesc - buf->buf) + 2, "%s", stesc + 1);
                   1310:                r->man->meta.rcsids |= rcsid;
                   1311:        }
                   1312:
                   1313:        /* Warn about trailing whitespace at the end of the comment. */
                   1314:
                   1315:        ep = strchr(stesc + 2, '\0') - 1;
                   1316:        if (*ep == '\n')
                   1317:                *ep-- = '\0';
                   1318:        if (*ep == ' ' || *ep == '\t')
                   1319:                mandoc_msg(MANDOCERR_SPACE_EOL,
                   1320:                    ln, (int)(ep - buf->buf), NULL);
                   1321:
                   1322:        /* Save comments preceding the title macro in the syntax tree. */
                   1323:
                   1324:        if (r->options & MPARSE_COMMENT) {
                   1325:                while (*ep == ' ' || *ep == '\t')
                   1326:                        ep--;
                   1327:                ep[1] = '\0';
                   1328:                n = roff_node_alloc(r->man, ln, stesc + 1 - buf->buf,
                   1329:                    ROFFT_COMMENT, TOKEN_NONE);
                   1330:                n->string = mandoc_strdup(stesc + 2);
                   1331:                roff_node_append(r->man, n);
                   1332:                n->flags |= NODE_VALID | NODE_ENDED;
                   1333:                r->man->next = ROFF_NEXT_SIBLING;
                   1334:        }
                   1335:
                   1336:        /* The comment requests line continuation. */
                   1337:
                   1338:        if (stesc[1] == '#') {
                   1339:                *stesc = '\0';
                   1340:                return ROFF_IGN | ROFF_APPEND;
                   1341:        }
                   1342:
                   1343:        /* Discard the comment including preceding whitespace. */
                   1344:
                   1345:        while (stesc > start && stesc[-1] == ' ' &&
                   1346:            (stesc == start + 1 || stesc[-2] != '\\'))
                   1347:                stesc--;
                   1348:        *stesc = '\0';
                   1349:        return ROFF_CONT;
                   1350: }
                   1351:
1.94      kristaps 1352: /*
1.355     schwarze 1353:  * In the current line, expand escape sequences that produce parsable
                   1354:  * input text.  Also check the syntax of the remaining escape sequences,
                   1355:  * which typically produce output glyphs or change formatter state.
1.154     kristaps 1356:  */
1.340     schwarze 1357: static int
1.388     schwarze 1358: roff_expand(struct roff *r, struct buf *buf, int ln, int pos, char ec)
1.94      kristaps 1359: {
1.388     schwarze 1360:        char             ubuf[24];      /* buffer to print a number */
                   1361:        struct mctx     *ctx;           /* current macro call context */
                   1362:        const char      *res;           /* the string to be pasted */
                   1363:        const char      *src;           /* source for copying */
                   1364:        char            *dst;           /* destination for copying */
                   1365:        int              iesc;          /* index of leading escape char */
                   1366:        int              inam;          /* index of the escape name */
                   1367:        int              iarg;          /* index beginning the argument */
                   1368:        int              iendarg;       /* index right after the argument */
                   1369:        int              iend;          /* index right after the sequence */
1.393     schwarze 1370:        int              isrc, idst;    /* to reduce \\ and \. in names */
1.388     schwarze 1371:        int              deftype;       /* type of definition to paste */
                   1372:        int              argi;          /* macro argument index */
                   1373:        int              quote_args;    /* true for \\$@, false for \\$* */
                   1374:        int              asz;           /* length of the replacement */
                   1375:        int              rsz;           /* length of the rest of the string */
                   1376:        int              npos;          /* position in numeric expression */
1.181     schwarze 1377:        int              expand_count;  /* to avoid infinite loops */
1.303     schwarze 1378:
1.170     schwarze 1379:        expand_count = 0;
1.388     schwarze 1380:        while (buf->buf[pos] != '\0') {
1.170     schwarze 1381:
1.388     schwarze 1382:                /*
                   1383:                 * Skip plain ASCII characters.
                   1384:                 * If we have a non-standard escape character,
                   1385:                 * escape literal backslashes because all processing in
                   1386:                 * subsequent functions uses the standard escaping rules.
                   1387:                 */
1.205     schwarze 1388:
1.388     schwarze 1389:                if (buf->buf[pos] != ec) {
1.397     schwarze 1390:                        if (buf->buf[pos] == '\\') {
1.388     schwarze 1391:                                roff_expand_patch(buf, pos, "\\e", pos + 1);
                   1392:                                pos++;
1.303     schwarze 1393:                        }
1.388     schwarze 1394:                        pos++;
1.205     schwarze 1395:                        continue;
1.303     schwarze 1396:                }
1.205     schwarze 1397:
1.388     schwarze 1398:                /*
                   1399:                 * Parse escape sequences,
                   1400:                 * issue diagnostic messages when appropriate,
                   1401:                 * and skip sequences that do not need expansion.
                   1402:                 * If we have a non-standard escape character, translate
                   1403:                 * it to backslashes and translate backslashes to \e.
                   1404:                 */
1.205     schwarze 1405:
1.392     schwarze 1406:                if (roff_escape(buf->buf, ln, pos, &iesc, &inam,
                   1407:                    &iarg, &iendarg, &iend) != ESCAPE_EXPAND) {
1.388     schwarze 1408:                        while (pos < iend) {
                   1409:                                if (buf->buf[pos] == ec) {
                   1410:                                        buf->buf[pos] = '\\';
                   1411:                                        if (pos + 1 < iend)
                   1412:                                                pos++;
                   1413:                                } else if (buf->buf[pos] == '\\') {
                   1414:                                        roff_expand_patch(buf,
                   1415:                                            pos, "\\e", pos + 1);
                   1416:                                        pos++;
                   1417:                                        iend++;
                   1418:                                }
                   1419:                                pos++;
1.352     schwarze 1420:                        }
1.205     schwarze 1421:                        continue;
1.152     kristaps 1422:                }
1.388     schwarze 1423:
1.393     schwarze 1424:                /* Reduce \\ and \. in names. */
                   1425:
                   1426:                if (buf->buf[inam] == '*' || buf->buf[inam] == 'n') {
                   1427:                        isrc = idst = iarg;
                   1428:                        while (isrc < iendarg) {
                   1429:                                if (isrc + 1 < iendarg &&
                   1430:                                    buf->buf[isrc] == '\\' &&
                   1431:                                    (buf->buf[isrc + 1] == '\\' ||
                   1432:                                     buf->buf[isrc + 1] == '.'))
                   1433:                                        isrc++;
                   1434:                                buf->buf[idst++] = buf->buf[isrc++];
                   1435:                        }
                   1436:                        iendarg -= isrc - idst;
                   1437:                }
                   1438:
1.388     schwarze 1439:                /* Handle expansion. */
                   1440:
                   1441:                res = NULL;
                   1442:                switch (buf->buf[inam]) {
                   1443:                case '*':
                   1444:                        if (iendarg == iarg)
1.206     schwarze 1445:                                break;
1.388     schwarze 1446:                        deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1447:                        if ((res = roff_getstrn(r, buf->buf + iarg,
                   1448:                            iendarg - iarg, &deftype)) != NULL)
1.206     schwarze 1449:                                break;
1.94      kristaps 1450:
1.388     schwarze 1451:                        /*
1.396     schwarze 1452:                         * If not overridden,
1.388     schwarze 1453:                         * let \*(.T through to the formatters.
                   1454:                         */
1.94      kristaps 1455:
1.388     schwarze 1456:                        if (iendarg - iarg == 2 &&
                   1457:                            buf->buf[iarg] == '.' &&
                   1458:                            buf->buf[iarg + 1] == 'T') {
                   1459:                                roff_setstrn(&r->strtab, ".T", 2, NULL, 0, 0);
                   1460:                                pos = iend;
1.253     schwarze 1461:                                continue;
                   1462:                        }
1.94      kristaps 1463:
1.388     schwarze 1464:                        mandoc_msg(MANDOCERR_STR_UNDEF, ln, iesc,
                   1465:                            "%.*s", iendarg - iarg, buf->buf + iarg);
                   1466:                        break;
1.108     schwarze 1467:
1.339     schwarze 1468:                case '$':
                   1469:                        if (r->mstackpos < 0) {
1.388     schwarze 1470:                                mandoc_msg(MANDOCERR_ARG_UNDEF, ln, iesc,
                   1471:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1472:                                break;
                   1473:                        }
                   1474:                        ctx = r->mstack + r->mstackpos;
1.388     schwarze 1475:                        argi = buf->buf[iarg] - '1';
                   1476:                        if (argi >= 0 && argi <= 8) {
                   1477:                                if (argi < ctx->argc)
                   1478:                                        res = ctx->argv[argi];
1.339     schwarze 1479:                                break;
                   1480:                        }
1.388     schwarze 1481:                        if (buf->buf[iarg] == '*')
1.339     schwarze 1482:                                quote_args = 0;
1.388     schwarze 1483:                        else if (buf->buf[iarg] == '@')
1.339     schwarze 1484:                                quote_args = 1;
                   1485:                        else {
1.388     schwarze 1486:                                mandoc_msg(MANDOCERR_ARG_NONUM, ln, iesc,
                   1487:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1488:                                break;
                   1489:                        }
                   1490:                        asz = 0;
1.388     schwarze 1491:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1492:                                if (argi)
1.339     schwarze 1493:                                        asz++;  /* blank */
                   1494:                                if (quote_args)
                   1495:                                        asz += 2;  /* quotes */
1.388     schwarze 1496:                                asz += strlen(ctx->argv[argi]);
1.339     schwarze 1497:                        }
1.388     schwarze 1498:                        if (asz != iend - iesc) {
                   1499:                                rsz = buf->sz - iend;
                   1500:                                if (asz < iend - iesc)
                   1501:                                        memmove(buf->buf + iesc + asz,
                   1502:                                            buf->buf + iend, rsz);
                   1503:                                buf->sz = iesc + asz + rsz;
                   1504:                                buf->buf = mandoc_realloc(buf->buf, buf->sz);
                   1505:                                if (asz > iend - iesc)
                   1506:                                        memmove(buf->buf + iesc + asz,
                   1507:                                            buf->buf + iend, rsz);
1.339     schwarze 1508:                        }
1.388     schwarze 1509:                        dst = buf->buf + iesc;
                   1510:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1511:                                if (argi)
                   1512:                                        *dst++ = ' ';
1.339     schwarze 1513:                                if (quote_args)
1.388     schwarze 1514:                                        *dst++ = '"';
                   1515:                                src = ctx->argv[argi];
                   1516:                                while (*src != '\0')
                   1517:                                        *dst++ = *src++;
1.339     schwarze 1518:                                if (quote_args)
1.388     schwarze 1519:                                        *dst++ = '"';
1.339     schwarze 1520:                        }
                   1521:                        continue;
1.391     schwarze 1522:                case 'A':
                   1523:                        ubuf[0] = iendarg > iarg ? '1' : '0';
                   1524:                        ubuf[1] = '\0';
                   1525:                        res = ubuf;
                   1526:                        break;
1.207     schwarze 1527:                case 'B':
1.206     schwarze 1528:                        npos = 0;
1.388     schwarze 1529:                        ubuf[0] = iendarg > iarg && iend > iendarg &&
                   1530:                            roff_evalnum(r, ln, buf->buf + iarg, &npos,
                   1531:                                         NULL, ROFFNUM_SCALE) &&
                   1532:                            npos == iendarg - iarg ? '1' : '0';
1.206     schwarze 1533:                        ubuf[1] = '\0';
1.388     schwarze 1534:                        res = ubuf;
1.206     schwarze 1535:                        break;
1.389     schwarze 1536:                case 'V':
                   1537:                        mandoc_msg(MANDOCERR_UNSUPP, ln, iesc,
                   1538:                            "%.*s", iend - iesc, buf->buf + iesc);
                   1539:                        roff_expand_patch(buf, iendarg, "}", iend);
                   1540:                        roff_expand_patch(buf, iesc, "${", iarg);
                   1541:                        continue;
1.390     schwarze 1542:                case 'g':
                   1543:                        break;
1.207     schwarze 1544:                case 'n':
1.388     schwarze 1545:                        if (iendarg > iarg)
1.218     schwarze 1546:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.388     schwarze 1547:                                    roff_getregn(r, buf->buf + iarg,
                   1548:                                    iendarg - iarg, buf->buf[inam + 1]));
1.218     schwarze 1549:                        else
                   1550:                                ubuf[0] = '\0';
1.388     schwarze 1551:                        res = ubuf;
1.206     schwarze 1552:                        break;
1.207     schwarze 1553:                case 'w':
1.388     schwarze 1554:                        (void)snprintf(ubuf, sizeof(ubuf),
                   1555:                            "%d", (iendarg - iarg) * 24);
                   1556:                        res = ubuf;
                   1557:                        break;
                   1558:                default:
1.206     schwarze 1559:                        break;
                   1560:                }
1.388     schwarze 1561:                if (res == NULL)
1.142     kristaps 1562:                        res = "";
1.388     schwarze 1563:                if (++expand_count > EXPAND_LIMIT ||
                   1564:                    buf->sz + strlen(res) > SHRT_MAX) {
                   1565:                        mandoc_msg(MANDOCERR_ROFFLOOP, ln, iesc, NULL);
1.277     schwarze 1566:                        return ROFF_IGN;
1.94      kristaps 1567:                }
1.388     schwarze 1568:                roff_expand_patch(buf, iesc, res, iend);
                   1569:        }
                   1570:        return ROFF_CONT;
                   1571: }
1.94      kristaps 1572:
1.388     schwarze 1573: /*
                   1574:  * Replace the substring from the start position (inclusive)
                   1575:  * to end position (exclusive) with the repl(acement) string.
                   1576:  */
                   1577: static void
                   1578: roff_expand_patch(struct buf *buf, int start, const char *repl, int end)
                   1579: {
                   1580:        char    *nbuf;
1.108     schwarze 1581:
1.389     schwarze 1582:        buf->sz = mandoc_asprintf(&nbuf, "%.*s%s%s", start, buf->buf,
                   1583:            repl, buf->buf + end) + 1;
1.388     schwarze 1584:        free(buf->buf);
                   1585:        buf->buf = nbuf;
1.154     kristaps 1586: }
1.353     schwarze 1587:
                   1588: /*
                   1589:  * Parse a quoted or unquoted roff-style request or macro argument.
                   1590:  * Return a pointer to the parsed argument, which is either the original
                   1591:  * pointer or advanced by one byte in case the argument is quoted.
                   1592:  * NUL-terminate the argument in place.
                   1593:  * Collapse pairs of quotes inside quoted arguments.
                   1594:  * Advance the argument pointer to the next argument,
                   1595:  * or to the NUL byte terminating the argument line.
                   1596:  */
                   1597: char *
1.355     schwarze 1598: roff_getarg(struct roff *r, char **cpp, int ln, int *pos)
1.353     schwarze 1599: {
1.355     schwarze 1600:        struct buf       buf;
1.366     schwarze 1601:        char            *cp, *start;
1.355     schwarze 1602:        int              newesc, pairs, quoted, white;
1.353     schwarze 1603:
                   1604:        /* Quoting can only start with a new word. */
                   1605:        start = *cpp;
                   1606:        quoted = 0;
                   1607:        if ('"' == *start) {
                   1608:                quoted = 1;
                   1609:                start++;
                   1610:        }
                   1611:
1.355     schwarze 1612:        newesc = pairs = white = 0;
1.353     schwarze 1613:        for (cp = start; '\0' != *cp; cp++) {
                   1614:
                   1615:                /*
                   1616:                 * Move the following text left
                   1617:                 * after quoted quotes and after "\\" and "\t".
                   1618:                 */
                   1619:                if (pairs)
                   1620:                        cp[-pairs] = cp[0];
                   1621:
                   1622:                if ('\\' == cp[0]) {
                   1623:                        /*
                   1624:                         * In copy mode, translate double to single
                   1625:                         * backslashes and backslash-t to literal tabs.
