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

Annotation of mandoc/roff.c, Revision 1.392

1.392   ! schwarze    1: /* $Id: roff.c,v 1.391 2022/05/31 20:23:05 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 */
                   1378:        int              deftype;       /* type of definition to paste */
                   1379:        int              argi;          /* macro argument index */
                   1380:        int              quote_args;    /* true for \\$@, false for \\$* */
                   1381:        int              asz;           /* length of the replacement */
                   1382:        int              rsz;           /* length of the rest of the string */
                   1383:        int              npos;          /* position in numeric expression */
1.181     schwarze 1384:        int              expand_count;  /* to avoid infinite loops */
1.303     schwarze 1385:
1.170     schwarze 1386:        expand_count = 0;
1.388     schwarze 1387:        while (buf->buf[pos] != '\0') {
1.170     schwarze 1388:
1.388     schwarze 1389:                /*
                   1390:                 * Skip plain ASCII characters.
                   1391:                 * If we have a non-standard escape character,
                   1392:                 * escape literal backslashes because all processing in
                   1393:                 * subsequent functions uses the standard escaping rules.
                   1394:                 */
1.205     schwarze 1395:
1.388     schwarze 1396:                if (buf->buf[pos] != ec) {
                   1397:                        if (ec != ASCII_ESC && buf->buf[pos] == '\\') {
                   1398:                                roff_expand_patch(buf, pos, "\\e", pos + 1);
                   1399:                                pos++;
1.303     schwarze 1400:                        }
1.388     schwarze 1401:                        pos++;
1.205     schwarze 1402:                        continue;
1.303     schwarze 1403:                }
1.205     schwarze 1404:
1.388     schwarze 1405:                /*
                   1406:                 * Parse escape sequences,
                   1407:                 * issue diagnostic messages when appropriate,
                   1408:                 * and skip sequences that do not need expansion.
                   1409:                 * If we have a non-standard escape character, translate
                   1410:                 * it to backslashes and translate backslashes to \e.
                   1411:                 */
1.205     schwarze 1412:
1.392   ! schwarze 1413:                if (roff_escape(buf->buf, ln, pos, &iesc, &inam,
        !          1414:                    &iarg, &iendarg, &iend) != ESCAPE_EXPAND) {
1.388     schwarze 1415:                        while (pos < iend) {
                   1416:                                if (buf->buf[pos] == ec) {
                   1417:                                        buf->buf[pos] = '\\';
                   1418:                                        if (pos + 1 < iend)
                   1419:                                                pos++;
                   1420:                                } else if (buf->buf[pos] == '\\') {
                   1421:                                        roff_expand_patch(buf,
                   1422:                                            pos, "\\e", pos + 1);
                   1423:                                        pos++;
                   1424:                                        iend++;
                   1425:                                }
                   1426:                                pos++;
1.352     schwarze 1427:                        }
1.205     schwarze 1428:                        continue;
1.152     kristaps 1429:                }
1.388     schwarze 1430:
                   1431:                /* Handle expansion. */
                   1432:
                   1433:                res = NULL;
                   1434:                switch (buf->buf[inam]) {
                   1435:                case '*':
                   1436:                        if (iendarg == iarg)
1.206     schwarze 1437:                                break;
1.388     schwarze 1438:                        deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1439:                        if ((res = roff_getstrn(r, buf->buf + iarg,
                   1440:                            iendarg - iarg, &deftype)) != NULL)
1.206     schwarze 1441:                                break;
1.94      kristaps 1442:
1.388     schwarze 1443:                        /*
                   1444:                         * If not overriden,
                   1445:                         * let \*(.T through to the formatters.
                   1446:                         */
1.94      kristaps 1447:
1.388     schwarze 1448:                        if (iendarg - iarg == 2 &&
                   1449:                            buf->buf[iarg] == '.' &&
                   1450:                            buf->buf[iarg + 1] == 'T') {
                   1451:                                roff_setstrn(&r->strtab, ".T", 2, NULL, 0, 0);
                   1452:                                pos = iend;
1.253     schwarze 1453:                                continue;
                   1454:                        }
1.94      kristaps 1455:
1.388     schwarze 1456:                        mandoc_msg(MANDOCERR_STR_UNDEF, ln, iesc,
                   1457:                            "%.*s", iendarg - iarg, buf->buf + iarg);
                   1458:                        break;
1.108     schwarze 1459:
1.339     schwarze 1460:                case '$':
                   1461:                        if (r->mstackpos < 0) {
1.388     schwarze 1462:                                mandoc_msg(MANDOCERR_ARG_UNDEF, ln, iesc,
                   1463:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1464:                                break;
                   1465:                        }
                   1466:                        ctx = r->mstack + r->mstackpos;
1.388     schwarze 1467:                        argi = buf->buf[iarg] - '1';
                   1468:                        if (argi >= 0 && argi <= 8) {
                   1469:                                if (argi < ctx->argc)
                   1470:                                        res = ctx->argv[argi];
1.339     schwarze 1471:                                break;
                   1472:                        }
1.388     schwarze 1473:                        if (buf->buf[iarg] == '*')
1.339     schwarze 1474:                                quote_args = 0;
1.388     schwarze 1475:                        else if (buf->buf[iarg] == '@')
1.339     schwarze 1476:                                quote_args = 1;
                   1477:                        else {
1.388     schwarze 1478:                                mandoc_msg(MANDOCERR_ARG_NONUM, ln, iesc,
                   1479:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1480:                                break;
                   1481:                        }
                   1482:                        asz = 0;
1.388     schwarze 1483:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1484:                                if (argi)
1.339     schwarze 1485:                                        asz++;  /* blank */
                   1486:                                if (quote_args)
                   1487:                                        asz += 2;  /* quotes */
1.388     schwarze 1488:                                asz += strlen(ctx->argv[argi]);
1.339     schwarze 1489:                        }
1.388     schwarze 1490:                        if (asz != iend - iesc) {
                   1491:                                rsz = buf->sz - iend;
                   1492:                                if (asz < iend - iesc)
                   1493:                                        memmove(buf->buf + iesc + asz,
                   1494:                                            buf->buf + iend, rsz);
                   1495:                                buf->sz = iesc + asz + rsz;
                   1496:                                buf->buf = mandoc_realloc(buf->buf, buf->sz);
                   1497:                                if (asz > iend - iesc)
                   1498:                                        memmove(buf->buf + iesc + asz,
                   1499:                                            buf->buf + iend, rsz);
1.339     schwarze 1500:                        }
1.388     schwarze 1501:                        dst = buf->buf + iesc;
                   1502:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1503:                                if (argi)
                   1504:                                        *dst++ = ' ';
1.339     schwarze 1505:                                if (quote_args)
1.388     schwarze 1506:                                        *dst++ = '"';
                   1507:                                src = ctx->argv[argi];
                   1508:                                while (*src != '\0')
                   1509:                                        *dst++ = *src++;
1.339     schwarze 1510:                                if (quote_args)
1.388     schwarze 1511:                                        *dst++ = '"';
1.339     schwarze 1512:                        }
                   1513:                        continue;
1.391     schwarze 1514:                case 'A':
                   1515:                        ubuf[0] = iendarg > iarg ? '1' : '0';
                   1516:                        ubuf[1] = '\0';
                   1517:                        res = ubuf;
                   1518:                        break;
1.207     schwarze 1519:                case 'B':
1.206     schwarze 1520:                        npos = 0;
1.388     schwarze 1521:                        ubuf[0] = iendarg > iarg && iend > iendarg &&
                   1522:                            roff_evalnum(r, ln, buf->buf + iarg, &npos,
                   1523:                                         NULL, ROFFNUM_SCALE) &&
                   1524:                            npos == iendarg - iarg ? '1' : '0';
1.206     schwarze 1525:                        ubuf[1] = '\0';
1.388     schwarze 1526:                        res = ubuf;
1.206     schwarze 1527:                        break;
1.389     schwarze 1528:                case 'V':
                   1529:                        mandoc_msg(MANDOCERR_UNSUPP, ln, iesc,
                   1530:                            "%.*s", iend - iesc, buf->buf + iesc);
                   1531:                        roff_expand_patch(buf, iendarg, "}", iend);
                   1532:                        roff_expand_patch(buf, iesc, "${", iarg);
                   1533:                        continue;
1.390     schwarze 1534:                case 'g':
                   1535:                        break;
1.207     schwarze 1536:                case 'n':
1.388     schwarze 1537:                        if (iendarg > iarg)
1.218     schwarze 1538:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.388     schwarze 1539:                                    roff_getregn(r, buf->buf + iarg,
                   1540:                                    iendarg - iarg, buf->buf[inam + 1]));
1.218     schwarze 1541:                        else
                   1542:                                ubuf[0] = '\0';
1.388     schwarze 1543:                        res = ubuf;
1.206     schwarze 1544:                        break;
1.207     schwarze 1545:                case 'w':
1.388     schwarze 1546:                        (void)snprintf(ubuf, sizeof(ubuf),
                   1547:                            "%d", (iendarg - iarg) * 24);
                   1548:                        res = ubuf;
                   1549:                        break;
                   1550:                default:
1.206     schwarze 1551:                        break;
                   1552:                }
1.388     schwarze 1553:                if (res == NULL)
1.142     kristaps 1554:                        res = "";
1.388     schwarze 1555:                if (++expand_count > EXPAND_LIMIT ||
                   1556:                    buf->sz + strlen(res) > SHRT_MAX) {
                   1557:                        mandoc_msg(MANDOCERR_ROFFLOOP, ln, iesc, NULL);
1.277     schwarze 1558:                        return ROFF_IGN;
1.94      kristaps 1559:                }
1.388     schwarze 1560:                roff_expand_patch(buf, iesc, res, iend);
                   1561:        }
                   1562:        return ROFF_CONT;
                   1563: }
1.94      kristaps 1564:
1.388     schwarze 1565: /*
                   1566:  * Replace the substring from the start position (inclusive)
                   1567:  * to end position (exclusive) with the repl(acement) string.
                   1568:  */
                   1569: static void
                   1570: roff_expand_patch(struct buf *buf, int start, const char *repl, int end)
                   1571: {
                   1572:        char    *nbuf;
1.108     schwarze 1573:
1.389     schwarze 1574:        buf->sz = mandoc_asprintf(&nbuf, "%.*s%s%s", start, buf->buf,
                   1575:            repl, buf->buf + end) + 1;
1.388     schwarze 1576:        free(buf->buf);
                   1577:        buf->buf = nbuf;
1.154     kristaps 1578: }
1.353     schwarze 1579:
                   1580: /*
                   1581:  * Parse a quoted or unquoted roff-style request or macro argument.
                   1582:  * Return a pointer to the parsed argument, which is either the original
                   1583:  * pointer or advanced by one byte in case the argument is quoted.
                   1584:  * NUL-terminate the argument in place.
                   1585:  * Collapse pairs of quotes inside quoted arguments.
                   1586:  * Advance the argument pointer to the next argument,
                   1587:  * or to the NUL byte terminating the argument line.
                   1588:  */
                   1589: char *
1.355     schwarze 1590: roff_getarg(struct roff *r, char **cpp, int ln, int *pos)
1.353     schwarze 1591: {
1.355     schwarze 1592:        struct buf       buf;
1.366     schwarze 1593:        char            *cp, *start;
1.355     schwarze 1594:        int              newesc, pairs, quoted, white;
1.353     schwarze 1595:
                   1596:        /* Quoting can only start with a new word. */
                   1597:        start = *cpp;
                   1598:        quoted = 0;
                   1599:        if ('"' == *start) {
                   1600:                quoted = 1;
                   1601:                start++;
                   1602:        }
                   1603:
1.355     schwarze 1604:        newesc = pairs = white = 0;
1.353     schwarze 1605:        for (cp = start; '\0' != *cp; cp++) {
                   1606:
                   1607:                /*
                   1608:                 * Move the following text left
                   1609:                 * after quoted quotes and after "\\" and "\t".
                   1610:                 */
                   1611:                if (pairs)
                   1612:                        cp[-pairs] = cp[0];
                   1613:
                   1614:                if ('\\' == cp[0]) {
                   1615:                        /*
                   1616:                         * In copy mode, translate double to single
                   1617:                         * backslashes and backslash-t to literal tabs.
                   1618:                         */
                   1619:                        switch (cp[1]) {
                   1620:                        case 'a':
                   1621:                        case 't':
1.354     schwarze 1622:                                cp[-pairs] = '\t';
1.355     schwarze 1623:                                pairs++;
                   1624:                                cp++;
                   1625:                                break;
1.353     schwarze 1626:                        case '\\':
1.355     schwarze 1627:                                newesc = 1;
                   1628:                                cp[-pairs] = ASCII_ESC;
1.353     schwarze 1629:                                pairs++;
                   1630:                                cp++;
                   1631:                                break;
                   1632:                        case ' ':
                   1633:                                /* Skip escaped blanks. */
                   1634:                                if (0 == quoted)
                   1635:                                        cp++;
                   1636:                                break;
                   1637:                        default:
                   1638:                                break;
                   1639:                        }
                   1640:                } else if (0 == quoted) {
                   1641:                        if (' ' == cp[0]) {
                   1642:                                /* Unescaped blanks end unquoted args. */
                   1643:                                white = 1;
                   1644:                                break;
                   1645:                        }
                   1646:                } else if ('"' == cp[0]) {
                   1647:                        if ('"' == cp[1]) {
                   1648:                                /* Quoted quotes collapse. */
                   1649:                                pairs++;
                   1650:                                cp++;
                   1651:                        } else {
                   1652:                                /* Unquoted quotes end quoted args. */
                   1653:                                quoted = 2;
                   1654:                                break;
                   1655:                        }
                   1656:                }
                   1657:        }
                   1658:
                   1659:        /* Quoted argument without a closing quote. */
                   1660:        if (1 == quoted)
                   1661:                mandoc_msg(MANDOCERR_ARG_QUOTE, ln, *pos, NULL);
                   1662:
                   1663:        /* NUL-terminate this argument and move to the next one. */
                   1664:        if (pairs)
                   1665:                cp[-pairs] = '\0';
                   1666:        if ('\0' != *cp) {
                   1667:                *cp++ = '\0';
                   1668:                while (' ' == *cp)
                   1669:                        cp++;
                   1670:        }
                   1671:        *pos += (int)(cp - start) + (quoted ? 1 : 0);
                   1672:        *cpp = cp;
                   1673:
                   1674:        if ('\0' == *cp && (white || ' ' == cp[-1]))
                   1675:                mandoc_msg(MANDOCERR_SPACE_EOL, ln, *pos, NULL);
                   1676:
1.355     schwarze 1677:        start = mandoc_strdup(start);
                   1678:        if (newesc == 0)
                   1679:                return start;
                   1680:
                   1681:        buf.buf = start;
                   1682:        buf.sz = strlen(start) + 1;
                   1683:        buf.next = NULL;
                   1684:        if (roff_expand(r, &buf, ln, 0, ASCII_ESC) & ROFF_IGN) {
                   1685:                free(buf.buf);
                   1686:                buf.buf = mandoc_strdup("");
                   1687:        }
                   1688:        return buf.buf;
1.353     schwarze 1689: }
                   1690:
1.154     kristaps 1691:
                   1692: /*
1.275     schwarze 1693:  * Process text streams.
