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

Annotation of mandoc/roff.c, Revision 1.395

1.395   ! schwarze    1: /* $Id: roff.c,v 1.394 2022/06/07 09:41:59 schwarze Exp $ */
1.1       kristaps    2: /*
1.381     schwarze    3:  * Copyright (c) 2010-2015, 2017-2022 Ingo Schwarze <schwarze@openbsd.org>
1.267     schwarze    4:  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.1       kristaps    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
1.66      kristaps    7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
1.1       kristaps    9:  *
1.106     kristaps   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66      kristaps   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106     kristaps   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66      kristaps   13:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     14:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     15:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     16:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.372     schwarze   17:  *
                     18:  * Implementation of the roff(7) parser for mandoc(1).
1.1       kristaps   19:  */
1.66      kristaps   20: #include "config.h"
1.225     schwarze   21:
                     22: #include <sys/types.h>
1.30      kristaps   23:
1.67      kristaps   24: #include <assert.h>
1.85      kristaps   25: #include <ctype.h>
1.245     schwarze   26: #include <limits.h>
1.295     schwarze   27: #include <stddef.h>
                     28: #include <stdint.h>
1.178     schwarze   29: #include <stdio.h>
1.1       kristaps   30: #include <stdlib.h>
1.67      kristaps   31: #include <string.h>
1.1       kristaps   32:
1.201     schwarze   33: #include "mandoc_aux.h"
1.295     schwarze   34: #include "mandoc_ohash.h"
1.345     schwarze   35: #include "mandoc.h"
1.265     schwarze   36: #include "roff.h"
1.349     schwarze   37: #include "mandoc_parse.h"
1.226     schwarze   38: #include "libmandoc.h"
1.266     schwarze   39: #include "roff_int.h"
1.346     schwarze   40: #include "tbl_parse.h"
1.347     schwarze   41: #include "eqn_parse.h"
1.82      kristaps   42:
1.170     schwarze   43: /* Maximum number of string expansions per line, to break infinite loops. */
                     44: #define        EXPAND_LIMIT    1000
                     45:
1.315     schwarze   46: /* Types of definitions of macros and strings. */
                     47: #define        ROFFDEF_USER    (1 << 1)  /* User-defined. */
                     48: #define        ROFFDEF_PRE     (1 << 2)  /* Predefined. */
                     49: #define        ROFFDEF_REN     (1 << 3)  /* Renamed standard macro. */
                     50: #define        ROFFDEF_STD     (1 << 4)  /* mdoc(7) or man(7) macro. */
                     51: #define        ROFFDEF_ANY     (ROFFDEF_USER | ROFFDEF_PRE | \
                     52:                         ROFFDEF_REN | ROFFDEF_STD)
1.325     schwarze   53: #define        ROFFDEF_UNDEF   (1 << 5)  /* Completely undefined. */
1.315     schwarze   54:
1.266     schwarze   55: /* --- data types --------------------------------------------------------- */
                     56:
1.147     kristaps   57: /*
1.167     kristaps   58:  * An incredibly-simple string buffer.
                     59:  */
1.94      kristaps   60: struct roffstr {
1.167     kristaps   61:        char            *p; /* nil-terminated buffer */
                     62:        size_t           sz; /* saved strlen(p) */
1.166     kristaps   63: };
                     64:
                     65: /*
1.167     kristaps   66:  * A key-value roffstr pair as part of a singly-linked list.
1.166     kristaps   67:  */
                     68: struct roffkv {
                     69:        struct roffstr   key;
                     70:        struct roffstr   val;
                     71:        struct roffkv   *next; /* next in list */
1.94      kristaps   72: };
                     73:
1.180     schwarze   74: /*
                     75:  * A single number register as part of a singly-linked list.
                     76:  */
                     77: struct roffreg {
                     78:        struct roffstr   key;
1.181     schwarze   79:        int              val;
1.327     schwarze   80:        int              step;
1.180     schwarze   81:        struct roffreg  *next;
                     82: };
                     83:
1.295     schwarze   84: /*
                     85:  * Association of request and macro names with token IDs.
                     86:  */
                     87: struct roffreq {
                     88:        enum roff_tok    tok;
                     89:        char             name[];
                     90: };
                     91:
1.339     schwarze   92: /*
                     93:  * A macro processing context.
                     94:  * More than one is needed when macro calls are nested.
                     95:  */
                     96: struct mctx {
                     97:        char            **argv;
                     98:        int              argc;
                     99:        int              argsz;
                    100: };
                    101:
1.67      kristaps  102: struct roff {
1.296     schwarze  103:        struct roff_man *man; /* mdoc or man parser */
1.67      kristaps  104:        struct roffnode *last; /* leaf of stack */
1.339     schwarze  105:        struct mctx     *mstack; /* stack of macro contexts */
1.223     schwarze  106:        int             *rstack; /* stack of inverted `ie' values */
1.295     schwarze  107:        struct ohash    *reqtab; /* request lookup table */
1.180     schwarze  108:        struct roffreg  *regtab; /* number registers */
1.166     kristaps  109:        struct roffkv   *strtab; /* user-defined strings & macros */
1.306     schwarze  110:        struct roffkv   *rentab; /* renamed strings & macros */
1.167     kristaps  111:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                    112:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.106     kristaps  113:        const char      *current_string; /* value of last called user macro */
1.118     kristaps  114:        struct tbl_node *first_tbl; /* first table parsed */
                    115:        struct tbl_node *last_tbl; /* last table parsed */
                    116:        struct tbl_node *tbl; /* current table being parsed */
1.319     schwarze  117:        struct eqn_node *last_eqn; /* equation parser */
                    118:        struct eqn_node *eqn; /* active equation parser */
1.230     schwarze  119:        int              eqn_inline; /* current equation is inline */
1.223     schwarze  120:        int              options; /* parse options */
1.339     schwarze  121:        int              mstacksz; /* current size of mstack */
                    122:        int              mstackpos; /* position in mstack */
1.223     schwarze  123:        int              rstacksz; /* current size limit of rstack */
                    124:        int              rstackpos; /* position in rstack */
1.227     schwarze  125:        int              format; /* current file in mdoc or man format */
1.223     schwarze  126:        char             control; /* control character */
1.303     schwarze  127:        char             escape; /* escape character */
1.79      kristaps  128: };
                    129:
1.364     schwarze  130: /*
                    131:  * A macro definition, condition, or ignored block.
                    132:  */
1.67      kristaps  133: struct roffnode {
1.294     schwarze  134:        enum roff_tok    tok; /* type of node */
1.67      kristaps  135:        struct roffnode *parent; /* up one in stack */
                    136:        int              line; /* parse line */
                    137:        int              col; /* parse col */
1.106     kristaps  138:        char            *name; /* node name, e.g. macro name */
1.364     schwarze  139:        char            *end; /* custom end macro of the block */
                    140:        int              endspan; /* scope to: 1=eol 2=next line -1=\} */
                    141:        int              rule; /* content is: 1=evaluated 0=skipped */
1.67      kristaps  142: };
                    143:
                    144: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.294     schwarze  145:                         enum roff_tok tok, /* tok of macro */ \
1.238     schwarze  146:                         struct buf *buf, /* input buffer */ \
1.67      kristaps  147:                         int ln, /* parse line */ \
1.75      kristaps  148:                         int ppos, /* original pos in buffer */ \
                    149:                         int pos, /* current pos in buffer */ \
1.74      kristaps  150:                         int *offs /* reset offset of buffer data */
1.67      kristaps  151:
1.340     schwarze  152: typedef        int (*roffproc)(ROFF_ARGS);
1.67      kristaps  153:
                    154: struct roffmac {
1.79      kristaps  155:        roffproc         proc; /* process new macro */
                    156:        roffproc         text; /* process as child text of macro */
                    157:        roffproc         sub; /* process as child of macro */
                    158:        int              flags;
                    159: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.67      kristaps  160: };
                    161:
1.141     kristaps  162: struct predef {
                    163:        const char      *name; /* predefined input name */
                    164:        const char      *str; /* replacement symbol */
                    165: };
                    166:
                    167: #define        PREDEF(__name, __str) \
                    168:        { (__name), (__str) },
                    169:
1.266     schwarze  170: /* --- function prototypes ------------------------------------------------ */
                    171:
1.340     schwarze  172: static int              roffnode_cleanscope(struct roff *);
                    173: static int              roffnode_pop(struct roff *);
1.294     schwarze  174: static void             roffnode_push(struct roff *, enum roff_tok,
1.155     kristaps  175:                                const char *, int, int);
1.346     schwarze  176: static void             roff_addtbl(struct roff_man *, int, struct tbl_node *);
1.340     schwarze  177: static int              roff_als(ROFF_ARGS);
                    178: static int              roff_block(ROFF_ARGS);
                    179: static int              roff_block_text(ROFF_ARGS);
                    180: static int              roff_block_sub(ROFF_ARGS);
1.364     schwarze  181: static int              roff_break(ROFF_ARGS);
1.340     schwarze  182: static int              roff_cblock(ROFF_ARGS);
                    183: static int              roff_cc(ROFF_ARGS);
                    184: static int              roff_ccond(struct roff *, int, int);
1.341     schwarze  185: static int              roff_char(ROFF_ARGS);
1.340     schwarze  186: static int              roff_cond(ROFF_ARGS);
1.375     schwarze  187: static int              roff_cond_checkend(ROFF_ARGS);
1.340     schwarze  188: static int              roff_cond_text(ROFF_ARGS);
                    189: static int              roff_cond_sub(ROFF_ARGS);
                    190: static int              roff_ds(ROFF_ARGS);
                    191: static int              roff_ec(ROFF_ARGS);
                    192: static int              roff_eo(ROFF_ARGS);
                    193: static int              roff_eqndelim(struct roff *, struct buf *, int);
1.372     schwarze  194: static int              roff_evalcond(struct roff *, int, char *, int *);
1.234     kristaps  195: static int              roff_evalnum(struct roff *, int,
                    196:                                const char *, int *, int *, int);
                    197: static int              roff_evalpar(struct roff *, int,
1.261     schwarze  198:                                const char *, int *, int *, int);
1.198     schwarze  199: static int              roff_evalstrcond(const char *, int *);
1.355     schwarze  200: static int              roff_expand(struct roff *, struct buf *,
                    201:                                int, int, char);
1.388     schwarze  202: static void             roff_expand_patch(struct buf *, int,
                    203:                                const char *, int);
1.155     kristaps  204: static void             roff_free1(struct roff *);
1.180     schwarze  205: static void             roff_freereg(struct roffreg *);
1.167     kristaps  206: static void             roff_freestr(struct roffkv *);
1.212     schwarze  207: static size_t           roff_getname(struct roff *, char **, int, int);
1.261     schwarze  208: static int              roff_getnum(const char *, int *, int *, int);
1.184     schwarze  209: static int              roff_getop(const char *, int *, char *);
1.327     schwarze  210: static int              roff_getregn(struct roff *,
                    211:                                const char *, size_t, char);
1.273     schwarze  212: static int              roff_getregro(const struct roff *,
                    213:                                const char *name);
1.325     schwarze  214: static const char      *roff_getstrn(struct roff *,
1.315     schwarze  215:                                const char *, size_t, int *);
1.271     schwarze  216: static int              roff_hasregn(const struct roff *,
                    217:                                const char *, size_t);
1.340     schwarze  218: static int              roff_insec(ROFF_ARGS);
                    219: static int              roff_it(ROFF_ARGS);
                    220: static int              roff_line_ignore(ROFF_ARGS);
1.265     schwarze  221: static void             roff_man_alloc1(struct roff_man *);
                    222: static void             roff_man_free1(struct roff_man *);
1.340     schwarze  223: static int              roff_manyarg(ROFF_ARGS);
1.384     schwarze  224: static int              roff_mc(ROFF_ARGS);
1.358     schwarze  225: static int              roff_noarg(ROFF_ARGS);
1.340     schwarze  226: static int              roff_nop(ROFF_ARGS);
                    227: static int              roff_nr(ROFF_ARGS);
                    228: static int              roff_onearg(ROFF_ARGS);
1.294     schwarze  229: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.214     schwarze  230:                                int, int);
1.387     schwarze  231: static int              roff_parse_comment(struct roff *, struct buf *,
                    232:                                int, int, char);
1.340     schwarze  233: static int              roff_parsetext(struct roff *, struct buf *,
1.305     schwarze  234:                                int, int *);
1.340     schwarze  235: static int              roff_renamed(ROFF_ARGS);
1.386     schwarze  236: static int              roff_req_or_macro(ROFF_ARGS);
1.340     schwarze  237: static int              roff_return(ROFF_ARGS);
                    238: static int              roff_rm(ROFF_ARGS);
                    239: static int              roff_rn(ROFF_ARGS);
                    240: static int              roff_rr(ROFF_ARGS);