                   1626:                         */
                   1627:                        switch (cp[1]) {
                   1628:                        case 'a':
                   1629:                        case 't':
1.354     schwarze 1630:                                cp[-pairs] = '\t';
1.355     schwarze 1631:                                pairs++;
                   1632:                                cp++;
                   1633:                                break;
1.353     schwarze 1634:                        case '\\':
1.397     schwarze 1635:                                cp[-pairs] = '\\';
1.355     schwarze 1636:                                newesc = 1;
1.353     schwarze 1637:                                pairs++;
                   1638:                                cp++;
                   1639:                                break;
                   1640:                        case ' ':
                   1641:                                /* Skip escaped blanks. */
                   1642:                                if (0 == quoted)
                   1643:                                        cp++;
                   1644:                                break;
                   1645:                        default:
                   1646:                                break;
                   1647:                        }
                   1648:                } else if (0 == quoted) {
                   1649:                        if (' ' == cp[0]) {
                   1650:                                /* Unescaped blanks end unquoted args. */
                   1651:                                white = 1;
                   1652:                                break;
                   1653:                        }
                   1654:                } else if ('"' == cp[0]) {
                   1655:                        if ('"' == cp[1]) {
                   1656:                                /* Quoted quotes collapse. */
                   1657:                                pairs++;
                   1658:                                cp++;
                   1659:                        } else {
                   1660:                                /* Unquoted quotes end quoted args. */
                   1661:                                quoted = 2;
                   1662:                                break;
                   1663:                        }
                   1664:                }
                   1665:        }
                   1666:
                   1667:        /* Quoted argument without a closing quote. */
                   1668:        if (1 == quoted)
                   1669:                mandoc_msg(MANDOCERR_ARG_QUOTE, ln, *pos, NULL);
                   1670:
                   1671:        /* NUL-terminate this argument and move to the next one. */
                   1672:        if (pairs)
                   1673:                cp[-pairs] = '\0';
                   1674:        if ('\0' != *cp) {
                   1675:                *cp++ = '\0';
                   1676:                while (' ' == *cp)
                   1677:                        cp++;
                   1678:        }
                   1679:        *pos += (int)(cp - start) + (quoted ? 1 : 0);
                   1680:        *cpp = cp;
                   1681:
                   1682:        if ('\0' == *cp && (white || ' ' == cp[-1]))
                   1683:                mandoc_msg(MANDOCERR_SPACE_EOL, ln, *pos, NULL);
                   1684:
1.355     schwarze 1685:        start = mandoc_strdup(start);
                   1686:        if (newesc == 0)
                   1687:                return start;
                   1688:
                   1689:        buf.buf = start;
                   1690:        buf.sz = strlen(start) + 1;
                   1691:        buf.next = NULL;
1.398   ! schwarze 1692:        if (roff_expand(r, &buf, ln, 0, '\\') == ROFF_IGN) {
1.355     schwarze 1693:                free(buf.buf);
                   1694:                buf.buf = mandoc_strdup("");
                   1695:        }
                   1696:        return buf.buf;
1.353     schwarze 1697: }
                   1698:
1.154     kristaps 1699:
                   1700: /*
1.275     schwarze 1701:  * Process text streams.
1.154     kristaps 1702:  */
1.340     schwarze 1703: static int
1.305     schwarze 1704: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154     kristaps 1705: {
                   1706:        size_t           sz;
                   1707:        const char      *start;
1.178     schwarze 1708:        char            *p;
                   1709:        int              isz;
1.154     kristaps 1710:        enum mandoc_esc  esc;
                   1711:
1.275     schwarze 1712:        /* Spring the input line trap. */
                   1713:
                   1714:        if (roffit_lines == 1) {
                   1715:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1716:                free(buf->buf);
                   1717:                buf->buf = p;
                   1718:                buf->sz = isz + 1;
                   1719:                *offs = 0;
                   1720:                free(roffit_macro);
                   1721:                roffit_lines = 0;
1.277     schwarze 1722:                return ROFF_REPARSE;
1.275     schwarze 1723:        } else if (roffit_lines > 1)
                   1724:                --roffit_lines;
                   1725:
1.305     schwarze 1726:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1727:                if (roffce_lines < 1) {
                   1728:                        r->man->last = roffce_node;
                   1729:                        r->man->next = ROFF_NEXT_SIBLING;
                   1730:                        roffce_lines = 0;
                   1731:                        roffce_node = NULL;
                   1732:                } else
                   1733:                        roffce_lines--;
                   1734:        }
                   1735:
1.275     schwarze 1736:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1737:
1.238     schwarze 1738:        start = p = buf->buf + pos;
1.154     kristaps 1739:
1.238     schwarze 1740:        while (*p != '\0') {
1.154     kristaps 1741:                sz = strcspn(p, "-\\");
                   1742:                p += sz;
                   1743:
1.238     schwarze 1744:                if (*p == '\0')
1.159     kristaps 1745:                        break;
                   1746:
1.238     schwarze 1747:                if (*p == '\\') {
1.154     kristaps 1748:                        /* Skip over escapes. */
                   1749:                        p++;
1.189     schwarze 1750:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238     schwarze 1751:                        if (esc == ESCAPE_ERROR)
1.154     kristaps 1752:                                break;
1.264     schwarze 1753:                        while (*p == '-')
                   1754:                                p++;
1.155     kristaps 1755:                        continue;
1.159     kristaps 1756:                } else if (p == start) {
1.158     kristaps 1757:                        p++;
1.155     kristaps 1758:                        continue;
1.158     kristaps 1759:                }
1.155     kristaps 1760:
1.171     schwarze 1761:                if (isalpha((unsigned char)p[-1]) &&
                   1762:                    isalpha((unsigned char)p[1]))
1.155     kristaps 1763:                        *p = ASCII_HYPH;
                   1764:                p++;
1.94      kristaps 1765:        }
1.277     schwarze 1766:        return ROFF_CONT;
1.94      kristaps 1767: }
                   1768:
1.340     schwarze 1769: int
1.377     schwarze 1770: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
1.67      kristaps 1771: {
1.294     schwarze 1772:        enum roff_tok    t;
1.340     schwarze 1773:        int              e;
1.238     schwarze 1774:        int              pos;   /* parse point */
1.243     schwarze 1775:        int              spos;  /* saved parse point for messages */
1.238     schwarze 1776:        int              ppos;  /* original offset in buf->buf */
                   1777:        int              ctl;   /* macro line (boolean) */
                   1778:
                   1779:        ppos = pos = *offs;
1.377     schwarze 1780:
                   1781:        if (len > 80 && r->tbl == NULL && r->eqn == NULL &&
                   1782:            (r->man->flags & ROFF_NOFILL) == 0 &&
                   1783:            strchr(" .\\", buf->buf[pos]) == NULL &&
                   1784:            buf->buf[pos] != r->control &&
                   1785:            strcspn(buf->buf, " ") < 80)
                   1786:                mandoc_msg(MANDOCERR_TEXT_LONG, ln, (int)len - 1,
                   1787:                    "%.20s...", buf->buf + pos);
1.79      kristaps 1788:
1.230     schwarze 1789:        /* Handle in-line equation delimiters. */
                   1790:
1.236     schwarze 1791:        if (r->tbl == NULL &&
                   1792:            r->last_eqn != NULL && r->last_eqn->delim &&
1.230     schwarze 1793:            (r->eqn == NULL || r->eqn_inline)) {
1.238     schwarze 1794:                e = roff_eqndelim(r, buf, pos);
1.230     schwarze 1795:                if (e == ROFF_REPARSE)
1.277     schwarze 1796:                        return e;
1.230     schwarze 1797:                assert(e == ROFF_CONT);
                   1798:        }
                   1799:
1.387     schwarze 1800:        /* Handle comments and escape sequences. */
                   1801:
                   1802:        e = roff_parse_comment(r, buf, ln, pos, r->escape);
                   1803:        if ((e & ROFF_MASK) == ROFF_IGN)
                   1804:                return e;
                   1805:        assert(e == ROFF_CONT);
1.94      kristaps 1806:
1.355     schwarze 1807:        e = roff_expand(r, buf, ln, pos, r->escape);
1.340     schwarze 1808:        if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1809:                return e;
1.238     schwarze 1810:        assert(e == ROFF_CONT);
1.94      kristaps 1811:
1.238     schwarze 1812:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.130     kristaps 1813:
1.94      kristaps 1814:        /*
1.79      kristaps 1815:         * First, if a scope is open and we're not a macro, pass the
1.252     schwarze 1816:         * text through the macro's filter.
                   1817:         * Equations process all content themselves.
                   1818:         * Tables process almost all content themselves, but we want
                   1819:         * to warn about macros before passing it there.
1.79      kristaps 1820:         */
1.74      kristaps 1821:
1.252     schwarze 1822:        if (r->last != NULL && ! ctl) {
1.78      kristaps 1823:                t = r->last->tok;
1.238     schwarze 1824:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.340     schwarze 1825:                if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1826:                        return e;
1.340     schwarze 1827:                e &= ~ROFF_MASK;
                   1828:        } else
                   1829:                e = ROFF_IGN;
1.319     schwarze 1830:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1831:                eqn_read(r->eqn, buf->buf + ppos);
1.340     schwarze 1832:                return e;
1.319     schwarze 1833:        }
1.321     schwarze 1834:        if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
                   1835:                tbl_read(r->tbl, ln, buf->buf, ppos);
1.346     schwarze 1836:                roff_addtbl(r->man, ln, r->tbl);
1.340     schwarze 1837:                return e;
1.321     schwarze 1838:        }
1.367     schwarze 1839:        if ( ! ctl) {
                   1840:                r->options &= ~MPARSE_COMMENT;
1.340     schwarze 1841:                return roff_parsetext(r, buf, pos, offs) | e;
1.367     schwarze 1842:        }
1.228     schwarze 1843:
                   1844:        /* Skip empty request lines. */
                   1845:
1.238     schwarze 1846:        if (buf->buf[pos] == '"') {
1.350     schwarze 1847:                mandoc_msg(MANDOCERR_COMMENT_BAD, ln, pos, NULL);
1.277     schwarze 1848:                return ROFF_IGN;
1.238     schwarze 1849:        } else if (buf->buf[pos] == '\0')
1.277     schwarze 1850:                return ROFF_IGN;
1.67      kristaps 1851:
1.79      kristaps 1852:        /*
                   1853:         * If a scope is open, go to the child handler for that macro,
                   1854:         * as it may want to preprocess before doing anything with it.
                   1855:         */
1.78      kristaps 1856:
1.79      kristaps 1857:        if (r->last) {
                   1858:                t = r->last->tok;
1.277     schwarze 1859:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1860:        }
1.78      kristaps 1861:
1.367     schwarze 1862:        r->options &= ~MPARSE_COMMENT;
1.243     schwarze 1863:        spos = pos;
                   1864:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.386     schwarze 1865:        return roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
                   1866: }
                   1867:
                   1868: /*
                   1869:  * Handle a new request or macro.
                   1870:  * May be called outside any scope or from inside a conditional scope.
                   1871:  */
                   1872: static int
                   1873: roff_req_or_macro(ROFF_ARGS) {
1.243     schwarze 1874:
1.386     schwarze 1875:        /* For now, tables ignore most macros and some request. */
1.243     schwarze 1876:
1.386     schwarze 1877:        if (r->tbl != NULL && (tok == TOKEN_NONE || tok == ROFF_TS ||
                   1878:            tok == ROFF_br || tok == ROFF_ce || tok == ROFF_rj ||
                   1879:            tok == ROFF_sp)) {
1.350     schwarze 1880:                mandoc_msg(MANDOCERR_TBLMACRO,
1.386     schwarze 1881:                    ln, ppos, "%s", buf->buf + ppos);
                   1882:                if (tok != TOKEN_NONE)
1.277     schwarze 1883:                        return ROFF_IGN;
1.257     schwarze 1884:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1885:                        pos++;
1.291     schwarze 1886:                while (buf->buf[pos] == ' ')
1.257     schwarze 1887:                        pos++;
1.321     schwarze 1888:                tbl_read(r->tbl, ln, buf->buf, pos);
1.346     schwarze 1889:                roff_addtbl(r->man, ln, r->tbl);
1.321     schwarze 1890:                return ROFF_IGN;
1.243     schwarze 1891:        }
                   1892:
1.305     schwarze 1893:        /* For now, let high level macros abort .ce mode. */
                   1894:
1.386     schwarze 1895:        if (roffce_node != NULL &&
                   1896:            (tok == TOKEN_NONE || tok == ROFF_Dd || tok == ROFF_EQ ||
                   1897:             tok == ROFF_TH || tok == ROFF_TS)) {
1.305     schwarze 1898:                r->man->last = roffce_node;
                   1899:                r->man->next = ROFF_NEXT_SIBLING;
                   1900:                roffce_lines = 0;
                   1901:                roffce_node = NULL;
                   1902:        }
                   1903:
1.79      kristaps 1904:        /*
1.243     schwarze 1905:         * This is neither a roff request nor a user-defined macro.
                   1906:         * Let the standard macro set parsers handle it.
1.79      kristaps 1907:         */
1.67      kristaps 1908:
1.386     schwarze 1909:        if (tok == TOKEN_NONE)
1.277     schwarze 1910:                return ROFF_CONT;
1.243     schwarze 1911:
1.386     schwarze 1912:        /* Execute a roff request or a user-defined macro. */
1.67      kristaps 1913:
1.386     schwarze 1914:        return (*roffs[tok].proc)(r, tok, buf, ln, ppos, pos, offs);
1.74      kristaps 1915: }
                   1916:
1.339     schwarze 1917: /*
                   1918:  * Internal interface function to tell the roff parser that execution
                   1919:  * of the current macro ended.  This is required because macro
                   1920:  * definitions usually do not end with a .return request.
                   1921:  */
                   1922: void
                   1923: roff_userret(struct roff *r)
                   1924: {
                   1925:        struct mctx     *ctx;
                   1926:        int              i;
                   1927:
                   1928:        assert(r->mstackpos >= 0);
                   1929:        ctx = r->mstack + r->mstackpos;
                   1930:        for (i = 0; i < ctx->argc; i++)
                   1931:                free(ctx->argv[i]);
                   1932:        ctx->argc = 0;
                   1933:        r->mstackpos--;
                   1934: }
                   1935:
1.117     kristaps 1936: void
1.74      kristaps 1937: roff_endparse(struct roff *r)
                   1938: {
1.321     schwarze 1939:        if (r->last != NULL)
1.350     schwarze 1940:                mandoc_msg(MANDOCERR_BLK_NOEND, r->last->line,
                   1941:                    r->last->col, "%s", roff_name[r->last->tok]);
1.117     kristaps 1942:
1.321     schwarze 1943:        if (r->eqn != NULL) {
1.350     schwarze 1944:                mandoc_msg(MANDOCERR_BLK_NOEND,
1.319     schwarze 1945:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1946:                eqn_parse(r->eqn);
                   1947:                r->eqn = NULL;
1.125     kristaps 1948:        }
                   1949:
1.321     schwarze 1950:        if (r->tbl != NULL) {
1.346     schwarze 1951:                tbl_end(r->tbl, 1);
1.321     schwarze 1952:                r->tbl = NULL;
1.117     kristaps 1953:        }
1.67      kristaps 1954: }
                   1955:
                   1956: /*
1.386     schwarze 1957:  * Parse the request or macro name at buf[*pos].
                   1958:  * Return ROFF_RENAMED, ROFF_USERDEF, or a ROFF_* token value.
                   1959:  * For empty, undefined, mdoc(7), and man(7) macros, return TOKEN_NONE.
                   1960:  * As a side effect, set r->current_string to the definition or to NULL.
1.67      kristaps 1961:  */
1.294     schwarze 1962: static enum roff_tok
1.214     schwarze 1963: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67      kristaps 1964: {
1.214     schwarze 1965:        char            *cp;
1.106     kristaps 1966:        const char      *mac;
                   1967:        size_t           maclen;
1.315     schwarze 1968:        int              deftype;
1.294     schwarze 1969:        enum roff_tok    t;
1.67      kristaps 1970:
1.214     schwarze 1971:        cp = buf + *pos;
                   1972:
                   1973:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294     schwarze 1974:                return TOKEN_NONE;
1.67      kristaps 1975:
1.214     schwarze 1976:        mac = cp;
                   1977:        maclen = roff_getname(r, &cp, ln, ppos);
1.67      kristaps 1978:
1.315     schwarze 1979:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1980:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1981:        switch (deftype) {
                   1982:        case ROFFDEF_USER:
                   1983:                t = ROFF_USERDEF;
                   1984:                break;
                   1985:        case ROFFDEF_REN:
                   1986:                t = ROFF_RENAMED;
                   1987:                break;
                   1988:        default:
                   1989:                t = roffhash_find(r->reqtab, mac, maclen);
                   1990:                break;
                   1991:        }
1.294     schwarze 1992:        if (t != TOKEN_NONE)
1.214     schwarze 1993:                *pos = cp - buf;
1.325     schwarze 1994:        else if (deftype == ROFFDEF_UNDEF) {
                   1995:                /* Using an undefined macro defines it to be empty. */
                   1996:                roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
                   1997:                roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
                   1998:        }
1.277     schwarze 1999:        return t;
1.67      kristaps 2000: }
                   2001:
1.266     schwarze 2002: /* --- handling of request blocks ----------------------------------------- */
                   2003:
1.375     schwarze 2004: /*
                   2005:  * Close a macro definition block or an "ignore" block.