1.154     kristaps 1694:  */
1.340     schwarze 1695: static int
1.305     schwarze 1696: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154     kristaps 1697: {
                   1698:        size_t           sz;
                   1699:        const char      *start;
1.178     schwarze 1700:        char            *p;
                   1701:        int              isz;
1.154     kristaps 1702:        enum mandoc_esc  esc;
                   1703:
1.275     schwarze 1704:        /* Spring the input line trap. */
                   1705:
                   1706:        if (roffit_lines == 1) {
                   1707:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1708:                free(buf->buf);
                   1709:                buf->buf = p;
                   1710:                buf->sz = isz + 1;
                   1711:                *offs = 0;
                   1712:                free(roffit_macro);
                   1713:                roffit_lines = 0;
1.277     schwarze 1714:                return ROFF_REPARSE;
1.275     schwarze 1715:        } else if (roffit_lines > 1)
                   1716:                --roffit_lines;
                   1717:
1.305     schwarze 1718:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1719:                if (roffce_lines < 1) {
                   1720:                        r->man->last = roffce_node;
                   1721:                        r->man->next = ROFF_NEXT_SIBLING;
                   1722:                        roffce_lines = 0;
                   1723:                        roffce_node = NULL;
                   1724:                } else
                   1725:                        roffce_lines--;
                   1726:        }
                   1727:
1.275     schwarze 1728:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1729:
1.238     schwarze 1730:        start = p = buf->buf + pos;
1.154     kristaps 1731:
1.238     schwarze 1732:        while (*p != '\0') {
1.154     kristaps 1733:                sz = strcspn(p, "-\\");
                   1734:                p += sz;
                   1735:
1.238     schwarze 1736:                if (*p == '\0')
1.159     kristaps 1737:                        break;
                   1738:
1.238     schwarze 1739:                if (*p == '\\') {
1.154     kristaps 1740:                        /* Skip over escapes. */
                   1741:                        p++;
1.189     schwarze 1742:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238     schwarze 1743:                        if (esc == ESCAPE_ERROR)
1.154     kristaps 1744:                                break;
1.264     schwarze 1745:                        while (*p == '-')
                   1746:                                p++;
1.155     kristaps 1747:                        continue;
1.159     kristaps 1748:                } else if (p == start) {
1.158     kristaps 1749:                        p++;
1.155     kristaps 1750:                        continue;
1.158     kristaps 1751:                }
1.155     kristaps 1752:
1.171     schwarze 1753:                if (isalpha((unsigned char)p[-1]) &&
                   1754:                    isalpha((unsigned char)p[1]))
1.155     kristaps 1755:                        *p = ASCII_HYPH;
                   1756:                p++;
1.94      kristaps 1757:        }
1.277     schwarze 1758:        return ROFF_CONT;
1.94      kristaps 1759: }
                   1760:
1.340     schwarze 1761: int
1.377     schwarze 1762: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
1.67      kristaps 1763: {
1.294     schwarze 1764:        enum roff_tok    t;
1.340     schwarze 1765:        int              e;
1.238     schwarze 1766:        int              pos;   /* parse point */
1.243     schwarze 1767:        int              spos;  /* saved parse point for messages */
1.238     schwarze 1768:        int              ppos;  /* original offset in buf->buf */
                   1769:        int              ctl;   /* macro line (boolean) */
                   1770:
                   1771:        ppos = pos = *offs;
1.377     schwarze 1772:
                   1773:        if (len > 80 && r->tbl == NULL && r->eqn == NULL &&
                   1774:            (r->man->flags & ROFF_NOFILL) == 0 &&
                   1775:            strchr(" .\\", buf->buf[pos]) == NULL &&
                   1776:            buf->buf[pos] != r->control &&
                   1777:            strcspn(buf->buf, " ") < 80)
                   1778:                mandoc_msg(MANDOCERR_TEXT_LONG, ln, (int)len - 1,
                   1779:                    "%.20s...", buf->buf + pos);
1.79      kristaps 1780:
1.230     schwarze 1781:        /* Handle in-line equation delimiters. */
                   1782:
1.236     schwarze 1783:        if (r->tbl == NULL &&
                   1784:            r->last_eqn != NULL && r->last_eqn->delim &&
1.230     schwarze 1785:            (r->eqn == NULL || r->eqn_inline)) {
1.238     schwarze 1786:                e = roff_eqndelim(r, buf, pos);
1.230     schwarze 1787:                if (e == ROFF_REPARSE)
1.277     schwarze 1788:                        return e;
1.230     schwarze 1789:                assert(e == ROFF_CONT);
                   1790:        }
                   1791:
1.387     schwarze 1792:        /* Handle comments and escape sequences. */
                   1793:
                   1794:        e = roff_parse_comment(r, buf, ln, pos, r->escape);
                   1795:        if ((e & ROFF_MASK) == ROFF_IGN)
                   1796:                return e;
                   1797:        assert(e == ROFF_CONT);
1.94      kristaps 1798:
1.355     schwarze 1799:        e = roff_expand(r, buf, ln, pos, r->escape);
1.340     schwarze 1800:        if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1801:                return e;
1.238     schwarze 1802:        assert(e == ROFF_CONT);
1.94      kristaps 1803:
1.238     schwarze 1804:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.130     kristaps 1805:
1.94      kristaps 1806:        /*
1.79      kristaps 1807:         * First, if a scope is open and we're not a macro, pass the
1.252     schwarze 1808:         * text through the macro's filter.
                   1809:         * Equations process all content themselves.
                   1810:         * Tables process almost all content themselves, but we want
                   1811:         * to warn about macros before passing it there.
1.79      kristaps 1812:         */
1.74      kristaps 1813:
1.252     schwarze 1814:        if (r->last != NULL && ! ctl) {
1.78      kristaps 1815:                t = r->last->tok;
1.238     schwarze 1816:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.340     schwarze 1817:                if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1818:                        return e;
1.340     schwarze 1819:                e &= ~ROFF_MASK;
                   1820:        } else
                   1821:                e = ROFF_IGN;
1.319     schwarze 1822:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1823:                eqn_read(r->eqn, buf->buf + ppos);
1.340     schwarze 1824:                return e;
1.319     schwarze 1825:        }
1.321     schwarze 1826:        if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
                   1827:                tbl_read(r->tbl, ln, buf->buf, ppos);
1.346     schwarze 1828:                roff_addtbl(r->man, ln, r->tbl);
1.340     schwarze 1829:                return e;
1.321     schwarze 1830:        }
1.367     schwarze 1831:        if ( ! ctl) {
                   1832:                r->options &= ~MPARSE_COMMENT;
1.340     schwarze 1833:                return roff_parsetext(r, buf, pos, offs) | e;
1.367     schwarze 1834:        }
1.228     schwarze 1835:
                   1836:        /* Skip empty request lines. */
                   1837:
1.238     schwarze 1838:        if (buf->buf[pos] == '"') {
1.350     schwarze 1839:                mandoc_msg(MANDOCERR_COMMENT_BAD, ln, pos, NULL);
1.277     schwarze 1840:                return ROFF_IGN;
1.238     schwarze 1841:        } else if (buf->buf[pos] == '\0')
1.277     schwarze 1842:                return ROFF_IGN;
1.67      kristaps 1843:
1.79      kristaps 1844:        /*
                   1845:         * If a scope is open, go to the child handler for that macro,
                   1846:         * as it may want to preprocess before doing anything with it.
                   1847:         */
1.78      kristaps 1848:
1.79      kristaps 1849:        if (r->last) {
                   1850:                t = r->last->tok;
1.277     schwarze 1851:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1852:        }
1.78      kristaps 1853:
1.367     schwarze 1854:        r->options &= ~MPARSE_COMMENT;
1.243     schwarze 1855:        spos = pos;
                   1856:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.386     schwarze 1857:        return roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
                   1858: }
                   1859:
                   1860: /*
                   1861:  * Handle a new request or macro.
                   1862:  * May be called outside any scope or from inside a conditional scope.
                   1863:  */
                   1864: static int
                   1865: roff_req_or_macro(ROFF_ARGS) {
1.243     schwarze 1866:
1.386     schwarze 1867:        /* For now, tables ignore most macros and some request. */
1.243     schwarze 1868:
1.386     schwarze 1869:        if (r->tbl != NULL && (tok == TOKEN_NONE || tok == ROFF_TS ||
                   1870:            tok == ROFF_br || tok == ROFF_ce || tok == ROFF_rj ||
                   1871:            tok == ROFF_sp)) {
1.350     schwarze 1872:                mandoc_msg(MANDOCERR_TBLMACRO,
1.386     schwarze 1873:                    ln, ppos, "%s", buf->buf + ppos);
                   1874:                if (tok != TOKEN_NONE)
1.277     schwarze 1875:                        return ROFF_IGN;
1.257     schwarze 1876:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1877:                        pos++;
1.291     schwarze 1878:                while (buf->buf[pos] == ' ')
1.257     schwarze 1879:                        pos++;
1.321     schwarze 1880:                tbl_read(r->tbl, ln, buf->buf, pos);
1.346     schwarze 1881:                roff_addtbl(r->man, ln, r->tbl);
1.321     schwarze 1882:                return ROFF_IGN;
1.243     schwarze 1883:        }
                   1884:
1.305     schwarze 1885:        /* For now, let high level macros abort .ce mode. */
                   1886:
1.386     schwarze 1887:        if (roffce_node != NULL &&
                   1888:            (tok == TOKEN_NONE || tok == ROFF_Dd || tok == ROFF_EQ ||
                   1889:             tok == ROFF_TH || tok == ROFF_TS)) {
1.305     schwarze 1890:                r->man->last = roffce_node;
                   1891:                r->man->next = ROFF_NEXT_SIBLING;
                   1892:                roffce_lines = 0;
                   1893:                roffce_node = NULL;
                   1894:        }
                   1895:
1.79      kristaps 1896:        /*
1.243     schwarze 1897:         * This is neither a roff request nor a user-defined macro.
                   1898:         * Let the standard macro set parsers handle it.
1.79      kristaps 1899:         */
1.67      kristaps 1900:
1.386     schwarze 1901:        if (tok == TOKEN_NONE)
1.277     schwarze 1902:                return ROFF_CONT;
1.243     schwarze 1903:
1.386     schwarze 1904:        /* Execute a roff request or a user-defined macro. */
1.67      kristaps 1905:
1.386     schwarze 1906:        return (*roffs[tok].proc)(r, tok, buf, ln, ppos, pos, offs);
1.74      kristaps 1907: }
                   1908:
1.339     schwarze 1909: /*
                   1910:  * Internal interface function to tell the roff parser that execution
                   1911:  * of the current macro ended.  This is required because macro
                   1912:  * definitions usually do not end with a .return request.
                   1913:  */
                   1914: void
                   1915: roff_userret(struct roff *r)
                   1916: {
                   1917:        struct mctx     *ctx;
                   1918:        int              i;
                   1919:
                   1920:        assert(r->mstackpos >= 0);
                   1921:        ctx = r->mstack + r->mstackpos;
                   1922:        for (i = 0; i < ctx->argc; i++)
                   1923:                free(ctx->argv[i]);
                   1924:        ctx->argc = 0;
                   1925:        r->mstackpos--;
                   1926: }
                   1927:
1.117     kristaps 1928: void
1.74      kristaps 1929: roff_endparse(struct roff *r)
                   1930: {
1.321     schwarze 1931:        if (r->last != NULL)
1.350     schwarze 1932:                mandoc_msg(MANDOCERR_BLK_NOEND, r->last->line,
                   1933:                    r->last->col, "%s", roff_name[r->last->tok]);
1.117     kristaps 1934:
1.321     schwarze 1935:        if (r->eqn != NULL) {
1.350     schwarze 1936:                mandoc_msg(MANDOCERR_BLK_NOEND,
1.319     schwarze 1937:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1938:                eqn_parse(r->eqn);
                   1939:                r->eqn = NULL;
1.125     kristaps 1940:        }
                   1941:
1.321     schwarze 1942:        if (r->tbl != NULL) {
1.346     schwarze 1943:                tbl_end(r->tbl, 1);
1.321     schwarze 1944:                r->tbl = NULL;
1.117     kristaps 1945:        }
1.67      kristaps 1946: }
                   1947:
                   1948: /*
1.386     schwarze 1949:  * Parse the request or macro name at buf[*pos].
                   1950:  * Return ROFF_RENAMED, ROFF_USERDEF, or a ROFF_* token value.
                   1951:  * For empty, undefined, mdoc(7), and man(7) macros, return TOKEN_NONE.
                   1952:  * As a side effect, set r->current_string to the definition or to NULL.
1.67      kristaps 1953:  */
1.294     schwarze 1954: static enum roff_tok
1.214     schwarze 1955: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67      kristaps 1956: {
1.214     schwarze 1957:        char            *cp;
1.106     kristaps 1958:        const char      *mac;
                   1959:        size_t           maclen;
1.315     schwarze 1960:        int              deftype;
1.294     schwarze 1961:        enum roff_tok    t;
1.67      kristaps 1962:
1.214     schwarze 1963:        cp = buf + *pos;
                   1964:
                   1965:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294     schwarze 1966:                return TOKEN_NONE;
1.67      kristaps 1967:
1.214     schwarze 1968:        mac = cp;
                   1969:        maclen = roff_getname(r, &cp, ln, ppos);
1.67      kristaps 1970:
1.315     schwarze 1971:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1972:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1973:        switch (deftype) {
                   1974:        case ROFFDEF_USER:
                   1975:                t = ROFF_USERDEF;
                   1976:                break;
                   1977:        case ROFFDEF_REN:
                   1978:                t = ROFF_RENAMED;
                   1979:                break;
                   1980:        default:
                   1981:                t = roffhash_find(r->reqtab, mac, maclen);
                   1982:                break;
                   1983:        }
1.294     schwarze 1984:        if (t != TOKEN_NONE)
1.214     schwarze 1985:                *pos = cp - buf;
1.325     schwarze 1986:        else if (deftype == ROFFDEF_UNDEF) {
                   1987:                /* Using an undefined macro defines it to be empty. */
                   1988:                roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
                   1989:                roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
                   1990:        }
1.277     schwarze 1991:        return t;
1.67      kristaps 1992: }
                   1993:
1.266     schwarze 1994: /* --- handling of request blocks ----------------------------------------- */
                   1995:
1.375     schwarze 1996: /*
                   1997:  * Close a macro definition block or an "ignore" block.
                   1998:  */
1.340     schwarze 1999: static int
1.76      kristaps 2000: roff_cblock(ROFF_ARGS)
1.67      kristaps 2001: {
1.375     schwarze 2002:        int      rr;
1.79      kristaps 2003:
1.238     schwarze 2004:        if (r->last == NULL) {
1.350     schwarze 2005:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.277     schwarze 2006:                return ROFF_IGN;
1.76      kristaps 2007:        }
1.67      kristaps 2008:
1.81      kristaps 2009:        switch (r->last->tok) {
1.207     schwarze 2010:        case ROFF_am:
                   2011:        case ROFF_ami:
                   2012:        case ROFF_de:
                   2013:        case ROFF_dei:
                   2014:        case ROFF_ig:
1.81      kristaps 2015:                break;
1.375     schwarze 2016:        case ROFF_am1:
                   2017:        case ROFF_de1:
                   2018:                /* Remapped in roff_block(). */
                   2019:                abort();
1.81      kristaps 2020:        default:
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:
1.375     schwarze 2025:        roffnode_pop(r);
                   2026:        roffnode_cleanscope(r);
                   2027:
                   2028:        /*
                   2029:         * If a conditional block with braces is still open,
                   2030:         * check for "\}" block end markers.
                   2031:         */
                   2032:
                   2033:        if (r->last != NULL && r->last->endspan < 0) {
                   2034:                rr = 1;  /* If arguments follow "\}", warn about them. */
                   2035:                roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
                   2036:        }
                   2037:
1.238     schwarze 2038:        if (buf->buf[pos] != '\0')
1.350     schwarze 2039:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 2040:                    ".. %s", buf->buf + pos);
1.71      kristaps 2041:
1.277     schwarze 2042:        return ROFF_IGN;
1.67      kristaps 2043: }
                   2044:
1.364     schwarze 2045: /*
                   2046:  * Pop all nodes ending at the end of the current input line.
                   2047:  * Return the number of loops ended.
                   2048:  */
1.340     schwarze 2049: static int
1.76      kristaps 2050: roffnode_cleanscope(struct roff *r)
1.67      kristaps 2051: {
1.340     schwarze 2052:        int inloop;
1.67      kristaps 2053:
1.340     schwarze 2054:        inloop = 0;
1.375     schwarze 2055:        while (r->last != NULL && r->last->endspan > 0) {
1.173     schwarze 2056:                if (--r->last->endspan != 0)
1.76      kristaps 2057:                        break;
1.340     schwarze 2058:                inloop += roffnode_pop(r);
1.76      kristaps 2059:        }
1.340     schwarze 2060:        return inloop;
1.67      kristaps 2061: }
                   2062:
1.364     schwarze 2063: /*
1.375     schwarze 2064:  * Handle the closing "\}" of a conditional block.
1.364     schwarze 2065:  * Apart from generating warnings, this only pops nodes.
                   2066:  * Return the number of loops ended.