1.326     schwarze  241: static void             roff_setregn(struct roff *, const char *,
1.327     schwarze  242:                                size_t, int, char, int);
1.94      kristaps  243: static void             roff_setstr(struct roff *,
1.106     kristaps  244:                                const char *, const char *, int);
1.207     schwarze  245: static void             roff_setstrn(struct roffkv **, const char *,
1.164     kristaps  246:                                size_t, const char *, size_t, int);
1.340     schwarze  247: static int              roff_shift(ROFF_ARGS);
                    248: static int              roff_so(ROFF_ARGS);
                    249: static int              roff_tr(ROFF_ARGS);
                    250: static int              roff_Dd(ROFF_ARGS);
                    251: static int              roff_TE(ROFF_ARGS);
                    252: static int              roff_TS(ROFF_ARGS);
                    253: static int              roff_EQ(ROFF_ARGS);
                    254: static int              roff_EN(ROFF_ARGS);
                    255: static int              roff_T_(ROFF_ARGS);
                    256: static int              roff_unsupp(ROFF_ARGS);
                    257: static int              roff_userdef(ROFF_ARGS);
1.67      kristaps  258:
1.266     schwarze  259: /* --- constant data ------------------------------------------------------ */
                    260:
1.261     schwarze  261: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    262: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    263:
1.294     schwarze  264: const char *__roff_name[MAN_MAX + 1] = {
1.358     schwarze  265:        "br",           "ce",           "fi",           "ft",
                    266:        "ll",           "mc",           "nf",
                    267:        "po",           "rj",           "sp",
1.312     schwarze  268:        "ta",           "ti",           NULL,
1.294     schwarze  269:        "ab",           "ad",           "af",           "aln",
                    270:        "als",          "am",           "am1",          "ami",
                    271:        "ami1",         "as",           "as1",          "asciify",
                    272:        "backtrace",    "bd",           "bleedat",      "blm",
                    273:         "box",         "boxa",         "bp",           "BP",
                    274:        "break",        "breakchar",    "brnl",         "brp",
1.305     schwarze  275:        "brpnl",        "c2",           "cc",
1.294     schwarze  276:        "cf",           "cflags",       "ch",           "char",
                    277:        "chop",         "class",        "close",        "CL",
                    278:        "color",        "composite",    "continue",     "cp",
                    279:        "cropat",       "cs",           "cu",           "da",
                    280:        "dch",          "Dd",           "de",           "de1",
                    281:        "defcolor",     "dei",          "dei1",         "device",
                    282:        "devicem",      "di",           "do",           "ds",
                    283:        "ds1",          "dwh",          "dt",           "ec",
                    284:        "ecr",          "ecs",          "el",           "em",
                    285:        "EN",           "eo",           "EP",           "EQ",
                    286:        "errprint",     "ev",           "evc",          "ex",
                    287:        "fallback",     "fam",          "fc",           "fchar",
                    288:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    289:        "fl",           "flig",         "fp",           "fps",
                    290:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    291:        "fzoom",        "gcolor",       "hc",           "hcode",
                    292:        "hidechar",     "hla",          "hlm",          "hpf",
                    293:        "hpfa",         "hpfcode",      "hw",           "hy",
                    294:        "hylang",       "hylen",        "hym",          "hypp",
                    295:        "hys",          "ie",           "if",           "ig",
                    296:        "index",        "it",           "itc",          "IX",
                    297:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    298:        "lc",           "lc_ctype",     "lds",          "length",
                    299:        "letadj",       "lf",           "lg",           "lhang",
                    300:        "linetabs",     "lnr",          "lnrf",         "lpfx",
1.304     schwarze  301:        "ls",           "lsm",          "lt",
1.294     schwarze  302:        "mediasize",    "minss",        "mk",           "mso",
                    303:        "na",           "ne",           "nh",           "nhychar",
                    304:        "nm",           "nn",           "nop",          "nr",
                    305:        "nrf",          "nroff",        "ns",           "nx",
                    306:        "open",         "opena",        "os",           "output",
                    307:        "padj",         "papersize",    "pc",           "pev",
                    308:        "pi",           "PI",           "pl",           "pm",
1.312     schwarze  309:        "pn",           "pnr",          "ps",
1.294     schwarze  310:        "psbb",         "pshape",       "pso",          "ptr",
                    311:        "pvs",          "rchar",        "rd",           "recursionlimit",
1.309     schwarze  312:        "return",       "rfschar",      "rhang",
1.294     schwarze  313:        "rm",           "rn",           "rnn",          "rr",
                    314:        "rs",           "rt",           "schar",        "sentchar",
                    315:        "shc",          "shift",        "sizes",        "so",
                    316:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    317:        "sty",          "substring",    "sv",           "sy",
1.300     schwarze  318:        "T&",           "tc",           "TE",
1.301     schwarze  319:        "TH",           "tkf",          "tl",
1.294     schwarze  320:        "tm",           "tm1",          "tmc",          "tr",
                    321:        "track",        "transchar",    "trf",          "trimat",
                    322:        "trin",         "trnt",         "troff",        "TS",
                    323:        "uf",           "ul",           "unformat",     "unwatch",
                    324:        "unwatchn",     "vpt",          "vs",           "warn",
                    325:        "warnscale",    "watch",        "watchlength",  "watchn",
                    326:        "wh",           "while",        "write",        "writec",
                    327:        "writem",       "xflag",        ".",            NULL,
1.306     schwarze  328:        NULL,           "text",
1.294     schwarze  329:        "Dd",           "Dt",           "Os",           "Sh",
                    330:        "Ss",           "Pp",           "D1",           "Dl",
                    331:        "Bd",           "Ed",           "Bl",           "El",
                    332:        "It",           "Ad",           "An",           "Ap",
                    333:        "Ar",           "Cd",           "Cm",           "Dv",
                    334:        "Er",           "Ev",           "Ex",           "Fa",
                    335:        "Fd",           "Fl",           "Fn",           "Ft",
                    336:        "Ic",           "In",           "Li",           "Nd",
                    337:        "Nm",           "Op",           "Ot",           "Pa",
                    338:        "Rv",           "St",           "Va",           "Vt",
                    339:        "Xr",           "%A",           "%B",           "%D",
                    340:        "%I",           "%J",           "%N",           "%O",
                    341:        "%P",           "%R",           "%T",           "%V",
                    342:        "Ac",           "Ao",           "Aq",           "At",
                    343:        "Bc",           "Bf",           "Bo",           "Bq",
                    344:        "Bsx",          "Bx",           "Db",           "Dc",
                    345:        "Do",           "Dq",           "Ec",           "Ef",
                    346:        "Em",           "Eo",           "Fx",           "Ms",
                    347:        "No",           "Ns",           "Nx",           "Ox",
                    348:        "Pc",           "Pf",           "Po",           "Pq",
                    349:        "Qc",           "Ql",           "Qo",           "Qq",
                    350:        "Re",           "Rs",           "Sc",           "So",
                    351:        "Sq",           "Sm",           "Sx",           "Sy",
                    352:        "Tn",           "Ux",           "Xc",           "Xo",
                    353:        "Fo",           "Fc",           "Oo",           "Oc",
                    354:        "Bk",           "Ek",           "Bt",           "Hf",
                    355:        "Fr",           "Ud",           "Lb",           "Lp",
                    356:        "Lk",           "Mt",           "Brq",          "Bro",
                    357:        "Brc",          "%C",           "Es",           "En",
1.299     schwarze  358:        "Dx",           "%Q",           "%U",           "Ta",
1.369     schwarze  359:        "Tg",           NULL,
1.294     schwarze  360:        "TH",           "SH",           "SS",           "TP",
1.332     schwarze  361:        "TQ",
1.294     schwarze  362:        "LP",           "PP",           "P",            "IP",
                    363:        "HP",           "SM",           "SB",           "BI",
                    364:        "IB",           "BR",           "RB",           "R",
                    365:        "B",            "I",            "IR",           "RI",
                    366:        "RE",           "RS",           "DT",           "UC",
1.297     schwarze  367:        "PD",           "AT",           "in",
1.333     schwarze  368:        "SY",           "YS",           "OP",
                    369:        "EX",           "EE",           "UR",
1.317     schwarze  370:        "UE",           "MT",           "ME",           NULL
1.294     schwarze  371: };
                    372: const  char *const *roff_name = __roff_name;
                    373:
                    374: static struct roffmac   roffs[TOKEN_NONE] = {
1.358     schwarze  375:        { roff_noarg, NULL, NULL, 0 },  /* br */
1.305     schwarze  376:        { roff_onearg, NULL, NULL, 0 },  /* ce */
1.358     schwarze  377:        { roff_noarg, NULL, NULL, 0 },  /* fi */
1.297     schwarze  378:        { roff_onearg, NULL, NULL, 0 },  /* ft */
1.298     schwarze  379:        { roff_onearg, NULL, NULL, 0 },  /* ll */
1.384     schwarze  380:        { roff_mc, NULL, NULL, 0 },  /* mc */
1.358     schwarze  381:        { roff_noarg, NULL, NULL, 0 },  /* nf */
1.312     schwarze  382:        { roff_onearg, NULL, NULL, 0 },  /* po */
1.309     schwarze  383:        { roff_onearg, NULL, NULL, 0 },  /* rj */
1.299     schwarze  384:        { roff_onearg, NULL, NULL, 0 },  /* sp */
1.300     schwarze  385:        { roff_manyarg, NULL, NULL, 0 },  /* ta */
1.301     schwarze  386:        { roff_onearg, NULL, NULL, 0 },  /* ti */
1.296     schwarze  387:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.295     schwarze  388:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    389:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    390:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    391:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
1.311     schwarze  392:        { roff_als, NULL, NULL, 0 },  /* als */
1.295     schwarze  393:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    394:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    395:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    396:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    397:        { roff_ds, NULL, NULL, 0 },  /* as */
                    398:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    399:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    400:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    401:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    402:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    403:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    404:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    405:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    406:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    407:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
1.364     schwarze  408:        { roff_break, NULL, NULL, 0 },  /* break */
1.295     schwarze  409:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    410:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.358     schwarze  411:        { roff_noarg, NULL, NULL, 0 },  /* brp */
1.295     schwarze  412:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    413:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    414:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    415:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    416:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    417:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
1.341     schwarze  418:        { roff_char, NULL, NULL, 0 },  /* char */
1.295     schwarze  419:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    421:        { roff_insec, NULL, NULL, 0 },  /* close */
                    422:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    424:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    425:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    426:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    427:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    428:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    429:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    430:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    431:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    432:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    433:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    434:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    435:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    436:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    437:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    438:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    439:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    440:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    441:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    442:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    443:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    444:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    445:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
1.303     schwarze  446:        { roff_ec, NULL, NULL, 0 },  /* ec */
1.295     schwarze  447:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    448:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    449:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    450:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    451:        { roff_EN, NULL, NULL, 0 },  /* EN */
