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

Annotation of mandoc/roff.c, Revision 1.394

1.394   ! schwarze    1: /* $Id: roff.c,v 1.393 2022/06/03 12:15:55 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:
1.355     schwarze   43: /*
                     44:  * ASCII_ESC is used to signal from roff_getarg() to roff_expand()
                     45:  * that an escape sequence resulted from copy-in processing and
                     46:  * needs to be checked or interpolated.  As it is used nowhere
                     47:  * else, it is defined here rather than in a header file.
                     48:  */
                     49: #define        ASCII_ESC       27
                     50:
1.170     schwarze   51: /* Maximum number of string expansions per line, to break infinite loops. */
                     52: #define        EXPAND_LIMIT    1000
                     53:
1.315     schwarze   54: /* Types of definitions of macros and strings. */
                     55: #define        ROFFDEF_USER    (1 << 1)  /* User-defined. */
                     56: #define        ROFFDEF_PRE     (1 << 2)  /* Predefined. */
                     57: #define        ROFFDEF_REN     (1 << 3)  /* Renamed standard macro. */
                     58: #define        ROFFDEF_STD     (1 << 4)  /* mdoc(7) or man(7) macro. */
                     59: #define        ROFFDEF_ANY     (ROFFDEF_USER | ROFFDEF_PRE | \
                     60:                         ROFFDEF_REN | ROFFDEF_STD)
1.325     schwarze   61: #define        ROFFDEF_UNDEF   (1 << 5)  /* Completely undefined. */
1.315     schwarze   62:
1.266     schwarze   63: /* --- data types --------------------------------------------------------- */
                     64:
1.147     kristaps   65: /*
1.167     kristaps   66:  * An incredibly-simple string buffer.
                     67:  */
1.94      kristaps   68: struct roffstr {
1.167     kristaps   69:        char            *p; /* nil-terminated buffer */
                     70:        size_t           sz; /* saved strlen(p) */
1.166     kristaps   71: };
                     72:
                     73: /*
1.167     kristaps   74:  * A key-value roffstr pair as part of a singly-linked list.
1.166     kristaps   75:  */
                     76: struct roffkv {
                     77:        struct roffstr   key;
                     78:        struct roffstr   val;
                     79:        struct roffkv   *next; /* next in list */
1.94      kristaps   80: };
                     81:
1.180     schwarze   82: /*
                     83:  * A single number register as part of a singly-linked list.
                     84:  */
                     85: struct roffreg {
                     86:        struct roffstr   key;
1.181     schwarze   87:        int              val;
1.327     schwarze   88:        int              step;
1.180     schwarze   89:        struct roffreg  *next;
                     90: };
                     91:
1.295     schwarze   92: /*
                     93:  * Association of request and macro names with token IDs.
                     94:  */
                     95: struct roffreq {
                     96:        enum roff_tok    tok;
                     97:        char             name[];
                     98: };
                     99:
1.339     schwarze  100: /*
                    101:  * A macro processing context.
                    102:  * More than one is needed when macro calls are nested.
                    103:  */
                    104: struct mctx {
                    105:        char            **argv;
                    106:        int              argc;
                    107:        int              argsz;
                    108: };
                    109:
1.67      kristaps  110: struct roff {
1.296     schwarze  111:        struct roff_man *man; /* mdoc or man parser */
1.67      kristaps  112:        struct roffnode *last; /* leaf of stack */
1.339     schwarze  113:        struct mctx     *mstack; /* stack of macro contexts */
1.223     schwarze  114:        int             *rstack; /* stack of inverted `ie' values */
1.295     schwarze  115:        struct ohash    *reqtab; /* request lookup table */
1.180     schwarze  116:        struct roffreg  *regtab; /* number registers */
1.166     kristaps  117:        struct roffkv   *strtab; /* user-defined strings & macros */
1.306     schwarze  118:        struct roffkv   *rentab; /* renamed strings & macros */
1.167     kristaps  119:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                    120:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.106     kristaps  121:        const char      *current_string; /* value of last called user macro */
1.118     kristaps  122:        struct tbl_node *first_tbl; /* first table parsed */
                    123:        struct tbl_node *last_tbl; /* last table parsed */
                    124:        struct tbl_node *tbl; /* current table being parsed */
1.319     schwarze  125:        struct eqn_node *last_eqn; /* equation parser */
                    126:        struct eqn_node *eqn; /* active equation parser */
1.230     schwarze  127:        int              eqn_inline; /* current equation is inline */
1.223     schwarze  128:        int              options; /* parse options */
1.339     schwarze  129:        int              mstacksz; /* current size of mstack */
                    130:        int              mstackpos; /* position in mstack */
1.223     schwarze  131:        int              rstacksz; /* current size limit of rstack */
                    132:        int              rstackpos; /* position in rstack */
1.227     schwarze  133:        int              format; /* current file in mdoc or man format */
1.223     schwarze  134:        char             control; /* control character */
1.303     schwarze  135:        char             escape; /* escape character */
1.79      kristaps  136: };
                    137:
1.364     schwarze  138: /*
                    139:  * A macro definition, condition, or ignored block.
                    140:  */
1.67      kristaps  141: struct roffnode {
1.294     schwarze  142:        enum roff_tok    tok; /* type of node */
1.67      kristaps  143:        struct roffnode *parent; /* up one in stack */
                    144:        int              line; /* parse line */
                    145:        int              col; /* parse col */
1.106     kristaps  146:        char            *name; /* node name, e.g. macro name */
1.364     schwarze  147:        char            *end; /* custom end macro of the block */
                    148:        int              endspan; /* scope to: 1=eol 2=next line -1=\} */
                    149:        int              rule; /* content is: 1=evaluated 0=skipped */
1.67      kristaps  150: };
                    151:
                    152: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.294     schwarze  153:                         enum roff_tok tok, /* tok of macro */ \
1.238     schwarze  154:                         struct buf *buf, /* input buffer */ \
1.67      kristaps  155:                         int ln, /* parse line */ \
1.75      kristaps  156:                         int ppos, /* original pos in buffer */ \
                    157:                         int pos, /* current pos in buffer */ \
1.74      kristaps  158:                         int *offs /* reset offset of buffer data */
1.67      kristaps  159:
1.340     schwarze  160: typedef        int (*roffproc)(ROFF_ARGS);
1.67      kristaps  161:
                    162: struct roffmac {
1.79      kristaps  163:        roffproc         proc; /* process new macro */
                    164:        roffproc         text; /* process as child text of macro */
                    165:        roffproc         sub; /* process as child of macro */
                    166:        int              flags;
                    167: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.67      kristaps  168: };
                    169:
1.141     kristaps  170: struct predef {
                    171:        const char      *name; /* predefined input name */
                    172:        const char      *str; /* replacement symbol */
                    173: };
                    174:
                    175: #define        PREDEF(__name, __str) \
                    176:        { (__name), (__str) },
                    177:
1.266     schwarze  178: /* --- function prototypes ------------------------------------------------ */
                    179:
1.340     schwarze  180: static int              roffnode_cleanscope(struct roff *);
                    181: static int              roffnode_pop(struct roff *);
1.294     schwarze  182: static void             roffnode_push(struct roff *, enum roff_tok,
1.155     kristaps  183:                                const char *, int, int);
1.346     schwarze  184: static void             roff_addtbl(struct roff_man *, int, struct tbl_node *);
1.340     schwarze  185: static int              roff_als(ROFF_ARGS);
                    186: static int              roff_block(ROFF_ARGS);
                    187: static int              roff_block_text(ROFF_ARGS);
                    188: static int              roff_block_sub(ROFF_ARGS);
1.364     schwarze  189: static int              roff_break(ROFF_ARGS);
1.340     schwarze  190: static int              roff_cblock(ROFF_ARGS);
                    191: static int              roff_cc(ROFF_ARGS);
                    192: static int              roff_ccond(struct roff *, int, int);
1.341     schwarze  193: static int              roff_char(ROFF_ARGS);
1.340     schwarze  194: static int              roff_cond(ROFF_ARGS);
1.375     schwarze  195: static int              roff_cond_checkend(ROFF_ARGS);
1.340     schwarze  196: static int              roff_cond_text(ROFF_ARGS);
                    197: static int              roff_cond_sub(ROFF_ARGS);
                    198: static int              roff_ds(ROFF_ARGS);
                    199: static int              roff_ec(ROFF_ARGS);
                    200: static int              roff_eo(ROFF_ARGS);
                    201: static int              roff_eqndelim(struct roff *, struct buf *, int);
1.372     schwarze  202: static int              roff_evalcond(struct roff *, int, char *, int *);
1.234     kristaps  203: static int              roff_evalnum(struct roff *, int,
                    204:                                const char *, int *, int *, int);
                    205: static int              roff_evalpar(struct roff *, int,
1.261     schwarze  206:                                const char *, int *, int *, int);
1.198     schwarze  207: static int              roff_evalstrcond(const char *, int *);
1.355     schwarze  208: static int              roff_expand(struct roff *, struct buf *,
                    209:                                int, int, char);
1.388     schwarze  210: static void             roff_expand_patch(struct buf *, int,
                    211:                                const char *, int);
1.155     kristaps  212: static void             roff_free1(struct roff *);
1.180     schwarze  213: static void             roff_freereg(struct roffreg *);
1.167     kristaps  214: static void             roff_freestr(struct roffkv *);
1.212     schwarze  215: static size_t           roff_getname(struct roff *, char **, int, int);
1.261     schwarze  216: static int              roff_getnum(const char *, int *, int *, int);
1.184     schwarze  217: static int              roff_getop(const char *, int *, char *);
1.327     schwarze  218: static int              roff_getregn(struct roff *,
                    219:                                const char *, size_t, char);
1.273     schwarze  220: static int              roff_getregro(const struct roff *,
                    221:                                const char *name);
1.325     schwarze  222: static const char      *roff_getstrn(struct roff *,
1.315     schwarze  223:                                const char *, size_t, int *);
1.271     schwarze  224: static int              roff_hasregn(const struct roff *,
                    225:                                const char *, size_t);
1.340     schwarze  226: static int              roff_insec(ROFF_ARGS);
                    227: static int              roff_it(ROFF_ARGS);
                    228: static int              roff_line_ignore(ROFF_ARGS);
1.265     schwarze  229: static void             roff_man_alloc1(struct roff_man *);
                    230: static void             roff_man_free1(struct roff_man *);
1.340     schwarze  231: static int              roff_manyarg(ROFF_ARGS);
1.384     schwarze  232: static int              roff_mc(ROFF_ARGS);
1.358     schwarze  233: static int              roff_noarg(ROFF_ARGS);
1.340     schwarze  234: static int              roff_nop(ROFF_ARGS);
                    235: static int              roff_nr(ROFF_ARGS);
                    236: static int              roff_onearg(ROFF_ARGS);
1.294     schwarze  237: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.214     schwarze  238:                                int, int);
1.387     schwarze  239: static int              roff_parse_comment(struct roff *, struct buf *,
                    240:                                int, int, char);
1.340     schwarze  241: static int              roff_parsetext(struct roff *, struct buf *,
1.305     schwarze  242:                                int, int *);
1.340     schwarze  243: static int              roff_renamed(ROFF_ARGS);
1.386     schwarze  244: static int              roff_req_or_macro(ROFF_ARGS);
1.340     schwarze  245: static int              roff_return(ROFF_ARGS);
                    246: static int              roff_rm(ROFF_ARGS);
                    247: static int              roff_rn(ROFF_ARGS);
                    248: static int              roff_rr(ROFF_ARGS);
1.326     schwarze  249: static void             roff_setregn(struct roff *, const char *,
1.327     schwarze  250:                                size_t, int, char, int);
1.94      kristaps  251: static void             roff_setstr(struct roff *,
1.106     kristaps  252:                                const char *, const char *, int);
1.207     schwarze  253: static void             roff_setstrn(struct roffkv **, const char *,
1.164     kristaps  254:                                size_t, const char *, size_t, int);
1.340     schwarze  255: static int              roff_shift(ROFF_ARGS);
                    256: static int              roff_so(ROFF_ARGS);
                    257: static int              roff_tr(ROFF_ARGS);
                    258: static int              roff_Dd(ROFF_ARGS);
                    259: static int              roff_TE(ROFF_ARGS);
                    260: static int              roff_TS(ROFF_ARGS);
                    261: static int              roff_EQ(ROFF_ARGS);
                    262: static int              roff_EN(ROFF_ARGS);
                    263: static int              roff_T_(ROFF_ARGS);
                    264: static int              roff_unsupp(ROFF_ARGS);
                    265: static int              roff_userdef(ROFF_ARGS);
1.67      kristaps  266:
1.266     schwarze  267: /* --- constant data ------------------------------------------------------ */
                    268:
1.261     schwarze  269: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    270: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    271:
1.294     schwarze  272: const char *__roff_name[MAN_MAX + 1] = {
1.358     schwarze  273:        "br",           "ce",           "fi",           "ft",
                    274:        "ll",           "mc",           "nf",
                    275:        "po",           "rj",           "sp",
1.312     schwarze  276:        "ta",           "ti",           NULL,
1.294     schwarze  277:        "ab",           "ad",           "af",           "aln",
                    278:        "als",          "am",           "am1",          "ami",
                    279:        "ami1",         "as",           "as1",          "asciify",
                    280:        "backtrace",    "bd",           "bleedat",      "blm",
                    281:         "box",         "boxa",         "bp",           "BP",
                    282:        "break",        "breakchar",    "brnl",         "brp",
1.305     schwarze  283:        "brpnl",        "c2",           "cc",
1.294     schwarze  284:        "cf",           "cflags",       "ch",           "char",
                    285:        "chop",         "class",        "close",        "CL",
                    286:        "color",        "composite",    "continue",     "cp",
                    287:        "cropat",       "cs",           "cu",           "da",
                    288:        "dch",          "Dd",           "de",           "de1",
                    289:        "defcolor",     "dei",          "dei1",         "device",
                    290:        "devicem",      "di",           "do",           "ds",
                    291:        "ds1",          "dwh",          "dt",           "ec",
                    292:        "ecr",          "ecs",          "el",           "em",
                    293:        "EN",           "eo",           "EP",           "EQ",
                    294:        "errprint",     "ev",           "evc",          "ex",
                    295:        "fallback",     "fam",          "fc",           "fchar",
                    296:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    297:        "fl",           "flig",         "fp",           "fps",
                    298:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    299:        "fzoom",        "gcolor",       "hc",           "hcode",
                    300:        "hidechar",     "hla",          "hlm",          "hpf",
                    301:        "hpfa",         "hpfcode",      "hw",           "hy",
                    302:        "hylang",       "hylen",        "hym",          "hypp",
                    303:        "hys",          "ie",           "if",           "ig",
                    304:        "index",        "it",           "itc",          "IX",
                    305:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    306:        "lc",           "lc_ctype",     "lds",          "length",
                    307:        "letadj",       "lf",           "lg",           "lhang",
                    308:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.304     schwarze  309:        "ls",           "lsm",          "lt",
1.294     schwarze  310:        "mediasize",    "minss",        "mk",           "mso",
                    311:        "na",           "ne",           "nh",           "nhychar",
                    312:        "nm",           "nn",           "nop",          "nr",
                    313:        "nrf",          "nroff",        "ns",           "nx",
                    314:        "open",         "opena",        "os",           "output",
                    315:        "padj",         "papersize",    "pc",           "pev",
                    316:        "pi",           "PI",           "pl",           "pm",
1.312     schwarze  317:        "pn",           "pnr",          "ps",
1.294     schwarze  318:        "psbb",         "pshape",       "pso",          "ptr",
                    319:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.309     schwarze  320:        "return",       "rfschar",      "rhang",
1.294     schwarze  321:        "rm",           "rn",           "rnn",          "rr",
                    322:        "rs",           "rt",           "schar",        "sentchar",
                    323:        "shc",          "shift",        "sizes",        "so",
                    324:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    325:        "sty",          "substring",    "sv",           "sy",
1.300     schwarze  326:        "T&",           "tc",           "TE",
1.301     schwarze  327:        "TH",           "tkf",          "tl",
1.294     schwarze  328:        "tm",           "tm1",          "tmc",          "tr",
                    329:        "track",        "transchar",    "trf",          "trimat",
                    330:        "trin",         "trnt",         "troff",        "TS",
                    331:        "uf",           "ul",           "unformat",     "unwatch",
                    332:        "unwatchn",     "vpt",          "vs",           "warn",
                    333:        "warnscale",    "watch",        "watchlength",  "watchn",
                    334:        "wh",           "while",        "write",        "writec",
                    335:        "writem",       "xflag",        ".",            NULL,
1.306     schwarze  336:        NULL,           "text",
1.294     schwarze  337:        "Dd",           "Dt",           "Os",           "Sh",
                    338:        "Ss",           "Pp",           "D1",           "Dl",
                    339:        "Bd",           "Ed",           "Bl",           "El",
                    340:        "It",           "Ad",           "An",           "Ap",
                    341:        "Ar",           "Cd",           "Cm",           "Dv",
                    342:        "Er",           "Ev",           "Ex",           "Fa",
                    343:        "Fd",           "Fl",           "Fn",           "Ft",
                    344:        "Ic",           "In",           "Li",           "Nd",
                    345:        "Nm",           "Op",           "Ot",           "Pa",
                    346:        "Rv",           "St",           "Va",           "Vt",
                    347:        "Xr",           "%A",           "%B",           "%D",
                    348:        "%I",           "%J",           "%N",           "%O",
                    349:        "%P",           "%R",           "%T",           "%V",
                    350:        "Ac",           "Ao",           "Aq",           "At",
                    351:        "Bc",           "Bf",           "Bo",           "Bq",
                    352:        "Bsx",          "Bx",           "Db",           "Dc",
                    353:        "Do",           "Dq",           "Ec",           "Ef",
                    354:        "Em",           "Eo",           "Fx",           "Ms",
                    355:        "No",           "Ns",           "Nx",           "Ox",
                    356:        "Pc",           "Pf",           "Po",           "Pq",
                    357:        "Qc",           "Ql",           "Qo",           "Qq",
                    358:        "Re",           "Rs",           "Sc",           "So",
                    359:        "Sq",           "Sm",           "Sx",           "Sy",
                    360:        "Tn",           "Ux",           "Xc",           "Xo",
                    361:        "Fo",           "Fc",           "Oo",           "Oc",
                    362:        "Bk",           "Ek",           "Bt",           "Hf",
                    363:        "Fr",           "Ud",           "Lb",           "Lp",
                    364:        "Lk",           "Mt",           "Brq",          "Bro",
                    365:        "Brc",          "%C",           "Es",           "En",
1.299     schwarze  366:        "Dx",           "%Q",           "%U",           "Ta",
1.369     schwarze  367:        "Tg",           NULL,
1.294     schwarze  368:        "TH",           "SH",           "SS",           "TP",
1.332     schwarze  369:        "TQ",
1.294     schwarze  370:        "LP",           "PP",           "P",            "IP",
                    371:        "HP",           "SM",           "SB",           "BI",
                    372:        "IB",           "BR",           "RB",           "R",
                    373:        "B",            "I",            "IR",           "RI",
                    374:        "RE",           "RS",           "DT",           "UC",
1.297     schwarze  375:        "PD",           "AT",           "in",
1.333     schwarze  376:        "SY",           "YS",           "OP",
                    377:        "EX",           "EE",           "UR",
1.317     schwarze  378:        "UE",           "MT",           "ME",           NULL
1.294     schwarze  379: };
                    380: const  char *const *roff_name = __roff_name;
                    381:
                    382: static struct roffmac   roffs[TOKEN_NONE] = {
1.358     schwarze  383:        { roff_noarg, NULL, NULL, 0 },  /* br */
1.305     schwarze  384:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.358     schwarze  385:        { roff_noarg, NULL, NULL, 0 },  /* fi */
1.297     schwarze  386:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.298     schwarze  387:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.384     schwarze  388:        { roff_mc, NULL, NULL, 0 },  /* mc */
1.358     schwarze  389:        { roff_noarg, NULL, NULL, 0 },  /* nf */
1.312     schwarze  390:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.309     schwarze  391:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.299     schwarze  392:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.300     schwarze  393:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.301     schwarze  394:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.296     schwarze  395:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.295     schwarze  396:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    397:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    398:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.311     schwarze  400:        { roff_als, NULL, NULL, 0 },  /* als */
1.295     schwarze  401:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    402:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    403:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    404:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    405:        { roff_ds, NULL, NULL, 0 },  /* as */
                    406:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    407:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    408:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    409:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    410:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    411:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    412:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    413:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    414:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    415:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
1.364     schwarze  416:        { roff_break, NULL, NULL, 0 },  /* break */
1.295     schwarze  417:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.358     schwarze  419:        { roff_noarg, NULL, NULL, 0 },  /* brp */
1.295     schwarze  420:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    421:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    422:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    423:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    424:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    425:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
1.341     schwarze  426:        { roff_char, NULL, NULL, 0 },  /* char */
1.295     schwarze  427:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    429:        { roff_insec, NULL, NULL, 0 },  /* close */
                    430:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    431:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    432:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    433:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    436:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    437:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    438:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    439:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    440:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    441:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    442:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    443:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    444:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    445:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    446:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    447:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    448:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    449:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    450:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    451:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    452:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    453:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.303     schwarze  454:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.295     schwarze  455:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    456:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    457:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    458:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    459:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.303     schwarze  460:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.295     schwarze  461:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    462:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    463:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    464:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    465:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    466:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    467:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    468:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    469:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    470:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    471:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    472:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    474:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    475:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    479:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    487:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    488:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    489:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    491:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    492:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    493:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    494:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    495:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    496:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    497:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    500:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    501:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    502:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    503:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    504:        { roff_it, NULL, NULL, 0 },  /* it */
                    505:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    506:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    508:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    511:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    512:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    513:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    514:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    516:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    517:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    518:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    519:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    520:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    521:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    522:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    523:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    524:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    526:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    529:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    530:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    533:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    534:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    535:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
1.330     schwarze  536:        { roff_nop, NULL, NULL, 0 },  /* nop */
1.295     schwarze  537:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    538:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    539:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    540:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    541:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    542:        { roff_insec, NULL, NULL, 0 },  /* open */
                    543:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    545:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    549:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    550:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    551:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    552:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    553:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    554:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    555:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    557:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    558:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    559:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    560:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    561:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    562:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    563:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
1.339     schwarze  565:        { roff_return, NULL, NULL, 0 },  /* return */
1.295     schwarze  566:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    568:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.306     schwarze  569:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.295     schwarze  570:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    571:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    572:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    573:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    574:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    575:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    576:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
1.339     schwarze  577:        { roff_shift, NULL, NULL, 0 },  /* shift */
1.295     schwarze  578:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    579:        { roff_so, NULL, NULL, 0 },  /* so */
                    580:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    581:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    582:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    583:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    584:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    585:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    586:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    587:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    588:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    589:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    590:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.315     schwarze  591:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.295     schwarze  592:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    593:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    594:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    595:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    596:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    597:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    598:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    599:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    600:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    601:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    602:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    603:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    604:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    605:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    606:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    607:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    608:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    609:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    610:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    611:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    612:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    613:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    614:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    615:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    616:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    617:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    618:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
1.340     schwarze  619:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /*while*/
1.295     schwarze  620:        { roff_insec, NULL, NULL, 0 },  /* write */
                    621:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    622:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    623:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    624:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.306     schwarze  625:        { roff_renamed, NULL, NULL, 0 },
1.295     schwarze  626:        { roff_userdef, NULL, NULL, 0 }
1.67      kristaps  627: };
                    628:
1.141     kristaps  629: /* Array of injected predefined strings. */
                    630: #define        PREDEFS_MAX      38
                    631: static const struct predef predefs[PREDEFS_MAX] = {
                    632: #include "predefs.in"
                    633: };
                    634:
1.305     schwarze  635: static int      roffce_lines;  /* number of input lines to center */
                    636: static struct roff_node *roffce_node;  /* active request */
1.178     schwarze  637: static int      roffit_lines;  /* number of lines to delay */
                    638: static char    *roffit_macro;  /* nil-terminated macro line */
                    639:
1.207     schwarze  640:
1.266     schwarze  641: /* --- request table ------------------------------------------------------ */
                    642:
1.295     schwarze  643: struct ohash *
                    644: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85      kristaps  645: {
1.295     schwarze  646:        struct ohash    *htab;
                    647:        struct roffreq  *req;
                    648:        enum roff_tok    tok;
                    649:        size_t           sz;
                    650:        unsigned int     slot;
1.85      kristaps  651:
1.295     schwarze  652:        htab = mandoc_malloc(sizeof(*htab));
                    653:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85      kristaps  654:
1.295     schwarze  655:        for (tok = mintok; tok < maxtok; tok++) {
1.296     schwarze  656:                if (roff_name[tok] == NULL)
                    657:                        continue;
1.295     schwarze  658:                sz = strlen(roff_name[tok]);
                    659:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    660:                req->tok = tok;
                    661:                memcpy(req->name, roff_name[tok], sz + 1);
                    662:                slot = ohash_qlookup(htab, req->name);
                    663:                ohash_insert(htab, slot, req);
1.85      kristaps  664:        }
1.295     schwarze  665:        return htab;
1.85      kristaps  666: }
                    667:
1.295     schwarze  668: void
                    669: roffhash_free(struct ohash *htab)
1.67      kristaps  670: {
1.295     schwarze  671:        struct roffreq  *req;
                    672:        unsigned int     slot;
1.67      kristaps  673:
1.295     schwarze  674:        if (htab == NULL)
                    675:                return;
                    676:        for (req = ohash_first(htab, &slot); req != NULL;
                    677:             req = ohash_next(htab, &slot))
                    678:                free(req);
                    679:        ohash_delete(htab);
                    680:        free(htab);
                    681: }
                    682:
                    683: enum roff_tok
                    684: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    685: {
                    686:        struct roffreq  *req;
                    687:        const char      *end;
                    688:
                    689:        if (sz) {
                    690:                end = name + sz;
                    691:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    692:        } else
                    693:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    694:        return req == NULL ? TOKEN_NONE : req->tok;
1.67      kristaps  695: }
                    696:
1.266     schwarze  697: /* --- stack of request blocks -------------------------------------------- */
                    698:
1.67      kristaps  699: /*
                    700:  * Pop the current node off of the stack of roff instructions currently
1.364     schwarze  701:  * pending.  Return 1 if it is a loop or 0 otherwise.