                   2006:  */
1.340     schwarze 2007: static int
1.76      kristaps 2008: roff_cblock(ROFF_ARGS)
1.67      kristaps 2009: {
1.375     schwarze 2010:        int      rr;
1.79      kristaps 2011:
1.238     schwarze 2012:        if (r->last == NULL) {
1.350     schwarze 2013:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.277     schwarze 2014:                return ROFF_IGN;
1.76      kristaps 2015:        }
1.67      kristaps 2016:
1.81      kristaps 2017:        switch (r->last->tok) {
1.207     schwarze 2018:        case ROFF_am:
                   2019:        case ROFF_ami:
                   2020:        case ROFF_de:
                   2021:        case ROFF_dei:
                   2022:        case ROFF_ig:
1.81      kristaps 2023:                break;
1.375     schwarze 2024:        case ROFF_am1:
                   2025:        case ROFF_de1:
                   2026:                /* Remapped in roff_block(). */
                   2027:                abort();
1.81      kristaps 2028:        default:
1.350     schwarze 2029:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.277     schwarze 2030:                return ROFF_IGN;
1.76      kristaps 2031:        }
1.67      kristaps 2032:
1.375     schwarze 2033:        roffnode_pop(r);
                   2034:        roffnode_cleanscope(r);
                   2035:
                   2036:        /*
                   2037:         * If a conditional block with braces is still open,
                   2038:         * check for "\}" block end markers.
                   2039:         */
                   2040:
                   2041:        if (r->last != NULL && r->last->endspan < 0) {
                   2042:                rr = 1;  /* If arguments follow "\}", warn about them. */
                   2043:                roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
                   2044:        }
                   2045:
1.238     schwarze 2046:        if (buf->buf[pos] != '\0')
1.350     schwarze 2047:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 2048:                    ".. %s", buf->buf + pos);
1.71      kristaps 2049:
1.277     schwarze 2050:        return ROFF_IGN;
1.67      kristaps 2051: }
                   2052:
1.364     schwarze 2053: /*
                   2054:  * Pop all nodes ending at the end of the current input line.
                   2055:  * Return the number of loops ended.
                   2056:  */
1.340     schwarze 2057: static int
1.76      kristaps 2058: roffnode_cleanscope(struct roff *r)
1.67      kristaps 2059: {
1.340     schwarze 2060:        int inloop;
1.67      kristaps 2061:
1.340     schwarze 2062:        inloop = 0;
1.375     schwarze 2063:        while (r->last != NULL && r->last->endspan > 0) {
1.173     schwarze 2064:                if (--r->last->endspan != 0)
1.76      kristaps 2065:                        break;
1.340     schwarze 2066:                inloop += roffnode_pop(r);
1.76      kristaps 2067:        }
1.340     schwarze 2068:        return inloop;
1.67      kristaps 2069: }
                   2070:
1.364     schwarze 2071: /*
1.375     schwarze 2072:  * Handle the closing "\}" of a conditional block.
1.364     schwarze 2073:  * Apart from generating warnings, this only pops nodes.
                   2074:  * Return the number of loops ended.
                   2075:  */
1.340     schwarze 2076: static int
1.195     schwarze 2077: roff_ccond(struct roff *r, int ln, int ppos)
1.74      kristaps 2078: {
1.76      kristaps 2079:        if (NULL == r->last) {
1.350     schwarze 2080:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2081:                return 0;
1.76      kristaps 2082:        }
                   2083:
1.82      kristaps 2084:        switch (r->last->tok) {
1.207     schwarze 2085:        case ROFF_el:
                   2086:        case ROFF_ie:
                   2087:        case ROFF_if:
1.340     schwarze 2088:        case ROFF_while:
1.82      kristaps 2089:                break;
                   2090:        default:
1.350     schwarze 2091:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2092:                return 0;
1.75      kristaps 2093:        }
                   2094:
1.76      kristaps 2095:        if (r->last->endspan > -1) {
1.350     schwarze 2096:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2097:                return 0;
1.76      kristaps 2098:        }
                   2099:
1.340     schwarze 2100:        return roffnode_pop(r) + roffnode_cleanscope(r);
1.76      kristaps 2101: }
                   2102:
1.340     schwarze 2103: static int
1.80      kristaps 2104: roff_block(ROFF_ARGS)
1.76      kristaps 2105: {
1.315     schwarze 2106:        const char      *name, *value;
                   2107:        char            *call, *cp, *iname, *rname;
                   2108:        size_t           csz, namesz, rsz;
                   2109:        int              deftype;
1.106     kristaps 2110:
1.220     schwarze 2111:        /* Ignore groff compatibility mode for now. */
1.76      kristaps 2112:
1.238     schwarze 2113:        if (tok == ROFF_de1)
1.220     schwarze 2114:                tok = ROFF_de;
1.251     schwarze 2115:        else if (tok == ROFF_dei1)
                   2116:                tok = ROFF_dei;
1.238     schwarze 2117:        else if (tok == ROFF_am1)
1.220     schwarze 2118:                tok = ROFF_am;
1.251     schwarze 2119:        else if (tok == ROFF_ami1)
                   2120:                tok = ROFF_ami;
1.220     schwarze 2121:
                   2122:        /* Parse the macro name argument. */
                   2123:
1.238     schwarze 2124:        cp = buf->buf + pos;
                   2125:        if (tok == ROFF_ig) {
1.220     schwarze 2126:                iname = NULL;
                   2127:                namesz = 0;
                   2128:        } else {
                   2129:                iname = cp;
                   2130:                namesz = roff_getname(r, &cp, ln, ppos);
                   2131:                iname[namesz] = '\0';
                   2132:        }
1.107     kristaps 2133:
1.220     schwarze 2134:        /* Resolve the macro name argument if it is indirect. */
1.107     kristaps 2135:
1.238     schwarze 2136:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2137:                deftype = ROFFDEF_USER;
                   2138:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2139:                if (name == NULL) {
1.350     schwarze 2140:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2141:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2142:                            "%.*s", (int)namesz, iname);
                   2143:                        namesz = 0;
                   2144:                } else
                   2145:                        namesz = strlen(name);
                   2146:        } else
                   2147:                name = iname;
1.107     kristaps 2148:
1.238     schwarze 2149:        if (namesz == 0 && tok != ROFF_ig) {
1.350     schwarze 2150:                mandoc_msg(MANDOCERR_REQ_EMPTY,
                   2151:                    ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2152:                return ROFF_IGN;
1.220     schwarze 2153:        }
1.80      kristaps 2154:
1.106     kristaps 2155:        roffnode_push(r, tok, name, ln, ppos);
                   2156:
                   2157:        /*
                   2158:         * At the beginning of a `de' macro, clear the existing string
                   2159:         * with the same name, if there is one.  New content will be
1.193     schwarze 2160:         * appended from roff_block_text() in multiline mode.
1.106     kristaps 2161:         */
1.107     kristaps 2162:
1.311     schwarze 2163:        if (tok == ROFF_de || tok == ROFF_dei) {
1.213     schwarze 2164:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311     schwarze 2165:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315     schwarze 2166:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   2167:                deftype = ROFFDEF_ANY;
                   2168:                value = roff_getstrn(r, iname, namesz, &deftype);
                   2169:                switch (deftype) {  /* Before appending, ... */
                   2170:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   2171:                        roff_setstrn(&r->strtab, name, namesz,
                   2172:                            value, strlen(value), 0);
                   2173:                        break;
                   2174:                case ROFFDEF_REN: /* call original standard macro. */
                   2175:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2176:                            (int)strlen(value), value);
                   2177:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2178:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   2179:                        free(call);
                   2180:                        break;
                   2181:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   2182:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   2183:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   2184:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2185:                            (int)rsz, rname);
                   2186:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2187:                        free(call);
                   2188:                        free(rname);
                   2189:                        break;
                   2190:                default:
                   2191:                        break;
                   2192:                }
1.311     schwarze 2193:        }
1.76      kristaps 2194:
1.238     schwarze 2195:        if (*cp == '\0')
1.277     schwarze 2196:                return ROFF_IGN;
1.78      kristaps 2197:
1.220     schwarze 2198:        /* Get the custom end marker. */
1.107     kristaps 2199:
1.220     schwarze 2200:        iname = cp;
1.213     schwarze 2201:        namesz = roff_getname(r, &cp, ln, ppos);
1.220     schwarze 2202:
                   2203:        /* Resolve the end marker if it is indirect. */
                   2204:
1.238     schwarze 2205:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2206:                deftype = ROFFDEF_USER;
                   2207:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2208:                if (name == NULL) {
1.350     schwarze 2209:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2210:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2211:                            "%.*s", (int)namesz, iname);
                   2212:                        namesz = 0;
                   2213:                } else
                   2214:                        namesz = strlen(name);
                   2215:        } else
                   2216:                name = iname;
                   2217:
1.213     schwarze 2218:        if (namesz)
                   2219:                r->last->end = mandoc_strndup(name, namesz);
1.78      kristaps 2220:
1.238     schwarze 2221:        if (*cp != '\0')
1.350     schwarze 2222:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.295     schwarze 2223:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74      kristaps 2224:
1.277     schwarze 2225:        return ROFF_IGN;
1.78      kristaps 2226: }
                   2227:
1.340     schwarze 2228: static int
1.80      kristaps 2229: roff_block_sub(ROFF_ARGS)
1.79      kristaps 2230: {
1.294     schwarze 2231:        enum roff_tok   t;
1.79      kristaps 2232:        int             i, j;
                   2233:
                   2234:        /*
1.385     schwarze 2235:         * If a custom end marker is a user-defined or predefined macro
                   2236:         * or a request, interpret it.
1.79      kristaps 2237:         */
                   2238:
                   2239:        if (r->last->end) {
1.130     kristaps 2240:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238     schwarze 2241:                        if (buf->buf[i] != r->last->end[j])
1.79      kristaps 2242:                                break;
                   2243:
1.238     schwarze 2244:                if (r->last->end[j] == '\0' &&
                   2245:                    (buf->buf[i] == '\0' ||
                   2246:                     buf->buf[i] == ' ' ||
                   2247:                     buf->buf[i] == '\t')) {
1.79      kristaps 2248:                        roffnode_pop(r);
                   2249:                        roffnode_cleanscope(r);
                   2250:
1.238     schwarze 2251:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130     kristaps 2252:                                i++;
                   2253:
                   2254:                        pos = i;
1.238     schwarze 2255:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294     schwarze 2256:                            TOKEN_NONE)
1.277     schwarze 2257:                                return ROFF_RERUN;
                   2258:                        return ROFF_IGN;
1.79      kristaps 2259:                }
                   2260:        }
                   2261:
1.385     schwarze 2262:        /* Handle the standard end marker. */
1.79      kristaps 2263:
1.238     schwarze 2264:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.385     schwarze 2265:        if (t == ROFF_cblock)
                   2266:                return roff_cblock(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 2267:
1.385     schwarze 2268:        /* Not an end marker, so append the line to the block. */
1.79      kristaps 2269:
1.385     schwarze 2270:        if (tok != ROFF_ig)
                   2271:                roff_setstr(r, r->last->name, buf->buf + ppos, 2);
                   2272:        return ROFF_IGN;
1.79      kristaps 2273: }
                   2274:
1.340     schwarze 2275: static int
1.80      kristaps 2276: roff_block_text(ROFF_ARGS)
1.78      kristaps 2277: {
                   2278:
1.238     schwarze 2279:        if (tok != ROFF_ig)
                   2280:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106     kristaps 2281:
1.277     schwarze 2282:        return ROFF_IGN;
1.78      kristaps 2283: }
                   2284:
1.375     schwarze 2285: /*
                   2286:  * Check for a closing "\}" and handle it.
                   2287:  * In this function, the final "int *offs" argument is used for
                   2288:  * different purposes than elsewhere:
                   2289:  * Input: *offs == 0: caller wants to discard arguments following \}
                   2290:  *        *offs == 1: caller wants to preserve text following \}
                   2291:  * Output: *offs = 0: tell caller to discard input line
                   2292:  *         *offs = 1: tell caller to use input line
                   2293:  */
1.340     schwarze 2294: static int
1.375     schwarze 2295: roff_cond_checkend(ROFF_ARGS)
1.82      kristaps 2296: {
1.340     schwarze 2297:        char            *ep;
                   2298:        int              endloop, irc, rr;
1.82      kristaps 2299:
1.340     schwarze 2300:        irc = ROFF_IGN;
1.82      kristaps 2301:        rr = r->last->rule;
1.340     schwarze 2302:        endloop = tok != ROFF_while ? ROFF_IGN :
                   2303:            rr ? ROFF_LOOPCONT : ROFF_LOOPEXIT;
                   2304:        if (roffnode_cleanscope(r))
                   2305:                irc |= endloop;
1.144     kristaps 2306:
1.196     schwarze 2307:        /*
1.375     schwarze 2308:         * If "\}" occurs on a macro line without a preceding macro or
                   2309:         * a text line contains nothing else, drop the line completely.
1.196     schwarze 2310:         */
                   2311:
1.238     schwarze 2312:        ep = buf->buf + pos;
1.375     schwarze 2313:        if (ep[0] == '\\' && ep[1] == '}' && (ep[2] == '\0' || *offs == 0))
1.198     schwarze 2314:                rr = 0;
1.196     schwarze 2315:
1.343     schwarze 2316:        /*
1.375     schwarze 2317:         * The closing delimiter "\}" rewinds the conditional scope
1.343     schwarze 2318:         * but is otherwise ignored when interpreting the line.
                   2319:         */
1.144     kristaps 2320:
1.238     schwarze 2321:        while ((ep = strchr(ep, '\\')) != NULL) {
1.318     schwarze 2322:                switch (ep[1]) {
                   2323:                case '}':
1.376     schwarze 2324:                        if (ep[2] == '\0')
                   2325:                                ep[0] = '\0';
                   2326:                        else if (rr)
1.375     schwarze 2327:                                ep[1] = '&';
                   2328:                        else
                   2329:                                memmove(ep, ep + 2, strlen(ep + 2) + 1);
1.340     schwarze 2330:                        if (roff_ccond(r, ln, ep - buf->buf))
                   2331:                                irc |= endloop;
1.318     schwarze 2332:                        break;
                   2333:                case '\0':
                   2334:                        ++ep;
                   2335:                        break;
                   2336:                default:
                   2337:                        ep += 2;
                   2338:                        break;
1.197     schwarze 2339:                }
1.177     schwarze 2340:        }
1.375     schwarze 2341:        *offs = rr;
                   2342:        return irc;
                   2343: }
1.318     schwarze 2344:
1.375     schwarze 2345: /*
                   2346:  * Parse and process a request or macro line in conditional scope.
                   2347:  */
                   2348: static int
                   2349: roff_cond_sub(ROFF_ARGS)
                   2350: {
                   2351:        struct roffnode *bl;
1.386     schwarze 2352:        int              irc, rr, spos;
1.375     schwarze 2353:        enum roff_tok    t;
                   2354:
                   2355:        rr = 0;  /* If arguments follow "\}", skip them. */
                   2356:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.386     schwarze 2357:        spos = pos;
1.368     schwarze 2358:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   2359:
1.318     schwarze 2360:        /*
1.386     schwarze 2361:         * Handle requests and macros if the conditional evaluated
                   2362:         * to true or if they are structurally required.
                   2363:         * The .break request is always handled specially.
1.318     schwarze 2364:         */
                   2365:
1.364     schwarze 2366:        if (t == ROFF_break) {
                   2367:                if (irc & ROFF_LOOPMASK)
                   2368:                        irc = ROFF_IGN | ROFF_LOOPEXIT;
                   2369:                else if (rr) {
                   2370:                        for (bl = r->last; bl != NULL; bl = bl->parent) {
                   2371:                                bl->rule = 0;
                   2372:                                if (bl->tok == ROFF_while)
                   2373:                                        break;
                   2374:                        }
                   2375:                }
1.386     schwarze 2376:        } else if (rr || (t < TOKEN_NONE && roffs[t].flags & ROFFMAC_STRUCT)) {
                   2377:                irc |= roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
1.383     schwarze 2378:                if (irc & ROFF_WHILE)
                   2379:                        irc &= ~(ROFF_LOOPCONT | ROFF_LOOPEXIT);
1.386     schwarze 2380:        }
1.340     schwarze 2381:        return irc;
1.82      kristaps 2382: }
                   2383:
1.375     schwarze 2384: /*
                   2385:  * Parse and process a text line in conditional scope.
                   2386:  */
1.340     schwarze 2387: static int
1.82      kristaps 2388: roff_cond_text(ROFF_ARGS)
1.78      kristaps 2389: {
1.375     schwarze 2390:        int      irc, rr;
1.343     schwarze 2391:
1.375     schwarze 2392:        rr = 1;  /* If arguments follow "\}", preserve them. */
                   2393:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.340     schwarze 2394:        if (rr)
                   2395:                irc |= ROFF_CONT;
                   2396:        return irc;
1.74      kristaps 2397: }
                   2398:
1.266     schwarze 2399: /* --- handling of numeric and conditional expressions -------------------- */
                   2400:
1.204     schwarze 2401: /*
                   2402:  * Parse a single signed integer number.  Stop at the first non-digit.