1.303     schwarze  452:        { roff_eo, NULL, NULL, 0 },  /* eo */
1.295     schwarze  453:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    454:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    455:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    456:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    457:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    458:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    459:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    460:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    461:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    462:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    463:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    464:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    465:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    466:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    467:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    468:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    469:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    470:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    471:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    472:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    474:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    475:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    476:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    483:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    484:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    485:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    487:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    488:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    489:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    491:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    492:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    493:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    494:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    495:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    496:        { roff_it, NULL, NULL, 0 },  /* it */
                    497:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    498:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    500:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    501:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    503:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    504:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    505:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    506:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    508:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    511:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    512:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    513:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    514:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    515:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    516:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    517:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    518:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    519:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    520:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    521:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    522:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    523:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    524:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    526:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    527:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
1.330     schwarze  528:        { roff_nop, NULL, NULL, 0 },  /* nop */
1.295     schwarze  529:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    530:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    531:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    532:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    533:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    534:        { roff_insec, NULL, NULL, 0 },  /* open */
                    535:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    536:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    537:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    538:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    539:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    540:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    541:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    542:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    543:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    547:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    548:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    549:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    550:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    551:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    552:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    553:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    554:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    555:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    556:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
1.339     schwarze  557:        { roff_return, NULL, NULL, 0 },  /* return */
1.295     schwarze  558:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    559:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    560:        { roff_rm, NULL, NULL, 0 },  /* rm */
1.306     schwarze  561:        { roff_rn, NULL, NULL, 0 },  /* rn */
1.295     schwarze  562:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    563:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    564:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    565:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    566:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    567:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    568:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
1.339     schwarze  569:        { roff_shift, NULL, NULL, 0 },  /* shift */
1.295     schwarze  570:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    571:        { roff_so, NULL, NULL, 0 },  /* so */
                    572:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    573:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    574:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    575:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    576:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    577:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    578:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    579:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    580:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    581:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    582:        { roff_TE, NULL, NULL, 0 },  /* TE */
1.315     schwarze  583:        { roff_Dd, NULL, NULL, 0 },  /* TH */
1.295     schwarze  584:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    585:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    586:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    587:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    588:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    589:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    590:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    591:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    592:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    593:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    594:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    595:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    596:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    597:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    598:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    599:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    600:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    601:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    602:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    603:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    604:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    605:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    606:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    607:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    608:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    609:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    610:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
1.340     schwarze  611:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /*while*/
1.295     schwarze  612:        { roff_insec, NULL, NULL, 0 },  /* write */
                    613:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    614:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    615:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    616:        { roff_cblock, NULL, NULL, 0 },  /* . */
1.306     schwarze  617:        { roff_renamed, NULL, NULL, 0 },
1.295     schwarze  618:        { roff_userdef, NULL, NULL, 0 }
1.67      kristaps  619: };
                    620:
1.141     kristaps  621: /* Array of injected predefined strings. */
                    622: #define        PREDEFS_MAX      38
                    623: static const struct predef predefs[PREDEFS_MAX] = {
                    624: #include "predefs.in"
                    625: };
                    626:
1.305     schwarze  627: static int      roffce_lines;  /* number of input lines to center */
                    628: static struct roff_node *roffce_node;  /* active request */
1.178     schwarze  629: static int      roffit_lines;  /* number of lines to delay */
                    630: static char    *roffit_macro;  /* nil-terminated macro line */
                    631:
1.207     schwarze  632:
1.266     schwarze  633: /* --- request table ------------------------------------------------------ */
                    634:
1.295     schwarze  635: struct ohash *
                    636: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85      kristaps  637: {
1.295     schwarze  638:        struct ohash    *htab;
                    639:        struct roffreq  *req;
                    640:        enum roff_tok    tok;
                    641:        size_t           sz;
                    642:        unsigned int     slot;
1.85      kristaps  643:
1.295     schwarze  644:        htab = mandoc_malloc(sizeof(*htab));
                    645:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85      kristaps  646:
1.295     schwarze  647:        for (tok = mintok; tok < maxtok; tok++) {
1.296     schwarze  648:                if (roff_name[tok] == NULL)
                    649:                        continue;
1.295     schwarze  650:                sz = strlen(roff_name[tok]);
                    651:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    652:                req->tok = tok;
                    653:                memcpy(req->name, roff_name[tok], sz + 1);
                    654:                slot = ohash_qlookup(htab, req->name);
                    655:                ohash_insert(htab, slot, req);
1.85      kristaps  656:        }
1.295     schwarze  657:        return htab;
1.85      kristaps  658: }
                    659:
1.295     schwarze  660: void
                    661: roffhash_free(struct ohash *htab)
1.67      kristaps  662: {
1.295     schwarze  663:        struct roffreq  *req;
                    664:        unsigned int     slot;
1.67      kristaps  665:
1.295     schwarze  666:        if (htab == NULL)
                    667:                return;
                    668:        for (req = ohash_first(htab, &slot); req != NULL;
                    669:             req = ohash_next(htab, &slot))
                    670:                free(req);
                    671:        ohash_delete(htab);
                    672:        free(htab);
                    673: }
                    674:
                    675: enum roff_tok
                    676: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    677: {
                    678:        struct roffreq  *req;
                    679:        const char      *end;
                    680:
                    681:        if (sz) {
                    682:                end = name + sz;
                    683:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    684:        } else
                    685:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    686:        return req == NULL ? TOKEN_NONE : req->tok;
1.67      kristaps  687: }
                    688:
1.266     schwarze  689: /* --- stack of request blocks -------------------------------------------- */
                    690:
1.67      kristaps  691: /*
                    692:  * Pop the current node off of the stack of roff instructions currently
1.364     schwarze  693:  * pending.  Return 1 if it is a loop or 0 otherwise.
1.67      kristaps  694:  */
1.340     schwarze  695: static int
1.67      kristaps  696: roffnode_pop(struct roff *r)
                    697: {
                    698:        struct roffnode *p;
1.340     schwarze  699:        int              inloop;
1.67      kristaps  700:
1.207     schwarze  701:        p = r->last;
1.340     schwarze  702:        inloop = p->tok == ROFF_while;
                    703:        r->last = p->parent;
1.106     kristaps  704:        free(p->name);
                    705:        free(p->end);
1.67      kristaps  706:        free(p);
1.340     schwarze  707:        return inloop;
1.67      kristaps  708: }
                    709:
                    710: /*
                    711:  * Push a roff node onto the instruction stack.  This must later be
                    712:  * removed with roffnode_pop().
                    713:  */
1.98      schwarze  714: static void
1.294     schwarze  715: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106     kristaps  716:                int line, int col)
1.67      kristaps  717: {
                    718:        struct roffnode *p;
                    719:
1.98      schwarze  720:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.67      kristaps  721:        p->tok = tok;
1.106     kristaps  722:        if (name)
                    723:                p->name = mandoc_strdup(name);
1.67      kristaps  724:        p->parent = r->last;
                    725:        p->line = line;
                    726:        p->col = col;
1.198     schwarze  727:        p->rule = p->parent ? p->parent->rule : 0;
1.67      kristaps  728:
                    729:        r->last = p;
                    730: }
                    731:
1.266     schwarze  732: /* --- roff parser state data management ---------------------------------- */
                    733:
1.67      kristaps  734: static void
                    735: roff_free1(struct roff *r)
                    736: {
1.167     kristaps  737:        int              i;
1.67      kristaps  738:
1.346     schwarze  739:        tbl_free(r->first_tbl);
1.113     kristaps  740:        r->first_tbl = r->last_tbl = r->tbl = NULL;
                    741:
1.347     schwarze  742:        eqn_free(r->last_eqn);
1.319     schwarze  743:        r->last_eqn = r->eqn = NULL;
1.125     kristaps  744:
1.339     schwarze  745:        while (r->mstackpos >= 0)
                    746:                roff_userret(r);
                    747:
1.67      kristaps  748:        while (r->last)
                    749:                roffnode_pop(r);
1.109     kristaps  750:
1.223     schwarze  751:        free (r->rstack);
                    752:        r->rstack = NULL;
                    753:        r->rstacksz = 0;
                    754:        r->rstackpos = -1;
                    755:
                    756:        roff_freereg(r->regtab);
                    757:        r->regtab = NULL;
                    758:
1.167     kristaps  759:        roff_freestr(r->strtab);
1.306     schwarze  760:        roff_freestr(r->rentab);
1.167     kristaps  761:        roff_freestr(r->xmbtab);
1.306     schwarze  762:        r->strtab = r->rentab = r->xmbtab = NULL;
1.167     kristaps  763:
                    764:        if (r->xtab)
                    765:                for (i = 0; i < 128; i++)
                    766:                        free(r->xtab[i].p);
                    767:        free(r->xtab);
                    768:        r->xtab = NULL;
1.67      kristaps  769: }
                    770:
                    771: void
                    772: roff_reset(struct roff *r)
                    773: {
                    774:        roff_free1(r);
1.367     schwarze  775:        r->options |= MPARSE_COMMENT;
1.227     schwarze  776:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303     schwarze  777:        r->control = '\0';
                    778:        r->escape = '\\';
1.307     schwarze  779:        roffce_lines = 0;
                    780:        roffce_node = NULL;
                    781:        roffit_lines = 0;
                    782:        roffit_macro = NULL;
1.67      kristaps  783: }
                    784:
                    785: void
                    786: roff_free(struct roff *r)
                    787: {
1.366     schwarze  788:        int              i;
1.339     schwarze  789:
1.67      kristaps  790:        roff_free1(r);
1.339     schwarze  791:        for (i = 0; i < r->mstacksz; i++)
                    792:                free(r->mstack[i].argv);
                    793:        free(r->mstack);
1.295     schwarze  794:        roffhash_free(r->reqtab);
1.67      kristaps  795:        free(r);
                    796: }
                    797:
                    798: struct roff *
1.351     schwarze  799: roff_alloc(int options)
1.67      kristaps  800: {
                    801:        struct roff     *r;
                    802:
1.98      schwarze  803:        r = mandoc_calloc(1, sizeof(struct roff));
1.328     schwarze  804:        r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.367     schwarze  805:        r->options = options | MPARSE_COMMENT;
1.227     schwarze  806:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.339     schwarze  807:        r->mstackpos = -1;
1.82      kristaps  808:        r->rstackpos = -1;
1.303     schwarze  809:        r->escape = '\\';
1.277     schwarze  810:        return r;
1.265     schwarze  811: }
                    812:
1.266     schwarze  813: /* --- syntax tree state data management ---------------------------------- */
                    814:
1.265     schwarze  815: static void
                    816: roff_man_free1(struct roff_man *man)
                    817: {
1.356     schwarze  818:        if (man->meta.first != NULL)
                    819:                roff_node_delete(man, man->meta.first);
1.265     schwarze  820:        free(man->meta.msec);
                    821:        free(man->meta.vol);
                    822:        free(man->meta.os);
                    823:        free(man->meta.arch);
                    824:        free(man->meta.title);
                    825:        free(man->meta.name);
                    826:        free(man->meta.date);
1.356     schwarze  827:        free(man->meta.sodest);
1.265     schwarze  828: }
                    829:
1.357     schwarze  830: void
                    831: roff_state_reset(struct roff_man *man)
                    832: {
                    833:        man->last = man->meta.first;
                    834:        man->last_es = NULL;
                    835:        man->flags = 0;
                    836:        man->lastsec = man->lastnamed = SEC_NONE;
                    837:        man->next = ROFF_NEXT_CHILD;
                    838:        roff_setreg(man->roff, "nS", 0, '=');
                    839: }
                    840:
1.265     schwarze  841: static void
                    842: roff_man_alloc1(struct roff_man *man)
                    843: {
                    844:        memset(&man->meta, 0, sizeof(man->meta));
1.356     schwarze  845:        man->meta.first = mandoc_calloc(1, sizeof(*man->meta.first));
                    846:        man->meta.first->type = ROFFT_ROOT;
                    847:        man->meta.macroset = MACROSET_NONE;
1.357     schwarze  848:        roff_state_reset(man);
1.265     schwarze  849: }
                    850:
                    851: void
                    852: roff_man_reset(struct roff_man *man)
                    853: {
                    854:        roff_man_free1(man);
                    855:        roff_man_alloc1(man);
                    856: }
                    857:
                    858: void
                    859: roff_man_free(struct roff_man *man)
                    860: {
                    861:        roff_man_free1(man);
1.380     schwarze  862:        free(man->os_r);
1.265     schwarze  863:        free(man);
                    864: }
                    865:
                    866: struct roff_man *
1.351     schwarze  867: roff_man_alloc(struct roff *roff, const char *os_s, int quick)
1.265     schwarze  868: {
                    869:        struct roff_man *man;
                    870:
                    871:        man = mandoc_calloc(1, sizeof(*man));
                    872:        man->roff = roff;
1.316     schwarze  873:        man->os_s = os_s;
1.265     schwarze  874:        man->quick = quick;
                    875:        roff_man_alloc1(man);
1.296     schwarze  876:        roff->man = man;
1.277     schwarze  877:        return man;
1.67      kristaps  878: }
                    879:
1.266     schwarze  880: /* --- syntax tree handling ----------------------------------------------- */
                    881:
                    882: struct roff_node *
                    883: roff_node_alloc(struct roff_man *man, int line, int pos,
                    884:        enum roff_type type, int tok)
                    885: {
                    886:        struct roff_node        *n;
                    887:
                    888:        n = mandoc_calloc(1, sizeof(*n));
                    889:        n->line = line;
                    890:        n->pos = pos;
                    891:        n->tok = tok;
                    892:        n->type = type;
                    893:        n->sec = man->lastsec;
                    894:
                    895:        if (man->flags & MDOC_SYNOPSIS)
1.285     schwarze  896:                n->flags |= NODE_SYNPRETTY;
1.266     schwarze  897:        else
1.285     schwarze  898:                n->flags &= ~NODE_SYNPRETTY;
1.360     schwarze  899:        if ((man->flags & (ROFF_NOFILL | ROFF_NONOFILL)) == ROFF_NOFILL)
1.359     schwarze  900:                n->flags |= NODE_NOFILL;
                    901:        else
                    902:                n->flags &= ~NODE_NOFILL;
1.266     schwarze  903:        if (man->flags & MDOC_NEWLINE)
1.285     schwarze  904:                n->flags |= NODE_LINE;
1.266     schwarze  905:        man->flags &= ~MDOC_NEWLINE;
                    906:
1.277     schwarze  907:        return n;
1.266     schwarze  908: }
                    909:
                    910: void
                    911: roff_node_append(struct roff_man *man, struct roff_node *n)
                    912: {
                    913:
                    914:        switch (man->next) {
                    915:        case ROFF_NEXT_SIBLING:
1.281     schwarze  916:                if (man->last->next != NULL) {
                    917:                        n->next = man->last->next;
                    918:                        man->last->next->prev = n;
                    919:                } else
                    920:                        man->last->parent->last = n;
1.266     schwarze  921:                man->last->next = n;
                    922:                n->prev = man->last;
                    923:                n->parent = man->last->parent;
                    924:                break;
                    925:        case ROFF_NEXT_CHILD:
1.287     schwarze  926:                if (man->last->child != NULL) {
                    927:                        n->next = man->last->child;
                    928:                        man->last->child->prev = n;
                    929:                } else
                    930:                        man->last->last = n;
1.266     schwarze  931:                man->last->child = n;
                    932:                n->parent = man->last;
                    933:                break;
                    934:        default:
                    935:                abort();
                    936:        }
1.282     schwarze  937:        man->last = n;
1.266     schwarze  938:
                    939:        switch (n->type) {
                    940:        case ROFFT_HEAD:
                    941:                n->parent->head = n;
                    942:                break;
                    943:        case ROFFT_BODY:
1.282     schwarze  944:                if (n->end != ENDBODY_NOT)
                    945:                        return;
1.266     schwarze  946:                n->parent->body = n;
                    947:                break;
                    948:        case ROFFT_TAIL:
                    949:                n->parent->tail = n;
                    950:                break;
                    951:        default:
1.282     schwarze  952:                return;
1.266     schwarze  953:        }
1.282     schwarze  954:
                    955:        /*
                    956:         * Copy over the normalised-data pointer of our parent.  Not
                    957:         * everybody has one, but copying a null pointer is fine.
                    958:         */
                    959:
                    960:        n->norm = n->parent->norm;
                    961:        assert(n->parent->type == ROFFT_BLOCK);
1.266     schwarze  962: }
                    963:
1.267     schwarze  964: void
                    965: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    966: {
                    967:        struct roff_node        *n;
                    968:
                    969:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    970:        n->string = roff_strdup(man->roff, word);
                    971:        roff_node_append(man, n);
1.286     schwarze  972:        n->flags |= NODE_VALID | NODE_ENDED;
1.267     schwarze  973:        man->next = ROFF_NEXT_SIBLING;
                    974: }
                    975:
                    976: void
                    977: roff_word_append(struct roff_man *man, const char *word)
                    978: {
                    979:        struct roff_node        *n;
                    980:        char                    *addstr, *newstr;
                    981:
                    982:        n = man->last;
                    983:        addstr = roff_strdup(man->roff, word);
                    984:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    985:        free(addstr);
                    986:        free(n->string);
                    987:        n->string = newstr;
                    988:        man->next = ROFF_NEXT_SIBLING;
1.268     schwarze  989: }
                    990:
                    991: void
                    992: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    993: {
                    994:        struct roff_node        *n;
                    995:
                    996:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    997:        roff_node_append(man, n);
                    998:        man->next = ROFF_NEXT_CHILD;
                    999: }
                   1000:
                   1001: struct roff_node *
                   1002: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                   1003: {
                   1004:        struct roff_node        *n;
                   1005:
                   1006:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                   1007:        roff_node_append(man, n);
                   1008:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1009:        return n;
1.267     schwarze 1010: }
                   1011:
1.266     schwarze 1012: struct roff_node *
                   1013: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                   1014: {
                   1015:        struct roff_node        *n;
                   1016:
                   1017:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                   1018:        roff_node_append(man, n);
                   1019:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1020:        return n;
1.266     schwarze 1021: }
                   1022:
                   1023: struct roff_node *
                   1024: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                   1025: {
                   1026:        struct roff_node        *n;
                   1027:
                   1028:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                   1029:        roff_node_append(man, n);
                   1030:        man->next = ROFF_NEXT_CHILD;
1.277     schwarze 1031:        return n;
1.267     schwarze 1032: }
                   1033:
1.321     schwarze 1034: static void
1.346     schwarze 1035: roff_addtbl(struct roff_man *man, int line, struct tbl_node *tbl)
1.267     schwarze 1036: {
                   1037:        struct roff_node        *n;
1.346     schwarze 1038:        struct tbl_span         *span;
1.267     schwarze 1039:
1.356     schwarze 1040:        if (man->meta.macroset == MACROSET_MAN)
1.302     schwarze 1041:                man_breakscope(man, ROFF_TS);
1.321     schwarze 1042:        while ((span = tbl_span(tbl)) != NULL) {
1.346     schwarze 1043:                n = roff_node_alloc(man, line, 0, ROFFT_TBL, TOKEN_NONE);
1.321     schwarze 1044:                n->span = span;
                   1045:                roff_node_append(man, n);
                   1046:                n->flags |= NODE_VALID | NODE_ENDED;
                   1047:                man->next = ROFF_NEXT_SIBLING;
                   1048:        }
1.266     schwarze 1049: }
                   1050:
                   1051: void
                   1052: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1053: {
                   1054:
                   1055:        /* Adjust siblings. */
                   1056:
                   1057:        if (n->prev)
                   1058:                n->prev->next = n->next;
                   1059:        if (n->next)
                   1060:                n->next->prev = n->prev;
                   1061:
                   1062:        /* Adjust parent. */
                   1063:
                   1064:        if (n->parent != NULL) {
                   1065:                if (n->parent->child == n)
                   1066:                        n->parent->child = n->next;
                   1067:                if (n->parent->last == n)
                   1068:                        n->parent->last = n->prev;
                   1069:        }
                   1070:
                   1071:        /* Adjust parse point. */
                   1072:
                   1073:        if (man == NULL)
                   1074:                return;
                   1075:        if (man->last == n) {
                   1076:                if (n->prev == NULL) {
                   1077:                        man->last = n->parent;
                   1078:                        man->next = ROFF_NEXT_CHILD;
                   1079:                } else {
                   1080:                        man->last = n->prev;
                   1081:                        man->next = ROFF_NEXT_SIBLING;
                   1082:                }
                   1083:        }
1.356     schwarze 1084:        if (man->meta.first == n)
                   1085:                man->meta.first = NULL;
1.344     schwarze 1086: }
                   1087:
                   1088: void
                   1089: roff_node_relink(struct roff_man *man, struct roff_node *n)
                   1090: {
                   1091:        roff_node_unlink(man, n);
                   1092:        n->prev = n->next = NULL;
                   1093:        roff_node_append(man, n);
1.266     schwarze 1094: }
                   1095:
                   1096: void
                   1097: roff_node_free(struct roff_node *n)
                   1098: {
                   1099:
                   1100:        if (n->args != NULL)
                   1101:                mdoc_argv_free(n->args);
                   1102:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1103:                free(n->norm);
1.347     schwarze 1104:        eqn_box_free(n->eqn);
1.266     schwarze 1105:        free(n->string);
1.374     schwarze 1106:        free(n->tag);
1.266     schwarze 1107:        free(n);
                   1108: }
                   1109:
                   1110: void
                   1111: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1112: {
                   1113:
                   1114:        while (n->child != NULL)
                   1115:                roff_node_delete(man, n->child);
                   1116:        roff_node_unlink(man, n);
                   1117:        roff_node_free(n);
1.370     schwarze 1118: }
                   1119:
                   1120: int
                   1121: roff_node_transparent(struct roff_node *n)
                   1122: {
                   1123:        if (n == NULL)
                   1124:                return 0;
                   1125:        if (n->type == ROFFT_COMMENT || n->flags & NODE_NOPRT)
                   1126:                return 1;
1.373     schwarze 1127:        return roff_tok_transparent(n->tok);
                   1128: }
                   1129:
                   1130: int
                   1131: roff_tok_transparent(enum roff_tok tok)
                   1132: {
                   1133:        switch (tok) {
1.370     schwarze 1134:        case ROFF_ft:
                   1135:        case ROFF_ll:
                   1136:        case ROFF_mc:
                   1137:        case ROFF_po:
                   1138:        case ROFF_ta:
                   1139:        case MDOC_Db:
                   1140:        case MDOC_Es:
                   1141:        case MDOC_Sm:
                   1142:        case MDOC_Tg:
                   1143:        case MAN_DT:
                   1144:        case MAN_UC:
                   1145:        case MAN_PD:
                   1146:        case MAN_AT:
                   1147:                return 1;
                   1148:        default:
                   1149:                return 0;
                   1150:        }
                   1151: }
                   1152:
                   1153: struct roff_node *
                   1154: roff_node_child(struct roff_node *n)
                   1155: {
                   1156:        for (n = n->child; roff_node_transparent(n); n = n->next)
                   1157:                continue;
                   1158:        return n;
                   1159: }
                   1160:
                   1161: struct roff_node *
                   1162: roff_node_prev(struct roff_node *n)
                   1163: {
                   1164:        do {
                   1165:                n = n->prev;
                   1166:        } while (roff_node_transparent(n));
                   1167:        return n;
                   1168: }
                   1169:
                   1170: struct roff_node *
                   1171: roff_node_next(struct roff_node *n)
                   1172: {
                   1173:        do {
                   1174:                n = n->next;
                   1175:        } while (roff_node_transparent(n));
                   1176:        return n;
1.269     schwarze 1177: }
                   1178:
                   1179: void
                   1180: deroff(char **dest, const struct roff_node *n)
                   1181: {
                   1182:        char    *cp;
                   1183:        size_t   sz;
                   1184:
1.371     schwarze 1185:        if (n->string == NULL) {
1.269     schwarze 1186:                for (n = n->child; n != NULL; n = n->next)
                   1187:                        deroff(dest, n);
                   1188:                return;
                   1189:        }
                   1190:
1.288     schwarze 1191:        /* Skip leading whitespace. */
1.269     schwarze 1192:
1.288     schwarze 1193:        for (cp = n->string; *cp != '\0'; cp++) {
1.289     schwarze 1194:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1195:                    strchr(" %&0^|~", cp[1]) != NULL)
1.269     schwarze 1196:                        cp++;
1.288     schwarze 1197:                else if ( ! isspace((unsigned char)*cp))
1.269     schwarze 1198:                        break;
                   1199:        }
                   1200:
1.289     schwarze 1201:        /* Skip trailing backslash. */
                   1202:
                   1203:        sz = strlen(cp);
1.290     schwarze 1204:        if (sz > 0 && cp[sz - 1] == '\\')
1.289     schwarze 1205:                sz--;
                   1206:
1.269     schwarze 1207:        /* Skip trailing whitespace. */
                   1208:
1.289     schwarze 1209:        for (; sz; sz--)
1.269     schwarze 1210:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1211:                        break;
                   1212:
                   1213:        /* Skip empty strings. */
                   1214:
                   1215:        if (sz == 0)
                   1216:                return;
                   1217:
                   1218:        if (*dest == NULL) {
                   1219:                *dest = mandoc_strndup(cp, sz);
                   1220:                return;
                   1221:        }
                   1222:
                   1223:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1224:        free(*dest);
                   1225:        *dest = cp;
1.266     schwarze 1226: }
                   1227:
                   1228: /* --- main functions of the roff parser ---------------------------------- */
                   1229:
1.388     schwarze 1230: /*
                   1231:  * Save comments preceding the title macro, for example in order to
                   1232:  * preserve Copyright and license headers in HTML output,
                   1233:  * provide diagnostics about RCS ids and trailing whitespace in comments,
                   1234:  * then discard comments including preceding whitespace.
                   1235:  * This function also handles input line continuation.
                   1236:  */
1.387     schwarze 1237: static int
1.388     schwarze 1238: roff_parse_comment(struct roff *r, struct buf *buf, int ln, int pos, char ec)
1.387     schwarze 1239: {
                   1240:        struct roff_node *n;    /* used for header comments */
                   1241:        const char      *start; /* start of the string to process */
                   1242:        const char      *cp;    /* for RCS id parsing */
                   1243:        char            *stesc; /* start of an escape sequence ('\\') */
                   1244:        char            *ep;    /* end of comment string */
                   1245:        int              rcsid; /* kind of RCS id seen */
                   1246:
                   1247:        for (start = stesc = buf->buf + pos;; stesc++) {
1.388     schwarze 1248:                /*
                   1249:                 * XXX Ugly hack: Remove the newline character that
                   1250:                 * mparse_buf_r() appended to mark the end of input
                   1251:                 * if it is not preceded by an escape character.
                   1252:                 */
                   1253:                if (stesc[0] == '\n') {
                   1254:                        assert(stesc[1] == '\0');
                   1255:                        stesc[0] = '\0';
                   1256:                }
                   1257:
1.387     schwarze 1258:                /* The line ends without continuation or comment. */
                   1259:                if (stesc[0] == '\0')
                   1260:                        return ROFF_CONT;
                   1261:
                   1262:                /* Unescaped byte: skip it. */
1.388     schwarze 1263:                if (stesc[0] != ec)
1.387     schwarze 1264:                        continue;
                   1265:
1.388     schwarze 1266:                /*
                   1267:                 * XXX Ugly hack: Do not attempt to append another line
                   1268:                 * if the function mparse_buf_r() appended a newline
                   1269:                 * character to indicate the end of input.
                   1270:                 */
                   1271:                if (stesc[1] == '\n') {
                   1272:                        assert(stesc[2] == '\0');
                   1273:                        stesc[0] = '\0';
                   1274:                        return ROFF_CONT;
                   1275:                }
                   1276:
                   1277:                /*
                   1278:                 * An escape character at the end of an input line
                   1279:                 * requests line continuation.
                   1280:                 */
1.387     schwarze 1281:                if (stesc[1] == '\0') {
                   1282:                        stesc[0] = '\0';
                   1283:                        return ROFF_IGN | ROFF_APPEND;
                   1284:                }
                   1285:
                   1286:                /* Found a comment: process it. */
                   1287:                if (stesc[1] == '"' || stesc[1] == '#')
                   1288:                        break;
                   1289:
                   1290:                /* Escaped escape character: skip them both. */
1.388     schwarze 1291:                if (stesc[1] == ec)
1.387     schwarze 1292:                        stesc++;
                   1293:        }
                   1294:
                   1295:        /* Look for an RCS id in the comment. */
                   1296:
                   1297:        rcsid = 0;
                   1298:        if ((cp = strstr(stesc + 2, "$" "OpenBSD")) != NULL) {
                   1299:                rcsid = 1 << MANDOC_OS_OPENBSD;
                   1300:                cp += 8;
                   1301:        } else if ((cp = strstr(stesc + 2, "$" "NetBSD")) != NULL) {
                   1302:                rcsid = 1 << MANDOC_OS_NETBSD;
                   1303:                cp += 7;
                   1304:        }
                   1305:        if (cp != NULL && isalnum((unsigned char)*cp) == 0 &&
                   1306:            strchr(cp, '$') != NULL) {
                   1307:                if (r->man->meta.rcsids & rcsid)
                   1308:                        mandoc_msg(MANDOCERR_RCS_REP, ln,
                   1309:                            (int)(stesc - buf->buf) + 2, "%s", stesc + 1);
                   1310:                r->man->meta.rcsids |= rcsid;
                   1311:        }
                   1312:
                   1313:        /* Warn about trailing whitespace at the end of the comment. */
                   1314:
                   1315:        ep = strchr(stesc + 2, '\0') - 1;
                   1316:        if (*ep == '\n')
                   1317:                *ep-- = '\0';
                   1318:        if (*ep == ' ' || *ep == '\t')
                   1319:                mandoc_msg(MANDOCERR_SPACE_EOL,
                   1320:                    ln, (int)(ep - buf->buf), NULL);
                   1321:
                   1322:        /* Save comments preceding the title macro in the syntax tree. */
                   1323:
                   1324:        if (r->options & MPARSE_COMMENT) {
                   1325:                while (*ep == ' ' || *ep == '\t')
                   1326:                        ep--;
                   1327:                ep[1] = '\0';
                   1328:                n = roff_node_alloc(r->man, ln, stesc + 1 - buf->buf,
                   1329:                    ROFFT_COMMENT, TOKEN_NONE);
                   1330:                n->string = mandoc_strdup(stesc + 2);
                   1331:                roff_node_append(r->man, n);
                   1332:                n->flags |= NODE_VALID | NODE_ENDED;
                   1333:                r->man->next = ROFF_NEXT_SIBLING;
                   1334:        }
                   1335:
                   1336:        /* The comment requests line continuation. */
                   1337:
                   1338:        if (stesc[1] == '#') {
                   1339:                *stesc = '\0';
                   1340:                return ROFF_IGN | ROFF_APPEND;
                   1341:        }
                   1342:
                   1343:        /* Discard the comment including preceding whitespace. */
                   1344:
                   1345:        while (stesc > start && stesc[-1] == ' ' &&
                   1346:            (stesc == start + 1 || stesc[-2] != '\\'))
                   1347:                stesc--;
                   1348:        *stesc = '\0';
                   1349:        return ROFF_CONT;
                   1350: }
                   1351:
1.94      kristaps 1352: /*
1.355     schwarze 1353:  * In the current line, expand escape sequences that produce parsable
                   1354:  * input text.  Also check the syntax of the remaining escape sequences,
                   1355:  * which typically produce output glyphs or change formatter state.