1.67      kristaps  702:  */
1.340     schwarze  703: static int
1.67      kristaps  704: roffnode_pop(struct roff *r)
                    705: {
                    706:        struct roffnode *p;
1.340     schwarze  707:        int              inloop;
1.67      kristaps  708:
1.207     schwarze  709:        p = r->last;
1.340     schwarze  710:        inloop = p->tok == ROFF_while;
                    711:        r->last = p->parent;
1.106     kristaps  712:        free(p->name);
                    713:        free(p->end);
1.67      kristaps  714:        free(p);
1.340     schwarze  715:        return inloop;
1.67      kristaps  716: }
                    717:
                    718: /*
                    719:  * Push a roff node onto the instruction stack.  This must later be
                    720:  * removed with roffnode_pop().
                    721:  */
1.98      schwarze  722: static void
1.294     schwarze  723: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106     kristaps  724:                int line, int col)
1.67      kristaps  725: {
                    726:        struct roffnode *p;
                    727:
1.98      schwarze  728:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.67      kristaps  729:        p->tok = tok;
1.106     kristaps  730:        if (name)
                    731:                p->name = mandoc_strdup(name);
1.67      kristaps  732:        p->parent = r->last;
                    733:        p->line = line;
                    734:        p->col = col;
1.198     schwarze  735:        p->rule = p->parent ? p->parent->rule : 0;
1.67      kristaps  736:
                    737:        r->last = p;
                    738: }
                    739:
1.266     schwarze  740: /* --- roff parser state data management ---------------------------------- */
                    741:
1.67      kristaps  742: static void
                    743: roff_free1(struct roff *r)
                    744: {
1.167     kristaps  745:        int              i;
1.67      kristaps  746:
1.346     schwarze  747:        tbl_free(r->first_tbl);
1.113     kristaps  748:        r->first_tbl = r->last_tbl = r->tbl = NULL;
                    749:
1.347     schwarze  750:        eqn_free(r->last_eqn);
1.319     schwarze  751:        r->last_eqn = r->eqn = NULL;
1.125     kristaps  752:
1.339     schwarze  753:        while (r->mstackpos >= 0)
                    754:                roff_userret(r);
                    755:
1.67      kristaps  756:        while (r->last)
                    757:                roffnode_pop(r);
1.109     kristaps  758:
1.223     schwarze  759:        free (r->rstack);
                    760:        r->rstack = NULL;
                    761:        r->rstacksz = 0;
                    762:        r->rstackpos = -1;
                    763:
                    764:        roff_freereg(r->regtab);
                    765:        r->regtab = NULL;
                    766:
1.167     kristaps  767:        roff_freestr(r->strtab);
1.306     schwarze  768:        roff_freestr(r->rentab);
1.167     kristaps  769:        roff_freestr(r->xmbtab);
1.306     schwarze  770:        r->strtab = r->rentab = r->xmbtab = NULL;
1.167     kristaps  771:
                    772:        if (r->xtab)
                    773:                for (i = 0; i < 128; i++)
                    774:                        free(r->xtab[i].p);
                    775:        free(r->xtab);
                    776:        r->xtab = NULL;
1.67      kristaps  777: }
                    778:
                    779: void
                    780: roff_reset(struct roff *r)
                    781: {
                    782:        roff_free1(r);
1.367     schwarze  783:        r->options |= MPARSE_COMMENT;
1.227     schwarze  784:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303     schwarze  785:        r->control = '\0';
                    786:        r->escape = '\\';
1.307     schwarze  787:        roffce_lines = 0;
                    788:        roffce_node = NULL;
                    789:        roffit_lines = 0;
                    790:        roffit_macro = NULL;
1.67      kristaps  791: }
                    792:
                    793: void
                    794: roff_free(struct roff *r)
                    795: {
1.366     schwarze  796:        int              i;
1.339     schwarze  797:
1.67      kristaps  798:        roff_free1(r);
1.339     schwarze  799:        for (i = 0; i < r->mstacksz; i++)
                    800:                free(r->mstack[i].argv);
                    801:        free(r->mstack);
1.295     schwarze  802:        roffhash_free(r->reqtab);
1.67      kristaps  803:        free(r);
                    804: }
                    805:
                    806: struct roff *
1.351     schwarze  807: roff_alloc(int options)
1.67      kristaps  808: {
                    809:        struct roff     *r;
                    810:
1.98      schwarze  811:        r = mandoc_calloc(1, sizeof(struct roff));
1.328     schwarze  812:        r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.367     schwarze  813:        r->options = options | MPARSE_COMMENT;
1.227     schwarze  814:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.339     schwarze  815:        r->mstackpos = -1;
1.82      kristaps  816:        r->rstackpos = -1;
1.303     schwarze  817:        r->escape = '\\';
1.277     schwarze  818:        return r;
1.265     schwarze  819: }
                    820:
1.266     schwarze  821: /* --- syntax tree state data management ---------------------------------- */
                    822:
1.265     schwarze  823: static void
                    824: roff_man_free1(struct roff_man *man)
                    825: {
1.356     schwarze  826:        if (man->meta.first != NULL)
                    827:                roff_node_delete(man, man->meta.first);
1.265     schwarze  828:        free(man->meta.msec);
                    829:        free(man->meta.vol);
                    830:        free(man->meta.os);
                    831:        free(man->meta.arch);
                    832:        free(man->meta.title);
                    833:        free(man->meta.name);
                    834:        free(man->meta.date);
1.356     schwarze  835:        free(man->meta.sodest);
1.265     schwarze  836: }
                    837:
1.357     schwarze  838: void
                    839: roff_state_reset(struct roff_man *man)
                    840: {
                    841:        man->last = man->meta.first;
                    842:        man->last_es = NULL;
                    843:        man->flags = 0;
                    844:        man->lastsec = man->lastnamed = SEC_NONE;
                    845:        man->next = ROFF_NEXT_CHILD;
                    846:        roff_setreg(man->roff, "nS", 0, '=');
                    847: }
                    848:
1.265     schwarze  849: static void
                    850: roff_man_alloc1(struct roff_man *man)
                    851: {
                    852:        memset(&man->meta, 0, sizeof(man->meta));
1.356     schwarze  853:        man->meta.first = mandoc_calloc(1, sizeof(*man->meta.first));
                    854:        man->meta.first->type = ROFFT_ROOT;
                    855:        man->meta.macroset = MACROSET_NONE;
1.357     schwarze  856:        roff_state_reset(man);
1.265     schwarze  857: }
                    858:
                    859: void
                    860: roff_man_reset(struct roff_man *man)
                    861: {
                    862:        roff_man_free1(man);
                    863:        roff_man_alloc1(man);
                    864: }
                    865:
                    866: void
                    867: roff_man_free(struct roff_man *man)
                    868: {
                    869:        roff_man_free1(man);
1.380     schwarze  870:        free(man->os_r);
1.265     schwarze  871:        free(man);
                    872: }
                    873:
                    874: struct roff_man *
1.351     schwarze  875: roff_man_alloc(struct roff *roff, const char *os_s, int quick)
1.265     schwarze  876: {
                    877:        struct roff_man *man;
                    878:
                    879:        man = mandoc_calloc(1, sizeof(*man));
                    880:        man->roff = roff;
1.316     schwarze  881:        man->os_s = os_s;
1.265     schwarze  882:        man->quick = quick;
                    883:        roff_man_alloc1(man);
1.296     schwarze  884:        roff->man = man;
1.277     schwarze  885:        return man;
1.67      kristaps  886: }
                    887:
1.266     schwarze  888: /* --- syntax tree handling ----------------------------------------------- */
                    889:
                    890: struct roff_node *
                    891: roff_node_alloc(struct roff_man *man, int line, int pos,
                    892:        enum roff_type type, int tok)
                    893: {
                    894:        struct roff_node        *n;
                    895:
                    896:        n = mandoc_calloc(1, sizeof(*n));
                    897:        n->line = line;
                    898:        n->pos = pos;
                    899:        n->tok = tok;
                    900:        n->type = type;
                    901:        n->sec = man->lastsec;
                    902:
                    903:        if (man->flags & MDOC_SYNOPSIS)
1.285     schwarze  904:                n->flags |= NODE_SYNPRETTY;
1.266     schwarze  905:        else
1.285     schwarze  906:                n->flags &= ~NODE_SYNPRETTY;
1.360     schwarze  907:        if ((man->flags & (ROFF_NOFILL | ROFF_NONOFILL)) == ROFF_NOFILL)
1.359     schwarze  908:                n->flags |= NODE_NOFILL;
                    909:        else
                    910:                n->flags &= ~NODE_NOFILL;
1.266     schwarze  911:        if (man->flags & MDOC_NEWLINE)
1.285     schwarze  912:                n->flags |= NODE_LINE;
1.266     schwarze  913:        man->flags &= ~MDOC_NEWLINE;
                    914:
1.277     schwarze  915:        return n;
1.266     schwarze  916: }
                    917:
                    918: void
                    919: roff_node_append(struct roff_man *man, struct roff_node *n)
                    920: {
                    921:
                    922:        switch (man->next) {
                    923:        case ROFF_NEXT_SIBLING:
1.281     schwarze  924:                if (man->last->next != NULL) {
                    925:                        n->next = man->last->next;
                    926:                        man->last->next->prev = n;
                    927:                } else
                    928:                        man->last->parent->last = n;
1.266     schwarze  929:                man->last->next = n;
                    930:                n->prev = man->last;
                    931:                n->parent = man->last->parent;
                    932:                break;
                    933:        case ROFF_NEXT_CHILD:
1.287     schwarze  934:                if (man->last->child != NULL) {
                    935:                        n->next = man->last->child;
                    936:                        man->last->child->prev = n;
                    937:                } else
                    938:                        man->last->last = n;
1.266     schwarze  939:                man->last->child = n;
                    940:                n->parent = man->last;
                    941:                break;
                    942:        default:
                    943:                abort();
                    944:        }
1.282     schwarze  945:        man->last = n;
1.266     schwarze  946:
                    947:        switch (n->type) {
                    948:        case ROFFT_HEAD:
                    949:                n->parent->head = n;
                    950:                break;
                    951:        case ROFFT_BODY:
1.282     schwarze  952:                if (n->end != ENDBODY_NOT)
                    953:                        return;
1.266     schwarze  954:                n->parent->body = n;
                    955:                break;
                    956:        case ROFFT_TAIL:
                    957:                n->parent->tail = n;
                    958:                break;
                    959:        default:
1.282     schwarze  960:                return;
1.266     schwarze  961:        }
1.282     schwarze  962:
                    963:        /*
                    964:         * Copy over the normalised-data pointer of our parent.  Not
                    965:         * everybody has one, but copying a null pointer is fine.
                    966:         */
                    967:
                    968:        n->norm = n->parent->norm;
                    969:        assert(n->parent->type == ROFFT_BLOCK);
1.266     schwarze  970: }
                    971:
1.267     schwarze  972: void
                    973: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    974: {
                    975:        struct roff_node        *n;
                    976:
                    977:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    978:        n->string = roff_strdup(man->roff, word);
                    979:        roff_node_append(man, n);
1.286     schwarze  980:        n->flags |= NODE_VALID | NODE_ENDED;
1.267     schwarze  981:        man->next = ROFF_NEXT_SIBLING;
                    982: }
                    983:
                    984: void
                    985: roff_word_append(struct roff_man *man, const char *word)
                    986: {
                    987:        struct roff_node        *n;
                    988:        char                    *addstr, *newstr;
                    989:
                    990:        n = man->last;
                    991:        addstr = roff_strdup(man->roff, word);
                    992:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    993:        free(addstr);
                    994:        free(n->string);
                    995:        n->string = newstr;
                    996:        man->next = ROFF_NEXT_SIBLING;
1.268     schwarze  997: }
                    998:
                    999: void
                   1000: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                   1001: {
                   1002:        struct roff_node        *n;
                   1003:
                   1004:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                   1005:        roff_node_append(man, n);
                   1006:        man->next = ROFF_NEXT_CHILD;
                   1007: }
                   1008:
                   1009: struct roff_node *
                   1010: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                   1011: {
                   1012:        struct roff_node        *n;
                   1013:
                   1014:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                   1015:        roff_node_append(man, n);
                   1016:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1017:        return n;
1.267     schwarze 1018: }
                   1019:
1.266     schwarze 1020: struct roff_node *
                   1021: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                   1022: {
                   1023:        struct roff_node        *n;
                   1024:
                   1025:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                   1026:        roff_node_append(man, n);
                   1027:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1028:        return n;
1.266     schwarze 1029: }
                   1030:
                   1031: struct roff_node *
                   1032: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                   1033: {
                   1034:        struct roff_node        *n;
                   1035:
                   1036:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                   1037:        roff_node_append(man, n);
                   1038:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1039:        return n;
1.267     schwarze 1040: }
                   1041:
1.321     schwarze 1042: static void
1.346     schwarze 1043: roff_addtbl(struct roff_man *man, int line, struct tbl_node *tbl)
1.267     schwarze 1044: {
                   1045:        struct roff_node        *n;
1.346     schwarze 1046:        struct tbl_span         *span;
1.267     schwarze 1047:
1.356     schwarze 1048:        if (man->meta.macroset == MACROSET_MAN)
1.302     schwarze 1049:                man_breakscope(man, ROFF_TS);
1.321     schwarze 1050:        while ((span = tbl_span(tbl)) != NULL) {
1.346     schwarze 1051:                n = roff_node_alloc(man, line, 0, ROFFT_TBL, TOKEN_NONE);
1.321     schwarze 1052:                n->span = span;
                   1053:                roff_node_append(man, n);
                   1054:                n->flags |= NODE_VALID | NODE_ENDED;
                   1055:                man->next = ROFF_NEXT_SIBLING;
                   1056:        }
1.266     schwarze 1057: }
                   1058:
                   1059: void
                   1060: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1061: {
                   1062:
                   1063:        /* Adjust siblings. */
                   1064:
                   1065:        if (n->prev)
                   1066:                n->prev->next = n->next;
                   1067:        if (n->next)
                   1068:                n->next->prev = n->prev;
                   1069:
                   1070:        /* Adjust parent. */
                   1071:
                   1072:        if (n->parent != NULL) {
                   1073:                if (n->parent->child == n)
                   1074:                        n->parent->child = n->next;
                   1075:                if (n->parent->last == n)
                   1076:                        n->parent->last = n->prev;
                   1077:        }
                   1078:
                   1079:        /* Adjust parse point. */
                   1080:
                   1081:        if (man == NULL)
                   1082:                return;
                   1083:        if (man->last == n) {
                   1084:                if (n->prev == NULL) {
                   1085:                        man->last = n->parent;
                   1086:                        man->next = ROFF_NEXT_CHILD;
                   1087:                } else {
                   1088:                        man->last = n->prev;
                   1089:                        man->next = ROFF_NEXT_SIBLING;
                   1090:                }
                   1091:        }
1.356     schwarze 1092:        if (man->meta.first == n)
                   1093:                man->meta.first = NULL;
1.344     schwarze 1094: }
                   1095:
                   1096: void
                   1097: roff_node_relink(struct roff_man *man, struct roff_node *n)
                   1098: {
                   1099:        roff_node_unlink(man, n);
                   1100:        n->prev = n->next = NULL;
                   1101:        roff_node_append(man, n);
1.266     schwarze 1102: }
                   1103:
                   1104: void
                   1105: roff_node_free(struct roff_node *n)
                   1106: {
                   1107:
                   1108:        if (n->args != NULL)
                   1109:                mdoc_argv_free(n->args);
                   1110:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1111:                free(n->norm);
1.347     schwarze 1112:        eqn_box_free(n->eqn);
1.266     schwarze 1113:        free(n->string);
1.374     schwarze 1114:        free(n->tag);
1.266     schwarze 1115:        free(n);
                   1116: }
                   1117:
                   1118: void
                   1119: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1120: {
                   1121:
                   1122:        while (n->child != NULL)
                   1123:                roff_node_delete(man, n->child);
                   1124:        roff_node_unlink(man, n);
                   1125:        roff_node_free(n);
1.370     schwarze 1126: }
                   1127:
                   1128: int
                   1129: roff_node_transparent(struct roff_node *n)
                   1130: {
                   1131:        if (n == NULL)
                   1132:                return 0;
                   1133:        if (n->type == ROFFT_COMMENT || n->flags & NODE_NOPRT)
                   1134:                return 1;
1.373     schwarze 1135:        return roff_tok_transparent(n->tok);
                   1136: }
                   1137:
                   1138: int
                   1139: roff_tok_transparent(enum roff_tok tok)
                   1140: {
                   1141:        switch (tok) {
1.370     schwarze 1142:        case ROFF_ft:
                   1143:        case ROFF_ll:
                   1144:        case ROFF_mc:
                   1145:        case ROFF_po:
                   1146:        case ROFF_ta:
                   1147:        case MDOC_Db:
                   1148:        case MDOC_Es:
                   1149:        case MDOC_Sm:
                   1150:        case MDOC_Tg:
                   1151:        case MAN_DT:
                   1152:        case MAN_UC:
                   1153:        case MAN_PD:
                   1154:        case MAN_AT:
                   1155:                return 1;
                   1156:        default:
                   1157:                return 0;
                   1158:        }
                   1159: }
                   1160:
                   1161: struct roff_node *
                   1162: roff_node_child(struct roff_node *n)
                   1163: {
                   1164:        for (n = n->child; roff_node_transparent(n); n = n->next)
                   1165:                continue;
                   1166:        return n;
                   1167: }
                   1168:
                   1169: struct roff_node *
                   1170: roff_node_prev(struct roff_node *n)
                   1171: {
                   1172:        do {
                   1173:                n = n->prev;
                   1174:        } while (roff_node_transparent(n));
                   1175:        return n;
                   1176: }
                   1177:
                   1178: struct roff_node *
                   1179: roff_node_next(struct roff_node *n)
                   1180: {
                   1181:        do {
                   1182:                n = n->next;
                   1183:        } while (roff_node_transparent(n));
                   1184:        return n;
1.269     schwarze 1185: }
                   1186:
                   1187: void
                   1188: deroff(char **dest, const struct roff_node *n)
                   1189: {
                   1190:        char    *cp;
                   1191:        size_t   sz;
                   1192:
1.371     schwarze 1193:        if (n->string == NULL) {
1.269     schwarze 1194:                for (n = n->child; n != NULL; n = n->next)
                   1195:                        deroff(dest, n);
                   1196:                return;
                   1197:        }
                   1198:
1.288     schwarze 1199:        /* Skip leading whitespace. */
1.269     schwarze 1200:
1.288     schwarze 1201:        for (cp = n->string; *cp != '\0'; cp++) {
1.289     schwarze 1202:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1203:                    strchr(" %&0^|~", cp[1]) != NULL)
1.269     schwarze 1204:                        cp++;
1.288     schwarze 1205:                else if ( ! isspace((unsigned char)*cp))
1.269     schwarze 1206:                        break;
                   1207:        }
                   1208:
1.289     schwarze 1209:        /* Skip trailing backslash. */
                   1210:
                   1211:        sz = strlen(cp);
1.290     schwarze 1212:        if (sz > 0 && cp[sz - 1] == '\\')
1.289     schwarze 1213:                sz--;
                   1214:
1.269     schwarze 1215:        /* Skip trailing whitespace. */
                   1216:
1.289     schwarze 1217:        for (; sz; sz--)
1.269     schwarze 1218:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1219:                        break;
                   1220:
                   1221:        /* Skip empty strings. */
                   1222:
                   1223:        if (sz == 0)
                   1224:                return;
                   1225:
                   1226:        if (*dest == NULL) {
                   1227:                *dest = mandoc_strndup(cp, sz);
                   1228:                return;
                   1229:        }
                   1230:
                   1231:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1232:        free(*dest);
                   1233:        *dest = cp;
1.266     schwarze 1234: }
                   1235:
                   1236: /* --- main functions of the roff parser ---------------------------------- */
                   1237:
1.388     schwarze 1238: /*
                   1239:  * Save comments preceding the title macro, for example in order to
                   1240:  * preserve Copyright and license headers in HTML output,
                   1241:  * provide diagnostics about RCS ids and trailing whitespace in comments,
                   1242:  * then discard comments including preceding whitespace.