                   2403:  * If there is at least one digit, return success and advance the
                   2404:  * parse point, else return failure and let the parse point unchanged.
                   2405:  * Ignore overflows, treat them just like the C language.
                   2406:  */
1.184     schwarze 2407: static int
1.261     schwarze 2408: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184     schwarze 2409: {
1.261     schwarze 2410:        int      myres, scaled, n, p;
1.206     schwarze 2411:
                   2412:        if (NULL == res)
                   2413:                res = &myres;
1.184     schwarze 2414:
                   2415:        p = *pos;
                   2416:        n = v[p] == '-';
1.261     schwarze 2417:        if (n || v[p] == '+')
1.184     schwarze 2418:                p++;
                   2419:
1.261     schwarze 2420:        if (flags & ROFFNUM_WHITE)
                   2421:                while (isspace((unsigned char)v[p]))
                   2422:                        p++;
                   2423:
1.184     schwarze 2424:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204     schwarze 2425:                *res = 10 * *res + v[p] - '0';
1.184     schwarze 2426:        if (p == *pos + n)
                   2427:                return 0;
                   2428:
                   2429:        if (n)
                   2430:                *res = -*res;
                   2431:
1.254     schwarze 2432:        /* Each number may be followed by one optional scaling unit. */
                   2433:
                   2434:        switch (v[p]) {
                   2435:        case 'f':
1.261     schwarze 2436:                scaled = *res * 65536;
1.254     schwarze 2437:                break;
                   2438:        case 'i':
1.261     schwarze 2439:                scaled = *res * 240;
1.254     schwarze 2440:                break;
                   2441:        case 'c':
1.261     schwarze 2442:                scaled = *res * 240 / 2.54;
1.254     schwarze 2443:                break;
                   2444:        case 'v':
                   2445:        case 'P':
1.261     schwarze 2446:                scaled = *res * 40;
1.254     schwarze 2447:                break;
                   2448:        case 'm':
                   2449:        case 'n':
1.261     schwarze 2450:                scaled = *res * 24;
1.254     schwarze 2451:                break;
                   2452:        case 'p':
1.261     schwarze 2453:                scaled = *res * 10 / 3;
1.254     schwarze 2454:                break;
                   2455:        case 'u':
1.261     schwarze 2456:                scaled = *res;
1.254     schwarze 2457:                break;
                   2458:        case 'M':
1.261     schwarze 2459:                scaled = *res * 6 / 25;
1.254     schwarze 2460:                break;
                   2461:        default:
1.261     schwarze 2462:                scaled = *res;
1.254     schwarze 2463:                p--;
                   2464:                break;
                   2465:        }
1.261     schwarze 2466:        if (flags & ROFFNUM_SCALE)
                   2467:                *res = scaled;
1.254     schwarze 2468:
                   2469:        *pos = p + 1;
1.277     schwarze 2470:        return 1;
1.184     schwarze 2471: }
                   2472:
1.198     schwarze 2473: /*
                   2474:  * Evaluate a string comparison condition.
                   2475:  * The first character is the delimiter.
                   2476:  * Succeed if the string up to its second occurrence
1.396     schwarze 2477:  * matches the string up to its third occurrence.
1.198     schwarze 2478:  * Advance the cursor after the third occurrence
                   2479:  * or lacking that, to the end of the line.
                   2480:  */
                   2481: static int
                   2482: roff_evalstrcond(const char *v, int *pos)
                   2483: {
                   2484:        const char      *s1, *s2, *s3;
                   2485:        int              match;
                   2486:
                   2487:        match = 0;
                   2488:        s1 = v + *pos;          /* initial delimiter */
                   2489:        s2 = s1 + 1;            /* for scanning the first string */
                   2490:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2491:
                   2492:        if (NULL == s3)         /* found no middle delimiter */
                   2493:                goto out;
                   2494:
                   2495:        while ('\0' != *++s3) {
                   2496:                if (*s2 != *s3) {  /* mismatch */
                   2497:                        s3 = strchr(s3, *s1);
                   2498:                        break;
                   2499:                }
                   2500:                if (*s3 == *s1) {  /* found the final delimiter */
                   2501:                        match = 1;
                   2502:                        break;
                   2503:                }
                   2504:                s2++;
                   2505:        }
                   2506:
                   2507: out:
                   2508:        if (NULL == s3)
                   2509:                s3 = strchr(s2, '\0');
1.242     schwarze 2510:        else if (*s3 != '\0')
1.198     schwarze 2511:                s3++;
                   2512:        *pos = s3 - v;
1.277     schwarze 2513:        return match;
1.198     schwarze 2514: }
                   2515:
1.204     schwarze 2516: /*
                   2517:  * Evaluate an optionally negated single character, numerical,
                   2518:  * or string condition.
                   2519:  */
1.198     schwarze 2520: static int
1.271     schwarze 2521: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88      kristaps 2522: {
1.336     schwarze 2523:        const char      *start, *end;
                   2524:        char            *cp, *name;
                   2525:        size_t           sz;
                   2526:        int              deftype, len, number, savepos, istrue, wanttrue;
1.88      kristaps 2527:
1.198     schwarze 2528:        if ('!' == v[*pos]) {
                   2529:                wanttrue = 0;
                   2530:                (*pos)++;
                   2531:        } else
                   2532:                wanttrue = 1;
                   2533:
1.88      kristaps 2534:        switch (v[*pos]) {
1.240     schwarze 2535:        case '\0':
1.277     schwarze 2536:                return 0;
1.207     schwarze 2537:        case 'n':
                   2538:        case 'o':
1.88      kristaps 2539:                (*pos)++;
1.277     schwarze 2540:                return wanttrue;
1.207     schwarze 2541:        case 'e':
                   2542:        case 't':
1.239     schwarze 2543:        case 'v':
1.88      kristaps 2544:                (*pos)++;
1.277     schwarze 2545:                return !wanttrue;
1.336     schwarze 2546:        case 'c':
                   2547:                do {
                   2548:                        (*pos)++;
                   2549:                } while (v[*pos] == ' ');
                   2550:
                   2551:                /*
                   2552:                 * Quirk for groff compatibility:
                   2553:                 * The horizontal tab is neither available nor unavailable.
                   2554:                 */
                   2555:
                   2556:                if (v[*pos] == '\t') {
                   2557:                        (*pos)++;
                   2558:                        return 0;
                   2559:                }
                   2560:
                   2561:                /* Printable ASCII characters are available. */
                   2562:
                   2563:                if (v[*pos] != '\\') {
                   2564:                        (*pos)++;
                   2565:                        return wanttrue;
                   2566:                }
                   2567:
                   2568:                end = v + ++*pos;
                   2569:                switch (mandoc_escape(&end, &start, &len)) {
                   2570:                case ESCAPE_SPECIAL:
                   2571:                        istrue = mchars_spec2cp(start, len) != -1;
                   2572:                        break;
                   2573:                case ESCAPE_UNICODE:
                   2574:                        istrue = 1;
                   2575:                        break;
                   2576:                case ESCAPE_NUMBERED:
                   2577:                        istrue = mchars_num2char(start, len) != -1;
                   2578:                        break;
                   2579:                default:
                   2580:                        istrue = !wanttrue;
                   2581:                        break;
                   2582:                }
                   2583:                *pos = end - v;
                   2584:                return istrue == wanttrue;
1.310     schwarze 2585:        case 'd':
1.271     schwarze 2586:        case 'r':
1.310     schwarze 2587:                cp = v + *pos + 1;
                   2588:                while (*cp == ' ')
                   2589:                        cp++;
                   2590:                name = cp;
                   2591:                sz = roff_getname(r, &cp, ln, cp - v);
1.315     schwarze 2592:                if (sz == 0)
                   2593:                        istrue = 0;
                   2594:                else if (v[*pos] == 'r')
                   2595:                        istrue = roff_hasregn(r, name, sz);
                   2596:                else {
                   2597:                        deftype = ROFFDEF_ANY;
                   2598:                        roff_getstrn(r, name, sz, &deftype);
                   2599:                        istrue = !!deftype;
                   2600:                }
1.363     schwarze 2601:                *pos = (name + sz) - v;
1.310     schwarze 2602:                return istrue == wanttrue;
1.88      kristaps 2603:        default:
                   2604:                break;
                   2605:        }
                   2606:
1.241     schwarze 2607:        savepos = *pos;
1.261     schwarze 2608:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277     schwarze 2609:                return (number > 0) == wanttrue;
1.241     schwarze 2610:        else if (*pos == savepos)
1.277     schwarze 2611:                return roff_evalstrcond(v, pos) == wanttrue;
1.204     schwarze 2612:        else
1.277     schwarze 2613:                return 0;
1.88      kristaps 2614: }
                   2615:
1.340     schwarze 2616: static int
1.103     kristaps 2617: roff_line_ignore(ROFF_ARGS)
1.89      kristaps 2618: {
1.123     schwarze 2619:
1.277     schwarze 2620:        return ROFF_IGN;
1.89      kristaps 2621: }
                   2622:
1.340     schwarze 2623: static int
1.251     schwarze 2624: roff_insec(ROFF_ARGS)
                   2625: {
                   2626:
1.350     schwarze 2627:        mandoc_msg(MANDOCERR_REQ_INSEC, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2628:        return ROFF_IGN;
1.251     schwarze 2629: }
                   2630:
1.340     schwarze 2631: static int
1.251     schwarze 2632: roff_unsupp(ROFF_ARGS)
                   2633: {
                   2634:
1.350     schwarze 2635:        mandoc_msg(MANDOCERR_REQ_UNSUPP, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2636:        return ROFF_IGN;
1.251     schwarze 2637: }
                   2638:
1.340     schwarze 2639: static int
1.82      kristaps 2640: roff_cond(ROFF_ARGS)
1.74      kristaps 2641: {
1.340     schwarze 2642:        int      irc;
1.173     schwarze 2643:
                   2644:        roffnode_push(r, tok, NULL, ln, ppos);
1.74      kristaps 2645:
1.207     schwarze 2646:        /*
1.134     kristaps 2647:         * An `.el' has no conditional body: it will consume the value
                   2648:         * of the current rstack entry set in prior `ie' calls or
1.207     schwarze 2649:         * defaults to DENY.
1.134     kristaps 2650:         *
                   2651:         * If we're not an `el', however, then evaluate the conditional.
                   2652:         */
1.133     kristaps 2653:
1.238     schwarze 2654:        r->last->rule = tok == ROFF_el ?
1.207     schwarze 2655:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238     schwarze 2656:            roff_evalcond(r, ln, buf->buf, &pos);
1.77      kristaps 2657:
1.134     kristaps 2658:        /*
                   2659:         * An if-else will put the NEGATION of the current evaluated
                   2660:         * conditional into the stack of rules.
                   2661:         */
                   2662:
1.238     schwarze 2663:        if (tok == ROFF_ie) {
1.223     schwarze 2664:                if (r->rstackpos + 1 == r->rstacksz) {
                   2665:                        r->rstacksz += 16;
                   2666:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2667:                            r->rstacksz, sizeof(int));
1.134     kristaps 2668:                }
1.198     schwarze 2669:                r->rstack[++r->rstackpos] = !r->last->rule;
1.82      kristaps 2670:        }
1.88      kristaps 2671:
                   2672:        /* If the parent has false as its rule, then so do we. */
                   2673:
1.198     schwarze 2674:        if (r->last->parent && !r->last->parent->rule)
                   2675:                r->last->rule = 0;
1.88      kristaps 2676:
                   2677:        /*
1.173     schwarze 2678:         * Determine scope.
                   2679:         * If there is nothing on the line after the conditional,
                   2680:         * not even whitespace, use next-line scope.
1.340     schwarze 2681:         * Except that .while does not support next-line scope.
1.88      kristaps 2682:         */
1.74      kristaps 2683:
1.340     schwarze 2684:        if (buf->buf[pos] == '\0' && tok != ROFF_while) {
1.173     schwarze 2685:                r->last->endspan = 2;
                   2686:                goto out;
                   2687:        }
                   2688:
1.238     schwarze 2689:        while (buf->buf[pos] == ' ')
1.173     schwarze 2690:                pos++;
                   2691:
                   2692:        /* An opening brace requests multiline scope. */
1.75      kristaps 2693:
1.238     schwarze 2694:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75      kristaps 2695:                r->last->endspan = -1;
                   2696:                pos += 2;
1.272     schwarze 2697:                while (buf->buf[pos] == ' ')
                   2698:                        pos++;
1.173     schwarze 2699:                goto out;
1.207     schwarze 2700:        }
1.74      kristaps 2701:
1.77      kristaps 2702:        /*
1.173     schwarze 2703:         * Anything else following the conditional causes
                   2704:         * single-line scope.  Warn if the scope contains
                   2705:         * nothing but trailing whitespace.
1.77      kristaps 2706:         */
                   2707:
1.238     schwarze 2708:        if (buf->buf[pos] == '\0')
1.350     schwarze 2709:                mandoc_msg(MANDOCERR_COND_EMPTY,
                   2710:                    ln, ppos, "%s", roff_name[tok]);
1.77      kristaps 2711:
1.173     schwarze 2712:        r->last->endspan = 1;
1.74      kristaps 2713:
1.173     schwarze 2714: out:
1.75      kristaps 2715:        *offs = pos;
1.340     schwarze 2716:        irc = ROFF_RERUN;
                   2717:        if (tok == ROFF_while)
                   2718:                irc |= ROFF_WHILE;
                   2719:        return irc;
1.83      schwarze 2720: }
                   2721:
1.340     schwarze 2722: static int
1.92      schwarze 2723: roff_ds(ROFF_ARGS)
                   2724: {
1.212     schwarze 2725:        char            *string;
                   2726:        const char      *name;
                   2727:        size_t           namesz;
1.96      kristaps 2728:
1.251     schwarze 2729:        /* Ignore groff compatibility mode for now. */
                   2730:
                   2731:        if (tok == ROFF_ds1)
                   2732:                tok = ROFF_ds;
                   2733:        else if (tok == ROFF_as1)
                   2734:                tok = ROFF_as;
                   2735:
1.96      kristaps 2736:        /*
1.212     schwarze 2737:         * The first word is the name of the string.
                   2738:         * If it is empty or terminated by an escape sequence,
                   2739:         * abort the `ds' request without defining anything.
1.96      kristaps 2740:         */
1.92      schwarze 2741:
1.238     schwarze 2742:        name = string = buf->buf + pos;
                   2743:        if (*name == '\0')
1.277     schwarze 2744:                return ROFF_IGN;
1.92      schwarze 2745:
1.212     schwarze 2746:        namesz = roff_getname(r, &string, ln, pos);
1.363     schwarze 2747:        switch (name[namesz]) {
                   2748:        case '\\':
1.277     schwarze 2749:                return ROFF_IGN;
1.363     schwarze 2750:        case '\t':
                   2751:                string = buf->buf + pos + namesz;
                   2752:                break;
                   2753:        default:
                   2754:                break;
                   2755:        }
1.212     schwarze 2756:
                   2757:        /* Read past the initial double-quote, if any. */
1.238     schwarze 2758:        if (*string == '"')
1.92      schwarze 2759:                string++;
                   2760:
1.96      kristaps 2761:        /* The rest is the value. */
1.212     schwarze 2762:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2763:            ROFF_as == tok);
1.311     schwarze 2764:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277     schwarze 2765:        return ROFF_IGN;
1.92      schwarze 2766: }
                   2767:
1.204     schwarze 2768: /*
                   2769:  * Parse a single operator, one or two characters long.
                   2770:  * If the operator is recognized, return success and advance the
                   2771:  * parse point, else return failure and let the parse point unchanged.
                   2772:  */
                   2773: static int
                   2774: roff_getop(const char *v, int *pos, char *res)
                   2775: {
                   2776:
                   2777:        *res = v[*pos];
                   2778:
                   2779:        switch (*res) {
1.207     schwarze 2780:        case '+':
                   2781:        case '-':
                   2782:        case '*':
                   2783:        case '/':
                   2784:        case '%':
                   2785:        case '&':
                   2786:        case ':':
1.204     schwarze 2787:                break;
                   2788:        case '<':
                   2789:                switch (v[*pos + 1]) {
1.207     schwarze 2790:                case '=':
1.204     schwarze 2791:                        *res = 'l';
                   2792:                        (*pos)++;
                   2793:                        break;
1.207     schwarze 2794:                case '>':
1.204     schwarze 2795:                        *res = '!';
                   2796:                        (*pos)++;
                   2797:                        break;
1.207     schwarze 2798:                case '?':
1.204     schwarze 2799:                        *res = 'i';
                   2800:                        (*pos)++;
                   2801:                        break;
                   2802:                default:
                   2803:                        break;
                   2804:                }
                   2805:                break;
                   2806:        case '>':
                   2807:                switch (v[*pos + 1]) {
1.207     schwarze 2808:                case '=':
1.204     schwarze 2809:                        *res = 'g';
                   2810:                        (*pos)++;
                   2811:                        break;
1.207     schwarze 2812:                case '?':
1.204     schwarze 2813:                        *res = 'a';
                   2814:                        (*pos)++;
                   2815:                        break;
                   2816:                default:
                   2817:                        break;
                   2818:                }
                   2819:                break;
                   2820:        case '=':
                   2821:                if ('=' == v[*pos + 1])
                   2822:                        (*pos)++;
                   2823:                break;
                   2824:        default:
1.277     schwarze 2825:                return 0;
1.204     schwarze 2826:        }
                   2827:        (*pos)++;
                   2828:
1.277     schwarze 2829:        return *res;
1.204     schwarze 2830: }
                   2831:
                   2832: /*
                   2833:  * Evaluate either a parenthesized numeric expression
                   2834:  * or a single signed integer number.