1.154     kristaps 1356:  */
1.340     schwarze 1357: static int
1.388     schwarze 1358: roff_expand(struct roff *r, struct buf *buf, int ln, int pos, char ec)
1.94      kristaps 1359: {
1.388     schwarze 1360:        char             ubuf[24];      /* buffer to print a number */
                   1361:        struct mctx     *ctx;           /* current macro call context */
                   1362:        const char      *res;           /* the string to be pasted */
                   1363:        const char      *src;           /* source for copying */
                   1364:        char            *dst;           /* destination for copying */
                   1365:        int              iesc;          /* index of leading escape char */
                   1366:        int              inam;          /* index of the escape name */
                   1367:        int              iarg;          /* index beginning the argument */
                   1368:        int              iendarg;       /* index right after the argument */
                   1369:        int              iend;          /* index right after the sequence */
1.393     schwarze 1370:        int              isrc, idst;    /* to reduce \\ and \. in names */
1.388     schwarze 1371:        int              deftype;       /* type of definition to paste */
                   1372:        int              argi;          /* macro argument index */
                   1373:        int              quote_args;    /* true for \\$@, false for \\$* */
                   1374:        int              asz;           /* length of the replacement */
                   1375:        int              rsz;           /* length of the rest of the string */
                   1376:        int              npos;          /* position in numeric expression */
1.181     schwarze 1377:        int              expand_count;  /* to avoid infinite loops */
1.303     schwarze 1378:
1.170     schwarze 1379:        expand_count = 0;
1.388     schwarze 1380:        while (buf->buf[pos] != '\0') {
1.170     schwarze 1381:
1.388     schwarze 1382:                /*
                   1383:                 * Skip plain ASCII characters.
                   1384:                 * If we have a non-standard escape character,
                   1385:                 * escape literal backslashes because all processing in
                   1386:                 * subsequent functions uses the standard escaping rules.
                   1387:                 */
1.205     schwarze 1388:
1.388     schwarze 1389:                if (buf->buf[pos] != ec) {
                   1390:                        if (ec != ASCII_ESC && buf->buf[pos] == '\\') {
                   1391:                                roff_expand_patch(buf, pos, "\\e", pos + 1);
                   1392:                                pos++;
1.303     schwarze 1393:                        }
1.388     schwarze 1394:                        pos++;
1.205     schwarze 1395:                        continue;
1.303     schwarze 1396:                }
1.205     schwarze 1397:
1.388     schwarze 1398:                /*
                   1399:                 * Parse escape sequences,
                   1400:                 * issue diagnostic messages when appropriate,
                   1401:                 * and skip sequences that do not need expansion.
                   1402:                 * If we have a non-standard escape character, translate
                   1403:                 * it to backslashes and translate backslashes to \e.
                   1404:                 */
1.205     schwarze 1405:
1.392     schwarze 1406:                if (roff_escape(buf->buf, ln, pos, &iesc, &inam,
                   1407:                    &iarg, &iendarg, &iend) != ESCAPE_EXPAND) {
1.388     schwarze 1408:                        while (pos < iend) {
                   1409:                                if (buf->buf[pos] == ec) {
                   1410:                                        buf->buf[pos] = '\\';
                   1411:                                        if (pos + 1 < iend)
                   1412:                                                pos++;
                   1413:                                } else if (buf->buf[pos] == '\\') {
                   1414:                                        roff_expand_patch(buf,
                   1415:                                            pos, "\\e", pos + 1);
                   1416:                                        pos++;
                   1417:                                        iend++;
                   1418:                                }
                   1419:                                pos++;
1.352     schwarze 1420:                        }
1.205     schwarze 1421:                        continue;
1.152     kristaps 1422:                }
1.388     schwarze 1423:
1.393     schwarze 1424:                /* Reduce \\ and \. in names. */
                   1425:
                   1426:                if (buf->buf[inam] == '*' || buf->buf[inam] == 'n') {
                   1427:                        isrc = idst = iarg;
                   1428:                        while (isrc < iendarg) {
                   1429:                                if (isrc + 1 < iendarg &&
                   1430:                                    buf->buf[isrc] == '\\' &&
                   1431:                                    (buf->buf[isrc + 1] == '\\' ||
                   1432:                                     buf->buf[isrc + 1] == '.'))
                   1433:                                        isrc++;
                   1434:                                buf->buf[idst++] = buf->buf[isrc++];
                   1435:                        }
                   1436:                        iendarg -= isrc - idst;
                   1437:                }
                   1438:
1.388     schwarze 1439:                /* Handle expansion. */
                   1440:
                   1441:                res = NULL;
                   1442:                switch (buf->buf[inam]) {
                   1443:                case '*':
                   1444:                        if (iendarg == iarg)
1.206     schwarze 1445:                                break;
1.388     schwarze 1446:                        deftype = ROFFDEF_USER | ROFFDEF_PRE;
                   1447:                        if ((res = roff_getstrn(r, buf->buf + iarg,
                   1448:                            iendarg - iarg, &deftype)) != NULL)
1.206     schwarze 1449:                                break;
1.94      kristaps 1450:
1.388     schwarze 1451:                        /*
                   1452:                         * If not overriden,
                   1453:                         * let \*(.T through to the formatters.
                   1454:                         */
1.94      kristaps 1455:
1.388     schwarze 1456:                        if (iendarg - iarg == 2 &&
                   1457:                            buf->buf[iarg] == '.' &&
                   1458:                            buf->buf[iarg + 1] == 'T') {
                   1459:                                roff_setstrn(&r->strtab, ".T", 2, NULL, 0, 0);
                   1460:                                pos = iend;
1.253     schwarze 1461:                                continue;
                   1462:                        }
1.94      kristaps 1463:
1.388     schwarze 1464:                        mandoc_msg(MANDOCERR_STR_UNDEF, ln, iesc,
                   1465:                            "%.*s", iendarg - iarg, buf->buf + iarg);
                   1466:                        break;
1.108     schwarze 1467:
1.339     schwarze 1468:                case '$':
                   1469:                        if (r->mstackpos < 0) {
1.388     schwarze 1470:                                mandoc_msg(MANDOCERR_ARG_UNDEF, ln, iesc,
                   1471:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1472:                                break;
                   1473:                        }
                   1474:                        ctx = r->mstack + r->mstackpos;
1.388     schwarze 1475:                        argi = buf->buf[iarg] - '1';
                   1476:                        if (argi >= 0 && argi <= 8) {
                   1477:                                if (argi < ctx->argc)
                   1478:                                        res = ctx->argv[argi];
1.339     schwarze 1479:                                break;
                   1480:                        }
1.388     schwarze 1481:                        if (buf->buf[iarg] == '*')
1.339     schwarze 1482:                                quote_args = 0;
1.388     schwarze 1483:                        else if (buf->buf[iarg] == '@')
1.339     schwarze 1484:                                quote_args = 1;
                   1485:                        else {
1.388     schwarze 1486:                                mandoc_msg(MANDOCERR_ARG_NONUM, ln, iesc,
                   1487:                                    "%.*s", iend - iesc, buf->buf + iesc);
1.339     schwarze 1488:                                break;
                   1489:                        }
                   1490:                        asz = 0;
1.388     schwarze 1491:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1492:                                if (argi)
1.339     schwarze 1493:                                        asz++;  /* blank */
                   1494:                                if (quote_args)
                   1495:                                        asz += 2;  /* quotes */
1.388     schwarze 1496:                                asz += strlen(ctx->argv[argi]);
1.339     schwarze 1497:                        }
1.388     schwarze 1498:                        if (asz != iend - iesc) {
                   1499:                                rsz = buf->sz - iend;
                   1500:                                if (asz < iend - iesc)
                   1501:                                        memmove(buf->buf + iesc + asz,
                   1502:                                            buf->buf + iend, rsz);
                   1503:                                buf->sz = iesc + asz + rsz;
                   1504:                                buf->buf = mandoc_realloc(buf->buf, buf->sz);
                   1505:                                if (asz > iend - iesc)
                   1506:                                        memmove(buf->buf + iesc + asz,
                   1507:                                            buf->buf + iend, rsz);
1.339     schwarze 1508:                        }
1.388     schwarze 1509:                        dst = buf->buf + iesc;
                   1510:                        for (argi = 0; argi < ctx->argc; argi++) {
                   1511:                                if (argi)
                   1512:                                        *dst++ = ' ';
1.339     schwarze 1513:                                if (quote_args)
1.388     schwarze 1514:                                        *dst++ = '"';
                   1515:                                src = ctx->argv[argi];
                   1516:                                while (*src != '\0')
                   1517:                                        *dst++ = *src++;
1.339     schwarze 1518:                                if (quote_args)
1.388     schwarze 1519:                                        *dst++ = '"';
1.339     schwarze 1520:                        }
                   1521:                        continue;
1.391     schwarze 1522:                case 'A':
                   1523:                        ubuf[0] = iendarg > iarg ? '1' : '0';
                   1524:                        ubuf[1] = '\0';
                   1525:                        res = ubuf;
                   1526:                        break;
1.207     schwarze 1527:                case 'B':
1.206     schwarze 1528:                        npos = 0;
1.388     schwarze 1529:                        ubuf[0] = iendarg > iarg && iend > iendarg &&
                   1530:                            roff_evalnum(r, ln, buf->buf + iarg, &npos,
                   1531:                                         NULL, ROFFNUM_SCALE) &&
                   1532:                            npos == iendarg - iarg ? '1' : '0';
1.206     schwarze 1533:                        ubuf[1] = '\0';
1.388     schwarze 1534:                        res = ubuf;
1.206     schwarze 1535:                        break;
1.389     schwarze 1536:                case 'V':
                   1537:                        mandoc_msg(MANDOCERR_UNSUPP, ln, iesc,
                   1538:                            "%.*s", iend - iesc, buf->buf + iesc);
                   1539:                        roff_expand_patch(buf, iendarg, "}", iend);
                   1540:                        roff_expand_patch(buf, iesc, "${", iarg);
                   1541:                        continue;
1.390     schwarze 1542:                case 'g':
                   1543:                        break;
1.207     schwarze 1544:                case 'n':
1.388     schwarze 1545:                        if (iendarg > iarg)
1.218     schwarze 1546:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.388     schwarze 1547:                                    roff_getregn(r, buf->buf + iarg,
                   1548:                                    iendarg - iarg, buf->buf[inam + 1]));
1.218     schwarze 1549:                        else
                   1550:                                ubuf[0] = '\0';
1.388     schwarze 1551:                        res = ubuf;
1.206     schwarze 1552:                        break;
1.207     schwarze 1553:                case 'w':
1.388     schwarze 1554:                        (void)snprintf(ubuf, sizeof(ubuf),
                   1555:                            "%d", (iendarg - iarg) * 24);
                   1556:                        res = ubuf;
                   1557:                        break;
                   1558:                default:
1.206     schwarze 1559:                        break;
                   1560:                }
1.388     schwarze 1561:                if (res == NULL)
1.142     kristaps 1562:                        res = "";
1.388     schwarze 1563:                if (++expand_count > EXPAND_LIMIT ||
                   1564:                    buf->sz + strlen(res) > SHRT_MAX) {
                   1565:                        mandoc_msg(MANDOCERR_ROFFLOOP, ln, iesc, NULL);
1.277     schwarze 1566:                        return ROFF_IGN;
1.94      kristaps 1567:                }
1.388     schwarze 1568:                roff_expand_patch(buf, iesc, res, iend);
                   1569:        }
                   1570:        return ROFF_CONT;
                   1571: }
1.94      kristaps 1572:
1.388     schwarze 1573: /*
                   1574:  * Replace the substring from the start position (inclusive)
                   1575:  * to end position (exclusive) with the repl(acement) string.