                   1243:  * This function also handles input line continuation.
                   1244:  */
1.387     schwarze 1245: static int
1.388     schwarze 1246: roff_parse_comment(struct roff *r, struct buf *buf, int ln, int pos, char ec)
1.387     schwarze 1247: {
                   1248:        struct roff_node *n;    /* used for header comments */
                   1249:        const char      *start; /* start of the string to process */
                   1250:        const char      *cp;    /* for RCS id parsing */
                   1251:        char            *stesc; /* start of an escape sequence ('\\') */
                   1252:        char            *ep;    /* end of comment string */
                   1253:        int              rcsid; /* kind of RCS id seen */
                   1254:
                   1255:        for (start = stesc = buf->buf + pos;; stesc++) {
1.388     schwarze 1256:                /*
                   1257:                 * XXX Ugly hack: Remove the newline character that
                   1258:                 * mparse_buf_r() appended to mark the end of input
                   1259:                 * if it is not preceded by an escape character.
                   1260:                 */
                   1261:                if (stesc[0] == '\n') {
                   1262:                        assert(stesc[1] == '\0');
                   1263:                        stesc[0] = '\0';
                   1264:                }
                   1265:
1.387     schwarze 1266:                /* The line ends without continuation or comment. */
                   1267:                if (stesc[0] == '\0')
                   1268:                        return ROFF_CONT;
                   1269:
                   1270:                /* Unescaped byte: skip it. */
1.388     schwarze 1271:                if (stesc[0] != ec)
1.387     schwarze 1272:                        continue;
                   1273:
1.388     schwarze 1274:                /*
                   1275:                 * XXX Ugly hack: Do not attempt to append another line
                   1276:                 * if the function mparse_buf_r() appended a newline
                   1277:                 * character to indicate the end of input.
                   1278:                 */
                   1279:                if (stesc[1] == '\n') {
                   1280:                        assert(stesc[2] == '\0');
                   1281:                        stesc[0] = '\0';
                   1282:                        return ROFF_CONT;
                   1283:                }
                   1284:
                   1285:                /*
                   1286:                 * An escape character at the end of an input line
                   1287:                 * requests line continuation.
                   1288:                 */
1.387     schwarze 1289:                if (stesc[1] == '\0') {
                   1290:                        stesc[0] = '\0';
                   1291:                        return ROFF_IGN | ROFF_APPEND;
                   1292:                }
                   1293:
                   1294:                /* Found a comment: process it. */
                   1295:                if (stesc[1] == '"' || stesc[1] == '#')
                   1296:                        break;
                   1297:
                   1298:                /* Escaped escape character: skip them both. */
1.388     schwarze 1299:                if (stesc[1] == ec)
1.387     schwarze 1300:                        stesc++;
                   1301:        }
                   1302:
                   1303:        /* Look for an RCS id in the comment. */
                   1304:
                   1305:        rcsid = 0;
                   1306:        if ((cp = strstr(stesc + 2, "$" "OpenBSD")) != NULL) {
                   1307:                rcsid = 1 << MANDOC_OS_OPENBSD;
                   1308:                cp += 8;
                   1309:        } else if ((cp = strstr(stesc + 2, "$" "NetBSD")) != NULL) {
                   1310:                rcsid = 1 << MANDOC_OS_NETBSD;
                   1311:                cp += 7;
                   1312:        }
                   1313:        if (cp != NULL && isalnum((unsigned char)*cp) == 0 &&
                   1314:            strchr(cp, '$') != NULL) {
                   1315:                if (r->man->meta.rcsids & rcsid)
                   1316:                        mandoc_msg(MANDOCERR_RCS_REP, ln,
                   1317:                            (int)(stesc - buf->buf) + 2, "%s", stesc + 1);
                   1318:                r->man->meta.rcsids |= rcsid;
                   1319:        }
                   1320:
                   1321:        /* Warn about trailing whitespace at the end of the comment. */
                   1322:
                   1323:        ep = strchr(stesc + 2, '\0') - 1;
                   1324:        if (*ep == '\n')
                   1325:                *ep-- = '\0';
                   1326:        if (*ep == ' ' || *ep == '\t')
                   1327:                mandoc_msg(MANDOCERR_SPACE_EOL,
                   1328:                    ln, (int)(ep - buf->buf), NULL);
                   1329:
                   1330:        /* Save comments preceding the title macro in the syntax tree. */
                   1331:
                   1332:        if (r->options & MPARSE_COMMENT) {
                   1333:                while (*ep == ' ' || *ep == '\t')
                   1334:                        ep--;
                   1335:                ep[1] = '\0';
                   1336:                n = roff_node_alloc(r->man, ln, stesc + 1 - buf->buf,
                   1337:                    ROFFT_COMMENT, TOKEN_NONE);
                   1338:                n->string = mandoc_strdup(stesc + 2);
                   1339:                roff_node_append(r->man, n);
                   1340:                n->flags |= NODE_VALID | NODE_ENDED;
                   1341:                r->man->next = ROFF_NEXT_SIBLING;
                   1342:        }
                   1343:
                   1344:        /* The comment requests line continuation. */
                   1345:
                   1346:        if (stesc[1] == '#') {
                   1347:                *stesc = '\0';
                   1348:                return ROFF_IGN | ROFF_APPEND;
                   1349:        }
                   1350:
                   1351:        /* Discard the comment including preceding whitespace. */
                   1352:
                   1353:        while (stesc > start && stesc[-1] == ' ' &&
                   1354:            (stesc == start + 1 || stesc[-2] != '\\'))
                   1355:                stesc--;
                   1356:        *stesc = '\0';
                   1357:        return ROFF_CONT;
                   1358: }
                   1359:
1.94      kristaps 1360: /*
1.355     schwarze 1361:  * In the current line, expand escape sequences that produce parsable
                   1362:  * input text.  Also check the syntax of the remaining escape sequences,
                   1363:  * which typically produce output glyphs or change formatter state.
1.154     kristaps 1364:  */
1.340     schwarze 1365: static int
1.388     schwarze 1366: roff_expand(struct roff *r, struct buf *buf, int ln, int pos, char ec)
1.94      kristaps 1367: {
1.388     schwarze 1368:        char             ubuf[24];      /* buffer to print a number */
                   1369:        struct mctx     *ctx;           /* current macro call context */
                   1370:        const char      *res;           /* the string to be pasted */
                   1371:        const char      *src;           /* source for copying */
                   1372:        char            *dst;           /* destination for copying */
                   1373:        int              iesc;          /* index of leading escape char */
                   1374:        int              inam;          /* index of the escape name */
                   1375:        int              iarg;          /* index beginning the argument */
                   1376:        int              iendarg;       /* index right after the argument */
                   1377:        int              iend;          /* index right after the sequence */
1.393     schwarze 1378:        int              isrc, idst;    /* to reduce \\ and \. in names */
1.388     schwarze 1379:        int              deftype;       /* type of definition to paste */
                   1380:        int              argi;          /* macro argument index */
                   1381:        int              quote_args;    /* true for \\$@, false for \\$* */
                   1382:        int              asz;           /* length of the replacement */
                   1383:        int              rsz;           /* length of the rest of the string */
                   1384:        int              npos;          /* position in numeric expression */
1.181     schwarze 1385:        int              expand_count;  /* to avoid infinite loops */
1.303     schwarze 1386:
1.170     schwarze 1387:        expand_count = 0;
1.388     schwarze 1388:        while (buf->buf[pos] != '\0') {
1.170     schwarze 1389:
1.388     schwarze 1390:                /*
                   1391:                 * Skip plain ASCII characters.
                   1392:                 * If we have a non-standard escape character,
                   1393:                 * escape literal backslashes because all processing in
                   1394:                 * subsequent functions uses the standard escaping rules.
                   1395:                 */
1.205     schwarze 1396:
1.388     schwarze 1397:                if (buf->buf[pos] != ec) {
                   1398:                        if (ec != ASCII_ESC && buf->buf[pos] == '\\') {
                   1399:                                roff_expand_patch(buf, pos, "\\e", pos + 1);
                   1400:                                pos++;
1.303     schwarze 1401:                        }
1.388     schwarze 1402:                        pos++;
1.205     schwarze 1403:                        continue;
1.303     schwarze 1404:                }
1.205     schwarze 1405:
1.388     schwarze 1406:                /*
                   1407:                 * Parse escape sequences,
                   1408:                 * issue diagnostic messages when appropriate,
                   1409:                 * and skip sequences that do not need expansion.
                   1410:                 * If we have a non-standard escape character, translate
                   1411:                 * it to backslashes and translate backslashes to \e.
                   1412:                 */
1.205     schwarze 1413:
1.392     schwarze 1414:                if (roff_escape(buf->buf, ln, pos, &iesc, &inam,
                   1415:                    &iarg, &iendarg, &iend) != ESCAPE_EXPAND) {
1.388     schwarze 1416:                        while (pos < iend) {
                   1417:                                if (buf->buf[pos] == ec) {
                   1418:                                        buf->buf[pos] = '\\';
                   1419:                                        if (pos + 1 < iend)
                   1420:                                                pos++;
                   1421:                                } else if (buf->buf[pos] == '\\') {
                   1422:                                        roff_expand_patch(buf,
                   1423:                                            pos, "\\e", pos + 1);
                   1424:                                        pos++;
                   1425:                                        iend++;
                   1426:                                }
                   1427:                                pos++;
1.352     schwarze 1428:                        }
1.205     schwarze 1429:                        continue;
1.152     kristaps 1430:                }
1.388     schwarze 1431:
1.393     schwarze 1432:                /* Reduce \\ and \. in names. */
                   1433:
                   1434:                if (buf->buf[inam] == '*' || buf->buf[inam] == 'n') {
                   1435:                        isrc = idst = iarg;
                   1436:                        while (isrc < iendarg) {
                   1437:                                if (isrc + 1 < iendarg &&
                   1438:                                    buf->buf[isrc] == '\\' &&
                   1439:                                    (buf->buf[isrc + 1] == '\\' ||
                   1440:                                     buf->buf[isrc + 1] == '.'))
                   1441:                                        isrc++;
                   1442:                                buf->buf[idst++] = buf->buf[isrc++];
                   1443:                        }
                   1444:                        iendarg -= isrc - idst;
                   1445:                }
                   1446:
1.388     schwarze 1447:                /* Handle expansion. */
                   1448:
                   1449:                res = NULL;
                   1450:                switch (buf->buf[inam]) {
                   1451:                case '*':
                   1452:                        if (iendarg == iarg)
1.206     schwarze 1453:                                break;
1.388     schwarze 1454:                        deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1455:                        if ((res = roff_getstrn(r, buf->buf + iarg,
                   1456:                            iendarg - iarg, &deftype)) != NULL)
1.206     schwarze 1457:                                break;
1.94      kristaps 1458:
1.388     schwarze 1459:                        /*
                   1460:                         * If not overriden,
                   1461:                         * let \*(.T through to the formatters.
                   1462:                         */
1.94      kristaps 1463:
1.388     schwarze 1464:                        if (iendarg - iarg == 2 &&
                   1465:                            buf->buf[iarg] == '.' &&
                   1466:                            buf->buf[iarg + 1] == 'T') {
                   1467:                                roff_setstrn(&r->strtab, ".T", 2, NULL, 0, 0);
                   1468:                                pos = iend;
1.253     schwarze 1469:                                continue;
                   1470:                        }
1.94      kristaps 1471:
1.388     schwarze 1472:                        mandoc_msg(MANDOCERR_STR_UNDEF, ln, iesc,
                   1473:                            "%.*s", iendarg - iarg, buf->buf + iarg);
                   1474:                        break;
1.108     schwarze 1475:
1.339     schwarze 1476:                case '$':
                   1477:                        if (r->mstackpos < 0) {
1.388     schwarze 1478:                                mandoc_msg(MANDOCERR_ARG_UNDEF, ln, iesc,
                   1479:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1480:                                break;
                   1481:                        }
                   1482:                        ctx = r->mstack + r->mstackpos;
1.388     schwarze 1483:                        argi = buf->buf[iarg] - '1';
                   1484:                        if (argi >= 0 && argi <= 8) {
                   1485:                                if (argi < ctx->argc)
                   1486:                                        res = ctx->argv[argi];
1.339     schwarze 1487:                                break;
                   1488:                        }
1.388     schwarze 1489:                        if (buf->buf[iarg] == '*')
1.339     schwarze 1490:                                quote_args = 0;
1.388     schwarze 1491:                        else if (buf->buf[iarg] == '@')
1.339     schwarze 1492:                                quote_args = 1;
                   1493:                        else {
1.388     schwarze 1494:                                mandoc_msg(MANDOCERR_ARG_NONUM, ln, iesc,
                   1495:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1496:                                break;
                   1497:                        }
                   1498:                        asz = 0;
1.388     schwarze 1499:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1500:                                if (argi)
1.339     schwarze 1501:                                        asz++;  /* blank */
                   1502:                                if (quote_args)
                   1503:                                        asz += 2;  /* quotes */
1.388     schwarze 1504:                                asz += strlen(ctx->argv[argi]);
1.339     schwarze 1505:                        }
1.388     schwarze 1506:                        if (asz != iend - iesc) {
                   1507:                                rsz = buf->sz - iend;
                   1508:                                if (asz < iend - iesc)
                   1509:                                        memmove(buf->buf + iesc + asz,
                   1510:                                            buf->buf + iend, rsz);
                   1511:                                buf->sz = iesc + asz + rsz;
                   1512:                                buf->buf = mandoc_realloc(buf->buf, buf->sz);
                   1513:                                if (asz > iend - iesc)
                   1514:                                        memmove(buf->buf + iesc + asz,
                   1515:                                            buf->buf + iend, rsz);
1.339     schwarze 1516:                        }
1.388     schwarze 1517:                        dst = buf->buf + iesc;
                   1518:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1519:                                if (argi)
                   1520:                                        *dst++ = ' ';
1.339     schwarze 1521:                                if (quote_args)
1.388     schwarze 1522:                                        *dst++ = '"';
                   1523:                                src = ctx->argv[argi];
                   1524:                                while (*src != '\0')
                   1525:                                        *dst++ = *src++;
1.339     schwarze 1526:                                if (quote_args)
1.388     schwarze 1527:                                        *dst++ = '"';
1.339     schwarze 1528:                        }
                   1529:                        continue;
1.391     schwarze 1530:                case 'A':
                   1531:                        ubuf[0] = iendarg > iarg ? '1' : '0';
                   1532:                        ubuf[1] = '\0';
                   1533:                        res = ubuf;
                   1534:                        break;
1.207     schwarze 1535:                case 'B':
1.206     schwarze 1536:                        npos = 0;
1.388     schwarze 1537:                        ubuf[0] = iendarg > iarg && iend > iendarg &&
                   1538:                            roff_evalnum(r, ln, buf->buf + iarg, &npos,
                   1539:                                         NULL, ROFFNUM_SCALE) &&
                   1540:                            npos == iendarg - iarg ? '1' : '0';
1.206     schwarze 1541:                        ubuf[1] = '\0';
1.388     schwarze 1542:                        res = ubuf;
1.206     schwarze 1543:                        break;
1.389     schwarze 1544:                case 'V':
                   1545:                        mandoc_msg(MANDOCERR_UNSUPP, ln, iesc,
                   1546:                            "%.*s", iend - iesc, buf->buf + iesc);
                   1547:                        roff_expand_patch(buf, iendarg, "}", iend);
                   1548:                        roff_expand_patch(buf, iesc, "${", iarg);
                   1549:                        continue;
1.390     schwarze 1550:                case 'g':
                   1551:                        break;
1.207     schwarze 1552:                case 'n':
1.388     schwarze 1553:                        if (iendarg > iarg)
1.218     schwarze 1554:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.388     schwarze 1555:                                    roff_getregn(r, buf->buf + iarg,
                   1556:                                    iendarg - iarg, buf->buf[inam + 1]));
1.218     schwarze 1557:                        else
                   1558:                                ubuf[0] = '\0';
1.388     schwarze 1559:                        res = ubuf;
1.206     schwarze 1560:                        break;
1.207     schwarze 1561:                case 'w':
1.388     schwarze 1562:                        (void)snprintf(ubuf, sizeof(ubuf),
                   1563:                            "%d", (iendarg - iarg) * 24);
                   1564:                        res = ubuf;
                   1565:                        break;
                   1566:                default:
1.206     schwarze 1567:                        break;
                   1568:                }
1.388     schwarze 1569:                if (res == NULL)
1.142     kristaps 1570:                        res = "";
1.388     schwarze 1571:                if (++expand_count > EXPAND_LIMIT ||
                   1572:                    buf->sz + strlen(res) > SHRT_MAX) {
                   1573:                        mandoc_msg(MANDOCERR_ROFFLOOP, ln, iesc, NULL);
1.277     schwarze 1574:                        return ROFF_IGN;
1.94      kristaps 1575:                }
1.388     schwarze 1576:                roff_expand_patch(buf, iesc, res, iend);
                   1577:        }
                   1578:        return ROFF_CONT;
                   1579: }
1.94      kristaps 1580:
1.388     schwarze 1581: /*
                   1582:  * Replace the substring from the start position (inclusive)
                   1583:  * to end position (exclusive) with the repl(acement) string.
                   1584:  */
                   1585: static void
                   1586: roff_expand_patch(struct buf *buf, int start, const char *repl, int end)
                   1587: {
                   1588:        char    *nbuf;
1.108     schwarze 1589:
1.389     schwarze 1590:        buf->sz = mandoc_asprintf(&nbuf, "%.*s%s%s", start, buf->buf,
                   1591:            repl, buf->buf + end) + 1;
1.388     schwarze 1592:        free(buf->buf);
                   1593:        buf->buf = nbuf;
1.154     kristaps 1594: }
1.353     schwarze 1595:
                   1596: /*
                   1597:  * Parse a quoted or unquoted roff-style request or macro argument.
                   1598:  * Return a pointer to the parsed argument, which is either the original
                   1599:  * pointer or advanced by one byte in case the argument is quoted.
                   1600:  * NUL-terminate the argument in place.
                   1601:  * Collapse pairs of quotes inside quoted arguments.
                   1602:  * Advance the argument pointer to the next argument,
                   1603:  * or to the NUL byte terminating the argument line.
                   1604:  */
                   1605: char *
1.355     schwarze 1606: roff_getarg(struct roff *r, char **cpp, int ln, int *pos)
1.353     schwarze 1607: {
1.355     schwarze 1608:        struct buf       buf;
1.366     schwarze 1609:        char            *cp, *start;
1.355     schwarze 1610:        int              newesc, pairs, quoted, white;
1.353     schwarze 1611:
                   1612:        /* Quoting can only start with a new word. */
                   1613:        start = *cpp;
                   1614:        quoted = 0;
                   1615:        if ('"' == *start) {
                   1616:                quoted = 1;
                   1617:                start++;
                   1618:        }
                   1619:
1.355     schwarze 1620:        newesc = pairs = white = 0;
1.353     schwarze 1621:        for (cp = start; '\0' != *cp; cp++) {
                   1622:
                   1623:                /*
                   1624:                 * Move the following text left
                   1625:                 * after quoted quotes and after "\\" and "\t".
                   1626:                 */
                   1627:                if (pairs)
                   1628:                        cp[-pairs] = cp[0];
                   1629:
                   1630:                if ('\\' == cp[0]) {
                   1631:                        /*
                   1632:                         * In copy mode, translate double to single
                   1633:                         * backslashes and backslash-t to literal tabs.
                   1634:                         */
                   1635:                        switch (cp[1]) {
                   1636:                        case 'a':
                   1637:                        case 't':
1.354     schwarze 1638:                                cp[-pairs] = '\t';
1.355     schwarze 1639:                                pairs++;
                   1640:                                cp++;
                   1641:                                break;
1.353     schwarze 1642:                        case '\\':
1.355     schwarze 1643:                                newesc = 1;
                   1644:                                cp[-pairs] = ASCII_ESC;
1.353     schwarze 1645:                                pairs++;
                   1646:                                cp++;
                   1647:                                break;
                   1648:                        case ' ':
                   1649:                                /* Skip escaped blanks. */
                   1650:                                if (0 == quoted)
                   1651:                                        cp++;
                   1652:                                break;
                   1653:                        default:
                   1654:                                break;
                   1655:                        }
                   1656:                } else if (0 == quoted) {
                   1657:                        if (' ' == cp[0]) {
                   1658:                                /* Unescaped blanks end unquoted args. */
                   1659:                                white = 1;
                   1660:                                break;
                   1661:                        }
                   1662:                } else if ('"' == cp[0]) {
                   1663:                        if ('"' == cp[1]) {
                   1664:                                /* Quoted quotes collapse. */
                   1665:                                pairs++;
                   1666:                                cp++;
                   1667:                        } else {
                   1668:                                /* Unquoted quotes end quoted args. */
                   1669:                                quoted = 2;
                   1670:                                break;
                   1671:                        }
                   1672:                }
                   1673:        }
                   1674:
                   1675:        /* Quoted argument without a closing quote. */
                   1676:        if (1 == quoted)
                   1677:                mandoc_msg(MANDOCERR_ARG_QUOTE, ln, *pos, NULL);
                   1678:
                   1679:        /* NUL-terminate this argument and move to the next one. */
                   1680:        if (pairs)
                   1681:                cp[-pairs] = '\0';
                   1682:        if ('\0' != *cp) {
                   1683:                *cp++ = '\0';
                   1684:                while (' ' == *cp)
                   1685:                        cp++;
                   1686:        }
                   1687:        *pos += (int)(cp - start) + (quoted ? 1 : 0);
                   1688:        *cpp = cp;
                   1689:
                   1690:        if ('\0' == *cp && (white || ' ' == cp[-1]))
                   1691:                mandoc_msg(MANDOCERR_SPACE_EOL, ln, *pos, NULL);
                   1692:
1.355     schwarze 1693:        start = mandoc_strdup(start);
                   1694:        if (newesc == 0)
                   1695:                return start;
                   1696:
                   1697:        buf.buf = start;
                   1698:        buf.sz = strlen(start) + 1;
                   1699:        buf.next = NULL;
                   1700:        if (roff_expand(r, &buf, ln, 0, ASCII_ESC) & ROFF_IGN) {
                   1701:                free(buf.buf);
                   1702:                buf.buf = mandoc_strdup("");
                   1703:        }
                   1704:        return buf.buf;
1.353     schwarze 1705: }
                   1706:
1.154     kristaps 1707:
                   1708: /*
1.275     schwarze 1709:  * Process text streams.