                   2067:  */
1.340     schwarze 2068: static int
1.195     schwarze 2069: roff_ccond(struct roff *r, int ln, int ppos)
1.74      kristaps 2070: {
1.76      kristaps 2071:        if (NULL == r->last) {
1.350     schwarze 2072:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2073:                return 0;
1.76      kristaps 2074:        }
                   2075:
1.82      kristaps 2076:        switch (r->last->tok) {
1.207     schwarze 2077:        case ROFF_el:
                   2078:        case ROFF_ie:
                   2079:        case ROFF_if:
1.340     schwarze 2080:        case ROFF_while:
1.82      kristaps 2081:                break;
                   2082:        default:
1.350     schwarze 2083:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2084:                return 0;
1.75      kristaps 2085:        }
                   2086:
1.76      kristaps 2087:        if (r->last->endspan > -1) {
1.350     schwarze 2088:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2089:                return 0;
1.76      kristaps 2090:        }
                   2091:
1.340     schwarze 2092:        return roffnode_pop(r) + roffnode_cleanscope(r);
1.76      kristaps 2093: }
                   2094:
1.340     schwarze 2095: static int
1.80      kristaps 2096: roff_block(ROFF_ARGS)
1.76      kristaps 2097: {
1.315     schwarze 2098:        const char      *name, *value;
                   2099:        char            *call, *cp, *iname, *rname;
                   2100:        size_t           csz, namesz, rsz;
                   2101:        int              deftype;
1.106     kristaps 2102:
1.220     schwarze 2103:        /* Ignore groff compatibility mode for now. */
1.76      kristaps 2104:
1.238     schwarze 2105:        if (tok == ROFF_de1)
1.220     schwarze 2106:                tok = ROFF_de;
1.251     schwarze 2107:        else if (tok == ROFF_dei1)
                   2108:                tok = ROFF_dei;
1.238     schwarze 2109:        else if (tok == ROFF_am1)
1.220     schwarze 2110:                tok = ROFF_am;
1.251     schwarze 2111:        else if (tok == ROFF_ami1)
                   2112:                tok = ROFF_ami;
1.220     schwarze 2113:
                   2114:        /* Parse the macro name argument. */
                   2115:
1.238     schwarze 2116:        cp = buf->buf + pos;
                   2117:        if (tok == ROFF_ig) {
1.220     schwarze 2118:                iname = NULL;
                   2119:                namesz = 0;
                   2120:        } else {
                   2121:                iname = cp;
                   2122:                namesz = roff_getname(r, &cp, ln, ppos);
                   2123:                iname[namesz] = '\0';
                   2124:        }
1.107     kristaps 2125:
1.220     schwarze 2126:        /* Resolve the macro name argument if it is indirect. */
1.107     kristaps 2127:
1.238     schwarze 2128:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2129:                deftype = ROFFDEF_USER;
                   2130:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2131:                if (name == NULL) {
1.350     schwarze 2132:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2133:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2134:                            "%.*s", (int)namesz, iname);
                   2135:                        namesz = 0;
                   2136:                } else
                   2137:                        namesz = strlen(name);
                   2138:        } else
                   2139:                name = iname;
1.107     kristaps 2140:
1.238     schwarze 2141:        if (namesz == 0 && tok != ROFF_ig) {
1.350     schwarze 2142:                mandoc_msg(MANDOCERR_REQ_EMPTY,
                   2143:                    ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2144:                return ROFF_IGN;
1.220     schwarze 2145:        }
1.80      kristaps 2146:
1.106     kristaps 2147:        roffnode_push(r, tok, name, ln, ppos);
                   2148:
                   2149:        /*
                   2150:         * At the beginning of a `de' macro, clear the existing string
                   2151:         * with the same name, if there is one.  New content will be
1.193     schwarze 2152:         * appended from roff_block_text() in multiline mode.
1.106     kristaps 2153:         */
1.107     kristaps 2154:
1.311     schwarze 2155:        if (tok == ROFF_de || tok == ROFF_dei) {
1.213     schwarze 2156:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311     schwarze 2157:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315     schwarze 2158:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   2159:                deftype = ROFFDEF_ANY;
                   2160:                value = roff_getstrn(r, iname, namesz, &deftype);
                   2161:                switch (deftype) {  /* Before appending, ... */
                   2162:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   2163:                        roff_setstrn(&r->strtab, name, namesz,
                   2164:                            value, strlen(value), 0);
                   2165:                        break;
                   2166:                case ROFFDEF_REN: /* call original standard macro. */
                   2167:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2168:                            (int)strlen(value), value);
                   2169:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2170:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   2171:                        free(call);
                   2172:                        break;
                   2173:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   2174:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   2175:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   2176:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2177:                            (int)rsz, rname);
                   2178:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2179:                        free(call);
                   2180:                        free(rname);
                   2181:                        break;
                   2182:                default:
                   2183:                        break;
                   2184:                }
1.311     schwarze 2185:        }
1.76      kristaps 2186:
1.238     schwarze 2187:        if (*cp == '\0')
1.277     schwarze 2188:                return ROFF_IGN;
1.78      kristaps 2189:
1.220     schwarze 2190:        /* Get the custom end marker. */
1.107     kristaps 2191:
1.220     schwarze 2192:        iname = cp;
1.213     schwarze 2193:        namesz = roff_getname(r, &cp, ln, ppos);
1.220     schwarze 2194:
                   2195:        /* Resolve the end marker if it is indirect. */
                   2196:
1.238     schwarze 2197:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2198:                deftype = ROFFDEF_USER;
                   2199:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2200:                if (name == NULL) {
1.350     schwarze 2201:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2202:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2203:                            "%.*s", (int)namesz, iname);
                   2204:                        namesz = 0;
                   2205:                } else
                   2206:                        namesz = strlen(name);
                   2207:        } else
                   2208:                name = iname;
                   2209:
1.213     schwarze 2210:        if (namesz)
                   2211:                r->last->end = mandoc_strndup(name, namesz);
1.78      kristaps 2212:
1.238     schwarze 2213:        if (*cp != '\0')
1.350     schwarze 2214:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.295     schwarze 2215:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74      kristaps 2216:
1.277     schwarze 2217:        return ROFF_IGN;
1.78      kristaps 2218: }
                   2219:
1.340     schwarze 2220: static int
1.80      kristaps 2221: roff_block_sub(ROFF_ARGS)
1.79      kristaps 2222: {
1.294     schwarze 2223:        enum roff_tok   t;
1.79      kristaps 2224:        int             i, j;
                   2225:
                   2226:        /*
1.385     schwarze 2227:         * If a custom end marker is a user-defined or predefined macro
                   2228:         * or a request, interpret it.
1.79      kristaps 2229:         */
                   2230:
                   2231:        if (r->last->end) {
1.130     kristaps 2232:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238     schwarze 2233:                        if (buf->buf[i] != r->last->end[j])
1.79      kristaps 2234:                                break;
                   2235:
1.238     schwarze 2236:                if (r->last->end[j] == '\0' &&
                   2237:                    (buf->buf[i] == '\0' ||
                   2238:                     buf->buf[i] == ' ' ||
                   2239:                     buf->buf[i] == '\t')) {
1.79      kristaps 2240:                        roffnode_pop(r);
                   2241:                        roffnode_cleanscope(r);
                   2242:
1.238     schwarze 2243:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130     kristaps 2244:                                i++;
                   2245:
                   2246:                        pos = i;
1.238     schwarze 2247:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294     schwarze 2248:                            TOKEN_NONE)
1.277     schwarze 2249:                                return ROFF_RERUN;
                   2250:                        return ROFF_IGN;
1.79      kristaps 2251:                }
                   2252:        }
                   2253:
1.385     schwarze 2254:        /* Handle the standard end marker. */
1.79      kristaps 2255:
1.238     schwarze 2256:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.385     schwarze 2257:        if (t == ROFF_cblock)
                   2258:                return roff_cblock(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 2259:
1.385     schwarze 2260:        /* Not an end marker, so append the line to the block. */
1.79      kristaps 2261:
1.385     schwarze 2262:        if (tok != ROFF_ig)
                   2263:                roff_setstr(r, r->last->name, buf->buf + ppos, 2);
                   2264:        return ROFF_IGN;
1.79      kristaps 2265: }
                   2266:
1.340     schwarze 2267: static int
1.80      kristaps 2268: roff_block_text(ROFF_ARGS)
1.78      kristaps 2269: {
                   2270:
1.238     schwarze 2271:        if (tok != ROFF_ig)
                   2272:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106     kristaps 2273:
1.277     schwarze 2274:        return ROFF_IGN;
1.78      kristaps 2275: }
                   2276:
1.375     schwarze 2277: /*
                   2278:  * Check for a closing "\}" and handle it.
                   2279:  * In this function, the final "int *offs" argument is used for
                   2280:  * different purposes than elsewhere:
                   2281:  * Input: *offs == 0: caller wants to discard arguments following \}
                   2282:  *        *offs == 1: caller wants to preserve text following \}
                   2283:  * Output: *offs = 0: tell caller to discard input line
                   2284:  *         *offs = 1: tell caller to use input line
                   2285:  */
1.340     schwarze 2286: static int
1.375     schwarze 2287: roff_cond_checkend(ROFF_ARGS)
1.82      kristaps 2288: {
1.340     schwarze 2289:        char            *ep;
                   2290:        int              endloop, irc, rr;
1.82      kristaps 2291:
1.340     schwarze 2292:        irc = ROFF_IGN;
1.82      kristaps 2293:        rr = r->last->rule;
1.340     schwarze 2294:        endloop = tok != ROFF_while ? ROFF_IGN :
                   2295:            rr ? ROFF_LOOPCONT : ROFF_LOOPEXIT;
                   2296:        if (roffnode_cleanscope(r))
                   2297:                irc |= endloop;
1.144     kristaps 2298:
1.196     schwarze 2299:        /*
1.375     schwarze 2300:         * If "\}" occurs on a macro line without a preceding macro or
                   2301:         * a text line contains nothing else, drop the line completely.
1.196     schwarze 2302:         */
                   2303:
1.238     schwarze 2304:        ep = buf->buf + pos;
1.375     schwarze 2305:        if (ep[0] == '\\' && ep[1] == '}' && (ep[2] == '\0' || *offs == 0))
1.198     schwarze 2306:                rr = 0;
1.196     schwarze 2307:
1.343     schwarze 2308:        /*
1.375     schwarze 2309:         * The closing delimiter "\}" rewinds the conditional scope
1.343     schwarze 2310:         * but is otherwise ignored when interpreting the line.
                   2311:         */
1.144     kristaps 2312:
1.238     schwarze 2313:        while ((ep = strchr(ep, '\\')) != NULL) {
1.318     schwarze 2314:                switch (ep[1]) {
                   2315:                case '}':
1.376     schwarze 2316:                        if (ep[2] == '\0')
                   2317:                                ep[0] = '\0';
                   2318:                        else if (rr)
1.375     schwarze 2319:                                ep[1] = '&';
                   2320:                        else
                   2321:                                memmove(ep, ep + 2, strlen(ep + 2) + 1);
1.340     schwarze 2322:                        if (roff_ccond(r, ln, ep - buf->buf))
                   2323:                                irc |= endloop;
1.318     schwarze 2324:                        break;
                   2325:                case '\0':
                   2326:                        ++ep;
                   2327:                        break;
                   2328:                default:
                   2329:                        ep += 2;
                   2330:                        break;
1.197     schwarze 2331:                }
1.177     schwarze 2332:        }
1.375     schwarze 2333:        *offs = rr;
                   2334:        return irc;
                   2335: }
1.318     schwarze 2336:
1.375     schwarze 2337: /*
                   2338:  * Parse and process a request or macro line in conditional scope.
                   2339:  */
                   2340: static int
                   2341: roff_cond_sub(ROFF_ARGS)
                   2342: {
                   2343:        struct roffnode *bl;
1.386     schwarze 2344:        int              irc, rr, spos;
1.375     schwarze 2345:        enum roff_tok    t;
                   2346:
                   2347:        rr = 0;  /* If arguments follow "\}", skip them. */
                   2348:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.386     schwarze 2349:        spos = pos;
1.368     schwarze 2350:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   2351:
1.318     schwarze 2352:        /*
1.386     schwarze 2353:         * Handle requests and macros if the conditional evaluated
                   2354:         * to true or if they are structurally required.
                   2355:         * The .break request is always handled specially.
1.318     schwarze 2356:         */
                   2357:
1.364     schwarze 2358:        if (t == ROFF_break) {
                   2359:                if (irc & ROFF_LOOPMASK)
                   2360:                        irc = ROFF_IGN | ROFF_LOOPEXIT;
                   2361:                else if (rr) {
                   2362:                        for (bl = r->last; bl != NULL; bl = bl->parent) {
                   2363:                                bl->rule = 0;
                   2364:                                if (bl->tok == ROFF_while)
                   2365:                                        break;
                   2366:                        }
                   2367:                }
1.386     schwarze 2368:        } else if (rr || (t < TOKEN_NONE && roffs[t].flags & ROFFMAC_STRUCT)) {
                   2369:                irc |= roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
1.383     schwarze 2370:                if (irc & ROFF_WHILE)
                   2371:                        irc &= ~(ROFF_LOOPCONT | ROFF_LOOPEXIT);
1.386     schwarze 2372:        }
1.340     schwarze 2373:        return irc;
1.82      kristaps 2374: }
                   2375:
1.375     schwarze 2376: /*
                   2377:  * Parse and process a text line in conditional scope.
                   2378:  */
1.340     schwarze 2379: static int
1.82      kristaps 2380: roff_cond_text(ROFF_ARGS)
1.78      kristaps 2381: {
1.375     schwarze 2382:        int      irc, rr;
1.343     schwarze 2383:
1.375     schwarze 2384:        rr = 1;  /* If arguments follow "\}", preserve them. */
                   2385:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.340     schwarze 2386:        if (rr)
                   2387:                irc |= ROFF_CONT;
                   2388:        return irc;
1.74      kristaps 2389: }
                   2390:
1.266     schwarze 2391: /* --- handling of numeric and conditional expressions -------------------- */
                   2392:
1.204     schwarze 2393: /*
                   2394:  * Parse a single signed integer number.  Stop at the first non-digit.
                   2395:  * If there is at least one digit, return success and advance the
                   2396:  * parse point, else return failure and let the parse point unchanged.
                   2397:  * Ignore overflows, treat them just like the C language.
                   2398:  */
1.184     schwarze 2399: static int
1.261     schwarze 2400: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184     schwarze 2401: {
1.261     schwarze 2402:        int      myres, scaled, n, p;
1.206     schwarze 2403:
                   2404:        if (NULL == res)
                   2405:                res = &myres;
1.184     schwarze 2406:
                   2407:        p = *pos;
                   2408:        n = v[p] == '-';
1.261     schwarze 2409:        if (n || v[p] == '+')
1.184     schwarze 2410:                p++;
                   2411:
1.261     schwarze 2412:        if (flags & ROFFNUM_WHITE)
                   2413:                while (isspace((unsigned char)v[p]))
                   2414:                        p++;
                   2415:
1.184     schwarze 2416:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204     schwarze 2417:                *res = 10 * *res + v[p] - '0';
1.184     schwarze 2418:        if (p == *pos + n)
                   2419:                return 0;
                   2420:
                   2421:        if (n)
                   2422:                *res = -*res;
                   2423:
1.254     schwarze 2424:        /* Each number may be followed by one optional scaling unit. */
                   2425:
                   2426:        switch (v[p]) {
                   2427:        case 'f':
1.261     schwarze 2428:                scaled = *res * 65536;
1.254     schwarze 2429:                break;
                   2430:        case 'i':
1.261     schwarze 2431:                scaled = *res * 240;
1.254     schwarze 2432:                break;
                   2433:        case 'c':
1.261     schwarze 2434:                scaled = *res * 240 / 2.54;
1.254     schwarze 2435:                break;
                   2436:        case 'v':
                   2437:        case 'P':
1.261     schwarze 2438:                scaled = *res * 40;
1.254     schwarze 2439:                break;
                   2440:        case 'm':
                   2441:        case 'n':
1.261     schwarze 2442:                scaled = *res * 24;
1.254     schwarze 2443:                break;
                   2444:        case 'p':
1.261     schwarze 2445:                scaled = *res * 10 / 3;
1.254     schwarze 2446:                break;
                   2447:        case 'u':
1.261     schwarze 2448:                scaled = *res;
1.254     schwarze 2449:                break;
                   2450:        case 'M':
1.261     schwarze 2451:                scaled = *res * 6 / 25;
1.254     schwarze 2452:                break;
                   2453:        default:
1.261     schwarze 2454:                scaled = *res;
1.254     schwarze 2455:                p--;
                   2456:                break;
                   2457:        }
1.261     schwarze 2458:        if (flags & ROFFNUM_SCALE)
                   2459:                *res = scaled;
1.254     schwarze 2460:
                   2461:        *pos = p + 1;
1.277     schwarze 2462:        return 1;
1.184     schwarze 2463: }
                   2464:
1.198     schwarze 2465: /*
                   2466:  * Evaluate a string comparison condition.