                   2835:  */
                   2836: static int
1.235     schwarze 2837: roff_evalpar(struct roff *r, int ln,
1.261     schwarze 2838:        const char *v, int *pos, int *res, int flags)
1.204     schwarze 2839: {
                   2840:
                   2841:        if ('(' != v[*pos])
1.277     schwarze 2842:                return roff_getnum(v, pos, res, flags);
1.204     schwarze 2843:
                   2844:        (*pos)++;
1.261     schwarze 2845:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277     schwarze 2846:                return 0;
1.204     schwarze 2847:
1.206     schwarze 2848:        /*
                   2849:         * Omission of the closing parenthesis
                   2850:         * is an error in validation mode,
                   2851:         * but ignored in evaluation mode.
                   2852:         */
                   2853:
1.204     schwarze 2854:        if (')' == v[*pos])
                   2855:                (*pos)++;
1.206     schwarze 2856:        else if (NULL == res)
1.277     schwarze 2857:                return 0;
1.204     schwarze 2858:
1.277     schwarze 2859:        return 1;
1.204     schwarze 2860: }
                   2861:
                   2862: /*
                   2863:  * Evaluate a complete numeric expression.
                   2864:  * Proceed left to right, there is no concept of precedence.
                   2865:  */
                   2866: static int
1.235     schwarze 2867: roff_evalnum(struct roff *r, int ln, const char *v,
1.261     schwarze 2868:        int *pos, int *res, int flags)
1.204     schwarze 2869: {
                   2870:        int              mypos, operand2;
                   2871:        char             operator;
                   2872:
                   2873:        if (NULL == pos) {
                   2874:                mypos = 0;
                   2875:                pos = &mypos;
                   2876:        }
                   2877:
1.261     schwarze 2878:        if (flags & ROFFNUM_WHITE)
1.204     schwarze 2879:                while (isspace((unsigned char)v[*pos]))
                   2880:                        (*pos)++;
                   2881:
1.261     schwarze 2882:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277     schwarze 2883:                return 0;
1.204     schwarze 2884:
                   2885:        while (1) {
1.261     schwarze 2886:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2887:                        while (isspace((unsigned char)v[*pos]))
                   2888:                                (*pos)++;
                   2889:
                   2890:                if ( ! roff_getop(v, pos, &operator))
                   2891:                        break;
                   2892:
1.261     schwarze 2893:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2894:                        while (isspace((unsigned char)v[*pos]))
                   2895:                                (*pos)++;
                   2896:
1.261     schwarze 2897:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277     schwarze 2898:                        return 0;
1.204     schwarze 2899:
1.261     schwarze 2900:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2901:                        while (isspace((unsigned char)v[*pos]))
                   2902:                                (*pos)++;
1.206     schwarze 2903:
                   2904:                if (NULL == res)
                   2905:                        continue;
1.204     schwarze 2906:
                   2907:                switch (operator) {
1.207     schwarze 2908:                case '+':
1.204     schwarze 2909:                        *res += operand2;
                   2910:                        break;
1.207     schwarze 2911:                case '-':
1.204     schwarze 2912:                        *res -= operand2;
                   2913:                        break;
1.207     schwarze 2914:                case '*':
1.204     schwarze 2915:                        *res *= operand2;
                   2916:                        break;
1.207     schwarze 2917:                case '/':
1.244     schwarze 2918:                        if (operand2 == 0) {
1.235     schwarze 2919:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2920:                                        ln, *pos, "%s", v);
1.234     kristaps 2921:                                *res = 0;
                   2922:                                break;
                   2923:                        }
1.204     schwarze 2924:                        *res /= operand2;
                   2925:                        break;
1.207     schwarze 2926:                case '%':
1.244     schwarze 2927:                        if (operand2 == 0) {
                   2928:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2929:                                        ln, *pos, "%s", v);
1.244     schwarze 2930:                                *res = 0;
                   2931:                                break;
                   2932:                        }
1.204     schwarze 2933:                        *res %= operand2;
                   2934:                        break;
1.207     schwarze 2935:                case '<':
1.204     schwarze 2936:                        *res = *res < operand2;
                   2937:                        break;
1.207     schwarze 2938:                case '>':
1.204     schwarze 2939:                        *res = *res > operand2;
                   2940:                        break;
1.207     schwarze 2941:                case 'l':
1.204     schwarze 2942:                        *res = *res <= operand2;
                   2943:                        break;
1.207     schwarze 2944:                case 'g':
1.204     schwarze 2945:                        *res = *res >= operand2;
                   2946:                        break;
1.207     schwarze 2947:                case '=':
1.204     schwarze 2948:                        *res = *res == operand2;
                   2949:                        break;
1.207     schwarze 2950:                case '!':
1.204     schwarze 2951:                        *res = *res != operand2;
                   2952:                        break;
1.207     schwarze 2953:                case '&':
1.204     schwarze 2954:                        *res = *res && operand2;
                   2955:                        break;
1.207     schwarze 2956:                case ':':
1.204     schwarze 2957:                        *res = *res || operand2;
                   2958:                        break;
1.207     schwarze 2959:                case 'i':
1.204     schwarze 2960:                        if (operand2 < *res)
                   2961:                                *res = operand2;
                   2962:                        break;
1.207     schwarze 2963:                case 'a':
1.204     schwarze 2964:                        if (operand2 > *res)
                   2965:                                *res = operand2;
                   2966:                        break;
                   2967:                default:
                   2968:                        abort();
                   2969:                }
                   2970:        }
1.277     schwarze 2971:        return 1;
1.204     schwarze 2972: }
                   2973:
1.266     schwarze 2974: /* --- register management ------------------------------------------------ */
                   2975:
1.180     schwarze 2976: void
1.187     schwarze 2977: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147     kristaps 2978: {
1.327     schwarze 2979:        roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.326     schwarze 2980: }
                   2981:
                   2982: static void
                   2983: roff_setregn(struct roff *r, const char *name, size_t len,
1.327     schwarze 2984:     int val, char sign, int step)
1.326     schwarze 2985: {
1.180     schwarze 2986:        struct roffreg  *reg;
                   2987:
                   2988:        /* Search for an existing register with the same name. */
                   2989:        reg = r->regtab;
                   2990:
1.326     schwarze 2991:        while (reg != NULL && (reg->key.sz != len ||
                   2992:            strncmp(reg->key.p, name, len) != 0))
1.180     schwarze 2993:                reg = reg->next;
1.147     kristaps 2994:
1.180     schwarze 2995:        if (NULL == reg) {
                   2996:                /* Create a new register. */
                   2997:                reg = mandoc_malloc(sizeof(struct roffreg));
1.326     schwarze 2998:                reg->key.p = mandoc_strndup(name, len);
                   2999:                reg->key.sz = len;
1.187     schwarze 3000:                reg->val = 0;
1.327     schwarze 3001:                reg->step = 0;
1.180     schwarze 3002:                reg->next = r->regtab;
                   3003:                r->regtab = reg;
                   3004:        }
                   3005:
1.187     schwarze 3006:        if ('+' == sign)
                   3007:                reg->val += val;
                   3008:        else if ('-' == sign)
                   3009:                reg->val -= val;
                   3010:        else
                   3011:                reg->val = val;
1.327     schwarze 3012:        if (step != INT_MIN)
                   3013:                reg->step = step;
1.147     kristaps 3014: }
                   3015:
1.192     schwarze 3016: /*
                   3017:  * Handle some predefined read-only number registers.
                   3018:  * For now, return -1 if the requested register is not predefined;
                   3019:  * in case a predefined read-only register having the value -1
                   3020:  * were to turn up, another special value would have to be chosen.
                   3021:  */
                   3022: static int
1.273     schwarze 3023: roff_getregro(const struct roff *r, const char *name)
1.192     schwarze 3024: {
                   3025:
                   3026:        switch (*name) {
1.273     schwarze 3027:        case '$':  /* Number of arguments of the last macro evaluated. */
1.339     schwarze 3028:                return r->mstackpos < 0 ? 0 : r->mstack[r->mstackpos].argc;
1.207     schwarze 3029:        case 'A':  /* ASCII approximation mode is always off. */
1.277     schwarze 3030:                return 0;
1.207     schwarze 3031:        case 'g':  /* Groff compatibility mode is always on. */
1.277     schwarze 3032:                return 1;
1.207     schwarze 3033:        case 'H':  /* Fixed horizontal resolution. */
1.277     schwarze 3034:                return 24;
1.207     schwarze 3035:        case 'j':  /* Always adjust left margin only. */
1.277     schwarze 3036:                return 0;
1.207     schwarze 3037:        case 'T':  /* Some output device is always defined. */
1.277     schwarze 3038:                return 1;
1.207     schwarze 3039:        case 'V':  /* Fixed vertical resolution. */
1.277     schwarze 3040:                return 40;
1.192     schwarze 3041:        default:
1.277     schwarze 3042:                return -1;
1.192     schwarze 3043:        }
                   3044: }
                   3045:
1.181     schwarze 3046: int
1.326     schwarze 3047: roff_getreg(struct roff *r, const char *name)
1.147     kristaps 3048: {
1.327     schwarze 3049:        return roff_getregn(r, name, strlen(name), '\0');
1.181     schwarze 3050: }
                   3051:
                   3052: static int
1.327     schwarze 3053: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.181     schwarze 3054: {
                   3055:        struct roffreg  *reg;
1.192     schwarze 3056:        int              val;
                   3057:
                   3058:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3059:                val = roff_getregro(r, name + 1);
1.192     schwarze 3060:                if (-1 != val)
1.277     schwarze 3061:                        return val;
1.192     schwarze 3062:        }
1.181     schwarze 3063:
1.327     schwarze 3064:        for (reg = r->regtab; reg; reg = reg->next) {
1.181     schwarze 3065:                if (len == reg->key.sz &&
1.327     schwarze 3066:                    0 == strncmp(name, reg->key.p, len)) {
                   3067:                        switch (sign) {
                   3068:                        case '+':
                   3069:                                reg->val += reg->step;
                   3070:                                break;
                   3071:                        case '-':
                   3072:                                reg->val -= reg->step;
                   3073:                                break;
                   3074:                        default:
                   3075:                                break;
                   3076:                        }
1.277     schwarze 3077:                        return reg->val;
1.327     schwarze 3078:                }
                   3079:        }
1.271     schwarze 3080:
1.327     schwarze 3081:        roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.277     schwarze 3082:        return 0;
1.271     schwarze 3083: }
                   3084:
                   3085: static int
                   3086: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   3087: {
                   3088:        struct roffreg  *reg;
                   3089:        int              val;
                   3090:
                   3091:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3092:                val = roff_getregro(r, name + 1);
1.271     schwarze 3093:                if (-1 != val)
1.277     schwarze 3094:                        return 1;
1.271     schwarze 3095:        }
                   3096:
                   3097:        for (reg = r->regtab; reg; reg = reg->next)
                   3098:                if (len == reg->key.sz &&
                   3099:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 3100:                        return 1;
1.147     kristaps 3101:
1.277     schwarze 3102:        return 0;
1.147     kristaps 3103: }
                   3104:
1.180     schwarze 3105: static void
                   3106: roff_freereg(struct roffreg *reg)
1.147     kristaps 3107: {
1.180     schwarze 3108:        struct roffreg  *old_reg;
1.147     kristaps 3109:
1.180     schwarze 3110:        while (NULL != reg) {
                   3111:                free(reg->key.p);
                   3112:                old_reg = reg;
                   3113:                reg = reg->next;
                   3114:                free(old_reg);
                   3115:        }
1.147     kristaps 3116: }
1.92      schwarze 3117:
1.340     schwarze 3118: static int
1.89      kristaps 3119: roff_nr(ROFF_ARGS)
1.83      schwarze 3120: {
1.327     schwarze 3121:        char            *key, *val, *step;
1.212     schwarze 3122:        size_t           keysz;
1.327     schwarze 3123:        int              iv, is, len;
1.187     schwarze 3124:        char             sign;
1.89      kristaps 3125:
1.238     schwarze 3126:        key = val = buf->buf + pos;
                   3127:        if (*key == '\0')
1.277     schwarze 3128:                return ROFF_IGN;
1.212     schwarze 3129:
                   3130:        keysz = roff_getname(r, &val, ln, pos);
1.363     schwarze 3131:        if (key[keysz] == '\\' || key[keysz] == '\t')
1.277     schwarze 3132:                return ROFF_IGN;
1.89      kristaps 3133:
1.187     schwarze 3134:        sign = *val;
1.238     schwarze 3135:        if (sign == '+' || sign == '-')
1.187     schwarze 3136:                val++;
                   3137:
1.327     schwarze 3138:        len = 0;
                   3139:        if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
                   3140:                return ROFF_IGN;
                   3141:
                   3142:        step = val + len;
                   3143:        while (isspace((unsigned char)*step))
                   3144:                step++;
                   3145:        if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
                   3146:                is = INT_MIN;
1.109     kristaps 3147:
1.327     schwarze 3148:        roff_setregn(r, key, keysz, iv, sign, is);
1.277     schwarze 3149:        return ROFF_IGN;
1.203     schwarze 3150: }
                   3151:
1.340     schwarze 3152: static int
1.203     schwarze 3153: roff_rr(ROFF_ARGS)
                   3154: {
                   3155:        struct roffreg  *reg, **prev;
1.212     schwarze 3156:        char            *name, *cp;
                   3157:        size_t           namesz;
1.203     schwarze 3158:
1.238     schwarze 3159:        name = cp = buf->buf + pos;
                   3160:        if (*name == '\0')
1.277     schwarze 3161:                return ROFF_IGN;
1.212     schwarze 3162:        namesz = roff_getname(r, &cp, ln, pos);
                   3163:        name[namesz] = '\0';
1.203     schwarze 3164:
                   3165:        prev = &r->regtab;
                   3166:        while (1) {
                   3167:                reg = *prev;
1.238     schwarze 3168:                if (reg == NULL || !strcmp(name, reg->key.p))
1.203     schwarze 3169:                        break;
                   3170:                prev = &reg->next;
                   3171:        }
1.238     schwarze 3172:        if (reg != NULL) {
1.203     schwarze 3173:                *prev = reg->next;
                   3174:                free(reg->key.p);
                   3175:                free(reg);
                   3176:        }
1.277     schwarze 3177:        return ROFF_IGN;
1.122     schwarze 3178: }
                   3179:
1.266     schwarze 3180: /* --- handler functions for roff requests -------------------------------- */
                   3181:
1.340     schwarze 3182: static int
1.122     schwarze 3183: roff_rm(ROFF_ARGS)
                   3184: {
                   3185:        const char       *name;
                   3186:        char             *cp;
1.212     schwarze 3187:        size_t            namesz;
1.122     schwarze 3188:
1.238     schwarze 3189:        cp = buf->buf + pos;
                   3190:        while (*cp != '\0') {
1.212     schwarze 3191:                name = cp;
1.238     schwarze 3192:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212     schwarze 3193:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311     schwarze 3194:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.363     schwarze 3195:                if (name[namesz] == '\\' || name[namesz] == '\t')
1.212     schwarze 3196:                        break;
1.122     schwarze 3197:        }
1.277     schwarze 3198:        return ROFF_IGN;
1.178     schwarze 3199: }
                   3200:
1.340     schwarze 3201: static int
1.178     schwarze 3202: roff_it(ROFF_ARGS)
                   3203: {
                   3204:        int              iv;
                   3205:
                   3206:        /* Parse the number of lines. */
1.261     schwarze 3207:
                   3208:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.350     schwarze 3209:                mandoc_msg(MANDOCERR_IT_NONUM,
                   3210:                    ln, ppos, "%s", buf->buf + 1);
1.277     schwarze 3211:                return ROFF_IGN;
1.178     schwarze 3212:        }
                   3213:
1.262     schwarze 3214:        while (isspace((unsigned char)buf->buf[pos]))
                   3215:                pos++;
                   3216:
                   3217:        /*
                   3218:         * Arm the input line trap.
                   3219:         * Special-casing "an-trap" is an ugly workaround to cope
                   3220:         * with DocBook stupidly fiddling with man(7) internals.