1.154     kristaps 1710:  */
1.340     schwarze 1711: static int
1.305     schwarze 1712: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154     kristaps 1713: {
                   1714:        size_t           sz;
                   1715:        const char      *start;
1.178     schwarze 1716:        char            *p;
                   1717:        int              isz;
1.154     kristaps 1718:        enum mandoc_esc  esc;
                   1719:
1.275     schwarze 1720:        /* Spring the input line trap. */
                   1721:
                   1722:        if (roffit_lines == 1) {
                   1723:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1724:                free(buf->buf);
                   1725:                buf->buf = p;
                   1726:                buf->sz = isz + 1;
                   1727:                *offs = 0;
                   1728:                free(roffit_macro);
                   1729:                roffit_lines = 0;
1.277     schwarze 1730:                return ROFF_REPARSE;
1.275     schwarze 1731:        } else if (roffit_lines > 1)
                   1732:                --roffit_lines;
                   1733:
1.305     schwarze 1734:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1735:                if (roffce_lines < 1) {
                   1736:                        r->man->last = roffce_node;
                   1737:                        r->man->next = ROFF_NEXT_SIBLING;
                   1738:                        roffce_lines = 0;
                   1739:                        roffce_node = NULL;
                   1740:                } else
                   1741:                        roffce_lines--;
                   1742:        }
                   1743:
1.275     schwarze 1744:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1745:
1.238     schwarze 1746:        start = p = buf->buf + pos;
1.154     kristaps 1747:
1.238     schwarze 1748:        while (*p != '\0') {
1.154     kristaps 1749:                sz = strcspn(p, "-\\");
                   1750:                p += sz;
                   1751:
1.238     schwarze 1752:                if (*p == '\0')
1.159     kristaps 1753:                        break;
                   1754:
1.238     schwarze 1755:                if (*p == '\\') {
1.154     kristaps 1756:                        /* Skip over escapes. */
                   1757:                        p++;
1.189     schwarze 1758:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238     schwarze 1759:                        if (esc == ESCAPE_ERROR)
1.154     kristaps 1760:                                break;
1.264     schwarze 1761:                        while (*p == '-')
                   1762:                                p++;
1.155     kristaps 1763:                        continue;
1.159     kristaps 1764:                } else if (p == start) {
1.158     kristaps 1765:                        p++;
1.155     kristaps 1766:                        continue;
1.158     kristaps 1767:                }
1.155     kristaps 1768:
1.171     schwarze 1769:                if (isalpha((unsigned char)p[-1]) &&
                   1770:                    isalpha((unsigned char)p[1]))
1.155     kristaps 1771:                        *p = ASCII_HYPH;
                   1772:                p++;
1.94      kristaps 1773:        }
1.277     schwarze 1774:        return ROFF_CONT;
1.94      kristaps 1775: }
                   1776:
1.340     schwarze 1777: int
1.377     schwarze 1778: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
1.67      kristaps 1779: {
1.294     schwarze 1780:        enum roff_tok    t;
1.340     schwarze 1781:        int              e;
1.238     schwarze 1782:        int              pos;   /* parse point */
1.243     schwarze 1783:        int              spos;  /* saved parse point for messages */
1.238     schwarze 1784:        int              ppos;  /* original offset in buf->buf */
                   1785:        int              ctl;   /* macro line (boolean) */
                   1786:
                   1787:        ppos = pos = *offs;
1.377     schwarze 1788:
                   1789:        if (len > 80 && r->tbl == NULL && r->eqn == NULL &&
                   1790:            (r->man->flags & ROFF_NOFILL) == 0 &&
                   1791:            strchr(" .\\", buf->buf[pos]) == NULL &&
                   1792:            buf->buf[pos] != r->control &&
                   1793:            strcspn(buf->buf, " ") < 80)
                   1794:                mandoc_msg(MANDOCERR_TEXT_LONG, ln, (int)len - 1,
                   1795:                    "%.20s...", buf->buf + pos);
1.79      kristaps 1796:
1.230     schwarze 1797:        /* Handle in-line equation delimiters. */
                   1798:
1.236     schwarze 1799:        if (r->tbl == NULL &&
                   1800:            r->last_eqn != NULL && r->last_eqn->delim &&
1.230     schwarze 1801:            (r->eqn == NULL || r->eqn_inline)) {
1.238     schwarze 1802:                e = roff_eqndelim(r, buf, pos);
1.230     schwarze 1803:                if (e == ROFF_REPARSE)
1.277     schwarze 1804:                        return e;
1.230     schwarze 1805:                assert(e == ROFF_CONT);
                   1806:        }
                   1807:
1.387     schwarze 1808:        /* Handle comments and escape sequences. */
                   1809:
                   1810:        e = roff_parse_comment(r, buf, ln, pos, r->escape);
                   1811:        if ((e & ROFF_MASK) == ROFF_IGN)
                   1812:                return e;
                   1813:        assert(e == ROFF_CONT);
1.94      kristaps 1814:
1.355     schwarze 1815:        e = roff_expand(r, buf, ln, pos, r->escape);
1.340     schwarze 1816:        if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1817:                return e;
1.238     schwarze 1818:        assert(e == ROFF_CONT);
1.94      kristaps 1819:
1.238     schwarze 1820:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.130     kristaps 1821:
1.94      kristaps 1822:        /*
1.79      kristaps 1823:         * First, if a scope is open and we're not a macro, pass the
1.252     schwarze 1824:         * text through the macro's filter.
                   1825:         * Equations process all content themselves.
                   1826:         * Tables process almost all content themselves, but we want
                   1827:         * to warn about macros before passing it there.
1.79      kristaps 1828:         */
1.74      kristaps 1829:
1.252     schwarze 1830:        if (r->last != NULL && ! ctl) {
1.78      kristaps 1831:                t = r->last->tok;
1.238     schwarze 1832:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.340     schwarze 1833:                if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1834:                        return e;
1.340     schwarze 1835:                e &= ~ROFF_MASK;
                   1836:        } else
                   1837:                e = ROFF_IGN;
1.319     schwarze 1838:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1839:                eqn_read(r->eqn, buf->buf + ppos);
1.340     schwarze 1840:                return e;
1.319     schwarze 1841:        }
1.321     schwarze 1842:        if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
                   1843:                tbl_read(r->tbl, ln, buf->buf, ppos);
1.346     schwarze 1844:                roff_addtbl(r->man, ln, r->tbl);
1.340     schwarze 1845:                return e;
1.321     schwarze 1846:        }
1.367     schwarze 1847:        if ( ! ctl) {
                   1848:                r->options &= ~MPARSE_COMMENT;
1.340     schwarze 1849:                return roff_parsetext(r, buf, pos, offs) | e;
1.367     schwarze 1850:        }
1.228     schwarze 1851:
                   1852:        /* Skip empty request lines. */
                   1853:
1.238     schwarze 1854:        if (buf->buf[pos] == '"') {
1.350     schwarze 1855:                mandoc_msg(MANDOCERR_COMMENT_BAD, ln, pos, NULL);
1.277     schwarze 1856:                return ROFF_IGN;
1.238     schwarze 1857:        } else if (buf->buf[pos] == '\0')
1.277     schwarze 1858:                return ROFF_IGN;
1.67      kristaps 1859:
1.79      kristaps 1860:        /*
                   1861:         * If a scope is open, go to the child handler for that macro,
                   1862:         * as it may want to preprocess before doing anything with it.
                   1863:         */
1.78      kristaps 1864:
1.79      kristaps 1865:        if (r->last) {
                   1866:                t = r->last->tok;
1.277     schwarze 1867:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1868:        }
1.78      kristaps 1869:
1.367     schwarze 1870:        r->options &= ~MPARSE_COMMENT;
1.243     schwarze 1871:        spos = pos;
                   1872:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.386     schwarze 1873:        return roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
                   1874: }
                   1875:
                   1876: /*
                   1877:  * Handle a new request or macro.
                   1878:  * May be called outside any scope or from inside a conditional scope.
                   1879:  */
                   1880: static int
                   1881: roff_req_or_macro(ROFF_ARGS) {
1.243     schwarze 1882:
1.386     schwarze 1883:        /* For now, tables ignore most macros and some request. */
1.243     schwarze 1884:
1.386     schwarze 1885:        if (r->tbl != NULL && (tok == TOKEN_NONE || tok == ROFF_TS ||
                   1886:            tok == ROFF_br || tok == ROFF_ce || tok == ROFF_rj ||
                   1887:            tok == ROFF_sp)) {
1.350     schwarze 1888:                mandoc_msg(MANDOCERR_TBLMACRO,
1.386     schwarze 1889:                    ln, ppos, "%s", buf->buf + ppos);
                   1890:                if (tok != TOKEN_NONE)
1.277     schwarze 1891:                        return ROFF_IGN;
1.257     schwarze 1892:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1893:                        pos++;
1.291     schwarze 1894:                while (buf->buf[pos] == ' ')
1.257     schwarze 1895:                        pos++;
1.321     schwarze 1896:                tbl_read(r->tbl, ln, buf->buf, pos);
1.346     schwarze 1897:                roff_addtbl(r->man, ln, r->tbl);
1.321     schwarze 1898:                return ROFF_IGN;
1.243     schwarze 1899:        }
                   1900:
1.305     schwarze 1901:        /* For now, let high level macros abort .ce mode. */
                   1902:
1.386     schwarze 1903:        if (roffce_node != NULL &&
                   1904:            (tok == TOKEN_NONE || tok == ROFF_Dd || tok == ROFF_EQ ||
                   1905:             tok == ROFF_TH || tok == ROFF_TS)) {
1.305     schwarze 1906:                r->man->last = roffce_node;
                   1907:                r->man->next = ROFF_NEXT_SIBLING;
                   1908:                roffce_lines = 0;
                   1909:                roffce_node = NULL;
                   1910:        }
                   1911:
1.79      kristaps 1912:        /*
1.243     schwarze 1913:         * This is neither a roff request nor a user-defined macro.
                   1914:         * Let the standard macro set parsers handle it.
1.79      kristaps 1915:         */
1.67      kristaps 1916:
1.386     schwarze 1917:        if (tok == TOKEN_NONE)
1.277     schwarze 1918:                return ROFF_CONT;
1.243     schwarze 1919:
1.386     schwarze 1920:        /* Execute a roff request or a user-defined macro. */
1.67      kristaps 1921:
1.386     schwarze 1922:        return (*roffs[tok].proc)(r, tok, buf, ln, ppos, pos, offs);
1.74      kristaps 1923: }
                   1924:
1.339     schwarze 1925: /*
                   1926:  * Internal interface function to tell the roff parser that execution
                   1927:  * of the current macro ended.  This is required because macro
                   1928:  * definitions usually do not end with a .return request.
                   1929:  */
                   1930: void
                   1931: roff_userret(struct roff *r)
                   1932: {
                   1933:        struct mctx     *ctx;
                   1934:        int              i;
                   1935:
                   1936:        assert(r->mstackpos >= 0);
                   1937:        ctx = r->mstack + r->mstackpos;
                   1938:        for (i = 0; i < ctx->argc; i++)
                   1939:                free(ctx->argv[i]);
                   1940:        ctx->argc = 0;
                   1941:        r->mstackpos--;
                   1942: }
                   1943:
1.117     kristaps 1944: void
1.74      kristaps 1945: roff_endparse(struct roff *r)
                   1946: {
1.321     schwarze 1947:        if (r->last != NULL)
1.350     schwarze 1948:                mandoc_msg(MANDOCERR_BLK_NOEND, r->last->line,
                   1949:                    r->last->col, "%s", roff_name[r->last->tok]);
1.117     kristaps 1950:
1.321     schwarze 1951:        if (r->eqn != NULL) {
1.350     schwarze 1952:                mandoc_msg(MANDOCERR_BLK_NOEND,
1.319     schwarze 1953:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1954:                eqn_parse(r->eqn);
                   1955:                r->eqn = NULL;
1.125     kristaps 1956:        }
                   1957:
1.321     schwarze 1958:        if (r->tbl != NULL) {
1.346     schwarze 1959:                tbl_end(r->tbl, 1);
1.321     schwarze 1960:                r->tbl = NULL;
1.117     kristaps 1961:        }
1.67      kristaps 1962: }
                   1963:
                   1964: /*
1.386     schwarze 1965:  * Parse the request or macro name at buf[*pos].
                   1966:  * Return ROFF_RENAMED, ROFF_USERDEF, or a ROFF_* token value.
                   1967:  * For empty, undefined, mdoc(7), and man(7) macros, return TOKEN_NONE.
                   1968:  * As a side effect, set r->current_string to the definition or to NULL.
1.67      kristaps 1969:  */
1.294     schwarze 1970: static enum roff_tok
1.214     schwarze 1971: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67      kristaps 1972: {
1.214     schwarze 1973:        char            *cp;
1.106     kristaps 1974:        const char      *mac;
                   1975:        size_t           maclen;
1.315     schwarze 1976:        int              deftype;
1.294     schwarze 1977:        enum roff_tok    t;
1.67      kristaps 1978:
1.214     schwarze 1979:        cp = buf + *pos;
                   1980:
                   1981:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294     schwarze 1982:                return TOKEN_NONE;
1.67      kristaps 1983:
1.214     schwarze 1984:        mac = cp;
                   1985:        maclen = roff_getname(r, &cp, ln, ppos);
1.67      kristaps 1986:
1.315     schwarze 1987:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1988:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1989:        switch (deftype) {
                   1990:        case ROFFDEF_USER:
                   1991:                t = ROFF_USERDEF;
                   1992:                break;
                   1993:        case ROFFDEF_REN:
                   1994:                t = ROFF_RENAMED;
                   1995:                break;
                   1996:        default:
                   1997:                t = roffhash_find(r->reqtab, mac, maclen);
                   1998:                break;
                   1999:        }
1.294     schwarze 2000:        if (t != TOKEN_NONE)
1.214     schwarze 2001:                *pos = cp - buf;
1.325     schwarze 2002:        else if (deftype == ROFFDEF_UNDEF) {
                   2003:                /* Using an undefined macro defines it to be empty. */
                   2004:                roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
                   2005:                roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
                   2006:        }
1.277     schwarze 2007:        return t;
1.67      kristaps 2008: }
                   2009:
1.266     schwarze 2010: /* --- handling of request blocks ----------------------------------------- */
                   2011:
1.375     schwarze 2012: /*
                   2013:  * Close a macro definition block or an "ignore" block.
                   2014:  */
1.340     schwarze 2015: static int
1.76      kristaps 2016: roff_cblock(ROFF_ARGS)
1.67      kristaps 2017: {
1.375     schwarze 2018:        int      rr;
1.79      kristaps 2019:
1.238     schwarze 2020:        if (r->last == NULL) {
1.350     schwarze 2021:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.277     schwarze 2022:                return ROFF_IGN;
1.76      kristaps 2023:        }
1.67      kristaps 2024:
1.81      kristaps 2025:        switch (r->last->tok) {
1.207     schwarze 2026:        case ROFF_am:
                   2027:        case ROFF_ami:
                   2028:        case ROFF_de:
                   2029:        case ROFF_dei:
                   2030:        case ROFF_ig:
1.81      kristaps 2031:                break;
1.375     schwarze 2032:        case ROFF_am1:
                   2033:        case ROFF_de1:
                   2034:                /* Remapped in roff_block(). */
                   2035:                abort();
1.81      kristaps 2036:        default:
1.350     schwarze 2037:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.277     schwarze 2038:                return ROFF_IGN;
1.76      kristaps 2039:        }
1.67      kristaps 2040:
1.375     schwarze 2041:        roffnode_pop(r);
                   2042:        roffnode_cleanscope(r);
                   2043:
                   2044:        /*
                   2045:         * If a conditional block with braces is still open,
                   2046:         * check for "\}" block end markers.
                   2047:         */
                   2048:
                   2049:        if (r->last != NULL && r->last->endspan < 0) {
                   2050:                rr = 1;  /* If arguments follow "\}", warn about them. */
                   2051:                roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
                   2052:        }
                   2053:
1.238     schwarze 2054:        if (buf->buf[pos] != '\0')
1.350     schwarze 2055:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 2056:                    ".. %s", buf->buf + pos);
1.71      kristaps 2057:
1.277     schwarze 2058:        return ROFF_IGN;
1.67      kristaps 2059: }
                   2060:
1.364     schwarze 2061: /*
                   2062:  * Pop all nodes ending at the end of the current input line.
                   2063:  * Return the number of loops ended.
                   2064:  */
1.340     schwarze 2065: static int
1.76      kristaps 2066: roffnode_cleanscope(struct roff *r)
1.67      kristaps 2067: {
1.340     schwarze 2068:        int inloop;
1.67      kristaps 2069:
1.340     schwarze 2070:        inloop = 0;
1.375     schwarze 2071:        while (r->last != NULL && r->last->endspan > 0) {
1.173     schwarze 2072:                if (--r->last->endspan != 0)
1.76      kristaps 2073:                        break;
1.340     schwarze 2074:                inloop += roffnode_pop(r);
1.76      kristaps 2075:        }
1.340     schwarze 2076:        return inloop;
1.67      kristaps 2077: }
                   2078:
1.364     schwarze 2079: /*
1.375     schwarze 2080:  * Handle the closing "\}" of a conditional block.
1.364     schwarze 2081:  * Apart from generating warnings, this only pops nodes.
                   2082:  * Return the number of loops ended.
                   2083:  */
1.340     schwarze 2084: static int
1.195     schwarze 2085: roff_ccond(struct roff *r, int ln, int ppos)
1.74      kristaps 2086: {
1.76      kristaps 2087:        if (NULL == r->last) {
1.350     schwarze 2088:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2089:                return 0;
1.76      kristaps 2090:        }
                   2091:
1.82      kristaps 2092:        switch (r->last->tok) {
1.207     schwarze 2093:        case ROFF_el:
                   2094:        case ROFF_ie:
                   2095:        case ROFF_if:
1.340     schwarze 2096:        case ROFF_while:
1.82      kristaps 2097:                break;
                   2098:        default:
1.350     schwarze 2099:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2100:                return 0;
1.75      kristaps 2101:        }
                   2102:
1.76      kristaps 2103:        if (r->last->endspan > -1) {
1.350     schwarze 2104:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2105:                return 0;
1.76      kristaps 2106:        }
                   2107:
1.340     schwarze 2108:        return roffnode_pop(r) + roffnode_cleanscope(r);
1.76      kristaps 2109: }
                   2110:
1.340     schwarze 2111: static int
1.80      kristaps 2112: roff_block(ROFF_ARGS)
1.76      kristaps 2113: {
1.315     schwarze 2114:        const char      *name, *value;
                   2115:        char            *call, *cp, *iname, *rname;
                   2116:        size_t           csz, namesz, rsz;
                   2117:        int              deftype;
1.106     kristaps 2118:
1.220     schwarze 2119:        /* Ignore groff compatibility mode for now. */
1.76      kristaps 2120:
1.238     schwarze 2121:        if (tok == ROFF_de1)
1.220     schwarze 2122:                tok = ROFF_de;
1.251     schwarze 2123:        else if (tok == ROFF_dei1)
                   2124:                tok = ROFF_dei;
1.238     schwarze 2125:        else if (tok == ROFF_am1)
1.220     schwarze 2126:                tok = ROFF_am;
1.251     schwarze 2127:        else if (tok == ROFF_ami1)
                   2128:                tok = ROFF_ami;
1.220     schwarze 2129:
                   2130:        /* Parse the macro name argument. */
                   2131:
1.238     schwarze 2132:        cp = buf->buf + pos;
                   2133:        if (tok == ROFF_ig) {
1.220     schwarze 2134:                iname = NULL;
                   2135:                namesz = 0;
                   2136:        } else {
                   2137:                iname = cp;
                   2138:                namesz = roff_getname(r, &cp, ln, ppos);
                   2139:                iname[namesz] = '\0';
                   2140:        }
1.107     kristaps 2141:
1.220     schwarze 2142:        /* Resolve the macro name argument if it is indirect. */
1.107     kristaps 2143:
1.238     schwarze 2144:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2145:                deftype = ROFFDEF_USER;
                   2146:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2147:                if (name == NULL) {
1.350     schwarze 2148:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2149:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2150:                            "%.*s", (int)namesz, iname);
                   2151:                        namesz = 0;
                   2152:                } else
                   2153:                        namesz = strlen(name);
                   2154:        } else
                   2155:                name = iname;
1.107     kristaps 2156:
1.238     schwarze 2157:        if (namesz == 0 && tok != ROFF_ig) {
1.350     schwarze 2158:                mandoc_msg(MANDOCERR_REQ_EMPTY,
                   2159:                    ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2160:                return ROFF_IGN;
1.220     schwarze 2161:        }
1.80      kristaps 2162:
1.106     kristaps 2163:        roffnode_push(r, tok, name, ln, ppos);
                   2164:
                   2165:        /*
                   2166:         * At the beginning of a `de' macro, clear the existing string
                   2167:         * with the same name, if there is one.  New content will be
1.193     schwarze 2168:         * appended from roff_block_text() in multiline mode.
1.106     kristaps 2169:         */
1.107     kristaps 2170:
1.311     schwarze 2171:        if (tok == ROFF_de || tok == ROFF_dei) {
1.213     schwarze 2172:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311     schwarze 2173:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315     schwarze 2174:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   2175:                deftype = ROFFDEF_ANY;
                   2176:                value = roff_getstrn(r, iname, namesz, &deftype);
                   2177:                switch (deftype) {  /* Before appending, ... */
                   2178:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   2179:                        roff_setstrn(&r->strtab, name, namesz,
                   2180:                            value, strlen(value), 0);
                   2181:                        break;
                   2182:                case ROFFDEF_REN: /* call original standard macro. */
                   2183:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2184:                            (int)strlen(value), value);
                   2185:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2186:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   2187:                        free(call);
                   2188:                        break;
                   2189:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   2190:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   2191:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   2192:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2193:                            (int)rsz, rname);
                   2194:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2195:                        free(call);
                   2196:                        free(rname);
                   2197:                        break;
                   2198:                default:
                   2199:                        break;
                   2200:                }
1.311     schwarze 2201:        }
1.76      kristaps 2202:
1.238     schwarze 2203:        if (*cp == '\0')
1.277     schwarze 2204:                return ROFF_IGN;
1.78      kristaps 2205:
1.220     schwarze 2206:        /* Get the custom end marker. */
1.107     kristaps 2207:
1.220     schwarze 2208:        iname = cp;
1.213     schwarze 2209:        namesz = roff_getname(r, &cp, ln, ppos);
1.220     schwarze 2210:
                   2211:        /* Resolve the end marker if it is indirect. */
                   2212:
1.238     schwarze 2213:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2214:                deftype = ROFFDEF_USER;
                   2215:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2216:                if (name == NULL) {
1.350     schwarze 2217:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2218:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2219:                            "%.*s", (int)namesz, iname);
                   2220:                        namesz = 0;
                   2221:                } else
                   2222:                        namesz = strlen(name);
                   2223:        } else
                   2224:                name = iname;
                   2225:
1.213     schwarze 2226:        if (namesz)
                   2227:                r->last->end = mandoc_strndup(name, namesz);
1.78      kristaps 2228:
1.238     schwarze 2229:        if (*cp != '\0')
1.350     schwarze 2230:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.295     schwarze 2231:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74      kristaps 2232:
1.277     schwarze 2233:        return ROFF_IGN;
1.78      kristaps 2234: }
                   2235:
1.340     schwarze 2236: static int
1.80      kristaps 2237: roff_block_sub(ROFF_ARGS)
1.79      kristaps 2238: {
1.294     schwarze 2239:        enum roff_tok   t;
1.79      kristaps 2240:        int             i, j;
                   2241:
                   2242:        /*
1.385     schwarze 2243:         * If a custom end marker is a user-defined or predefined macro
                   2244:         * or a request, interpret it.