                   2467:  * The first character is the delimiter.
                   2468:  * Succeed if the string up to its second occurrence
                   2469:  * matches the string up to its third occurence.
                   2470:  * Advance the cursor after the third occurrence
                   2471:  * or lacking that, to the end of the line.
                   2472:  */
                   2473: static int
                   2474: roff_evalstrcond(const char *v, int *pos)
                   2475: {
                   2476:        const char      *s1, *s2, *s3;
                   2477:        int              match;
                   2478:
                   2479:        match = 0;
                   2480:        s1 = v + *pos;          /* initial delimiter */
                   2481:        s2 = s1 + 1;            /* for scanning the first string */
                   2482:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2483:
                   2484:        if (NULL == s3)         /* found no middle delimiter */
                   2485:                goto out;
                   2486:
                   2487:        while ('\0' != *++s3) {
                   2488:                if (*s2 != *s3) {  /* mismatch */
                   2489:                        s3 = strchr(s3, *s1);
                   2490:                        break;
                   2491:                }
                   2492:                if (*s3 == *s1) {  /* found the final delimiter */
                   2493:                        match = 1;
                   2494:                        break;
                   2495:                }
                   2496:                s2++;
                   2497:        }
                   2498:
                   2499: out:
                   2500:        if (NULL == s3)
                   2501:                s3 = strchr(s2, '\0');
1.242     schwarze 2502:        else if (*s3 != '\0')
1.198     schwarze 2503:                s3++;
                   2504:        *pos = s3 - v;
1.277     schwarze 2505:        return match;
1.198     schwarze 2506: }
                   2507:
1.204     schwarze 2508: /*
                   2509:  * Evaluate an optionally negated single character, numerical,
                   2510:  * or string condition.
                   2511:  */
1.198     schwarze 2512: static int
1.271     schwarze 2513: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88      kristaps 2514: {
1.336     schwarze 2515:        const char      *start, *end;
                   2516:        char            *cp, *name;
                   2517:        size_t           sz;
                   2518:        int              deftype, len, number, savepos, istrue, wanttrue;
1.88      kristaps 2519:
1.198     schwarze 2520:        if ('!' == v[*pos]) {
                   2521:                wanttrue = 0;
                   2522:                (*pos)++;
                   2523:        } else
                   2524:                wanttrue = 1;
                   2525:
1.88      kristaps 2526:        switch (v[*pos]) {
1.240     schwarze 2527:        case '\0':
1.277     schwarze 2528:                return 0;
1.207     schwarze 2529:        case 'n':
                   2530:        case 'o':
1.88      kristaps 2531:                (*pos)++;
1.277     schwarze 2532:                return wanttrue;
1.207     schwarze 2533:        case 'e':
                   2534:        case 't':
1.239     schwarze 2535:        case 'v':
1.88      kristaps 2536:                (*pos)++;
1.277     schwarze 2537:                return !wanttrue;
1.336     schwarze 2538:        case 'c':
                   2539:                do {
                   2540:                        (*pos)++;
                   2541:                } while (v[*pos] == ' ');
                   2542:
                   2543:                /*
                   2544:                 * Quirk for groff compatibility:
                   2545:                 * The horizontal tab is neither available nor unavailable.
                   2546:                 */
                   2547:
                   2548:                if (v[*pos] == '\t') {
                   2549:                        (*pos)++;
                   2550:                        return 0;
                   2551:                }
                   2552:
                   2553:                /* Printable ASCII characters are available. */
                   2554:
                   2555:                if (v[*pos] != '\\') {
                   2556:                        (*pos)++;
                   2557:                        return wanttrue;
                   2558:                }
                   2559:
                   2560:                end = v + ++*pos;
                   2561:                switch (mandoc_escape(&end, &start, &len)) {
                   2562:                case ESCAPE_SPECIAL:
                   2563:                        istrue = mchars_spec2cp(start, len) != -1;
                   2564:                        break;
                   2565:                case ESCAPE_UNICODE:
                   2566:                        istrue = 1;
                   2567:                        break;
                   2568:                case ESCAPE_NUMBERED:
                   2569:                        istrue = mchars_num2char(start, len) != -1;
                   2570:                        break;
                   2571:                default:
                   2572:                        istrue = !wanttrue;
                   2573:                        break;
                   2574:                }
                   2575:                *pos = end - v;
                   2576:                return istrue == wanttrue;
1.310     schwarze 2577:        case 'd':
1.271     schwarze 2578:        case 'r':
1.310     schwarze 2579:                cp = v + *pos + 1;
                   2580:                while (*cp == ' ')
                   2581:                        cp++;
                   2582:                name = cp;
                   2583:                sz = roff_getname(r, &cp, ln, cp - v);
1.315     schwarze 2584:                if (sz == 0)
                   2585:                        istrue = 0;
                   2586:                else if (v[*pos] == 'r')
                   2587:                        istrue = roff_hasregn(r, name, sz);
                   2588:                else {
                   2589:                        deftype = ROFFDEF_ANY;
                   2590:                        roff_getstrn(r, name, sz, &deftype);
                   2591:                        istrue = !!deftype;
                   2592:                }
1.363     schwarze 2593:                *pos = (name + sz) - v;
1.310     schwarze 2594:                return istrue == wanttrue;
1.88      kristaps 2595:        default:
                   2596:                break;
                   2597:        }
                   2598:
1.241     schwarze 2599:        savepos = *pos;
1.261     schwarze 2600:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277     schwarze 2601:                return (number > 0) == wanttrue;
1.241     schwarze 2602:        else if (*pos == savepos)
1.277     schwarze 2603:                return roff_evalstrcond(v, pos) == wanttrue;
1.204     schwarze 2604:        else
1.277     schwarze 2605:                return 0;
1.88      kristaps 2606: }
                   2607:
1.340     schwarze 2608: static int
1.103     kristaps 2609: roff_line_ignore(ROFF_ARGS)
1.89      kristaps 2610: {
1.123     schwarze 2611:
1.277     schwarze 2612:        return ROFF_IGN;
1.89      kristaps 2613: }
                   2614:
1.340     schwarze 2615: static int
1.251     schwarze 2616: roff_insec(ROFF_ARGS)
                   2617: {
                   2618:
1.350     schwarze 2619:        mandoc_msg(MANDOCERR_REQ_INSEC, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2620:        return ROFF_IGN;
1.251     schwarze 2621: }
                   2622:
1.340     schwarze 2623: static int
1.251     schwarze 2624: roff_unsupp(ROFF_ARGS)
                   2625: {
                   2626:
1.350     schwarze 2627:        mandoc_msg(MANDOCERR_REQ_UNSUPP, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2628:        return ROFF_IGN;
1.251     schwarze 2629: }
                   2630:
1.340     schwarze 2631: static int
1.82      kristaps 2632: roff_cond(ROFF_ARGS)
1.74      kristaps 2633: {
1.340     schwarze 2634:        int      irc;
1.173     schwarze 2635:
                   2636:        roffnode_push(r, tok, NULL, ln, ppos);
1.74      kristaps 2637:
1.207     schwarze 2638:        /*
1.134     kristaps 2639:         * An `.el' has no conditional body: it will consume the value
                   2640:         * of the current rstack entry set in prior `ie' calls or
1.207     schwarze 2641:         * defaults to DENY.
1.134     kristaps 2642:         *
                   2643:         * If we're not an `el', however, then evaluate the conditional.
                   2644:         */
1.133     kristaps 2645:
1.238     schwarze 2646:        r->last->rule = tok == ROFF_el ?
1.207     schwarze 2647:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238     schwarze 2648:            roff_evalcond(r, ln, buf->buf, &pos);
1.77      kristaps 2649:
1.134     kristaps 2650:        /*
                   2651:         * An if-else will put the NEGATION of the current evaluated
                   2652:         * conditional into the stack of rules.
                   2653:         */
                   2654:
1.238     schwarze 2655:        if (tok == ROFF_ie) {
1.223     schwarze 2656:                if (r->rstackpos + 1 == r->rstacksz) {
                   2657:                        r->rstacksz += 16;
                   2658:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2659:                            r->rstacksz, sizeof(int));
1.134     kristaps 2660:                }
1.198     schwarze 2661:                r->rstack[++r->rstackpos] = !r->last->rule;
1.82      kristaps 2662:        }
1.88      kristaps 2663:
                   2664:        /* If the parent has false as its rule, then so do we. */
                   2665:
1.198     schwarze 2666:        if (r->last->parent && !r->last->parent->rule)
                   2667:                r->last->rule = 0;
1.88      kristaps 2668:
                   2669:        /*
1.173     schwarze 2670:         * Determine scope.
                   2671:         * If there is nothing on the line after the conditional,
                   2672:         * not even whitespace, use next-line scope.
1.340     schwarze 2673:         * Except that .while does not support next-line scope.
1.88      kristaps 2674:         */
1.74      kristaps 2675:
1.340     schwarze 2676:        if (buf->buf[pos] == '\0' && tok != ROFF_while) {
1.173     schwarze 2677:                r->last->endspan = 2;
                   2678:                goto out;
                   2679:        }
                   2680:
1.238     schwarze 2681:        while (buf->buf[pos] == ' ')
1.173     schwarze 2682:                pos++;
                   2683:
                   2684:        /* An opening brace requests multiline scope. */
1.75      kristaps 2685:
1.238     schwarze 2686:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75      kristaps 2687:                r->last->endspan = -1;
                   2688:                pos += 2;
1.272     schwarze 2689:                while (buf->buf[pos] == ' ')
                   2690:                        pos++;
1.173     schwarze 2691:                goto out;
1.207     schwarze 2692:        }
1.74      kristaps 2693:
1.77      kristaps 2694:        /*
1.173     schwarze 2695:         * Anything else following the conditional causes
                   2696:         * single-line scope.  Warn if the scope contains
                   2697:         * nothing but trailing whitespace.
1.77      kristaps 2698:         */
                   2699:
1.238     schwarze 2700:        if (buf->buf[pos] == '\0')
1.350     schwarze 2701:                mandoc_msg(MANDOCERR_COND_EMPTY,
                   2702:                    ln, ppos, "%s", roff_name[tok]);
1.77      kristaps 2703:
1.173     schwarze 2704:        r->last->endspan = 1;
1.74      kristaps 2705:
1.173     schwarze 2706: out:
1.75      kristaps 2707:        *offs = pos;
1.340     schwarze 2708:        irc = ROFF_RERUN;
                   2709:        if (tok == ROFF_while)
                   2710:                irc |= ROFF_WHILE;
                   2711:        return irc;
1.83      schwarze 2712: }
                   2713:
1.340     schwarze 2714: static int
1.92      schwarze 2715: roff_ds(ROFF_ARGS)
                   2716: {
1.212     schwarze 2717:        char            *string;
                   2718:        const char      *name;
                   2719:        size_t           namesz;
1.96      kristaps 2720:
1.251     schwarze 2721:        /* Ignore groff compatibility mode for now. */
                   2722:
                   2723:        if (tok == ROFF_ds1)
                   2724:                tok = ROFF_ds;
                   2725:        else if (tok == ROFF_as1)
                   2726:                tok = ROFF_as;
                   2727:
1.96      kristaps 2728:        /*
1.212     schwarze 2729:         * The first word is the name of the string.
                   2730:         * If it is empty or terminated by an escape sequence,
                   2731:         * abort the `ds' request without defining anything.
1.96      kristaps 2732:         */
1.92      schwarze 2733:
1.238     schwarze 2734:        name = string = buf->buf + pos;
                   2735:        if (*name == '\0')
1.277     schwarze 2736:                return ROFF_IGN;
1.92      schwarze 2737:
1.212     schwarze 2738:        namesz = roff_getname(r, &string, ln, pos);
1.363     schwarze 2739:        switch (name[namesz]) {
                   2740:        case '\\':
1.277     schwarze 2741:                return ROFF_IGN;
1.363     schwarze 2742:        case '\t':
                   2743:                string = buf->buf + pos + namesz;
                   2744:                break;
                   2745:        default:
                   2746:                break;
                   2747:        }
1.212     schwarze 2748:
                   2749:        /* Read past the initial double-quote, if any. */
1.238     schwarze 2750:        if (*string == '"')
1.92      schwarze 2751:                string++;
                   2752:
1.96      kristaps 2753:        /* The rest is the value. */
1.212     schwarze 2754:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2755:            ROFF_as == tok);
1.311     schwarze 2756:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277     schwarze 2757:        return ROFF_IGN;
1.92      schwarze 2758: }
                   2759:
1.204     schwarze 2760: /*
                   2761:  * Parse a single operator, one or two characters long.
                   2762:  * If the operator is recognized, return success and advance the
                   2763:  * parse point, else return failure and let the parse point unchanged.
                   2764:  */
                   2765: static int
                   2766: roff_getop(const char *v, int *pos, char *res)
                   2767: {
                   2768:
                   2769:        *res = v[*pos];
                   2770:
                   2771:        switch (*res) {
1.207     schwarze 2772:        case '+':
                   2773:        case '-':
                   2774:        case '*':
                   2775:        case '/':
                   2776:        case '%':
                   2777:        case '&':
                   2778:        case ':':
1.204     schwarze 2779:                break;
                   2780:        case '<':
                   2781:                switch (v[*pos + 1]) {
1.207     schwarze 2782:                case '=':
1.204     schwarze 2783:                        *res = 'l';
                   2784:                        (*pos)++;
                   2785:                        break;
1.207     schwarze 2786:                case '>':
1.204     schwarze 2787:                        *res = '!';
                   2788:                        (*pos)++;
                   2789:                        break;
1.207     schwarze 2790:                case '?':
1.204     schwarze 2791:                        *res = 'i';
                   2792:                        (*pos)++;
                   2793:                        break;
                   2794:                default:
                   2795:                        break;
                   2796:                }
                   2797:                break;
                   2798:        case '>':
                   2799:                switch (v[*pos + 1]) {
1.207     schwarze 2800:                case '=':
1.204     schwarze 2801:                        *res = 'g';
                   2802:                        (*pos)++;
                   2803:                        break;
1.207     schwarze 2804:                case '?':
1.204     schwarze 2805:                        *res = 'a';
                   2806:                        (*pos)++;
                   2807:                        break;
                   2808:                default:
                   2809:                        break;
                   2810:                }
                   2811:                break;
                   2812:        case '=':
                   2813:                if ('=' == v[*pos + 1])
                   2814:                        (*pos)++;
                   2815:                break;
                   2816:        default:
1.277     schwarze 2817:                return 0;
1.204     schwarze 2818:        }
                   2819:        (*pos)++;
                   2820:
1.277     schwarze 2821:        return *res;
1.204     schwarze 2822: }
                   2823:
                   2824: /*
                   2825:  * Evaluate either a parenthesized numeric expression
                   2826:  * or a single signed integer number.
                   2827:  */
                   2828: static int
1.235     schwarze 2829: roff_evalpar(struct roff *r, int ln,
1.261     schwarze 2830:        const char *v, int *pos, int *res, int flags)
1.204     schwarze 2831: {
                   2832:
                   2833:        if ('(' != v[*pos])
1.277     schwarze 2834:                return roff_getnum(v, pos, res, flags);
1.204     schwarze 2835:
                   2836:        (*pos)++;
1.261     schwarze 2837:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277     schwarze 2838:                return 0;
1.204     schwarze 2839:
1.206     schwarze 2840:        /*
                   2841:         * Omission of the closing parenthesis
                   2842:         * is an error in validation mode,
                   2843:         * but ignored in evaluation mode.
                   2844:         */
                   2845:
1.204     schwarze 2846:        if (')' == v[*pos])
                   2847:                (*pos)++;
1.206     schwarze 2848:        else if (NULL == res)
1.277     schwarze 2849:                return 0;
1.204     schwarze 2850:
1.277     schwarze 2851:        return 1;
1.204     schwarze 2852: }
                   2853:
                   2854: /*
                   2855:  * Evaluate a complete numeric expression.
                   2856:  * Proceed left to right, there is no concept of precedence.