                   3221:         */
1.261     schwarze 3222:
1.178     schwarze 3223:        roffit_lines = iv;
1.262     schwarze 3224:        roffit_macro = mandoc_strdup(iv != 1 ||
                   3225:            strcmp(buf->buf + pos, "an-trap") ?
                   3226:            buf->buf + pos : "br");
1.277     schwarze 3227:        return ROFF_IGN;
1.175     schwarze 3228: }
                   3229:
1.340     schwarze 3230: static int
1.175     schwarze 3231: roff_Dd(ROFF_ARGS)
                   3232: {
1.315     schwarze 3233:        int              mask;
                   3234:        enum roff_tok    t, te;
1.227     schwarze 3235:
1.315     schwarze 3236:        switch (tok) {
                   3237:        case ROFF_Dd:
                   3238:                tok = MDOC_Dd;
                   3239:                te = MDOC_MAX;
                   3240:                if (r->format == 0)
                   3241:                        r->format = MPARSE_MDOC;
                   3242:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   3243:                break;
                   3244:        case ROFF_TH:
                   3245:                tok = MAN_TH;
                   3246:                te = MAN_MAX;
                   3247:                if (r->format == 0)
                   3248:                        r->format = MPARSE_MAN;
                   3249:                mask = MPARSE_QUICK;
                   3250:                break;
                   3251:        default:
                   3252:                abort();
                   3253:        }
                   3254:        if ((r->options & mask) == 0)
                   3255:                for (t = tok; t < te; t++)
                   3256:                        roff_setstr(r, roff_name[t], NULL, 0);
1.277     schwarze 3257:        return ROFF_CONT;
1.109     kristaps 3258: }
                   3259:
1.340     schwarze 3260: static int
1.109     kristaps 3261: roff_TE(ROFF_ARGS)
                   3262: {
1.361     schwarze 3263:        r->man->flags &= ~ROFF_NONOFILL;
1.321     schwarze 3264:        if (r->tbl == NULL) {
1.350     schwarze 3265:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "TE");
1.321     schwarze 3266:                return ROFF_IGN;
                   3267:        }
1.346     schwarze 3268:        if (tbl_end(r->tbl, 0) == 0) {
1.321     schwarze 3269:                r->tbl = NULL;
1.258     schwarze 3270:                free(buf->buf);
                   3271:                buf->buf = mandoc_strdup(".sp");
                   3272:                buf->sz = 4;
1.329     schwarze 3273:                *offs = 0;
1.277     schwarze 3274:                return ROFF_REPARSE;
1.258     schwarze 3275:        }
1.321     schwarze 3276:        r->tbl = NULL;
1.277     schwarze 3277:        return ROFF_IGN;
1.112     kristaps 3278: }
                   3279:
1.340     schwarze 3280: static int
1.112     kristaps 3281: roff_T_(ROFF_ARGS)
                   3282: {
                   3283:
                   3284:        if (NULL == r->tbl)
1.350     schwarze 3285:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "T&");
1.112     kristaps 3286:        else
1.296     schwarze 3287:                tbl_restart(ln, ppos, r->tbl);
1.112     kristaps 3288:
1.277     schwarze 3289:        return ROFF_IGN;
1.109     kristaps 3290: }
                   3291:
1.230     schwarze 3292: /*
                   3293:  * Handle in-line equation delimiters.
                   3294:  */
1.340     schwarze 3295: static int
1.238     schwarze 3296: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151     kristaps 3297: {
1.233     schwarze 3298:        char            *cp1, *cp2;
                   3299:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151     kristaps 3300:
1.230     schwarze 3301:        /*
                   3302:         * Outside equations, look for an opening delimiter.
                   3303:         * If we are inside an equation, we already know it is
                   3304:         * in-line, or this function wouldn't have been called;
                   3305:         * so look for a closing delimiter.
                   3306:         */
                   3307:
1.238     schwarze 3308:        cp1 = buf->buf + pos;
1.230     schwarze 3309:        cp2 = strchr(cp1, r->eqn == NULL ?
                   3310:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   3311:        if (cp2 == NULL)
1.277     schwarze 3312:                return ROFF_CONT;
1.230     schwarze 3313:
1.233     schwarze 3314:        *cp2++ = '\0';
                   3315:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   3316:
                   3317:        /* Handle preceding text, protecting whitespace. */
                   3318:
1.238     schwarze 3319:        if (*buf->buf != '\0') {
1.233     schwarze 3320:                if (r->eqn == NULL)
                   3321:                        bef_pr = "\\&";
                   3322:                bef_nl = "\n";
                   3323:        }
                   3324:
                   3325:        /*
                   3326:         * Prepare replacing the delimiter with an equation macro
                   3327:         * and drop leading white space from the equation.
                   3328:         */
1.230     schwarze 3329:
1.233     schwarze 3330:        if (r->eqn == NULL) {
                   3331:                while (*cp2 == ' ')
                   3332:                        cp2++;
                   3333:                mac = ".EQ";
                   3334:        } else
                   3335:                mac = ".EN";
                   3336:
                   3337:        /* Handle following text, protecting whitespace. */
                   3338:
                   3339:        if (*cp2 != '\0') {
                   3340:                aft_nl = "\n";
                   3341:                if (r->eqn != NULL)
                   3342:                        aft_pr = "\\&";
                   3343:        }
                   3344:
                   3345:        /* Do the actual replacement. */
                   3346:
1.238     schwarze 3347:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233     schwarze 3348:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238     schwarze 3349:        free(buf->buf);
                   3350:        buf->buf = cp1;
1.230     schwarze 3351:
                   3352:        /* Toggle the in-line state of the eqn subsystem. */
                   3353:
                   3354:        r->eqn_inline = r->eqn == NULL;
1.277     schwarze 3355:        return ROFF_REPARSE;
1.151     kristaps 3356: }
                   3357:
1.340     schwarze 3358: static int
1.235     schwarze 3359: roff_EQ(ROFF_ARGS)
1.125     kristaps 3360: {
1.319     schwarze 3361:        struct roff_node        *n;
                   3362:
1.356     schwarze 3363:        if (r->man->meta.macroset == MACROSET_MAN)
1.322     schwarze 3364:                man_breakscope(r->man, ROFF_EQ);
1.319     schwarze 3365:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   3366:        if (ln > r->man->last->line)
                   3367:                n->flags |= NODE_LINE;
1.348     schwarze 3368:        n->eqn = eqn_box_new();
1.319     schwarze 3369:        roff_node_append(r->man, n);
                   3370:        r->man->next = ROFF_NEXT_SIBLING;
1.125     kristaps 3371:
1.238     schwarze 3372:        assert(r->eqn == NULL);
1.319     schwarze 3373:        if (r->last_eqn == NULL)
1.351     schwarze 3374:                r->last_eqn = eqn_alloc();
1.319     schwarze 3375:        else
                   3376:                eqn_reset(r->last_eqn);
                   3377:        r->eqn = r->last_eqn;
                   3378:        r->eqn->node = n;
1.151     kristaps 3379:
1.238     schwarze 3380:        if (buf->buf[pos] != '\0')
1.350     schwarze 3381:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 3382:                    ".EQ %s", buf->buf + pos);
1.151     kristaps 3383:
1.277     schwarze 3384:        return ROFF_IGN;
1.125     kristaps 3385: }
                   3386:
1.340     schwarze 3387: static int
1.125     kristaps 3388: roff_EN(ROFF_ARGS)
                   3389: {
1.319     schwarze 3390:        if (r->eqn != NULL) {
                   3391:                eqn_parse(r->eqn);
                   3392:                r->eqn = NULL;
                   3393:        } else
1.350     schwarze 3394:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "EN");
1.319     schwarze 3395:        if (buf->buf[pos] != '\0')
1.350     schwarze 3396:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.319     schwarze 3397:                    "EN %s", buf->buf + pos);
1.277     schwarze 3398:        return ROFF_IGN;
1.125     kristaps 3399: }
                   3400:
1.340     schwarze 3401: static int
1.109     kristaps 3402: roff_TS(ROFF_ARGS)
                   3403: {
1.321     schwarze 3404:        if (r->tbl != NULL) {
1.350     schwarze 3405:                mandoc_msg(MANDOCERR_BLK_BROKEN, ln, ppos, "TS breaks TS");
1.346     schwarze 3406:                tbl_end(r->tbl, 0);
1.115     kristaps 3407:        }
1.361     schwarze 3408:        r->man->flags |= ROFF_NONOFILL;
1.351     schwarze 3409:        r->tbl = tbl_alloc(ppos, ln, r->last_tbl);
1.346     schwarze 3410:        if (r->last_tbl == NULL)
1.321     schwarze 3411:                r->first_tbl = r->tbl;
                   3412:        r->last_tbl = r->tbl;
1.297     schwarze 3413:        return ROFF_IGN;
                   3414: }
                   3415:
1.340     schwarze 3416: static int
1.358     schwarze 3417: roff_noarg(ROFF_ARGS)
                   3418: {
                   3419:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3420:                man_breakscope(r->man, tok);
                   3421:        if (tok == ROFF_brp)
                   3422:                tok = ROFF_br;
                   3423:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3424:        if (buf->buf[pos] != '\0')
                   3425:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
                   3426:                   "%s %s", roff_name[tok], buf->buf + pos);
                   3427:        if (tok == ROFF_nf)
                   3428:                r->man->flags |= ROFF_NOFILL;
                   3429:        else if (tok == ROFF_fi)
                   3430:                r->man->flags &= ~ROFF_NOFILL;
                   3431:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3432:        r->man->next = ROFF_NEXT_SIBLING;
                   3433:        return ROFF_IGN;
                   3434: }
                   3435:
                   3436: static int
1.297     schwarze 3437: roff_onearg(ROFF_ARGS)
                   3438: {
                   3439:        struct roff_node        *n;
                   3440:        char                    *cp;
1.305     schwarze 3441:        int                      npos;
1.297     schwarze 3442:
1.302     schwarze 3443:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320     schwarze 3444:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
                   3445:             tok == ROFF_ti))
1.302     schwarze 3446:                man_breakscope(r->man, tok);
                   3447:
1.309     schwarze 3448:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305     schwarze 3449:                r->man->last = roffce_node;
                   3450:                r->man->next = ROFF_NEXT_SIBLING;
                   3451:        }
                   3452:
1.297     schwarze 3453:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3454:        n = r->man->last;
                   3455:
                   3456:        cp = buf->buf + pos;
                   3457:        if (*cp != '\0') {
                   3458:                while (*cp != '\0' && *cp != ' ')
                   3459:                        cp++;
                   3460:                while (*cp == ' ')
                   3461:                        *cp++ = '\0';
                   3462:                if (*cp != '\0')
1.350     schwarze 3463:                        mandoc_msg(MANDOCERR_ARG_EXCESS,
                   3464:                            ln, (int)(cp - buf->buf),
1.297     schwarze 3465:                            "%s ... %s", roff_name[tok], cp);
                   3466:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300     schwarze 3467:        }
                   3468:
1.309     schwarze 3469:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   3470:                if (r->man->last->type == ROFFT_ELEM) {
1.305     schwarze 3471:                        roff_word_alloc(r->man, ln, pos, "1");
                   3472:                        r->man->last->flags |= NODE_NOSRC;
                   3473:                }
                   3474:                npos = 0;
                   3475:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   3476:                    &roffce_lines, 0) == 0) {
1.350     schwarze 3477:                        mandoc_msg(MANDOCERR_CE_NONUM,
                   3478:                            ln, pos, "ce %s", buf->buf + pos);
1.305     schwarze 3479:                        roffce_lines = 1;
                   3480:                }
                   3481:                if (roffce_lines < 1) {
                   3482:                        r->man->last = r->man->last->parent;
                   3483:                        roffce_node = NULL;
                   3484:                        roffce_lines = 0;
                   3485:                } else
                   3486:                        roffce_node = r->man->last->parent;
                   3487:        } else {
                   3488:                n->flags |= NODE_VALID | NODE_ENDED;
                   3489:                r->man->last = n;
                   3490:        }
                   3491:        n->flags |= NODE_LINE;
1.300     schwarze 3492:        r->man->next = ROFF_NEXT_SIBLING;
                   3493:        return ROFF_IGN;
                   3494: }
                   3495:
1.340     schwarze 3496: static int
1.300     schwarze 3497: roff_manyarg(ROFF_ARGS)
                   3498: {
                   3499:        struct roff_node        *n;
                   3500:        char                    *sp, *ep;
                   3501:
                   3502:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3503:        n = r->man->last;
                   3504:
                   3505:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   3506:                while (*ep != '\0' && *ep != ' ')
                   3507:                        ep++;
                   3508:                while (*ep == ' ')
                   3509:                        *ep++ = '\0';
                   3510:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297     schwarze 3511:        }
                   3512:
                   3513:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3514:        r->man->last = n;
                   3515:        r->man->next = ROFF_NEXT_SIBLING;
1.277     schwarze 3516:        return ROFF_IGN;
1.251     schwarze 3517: }
                   3518:
1.340     schwarze 3519: static int
1.311     schwarze 3520: roff_als(ROFF_ARGS)
                   3521: {
                   3522:        char            *oldn, *newn, *end, *value;
                   3523:        size_t           oldsz, newsz, valsz;
                   3524:
                   3525:        newn = oldn = buf->buf + pos;
                   3526:        if (*newn == '\0')
                   3527:                return ROFF_IGN;
                   3528:
                   3529:        newsz = roff_getname(r, &oldn, ln, pos);
1.363     schwarze 3530:        if (newn[newsz] == '\\' || newn[newsz] == '\t' || *oldn == '\0')
1.311     schwarze 3531:                return ROFF_IGN;
                   3532:
                   3533:        end = oldn;
                   3534:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3535:        if (oldsz == 0)
                   3536:                return ROFF_IGN;
                   3537:
1.338     schwarze 3538:        valsz = mandoc_asprintf(&value, ".%.*s \\$@\\\"\n",
1.315     schwarze 3539:            (int)oldsz, oldn);
1.311     schwarze 3540:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3541:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3542:        free(value);
1.364     schwarze 3543:        return ROFF_IGN;
                   3544: }
                   3545:
                   3546: /*
                   3547:  * The .break request only makes sense inside conditionals,
                   3548:  * and that case is already handled in roff_cond_sub().
                   3549:  */
                   3550: static int
                   3551: roff_break(ROFF_ARGS)
                   3552: {
                   3553:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, pos, "break");
1.296     schwarze 3554:        return ROFF_IGN;
1.92      schwarze 3555: }
                   3556:
1.340     schwarze 3557: static int
1.174     kristaps 3558: roff_cc(ROFF_ARGS)
                   3559: {
                   3560:        const char      *p;
                   3561:
1.238     schwarze 3562:        p = buf->buf + pos;
1.174     kristaps 3563:
1.238     schwarze 3564:        if (*p == '\0' || (r->control = *p++) == '.')
1.303     schwarze 3565:                r->control = '\0';
1.174     kristaps 3566:
1.238     schwarze 3567:        if (*p != '\0')
1.350     schwarze 3568:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.260     schwarze 3569:                    ln, p - buf->buf, "cc ... %s", p);
1.174     kristaps 3570:
1.341     schwarze 3571:        return ROFF_IGN;
                   3572: }
                   3573:
                   3574: static int
                   3575: roff_char(ROFF_ARGS)
                   3576: {
                   3577:        const char      *p, *kp, *vp;
                   3578:        size_t           ksz, vsz;
                   3579:        int              font;
                   3580:
                   3581:        /* Parse the character to be replaced. */
                   3582:
                   3583:        kp = buf->buf + pos;
                   3584:        p = kp + 1;
                   3585:        if (*kp == '\0' || (*kp == '\\' &&
                   3586:             mandoc_escape(&p, NULL, NULL) != ESCAPE_SPECIAL) ||
                   3587:            (*p != ' ' && *p != '\0')) {
1.350     schwarze 3588:                mandoc_msg(MANDOCERR_CHAR_ARG, ln, pos, "char %s", kp);
1.341     schwarze 3589:                return ROFF_IGN;
                   3590:        }
                   3591:        ksz = p - kp;
                   3592:        while (*p == ' ')
                   3593:                p++;
                   3594:
                   3595:        /*
                   3596:         * If the replacement string contains a font escape sequence,
                   3597:         * we have to restore the font at the end.
                   3598:         */
                   3599:
                   3600:        vp = p;
                   3601:        vsz = strlen(p);
                   3602:        font = 0;
                   3603:        while (*p != '\0') {
                   3604:                if (*p++ != '\\')
                   3605:                        continue;
                   3606:                switch (mandoc_escape(&p, NULL, NULL)) {
                   3607:                case ESCAPE_FONT:
                   3608:                case ESCAPE_FONTROMAN:
                   3609:                case ESCAPE_FONTITALIC:
                   3610:                case ESCAPE_FONTBOLD:
                   3611:                case ESCAPE_FONTBI:
1.378     schwarze 3612:                case ESCAPE_FONTCR:
                   3613:                case ESCAPE_FONTCB:
                   3614:                case ESCAPE_FONTCI:
1.341     schwarze 3615:                case ESCAPE_FONTPREV:
                   3616:                        font++;
                   3617:                        break;
                   3618:                default:
                   3619:                        break;
                   3620:                }
                   3621:        }
                   3622:        if (font > 1)
1.350     schwarze 3623:                mandoc_msg(MANDOCERR_CHAR_FONT,
                   3624:                    ln, (int)(vp - buf->buf), "%s", vp);
1.341     schwarze 3625:
                   3626:        /*
                   3627:         * Approximate the effect of .char using the .tr tables.