                   1576:  */
                   1577: static void
                   1578: roff_expand_patch(struct buf *buf, int start, const char *repl, int end)
                   1579: {
                   1580:        char    *nbuf;
1.108     schwarze 1581:
1.389     schwarze 1582:        buf->sz = mandoc_asprintf(&nbuf, "%.*s%s%s", start, buf->buf,
                   1583:            repl, buf->buf + end) + 1;
1.388     schwarze 1584:        free(buf->buf);
                   1585:        buf->buf = nbuf;
1.154     kristaps 1586: }
1.353     schwarze 1587:
                   1588: /*
                   1589:  * Parse a quoted or unquoted roff-style request or macro argument.
                   1590:  * Return a pointer to the parsed argument, which is either the original
                   1591:  * pointer or advanced by one byte in case the argument is quoted.
                   1592:  * NUL-terminate the argument in place.
                   1593:  * Collapse pairs of quotes inside quoted arguments.
                   1594:  * Advance the argument pointer to the next argument,
                   1595:  * or to the NUL byte terminating the argument line.
                   1596:  */
                   1597: char *
1.355     schwarze 1598: roff_getarg(struct roff *r, char **cpp, int ln, int *pos)
1.353     schwarze 1599: {
1.355     schwarze 1600:        struct buf       buf;
1.366     schwarze 1601:        char            *cp, *start;
1.355     schwarze 1602:        int              newesc, pairs, quoted, white;
1.353     schwarze 1603:
                   1604:        /* Quoting can only start with a new word. */
                   1605:        start = *cpp;
                   1606:        quoted = 0;
                   1607:        if ('"' == *start) {
                   1608:                quoted = 1;
                   1609:                start++;
                   1610:        }
                   1611:
1.355     schwarze 1612:        newesc = pairs = white = 0;
1.353     schwarze 1613:        for (cp = start; '\0' != *cp; cp++) {
                   1614:
                   1615:                /*
                   1616:                 * Move the following text left
                   1617:                 * after quoted quotes and after "\\" and "\t".
                   1618:                 */
                   1619:                if (pairs)
                   1620:                        cp[-pairs] = cp[0];
                   1621:
                   1622:                if ('\\' == cp[0]) {
                   1623:                        /*
                   1624:                         * In copy mode, translate double to single
                   1625:                         * backslashes and backslash-t to literal tabs.
                   1626:                         */
                   1627:                        switch (cp[1]) {
                   1628:                        case 'a':
                   1629:                        case 't':
1.354     schwarze 1630:                                cp[-pairs] = '\t';
1.355     schwarze 1631:                                pairs++;
                   1632:                                cp++;
                   1633:                                break;
1.353     schwarze 1634:                        case '\\':
1.395   ! schwarze 1635:                                /*
        !          1636:                                 * Signal to roff_expand() that an escape
        !          1637:                                 * sequence resulted from copy-in processing
        !          1638:                                 * and needs to be checked or interpolated.
        !          1639:                                 */
        !          1640:                                cp[-pairs] = ASCII_ESC;
1.355     schwarze 1641:                                newesc = 1;
1.353     schwarze 1642:                                pairs++;
                   1643:                                cp++;
                   1644:                                break;
                   1645:                        case ' ':
                   1646:                                /* Skip escaped blanks. */
                   1647:                                if (0 == quoted)
                   1648:                                        cp++;
                   1649:                                break;
                   1650:                        default:
                   1651:                                break;
                   1652:                        }
                   1653:                } else if (0 == quoted) {
                   1654:                        if (' ' == cp[0]) {
                   1655:                                /* Unescaped blanks end unquoted args. */
                   1656:                                white = 1;
                   1657:                                break;
                   1658:                        }
                   1659:                } else if ('"' == cp[0]) {
                   1660:                        if ('"' == cp[1]) {
                   1661:                                /* Quoted quotes collapse. */
                   1662:                                pairs++;
                   1663:                                cp++;
                   1664:                        } else {
                   1665:                                /* Unquoted quotes end quoted args. */
                   1666:                                quoted = 2;
                   1667:                                break;
                   1668:                        }
                   1669:                }
                   1670:        }
                   1671:
                   1672:        /* Quoted argument without a closing quote. */
                   1673:        if (1 == quoted)
                   1674:                mandoc_msg(MANDOCERR_ARG_QUOTE, ln, *pos, NULL);
                   1675:
                   1676:        /* NUL-terminate this argument and move to the next one. */
                   1677:        if (pairs)
                   1678:                cp[-pairs] = '\0';
                   1679:        if ('\0' != *cp) {
                   1680:                *cp++ = '\0';
                   1681:                while (' ' == *cp)
                   1682:                        cp++;
                   1683:        }
                   1684:        *pos += (int)(cp - start) + (quoted ? 1 : 0);
                   1685:        *cpp = cp;
                   1686:
                   1687:        if ('\0' == *cp && (white || ' ' == cp[-1]))
                   1688:                mandoc_msg(MANDOCERR_SPACE_EOL, ln, *pos, NULL);
                   1689:
1.355     schwarze 1690:        start = mandoc_strdup(start);
                   1691:        if (newesc == 0)
                   1692:                return start;
                   1693:
                   1694:        buf.buf = start;
                   1695:        buf.sz = strlen(start) + 1;
                   1696:        buf.next = NULL;
                   1697:        if (roff_expand(r, &buf, ln, 0, ASCII_ESC) & ROFF_IGN) {
                   1698:                free(buf.buf);
                   1699:                buf.buf = mandoc_strdup("");
                   1700:        }
                   1701:        return buf.buf;
1.353     schwarze 1702: }
                   1703:
1.154     kristaps 1704:
                   1705: /*
1.275     schwarze 1706:  * Process text streams.
1.154     kristaps 1707:  */
1.340     schwarze 1708: static int
1.305     schwarze 1709: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154     kristaps 1710: {
                   1711:        size_t           sz;
                   1712:        const char      *start;
1.178     schwarze 1713:        char            *p;
                   1714:        int              isz;
1.154     kristaps 1715:        enum mandoc_esc  esc;
                   1716:
1.275     schwarze 1717:        /* Spring the input line trap. */
                   1718:
                   1719:        if (roffit_lines == 1) {
                   1720:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1721:                free(buf->buf);
                   1722:                buf->buf = p;
                   1723:                buf->sz = isz + 1;
                   1724:                *offs = 0;
                   1725:                free(roffit_macro);
                   1726:                roffit_lines = 0;
1.277     schwarze 1727:                return ROFF_REPARSE;
1.275     schwarze 1728:        } else if (roffit_lines > 1)
                   1729:                --roffit_lines;
                   1730:
1.305     schwarze 1731:        if (roffce_node != NULL && buf->buf[pos] != '\0') {
                   1732:                if (roffce_lines < 1) {
                   1733:                        r->man->last = roffce_node;
                   1734:                        r->man->next = ROFF_NEXT_SIBLING;
                   1735:                        roffce_lines = 0;
                   1736:                        roffce_node = NULL;
                   1737:                } else
                   1738:                        roffce_lines--;
                   1739:        }
                   1740:
1.275     schwarze 1741:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1742:
1.238     schwarze 1743:        start = p = buf->buf + pos;
1.154     kristaps 1744:
1.238     schwarze 1745:        while (*p != '\0') {
1.154     kristaps 1746:                sz = strcspn(p, "-\\");
                   1747:                p += sz;
                   1748:
1.238     schwarze 1749:                if (*p == '\0')
1.159     kristaps 1750:                        break;
                   1751:
1.238     schwarze 1752:                if (*p == '\\') {
1.154     kristaps 1753:                        /* Skip over escapes. */
                   1754:                        p++;
1.189     schwarze 1755:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238     schwarze 1756:                        if (esc == ESCAPE_ERROR)
1.154     kristaps 1757:                                break;
1.264     schwarze 1758:                        while (*p == '-')
                   1759:                                p++;
1.155     kristaps 1760:                        continue;
1.159     kristaps 1761:                } else if (p == start) {
1.158     kristaps 1762:                        p++;
1.155     kristaps 1763:                        continue;
1.158     kristaps 1764:                }
1.155     kristaps 1765:
1.171     schwarze 1766:                if (isalpha((unsigned char)p[-1]) &&
                   1767:                    isalpha((unsigned char)p[1]))
1.155     kristaps 1768:                        *p = ASCII_HYPH;
                   1769:                p++;
1.94      kristaps 1770:        }
1.277     schwarze 1771:        return ROFF_CONT;
1.94      kristaps 1772: }
                   1773:
1.340     schwarze 1774: int
1.377     schwarze 1775: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
1.67      kristaps 1776: {
1.294     schwarze 1777:        enum roff_tok    t;
1.340     schwarze 1778:        int              e;
1.238     schwarze 1779:        int              pos;   /* parse point */
1.243     schwarze 1780:        int              spos;  /* saved parse point for messages */
1.238     schwarze 1781:        int              ppos;  /* original offset in buf->buf */
                   1782:        int              ctl;   /* macro line (boolean) */
                   1783:
                   1784:        ppos = pos = *offs;
1.377     schwarze 1785:
                   1786:        if (len > 80 && r->tbl == NULL && r->eqn == NULL &&
                   1787:            (r->man->flags & ROFF_NOFILL) == 0 &&
                   1788:            strchr(" .\\", buf->buf[pos]) == NULL &&
                   1789:            buf->buf[pos] != r->control &&
                   1790:            strcspn(buf->buf, " ") < 80)
                   1791:                mandoc_msg(MANDOCERR_TEXT_LONG, ln, (int)len - 1,
                   1792:                    "%.20s...", buf->buf + pos);
1.79      kristaps 1793:
1.230     schwarze 1794:        /* Handle in-line equation delimiters. */
                   1795:
1.236     schwarze 1796:        if (r->tbl == NULL &&
                   1797:            r->last_eqn != NULL && r->last_eqn->delim &&
1.230     schwarze 1798:            (r->eqn == NULL || r->eqn_inline)) {
1.238     schwarze 1799:                e = roff_eqndelim(r, buf, pos);
1.230     schwarze 1800:                if (e == ROFF_REPARSE)
1.277     schwarze 1801:                        return e;
1.230     schwarze 1802:                assert(e == ROFF_CONT);
                   1803:        }
                   1804:
1.387     schwarze 1805:        /* Handle comments and escape sequences. */
                   1806:
                   1807:        e = roff_parse_comment(r, buf, ln, pos, r->escape);
                   1808:        if ((e & ROFF_MASK) == ROFF_IGN)
                   1809:                return e;
                   1810:        assert(e == ROFF_CONT);
1.94      kristaps 1811:
1.355     schwarze 1812:        e = roff_expand(r, buf, ln, pos, r->escape);
1.340     schwarze 1813:        if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1814:                return e;
1.238     schwarze 1815:        assert(e == ROFF_CONT);
1.94      kristaps 1816:
1.238     schwarze 1817:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.130     kristaps 1818:
1.94      kristaps 1819:        /*
1.79      kristaps 1820:         * First, if a scope is open and we're not a macro, pass the
1.252     schwarze 1821:         * text through the macro's filter.
                   1822:         * Equations process all content themselves.
                   1823:         * Tables process almost all content themselves, but we want
                   1824:         * to warn about macros before passing it there.