1.79      kristaps 2245:         */
                   2246:
                   2247:        if (r->last->end) {
1.130     kristaps 2248:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238     schwarze 2249:                        if (buf->buf[i] != r->last->end[j])
1.79      kristaps 2250:                                break;
                   2251:
1.238     schwarze 2252:                if (r->last->end[j] == '\0' &&
                   2253:                    (buf->buf[i] == '\0' ||
                   2254:                     buf->buf[i] == ' ' ||
                   2255:                     buf->buf[i] == '\t')) {
1.79      kristaps 2256:                        roffnode_pop(r);
                   2257:                        roffnode_cleanscope(r);
                   2258:
1.238     schwarze 2259:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130     kristaps 2260:                                i++;
                   2261:
                   2262:                        pos = i;
1.238     schwarze 2263:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294     schwarze 2264:                            TOKEN_NONE)
1.277     schwarze 2265:                                return ROFF_RERUN;
                   2266:                        return ROFF_IGN;
1.79      kristaps 2267:                }
                   2268:        }
                   2269:
1.385     schwarze 2270:        /* Handle the standard end marker. */
1.79      kristaps 2271:
1.238     schwarze 2272:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.385     schwarze 2273:        if (t == ROFF_cblock)
                   2274:                return roff_cblock(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 2275:
1.385     schwarze 2276:        /* Not an end marker, so append the line to the block. */
1.79      kristaps 2277:
1.385     schwarze 2278:        if (tok != ROFF_ig)
                   2279:                roff_setstr(r, r->last->name, buf->buf + ppos, 2);
                   2280:        return ROFF_IGN;
1.79      kristaps 2281: }
                   2282:
1.340     schwarze 2283: static int
1.80      kristaps 2284: roff_block_text(ROFF_ARGS)
1.78      kristaps 2285: {
                   2286:
1.238     schwarze 2287:        if (tok != ROFF_ig)
                   2288:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106     kristaps 2289:
1.277     schwarze 2290:        return ROFF_IGN;
1.78      kristaps 2291: }
                   2292:
1.375     schwarze 2293: /*
                   2294:  * Check for a closing "\}" and handle it.
                   2295:  * In this function, the final "int *offs" argument is used for
                   2296:  * different purposes than elsewhere:
                   2297:  * Input: *offs == 0: caller wants to discard arguments following \}
                   2298:  *        *offs == 1: caller wants to preserve text following \}
                   2299:  * Output: *offs = 0: tell caller to discard input line
                   2300:  *         *offs = 1: tell caller to use input line
                   2301:  */
1.340     schwarze 2302: static int
1.375     schwarze 2303: roff_cond_checkend(ROFF_ARGS)
1.82      kristaps 2304: {
1.340     schwarze 2305:        char            *ep;
                   2306:        int              endloop, irc, rr;
1.82      kristaps 2307:
1.340     schwarze 2308:        irc = ROFF_IGN;
1.82      kristaps 2309:        rr = r->last->rule;
1.340     schwarze 2310:        endloop = tok != ROFF_while ? ROFF_IGN :
                   2311:            rr ? ROFF_LOOPCONT : ROFF_LOOPEXIT;
                   2312:        if (roffnode_cleanscope(r))
                   2313:                irc |= endloop;
1.144     kristaps 2314:
1.196     schwarze 2315:        /*
1.375     schwarze 2316:         * If "\}" occurs on a macro line without a preceding macro or
                   2317:         * a text line contains nothing else, drop the line completely.
1.196     schwarze 2318:         */
                   2319:
1.238     schwarze 2320:        ep = buf->buf + pos;
1.375     schwarze 2321:        if (ep[0] == '\\' && ep[1] == '}' && (ep[2] == '\0' || *offs == 0))
1.198     schwarze 2322:                rr = 0;
1.196     schwarze 2323:
1.343     schwarze 2324:        /*
1.375     schwarze 2325:         * The closing delimiter "\}" rewinds the conditional scope
1.343     schwarze 2326:         * but is otherwise ignored when interpreting the line.
                   2327:         */
1.144     kristaps 2328:
1.238     schwarze 2329:        while ((ep = strchr(ep, '\\')) != NULL) {
1.318     schwarze 2330:                switch (ep[1]) {
                   2331:                case '}':
1.376     schwarze 2332:                        if (ep[2] == '\0')
                   2333:                                ep[0] = '\0';
                   2334:                        else if (rr)
1.375     schwarze 2335:                                ep[1] = '&';
                   2336:                        else
                   2337:                                memmove(ep, ep + 2, strlen(ep + 2) + 1);
1.340     schwarze 2338:                        if (roff_ccond(r, ln, ep - buf->buf))
                   2339:                                irc |= endloop;
1.318     schwarze 2340:                        break;
                   2341:                case '\0':
                   2342:                        ++ep;
                   2343:                        break;
                   2344:                default:
                   2345:                        ep += 2;
                   2346:                        break;
1.197     schwarze 2347:                }
1.177     schwarze 2348:        }
1.375     schwarze 2349:        *offs = rr;
                   2350:        return irc;
                   2351: }
1.318     schwarze 2352:
1.375     schwarze 2353: /*
                   2354:  * Parse and process a request or macro line in conditional scope.
                   2355:  */
                   2356: static int
                   2357: roff_cond_sub(ROFF_ARGS)
                   2358: {
                   2359:        struct roffnode *bl;
1.386     schwarze 2360:        int              irc, rr, spos;
1.375     schwarze 2361:        enum roff_tok    t;
                   2362:
                   2363:        rr = 0;  /* If arguments follow "\}", skip them. */
                   2364:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.386     schwarze 2365:        spos = pos;
1.368     schwarze 2366:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   2367:
1.318     schwarze 2368:        /*
1.386     schwarze 2369:         * Handle requests and macros if the conditional evaluated
                   2370:         * to true or if they are structurally required.
                   2371:         * The .break request is always handled specially.
1.318     schwarze 2372:         */
                   2373:
1.364     schwarze 2374:        if (t == ROFF_break) {
                   2375:                if (irc & ROFF_LOOPMASK)
                   2376:                        irc = ROFF_IGN | ROFF_LOOPEXIT;
                   2377:                else if (rr) {
                   2378:                        for (bl = r->last; bl != NULL; bl = bl->parent) {
                   2379:                                bl->rule = 0;
                   2380:                                if (bl->tok == ROFF_while)
                   2381:                                        break;
                   2382:                        }
                   2383:                }
1.386     schwarze 2384:        } else if (rr || (t < TOKEN_NONE && roffs[t].flags & ROFFMAC_STRUCT)) {
                   2385:                irc |= roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
1.383     schwarze 2386:                if (irc & ROFF_WHILE)
                   2387:                        irc &= ~(ROFF_LOOPCONT | ROFF_LOOPEXIT);
1.386     schwarze 2388:        }
1.340     schwarze 2389:        return irc;
1.82      kristaps 2390: }
                   2391:
1.375     schwarze 2392: /*
                   2393:  * Parse and process a text line in conditional scope.
                   2394:  */
1.340     schwarze 2395: static int
1.82      kristaps 2396: roff_cond_text(ROFF_ARGS)
1.78      kristaps 2397: {
1.375     schwarze 2398:        int      irc, rr;
1.343     schwarze 2399:
1.375     schwarze 2400:        rr = 1;  /* If arguments follow "\}", preserve them. */
                   2401:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.340     schwarze 2402:        if (rr)
                   2403:                irc |= ROFF_CONT;
                   2404:        return irc;
1.74      kristaps 2405: }
                   2406:
1.266     schwarze 2407: /* --- handling of numeric and conditional expressions -------------------- */
                   2408:
1.204     schwarze 2409: /*
                   2410:  * Parse a single signed integer number.  Stop at the first non-digit.
                   2411:  * If there is at least one digit, return success and advance the
                   2412:  * parse point, else return failure and let the parse point unchanged.
                   2413:  * Ignore overflows, treat them just like the C language.
                   2414:  */
1.184     schwarze 2415: static int
1.261     schwarze 2416: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184     schwarze 2417: {
1.261     schwarze 2418:        int      myres, scaled, n, p;
1.206     schwarze 2419:
                   2420:        if (NULL == res)
                   2421:                res = &myres;
1.184     schwarze 2422:
                   2423:        p = *pos;
                   2424:        n = v[p] == '-';
1.261     schwarze 2425:        if (n || v[p] == '+')
1.184     schwarze 2426:                p++;
                   2427:
1.261     schwarze 2428:        if (flags & ROFFNUM_WHITE)
                   2429:                while (isspace((unsigned char)v[p]))
                   2430:                        p++;
                   2431:
1.184     schwarze 2432:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204     schwarze 2433:                *res = 10 * *res + v[p] - '0';
1.184     schwarze 2434:        if (p == *pos + n)
                   2435:                return 0;
                   2436:
                   2437:        if (n)
                   2438:                *res = -*res;
                   2439:
1.254     schwarze 2440:        /* Each number may be followed by one optional scaling unit. */
                   2441:
                   2442:        switch (v[p]) {
                   2443:        case 'f':
1.261     schwarze 2444:                scaled = *res * 65536;
1.254     schwarze 2445:                break;
                   2446:        case 'i':
1.261     schwarze 2447:                scaled = *res * 240;
1.254     schwarze 2448:                break;
                   2449:        case 'c':
1.261     schwarze 2450:                scaled = *res * 240 / 2.54;
1.254     schwarze 2451:                break;
                   2452:        case 'v':
                   2453:        case 'P':
1.261     schwarze 2454:                scaled = *res * 40;
1.254     schwarze 2455:                break;
                   2456:        case 'm':
                   2457:        case 'n':
1.261     schwarze 2458:                scaled = *res * 24;
1.254     schwarze 2459:                break;
                   2460:        case 'p':
1.261     schwarze 2461:                scaled = *res * 10 / 3;
1.254     schwarze 2462:                break;
                   2463:        case 'u':
1.261     schwarze 2464:                scaled = *res;
1.254     schwarze 2465:                break;
                   2466:        case 'M':
1.261     schwarze 2467:                scaled = *res * 6 / 25;
1.254     schwarze 2468:                break;
                   2469:        default:
1.261     schwarze 2470:                scaled = *res;
1.254     schwarze 2471:                p--;
                   2472:                break;
                   2473:        }
1.261     schwarze 2474:        if (flags & ROFFNUM_SCALE)
                   2475:                *res = scaled;
1.254     schwarze 2476:
                   2477:        *pos = p + 1;
1.277     schwarze 2478:        return 1;
1.184     schwarze 2479: }
                   2480:
1.198     schwarze 2481: /*
                   2482:  * Evaluate a string comparison condition.
                   2483:  * The first character is the delimiter.
                   2484:  * Succeed if the string up to its second occurrence
                   2485:  * matches the string up to its third occurence.
                   2486:  * Advance the cursor after the third occurrence
                   2487:  * or lacking that, to the end of the line.
                   2488:  */
                   2489: static int
                   2490: roff_evalstrcond(const char *v, int *pos)
                   2491: {
                   2492:        const char      *s1, *s2, *s3;
                   2493:        int              match;
                   2494:
                   2495:        match = 0;
                   2496:        s1 = v + *pos;          /* initial delimiter */
                   2497:        s2 = s1 + 1;            /* for scanning the first string */
                   2498:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2499:
                   2500:        if (NULL == s3)         /* found no middle delimiter */
                   2501:                goto out;
                   2502:
                   2503:        while ('\0' != *++s3) {
                   2504:                if (*s2 != *s3) {  /* mismatch */
                   2505:                        s3 = strchr(s3, *s1);
                   2506:                        break;
                   2507:                }
                   2508:                if (*s3 == *s1) {  /* found the final delimiter */
                   2509:                        match = 1;
                   2510:                        break;
                   2511:                }
                   2512:                s2++;
                   2513:        }
                   2514:
                   2515: out:
                   2516:        if (NULL == s3)
                   2517:                s3 = strchr(s2, '\0');
1.242     schwarze 2518:        else if (*s3 != '\0')
1.198     schwarze 2519:                s3++;
                   2520:        *pos = s3 - v;
1.277     schwarze 2521:        return match;
1.198     schwarze 2522: }
                   2523:
1.204     schwarze 2524: /*
                   2525:  * Evaluate an optionally negated single character, numerical,
                   2526:  * or string condition.
                   2527:  */
1.198     schwarze 2528: static int
1.271     schwarze 2529: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88      kristaps 2530: {
1.336     schwarze 2531:        const char      *start, *end;
                   2532:        char            *cp, *name;
                   2533:        size_t           sz;
                   2534:        int              deftype, len, number, savepos, istrue, wanttrue;
1.88      kristaps 2535:
1.198     schwarze 2536:        if ('!' == v[*pos]) {
                   2537:                wanttrue = 0;
                   2538:                (*pos)++;
                   2539:        } else
                   2540:                wanttrue = 1;
                   2541:
1.88      kristaps 2542:        switch (v[*pos]) {
1.240     schwarze 2543:        case '\0':
1.277     schwarze 2544:                return 0;
1.207     schwarze 2545:        case 'n':
                   2546:        case 'o':
1.88      kristaps 2547:                (*pos)++;
1.277     schwarze 2548:                return wanttrue;
1.207     schwarze 2549:        case 'e':
                   2550:        case 't':
1.239     schwarze 2551:        case 'v':
1.88      kristaps 2552:                (*pos)++;
1.277     schwarze 2553:                return !wanttrue;
1.336     schwarze 2554:        case 'c':
                   2555:                do {
                   2556:                        (*pos)++;
                   2557:                } while (v[*pos] == ' ');
                   2558:
                   2559:                /*
                   2560:                 * Quirk for groff compatibility:
                   2561:                 * The horizontal tab is neither available nor unavailable.
                   2562:                 */
                   2563:
                   2564:                if (v[*pos] == '\t') {
                   2565:                        (*pos)++;
                   2566:                        return 0;
                   2567:                }
                   2568:
                   2569:                /* Printable ASCII characters are available. */
                   2570:
                   2571:                if (v[*pos] != '\\') {
                   2572:                        (*pos)++;
                   2573:                        return wanttrue;
                   2574:                }
                   2575:
                   2576:                end = v + ++*pos;
                   2577:                switch (mandoc_escape(&end, &start, &len)) {
                   2578:                case ESCAPE_SPECIAL:
                   2579:                        istrue = mchars_spec2cp(start, len) != -1;
                   2580:                        break;
                   2581:                case ESCAPE_UNICODE:
                   2582:                        istrue = 1;
                   2583:                        break;
                   2584:                case ESCAPE_NUMBERED:
                   2585:                        istrue = mchars_num2char(start, len) != -1;
                   2586:                        break;
                   2587:                default:
                   2588:                        istrue = !wanttrue;
                   2589:                        break;
                   2590:                }
                   2591:                *pos = end - v;
                   2592:                return istrue == wanttrue;
1.310     schwarze 2593:        case 'd':
1.271     schwarze 2594:        case 'r':
1.310     schwarze 2595:                cp = v + *pos + 1;
                   2596:                while (*cp == ' ')
                   2597:                        cp++;
                   2598:                name = cp;
                   2599:                sz = roff_getname(r, &cp, ln, cp - v);
1.315     schwarze 2600:                if (sz == 0)
                   2601:                        istrue = 0;
                   2602:                else if (v[*pos] == 'r')
                   2603:                        istrue = roff_hasregn(r, name, sz);
                   2604:                else {
                   2605:                        deftype = ROFFDEF_ANY;
                   2606:                        roff_getstrn(r, name, sz, &deftype);
                   2607:                        istrue = !!deftype;
                   2608:                }
1.363     schwarze 2609:                *pos = (name + sz) - v;
1.310     schwarze 2610:                return istrue == wanttrue;
1.88      kristaps 2611:        default:
                   2612:                break;
                   2613:        }
                   2614:
1.241     schwarze 2615:        savepos = *pos;
1.261     schwarze 2616:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277     schwarze 2617:                return (number > 0) == wanttrue;
1.241     schwarze 2618:        else if (*pos == savepos)
1.277     schwarze 2619:                return roff_evalstrcond(v, pos) == wanttrue;
1.204     schwarze 2620:        else
1.277     schwarze 2621:                return 0;
1.88      kristaps 2622: }
                   2623:
1.340     schwarze 2624: static int
1.103     kristaps 2625: roff_line_ignore(ROFF_ARGS)
1.89      kristaps 2626: {
1.123     schwarze 2627:
1.277     schwarze 2628:        return ROFF_IGN;
1.89      kristaps 2629: }
                   2630:
1.340     schwarze 2631: static int
1.251     schwarze 2632: roff_insec(ROFF_ARGS)
                   2633: {
                   2634:
1.350     schwarze 2635:        mandoc_msg(MANDOCERR_REQ_INSEC, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2636:        return ROFF_IGN;
1.251     schwarze 2637: }
                   2638:
1.340     schwarze 2639: static int
1.251     schwarze 2640: roff_unsupp(ROFF_ARGS)
                   2641: {
                   2642:
1.350     schwarze 2643:        mandoc_msg(MANDOCERR_REQ_UNSUPP, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2644:        return ROFF_IGN;
1.251     schwarze 2645: }
                   2646:
1.340     schwarze 2647: static int
1.82      kristaps 2648: roff_cond(ROFF_ARGS)
1.74      kristaps 2649: {
1.340     schwarze 2650:        int      irc;
1.173     schwarze 2651:
                   2652:        roffnode_push(r, tok, NULL, ln, ppos);
1.74      kristaps 2653:
1.207     schwarze 2654:        /*
1.134     kristaps 2655:         * An `.el' has no conditional body: it will consume the value
                   2656:         * of the current rstack entry set in prior `ie' calls or
1.207     schwarze 2657:         * defaults to DENY.
1.134     kristaps 2658:         *
                   2659:         * If we're not an `el', however, then evaluate the conditional.
                   2660:         */
1.133     kristaps 2661:
1.238     schwarze 2662:        r->last->rule = tok == ROFF_el ?
1.207     schwarze 2663:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238     schwarze 2664:            roff_evalcond(r, ln, buf->buf, &pos);
1.77      kristaps 2665:
1.134     kristaps 2666:        /*
                   2667:         * An if-else will put the NEGATION of the current evaluated
                   2668:         * conditional into the stack of rules.
                   2669:         */
                   2670:
1.238     schwarze 2671:        if (tok == ROFF_ie) {
1.223     schwarze 2672:                if (r->rstackpos + 1 == r->rstacksz) {
                   2673:                        r->rstacksz += 16;
                   2674:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2675:                            r->rstacksz, sizeof(int));
1.134     kristaps 2676:                }
1.198     schwarze 2677:                r->rstack[++r->rstackpos] = !r->last->rule;
1.82      kristaps 2678:        }
1.88      kristaps 2679:
                   2680:        /* If the parent has false as its rule, then so do we. */
                   2681:
1.198     schwarze 2682:        if (r->last->parent && !r->last->parent->rule)
                   2683:                r->last->rule = 0;
1.88      kristaps 2684:
                   2685:        /*
1.173     schwarze 2686:         * Determine scope.
                   2687:         * If there is nothing on the line after the conditional,
                   2688:         * not even whitespace, use next-line scope.
1.340     schwarze 2689:         * Except that .while does not support next-line scope.
1.88      kristaps 2690:         */
1.74      kristaps 2691:
1.340     schwarze 2692:        if (buf->buf[pos] == '\0' && tok != ROFF_while) {
1.173     schwarze 2693:                r->last->endspan = 2;
                   2694:                goto out;
                   2695:        }
                   2696:
1.238     schwarze 2697:        while (buf->buf[pos] == ' ')
1.173     schwarze 2698:                pos++;
                   2699:
                   2700:        /* An opening brace requests multiline scope. */
1.75      kristaps 2701:
1.238     schwarze 2702:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75      kristaps 2703:                r->last->endspan = -1;
                   2704:                pos += 2;
1.272     schwarze 2705:                while (buf->buf[pos] == ' ')
                   2706:                        pos++;
1.173     schwarze 2707:                goto out;
1.207     schwarze 2708:        }
1.74      kristaps 2709:
1.77      kristaps 2710:        /*
1.173     schwarze 2711:         * Anything else following the conditional causes
                   2712:         * single-line scope.  Warn if the scope contains
                   2713:         * nothing but trailing whitespace.
1.77      kristaps 2714:         */
                   2715:
1.238     schwarze 2716:        if (buf->buf[pos] == '\0')
1.350     schwarze 2717:                mandoc_msg(MANDOCERR_COND_EMPTY,
                   2718:                    ln, ppos, "%s", roff_name[tok]);
1.77      kristaps 2719:
1.173     schwarze 2720:        r->last->endspan = 1;
1.74      kristaps 2721:
1.173     schwarze 2722: out:
1.75      kristaps 2723:        *offs = pos;
1.340     schwarze 2724:        irc = ROFF_RERUN;
                   2725:        if (tok == ROFF_while)
                   2726:                irc |= ROFF_WHILE;
                   2727:        return irc;
1.83      schwarze 2728: }
                   2729:
1.340     schwarze 2730: static int
1.92      schwarze 2731: roff_ds(ROFF_ARGS)
                   2732: {
1.212     schwarze 2733:        char            *string;
                   2734:        const char      *name;
                   2735:        size_t           namesz;
1.96      kristaps 2736:
1.251     schwarze 2737:        /* Ignore groff compatibility mode for now. */
                   2738:
                   2739:        if (tok == ROFF_ds1)
                   2740:                tok = ROFF_ds;
                   2741:        else if (tok == ROFF_as1)
                   2742:                tok = ROFF_as;
                   2743:
1.96      kristaps 2744:        /*
1.212     schwarze 2745:         * The first word is the name of the string.
                   2746:         * If it is empty or terminated by an escape sequence,
                   2747:         * abort the `ds' request without defining anything.
1.96      kristaps 2748:         */
1.92      schwarze 2749:
1.238     schwarze 2750:        name = string = buf->buf + pos;
                   2751:        if (*name == '\0')
1.277     schwarze 2752:                return ROFF_IGN;
1.92      schwarze 2753:
1.212     schwarze 2754:        namesz = roff_getname(r, &string, ln, pos);
1.363     schwarze 2755:        switch (name[namesz]) {
                   2756:        case '\\':
1.277     schwarze 2757:                return ROFF_IGN;
1.363     schwarze 2758:        case '\t':
                   2759:                string = buf->buf + pos + namesz;
                   2760:                break;
                   2761:        default:
                   2762:                break;
                   2763:        }
1.212     schwarze 2764:
                   2765:        /* Read past the initial double-quote, if any. */
1.238     schwarze 2766:        if (*string == '"')
1.92      schwarze 2767:                string++;
                   2768:
1.96      kristaps 2769:        /* The rest is the value. */
1.212     schwarze 2770:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2771:            ROFF_as == tok);
1.311     schwarze 2772:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277     schwarze 2773:        return ROFF_IGN;
1.92      schwarze 2774: }
                   2775:
1.204     schwarze 2776: /*
                   2777:  * Parse a single operator, one or two characters long.