                   3628:         * XXX In groff, .char and .tr interact differently.
                   3629:         */
                   3630:
                   3631:        if (ksz == 1) {
                   3632:                if (r->xtab == NULL)
                   3633:                        r->xtab = mandoc_calloc(128, sizeof(*r->xtab));
                   3634:                assert((unsigned int)*kp < 128);
                   3635:                free(r->xtab[(int)*kp].p);
                   3636:                r->xtab[(int)*kp].sz = mandoc_asprintf(&r->xtab[(int)*kp].p,
                   3637:                    "%s%s", vp, font ? "\fP" : "");
                   3638:        } else {
                   3639:                roff_setstrn(&r->xmbtab, kp, ksz, vp, vsz, 0);
                   3640:                if (font)
                   3641:                        roff_setstrn(&r->xmbtab, kp, ksz, "\\fP", 3, 1);
                   3642:        }
1.277     schwarze 3643:        return ROFF_IGN;
1.174     kristaps 3644: }
                   3645:
1.340     schwarze 3646: static int
1.303     schwarze 3647: roff_ec(ROFF_ARGS)
                   3648: {
                   3649:        const char      *p;
                   3650:
                   3651:        p = buf->buf + pos;
                   3652:        if (*p == '\0')
                   3653:                r->escape = '\\';
                   3654:        else {
                   3655:                r->escape = *p;
                   3656:                if (*++p != '\0')
1.350     schwarze 3657:                        mandoc_msg(MANDOCERR_ARG_EXCESS, ln,
                   3658:                            (int)(p - buf->buf), "ec ... %s", p);
1.303     schwarze 3659:        }
                   3660:        return ROFF_IGN;
                   3661: }
                   3662:
1.340     schwarze 3663: static int
1.303     schwarze 3664: roff_eo(ROFF_ARGS)
                   3665: {
                   3666:        r->escape = '\0';
                   3667:        if (buf->buf[pos] != '\0')
1.350     schwarze 3668:                mandoc_msg(MANDOCERR_ARG_SKIP,
1.303     schwarze 3669:                    ln, pos, "eo %s", buf->buf + pos);
1.384     schwarze 3670:        return ROFF_IGN;
                   3671: }
                   3672:
                   3673: static int
                   3674: roff_mc(ROFF_ARGS)
                   3675: {
                   3676:        struct roff_node        *n;
                   3677:        char                    *cp;
                   3678:
                   3679:        /* Parse the first argument. */
                   3680:
                   3681:        cp = buf->buf + pos;
                   3682:        if (*cp != '\0')
                   3683:                cp++;
                   3684:        if (buf->buf[pos] == '\\') {
                   3685:                switch (mandoc_escape((const char **)&cp, NULL, NULL)) {
                   3686:                case ESCAPE_SPECIAL:
                   3687:                case ESCAPE_UNICODE:
                   3688:                case ESCAPE_NUMBERED:
                   3689:                        break;
                   3690:                default:
                   3691:                        *cp = '\0';
                   3692:                        mandoc_msg(MANDOCERR_MC_ESC, ln, pos,
                   3693:                            "mc %s", buf->buf + pos);
                   3694:                        buf->buf[pos] = '\0';
                   3695:                        break;
                   3696:                }
                   3697:        }
                   3698:
                   3699:        /* Ignore additional arguments. */
                   3700:
                   3701:        while (*cp == ' ')
                   3702:                *cp++ = '\0';
                   3703:        if (*cp != '\0') {
                   3704:                mandoc_msg(MANDOCERR_MC_DIST, ln, (int)(cp - buf->buf),
                   3705:                    "mc ... %s", cp);
                   3706:                *cp = '\0';
                   3707:        }
                   3708:
                   3709:        /* Create the .mc node. */
                   3710:
                   3711:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3712:        n = r->man->last;
                   3713:        if (buf->buf[pos] != '\0')
                   3714:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
                   3715:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3716:        r->man->last = n;
                   3717:        r->man->next = ROFF_NEXT_SIBLING;
1.303     schwarze 3718:        return ROFF_IGN;
1.330     schwarze 3719: }
                   3720:
1.340     schwarze 3721: static int
1.330     schwarze 3722: roff_nop(ROFF_ARGS)
                   3723: {
                   3724:        while (buf->buf[pos] == ' ')
                   3725:                pos++;
                   3726:        *offs = pos;
                   3727:        return ROFF_RERUN;
1.303     schwarze 3728: }
                   3729:
1.340     schwarze 3730: static int
1.164     kristaps 3731: roff_tr(ROFF_ARGS)
                   3732: {
                   3733:        const char      *p, *first, *second;
                   3734:        size_t           fsz, ssz;
                   3735:
1.238     schwarze 3736:        p = buf->buf + pos;
1.164     kristaps 3737:
1.238     schwarze 3738:        if (*p == '\0') {
1.350     schwarze 3739:                mandoc_msg(MANDOCERR_REQ_EMPTY, ln, ppos, "tr");
1.277     schwarze 3740:                return ROFF_IGN;
1.164     kristaps 3741:        }
                   3742:
1.238     schwarze 3743:        while (*p != '\0') {
1.164     kristaps 3744:                fsz = ssz = 1;
                   3745:
                   3746:                first = p++;
1.238     schwarze 3747:                if (*first == '\\') {
1.394     schwarze 3748:                        if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR)
1.277     schwarze 3749:                                return ROFF_IGN;
1.164     kristaps 3750:                        fsz = (size_t)(p - first);
                   3751:                }
                   3752:
                   3753:                second = p++;
1.238     schwarze 3754:                if (*second == '\\') {
1.394     schwarze 3755:                        if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR)
1.277     schwarze 3756:                                return ROFF_IGN;
1.164     kristaps 3757:                        ssz = (size_t)(p - second);
1.238     schwarze 3758:                } else if (*second == '\0') {
1.350     schwarze 3759:                        mandoc_msg(MANDOCERR_TR_ODD, ln,
                   3760:                            (int)(first - buf->buf), "tr %s", first);
1.164     kristaps 3761:                        second = " ";
1.165     kristaps 3762:                        p--;
1.164     kristaps 3763:                }
                   3764:
1.167     kristaps 3765:                if (fsz > 1) {
1.207     schwarze 3766:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3767:                            second, ssz, 0);
1.167     kristaps 3768:                        continue;
                   3769:                }
                   3770:
1.238     schwarze 3771:                if (r->xtab == NULL)
1.207     schwarze 3772:                        r->xtab = mandoc_calloc(128,
                   3773:                            sizeof(struct roffstr));
1.167     kristaps 3774:
                   3775:                free(r->xtab[(int)*first].p);
                   3776:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3777:                r->xtab[(int)*first].sz = ssz;
1.164     kristaps 3778:        }
                   3779:
1.277     schwarze 3780:        return ROFF_IGN;
1.164     kristaps 3781: }
                   3782:
1.339     schwarze 3783: /*
                   3784:  * Implementation of the .return request.
                   3785:  * There is no need to call roff_userret() from here.
                   3786:  * The read module will call that after rewinding the reader stack
                   3787:  * to the place from where the current macro was called.
                   3788:  */
1.340     schwarze 3789: static int
1.339     schwarze 3790: roff_return(ROFF_ARGS)
                   3791: {
                   3792:        if (r->mstackpos >= 0)
1.340     schwarze 3793:                return ROFF_IGN | ROFF_USERRET;
1.339     schwarze 3794:
1.350     schwarze 3795:        mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "return");
1.339     schwarze 3796:        return ROFF_IGN;
                   3797: }
                   3798:
1.340     schwarze 3799: static int
1.306     schwarze 3800: roff_rn(ROFF_ARGS)
                   3801: {
                   3802:        const char      *value;
                   3803:        char            *oldn, *newn, *end;
                   3804:        size_t           oldsz, newsz;
1.315     schwarze 3805:        int              deftype;
1.306     schwarze 3806:
                   3807:        oldn = newn = buf->buf + pos;
                   3808:        if (*oldn == '\0')
                   3809:                return ROFF_IGN;
                   3810:
                   3811:        oldsz = roff_getname(r, &newn, ln, pos);
1.363     schwarze 3812:        if (oldn[oldsz] == '\\' || oldn[oldsz] == '\t' || *newn == '\0')
1.306     schwarze 3813:                return ROFF_IGN;
                   3814:
                   3815:        end = newn;
                   3816:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3817:        if (newsz == 0)
                   3818:                return ROFF_IGN;
                   3819:
1.315     schwarze 3820:        deftype = ROFFDEF_ANY;
                   3821:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3822:        switch (deftype) {
                   3823:        case ROFFDEF_USER:
1.306     schwarze 3824:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3825:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3826:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315     schwarze 3827:                break;
                   3828:        case ROFFDEF_PRE:
                   3829:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3830:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3831:                break;
                   3832:        case ROFFDEF_REN:
1.306     schwarze 3833:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3834:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315     schwarze 3835:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3836:                break;
                   3837:        case ROFFDEF_STD:
1.306     schwarze 3838:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315     schwarze 3839:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3840:                break;
                   3841:        default:
                   3842:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3843:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3844:                break;
                   3845:        }
1.306     schwarze 3846:        return ROFF_IGN;
                   3847: }
                   3848:
1.340     schwarze 3849: static int
1.339     schwarze 3850: roff_shift(ROFF_ARGS)
                   3851: {
                   3852:        struct mctx     *ctx;
1.382     schwarze 3853:        int              argpos, levels, i;
1.339     schwarze 3854:
1.382     schwarze 3855:        argpos = pos;
1.339     schwarze 3856:        levels = 1;
                   3857:        if (buf->buf[pos] != '\0' &&
                   3858:            roff_evalnum(r, ln, buf->buf, &pos, &levels, 0) == 0) {
1.350     schwarze 3859:                mandoc_msg(MANDOCERR_CE_NONUM,
1.339     schwarze 3860:                    ln, pos, "shift %s", buf->buf + pos);
                   3861:                levels = 1;
                   3862:        }
                   3863:        if (r->mstackpos < 0) {
1.350     schwarze 3864:                mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "shift");
1.339     schwarze 3865:                return ROFF_IGN;
                   3866:        }
                   3867:        ctx = r->mstack + r->mstackpos;
                   3868:        if (levels > ctx->argc) {
1.350     schwarze 3869:                mandoc_msg(MANDOCERR_SHIFT,
1.382     schwarze 3870:                    ln, argpos, "%d, but max is %d", levels, ctx->argc);
1.339     schwarze 3871:                levels = ctx->argc;
1.382     schwarze 3872:        }
                   3873:        if (levels < 0) {
                   3874:                mandoc_msg(MANDOCERR_ARG_NEG, ln, argpos, "shift %d", levels);
                   3875:                levels = 0;
1.339     schwarze 3876:        }
                   3877:        if (levels == 0)
                   3878:                return ROFF_IGN;
                   3879:        for (i = 0; i < levels; i++)
                   3880:                free(ctx->argv[i]);
                   3881:        ctx->argc -= levels;
                   3882:        for (i = 0; i < ctx->argc; i++)
                   3883:                ctx->argv[i] = ctx->argv[i + levels];
                   3884:        return ROFF_IGN;
                   3885: }
                   3886:
1.340     schwarze 3887: static int
1.105     kristaps 3888: roff_so(ROFF_ARGS)
                   3889: {
1.249     schwarze 3890:        char *name, *cp;
1.105     kristaps 3891:
1.238     schwarze 3892:        name = buf->buf + pos;
1.350     schwarze 3893:        mandoc_msg(MANDOCERR_SO, ln, ppos, "so %s", name);
1.105     kristaps 3894:
                   3895:        /*
                   3896:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3897:         * opening anything that's not in our cwd or anything beneath
                   3898:         * it.  Thus, explicitly disallow traversing up the file-system
                   3899:         * or using absolute paths.
                   3900:         */
                   3901:
1.238     schwarze 3902:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.350     schwarze 3903:                mandoc_msg(MANDOCERR_SO_PATH, ln, ppos, ".so %s", name);
1.249     schwarze 3904:                buf->sz = mandoc_asprintf(&cp,
                   3905:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3906:                free(buf->buf);
                   3907:                buf->buf = cp;
                   3908:                *offs = 0;
1.277     schwarze 3909:                return ROFF_REPARSE;
1.105     kristaps 3910:        }
                   3911:
                   3912:        *offs = pos;
1.277     schwarze 3913:        return ROFF_SO;
1.105     kristaps 3914: }
1.92      schwarze 3915:
1.266     schwarze 3916: /* --- user defined strings and macros ------------------------------------ */
                   3917:
1.340     schwarze 3918: static int
1.106     kristaps 3919: roff_userdef(ROFF_ARGS)
1.99      kristaps 3920: {
1.339     schwarze 3921:        struct mctx      *ctx;
                   3922:        char             *arg, *ap, *dst, *src;
                   3923:        size_t            sz;
                   3924:
1.365     schwarze 3925:        /* If the macro is empty, ignore it altogether. */
                   3926:
                   3927:        if (*r->current_string == '\0')
                   3928:                return ROFF_IGN;
                   3929:
1.339     schwarze 3930:        /* Initialize a new macro stack context. */
                   3931:
                   3932:        if (++r->mstackpos == r->mstacksz) {
                   3933:                r->mstack = mandoc_recallocarray(r->mstack,
                   3934:                    r->mstacksz, r->mstacksz + 8, sizeof(*r->mstack));
                   3935:                r->mstacksz += 8;
                   3936:        }
                   3937:        ctx = r->mstack + r->mstackpos;
                   3938:        ctx->argc = 0;
                   3939:
                   3940:        /*
                   3941:         * Collect pointers to macro argument strings,
                   3942:         * NUL-terminating them and escaping quotes.
                   3943:         */
                   3944:
                   3945:        src = buf->buf + pos;
                   3946:        while (*src != '\0') {
                   3947:                if (ctx->argc == ctx->argsz) {
                   3948:                        ctx->argsz += 8;
                   3949:                        ctx->argv = mandoc_reallocarray(ctx->argv,
                   3950:                            ctx->argsz, sizeof(*ctx->argv));
                   3951:                }
1.355     schwarze 3952:                arg = roff_getarg(r, &src, ln, &pos);
1.339     schwarze 3953:                sz = 1;  /* For the terminating NUL. */
                   3954:                for (ap = arg; *ap != '\0'; ap++)
                   3955:                        sz += *ap == '"' ? 4 : 1;
                   3956:                ctx->argv[ctx->argc++] = dst = mandoc_malloc(sz);
                   3957:                for (ap = arg; *ap != '\0'; ap++) {
                   3958:                        if (*ap == '"') {
                   3959:                                memcpy(dst, "\\(dq", 4);
                   3960:                                dst += 4;
                   3961:                        } else
                   3962:                                *dst++ = *ap;
1.106     kristaps 3963:                }
1.339     schwarze 3964:                *dst = '\0';
1.355     schwarze 3965:                free(arg);
1.337     schwarze 3966:        }
                   3967:
1.339     schwarze 3968:        /* Replace the macro invocation by the macro definition. */
1.263     schwarze 3969:
1.238     schwarze 3970:        free(buf->buf);
1.339     schwarze 3971:        buf->buf = mandoc_strdup(r->current_string);
                   3972:        buf->sz = strlen(buf->buf) + 1;
1.248     schwarze 3973:        *offs = 0;
1.106     kristaps 3974:
1.365     schwarze 3975:        return buf->buf[buf->sz - 2] == '\n' ?
1.340     schwarze 3976:            ROFF_REPARSE | ROFF_USERCALL : ROFF_IGN | ROFF_APPEND;
1.99      kristaps 3977: }
1.121     schwarze 3978:
1.306     schwarze 3979: /*
                   3980:  * Calling a high-level macro that was renamed with .rn.
                   3981:  * r->current_string has already been set up by roff_parse().
                   3982:  */
1.340     schwarze 3983: static int
1.306     schwarze 3984: roff_renamed(ROFF_ARGS)
                   3985: {
                   3986:        char    *nbuf;
                   3987:
1.315     schwarze 3988:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3989:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306     schwarze 3990:        free(buf->buf);
                   3991:        buf->buf = nbuf;
1.329     schwarze 3992:        *offs = 0;
1.306     schwarze 3993:        return ROFF_CONT;
                   3994: }
                   3995:
1.362     schwarze 3996: /*
                   3997:  * Measure the length in bytes of the roff identifier at *cpp
                   3998:  * and advance the pointer to the next word.