                   2857:  */
                   2858: static int
1.235     schwarze 2859: roff_evalnum(struct roff *r, int ln, const char *v,
1.261     schwarze 2860:        int *pos, int *res, int flags)
1.204     schwarze 2861: {
                   2862:        int              mypos, operand2;
                   2863:        char             operator;
                   2864:
                   2865:        if (NULL == pos) {
                   2866:                mypos = 0;
                   2867:                pos = &mypos;
                   2868:        }
                   2869:
1.261     schwarze 2870:        if (flags & ROFFNUM_WHITE)
1.204     schwarze 2871:                while (isspace((unsigned char)v[*pos]))
                   2872:                        (*pos)++;
                   2873:
1.261     schwarze 2874:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277     schwarze 2875:                return 0;
1.204     schwarze 2876:
                   2877:        while (1) {
1.261     schwarze 2878:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2879:                        while (isspace((unsigned char)v[*pos]))
                   2880:                                (*pos)++;
                   2881:
                   2882:                if ( ! roff_getop(v, pos, &operator))
                   2883:                        break;
                   2884:
1.261     schwarze 2885:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2886:                        while (isspace((unsigned char)v[*pos]))
                   2887:                                (*pos)++;
                   2888:
1.261     schwarze 2889:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277     schwarze 2890:                        return 0;
1.204     schwarze 2891:
1.261     schwarze 2892:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2893:                        while (isspace((unsigned char)v[*pos]))
                   2894:                                (*pos)++;
1.206     schwarze 2895:
                   2896:                if (NULL == res)
                   2897:                        continue;
1.204     schwarze 2898:
                   2899:                switch (operator) {
1.207     schwarze 2900:                case '+':
1.204     schwarze 2901:                        *res += operand2;
                   2902:                        break;
1.207     schwarze 2903:                case '-':
1.204     schwarze 2904:                        *res -= operand2;
                   2905:                        break;
1.207     schwarze 2906:                case '*':
1.204     schwarze 2907:                        *res *= operand2;
                   2908:                        break;
1.207     schwarze 2909:                case '/':
1.244     schwarze 2910:                        if (operand2 == 0) {
1.235     schwarze 2911:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2912:                                        ln, *pos, "%s", v);
1.234     kristaps 2913:                                *res = 0;
                   2914:                                break;
                   2915:                        }
1.204     schwarze 2916:                        *res /= operand2;
                   2917:                        break;
1.207     schwarze 2918:                case '%':
1.244     schwarze 2919:                        if (operand2 == 0) {
                   2920:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2921:                                        ln, *pos, "%s", v);
1.244     schwarze 2922:                                *res = 0;
                   2923:                                break;
                   2924:                        }
1.204     schwarze 2925:                        *res %= operand2;
                   2926:                        break;
1.207     schwarze 2927:                case '<':
1.204     schwarze 2928:                        *res = *res < operand2;
                   2929:                        break;
1.207     schwarze 2930:                case '>':
1.204     schwarze 2931:                        *res = *res > operand2;
                   2932:                        break;
1.207     schwarze 2933:                case 'l':
1.204     schwarze 2934:                        *res = *res <= operand2;
                   2935:                        break;
1.207     schwarze 2936:                case 'g':
1.204     schwarze 2937:                        *res = *res >= operand2;
                   2938:                        break;
1.207     schwarze 2939:                case '=':
1.204     schwarze 2940:                        *res = *res == operand2;
                   2941:                        break;
1.207     schwarze 2942:                case '!':
1.204     schwarze 2943:                        *res = *res != operand2;
                   2944:                        break;
1.207     schwarze 2945:                case '&':
1.204     schwarze 2946:                        *res = *res && operand2;
                   2947:                        break;
1.207     schwarze 2948:                case ':':
1.204     schwarze 2949:                        *res = *res || operand2;
                   2950:                        break;
1.207     schwarze 2951:                case 'i':
1.204     schwarze 2952:                        if (operand2 < *res)
                   2953:                                *res = operand2;
                   2954:                        break;
1.207     schwarze 2955:                case 'a':
1.204     schwarze 2956:                        if (operand2 > *res)
                   2957:                                *res = operand2;
                   2958:                        break;
                   2959:                default:
                   2960:                        abort();
                   2961:                }
                   2962:        }
1.277     schwarze 2963:        return 1;
1.204     schwarze 2964: }
                   2965:
1.266     schwarze 2966: /* --- register management ------------------------------------------------ */
                   2967:
1.180     schwarze 2968: void
1.187     schwarze 2969: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147     kristaps 2970: {
1.327     schwarze 2971:        roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.326     schwarze 2972: }
                   2973:
                   2974: static void
                   2975: roff_setregn(struct roff *r, const char *name, size_t len,
1.327     schwarze 2976:     int val, char sign, int step)
1.326     schwarze 2977: {
1.180     schwarze 2978:        struct roffreg  *reg;
                   2979:
                   2980:        /* Search for an existing register with the same name. */
                   2981:        reg = r->regtab;
                   2982:
1.326     schwarze 2983:        while (reg != NULL && (reg->key.sz != len ||
                   2984:            strncmp(reg->key.p, name, len) != 0))
1.180     schwarze 2985:                reg = reg->next;
1.147     kristaps 2986:
1.180     schwarze 2987:        if (NULL == reg) {
                   2988:                /* Create a new register. */
                   2989:                reg = mandoc_malloc(sizeof(struct roffreg));
1.326     schwarze 2990:                reg->key.p = mandoc_strndup(name, len);
                   2991:                reg->key.sz = len;
1.187     schwarze 2992:                reg->val = 0;
1.327     schwarze 2993:                reg->step = 0;
1.180     schwarze 2994:                reg->next = r->regtab;
                   2995:                r->regtab = reg;
                   2996:        }
                   2997:
1.187     schwarze 2998:        if ('+' == sign)
                   2999:                reg->val += val;
                   3000:        else if ('-' == sign)
                   3001:                reg->val -= val;
                   3002:        else
                   3003:                reg->val = val;
1.327     schwarze 3004:        if (step != INT_MIN)
                   3005:                reg->step = step;
1.147     kristaps 3006: }
                   3007:
1.192     schwarze 3008: /*
                   3009:  * Handle some predefined read-only number registers.
                   3010:  * For now, return -1 if the requested register is not predefined;
                   3011:  * in case a predefined read-only register having the value -1
                   3012:  * were to turn up, another special value would have to be chosen.
                   3013:  */
                   3014: static int
1.273     schwarze 3015: roff_getregro(const struct roff *r, const char *name)
1.192     schwarze 3016: {
                   3017:
                   3018:        switch (*name) {
1.273     schwarze 3019:        case '$':  /* Number of arguments of the last macro evaluated. */
1.339     schwarze 3020:                return r->mstackpos < 0 ? 0 : r->mstack[r->mstackpos].argc;
1.207     schwarze 3021:        case 'A':  /* ASCII approximation mode is always off. */
1.277     schwarze 3022:                return 0;
1.207     schwarze 3023:        case 'g':  /* Groff compatibility mode is always on. */
1.277     schwarze 3024:                return 1;
1.207     schwarze 3025:        case 'H':  /* Fixed horizontal resolution. */
1.277     schwarze 3026:                return 24;
1.207     schwarze 3027:        case 'j':  /* Always adjust left margin only. */
1.277     schwarze 3028:                return 0;
1.207     schwarze 3029:        case 'T':  /* Some output device is always defined. */
1.277     schwarze 3030:                return 1;
1.207     schwarze 3031:        case 'V':  /* Fixed vertical resolution. */
1.277     schwarze 3032:                return 40;
1.192     schwarze 3033:        default:
1.277     schwarze 3034:                return -1;
1.192     schwarze 3035:        }
                   3036: }
                   3037:
1.181     schwarze 3038: int
1.326     schwarze 3039: roff_getreg(struct roff *r, const char *name)
1.147     kristaps 3040: {
1.327     schwarze 3041:        return roff_getregn(r, name, strlen(name), '\0');
1.181     schwarze 3042: }
                   3043:
                   3044: static int
1.327     schwarze 3045: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.181     schwarze 3046: {
                   3047:        struct roffreg  *reg;
1.192     schwarze 3048:        int              val;
                   3049:
                   3050:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3051:                val = roff_getregro(r, name + 1);
1.192     schwarze 3052:                if (-1 != val)
1.277     schwarze 3053:                        return val;
1.192     schwarze 3054:        }
1.181     schwarze 3055:
1.327     schwarze 3056:        for (reg = r->regtab; reg; reg = reg->next) {
1.181     schwarze 3057:                if (len == reg->key.sz &&
1.327     schwarze 3058:                    0 == strncmp(name, reg->key.p, len)) {
                   3059:                        switch (sign) {
                   3060:                        case '+':
                   3061:                                reg->val += reg->step;
                   3062:                                break;
                   3063:                        case '-':
                   3064:                                reg->val -= reg->step;
                   3065:                                break;
                   3066:                        default:
                   3067:                                break;
                   3068:                        }
1.277     schwarze 3069:                        return reg->val;
1.327     schwarze 3070:                }
                   3071:        }
1.271     schwarze 3072:
1.327     schwarze 3073:        roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.277     schwarze 3074:        return 0;
1.271     schwarze 3075: }
                   3076:
                   3077: static int
                   3078: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   3079: {
                   3080:        struct roffreg  *reg;
                   3081:        int              val;
                   3082:
                   3083:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3084:                val = roff_getregro(r, name + 1);
1.271     schwarze 3085:                if (-1 != val)
1.277     schwarze 3086:                        return 1;
1.271     schwarze 3087:        }
                   3088:
                   3089:        for (reg = r->regtab; reg; reg = reg->next)
                   3090:                if (len == reg->key.sz &&
                   3091:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 3092:                        return 1;
1.147     kristaps 3093:
1.277     schwarze 3094:        return 0;
1.147     kristaps 3095: }
                   3096:
1.180     schwarze 3097: static void
                   3098: roff_freereg(struct roffreg *reg)
1.147     kristaps 3099: {
1.180     schwarze 3100:        struct roffreg  *old_reg;
1.147     kristaps 3101:
1.180     schwarze 3102:        while (NULL != reg) {
                   3103:                free(reg->key.p);
                   3104:                old_reg = reg;
                   3105:                reg = reg->next;
                   3106:                free(old_reg);
                   3107:        }
1.147     kristaps 3108: }
1.92      schwarze 3109:
1.340     schwarze 3110: static int
1.89      kristaps 3111: roff_nr(ROFF_ARGS)
1.83      schwarze 3112: {
1.327     schwarze 3113:        char            *key, *val, *step;
1.212     schwarze 3114:        size_t           keysz;
1.327     schwarze 3115:        int              iv, is, len;
1.187     schwarze 3116:        char             sign;
1.89      kristaps 3117:
1.238     schwarze 3118:        key = val = buf->buf + pos;
                   3119:        if (*key == '\0')
1.277     schwarze 3120:                return ROFF_IGN;
1.212     schwarze 3121:
                   3122:        keysz = roff_getname(r, &val, ln, pos);
1.363     schwarze 3123:        if (key[keysz] == '\\' || key[keysz] == '\t')
1.277     schwarze 3124:                return ROFF_IGN;
1.89      kristaps 3125:
1.187     schwarze 3126:        sign = *val;
1.238     schwarze 3127:        if (sign == '+' || sign == '-')
1.187     schwarze 3128:                val++;
                   3129:
1.327     schwarze 3130:        len = 0;
                   3131:        if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
                   3132:                return ROFF_IGN;
                   3133:
                   3134:        step = val + len;
                   3135:        while (isspace((unsigned char)*step))
                   3136:                step++;
                   3137:        if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
                   3138:                is = INT_MIN;
1.109     kristaps 3139:
1.327     schwarze 3140:        roff_setregn(r, key, keysz, iv, sign, is);
1.277     schwarze 3141:        return ROFF_IGN;
1.203     schwarze 3142: }
                   3143:
1.340     schwarze 3144: static int
1.203     schwarze 3145: roff_rr(ROFF_ARGS)
                   3146: {
                   3147:        struct roffreg  *reg, **prev;
1.212     schwarze 3148:        char            *name, *cp;
                   3149:        size_t           namesz;
1.203     schwarze 3150:
1.238     schwarze 3151:        name = cp = buf->buf + pos;
                   3152:        if (*name == '\0')
1.277     schwarze 3153:                return ROFF_IGN;
1.212     schwarze 3154:        namesz = roff_getname(r, &cp, ln, pos);
                   3155:        name[namesz] = '\0';
1.203     schwarze 3156:
                   3157:        prev = &r->regtab;
                   3158:        while (1) {
                   3159:                reg = *prev;
1.238     schwarze 3160:                if (reg == NULL || !strcmp(name, reg->key.p))
1.203     schwarze 3161:                        break;
                   3162:                prev = &reg->next;
                   3163:        }
1.238     schwarze 3164:        if (reg != NULL) {
1.203     schwarze 3165:                *prev = reg->next;
                   3166:                free(reg->key.p);
                   3167:                free(reg);
                   3168:        }
1.277     schwarze 3169:        return ROFF_IGN;
1.122     schwarze 3170: }
                   3171:
1.266     schwarze 3172: /* --- handler functions for roff requests -------------------------------- */
                   3173:
1.340     schwarze 3174: static int
1.122     schwarze 3175: roff_rm(ROFF_ARGS)
                   3176: {
                   3177:        const char       *name;
                   3178:        char             *cp;
1.212     schwarze 3179:        size_t            namesz;
1.122     schwarze 3180:
1.238     schwarze 3181:        cp = buf->buf + pos;
                   3182:        while (*cp != '\0') {
1.212     schwarze 3183:                name = cp;
1.238     schwarze 3184:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212     schwarze 3185:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311     schwarze 3186:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.363     schwarze 3187:                if (name[namesz] == '\\' || name[namesz] == '\t')
1.212     schwarze 3188:                        break;
1.122     schwarze 3189:        }
1.277     schwarze 3190:        return ROFF_IGN;
1.178     schwarze 3191: }
                   3192:
1.340     schwarze 3193: static int
1.178     schwarze 3194: roff_it(ROFF_ARGS)
                   3195: {
                   3196:        int              iv;
                   3197:
                   3198:        /* Parse the number of lines. */
1.261     schwarze 3199:
                   3200:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.350     schwarze 3201:                mandoc_msg(MANDOCERR_IT_NONUM,
                   3202:                    ln, ppos, "%s", buf->buf + 1);
1.277     schwarze 3203:                return ROFF_IGN;
1.178     schwarze 3204:        }
                   3205:
1.262     schwarze 3206:        while (isspace((unsigned char)buf->buf[pos]))
                   3207:                pos++;
                   3208:
                   3209:        /*
                   3210:         * Arm the input line trap.
                   3211:         * Special-casing "an-trap" is an ugly workaround to cope
                   3212:         * with DocBook stupidly fiddling with man(7) internals.
                   3213:         */
1.261     schwarze 3214:
1.178     schwarze 3215:        roffit_lines = iv;
1.262     schwarze 3216:        roffit_macro = mandoc_strdup(iv != 1 ||
                   3217:            strcmp(buf->buf + pos, "an-trap") ?