                   2778:  * If the operator is recognized, return success and advance the
                   2779:  * parse point, else return failure and let the parse point unchanged.
                   2780:  */
                   2781: static int
                   2782: roff_getop(const char *v, int *pos, char *res)
                   2783: {
                   2784:
                   2785:        *res = v[*pos];
                   2786:
                   2787:        switch (*res) {
1.207     schwarze 2788:        case '+':
                   2789:        case '-':
                   2790:        case '*':
                   2791:        case '/':
                   2792:        case '%':
                   2793:        case '&':
                   2794:        case ':':
1.204     schwarze 2795:                break;
                   2796:        case '<':
                   2797:                switch (v[*pos + 1]) {
1.207     schwarze 2798:                case '=':
1.204     schwarze 2799:                        *res = 'l';
                   2800:                        (*pos)++;
                   2801:                        break;
1.207     schwarze 2802:                case '>':
1.204     schwarze 2803:                        *res = '!';
                   2804:                        (*pos)++;
                   2805:                        break;
1.207     schwarze 2806:                case '?':
1.204     schwarze 2807:                        *res = 'i';
                   2808:                        (*pos)++;
                   2809:                        break;
                   2810:                default:
                   2811:                        break;
                   2812:                }
                   2813:                break;
                   2814:        case '>':
                   2815:                switch (v[*pos + 1]) {
1.207     schwarze 2816:                case '=':
1.204     schwarze 2817:                        *res = 'g';
                   2818:                        (*pos)++;
                   2819:                        break;
1.207     schwarze 2820:                case '?':
1.204     schwarze 2821:                        *res = 'a';
                   2822:                        (*pos)++;
                   2823:                        break;
                   2824:                default:
                   2825:                        break;
                   2826:                }
                   2827:                break;
                   2828:        case '=':
                   2829:                if ('=' == v[*pos + 1])
                   2830:                        (*pos)++;
                   2831:                break;
                   2832:        default:
1.277     schwarze 2833:                return 0;
1.204     schwarze 2834:        }
                   2835:        (*pos)++;
                   2836:
1.277     schwarze 2837:        return *res;
1.204     schwarze 2838: }
                   2839:
                   2840: /*
                   2841:  * Evaluate either a parenthesized numeric expression
                   2842:  * or a single signed integer number.
                   2843:  */
                   2844: static int
1.235     schwarze 2845: roff_evalpar(struct roff *r, int ln,
1.261     schwarze 2846:        const char *v, int *pos, int *res, int flags)
1.204     schwarze 2847: {
                   2848:
                   2849:        if ('(' != v[*pos])
1.277     schwarze 2850:                return roff_getnum(v, pos, res, flags);
1.204     schwarze 2851:
                   2852:        (*pos)++;
1.261     schwarze 2853:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277     schwarze 2854:                return 0;
1.204     schwarze 2855:
1.206     schwarze 2856:        /*
                   2857:         * Omission of the closing parenthesis
                   2858:         * is an error in validation mode,
                   2859:         * but ignored in evaluation mode.
                   2860:         */
                   2861:
1.204     schwarze 2862:        if (')' == v[*pos])
                   2863:                (*pos)++;
1.206     schwarze 2864:        else if (NULL == res)
1.277     schwarze 2865:                return 0;
1.204     schwarze 2866:
1.277     schwarze 2867:        return 1;
1.204     schwarze 2868: }
                   2869:
                   2870: /*
                   2871:  * Evaluate a complete numeric expression.
                   2872:  * Proceed left to right, there is no concept of precedence.
                   2873:  */
                   2874: static int
1.235     schwarze 2875: roff_evalnum(struct roff *r, int ln, const char *v,
1.261     schwarze 2876:        int *pos, int *res, int flags)
1.204     schwarze 2877: {
                   2878:        int              mypos, operand2;
                   2879:        char             operator;
                   2880:
                   2881:        if (NULL == pos) {
                   2882:                mypos = 0;
                   2883:                pos = &mypos;
                   2884:        }
                   2885:
1.261     schwarze 2886:        if (flags & ROFFNUM_WHITE)
1.204     schwarze 2887:                while (isspace((unsigned char)v[*pos]))
                   2888:                        (*pos)++;
                   2889:
1.261     schwarze 2890:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277     schwarze 2891:                return 0;
1.204     schwarze 2892:
                   2893:        while (1) {
1.261     schwarze 2894:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2895:                        while (isspace((unsigned char)v[*pos]))
                   2896:                                (*pos)++;
                   2897:
                   2898:                if ( ! roff_getop(v, pos, &operator))
                   2899:                        break;
                   2900:
1.261     schwarze 2901:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2902:                        while (isspace((unsigned char)v[*pos]))
                   2903:                                (*pos)++;
                   2904:
1.261     schwarze 2905:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277     schwarze 2906:                        return 0;
1.204     schwarze 2907:
1.261     schwarze 2908:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2909:                        while (isspace((unsigned char)v[*pos]))
                   2910:                                (*pos)++;
1.206     schwarze 2911:
                   2912:                if (NULL == res)
                   2913:                        continue;
1.204     schwarze 2914:
                   2915:                switch (operator) {
1.207     schwarze 2916:                case '+':
1.204     schwarze 2917:                        *res += operand2;
                   2918:                        break;
1.207     schwarze 2919:                case '-':
1.204     schwarze 2920:                        *res -= operand2;
                   2921:                        break;
1.207     schwarze 2922:                case '*':
1.204     schwarze 2923:                        *res *= operand2;
                   2924:                        break;
1.207     schwarze 2925:                case '/':
1.244     schwarze 2926:                        if (operand2 == 0) {
1.235     schwarze 2927:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2928:                                        ln, *pos, "%s", v);
1.234     kristaps 2929:                                *res = 0;
                   2930:                                break;
                   2931:                        }
1.204     schwarze 2932:                        *res /= operand2;
                   2933:                        break;
1.207     schwarze 2934:                case '%':
1.244     schwarze 2935:                        if (operand2 == 0) {
                   2936:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2937:                                        ln, *pos, "%s", v);
1.244     schwarze 2938:                                *res = 0;
                   2939:                                break;
                   2940:                        }
1.204     schwarze 2941:                        *res %= operand2;
                   2942:                        break;
1.207     schwarze 2943:                case '<':
1.204     schwarze 2944:                        *res = *res < operand2;
                   2945:                        break;
1.207     schwarze 2946:                case '>':
1.204     schwarze 2947:                        *res = *res > operand2;
                   2948:                        break;
1.207     schwarze 2949:                case 'l':
1.204     schwarze 2950:                        *res = *res <= operand2;
                   2951:                        break;
1.207     schwarze 2952:                case 'g':
1.204     schwarze 2953:                        *res = *res >= operand2;
                   2954:                        break;
1.207     schwarze 2955:                case '=':
1.204     schwarze 2956:                        *res = *res == operand2;
                   2957:                        break;
1.207     schwarze 2958:                case '!':
1.204     schwarze 2959:                        *res = *res != operand2;
                   2960:                        break;
1.207     schwarze 2961:                case '&':
1.204     schwarze 2962:                        *res = *res && operand2;
                   2963:                        break;
1.207     schwarze 2964:                case ':':
1.204     schwarze 2965:                        *res = *res || operand2;
                   2966:                        break;
1.207     schwarze 2967:                case 'i':
1.204     schwarze 2968:                        if (operand2 < *res)
                   2969:                                *res = operand2;
                   2970:                        break;
1.207     schwarze 2971:                case 'a':
1.204     schwarze 2972:                        if (operand2 > *res)
                   2973:                                *res = operand2;
                   2974:                        break;
                   2975:                default:
                   2976:                        abort();
                   2977:                }
                   2978:        }
1.277     schwarze 2979:        return 1;
1.204     schwarze 2980: }
                   2981:
1.266     schwarze 2982: /* --- register management ------------------------------------------------ */
                   2983:
1.180     schwarze 2984: void
1.187     schwarze 2985: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147     kristaps 2986: {
1.327     schwarze 2987:        roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.326     schwarze 2988: }
                   2989:
                   2990: static void
                   2991: roff_setregn(struct roff *r, const char *name, size_t len,
1.327     schwarze 2992:     int val, char sign, int step)
1.326     schwarze 2993: {
1.180     schwarze 2994:        struct roffreg  *reg;
                   2995:
                   2996:        /* Search for an existing register with the same name. */
                   2997:        reg = r->regtab;
                   2998:
1.326     schwarze 2999:        while (reg != NULL && (reg->key.sz != len ||
                   3000:            strncmp(reg->key.p, name, len) != 0))
1.180     schwarze 3001:                reg = reg->next;
1.147     kristaps 3002:
1.180     schwarze 3003:        if (NULL == reg) {
                   3004:                /* Create a new register. */
                   3005:                reg = mandoc_malloc(sizeof(struct roffreg));
1.326     schwarze 3006:                reg->key.p = mandoc_strndup(name, len);
                   3007:                reg->key.sz = len;
1.187     schwarze 3008:                reg->val = 0;
1.327     schwarze 3009:                reg->step = 0;
1.180     schwarze 3010:                reg->next = r->regtab;
                   3011:                r->regtab = reg;
                   3012:        }
                   3013:
1.187     schwarze 3014:        if ('+' == sign)
                   3015:                reg->val += val;
                   3016:        else if ('-' == sign)
                   3017:                reg->val -= val;
                   3018:        else
                   3019:                reg->val = val;
1.327     schwarze 3020:        if (step != INT_MIN)
                   3021:                reg->step = step;
1.147     kristaps 3022: }
                   3023:
1.192     schwarze 3024: /*
                   3025:  * Handle some predefined read-only number registers.
                   3026:  * For now, return -1 if the requested register is not predefined;
                   3027:  * in case a predefined read-only register having the value -1
                   3028:  * were to turn up, another special value would have to be chosen.
                   3029:  */
                   3030: static int
1.273     schwarze 3031: roff_getregro(const struct roff *r, const char *name)
1.192     schwarze 3032: {
                   3033:
                   3034:        switch (*name) {
1.273     schwarze 3035:        case '$':  /* Number of arguments of the last macro evaluated. */
1.339     schwarze 3036:                return r->mstackpos < 0 ? 0 : r->mstack[r->mstackpos].argc;
1.207     schwarze 3037:        case 'A':  /* ASCII approximation mode is always off. */
1.277     schwarze 3038:                return 0;
1.207     schwarze 3039:        case 'g':  /* Groff compatibility mode is always on. */
1.277     schwarze 3040:                return 1;
1.207     schwarze 3041:        case 'H':  /* Fixed horizontal resolution. */
1.277     schwarze 3042:                return 24;
1.207     schwarze 3043:        case 'j':  /* Always adjust left margin only. */
1.277     schwarze 3044:                return 0;
1.207     schwarze 3045:        case 'T':  /* Some output device is always defined. */
1.277     schwarze 3046:                return 1;
1.207     schwarze 3047:        case 'V':  /* Fixed vertical resolution. */
1.277     schwarze 3048:                return 40;
1.192     schwarze 3049:        default:
1.277     schwarze 3050:                return -1;
1.192     schwarze 3051:        }
                   3052: }
                   3053:
1.181     schwarze 3054: int
1.326     schwarze 3055: roff_getreg(struct roff *r, const char *name)
1.147     kristaps 3056: {
1.327     schwarze 3057:        return roff_getregn(r, name, strlen(name), '\0');
1.181     schwarze 3058: }
                   3059:
                   3060: static int
1.327     schwarze 3061: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.181     schwarze 3062: {
                   3063:        struct roffreg  *reg;
1.192     schwarze 3064:        int              val;
                   3065:
                   3066:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3067:                val = roff_getregro(r, name + 1);
1.192     schwarze 3068:                if (-1 != val)
1.277     schwarze 3069:                        return val;
1.192     schwarze 3070:        }
1.181     schwarze 3071:
1.327     schwarze 3072:        for (reg = r->regtab; reg; reg = reg->next) {
1.181     schwarze 3073:                if (len == reg->key.sz &&
1.327     schwarze 3074:                    0 == strncmp(name, reg->key.p, len)) {
                   3075:                        switch (sign) {
                   3076:                        case '+':
                   3077:                                reg->val += reg->step;
                   3078:                                break;
                   3079:                        case '-':
                   3080:                                reg->val -= reg->step;
                   3081:                                break;
                   3082:                        default:
                   3083:                                break;
                   3084:                        }
1.277     schwarze 3085:                        return reg->val;
1.327     schwarze 3086:                }
                   3087:        }
1.271     schwarze 3088:
1.327     schwarze 3089:        roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.277     schwarze 3090:        return 0;
1.271     schwarze 3091: }
                   3092:
                   3093: static int
                   3094: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   3095: {
                   3096:        struct roffreg  *reg;
                   3097:        int              val;
                   3098:
                   3099:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3100:                val = roff_getregro(r, name + 1);
1.271     schwarze 3101:                if (-1 != val)
1.277     schwarze 3102:                        return 1;
1.271     schwarze 3103:        }
                   3104:
                   3105:        for (reg = r->regtab; reg; reg = reg->next)
                   3106:                if (len == reg->key.sz &&
                   3107:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 3108:                        return 1;
1.147     kristaps 3109:
1.277     schwarze 3110:        return 0;
1.147     kristaps 3111: }
                   3112:
1.180     schwarze 3113: static void
                   3114: roff_freereg(struct roffreg *reg)
1.147     kristaps 3115: {
1.180     schwarze 3116:        struct roffreg  *old_reg;
1.147     kristaps 3117:
1.180     schwarze 3118:        while (NULL != reg) {
                   3119:                free(reg->key.p);
                   3120:                old_reg = reg;
                   3121:                reg = reg->next;
                   3122:                free(old_reg);
                   3123:        }
1.147     kristaps 3124: }
1.92      schwarze 3125:
1.340     schwarze 3126: static int
1.89      kristaps 3127: roff_nr(ROFF_ARGS)
1.83      schwarze 3128: {
1.327     schwarze 3129:        char            *key, *val, *step;
1.212     schwarze 3130:        size_t           keysz;
1.327     schwarze 3131:        int              iv, is, len;
1.187     schwarze 3132:        char             sign;
1.89      kristaps 3133:
1.238     schwarze 3134:        key = val = buf->buf + pos;
                   3135:        if (*key == '\0')
1.277     schwarze 3136:                return ROFF_IGN;
1.212     schwarze 3137:
                   3138:        keysz = roff_getname(r, &val, ln, pos);
1.363     schwarze 3139:        if (key[keysz] == '\\' || key[keysz] == '\t')
1.277     schwarze 3140:                return ROFF_IGN;
1.89      kristaps 3141:
1.187     schwarze 3142:        sign = *val;
1.238     schwarze 3143:        if (sign == '+' || sign == '-')
1.187     schwarze 3144:                val++;
                   3145:
1.327     schwarze 3146:        len = 0;
                   3147:        if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
                   3148:                return ROFF_IGN;
                   3149:
                   3150:        step = val + len;
                   3151:        while (isspace((unsigned char)*step))
                   3152:                step++;
                   3153:        if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
                   3154:                is = INT_MIN;
1.109     kristaps 3155:
1.327     schwarze 3156:        roff_setregn(r, key, keysz, iv, sign, is);
1.277     schwarze 3157:        return ROFF_IGN;
1.203     schwarze 3158: }
                   3159:
1.340     schwarze 3160: static int
1.203     schwarze 3161: roff_rr(ROFF_ARGS)
                   3162: {
                   3163:        struct roffreg  *reg, **prev;
1.212     schwarze 3164:        char            *name, *cp;
                   3165:        size_t           namesz;
1.203     schwarze 3166:
1.238     schwarze 3167:        name = cp = buf->buf + pos;
                   3168:        if (*name == '\0')
1.277     schwarze 3169:                return ROFF_IGN;
1.212     schwarze 3170:        namesz = roff_getname(r, &cp, ln, pos);
                   3171:        name[namesz] = '\0';
1.203     schwarze 3172:
                   3173:        prev = &r->regtab;
                   3174:        while (1) {
                   3175:                reg = *prev;
1.238     schwarze 3176:                if (reg == NULL || !strcmp(name, reg->key.p))
1.203     schwarze 3177:                        break;
                   3178:                prev = &reg->next;
                   3179:        }
1.238     schwarze 3180:        if (reg != NULL) {
1.203     schwarze 3181:                *prev = reg->next;
                   3182:                free(reg->key.p);
                   3183:                free(reg);
                   3184:        }
1.277     schwarze 3185:        return ROFF_IGN;
1.122     schwarze 3186: }
                   3187:
1.266     schwarze 3188: /* --- handler functions for roff requests -------------------------------- */
                   3189:
1.340     schwarze 3190: static int
1.122     schwarze 3191: roff_rm(ROFF_ARGS)
                   3192: {
                   3193:        const char       *name;
                   3194:        char             *cp;
1.212     schwarze 3195:        size_t            namesz;
1.122     schwarze 3196:
1.238     schwarze 3197:        cp = buf->buf + pos;
                   3198:        while (*cp != '\0') {
1.212     schwarze 3199:                name = cp;
1.238     schwarze 3200:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212     schwarze 3201:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311     schwarze 3202:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.363     schwarze 3203:                if (name[namesz] == '\\' || name[namesz] == '\t')
1.212     schwarze 3204:                        break;
1.122     schwarze 3205:        }
1.277     schwarze 3206:        return ROFF_IGN;
1.178     schwarze 3207: }
                   3208:
1.340     schwarze 3209: static int
1.178     schwarze 3210: roff_it(ROFF_ARGS)
                   3211: {
                   3212:        int              iv;
                   3213:
                   3214:        /* Parse the number of lines. */
1.261     schwarze 3215:
                   3216:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.350     schwarze 3217:                mandoc_msg(MANDOCERR_IT_NONUM,
                   3218:                    ln, ppos, "%s", buf->buf + 1);
1.277     schwarze 3219:                return ROFF_IGN;
1.178     schwarze 3220:        }
                   3221:
1.262     schwarze 3222:        while (isspace((unsigned char)buf->buf[pos]))
                   3223:                pos++;
                   3224:
                   3225:        /*
                   3226:         * Arm the input line trap.
                   3227:         * Special-casing "an-trap" is an ugly workaround to cope
                   3228:         * with DocBook stupidly fiddling with man(7) internals.
                   3229:         */
1.261     schwarze 3230:
1.178     schwarze 3231:        roffit_lines = iv;
1.262     schwarze 3232:        roffit_macro = mandoc_strdup(iv != 1 ||
                   3233:            strcmp(buf->buf + pos, "an-trap") ?
                   3234:            buf->buf + pos : "br");
1.277     schwarze 3235:        return ROFF_IGN;
1.175     schwarze 3236: }
                   3237:
1.340     schwarze 3238: static int
1.175     schwarze 3239: roff_Dd(ROFF_ARGS)
                   3240: {
1.315     schwarze 3241:        int              mask;
                   3242:        enum roff_tok    t, te;
1.227     schwarze 3243:
1.315     schwarze 3244:        switch (tok) {
                   3245:        case ROFF_Dd:
                   3246:                tok = MDOC_Dd;
                   3247:                te = MDOC_MAX;
                   3248:                if (r->format == 0)
                   3249:                        r->format = MPARSE_MDOC;
                   3250:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   3251:                break;
                   3252:        case ROFF_TH:
                   3253:                tok = MAN_TH;
                   3254:                te = MAN_MAX;
                   3255:                if (r->format == 0)
                   3256:                        r->format = MPARSE_MAN;
                   3257:                mask = MPARSE_QUICK;
                   3258:                break;
                   3259:        default:
                   3260:                abort();
                   3261:        }
                   3262:        if ((r->options & mask) == 0)
                   3263:                for (t = tok; t < te; t++)
                   3264:                        roff_setstr(r, roff_name[t], NULL, 0);
1.277     schwarze 3265:        return ROFF_CONT;
1.109     kristaps 3266: }
                   3267:
1.340     schwarze 3268: static int
1.109     kristaps 3269: roff_TE(ROFF_ARGS)
                   3270: {
1.361     schwarze 3271:        r->man->flags &= ~ROFF_NONOFILL;
1.321     schwarze 3272:        if (r->tbl == NULL) {
1.350     schwarze 3273:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "TE");
1.321     schwarze 3274:                return ROFF_IGN;
                   3275:        }
1.346     schwarze 3276:        if (tbl_end(r->tbl, 0) == 0) {
1.321     schwarze 3277:                r->tbl = NULL;
1.258     schwarze 3278:                free(buf->buf);
                   3279:                buf->buf = mandoc_strdup(".sp");
                   3280:                buf->sz = 4;
1.329     schwarze 3281:                *offs = 0;
1.277     schwarze 3282:                return ROFF_REPARSE;
1.258     schwarze 3283:        }
1.321     schwarze 3284:        r->tbl = NULL;
1.277     schwarze 3285:        return ROFF_IGN;
1.112     kristaps 3286: }
                   3287:
1.340     schwarze 3288: static int
1.112     kristaps 3289: roff_T_(ROFF_ARGS)
                   3290: {
                   3291:
                   3292:        if (NULL == r->tbl)
1.350     schwarze 3293:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "T&");
1.112     kristaps 3294:        else
1.296     schwarze 3295:                tbl_restart(ln, ppos, r->tbl);
1.112     kristaps 3296:
1.277     schwarze 3297:        return ROFF_IGN;
1.109     kristaps 3298: }
                   3299:
1.230     schwarze 3300: /*
                   3301:  * Handle in-line equation delimiters.
                   3302:  */
1.340     schwarze 3303: static int
1.238     schwarze 3304: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151     kristaps 3305: {
1.233     schwarze 3306:        char            *cp1, *cp2;
                   3307:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151     kristaps 3308:
1.230     schwarze 3309:        /*
                   3310:         * Outside equations, look for an opening delimiter.
                   3311:         * If we are inside an equation, we already know it is
                   3312:         * in-line, or this function wouldn't have been called;
                   3313:         * so look for a closing delimiter.
                   3314:         */
                   3315:
1.238     schwarze 3316:        cp1 = buf->buf + pos;
1.230     schwarze 3317:        cp2 = strchr(cp1, r->eqn == NULL ?
                   3318:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   3319:        if (cp2 == NULL)
1.277     schwarze 3320:                return ROFF_CONT;
1.230     schwarze 3321:
1.233     schwarze 3322:        *cp2++ = '\0';
                   3323:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   3324:
                   3325:        /* Handle preceding text, protecting whitespace. */
                   3326:
1.238     schwarze 3327:        if (*buf->buf != '\0') {
1.233     schwarze 3328:                if (r->eqn == NULL)
                   3329:                        bef_pr = "\\&";
                   3330:                bef_nl = "\n";
                   3331:        }
                   3332:
                   3333:        /*
                   3334:         * Prepare replacing the delimiter with an equation macro
                   3335:         * and drop leading white space from the equation.