                   3999:  */
1.212     schwarze 4000: static size_t
1.121     schwarze 4001: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   4002: {
                   4003:        char     *name, *cp;
1.393     schwarze 4004:        int       namesz, inam, iend;
1.121     schwarze 4005:
                   4006:        name = *cpp;
1.362     schwarze 4007:        if (*name == '\0')
1.277     schwarze 4008:                return 0;
1.121     schwarze 4009:
1.362     schwarze 4010:        /* Advance cp to the byte after the end of the name. */
                   4011:
1.393     schwarze 4012:        cp = name;
                   4013:        namesz = 0;
                   4014:        for (;;) {
1.363     schwarze 4015:                if (*cp == '\0')
1.212     schwarze 4016:                        break;
1.363     schwarze 4017:                if (*cp == ' ' || *cp == '\t') {
                   4018:                        cp++;
                   4019:                        break;
                   4020:                }
1.393     schwarze 4021:                if (*cp != '\\') {
                   4022:                        if (name + namesz < cp) {
                   4023:                                name[namesz] = *cp;
                   4024:                                *cp = ' ';
                   4025:                        }
                   4026:                        namesz++;
                   4027:                        cp++;
1.121     schwarze 4028:                        continue;
1.393     schwarze 4029:                }
1.362     schwarze 4030:                if (cp[1] == '{' || cp[1] == '}')
1.215     schwarze 4031:                        break;
1.393     schwarze 4032:                if (roff_escape(cp, 0, 0, NULL, &inam,
                   4033:                    NULL, NULL, &iend) != ESCAPE_UNDEF) {
                   4034:                        mandoc_msg(MANDOCERR_NAMESC, ln, pos,
                   4035:                            "%.*s%.*s", namesz, name, iend, cp);
                   4036:                        cp += iend;
                   4037:                        break;
                   4038:                }
                   4039:
                   4040:                /*
                   4041:                 * In an identifier, \\, \., \G and so on
                   4042:                 * are reduced to \, ., G and so on,
                   4043:                 * vaguely similar to copy mode.
                   4044:                 */
                   4045:
                   4046:                name[namesz++] = cp[inam];
                   4047:                while (iend--) {
                   4048:                        if (cp >= name + namesz)
                   4049:                                *cp = ' ';
                   4050:                        cp++;
                   4051:                }
1.121     schwarze 4052:        }
                   4053:
                   4054:        /* Read past spaces. */
1.362     schwarze 4055:
                   4056:        while (*cp == ' ')
1.121     schwarze 4057:                cp++;
                   4058:
                   4059:        *cpp = cp;
1.277     schwarze 4060:        return namesz;
1.121     schwarze 4061: }
                   4062:
1.106     kristaps 4063: /*
                   4064:  * Store *string into the user-defined string called *name.
                   4065:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193     schwarze 4066:  * append == 0: replace mode
                   4067:  * append == 1: single-line append mode
                   4068:  * append == 2: multiline append mode, append '\n' after each call
1.106     kristaps 4069:  */
1.94      kristaps 4070: static void
1.106     kristaps 4071: roff_setstr(struct roff *r, const char *name, const char *string,
1.193     schwarze 4072:        int append)
1.92      schwarze 4073: {
1.315     schwarze 4074:        size_t   namesz;
1.164     kristaps 4075:
1.315     schwarze 4076:        namesz = strlen(name);
                   4077:        roff_setstrn(&r->strtab, name, namesz, string,
1.207     schwarze 4078:            string ? strlen(string) : 0, append);
1.315     schwarze 4079:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164     kristaps 4080: }
                   4081:
                   4082: static void
1.166     kristaps 4083: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193     schwarze 4084:                const char *string, size_t stringsz, int append)
1.164     kristaps 4085: {
1.166     kristaps 4086:        struct roffkv   *n;
1.164     kristaps 4087:        char            *c;
                   4088:        int              i;
                   4089:        size_t           oldch, newch;
1.92      schwarze 4090:
1.106     kristaps 4091:        /* Search for an existing string with the same name. */
1.164     kristaps 4092:        n = *r;
                   4093:
1.211     schwarze 4094:        while (n && (namesz != n->key.sz ||
                   4095:                        strncmp(n->key.p, name, namesz)))
1.92      schwarze 4096:                n = n->next;
1.94      kristaps 4097:
                   4098:        if (NULL == n) {
1.106     kristaps 4099:                /* Create a new string table entry. */
1.166     kristaps 4100:                n = mandoc_malloc(sizeof(struct roffkv));
                   4101:                n->key.p = mandoc_strndup(name, namesz);
                   4102:                n->key.sz = namesz;
                   4103:                n->val.p = NULL;
                   4104:                n->val.sz = 0;
1.164     kristaps 4105:                n->next = *r;
                   4106:                *r = n;
1.193     schwarze 4107:        } else if (0 == append) {
1.166     kristaps 4108:                free(n->val.p);
                   4109:                n->val.p = NULL;
                   4110:                n->val.sz = 0;
1.106     kristaps 4111:        }
                   4112:
                   4113:        if (NULL == string)
                   4114:                return;
                   4115:
                   4116:        /*
                   4117:         * One additional byte for the '\n' in multiline mode,
                   4118:         * and one for the terminating '\0'.
                   4119:         */
1.193     schwarze 4120:        newch = stringsz + (1 < append ? 2u : 1u);
1.164     kristaps 4121:
1.166     kristaps 4122:        if (NULL == n->val.p) {
                   4123:                n->val.p = mandoc_malloc(newch);
                   4124:                *n->val.p = '\0';
1.106     kristaps 4125:                oldch = 0;
                   4126:        } else {
1.166     kristaps 4127:                oldch = n->val.sz;
                   4128:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106     kristaps 4129:        }
                   4130:
                   4131:        /* Skip existing content in the destination buffer. */
1.166     kristaps 4132:        c = n->val.p + (int)oldch;
1.106     kristaps 4133:
                   4134:        /* Append new content to the destination buffer. */
1.164     kristaps 4135:        i = 0;
                   4136:        while (i < (int)stringsz) {
1.106     kristaps 4137:                /*
                   4138:                 * Rudimentary roff copy mode:
                   4139:                 * Handle escaped backslashes.
                   4140:                 */
1.164     kristaps 4141:                if ('\\' == string[i] && '\\' == string[i + 1])
                   4142:                        i++;
                   4143:                *c++ = string[i++];
1.106     kristaps 4144:        }
1.94      kristaps 4145:
1.106     kristaps 4146:        /* Append terminating bytes. */
1.193     schwarze 4147:        if (1 < append)
1.106     kristaps 4148:                *c++ = '\n';
1.163     kristaps 4149:
1.106     kristaps 4150:        *c = '\0';
1.166     kristaps 4151:        n->val.sz = (int)(c - n->val.p);
1.92      schwarze 4152: }
                   4153:
1.94      kristaps 4154: static const char *
1.325     schwarze 4155: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315     schwarze 4156:     int *deftype)
1.92      schwarze 4157: {
1.315     schwarze 4158:        const struct roffkv     *n;
1.325     schwarze 4159:        int                      found, i;
1.315     schwarze 4160:        enum roff_tok            tok;
                   4161:
1.325     schwarze 4162:        found = 0;
                   4163:        for (n = r->strtab; n != NULL; n = n->next) {
                   4164:                if (strncmp(name, n->key.p, len) != 0 ||
                   4165:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4166:                        continue;
                   4167:                if (*deftype & ROFFDEF_USER) {
                   4168:                        *deftype = ROFFDEF_USER;
                   4169:                        return n->val.p;
                   4170:                } else {
                   4171:                        found = 1;
                   4172:                        break;
                   4173:                }
                   4174:        }
                   4175:        for (n = r->rentab; n != NULL; n = n->next) {
                   4176:                if (strncmp(name, n->key.p, len) != 0 ||
                   4177:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4178:                        continue;
                   4179:                if (*deftype & ROFFDEF_REN) {
                   4180:                        *deftype = ROFFDEF_REN;
                   4181:                        return n->val.p;
                   4182:                } else {
                   4183:                        found = 1;
                   4184:                        break;
1.315     schwarze 4185:                }
                   4186:        }
1.325     schwarze 4187:        for (i = 0; i < PREDEFS_MAX; i++) {
                   4188:                if (strncmp(name, predefs[i].name, len) != 0 ||
                   4189:                    predefs[i].name[len] != '\0')
                   4190:                        continue;
                   4191:                if (*deftype & ROFFDEF_PRE) {
                   4192:                        *deftype = ROFFDEF_PRE;
                   4193:                        return predefs[i].str;
                   4194:                } else {
                   4195:                        found = 1;
                   4196:                        break;
1.315     schwarze 4197:                }
                   4198:        }
1.356     schwarze 4199:        if (r->man->meta.macroset != MACROSET_MAN) {
1.325     schwarze 4200:                for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
                   4201:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4202:                            roff_name[tok][len] != '\0')
                   4203:                                continue;
                   4204:                        if (*deftype & ROFFDEF_STD) {
                   4205:                                *deftype = ROFFDEF_STD;
                   4206:                                return NULL;
                   4207:                        } else {
                   4208:                                found = 1;
                   4209:                                break;
1.315     schwarze 4210:                        }
                   4211:                }
                   4212:        }
1.356     schwarze 4213:        if (r->man->meta.macroset != MACROSET_MDOC) {
1.325     schwarze 4214:                for (tok = MAN_TH; tok < MAN_MAX; tok++) {
                   4215:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4216:                            roff_name[tok][len] != '\0')
                   4217:                                continue;
                   4218:                        if (*deftype & ROFFDEF_STD) {
                   4219:                                *deftype = ROFFDEF_STD;
                   4220:                                return NULL;
                   4221:                        } else {
                   4222:                                found = 1;
                   4223:                                break;
1.315     schwarze 4224:                        }
                   4225:                }
1.325     schwarze 4226:        }
                   4227:
                   4228:        if (found == 0 && *deftype != ROFFDEF_ANY) {
                   4229:                if (*deftype & ROFFDEF_REN) {
                   4230:                        /*
                   4231:                         * This might still be a request,
                   4232:                         * so do not treat it as undefined yet.
                   4233:                         */
                   4234:                        *deftype = ROFFDEF_UNDEF;
                   4235:                        return NULL;
1.315     schwarze 4236:                }
1.325     schwarze 4237:
                   4238:                /* Using an undefined string defines it to be empty. */
                   4239:
                   4240:                roff_setstrn(&r->strtab, name, len, "", 0, 0);
                   4241:                roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315     schwarze 4242:        }
1.325     schwarze 4243:
1.315     schwarze 4244:        *deftype = 0;
1.277     schwarze 4245:        return NULL;
1.92      schwarze 4246: }
                   4247:
1.94      kristaps 4248: static void
1.167     kristaps 4249: roff_freestr(struct roffkv *r)
1.92      schwarze 4250: {
1.166     kristaps 4251:        struct roffkv    *n, *nn;
1.92      schwarze 4252:
1.167     kristaps 4253:        for (n = r; n; n = nn) {
1.166     kristaps 4254:                free(n->key.p);
                   4255:                free(n->val.p);
1.92      schwarze 4256:                nn = n->next;
                   4257:                free(n);
                   4258:        }
1.114     kristaps 4259: }
1.266     schwarze 4260:
                   4261: /* --- accessors and utility functions ------------------------------------ */
1.164     kristaps 4262:
                   4263: /*
                   4264:  * Duplicate an input string, making the appropriate character
                   4265:  * conversations (as stipulated by `tr') along the way.
                   4266:  * Returns a heap-allocated string with all the replacements made.
                   4267:  */
                   4268: char *
                   4269: roff_strdup(const struct roff *r, const char *p)
                   4270: {
1.166     kristaps 4271:        const struct roffkv *cp;
1.164     kristaps 4272:        char            *res;
                   4273:        const char      *pp;
                   4274:        size_t           ssz, sz;
                   4275:        enum mandoc_esc  esc;
                   4276:
1.167     kristaps 4277:        if (NULL == r->xmbtab && NULL == r->xtab)
1.277     schwarze 4278:                return mandoc_strdup(p);
1.164     kristaps 4279:        else if ('\0' == *p)
1.277     schwarze 4280:                return mandoc_strdup("");
1.164     kristaps 4281:
                   4282:        /*
                   4283:         * Step through each character looking for term matches
                   4284:         * (remember that a `tr' can be invoked with an escape, which is
                   4285:         * a glyph but the escape is multi-character).
                   4286:         * We only do this if the character hash has been initialised
                   4287:         * and the string is >0 length.
                   4288:         */
                   4289:
                   4290:        res = NULL;
                   4291:        ssz = 0;
                   4292:
                   4293:        while ('\0' != *p) {
1.289     schwarze 4294:                assert((unsigned int)*p < 128);
                   4295:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167     kristaps 4296:                        sz = r->xtab[(int)*p].sz;
                   4297:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4298:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   4299:                        ssz += sz;
                   4300:                        p++;
                   4301:                        continue;
                   4302:                } else if ('\\' != *p) {
                   4303:                        res = mandoc_realloc(res, ssz + 2);
                   4304:                        res[ssz++] = *p++;
                   4305:                        continue;
                   4306:                }
                   4307:
1.164     kristaps 4308:                /* Search for term matches. */
1.167     kristaps 4309:                for (cp = r->xmbtab; cp; cp = cp->next)
1.166     kristaps 4310:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164     kristaps 4311:                                break;
                   4312:
                   4313:                if (NULL != cp) {
                   4314:                        /*
                   4315:                         * A match has been found.
                   4316:                         * Append the match to the array and move
                   4317:                         * forward by its keysize.
                   4318:                         */
1.207     schwarze 4319:                        res = mandoc_realloc(res,
                   4320:                            ssz + cp->val.sz + 1);
1.166     kristaps 4321:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   4322:                        ssz += cp->val.sz;
                   4323:                        p += (int)cp->key.sz;
1.164     kristaps 4324:                        continue;
                   4325:                }
                   4326:
1.167     kristaps 4327:                /*
                   4328:                 * Handle escapes carefully: we need to copy
                   4329:                 * over just the escape itself, or else we might
                   4330:                 * do replacements within the escape itself.
                   4331:                 * Make sure to pass along the bogus string.
                   4332:                 */
                   4333:                pp = p++;
                   4334:                esc = mandoc_escape(&p, NULL, NULL);
                   4335:                if (ESCAPE_ERROR == esc) {
                   4336:                        sz = strlen(pp);
1.164     kristaps 4337:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4338:                        memcpy(res + ssz, pp, sz);
1.167     kristaps 4339:                        break;
1.164     kristaps 4340:                }
1.207     schwarze 4341:                /*
                   4342:                 * We bail out on bad escapes.
1.167     kristaps 4343:                 * No need to warn: we already did so when
1.355     schwarze 4344:                 * roff_expand() was called.
1.167     kristaps 4345:                 */
                   4346:                sz = (int)(p - pp);
                   4347:                res = mandoc_realloc(res, ssz + sz + 1);
                   4348:                memcpy(res + ssz, pp, sz);
                   4349:                ssz += sz;
1.164     kristaps 4350:        }
                   4351:
                   4352:        res[(int)ssz] = '\0';
1.277     schwarze 4353:        return res;
1.227     schwarze 4354: }
                   4355:
                   4356: int
                   4357: roff_getformat(const struct roff *r)
                   4358: {
                   4359:
1.277     schwarze 4360:        return r->format;
1.174     kristaps 4361: }
                   4362:
                   4363: /*
1.207     schwarze 4364:  * Find out whether a line is a macro line or not.
1.174     kristaps 4365:  * If it is, adjust the current position and return one; if it isn't,
                   4366:  * return zero and don't change the current position.
                   4367:  * If the control character has been set with `.cc', then let that grain
                   4368:  * precedence.
1.396     schwarze 4369:  * This is slightly contrary to groff, where using the non-breaking
1.174     kristaps 4370:  * control character when `cc' has been invoked will cause the
                   4371:  * non-breaking macro contents to be printed verbatim.
                   4372:  */
                   4373: int
                   4374: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   4375: {
                   4376:        int             pos;
                   4377:
                   4378:        pos = *ppos;
                   4379:
1.303     schwarze 4380:        if (r->control != '\0' && cp[pos] == r->control)
1.174     kristaps 4381:                pos++;
1.303     schwarze 4382:        else if (r->control != '\0')
1.277     schwarze 4383:                return 0;
1.174     kristaps 4384:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   4385:                pos += 2;
                   4386:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   4387:                pos++;
                   4388:        else
1.277     schwarze 4389:                return 0;
1.174     kristaps 4390:
                   4391:        while (' ' == cp[pos] || '\t' == cp[pos])
                   4392:                pos++;
                   4393:
                   4394:        *ppos = pos;
1.277     schwarze 4395:        return 1;
1.74      kristaps 4396: }

CVSweb