                   3218:            buf->buf + pos : "br");
1.277     schwarze 3219:        return ROFF_IGN;
1.175     schwarze 3220: }
                   3221:
1.340     schwarze 3222: static int
1.175     schwarze 3223: roff_Dd(ROFF_ARGS)
                   3224: {
1.315     schwarze 3225:        int              mask;
                   3226:        enum roff_tok    t, te;
1.227     schwarze 3227:
1.315     schwarze 3228:        switch (tok) {
                   3229:        case ROFF_Dd:
                   3230:                tok = MDOC_Dd;
                   3231:                te = MDOC_MAX;
                   3232:                if (r->format == 0)
                   3233:                        r->format = MPARSE_MDOC;
                   3234:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   3235:                break;
                   3236:        case ROFF_TH:
                   3237:                tok = MAN_TH;
                   3238:                te = MAN_MAX;
                   3239:                if (r->format == 0)
                   3240:                        r->format = MPARSE_MAN;
                   3241:                mask = MPARSE_QUICK;
                   3242:                break;
                   3243:        default:
                   3244:                abort();
                   3245:        }
                   3246:        if ((r->options & mask) == 0)
                   3247:                for (t = tok; t < te; t++)
                   3248:                        roff_setstr(r, roff_name[t], NULL, 0);
1.277     schwarze 3249:        return ROFF_CONT;
1.109     kristaps 3250: }
                   3251:
1.340     schwarze 3252: static int
1.109     kristaps 3253: roff_TE(ROFF_ARGS)
                   3254: {
1.361     schwarze 3255:        r->man->flags &= ~ROFF_NONOFILL;
1.321     schwarze 3256:        if (r->tbl == NULL) {
1.350     schwarze 3257:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "TE");
1.321     schwarze 3258:                return ROFF_IGN;
                   3259:        }
1.346     schwarze 3260:        if (tbl_end(r->tbl, 0) == 0) {
1.321     schwarze 3261:                r->tbl = NULL;
1.258     schwarze 3262:                free(buf->buf);
                   3263:                buf->buf = mandoc_strdup(".sp");
                   3264:                buf->sz = 4;
1.329     schwarze 3265:                *offs = 0;
1.277     schwarze 3266:                return ROFF_REPARSE;
1.258     schwarze 3267:        }
1.321     schwarze 3268:        r->tbl = NULL;
1.277     schwarze 3269:        return ROFF_IGN;
1.112     kristaps 3270: }
                   3271:
1.340     schwarze 3272: static int
1.112     kristaps 3273: roff_T_(ROFF_ARGS)
                   3274: {
                   3275:
                   3276:        if (NULL == r->tbl)
1.350     schwarze 3277:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "T&");
1.112     kristaps 3278:        else
1.296     schwarze 3279:                tbl_restart(ln, ppos, r->tbl);
1.112     kristaps 3280:
1.277     schwarze 3281:        return ROFF_IGN;
1.109     kristaps 3282: }
                   3283:
1.230     schwarze 3284: /*
                   3285:  * Handle in-line equation delimiters.
                   3286:  */
1.340     schwarze 3287: static int
1.238     schwarze 3288: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151     kristaps 3289: {
1.233     schwarze 3290:        char            *cp1, *cp2;
                   3291:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151     kristaps 3292:
1.230     schwarze 3293:        /*
                   3294:         * Outside equations, look for an opening delimiter.
                   3295:         * If we are inside an equation, we already know it is
                   3296:         * in-line, or this function wouldn't have been called;
                   3297:         * so look for a closing delimiter.
                   3298:         */
                   3299:
1.238     schwarze 3300:        cp1 = buf->buf + pos;
1.230     schwarze 3301:        cp2 = strchr(cp1, r->eqn == NULL ?
                   3302:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   3303:        if (cp2 == NULL)
1.277     schwarze 3304:                return ROFF_CONT;
1.230     schwarze 3305:
1.233     schwarze 3306:        *cp2++ = '\0';
                   3307:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   3308:
                   3309:        /* Handle preceding text, protecting whitespace. */
                   3310:
1.238     schwarze 3311:        if (*buf->buf != '\0') {
1.233     schwarze 3312:                if (r->eqn == NULL)
                   3313:                        bef_pr = "\\&";
                   3314:                bef_nl = "\n";
                   3315:        }
                   3316:
                   3317:        /*
                   3318:         * Prepare replacing the delimiter with an equation macro
                   3319:         * and drop leading white space from the equation.
                   3320:         */
1.230     schwarze 3321:
1.233     schwarze 3322:        if (r->eqn == NULL) {
                   3323:                while (*cp2 == ' ')
                   3324:                        cp2++;
                   3325:                mac = ".EQ";
                   3326:        } else
                   3327:                mac = ".EN";
                   3328:
                   3329:        /* Handle following text, protecting whitespace. */
                   3330:
                   3331:        if (*cp2 != '\0') {
                   3332:                aft_nl = "\n";
                   3333:                if (r->eqn != NULL)
                   3334:                        aft_pr = "\\&";
                   3335:        }
                   3336:
                   3337:        /* Do the actual replacement. */
                   3338:
1.238     schwarze 3339:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233     schwarze 3340:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238     schwarze 3341:        free(buf->buf);
                   3342:        buf->buf = cp1;
1.230     schwarze 3343:
                   3344:        /* Toggle the in-line state of the eqn subsystem. */
                   3345:
                   3346:        r->eqn_inline = r->eqn == NULL;
1.277     schwarze 3347:        return ROFF_REPARSE;
1.151     kristaps 3348: }
                   3349:
1.340     schwarze 3350: static int
1.235     schwarze 3351: roff_EQ(ROFF_ARGS)
1.125     kristaps 3352: {
1.319     schwarze 3353:        struct roff_node        *n;
                   3354:
1.356     schwarze 3355:        if (r->man->meta.macroset == MACROSET_MAN)
1.322     schwarze 3356:                man_breakscope(r->man, ROFF_EQ);
1.319     schwarze 3357:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   3358:        if (ln > r->man->last->line)
                   3359:                n->flags |= NODE_LINE;
1.348     schwarze 3360:        n->eqn = eqn_box_new();
1.319     schwarze 3361:        roff_node_append(r->man, n);
                   3362:        r->man->next = ROFF_NEXT_SIBLING;
1.125     kristaps 3363:
1.238     schwarze 3364:        assert(r->eqn == NULL);
1.319     schwarze 3365:        if (r->last_eqn == NULL)
1.351     schwarze 3366:                r->last_eqn = eqn_alloc();
1.319     schwarze 3367:        else
                   3368:                eqn_reset(r->last_eqn);
                   3369:        r->eqn = r->last_eqn;
                   3370:        r->eqn->node = n;
1.151     kristaps 3371:
1.238     schwarze 3372:        if (buf->buf[pos] != '\0')
1.350     schwarze 3373:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 3374:                    ".EQ %s", buf->buf + pos);
1.151     kristaps 3375:
1.277     schwarze 3376:        return ROFF_IGN;
1.125     kristaps 3377: }
                   3378:
1.340     schwarze 3379: static int
1.125     kristaps 3380: roff_EN(ROFF_ARGS)
                   3381: {
1.319     schwarze 3382:        if (r->eqn != NULL) {
                   3383:                eqn_parse(r->eqn);
                   3384:                r->eqn = NULL;
                   3385:        } else
1.350     schwarze 3386:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "EN");
1.319     schwarze 3387:        if (buf->buf[pos] != '\0')
1.350     schwarze 3388:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.319     schwarze 3389:                    "EN %s", buf->buf + pos);
1.277     schwarze 3390:        return ROFF_IGN;
1.125     kristaps 3391: }
                   3392:
1.340     schwarze 3393: static int
1.109     kristaps 3394: roff_TS(ROFF_ARGS)
                   3395: {
1.321     schwarze 3396:        if (r->tbl != NULL) {
1.350     schwarze 3397:                mandoc_msg(MANDOCERR_BLK_BROKEN, ln, ppos, "TS breaks TS");
1.346     schwarze 3398:                tbl_end(r->tbl, 0);
1.115     kristaps 3399:        }
1.361     schwarze 3400:        r->man->flags |= ROFF_NONOFILL;
1.351     schwarze 3401:        r->tbl = tbl_alloc(ppos, ln, r->last_tbl);
1.346     schwarze 3402:        if (r->last_tbl == NULL)
1.321     schwarze 3403:                r->first_tbl = r->tbl;
                   3404:        r->last_tbl = r->tbl;
1.297     schwarze 3405:        return ROFF_IGN;
                   3406: }
                   3407:
1.340     schwarze 3408: static int
1.358     schwarze 3409: roff_noarg(ROFF_ARGS)
                   3410: {
                   3411:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3412:                man_breakscope(r->man, tok);
                   3413:        if (tok == ROFF_brp)
                   3414:                tok = ROFF_br;
                   3415:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3416:        if (buf->buf[pos] != '\0')
                   3417:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
                   3418:                   "%s %s", roff_name[tok], buf->buf + pos);
                   3419:        if (tok == ROFF_nf)
                   3420:                r->man->flags |= ROFF_NOFILL;
                   3421:        else if (tok == ROFF_fi)
                   3422:                r->man->flags &= ~ROFF_NOFILL;
                   3423:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3424:        r->man->next = ROFF_NEXT_SIBLING;
                   3425:        return ROFF_IGN;
                   3426: }
                   3427:
                   3428: static int
1.297     schwarze 3429: roff_onearg(ROFF_ARGS)
                   3430: {
                   3431:        struct roff_node        *n;
                   3432:        char                    *cp;
1.305     schwarze 3433:        int                      npos;
1.297     schwarze 3434:
1.302     schwarze 3435:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320     schwarze 3436:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
                   3437:             tok == ROFF_ti))
1.302     schwarze 3438:                man_breakscope(r->man, tok);
                   3439:
1.309     schwarze 3440:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305     schwarze 3441:                r->man->last = roffce_node;
                   3442:                r->man->next = ROFF_NEXT_SIBLING;
                   3443:        }
                   3444:
1.297     schwarze 3445:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3446:        n = r->man->last;
                   3447:
                   3448:        cp = buf->buf + pos;
                   3449:        if (*cp != '\0') {
                   3450:                while (*cp != '\0' && *cp != ' ')
                   3451:                        cp++;
                   3452:                while (*cp == ' ')
                   3453:                        *cp++ = '\0';
                   3454:                if (*cp != '\0')
1.350     schwarze 3455:                        mandoc_msg(MANDOCERR_ARG_EXCESS,
                   3456:                            ln, (int)(cp - buf->buf),
1.297     schwarze 3457:                            "%s ... %s", roff_name[tok], cp);
                   3458:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300     schwarze 3459:        }
                   3460:
1.309     schwarze 3461:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   3462:                if (r->man->last->type == ROFFT_ELEM) {
1.305     schwarze 3463:                        roff_word_alloc(r->man, ln, pos, "1");
                   3464:                        r->man->last->flags |= NODE_NOSRC;
                   3465:                }
                   3466:                npos = 0;
                   3467:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   3468:                    &roffce_lines, 0) == 0) {
1.350     schwarze 3469:                        mandoc_msg(MANDOCERR_CE_NONUM,
                   3470:                            ln, pos, "ce %s", buf->buf + pos);
1.305     schwarze 3471:                        roffce_lines = 1;
                   3472:                }
                   3473:                if (roffce_lines < 1) {
                   3474:                        r->man->last = r->man->last->parent;
                   3475:                        roffce_node = NULL;
                   3476:                        roffce_lines = 0;
                   3477:                } else
                   3478:                        roffce_node = r->man->last->parent;
                   3479:        } else {
                   3480:                n->flags |= NODE_VALID | NODE_ENDED;
                   3481:                r->man->last = n;
                   3482:        }
                   3483:        n->flags |= NODE_LINE;
1.300     schwarze 3484:        r->man->next = ROFF_NEXT_SIBLING;
                   3485:        return ROFF_IGN;
                   3486: }
                   3487:
1.340     schwarze 3488: static int
1.300     schwarze 3489: roff_manyarg(ROFF_ARGS)
                   3490: {
                   3491:        struct roff_node        *n;
                   3492:        char                    *sp, *ep;
                   3493:
                   3494:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3495:        n = r->man->last;
                   3496:
                   3497:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   3498:                while (*ep != '\0' && *ep != ' ')
                   3499:                        ep++;
                   3500:                while (*ep == ' ')
                   3501:                        *ep++ = '\0';
                   3502:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297     schwarze 3503:        }
                   3504:
                   3505:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3506:        r->man->last = n;
                   3507:        r->man->next = ROFF_NEXT_SIBLING;
1.277     schwarze 3508:        return ROFF_IGN;
1.251     schwarze 3509: }
                   3510:
1.340     schwarze 3511: static int
1.311     schwarze 3512: roff_als(ROFF_ARGS)
                   3513: {
                   3514:        char            *oldn, *newn, *end, *value;
                   3515:        size_t           oldsz, newsz, valsz;
                   3516:
                   3517:        newn = oldn = buf->buf + pos;
                   3518:        if (*newn == '\0')
                   3519:                return ROFF_IGN;
                   3520:
                   3521:        newsz = roff_getname(r, &oldn, ln, pos);
1.363     schwarze 3522:        if (newn[newsz] == '\\' || newn[newsz] == '\t' || *oldn == '\0')
1.311     schwarze 3523:                return ROFF_IGN;
                   3524:
                   3525:        end = oldn;
                   3526:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3527:        if (oldsz == 0)
                   3528:                return ROFF_IGN;
                   3529:
1.338     schwarze 3530:        valsz = mandoc_asprintf(&value, ".%.*s \\$@\\\"\n",
1.315     schwarze 3531:            (int)oldsz, oldn);
1.311     schwarze 3532:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3533:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3534:        free(value);
1.364     schwarze 3535:        return ROFF_IGN;
                   3536: }
                   3537:
                   3538: /*
                   3539:  * The .break request only makes sense inside conditionals,
                   3540:  * and that case is already handled in roff_cond_sub().
                   3541:  */
                   3542: static int
                   3543: roff_break(ROFF_ARGS)
                   3544: {
                   3545:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, pos, "break");
1.296     schwarze 3546:        return ROFF_IGN;
1.92      schwarze 3547: }
                   3548:
1.340     schwarze 3549: static int
1.174     kristaps 3550: roff_cc(ROFF_ARGS)
                   3551: {
                   3552:        const char      *p;
                   3553:
1.238     schwarze 3554:        p = buf->buf + pos;
1.174     kristaps 3555:
1.238     schwarze 3556:        if (*p == '\0' || (r->control = *p++) == '.')
1.303     schwarze 3557:                r->control = '\0';
1.174     kristaps 3558:
1.238     schwarze 3559:        if (*p != '\0')
1.350     schwarze 3560:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.260     schwarze 3561:                    ln, p - buf->buf, "cc ... %s", p);
1.174     kristaps 3562:
1.341     schwarze 3563:        return ROFF_IGN;
                   3564: }
                   3565:
                   3566: static int
                   3567: roff_char(ROFF_ARGS)
                   3568: {
                   3569:        const char      *p, *kp, *vp;
                   3570:        size_t           ksz, vsz;
                   3571:        int              font;
                   3572:
                   3573:        /* Parse the character to be replaced. */
                   3574:
                   3575:        kp = buf->buf + pos;
                   3576:        p = kp + 1;
                   3577:        if (*kp == '\0' || (*kp == '\\' &&
                   3578:             mandoc_escape(&p, NULL, NULL) != ESCAPE_SPECIAL) ||
                   3579:            (*p != ' ' && *p != '\0')) {
1.350     schwarze 3580:                mandoc_msg(MANDOCERR_CHAR_ARG, ln, pos, "char %s", kp);
1.341     schwarze 3581:                return ROFF_IGN;
                   3582:        }
                   3583:        ksz = p - kp;
                   3584:        while (*p == ' ')
                   3585:                p++;
                   3586:
                   3587:        /*
                   3588:         * If the replacement string contains a font escape sequence,
                   3589:         * we have to restore the font at the end.
                   3590:         */
                   3591:
                   3592:        vp = p;
                   3593:        vsz = strlen(p);
                   3594:        font = 0;
                   3595:        while (*p != '\0') {
                   3596:                if (*p++ != '\\')
                   3597:                        continue;
                   3598:                switch (mandoc_escape(&p, NULL, NULL)) {
                   3599:                case ESCAPE_FONT:
                   3600:                case ESCAPE_FONTROMAN:
                   3601:                case ESCAPE_FONTITALIC:
                   3602:                case ESCAPE_FONTBOLD:
                   3603:                case ESCAPE_FONTBI:
1.378     schwarze 3604:                case ESCAPE_FONTCR:
                   3605:                case ESCAPE_FONTCB:
                   3606:                case ESCAPE_FONTCI:
1.341     schwarze 3607:                case ESCAPE_FONTPREV:
                   3608:                        font++;
                   3609:                        break;
                   3610:                default:
                   3611:                        break;
                   3612:                }
                   3613:        }
                   3614:        if (font > 1)
1.350     schwarze 3615:                mandoc_msg(MANDOCERR_CHAR_FONT,
                   3616:                    ln, (int)(vp - buf->buf), "%s", vp);
1.341     schwarze 3617:
                   3618:        /*
                   3619:         * Approximate the effect of .char using the .tr tables.
                   3620:         * XXX In groff, .char and .tr interact differently.