                   3336:         */
1.230     schwarze 3337:
1.233     schwarze 3338:        if (r->eqn == NULL) {
                   3339:                while (*cp2 == ' ')
                   3340:                        cp2++;
                   3341:                mac = ".EQ";
                   3342:        } else
                   3343:                mac = ".EN";
                   3344:
                   3345:        /* Handle following text, protecting whitespace. */
                   3346:
                   3347:        if (*cp2 != '\0') {
                   3348:                aft_nl = "\n";
                   3349:                if (r->eqn != NULL)
                   3350:                        aft_pr = "\\&";
                   3351:        }
                   3352:
                   3353:        /* Do the actual replacement. */
                   3354:
1.238     schwarze 3355:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233     schwarze 3356:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238     schwarze 3357:        free(buf->buf);
                   3358:        buf->buf = cp1;
1.230     schwarze 3359:
                   3360:        /* Toggle the in-line state of the eqn subsystem. */
                   3361:
                   3362:        r->eqn_inline = r->eqn == NULL;
1.277     schwarze 3363:        return ROFF_REPARSE;
1.151     kristaps 3364: }
                   3365:
1.340     schwarze 3366: static int
1.235     schwarze 3367: roff_EQ(ROFF_ARGS)
1.125     kristaps 3368: {
1.319     schwarze 3369:        struct roff_node        *n;
                   3370:
1.356     schwarze 3371:        if (r->man->meta.macroset == MACROSET_MAN)
1.322     schwarze 3372:                man_breakscope(r->man, ROFF_EQ);
1.319     schwarze 3373:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   3374:        if (ln > r->man->last->line)
                   3375:                n->flags |= NODE_LINE;
1.348     schwarze 3376:        n->eqn = eqn_box_new();
1.319     schwarze 3377:        roff_node_append(r->man, n);
                   3378:        r->man->next = ROFF_NEXT_SIBLING;
1.125     kristaps 3379:
1.238     schwarze 3380:        assert(r->eqn == NULL);
1.319     schwarze 3381:        if (r->last_eqn == NULL)
1.351     schwarze 3382:                r->last_eqn = eqn_alloc();
1.319     schwarze 3383:        else
                   3384:                eqn_reset(r->last_eqn);
                   3385:        r->eqn = r->last_eqn;
                   3386:        r->eqn->node = n;
1.151     kristaps 3387:
1.238     schwarze 3388:        if (buf->buf[pos] != '\0')
1.350     schwarze 3389:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 3390:                    ".EQ %s", buf->buf + pos);
1.151     kristaps 3391:
1.277     schwarze 3392:        return ROFF_IGN;
1.125     kristaps 3393: }
                   3394:
1.340     schwarze 3395: static int
1.125     kristaps 3396: roff_EN(ROFF_ARGS)
                   3397: {
1.319     schwarze 3398:        if (r->eqn != NULL) {
                   3399:                eqn_parse(r->eqn);
                   3400:                r->eqn = NULL;
                   3401:        } else
1.350     schwarze 3402:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "EN");
1.319     schwarze 3403:        if (buf->buf[pos] != '\0')
1.350     schwarze 3404:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.319     schwarze 3405:                    "EN %s", buf->buf + pos);
1.277     schwarze 3406:        return ROFF_IGN;
1.125     kristaps 3407: }
                   3408:
1.340     schwarze 3409: static int
1.109     kristaps 3410: roff_TS(ROFF_ARGS)
                   3411: {
1.321     schwarze 3412:        if (r->tbl != NULL) {
1.350     schwarze 3413:                mandoc_msg(MANDOCERR_BLK_BROKEN, ln, ppos, "TS breaks TS");
1.346     schwarze 3414:                tbl_end(r->tbl, 0);
1.115     kristaps 3415:        }
1.361     schwarze 3416:        r->man->flags |= ROFF_NONOFILL;
1.351     schwarze 3417:        r->tbl = tbl_alloc(ppos, ln, r->last_tbl);
1.346     schwarze 3418:        if (r->last_tbl == NULL)
1.321     schwarze 3419:                r->first_tbl = r->tbl;
                   3420:        r->last_tbl = r->tbl;
1.297     schwarze 3421:        return ROFF_IGN;
                   3422: }
                   3423:
1.340     schwarze 3424: static int
1.358     schwarze 3425: roff_noarg(ROFF_ARGS)
                   3426: {
                   3427:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3428:                man_breakscope(r->man, tok);
                   3429:        if (tok == ROFF_brp)
                   3430:                tok = ROFF_br;
                   3431:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3432:        if (buf->buf[pos] != '\0')
                   3433:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
                   3434:                   "%s %s", roff_name[tok], buf->buf + pos);
                   3435:        if (tok == ROFF_nf)
                   3436:                r->man->flags |= ROFF_NOFILL;
                   3437:        else if (tok == ROFF_fi)
                   3438:                r->man->flags &= ~ROFF_NOFILL;
                   3439:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3440:        r->man->next = ROFF_NEXT_SIBLING;
                   3441:        return ROFF_IGN;
                   3442: }
                   3443:
                   3444: static int
1.297     schwarze 3445: roff_onearg(ROFF_ARGS)
                   3446: {
                   3447:        struct roff_node        *n;
                   3448:        char                    *cp;
1.305     schwarze 3449:        int                      npos;
1.297     schwarze 3450:
1.302     schwarze 3451:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320     schwarze 3452:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
                   3453:             tok == ROFF_ti))
1.302     schwarze 3454:                man_breakscope(r->man, tok);
                   3455:
1.309     schwarze 3456:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305     schwarze 3457:                r->man->last = roffce_node;
                   3458:                r->man->next = ROFF_NEXT_SIBLING;
                   3459:        }
                   3460:
1.297     schwarze 3461:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3462:        n = r->man->last;
                   3463:
                   3464:        cp = buf->buf + pos;
                   3465:        if (*cp != '\0') {
                   3466:                while (*cp != '\0' && *cp != ' ')
                   3467:                        cp++;
                   3468:                while (*cp == ' ')
                   3469:                        *cp++ = '\0';
                   3470:                if (*cp != '\0')
1.350     schwarze 3471:                        mandoc_msg(MANDOCERR_ARG_EXCESS,
                   3472:                            ln, (int)(cp - buf->buf),
1.297     schwarze 3473:                            "%s ... %s", roff_name[tok], cp);
                   3474:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300     schwarze 3475:        }
                   3476:
1.309     schwarze 3477:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   3478:                if (r->man->last->type == ROFFT_ELEM) {
1.305     schwarze 3479:                        roff_word_alloc(r->man, ln, pos, "1");
                   3480:                        r->man->last->flags |= NODE_NOSRC;
                   3481:                }
                   3482:                npos = 0;
                   3483:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   3484:                    &roffce_lines, 0) == 0) {
1.350     schwarze 3485:                        mandoc_msg(MANDOCERR_CE_NONUM,
                   3486:                            ln, pos, "ce %s", buf->buf + pos);
1.305     schwarze 3487:                        roffce_lines = 1;
                   3488:                }
                   3489:                if (roffce_lines < 1) {
                   3490:                        r->man->last = r->man->last->parent;
                   3491:                        roffce_node = NULL;
                   3492:                        roffce_lines = 0;
                   3493:                } else
                   3494:                        roffce_node = r->man->last->parent;
                   3495:        } else {
                   3496:                n->flags |= NODE_VALID | NODE_ENDED;
                   3497:                r->man->last = n;
                   3498:        }
                   3499:        n->flags |= NODE_LINE;
1.300     schwarze 3500:        r->man->next = ROFF_NEXT_SIBLING;
                   3501:        return ROFF_IGN;
                   3502: }
                   3503:
1.340     schwarze 3504: static int
1.300     schwarze 3505: roff_manyarg(ROFF_ARGS)
                   3506: {
                   3507:        struct roff_node        *n;
                   3508:        char                    *sp, *ep;
                   3509:
                   3510:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3511:        n = r->man->last;
                   3512:
                   3513:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   3514:                while (*ep != '\0' && *ep != ' ')
                   3515:                        ep++;
                   3516:                while (*ep == ' ')
                   3517:                        *ep++ = '\0';
                   3518:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297     schwarze 3519:        }
                   3520:
                   3521:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3522:        r->man->last = n;
                   3523:        r->man->next = ROFF_NEXT_SIBLING;
1.277     schwarze 3524:        return ROFF_IGN;
1.251     schwarze 3525: }
                   3526:
1.340     schwarze 3527: static int
1.311     schwarze 3528: roff_als(ROFF_ARGS)
                   3529: {
                   3530:        char            *oldn, *newn, *end, *value;
                   3531:        size_t           oldsz, newsz, valsz;
                   3532:
                   3533:        newn = oldn = buf->buf + pos;
                   3534:        if (*newn == '\0')
                   3535:                return ROFF_IGN;
                   3536:
                   3537:        newsz = roff_getname(r, &oldn, ln, pos);
1.363     schwarze 3538:        if (newn[newsz] == '\\' || newn[newsz] == '\t' || *oldn == '\0')
1.311     schwarze 3539:                return ROFF_IGN;
                   3540:
                   3541:        end = oldn;
                   3542:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3543:        if (oldsz == 0)
                   3544:                return ROFF_IGN;
                   3545:
1.338     schwarze 3546:        valsz = mandoc_asprintf(&value, ".%.*s \\$@\\\"\n",
1.315     schwarze 3547:            (int)oldsz, oldn);
1.311     schwarze 3548:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3549:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3550:        free(value);
1.364     schwarze 3551:        return ROFF_IGN;
                   3552: }
                   3553:
                   3554: /*
                   3555:  * The .break request only makes sense inside conditionals,
                   3556:  * and that case is already handled in roff_cond_sub().
                   3557:  */
                   3558: static int
                   3559: roff_break(ROFF_ARGS)
                   3560: {
                   3561:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, pos, "break");
1.296     schwarze 3562:        return ROFF_IGN;
1.92      schwarze 3563: }
                   3564:
1.340     schwarze 3565: static int
1.174     kristaps 3566: roff_cc(ROFF_ARGS)
                   3567: {
                   3568:        const char      *p;
                   3569:
1.238     schwarze 3570:        p = buf->buf + pos;
1.174     kristaps 3571:
1.238     schwarze 3572:        if (*p == '\0' || (r->control = *p++) == '.')
1.303     schwarze 3573:                r->control = '\0';
1.174     kristaps 3574:
1.238     schwarze 3575:        if (*p != '\0')
1.350     schwarze 3576:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.260     schwarze 3577:                    ln, p - buf->buf, "cc ... %s", p);
1.174     kristaps 3578:
1.341     schwarze 3579:        return ROFF_IGN;
                   3580: }
                   3581:
                   3582: static int
                   3583: roff_char(ROFF_ARGS)
                   3584: {
                   3585:        const char      *p, *kp, *vp;
                   3586:        size_t           ksz, vsz;
                   3587:        int              font;
                   3588:
                   3589:        /* Parse the character to be replaced. */
                   3590:
                   3591:        kp = buf->buf + pos;
                   3592:        p = kp + 1;
                   3593:        if (*kp == '\0' || (*kp == '\\' &&
                   3594:             mandoc_escape(&p, NULL, NULL) != ESCAPE_SPECIAL) ||
                   3595:            (*p != ' ' && *p != '\0')) {
1.350     schwarze 3596:                mandoc_msg(MANDOCERR_CHAR_ARG, ln, pos, "char %s", kp);
1.341     schwarze 3597:                return ROFF_IGN;
                   3598:        }
                   3599:        ksz = p - kp;
                   3600:        while (*p == ' ')
                   3601:                p++;
                   3602:
                   3603:        /*
                   3604:         * If the replacement string contains a font escape sequence,
                   3605:         * we have to restore the font at the end.
                   3606:         */
                   3607:
                   3608:        vp = p;
                   3609:        vsz = strlen(p);
                   3610:        font = 0;
                   3611:        while (*p != '\0') {
                   3612:                if (*p++ != '\\')
                   3613:                        continue;
                   3614:                switch (mandoc_escape(&p, NULL, NULL)) {
                   3615:                case ESCAPE_FONT:
                   3616:                case ESCAPE_FONTROMAN:
                   3617:                case ESCAPE_FONTITALIC:
                   3618:                case ESCAPE_FONTBOLD:
                   3619:                case ESCAPE_FONTBI:
1.378     schwarze 3620:                case ESCAPE_FONTCR:
                   3621:                case ESCAPE_FONTCB:
                   3622:                case ESCAPE_FONTCI:
1.341     schwarze 3623:                case ESCAPE_FONTPREV:
                   3624:                        font++;
                   3625:                        break;
                   3626:                default:
                   3627:                        break;
                   3628:                }
                   3629:        }
                   3630:        if (font > 1)
1.350     schwarze 3631:                mandoc_msg(MANDOCERR_CHAR_FONT,
                   3632:                    ln, (int)(vp - buf->buf), "%s", vp);
1.341     schwarze 3633:
                   3634:        /*
                   3635:         * Approximate the effect of .char using the .tr tables.
                   3636:         * XXX In groff, .char and .tr interact differently.
                   3637:         */
                   3638:
                   3639:        if (ksz == 1) {
                   3640:                if (r->xtab == NULL)
                   3641:                        r->xtab = mandoc_calloc(128, sizeof(*r->xtab));
                   3642:                assert((unsigned int)*kp < 128);
                   3643:                free(r->xtab[(int)*kp].p);
                   3644:                r->xtab[(int)*kp].sz = mandoc_asprintf(&r->xtab[(int)*kp].p,
                   3645:                    "%s%s", vp, font ? "\fP" : "");
                   3646:        } else {
                   3647:                roff_setstrn(&r->xmbtab, kp, ksz, vp, vsz, 0);
                   3648:                if (font)
                   3649:                        roff_setstrn(&r->xmbtab, kp, ksz, "\\fP", 3, 1);
                   3650:        }
1.277     schwarze 3651:        return ROFF_IGN;
1.174     kristaps 3652: }
                   3653:
1.340     schwarze 3654: static int
1.303     schwarze 3655: roff_ec(ROFF_ARGS)
                   3656: {
                   3657:        const char      *p;
                   3658:
                   3659:        p = buf->buf + pos;
                   3660:        if (*p == '\0')
                   3661:                r->escape = '\\';
                   3662:        else {
                   3663:                r->escape = *p;
                   3664:                if (*++p != '\0')
1.350     schwarze 3665:                        mandoc_msg(MANDOCERR_ARG_EXCESS, ln,
                   3666:                            (int)(p - buf->buf), "ec ... %s", p);
1.303     schwarze 3667:        }
                   3668:        return ROFF_IGN;
                   3669: }
                   3670:
1.340     schwarze 3671: static int
1.303     schwarze 3672: roff_eo(ROFF_ARGS)
                   3673: {
                   3674:        r->escape = '\0';
                   3675:        if (buf->buf[pos] != '\0')
1.350     schwarze 3676:                mandoc_msg(MANDOCERR_ARG_SKIP,
1.303     schwarze 3677:                    ln, pos, "eo %s", buf->buf + pos);
1.384     schwarze 3678:        return ROFF_IGN;
                   3679: }
                   3680:
                   3681: static int
                   3682: roff_mc(ROFF_ARGS)
                   3683: {
                   3684:        struct roff_node        *n;
                   3685:        char                    *cp;
                   3686:
                   3687:        /* Parse the first argument. */
                   3688:
                   3689:        cp = buf->buf + pos;
                   3690:        if (*cp != '\0')
                   3691:                cp++;
                   3692:        if (buf->buf[pos] == '\\') {
                   3693:                switch (mandoc_escape((const char **)&cp, NULL, NULL)) {
                   3694:                case ESCAPE_SPECIAL:
                   3695:                case ESCAPE_UNICODE:
                   3696:                case ESCAPE_NUMBERED:
                   3697:                        break;
                   3698:                default:
                   3699:                        *cp = '\0';
                   3700:                        mandoc_msg(MANDOCERR_MC_ESC, ln, pos,
                   3701:                            "mc %s", buf->buf + pos);
                   3702:                        buf->buf[pos] = '\0';
                   3703:                        break;
                   3704:                }
                   3705:        }
                   3706:
                   3707:        /* Ignore additional arguments. */
                   3708:
                   3709:        while (*cp == ' ')
                   3710:                *cp++ = '\0';
                   3711:        if (*cp != '\0') {
                   3712:                mandoc_msg(MANDOCERR_MC_DIST, ln, (int)(cp - buf->buf),
                   3713:                    "mc ... %s", cp);
                   3714:                *cp = '\0';
                   3715:        }
                   3716:
                   3717:        /* Create the .mc node. */
                   3718:
                   3719:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3720:        n = r->man->last;
                   3721:        if (buf->buf[pos] != '\0')
                   3722:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
                   3723:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3724:        r->man->last = n;
                   3725:        r->man->next = ROFF_NEXT_SIBLING;
1.303     schwarze 3726:        return ROFF_IGN;
1.330     schwarze 3727: }
                   3728:
1.340     schwarze 3729: static int
1.330     schwarze 3730: roff_nop(ROFF_ARGS)
                   3731: {
                   3732:        while (buf->buf[pos] == ' ')
                   3733:                pos++;
                   3734:        *offs = pos;
                   3735:        return ROFF_RERUN;
1.303     schwarze 3736: }
                   3737:
1.340     schwarze 3738: static int
1.164     kristaps 3739: roff_tr(ROFF_ARGS)
                   3740: {
                   3741:        const char      *p, *first, *second;
                   3742:        size_t           fsz, ssz;
                   3743:
1.238     schwarze 3744:        p = buf->buf + pos;
1.164     kristaps 3745:
1.238     schwarze 3746:        if (*p == '\0') {
1.350     schwarze 3747:                mandoc_msg(MANDOCERR_REQ_EMPTY, ln, ppos, "tr");
1.277     schwarze 3748:                return ROFF_IGN;
1.164     kristaps 3749:        }
                   3750:
1.238     schwarze 3751:        while (*p != '\0') {
1.164     kristaps 3752:                fsz = ssz = 1;
                   3753:
                   3754:                first = p++;
1.238     schwarze 3755:                if (*first == '\\') {
1.394   ! schwarze 3756:                        if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR)
1.277     schwarze 3757:                                return ROFF_IGN;
1.164     kristaps 3758:                        fsz = (size_t)(p - first);
                   3759:                }
                   3760:
                   3761:                second = p++;
1.238     schwarze 3762:                if (*second == '\\') {
1.394   ! schwarze 3763:                        if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR)
1.277     schwarze 3764:                                return ROFF_IGN;
1.164     kristaps 3765:                        ssz = (size_t)(p - second);
1.238     schwarze 3766:                } else if (*second == '\0') {
1.350     schwarze 3767:                        mandoc_msg(MANDOCERR_TR_ODD, ln,
                   3768:                            (int)(first - buf->buf), "tr %s", first);
1.164     kristaps 3769:                        second = " ";
1.165     kristaps 3770:                        p--;
1.164     kristaps 3771:                }
                   3772:
1.167     kristaps 3773:                if (fsz > 1) {
1.207     schwarze 3774:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3775:                            second, ssz, 0);
1.167     kristaps 3776:                        continue;
                   3777:                }
                   3778:
1.238     schwarze 3779:                if (r->xtab == NULL)
1.207     schwarze 3780:                        r->xtab = mandoc_calloc(128,
                   3781:                            sizeof(struct roffstr));
1.167     kristaps 3782:
                   3783:                free(r->xtab[(int)*first].p);
                   3784:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3785:                r->xtab[(int)*first].sz = ssz;
1.164     kristaps 3786:        }
                   3787:
1.277     schwarze 3788:        return ROFF_IGN;
1.164     kristaps 3789: }
                   3790:
1.339     schwarze 3791: /*
                   3792:  * Implementation of the .return request.
                   3793:  * There is no need to call roff_userret() from here.
                   3794:  * The read module will call that after rewinding the reader stack
                   3795:  * to the place from where the current macro was called.
                   3796:  */
1.340     schwarze 3797: static int
1.339     schwarze 3798: roff_return(ROFF_ARGS)
                   3799: {
                   3800:        if (r->mstackpos >= 0)
1.340     schwarze 3801:                return ROFF_IGN | ROFF_USERRET;
1.339     schwarze 3802:
1.350     schwarze 3803:        mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "return");
1.339     schwarze 3804:        return ROFF_IGN;
                   3805: }
                   3806:
1.340     schwarze 3807: static int
1.306     schwarze 3808: roff_rn(ROFF_ARGS)
                   3809: {
                   3810:        const char      *value;
                   3811:        char            *oldn, *newn, *end;
                   3812:        size_t           oldsz, newsz;
1.315     schwarze 3813:        int              deftype;
1.306     schwarze 3814:
                   3815:        oldn = newn = buf->buf + pos;
                   3816:        if (*oldn == '\0')
                   3817:                return ROFF_IGN;
                   3818:
                   3819:        oldsz = roff_getname(r, &newn, ln, pos);
1.363     schwarze 3820:        if (oldn[oldsz] == '\\' || oldn[oldsz] == '\t' || *newn == '\0')
1.306     schwarze 3821:                return ROFF_IGN;
                   3822:
                   3823:        end = newn;
                   3824:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3825:        if (newsz == 0)
                   3826:                return ROFF_IGN;
                   3827:
1.315     schwarze 3828:        deftype = ROFFDEF_ANY;
                   3829:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3830:        switch (deftype) {
                   3831:        case ROFFDEF_USER:
1.306     schwarze 3832:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3833:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3834:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315     schwarze 3835:                break;
                   3836:        case ROFFDEF_PRE:
                   3837:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3838:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3839:                break;
                   3840:        case ROFFDEF_REN:
1.306     schwarze 3841:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3842:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315     schwarze 3843:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3844:                break;
                   3845:        case ROFFDEF_STD:
1.306     schwarze 3846:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315     schwarze 3847:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3848:                break;
                   3849:        default:
                   3850:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3851:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3852:                break;
                   3853:        }
1.306     schwarze 3854:        return ROFF_IGN;
                   3855: }
                   3856:
1.340     schwarze 3857: static int
1.339     schwarze 3858: roff_shift(ROFF_ARGS)
                   3859: {
                   3860:        struct mctx     *ctx;
1.382     schwarze 3861:        int              argpos, levels, i;
1.339     schwarze 3862:
1.382     schwarze 3863:        argpos = pos;
1.339     schwarze 3864:        levels = 1;
                   3865:        if (buf->buf[pos] != '\0' &&
                   3866:            roff_evalnum(r, ln, buf->buf, &pos, &levels, 0) == 0) {
1.350     schwarze 3867:                mandoc_msg(MANDOCERR_CE_NONUM,
1.339     schwarze 3868:                    ln, pos, "shift %s", buf->buf + pos);
                   3869:                levels = 1;
                   3870:        }
                   3871:        if (r->mstackpos < 0) {
1.350     schwarze 3872:                mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "shift");
1.339     schwarze 3873:                return ROFF_IGN;
                   3874:        }
                   3875:        ctx = r->mstack + r->mstackpos;
                   3876:        if (levels > ctx->argc) {
1.350     schwarze 3877:                mandoc_msg(MANDOCERR_SHIFT,
1.382     schwarze 3878:                    ln, argpos, "%d, but max is %d", levels, ctx->argc);
1.339     schwarze 3879:                levels = ctx->argc;
1.382     schwarze 3880:        }
                   3881:        if (levels < 0) {
                   3882:                mandoc_msg(MANDOCERR_ARG_NEG, ln, argpos, "shift %d", levels);
                   3883:                levels = 0;
1.339     schwarze 3884:        }
                   3885:        if (levels == 0)
                   3886:                return ROFF_IGN;
                   3887:        for (i = 0; i < levels; i++)
                   3888:                free(ctx->argv[i]);
                   3889:        ctx->argc -= levels;
                   3890:        for (i = 0; i < ctx->argc; i++)
                   3891:                ctx->argv[i] = ctx->argv[i + levels];
                   3892:        return ROFF_IGN;
                   3893: }
                   3894:
1.340     schwarze 3895: static int
1.105     kristaps 3896: roff_so(ROFF_ARGS)
                   3897: {
1.249     schwarze 3898:        char *name, *cp;
1.105     kristaps 3899:
1.238     schwarze 3900:        name = buf->buf + pos;
1.350     schwarze 3901:        mandoc_msg(MANDOCERR_SO, ln, ppos, "so %s", name);
1.105     kristaps 3902:
                   3903:        /*
                   3904:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3905:         * opening anything that's not in our cwd or anything beneath
                   3906:         * it.  Thus, explicitly disallow traversing up the file-system
                   3907:         * or using absolute paths.