                   3621:         */
                   3622:
                   3623:        if (ksz == 1) {
                   3624:                if (r->xtab == NULL)
                   3625:                        r->xtab = mandoc_calloc(128, sizeof(*r->xtab));
                   3626:                assert((unsigned int)*kp < 128);
                   3627:                free(r->xtab[(int)*kp].p);
                   3628:                r->xtab[(int)*kp].sz = mandoc_asprintf(&r->xtab[(int)*kp].p,
                   3629:                    "%s%s", vp, font ? "\fP" : "");
                   3630:        } else {
                   3631:                roff_setstrn(&r->xmbtab, kp, ksz, vp, vsz, 0);
                   3632:                if (font)
                   3633:                        roff_setstrn(&r->xmbtab, kp, ksz, "\\fP", 3, 1);
                   3634:        }
1.277     schwarze 3635:        return ROFF_IGN;
1.174     kristaps 3636: }
                   3637:
1.340     schwarze 3638: static int
1.303     schwarze 3639: roff_ec(ROFF_ARGS)
                   3640: {
                   3641:        const char      *p;
                   3642:
                   3643:        p = buf->buf + pos;
                   3644:        if (*p == '\0')
                   3645:                r->escape = '\\';
                   3646:        else {
                   3647:                r->escape = *p;
                   3648:                if (*++p != '\0')
1.350     schwarze 3649:                        mandoc_msg(MANDOCERR_ARG_EXCESS, ln,
                   3650:                            (int)(p - buf->buf), "ec ... %s", p);
1.303     schwarze 3651:        }
                   3652:        return ROFF_IGN;
                   3653: }
                   3654:
1.340     schwarze 3655: static int
1.303     schwarze 3656: roff_eo(ROFF_ARGS)
                   3657: {
                   3658:        r->escape = '\0';
                   3659:        if (buf->buf[pos] != '\0')
1.350     schwarze 3660:                mandoc_msg(MANDOCERR_ARG_SKIP,
1.303     schwarze 3661:                    ln, pos, "eo %s", buf->buf + pos);
1.384     schwarze 3662:        return ROFF_IGN;
                   3663: }
                   3664:
                   3665: static int
                   3666: roff_mc(ROFF_ARGS)
                   3667: {
                   3668:        struct roff_node        *n;
                   3669:        char                    *cp;
                   3670:
                   3671:        /* Parse the first argument. */
                   3672:
                   3673:        cp = buf->buf + pos;
                   3674:        if (*cp != '\0')
                   3675:                cp++;
                   3676:        if (buf->buf[pos] == '\\') {
                   3677:                switch (mandoc_escape((const char **)&cp, NULL, NULL)) {
                   3678:                case ESCAPE_SPECIAL:
                   3679:                case ESCAPE_UNICODE:
                   3680:                case ESCAPE_NUMBERED:
                   3681:                        break;
                   3682:                default:
                   3683:                        *cp = '\0';
                   3684:                        mandoc_msg(MANDOCERR_MC_ESC, ln, pos,
                   3685:                            "mc %s", buf->buf + pos);
                   3686:                        buf->buf[pos] = '\0';
                   3687:                        break;
                   3688:                }
                   3689:        }
                   3690:
                   3691:        /* Ignore additional arguments. */
                   3692:
                   3693:        while (*cp == ' ')
                   3694:                *cp++ = '\0';
                   3695:        if (*cp != '\0') {
                   3696:                mandoc_msg(MANDOCERR_MC_DIST, ln, (int)(cp - buf->buf),
                   3697:                    "mc ... %s", cp);
                   3698:                *cp = '\0';
                   3699:        }
                   3700:
                   3701:        /* Create the .mc node. */
                   3702:
                   3703:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3704:        n = r->man->last;
                   3705:        if (buf->buf[pos] != '\0')
                   3706:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
                   3707:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3708:        r->man->last = n;
                   3709:        r->man->next = ROFF_NEXT_SIBLING;
1.303     schwarze 3710:        return ROFF_IGN;
1.330     schwarze 3711: }
                   3712:
1.340     schwarze 3713: static int
1.330     schwarze 3714: roff_nop(ROFF_ARGS)
                   3715: {
                   3716:        while (buf->buf[pos] == ' ')
                   3717:                pos++;
                   3718:        *offs = pos;
                   3719:        return ROFF_RERUN;
1.303     schwarze 3720: }
                   3721:
1.340     schwarze 3722: static int
1.164     kristaps 3723: roff_tr(ROFF_ARGS)
                   3724: {
                   3725:        const char      *p, *first, *second;
                   3726:        size_t           fsz, ssz;
                   3727:        enum mandoc_esc  esc;
                   3728:
1.238     schwarze 3729:        p = buf->buf + pos;
1.164     kristaps 3730:
1.238     schwarze 3731:        if (*p == '\0') {
1.350     schwarze 3732:                mandoc_msg(MANDOCERR_REQ_EMPTY, ln, ppos, "tr");
1.277     schwarze 3733:                return ROFF_IGN;
1.164     kristaps 3734:        }
                   3735:
1.238     schwarze 3736:        while (*p != '\0') {
1.164     kristaps 3737:                fsz = ssz = 1;
                   3738:
                   3739:                first = p++;
1.238     schwarze 3740:                if (*first == '\\') {
1.164     kristaps 3741:                        esc = mandoc_escape(&p, NULL, NULL);
1.238     schwarze 3742:                        if (esc == ESCAPE_ERROR) {
1.350     schwarze 3743:                                mandoc_msg(MANDOCERR_ESC_BAD, ln,
                   3744:                                    (int)(p - buf->buf), "%s", first);
1.277     schwarze 3745:                                return ROFF_IGN;
1.164     kristaps 3746:                        }
                   3747:                        fsz = (size_t)(p - first);
                   3748:                }
                   3749:
                   3750:                second = p++;
1.238     schwarze 3751:                if (*second == '\\') {
1.164     kristaps 3752:                        esc = mandoc_escape(&p, NULL, NULL);
1.238     schwarze 3753:                        if (esc == ESCAPE_ERROR) {
1.350     schwarze 3754:                                mandoc_msg(MANDOCERR_ESC_BAD, ln,
                   3755:                                    (int)(p - buf->buf), "%s", second);
1.277     schwarze 3756:                                return ROFF_IGN;
1.164     kristaps 3757:                        }
                   3758:                        ssz = (size_t)(p - second);
1.238     schwarze 3759:                } else if (*second == '\0') {
1.350     schwarze 3760:                        mandoc_msg(MANDOCERR_TR_ODD, ln,
                   3761:                            (int)(first - buf->buf), "tr %s", first);
1.164     kristaps 3762:                        second = " ";
1.165     kristaps 3763:                        p--;
1.164     kristaps 3764:                }
                   3765:
1.167     kristaps 3766:                if (fsz > 1) {
1.207     schwarze 3767:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3768:                            second, ssz, 0);
1.167     kristaps 3769:                        continue;
                   3770:                }
                   3771:
1.238     schwarze 3772:                if (r->xtab == NULL)
1.207     schwarze 3773:                        r->xtab = mandoc_calloc(128,
                   3774:                            sizeof(struct roffstr));
1.167     kristaps 3775:
                   3776:                free(r->xtab[(int)*first].p);
                   3777:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3778:                r->xtab[(int)*first].sz = ssz;
1.164     kristaps 3779:        }
                   3780:
1.277     schwarze 3781:        return ROFF_IGN;
1.164     kristaps 3782: }
                   3783:
1.339     schwarze 3784: /*
                   3785:  * Implementation of the .return request.
                   3786:  * There is no need to call roff_userret() from here.
                   3787:  * The read module will call that after rewinding the reader stack
                   3788:  * to the place from where the current macro was called.
                   3789:  */
1.340     schwarze 3790: static int
1.339     schwarze 3791: roff_return(ROFF_ARGS)
                   3792: {
                   3793:        if (r->mstackpos >= 0)
1.340     schwarze 3794:                return ROFF_IGN | ROFF_USERRET;
1.339     schwarze 3795:
1.350     schwarze 3796:        mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "return");
1.339     schwarze 3797:        return ROFF_IGN;
                   3798: }
                   3799:
1.340     schwarze 3800: static int
1.306     schwarze 3801: roff_rn(ROFF_ARGS)
                   3802: {
                   3803:        const char      *value;
                   3804:        char            *oldn, *newn, *end;
                   3805:        size_t           oldsz, newsz;
1.315     schwarze 3806:        int              deftype;
1.306     schwarze 3807:
                   3808:        oldn = newn = buf->buf + pos;
                   3809:        if (*oldn == '\0')
                   3810:                return ROFF_IGN;
                   3811:
                   3812:        oldsz = roff_getname(r, &newn, ln, pos);
1.363     schwarze 3813:        if (oldn[oldsz] == '\\' || oldn[oldsz] == '\t' || *newn == '\0')
1.306     schwarze 3814:                return ROFF_IGN;
                   3815:
                   3816:        end = newn;
                   3817:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3818:        if (newsz == 0)
                   3819:                return ROFF_IGN;
                   3820:
1.315     schwarze 3821:        deftype = ROFFDEF_ANY;
                   3822:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3823:        switch (deftype) {
                   3824:        case ROFFDEF_USER:
1.306     schwarze 3825:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3826:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3827:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315     schwarze 3828:                break;
                   3829:        case ROFFDEF_PRE:
                   3830:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3831:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3832:                break;
                   3833:        case ROFFDEF_REN:
1.306     schwarze 3834:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3835:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315     schwarze 3836:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3837:                break;
                   3838:        case ROFFDEF_STD:
1.306     schwarze 3839:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315     schwarze 3840:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3841:                break;
                   3842:        default:
                   3843:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3844:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3845:                break;
                   3846:        }
1.306     schwarze 3847:        return ROFF_IGN;
                   3848: }
                   3849:
1.340     schwarze 3850: static int
1.339     schwarze 3851: roff_shift(ROFF_ARGS)
                   3852: {
                   3853:        struct mctx     *ctx;
1.382     schwarze 3854:        int              argpos, levels, i;
1.339     schwarze 3855:
1.382     schwarze 3856:        argpos = pos;
1.339     schwarze 3857:        levels = 1;
                   3858:        if (buf->buf[pos] != '\0' &&
                   3859:            roff_evalnum(r, ln, buf->buf, &pos, &levels, 0) == 0) {
1.350     schwarze 3860:                mandoc_msg(MANDOCERR_CE_NONUM,
1.339     schwarze 3861:                    ln, pos, "shift %s", buf->buf + pos);
                   3862:                levels = 1;
                   3863:        }
                   3864:        if (r->mstackpos < 0) {
1.350     schwarze 3865:                mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "shift");
1.339     schwarze 3866:                return ROFF_IGN;
                   3867:        }
                   3868:        ctx = r->mstack + r->mstackpos;
                   3869:        if (levels > ctx->argc) {
1.350     schwarze 3870:                mandoc_msg(MANDOCERR_SHIFT,
1.382     schwarze 3871:                    ln, argpos, "%d, but max is %d", levels, ctx->argc);
1.339     schwarze 3872:                levels = ctx->argc;
1.382     schwarze 3873:        }
                   3874:        if (levels < 0) {
                   3875:                mandoc_msg(MANDOCERR_ARG_NEG, ln, argpos, "shift %d", levels);
                   3876:                levels = 0;
1.339     schwarze 3877:        }
                   3878:        if (levels == 0)
                   3879:                return ROFF_IGN;
                   3880:        for (i = 0; i < levels; i++)
                   3881:                free(ctx->argv[i]);
                   3882:        ctx->argc -= levels;
                   3883:        for (i = 0; i < ctx->argc; i++)
                   3884:                ctx->argv[i] = ctx->argv[i + levels];
                   3885:        return ROFF_IGN;
                   3886: }
                   3887:
1.340     schwarze 3888: static int
1.105     kristaps 3889: roff_so(ROFF_ARGS)
                   3890: {
1.249     schwarze 3891:        char *name, *cp;
1.105     kristaps 3892:
1.238     schwarze 3893:        name = buf->buf + pos;
1.350     schwarze 3894:        mandoc_msg(MANDOCERR_SO, ln, ppos, "so %s", name);
1.105     kristaps 3895:
                   3896:        /*
                   3897:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3898:         * opening anything that's not in our cwd or anything beneath
                   3899:         * it.  Thus, explicitly disallow traversing up the file-system
                   3900:         * or using absolute paths.
                   3901:         */
                   3902:
1.238     schwarze 3903:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.350     schwarze 3904:                mandoc_msg(MANDOCERR_SO_PATH, ln, ppos, ".so %s", name);
1.249     schwarze 3905:                buf->sz = mandoc_asprintf(&cp,
                   3906:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3907:                free(buf->buf);
                   3908:                buf->buf = cp;
                   3909:                *offs = 0;
1.277     schwarze 3910:                return ROFF_REPARSE;
1.105     kristaps 3911:        }
                   3912:
                   3913:        *offs = pos;
1.277     schwarze 3914:        return ROFF_SO;
1.105     kristaps 3915: }
1.92      schwarze 3916:
1.266     schwarze 3917: /* --- user defined strings and macros ------------------------------------ */
                   3918:
1.340     schwarze 3919: static int
1.106     kristaps 3920: roff_userdef(ROFF_ARGS)
1.99      kristaps 3921: {
1.339     schwarze 3922:        struct mctx      *ctx;
                   3923:        char             *arg, *ap, *dst, *src;
                   3924:        size_t            sz;
                   3925:
1.365     schwarze 3926:        /* If the macro is empty, ignore it altogether. */
                   3927:
                   3928:        if (*r->current_string == '\0')
                   3929:                return ROFF_IGN;
                   3930:
1.339     schwarze 3931:        /* Initialize a new macro stack context. */
                   3932:
                   3933:        if (++r->mstackpos == r->mstacksz) {
                   3934:                r->mstack = mandoc_recallocarray(r->mstack,
                   3935:                    r->mstacksz, r->mstacksz + 8, sizeof(*r->mstack));
                   3936:                r->mstacksz += 8;
                   3937:        }
                   3938:        ctx = r->mstack + r->mstackpos;
                   3939:        ctx->argc = 0;
                   3940:
                   3941:        /*
                   3942:         * Collect pointers to macro argument strings,
                   3943:         * NUL-terminating them and escaping quotes.
                   3944:         */
                   3945:
                   3946:        src = buf->buf + pos;
                   3947:        while (*src != '\0') {
                   3948:                if (ctx->argc == ctx->argsz) {
                   3949:                        ctx->argsz += 8;
                   3950:                        ctx->argv = mandoc_reallocarray(ctx->argv,
                   3951:                            ctx->argsz, sizeof(*ctx->argv));
                   3952:                }
1.355     schwarze 3953:                arg = roff_getarg(r, &src, ln, &pos);
1.339     schwarze 3954:                sz = 1;  /* For the terminating NUL. */
                   3955:                for (ap = arg; *ap != '\0'; ap++)
                   3956:                        sz += *ap == '"' ? 4 : 1;
                   3957:                ctx->argv[ctx->argc++] = dst = mandoc_malloc(sz);
                   3958:                for (ap = arg; *ap != '\0'; ap++) {
                   3959:                        if (*ap == '"') {
                   3960:                                memcpy(dst, "\\(dq", 4);
                   3961:                                dst += 4;
                   3962:                        } else
                   3963:                                *dst++ = *ap;
1.106     kristaps 3964:                }
1.339     schwarze 3965:                *dst = '\0';
1.355     schwarze 3966:                free(arg);
1.337     schwarze 3967:        }
                   3968:
1.339     schwarze 3969:        /* Replace the macro invocation by the macro definition. */
1.263     schwarze 3970:
1.238     schwarze 3971:        free(buf->buf);
1.339     schwarze 3972:        buf->buf = mandoc_strdup(r->current_string);
                   3973:        buf->sz = strlen(buf->buf) + 1;
1.248     schwarze 3974:        *offs = 0;
1.106     kristaps 3975:
1.365     schwarze 3976:        return buf->buf[buf->sz - 2] == '\n' ?
1.340     schwarze 3977:            ROFF_REPARSE | ROFF_USERCALL : ROFF_IGN | ROFF_APPEND;
1.99      kristaps 3978: }
1.121     schwarze 3979:
1.306     schwarze 3980: /*
                   3981:  * Calling a high-level macro that was renamed with .rn.
                   3982:  * r->current_string has already been set up by roff_parse().