                   3908:         */
                   3909:
1.238     schwarze 3910:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.350     schwarze 3911:                mandoc_msg(MANDOCERR_SO_PATH, ln, ppos, ".so %s", name);
1.249     schwarze 3912:                buf->sz = mandoc_asprintf(&cp,
                   3913:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3914:                free(buf->buf);
                   3915:                buf->buf = cp;
                   3916:                *offs = 0;
1.277     schwarze 3917:                return ROFF_REPARSE;
1.105     kristaps 3918:        }
                   3919:
                   3920:        *offs = pos;
1.277     schwarze 3921:        return ROFF_SO;
1.105     kristaps 3922: }
1.92      schwarze 3923:
1.266     schwarze 3924: /* --- user defined strings and macros ------------------------------------ */
                   3925:
1.340     schwarze 3926: static int
1.106     kristaps 3927: roff_userdef(ROFF_ARGS)
1.99      kristaps 3928: {
1.339     schwarze 3929:        struct mctx      *ctx;
                   3930:        char             *arg, *ap, *dst, *src;
                   3931:        size_t            sz;
                   3932:
1.365     schwarze 3933:        /* If the macro is empty, ignore it altogether. */
                   3934:
                   3935:        if (*r->current_string == '\0')
                   3936:                return ROFF_IGN;
                   3937:
1.339     schwarze 3938:        /* Initialize a new macro stack context. */
                   3939:
                   3940:        if (++r->mstackpos == r->mstacksz) {
                   3941:                r->mstack = mandoc_recallocarray(r->mstack,
                   3942:                    r->mstacksz, r->mstacksz + 8, sizeof(*r->mstack));
                   3943:                r->mstacksz += 8;
                   3944:        }
                   3945:        ctx = r->mstack + r->mstackpos;
                   3946:        ctx->argc = 0;
                   3947:
                   3948:        /*
                   3949:         * Collect pointers to macro argument strings,
                   3950:         * NUL-terminating them and escaping quotes.
                   3951:         */
                   3952:
                   3953:        src = buf->buf + pos;
                   3954:        while (*src != '\0') {
                   3955:                if (ctx->argc == ctx->argsz) {
                   3956:                        ctx->argsz += 8;
                   3957:                        ctx->argv = mandoc_reallocarray(ctx->argv,
                   3958:                            ctx->argsz, sizeof(*ctx->argv));
                   3959:                }
1.355     schwarze 3960:                arg = roff_getarg(r, &src, ln, &pos);
1.339     schwarze 3961:                sz = 1;  /* For the terminating NUL. */
                   3962:                for (ap = arg; *ap != '\0'; ap++)
                   3963:                        sz += *ap == '"' ? 4 : 1;
                   3964:                ctx->argv[ctx->argc++] = dst = mandoc_malloc(sz);
                   3965:                for (ap = arg; *ap != '\0'; ap++) {
                   3966:                        if (*ap == '"') {
                   3967:                                memcpy(dst, "\\(dq", 4);
                   3968:                                dst += 4;
                   3969:                        } else
                   3970:                                *dst++ = *ap;
1.106     kristaps 3971:                }
1.339     schwarze 3972:                *dst = '\0';
1.355     schwarze 3973:                free(arg);
1.337     schwarze 3974:        }
                   3975:
1.339     schwarze 3976:        /* Replace the macro invocation by the macro definition. */
1.263     schwarze 3977:
1.238     schwarze 3978:        free(buf->buf);
1.339     schwarze 3979:        buf->buf = mandoc_strdup(r->current_string);
                   3980:        buf->sz = strlen(buf->buf) + 1;
1.248     schwarze 3981:        *offs = 0;
1.106     kristaps 3982:
1.365     schwarze 3983:        return buf->buf[buf->sz - 2] == '\n' ?
1.340     schwarze 3984:            ROFF_REPARSE | ROFF_USERCALL : ROFF_IGN | ROFF_APPEND;
1.99      kristaps 3985: }
1.121     schwarze 3986:
1.306     schwarze 3987: /*
                   3988:  * Calling a high-level macro that was renamed with .rn.
                   3989:  * r->current_string has already been set up by roff_parse().
                   3990:  */
1.340     schwarze 3991: static int
1.306     schwarze 3992: roff_renamed(ROFF_ARGS)
                   3993: {
                   3994:        char    *nbuf;
                   3995:
1.315     schwarze 3996:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3997:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306     schwarze 3998:        free(buf->buf);
                   3999:        buf->buf = nbuf;
1.329     schwarze 4000:        *offs = 0;
1.306     schwarze 4001:        return ROFF_CONT;
                   4002: }
                   4003:
1.362     schwarze 4004: /*
                   4005:  * Measure the length in bytes of the roff identifier at *cpp
                   4006:  * and advance the pointer to the next word.
                   4007:  */
1.212     schwarze 4008: static size_t
1.121     schwarze 4009: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   4010: {
                   4011:        char     *name, *cp;
1.393     schwarze 4012:        int       namesz, inam, iend;
1.121     schwarze 4013:
                   4014:        name = *cpp;
1.362     schwarze 4015:        if (*name == '\0')
1.277     schwarze 4016:                return 0;
1.121     schwarze 4017:
1.362     schwarze 4018:        /* Advance cp to the byte after the end of the name. */
                   4019:
1.393     schwarze 4020:        cp = name;
                   4021:        namesz = 0;
                   4022:        for (;;) {
1.363     schwarze 4023:                if (*cp == '\0')
1.212     schwarze 4024:                        break;
1.363     schwarze 4025:                if (*cp == ' ' || *cp == '\t') {
                   4026:                        cp++;
                   4027:                        break;
                   4028:                }
1.393     schwarze 4029:                if (*cp != '\\') {
                   4030:                        if (name + namesz < cp) {
                   4031:                                name[namesz] = *cp;
                   4032:                                *cp = ' ';
                   4033:                        }
                   4034:                        namesz++;
                   4035:                        cp++;
1.121     schwarze 4036:                        continue;
1.393     schwarze 4037:                }
1.362     schwarze 4038:                if (cp[1] == '{' || cp[1] == '}')
1.215     schwarze 4039:                        break;
1.393     schwarze 4040:                if (roff_escape(cp, 0, 0, NULL, &inam,
                   4041:                    NULL, NULL, &iend) != ESCAPE_UNDEF) {
                   4042:                        mandoc_msg(MANDOCERR_NAMESC, ln, pos,
                   4043:                            "%.*s%.*s", namesz, name, iend, cp);
                   4044:                        cp += iend;
                   4045:                        break;
                   4046:                }
                   4047:
                   4048:                /*
                   4049:                 * In an identifier, \\, \., \G and so on
                   4050:                 * are reduced to \, ., G and so on,
                   4051:                 * vaguely similar to copy mode.
                   4052:                 */
                   4053:
                   4054:                name[namesz++] = cp[inam];
                   4055:                while (iend--) {
                   4056:                        if (cp >= name + namesz)
                   4057:                                *cp = ' ';
                   4058:                        cp++;
                   4059:                }
1.121     schwarze 4060:        }
                   4061:
                   4062:        /* Read past spaces. */
1.362     schwarze 4063:
                   4064:        while (*cp == ' ')
1.121     schwarze 4065:                cp++;
                   4066:
                   4067:        *cpp = cp;
1.277     schwarze 4068:        return namesz;
1.121     schwarze 4069: }
                   4070:
1.106     kristaps 4071: /*
                   4072:  * Store *string into the user-defined string called *name.
                   4073:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193     schwarze 4074:  * append == 0: replace mode
                   4075:  * append == 1: single-line append mode
                   4076:  * append == 2: multiline append mode, append '\n' after each call
1.106     kristaps 4077:  */
1.94      kristaps 4078: static void
1.106     kristaps 4079: roff_setstr(struct roff *r, const char *name, const char *string,
1.193     schwarze 4080:        int append)
1.92      schwarze 4081: {
1.315     schwarze 4082:        size_t   namesz;
1.164     kristaps 4083:
1.315     schwarze 4084:        namesz = strlen(name);
                   4085:        roff_setstrn(&r->strtab, name, namesz, string,
1.207     schwarze 4086:            string ? strlen(string) : 0, append);
1.315     schwarze 4087:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164     kristaps 4088: }
                   4089:
                   4090: static void
1.166     kristaps 4091: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193     schwarze 4092:                const char *string, size_t stringsz, int append)
1.164     kristaps 4093: {
1.166     kristaps 4094:        struct roffkv   *n;
1.164     kristaps 4095:        char            *c;
                   4096:        int              i;
                   4097:        size_t           oldch, newch;
1.92      schwarze 4098:
1.106     kristaps 4099:        /* Search for an existing string with the same name. */
1.164     kristaps 4100:        n = *r;
                   4101:
1.211     schwarze 4102:        while (n && (namesz != n->key.sz ||
                   4103:                        strncmp(n->key.p, name, namesz)))
1.92      schwarze 4104:                n = n->next;
1.94      kristaps 4105:
                   4106:        if (NULL == n) {
1.106     kristaps 4107:                /* Create a new string table entry. */
1.166     kristaps 4108:                n = mandoc_malloc(sizeof(struct roffkv));
                   4109:                n->key.p = mandoc_strndup(name, namesz);
                   4110:                n->key.sz = namesz;
                   4111:                n->val.p = NULL;
                   4112:                n->val.sz = 0;
1.164     kristaps 4113:                n->next = *r;
                   4114:                *r = n;
1.193     schwarze 4115:        } else if (0 == append) {
1.166     kristaps 4116:                free(n->val.p);
                   4117:                n->val.p = NULL;
                   4118:                n->val.sz = 0;
1.106     kristaps 4119:        }
                   4120:
                   4121:        if (NULL == string)
                   4122:                return;
                   4123:
                   4124:        /*
                   4125:         * One additional byte for the '\n' in multiline mode,
                   4126:         * and one for the terminating '\0'.
                   4127:         */
1.193     schwarze 4128:        newch = stringsz + (1 < append ? 2u : 1u);
1.164     kristaps 4129:
1.166     kristaps 4130:        if (NULL == n->val.p) {
                   4131:                n->val.p = mandoc_malloc(newch);
                   4132:                *n->val.p = '\0';
1.106     kristaps 4133:                oldch = 0;
                   4134:        } else {
1.166     kristaps 4135:                oldch = n->val.sz;
                   4136:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106     kristaps 4137:        }
                   4138:
                   4139:        /* Skip existing content in the destination buffer. */
1.166     kristaps 4140:        c = n->val.p + (int)oldch;
1.106     kristaps 4141:
                   4142:        /* Append new content to the destination buffer. */
1.164     kristaps 4143:        i = 0;
                   4144:        while (i < (int)stringsz) {
1.106     kristaps 4145:                /*
                   4146:                 * Rudimentary roff copy mode:
                   4147:                 * Handle escaped backslashes.
                   4148:                 */
1.164     kristaps 4149:                if ('\\' == string[i] && '\\' == string[i + 1])
                   4150:                        i++;
                   4151:                *c++ = string[i++];
1.106     kristaps 4152:        }
1.94      kristaps 4153:
1.106     kristaps 4154:        /* Append terminating bytes. */
1.193     schwarze 4155:        if (1 < append)
1.106     kristaps 4156:                *c++ = '\n';
1.163     kristaps 4157:
1.106     kristaps 4158:        *c = '\0';
1.166     kristaps 4159:        n->val.sz = (int)(c - n->val.p);
1.92      schwarze 4160: }
                   4161:
1.94      kristaps 4162: static const char *
1.325     schwarze 4163: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315     schwarze 4164:     int *deftype)
1.92      schwarze 4165: {
1.315     schwarze 4166:        const struct roffkv     *n;
1.325     schwarze 4167:        int                      found, i;
1.315     schwarze 4168:        enum roff_tok            tok;
                   4169:
1.325     schwarze 4170:        found = 0;
                   4171:        for (n = r->strtab; n != NULL; n = n->next) {
                   4172:                if (strncmp(name, n->key.p, len) != 0 ||
                   4173:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4174:                        continue;
                   4175:                if (*deftype & ROFFDEF_USER) {
                   4176:                        *deftype = ROFFDEF_USER;
                   4177:                        return n->val.p;
                   4178:                } else {
                   4179:                        found = 1;
                   4180:                        break;
                   4181:                }
                   4182:        }
                   4183:        for (n = r->rentab; n != NULL; n = n->next) {
                   4184:                if (strncmp(name, n->key.p, len) != 0 ||
                   4185:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4186:                        continue;
                   4187:                if (*deftype & ROFFDEF_REN) {
                   4188:                        *deftype = ROFFDEF_REN;
                   4189:                        return n->val.p;
                   4190:                } else {
                   4191:                        found = 1;
                   4192:                        break;
1.315     schwarze 4193:                }
                   4194:        }
1.325     schwarze 4195:        for (i = 0; i < PREDEFS_MAX; i++) {
                   4196:                if (strncmp(name, predefs[i].name, len) != 0 ||
                   4197:                    predefs[i].name[len] != '\0')
                   4198:                        continue;
                   4199:                if (*deftype & ROFFDEF_PRE) {
                   4200:                        *deftype = ROFFDEF_PRE;
                   4201:                        return predefs[i].str;
                   4202:                } else {
                   4203:                        found = 1;
                   4204:                        break;
1.315     schwarze 4205:                }
                   4206:        }
1.356     schwarze 4207:        if (r->man->meta.macroset != MACROSET_MAN) {
1.325     schwarze 4208:                for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
                   4209:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4210:                            roff_name[tok][len] != '\0')
                   4211:                                continue;
                   4212:                        if (*deftype & ROFFDEF_STD) {
                   4213:                                *deftype = ROFFDEF_STD;
                   4214:                                return NULL;
                   4215:                        } else {
                   4216:                                found = 1;
                   4217:                                break;
1.315     schwarze 4218:                        }
                   4219:                }
                   4220:        }
1.356     schwarze 4221:        if (r->man->meta.macroset != MACROSET_MDOC) {
1.325     schwarze 4222:                for (tok = MAN_TH; tok < MAN_MAX; tok++) {
                   4223:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4224:                            roff_name[tok][len] != '\0')
                   4225:                                continue;
                   4226:                        if (*deftype & ROFFDEF_STD) {
                   4227:                                *deftype = ROFFDEF_STD;
                   4228:                                return NULL;
                   4229:                        } else {
                   4230:                                found = 1;
                   4231:                                break;
1.315     schwarze 4232:                        }
                   4233:                }
1.325     schwarze 4234:        }
                   4235:
                   4236:        if (found == 0 && *deftype != ROFFDEF_ANY) {
                   4237:                if (*deftype & ROFFDEF_REN) {
                   4238:                        /*
                   4239:                         * This might still be a request,
                   4240:                         * so do not treat it as undefined yet.
                   4241:                         */
                   4242:                        *deftype = ROFFDEF_UNDEF;
                   4243:                        return NULL;
1.315     schwarze 4244:                }
1.325     schwarze 4245:
                   4246:                /* Using an undefined string defines it to be empty. */
                   4247:
                   4248:                roff_setstrn(&r->strtab, name, len, "", 0, 0);
                   4249:                roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315     schwarze 4250:        }
1.325     schwarze 4251:
1.315     schwarze 4252:        *deftype = 0;
1.277     schwarze 4253:        return NULL;
1.92      schwarze 4254: }
                   4255:
1.94      kristaps 4256: static void
1.167     kristaps 4257: roff_freestr(struct roffkv *r)
1.92      schwarze 4258: {
1.166     kristaps 4259:        struct roffkv    *n, *nn;
1.92      schwarze 4260:
1.167     kristaps 4261:        for (n = r; n; n = nn) {
1.166     kristaps 4262:                free(n->key.p);
                   4263:                free(n->val.p);
1.92      schwarze 4264:                nn = n->next;
                   4265:                free(n);
                   4266:        }
1.114     kristaps 4267: }
1.266     schwarze 4268:
                   4269: /* --- accessors and utility functions ------------------------------------ */
1.164     kristaps 4270:
                   4271: /*
                   4272:  * Duplicate an input string, making the appropriate character
                   4273:  * conversations (as stipulated by `tr') along the way.
                   4274:  * Returns a heap-allocated string with all the replacements made.
                   4275:  */
                   4276: char *
                   4277: roff_strdup(const struct roff *r, const char *p)
                   4278: {
1.166     kristaps 4279:        const struct roffkv *cp;
1.164     kristaps 4280:        char            *res;
                   4281:        const char      *pp;
                   4282:        size_t           ssz, sz;
                   4283:        enum mandoc_esc  esc;
                   4284:
1.167     kristaps 4285:        if (NULL == r->xmbtab && NULL == r->xtab)
1.277     schwarze 4286:                return mandoc_strdup(p);
1.164     kristaps 4287:        else if ('\0' == *p)
1.277     schwarze 4288:                return mandoc_strdup("");
1.164     kristaps 4289:
                   4290:        /*
                   4291:         * Step through each character looking for term matches
                   4292:         * (remember that a `tr' can be invoked with an escape, which is
                   4293:         * a glyph but the escape is multi-character).
                   4294:         * We only do this if the character hash has been initialised
                   4295:         * and the string is >0 length.
                   4296:         */
                   4297:
                   4298:        res = NULL;
                   4299:        ssz = 0;
                   4300:
                   4301:        while ('\0' != *p) {
1.289     schwarze 4302:                assert((unsigned int)*p < 128);
                   4303:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167     kristaps 4304:                        sz = r->xtab[(int)*p].sz;
                   4305:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4306:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   4307:                        ssz += sz;
                   4308:                        p++;
                   4309:                        continue;
                   4310:                } else if ('\\' != *p) {
                   4311:                        res = mandoc_realloc(res, ssz + 2);
                   4312:                        res[ssz++] = *p++;
                   4313:                        continue;
                   4314:                }
                   4315:
1.164     kristaps 4316:                /* Search for term matches. */
1.167     kristaps 4317:                for (cp = r->xmbtab; cp; cp = cp->next)
1.166     kristaps 4318:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164     kristaps 4319:                                break;
                   4320:
                   4321:                if (NULL != cp) {
                   4322:                        /*
                   4323:                         * A match has been found.
                   4324:                         * Append the match to the array and move
                   4325:                         * forward by its keysize.
                   4326:                         */
1.207     schwarze 4327:                        res = mandoc_realloc(res,
                   4328:                            ssz + cp->val.sz + 1);
1.166     kristaps 4329:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   4330:                        ssz += cp->val.sz;
                   4331:                        p += (int)cp->key.sz;
1.164     kristaps 4332:                        continue;
                   4333:                }
                   4334:
1.167     kristaps 4335:                /*
                   4336:                 * Handle escapes carefully: we need to copy
                   4337:                 * over just the escape itself, or else we might
                   4338:                 * do replacements within the escape itself.
                   4339:                 * Make sure to pass along the bogus string.
                   4340:                 */
                   4341:                pp = p++;
                   4342:                esc = mandoc_escape(&p, NULL, NULL);
                   4343:                if (ESCAPE_ERROR == esc) {
                   4344:                        sz = strlen(pp);
1.164     kristaps 4345:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4346:                        memcpy(res + ssz, pp, sz);
1.167     kristaps 4347:                        break;
1.164     kristaps 4348:                }
1.207     schwarze 4349:                /*
                   4350:                 * We bail out on bad escapes.
1.167     kristaps 4351:                 * No need to warn: we already did so when
1.355     schwarze 4352:                 * roff_expand() was called.
1.167     kristaps 4353:                 */
                   4354:                sz = (int)(p - pp);
                   4355:                res = mandoc_realloc(res, ssz + sz + 1);
                   4356:                memcpy(res + ssz, pp, sz);
                   4357:                ssz += sz;
1.164     kristaps 4358:        }
                   4359:
                   4360:        res[(int)ssz] = '\0';
1.277     schwarze 4361:        return res;
1.227     schwarze 4362: }
                   4363:
                   4364: int
                   4365: roff_getformat(const struct roff *r)
                   4366: {
                   4367:
1.277     schwarze 4368:        return r->format;
1.174     kristaps 4369: }
                   4370:
                   4371: /*
1.207     schwarze 4372:  * Find out whether a line is a macro line or not.
1.174     kristaps 4373:  * If it is, adjust the current position and return one; if it isn't,
                   4374:  * return zero and don't change the current position.
                   4375:  * If the control character has been set with `.cc', then let that grain
                   4376:  * precedence.
                   4377:  * This is slighly contrary to groff, where using the non-breaking
                   4378:  * control character when `cc' has been invoked will cause the
                   4379:  * non-breaking macro contents to be printed verbatim.
                   4380:  */
                   4381: int
                   4382: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   4383: {
                   4384:        int             pos;
                   4385:
                   4386:        pos = *ppos;
                   4387:
1.303     schwarze 4388:        if (r->control != '\0' && cp[pos] == r->control)
1.174     kristaps 4389:                pos++;
1.303     schwarze 4390:        else if (r->control != '\0')
1.277     schwarze 4391:                return 0;
1.174     kristaps 4392:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   4393:                pos += 2;
                   4394:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   4395:                pos++;
                   4396:        else
1.277     schwarze 4397:                return 0;
1.174     kristaps 4398:
                   4399:        while (' ' == cp[pos] || '\t' == cp[pos])
                   4400:                pos++;
                   4401:
                   4402:        *ppos = pos;
1.277     schwarze 4403:        return 1;
1.74      kristaps 4404: }

CVSweb