                   3983:  */
1.340     schwarze 3984: static int
1.306     schwarze 3985: roff_renamed(ROFF_ARGS)
                   3986: {
                   3987:        char    *nbuf;
                   3988:
1.315     schwarze 3989:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3990:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306     schwarze 3991:        free(buf->buf);
                   3992:        buf->buf = nbuf;
1.329     schwarze 3993:        *offs = 0;
1.306     schwarze 3994:        return ROFF_CONT;
                   3995: }
                   3996:
1.362     schwarze 3997: /*
                   3998:  * Measure the length in bytes of the roff identifier at *cpp
                   3999:  * and advance the pointer to the next word.
                   4000:  */
1.212     schwarze 4001: static size_t
1.121     schwarze 4002: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   4003: {
                   4004:        char     *name, *cp;
1.212     schwarze 4005:        size_t    namesz;
1.121     schwarze 4006:
                   4007:        name = *cpp;
1.362     schwarze 4008:        if (*name == '\0')
1.277     schwarze 4009:                return 0;
1.121     schwarze 4010:
1.362     schwarze 4011:        /* Advance cp to the byte after the end of the name. */
                   4012:
1.212     schwarze 4013:        for (cp = name; 1; cp++) {
1.362     schwarze 4014:                namesz = cp - name;
1.363     schwarze 4015:                if (*cp == '\0')
1.212     schwarze 4016:                        break;
1.363     schwarze 4017:                if (*cp == ' ' || *cp == '\t') {
                   4018:                        cp++;
                   4019:                        break;
                   4020:                }
1.362     schwarze 4021:                if (*cp != '\\')
1.121     schwarze 4022:                        continue;
1.362     schwarze 4023:                if (cp[1] == '{' || cp[1] == '}')
1.215     schwarze 4024:                        break;
1.362     schwarze 4025:                if (*++cp == '\\')
1.121     schwarze 4026:                        continue;
1.350     schwarze 4027:                mandoc_msg(MANDOCERR_NAMESC, ln, pos,
1.224     schwarze 4028:                    "%.*s", (int)(cp - name + 1), name);
1.212     schwarze 4029:                mandoc_escape((const char **)&cp, NULL, NULL);
                   4030:                break;
1.121     schwarze 4031:        }
                   4032:
                   4033:        /* Read past spaces. */
1.362     schwarze 4034:
                   4035:        while (*cp == ' ')
1.121     schwarze 4036:                cp++;
                   4037:
                   4038:        *cpp = cp;
1.277     schwarze 4039:        return namesz;
1.121     schwarze 4040: }
                   4041:
1.106     kristaps 4042: /*
                   4043:  * Store *string into the user-defined string called *name.
                   4044:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193     schwarze 4045:  * append == 0: replace mode
                   4046:  * append == 1: single-line append mode
                   4047:  * append == 2: multiline append mode, append '\n' after each call
1.106     kristaps 4048:  */
1.94      kristaps 4049: static void
1.106     kristaps 4050: roff_setstr(struct roff *r, const char *name, const char *string,
1.193     schwarze 4051:        int append)
1.92      schwarze 4052: {
1.315     schwarze 4053:        size_t   namesz;
1.164     kristaps 4054:
1.315     schwarze 4055:        namesz = strlen(name);
                   4056:        roff_setstrn(&r->strtab, name, namesz, string,
1.207     schwarze 4057:            string ? strlen(string) : 0, append);
1.315     schwarze 4058:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164     kristaps 4059: }
                   4060:
                   4061: static void
1.166     kristaps 4062: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193     schwarze 4063:                const char *string, size_t stringsz, int append)
1.164     kristaps 4064: {
1.166     kristaps 4065:        struct roffkv   *n;
1.164     kristaps 4066:        char            *c;
                   4067:        int              i;
                   4068:        size_t           oldch, newch;
1.92      schwarze 4069:
1.106     kristaps 4070:        /* Search for an existing string with the same name. */
1.164     kristaps 4071:        n = *r;
                   4072:
1.211     schwarze 4073:        while (n && (namesz != n->key.sz ||
                   4074:                        strncmp(n->key.p, name, namesz)))
1.92      schwarze 4075:                n = n->next;
1.94      kristaps 4076:
                   4077:        if (NULL == n) {
1.106     kristaps 4078:                /* Create a new string table entry. */
1.166     kristaps 4079:                n = mandoc_malloc(sizeof(struct roffkv));
                   4080:                n->key.p = mandoc_strndup(name, namesz);
                   4081:                n->key.sz = namesz;
                   4082:                n->val.p = NULL;
                   4083:                n->val.sz = 0;
1.164     kristaps 4084:                n->next = *r;
                   4085:                *r = n;
1.193     schwarze 4086:        } else if (0 == append) {
1.166     kristaps 4087:                free(n->val.p);
                   4088:                n->val.p = NULL;
                   4089:                n->val.sz = 0;
1.106     kristaps 4090:        }
                   4091:
                   4092:        if (NULL == string)
                   4093:                return;
                   4094:
                   4095:        /*
                   4096:         * One additional byte for the '\n' in multiline mode,
                   4097:         * and one for the terminating '\0'.
                   4098:         */
1.193     schwarze 4099:        newch = stringsz + (1 < append ? 2u : 1u);
1.164     kristaps 4100:
1.166     kristaps 4101:        if (NULL == n->val.p) {
                   4102:                n->val.p = mandoc_malloc(newch);
                   4103:                *n->val.p = '\0';
1.106     kristaps 4104:                oldch = 0;
                   4105:        } else {
1.166     kristaps 4106:                oldch = n->val.sz;
                   4107:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106     kristaps 4108:        }
                   4109:
                   4110:        /* Skip existing content in the destination buffer. */
1.166     kristaps 4111:        c = n->val.p + (int)oldch;
1.106     kristaps 4112:
                   4113:        /* Append new content to the destination buffer. */
1.164     kristaps 4114:        i = 0;
                   4115:        while (i < (int)stringsz) {
1.106     kristaps 4116:                /*
                   4117:                 * Rudimentary roff copy mode:
                   4118:                 * Handle escaped backslashes.
                   4119:                 */
1.164     kristaps 4120:                if ('\\' == string[i] && '\\' == string[i + 1])
                   4121:                        i++;
                   4122:                *c++ = string[i++];
1.106     kristaps 4123:        }
1.94      kristaps 4124:
1.106     kristaps 4125:        /* Append terminating bytes. */
1.193     schwarze 4126:        if (1 < append)
1.106     kristaps 4127:                *c++ = '\n';
1.163     kristaps 4128:
1.106     kristaps 4129:        *c = '\0';
1.166     kristaps 4130:        n->val.sz = (int)(c - n->val.p);
1.92      schwarze 4131: }
                   4132:
1.94      kristaps 4133: static const char *
1.325     schwarze 4134: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315     schwarze 4135:     int *deftype)
1.92      schwarze 4136: {
1.315     schwarze 4137:        const struct roffkv     *n;
1.325     schwarze 4138:        int                      found, i;
1.315     schwarze 4139:        enum roff_tok            tok;
                   4140:
1.325     schwarze 4141:        found = 0;
                   4142:        for (n = r->strtab; n != NULL; n = n->next) {
                   4143:                if (strncmp(name, n->key.p, len) != 0 ||
                   4144:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4145:                        continue;
                   4146:                if (*deftype & ROFFDEF_USER) {
                   4147:                        *deftype = ROFFDEF_USER;
                   4148:                        return n->val.p;
                   4149:                } else {
                   4150:                        found = 1;
                   4151:                        break;
                   4152:                }
                   4153:        }
                   4154:        for (n = r->rentab; n != NULL; n = n->next) {
                   4155:                if (strncmp(name, n->key.p, len) != 0 ||
                   4156:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4157:                        continue;
                   4158:                if (*deftype & ROFFDEF_REN) {
                   4159:                        *deftype = ROFFDEF_REN;
                   4160:                        return n->val.p;
                   4161:                } else {
                   4162:                        found = 1;
                   4163:                        break;
1.315     schwarze 4164:                }
                   4165:        }
1.325     schwarze 4166:        for (i = 0; i < PREDEFS_MAX; i++) {
                   4167:                if (strncmp(name, predefs[i].name, len) != 0 ||
                   4168:                    predefs[i].name[len] != '\0')
                   4169:                        continue;
                   4170:                if (*deftype & ROFFDEF_PRE) {
                   4171:                        *deftype = ROFFDEF_PRE;
                   4172:                        return predefs[i].str;
                   4173:                } else {
                   4174:                        found = 1;
                   4175:                        break;
1.315     schwarze 4176:                }
                   4177:        }
1.356     schwarze 4178:        if (r->man->meta.macroset != MACROSET_MAN) {
1.325     schwarze 4179:                for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
                   4180:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4181:                            roff_name[tok][len] != '\0')
                   4182:                                continue;
                   4183:                        if (*deftype & ROFFDEF_STD) {
                   4184:                                *deftype = ROFFDEF_STD;
                   4185:                                return NULL;
                   4186:                        } else {
                   4187:                                found = 1;
                   4188:                                break;
1.315     schwarze 4189:                        }
                   4190:                }
                   4191:        }
1.356     schwarze 4192:        if (r->man->meta.macroset != MACROSET_MDOC) {
1.325     schwarze 4193:                for (tok = MAN_TH; tok < MAN_MAX; tok++) {
                   4194:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4195:                            roff_name[tok][len] != '\0')
                   4196:                                continue;
                   4197:                        if (*deftype & ROFFDEF_STD) {
                   4198:                                *deftype = ROFFDEF_STD;
                   4199:                                return NULL;
                   4200:                        } else {
                   4201:                                found = 1;
                   4202:                                break;
1.315     schwarze 4203:                        }
                   4204:                }
1.325     schwarze 4205:        }
                   4206:
                   4207:        if (found == 0 && *deftype != ROFFDEF_ANY) {
                   4208:                if (*deftype & ROFFDEF_REN) {
                   4209:                        /*
                   4210:                         * This might still be a request,
                   4211:                         * so do not treat it as undefined yet.
                   4212:                         */
                   4213:                        *deftype = ROFFDEF_UNDEF;
                   4214:                        return NULL;
1.315     schwarze 4215:                }
1.325     schwarze 4216:
                   4217:                /* Using an undefined string defines it to be empty. */
                   4218:
                   4219:                roff_setstrn(&r->strtab, name, len, "", 0, 0);
                   4220:                roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315     schwarze 4221:        }
1.325     schwarze 4222:
1.315     schwarze 4223:        *deftype = 0;
1.277     schwarze 4224:        return NULL;
1.92      schwarze 4225: }
                   4226:
1.94      kristaps 4227: static void
1.167     kristaps 4228: roff_freestr(struct roffkv *r)
1.92      schwarze 4229: {
1.166     kristaps 4230:        struct roffkv    *n, *nn;
1.92      schwarze 4231:
1.167     kristaps 4232:        for (n = r; n; n = nn) {
1.166     kristaps 4233:                free(n->key.p);
                   4234:                free(n->val.p);
1.92      schwarze 4235:                nn = n->next;
                   4236:                free(n);
                   4237:        }
1.114     kristaps 4238: }
1.266     schwarze 4239:
                   4240: /* --- accessors and utility functions ------------------------------------ */
1.164     kristaps 4241:
                   4242: /*
                   4243:  * Duplicate an input string, making the appropriate character
                   4244:  * conversations (as stipulated by `tr') along the way.
                   4245:  * Returns a heap-allocated string with all the replacements made.
                   4246:  */
                   4247: char *
                   4248: roff_strdup(const struct roff *r, const char *p)
                   4249: {
1.166     kristaps 4250:        const struct roffkv *cp;
1.164     kristaps 4251:        char            *res;
                   4252:        const char      *pp;
                   4253:        size_t           ssz, sz;
                   4254:        enum mandoc_esc  esc;
                   4255:
1.167     kristaps 4256:        if (NULL == r->xmbtab && NULL == r->xtab)
1.277     schwarze 4257:                return mandoc_strdup(p);
1.164     kristaps 4258:        else if ('\0' == *p)
1.277     schwarze 4259:                return mandoc_strdup("");
1.164     kristaps 4260:
                   4261:        /*
                   4262:         * Step through each character looking for term matches
                   4263:         * (remember that a `tr' can be invoked with an escape, which is
                   4264:         * a glyph but the escape is multi-character).
                   4265:         * We only do this if the character hash has been initialised
                   4266:         * and the string is >0 length.
                   4267:         */
                   4268:
                   4269:        res = NULL;
                   4270:        ssz = 0;
                   4271:
                   4272:        while ('\0' != *p) {
1.289     schwarze 4273:                assert((unsigned int)*p < 128);
                   4274:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167     kristaps 4275:                        sz = r->xtab[(int)*p].sz;
                   4276:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4277:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   4278:                        ssz += sz;
                   4279:                        p++;
                   4280:                        continue;
                   4281:                } else if ('\\' != *p) {
                   4282:                        res = mandoc_realloc(res, ssz + 2);
                   4283:                        res[ssz++] = *p++;
                   4284:                        continue;
                   4285:                }
                   4286:
1.164     kristaps 4287:                /* Search for term matches. */
1.167     kristaps 4288:                for (cp = r->xmbtab; cp; cp = cp->next)
1.166     kristaps 4289:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164     kristaps 4290:                                break;
                   4291:
                   4292:                if (NULL != cp) {
                   4293:                        /*
                   4294:                         * A match has been found.
                   4295:                         * Append the match to the array and move
                   4296:                         * forward by its keysize.
                   4297:                         */
1.207     schwarze 4298:                        res = mandoc_realloc(res,
                   4299:                            ssz + cp->val.sz + 1);
1.166     kristaps 4300:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   4301:                        ssz += cp->val.sz;
                   4302:                        p += (int)cp->key.sz;
1.164     kristaps 4303:                        continue;
                   4304:                }
                   4305:
1.167     kristaps 4306:                /*
                   4307:                 * Handle escapes carefully: we need to copy
                   4308:                 * over just the escape itself, or else we might
                   4309:                 * do replacements within the escape itself.
                   4310:                 * Make sure to pass along the bogus string.
                   4311:                 */
                   4312:                pp = p++;
                   4313:                esc = mandoc_escape(&p, NULL, NULL);
                   4314:                if (ESCAPE_ERROR == esc) {
                   4315:                        sz = strlen(pp);
1.164     kristaps 4316:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4317:                        memcpy(res + ssz, pp, sz);
1.167     kristaps 4318:                        break;
1.164     kristaps 4319:                }
1.207     schwarze 4320:                /*
                   4321:                 * We bail out on bad escapes.
1.167     kristaps 4322:                 * No need to warn: we already did so when
1.355     schwarze 4323:                 * roff_expand() was called.
1.167     kristaps 4324:                 */
                   4325:                sz = (int)(p - pp);
                   4326:                res = mandoc_realloc(res, ssz + sz + 1);
                   4327:                memcpy(res + ssz, pp, sz);
                   4328:                ssz += sz;
1.164     kristaps 4329:        }
                   4330:
                   4331:        res[(int)ssz] = '\0';
1.277     schwarze 4332:        return res;
1.227     schwarze 4333: }
                   4334:
                   4335: int
                   4336: roff_getformat(const struct roff *r)
                   4337: {
                   4338:
1.277     schwarze 4339:        return r->format;
1.174     kristaps 4340: }
                   4341:
                   4342: /*
1.207     schwarze 4343:  * Find out whether a line is a macro line or not.
1.174     kristaps 4344:  * If it is, adjust the current position and return one; if it isn't,
                   4345:  * return zero and don't change the current position.
                   4346:  * If the control character has been set with `.cc', then let that grain
                   4347:  * precedence.
                   4348:  * This is slighly contrary to groff, where using the non-breaking
                   4349:  * control character when `cc' has been invoked will cause the
                   4350:  * non-breaking macro contents to be printed verbatim.
                   4351:  */
                   4352: int
                   4353: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   4354: {
                   4355:        int             pos;
                   4356:
                   4357:        pos = *ppos;
                   4358:
1.303     schwarze 4359:        if (r->control != '\0' && cp[pos] == r->control)
1.174     kristaps 4360:                pos++;
1.303     schwarze 4361:        else if (r->control != '\0')
1.277     schwarze 4362:                return 0;
1.174     kristaps 4363:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   4364:                pos += 2;
                   4365:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   4366:                pos++;
                   4367:        else
1.277     schwarze 4368:                return 0;
1.174     kristaps 4369:
                   4370:        while (' ' == cp[pos] || '\t' == cp[pos])
                   4371:                pos++;
                   4372:
                   4373:        *ppos = pos;
1.277     schwarze 4374:        return 1;
1.74      kristaps 4375: }

CVSweb