1.79      kristaps 1825:         */
1.74      kristaps 1826:
1.252     schwarze 1827:        if (r->last != NULL && ! ctl) {
1.78      kristaps 1828:                t = r->last->tok;
1.238     schwarze 1829:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.340     schwarze 1830:                if ((e & ROFF_MASK) == ROFF_IGN)
1.277     schwarze 1831:                        return e;
1.340     schwarze 1832:                e &= ~ROFF_MASK;
                   1833:        } else
                   1834:                e = ROFF_IGN;
1.319     schwarze 1835:        if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
                   1836:                eqn_read(r->eqn, buf->buf + ppos);
1.340     schwarze 1837:                return e;
1.319     schwarze 1838:        }
1.321     schwarze 1839:        if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
                   1840:                tbl_read(r->tbl, ln, buf->buf, ppos);
1.346     schwarze 1841:                roff_addtbl(r->man, ln, r->tbl);
1.340     schwarze 1842:                return e;
1.321     schwarze 1843:        }
1.367     schwarze 1844:        if ( ! ctl) {
                   1845:                r->options &= ~MPARSE_COMMENT;
1.340     schwarze 1846:                return roff_parsetext(r, buf, pos, offs) | e;
1.367     schwarze 1847:        }
1.228     schwarze 1848:
                   1849:        /* Skip empty request lines. */
                   1850:
1.238     schwarze 1851:        if (buf->buf[pos] == '"') {
1.350     schwarze 1852:                mandoc_msg(MANDOCERR_COMMENT_BAD, ln, pos, NULL);
1.277     schwarze 1853:                return ROFF_IGN;
1.238     schwarze 1854:        } else if (buf->buf[pos] == '\0')
1.277     schwarze 1855:                return ROFF_IGN;
1.67      kristaps 1856:
1.79      kristaps 1857:        /*
                   1858:         * If a scope is open, go to the child handler for that macro,
                   1859:         * as it may want to preprocess before doing anything with it.
                   1860:         */
1.78      kristaps 1861:
1.79      kristaps 1862:        if (r->last) {
                   1863:                t = r->last->tok;
1.277     schwarze 1864:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 1865:        }
1.78      kristaps 1866:
1.367     schwarze 1867:        r->options &= ~MPARSE_COMMENT;
1.243     schwarze 1868:        spos = pos;
                   1869:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.386     schwarze 1870:        return roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
                   1871: }
                   1872:
                   1873: /*
                   1874:  * Handle a new request or macro.
                   1875:  * May be called outside any scope or from inside a conditional scope.
                   1876:  */
                   1877: static int
                   1878: roff_req_or_macro(ROFF_ARGS) {
1.243     schwarze 1879:
1.386     schwarze 1880:        /* For now, tables ignore most macros and some request. */
1.243     schwarze 1881:
1.386     schwarze 1882:        if (r->tbl != NULL && (tok == TOKEN_NONE || tok == ROFF_TS ||
                   1883:            tok == ROFF_br || tok == ROFF_ce || tok == ROFF_rj ||
                   1884:            tok == ROFF_sp)) {
1.350     schwarze 1885:                mandoc_msg(MANDOCERR_TBLMACRO,
1.386     schwarze 1886:                    ln, ppos, "%s", buf->buf + ppos);
                   1887:                if (tok != TOKEN_NONE)
1.277     schwarze 1888:                        return ROFF_IGN;
1.257     schwarze 1889:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1890:                        pos++;
1.291     schwarze 1891:                while (buf->buf[pos] == ' ')
1.257     schwarze 1892:                        pos++;
1.321     schwarze 1893:                tbl_read(r->tbl, ln, buf->buf, pos);
1.346     schwarze 1894:                roff_addtbl(r->man, ln, r->tbl);
1.321     schwarze 1895:                return ROFF_IGN;
1.243     schwarze 1896:        }
                   1897:
1.305     schwarze 1898:        /* For now, let high level macros abort .ce mode. */
                   1899:
1.386     schwarze 1900:        if (roffce_node != NULL &&
                   1901:            (tok == TOKEN_NONE || tok == ROFF_Dd || tok == ROFF_EQ ||
                   1902:             tok == ROFF_TH || tok == ROFF_TS)) {
1.305     schwarze 1903:                r->man->last = roffce_node;
                   1904:                r->man->next = ROFF_NEXT_SIBLING;
                   1905:                roffce_lines = 0;
                   1906:                roffce_node = NULL;
                   1907:        }
                   1908:
1.79      kristaps 1909:        /*
1.243     schwarze 1910:         * This is neither a roff request nor a user-defined macro.
                   1911:         * Let the standard macro set parsers handle it.
1.79      kristaps 1912:         */
1.67      kristaps 1913:
1.386     schwarze 1914:        if (tok == TOKEN_NONE)
1.277     schwarze 1915:                return ROFF_CONT;
1.243     schwarze 1916:
1.386     schwarze 1917:        /* Execute a roff request or a user-defined macro. */
1.67      kristaps 1918:
1.386     schwarze 1919:        return (*roffs[tok].proc)(r, tok, buf, ln, ppos, pos, offs);
1.74      kristaps 1920: }
                   1921:
1.339     schwarze 1922: /*
                   1923:  * Internal interface function to tell the roff parser that execution
                   1924:  * of the current macro ended.  This is required because macro
                   1925:  * definitions usually do not end with a .return request.
                   1926:  */
                   1927: void
                   1928: roff_userret(struct roff *r)
                   1929: {
                   1930:        struct mctx     *ctx;
                   1931:        int              i;
                   1932:
                   1933:        assert(r->mstackpos >= 0);
                   1934:        ctx = r->mstack + r->mstackpos;
                   1935:        for (i = 0; i < ctx->argc; i++)
                   1936:                free(ctx->argv[i]);
                   1937:        ctx->argc = 0;
                   1938:        r->mstackpos--;
                   1939: }
                   1940:
1.117     kristaps 1941: void
1.74      kristaps 1942: roff_endparse(struct roff *r)
                   1943: {
1.321     schwarze 1944:        if (r->last != NULL)
1.350     schwarze 1945:                mandoc_msg(MANDOCERR_BLK_NOEND, r->last->line,
                   1946:                    r->last->col, "%s", roff_name[r->last->tok]);
1.117     kristaps 1947:
1.321     schwarze 1948:        if (r->eqn != NULL) {
1.350     schwarze 1949:                mandoc_msg(MANDOCERR_BLK_NOEND,
1.319     schwarze 1950:                    r->eqn->node->line, r->eqn->node->pos, "EQ");
                   1951:                eqn_parse(r->eqn);
                   1952:                r->eqn = NULL;
1.125     kristaps 1953:        }
                   1954:
1.321     schwarze 1955:        if (r->tbl != NULL) {
1.346     schwarze 1956:                tbl_end(r->tbl, 1);
1.321     schwarze 1957:                r->tbl = NULL;
1.117     kristaps 1958:        }
1.67      kristaps 1959: }
                   1960:
                   1961: /*
1.386     schwarze 1962:  * Parse the request or macro name at buf[*pos].
                   1963:  * Return ROFF_RENAMED, ROFF_USERDEF, or a ROFF_* token value.
                   1964:  * For empty, undefined, mdoc(7), and man(7) macros, return TOKEN_NONE.
                   1965:  * As a side effect, set r->current_string to the definition or to NULL.
1.67      kristaps 1966:  */
1.294     schwarze 1967: static enum roff_tok
1.214     schwarze 1968: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67      kristaps 1969: {
1.214     schwarze 1970:        char            *cp;
1.106     kristaps 1971:        const char      *mac;
                   1972:        size_t           maclen;
1.315     schwarze 1973:        int              deftype;
1.294     schwarze 1974:        enum roff_tok    t;
1.67      kristaps 1975:
1.214     schwarze 1976:        cp = buf + *pos;
                   1977:
                   1978:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294     schwarze 1979:                return TOKEN_NONE;
1.67      kristaps 1980:
1.214     schwarze 1981:        mac = cp;
                   1982:        maclen = roff_getname(r, &cp, ln, ppos);
1.67      kristaps 1983:
1.315     schwarze 1984:        deftype = ROFFDEF_USER | ROFFDEF_REN;
                   1985:        r->current_string = roff_getstrn(r, mac, maclen, &deftype);
                   1986:        switch (deftype) {
                   1987:        case ROFFDEF_USER:
                   1988:                t = ROFF_USERDEF;
                   1989:                break;
                   1990:        case ROFFDEF_REN:
                   1991:                t = ROFF_RENAMED;
                   1992:                break;
                   1993:        default:
                   1994:                t = roffhash_find(r->reqtab, mac, maclen);
                   1995:                break;
                   1996:        }
1.294     schwarze 1997:        if (t != TOKEN_NONE)
1.214     schwarze 1998:                *pos = cp - buf;
1.325     schwarze 1999:        else if (deftype == ROFFDEF_UNDEF) {
                   2000:                /* Using an undefined macro defines it to be empty. */
                   2001:                roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
                   2002:                roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
                   2003:        }
1.277     schwarze 2004:        return t;
1.67      kristaps 2005: }
                   2006:
1.266     schwarze 2007: /* --- handling of request blocks ----------------------------------------- */
                   2008:
1.375     schwarze 2009: /*
                   2010:  * Close a macro definition block or an "ignore" block.
                   2011:  */
1.340     schwarze 2012: static int
1.76      kristaps 2013: roff_cblock(ROFF_ARGS)
1.67      kristaps 2014: {
1.375     schwarze 2015:        int      rr;
1.79      kristaps 2016:
1.238     schwarze 2017:        if (r->last == NULL) {
1.350     schwarze 2018:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.277     schwarze 2019:                return ROFF_IGN;
1.76      kristaps 2020:        }
1.67      kristaps 2021:
1.81      kristaps 2022:        switch (r->last->tok) {
1.207     schwarze 2023:        case ROFF_am:
                   2024:        case ROFF_ami:
                   2025:        case ROFF_de:
                   2026:        case ROFF_dei:
                   2027:        case ROFF_ig:
1.81      kristaps 2028:                break;
1.375     schwarze 2029:        case ROFF_am1:
                   2030:        case ROFF_de1:
                   2031:                /* Remapped in roff_block(). */
                   2032:                abort();
1.81      kristaps 2033:        default:
1.350     schwarze 2034:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.277     schwarze 2035:                return ROFF_IGN;
1.76      kristaps 2036:        }
1.67      kristaps 2037:
1.375     schwarze 2038:        roffnode_pop(r);
                   2039:        roffnode_cleanscope(r);
                   2040:
                   2041:        /*
                   2042:         * If a conditional block with braces is still open,
                   2043:         * check for "\}" block end markers.
                   2044:         */
                   2045:
                   2046:        if (r->last != NULL && r->last->endspan < 0) {
                   2047:                rr = 1;  /* If arguments follow "\}", warn about them. */
                   2048:                roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
                   2049:        }
                   2050:
1.238     schwarze 2051:        if (buf->buf[pos] != '\0')
1.350     schwarze 2052:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 2053:                    ".. %s", buf->buf + pos);
1.71      kristaps 2054:
1.277     schwarze 2055:        return ROFF_IGN;
1.67      kristaps 2056: }
                   2057:
1.364     schwarze 2058: /*
                   2059:  * Pop all nodes ending at the end of the current input line.
                   2060:  * Return the number of loops ended.
                   2061:  */
1.340     schwarze 2062: static int
1.76      kristaps 2063: roffnode_cleanscope(struct roff *r)
1.67      kristaps 2064: {
1.340     schwarze 2065:        int inloop;
1.67      kristaps 2066:
1.340     schwarze 2067:        inloop = 0;
1.375     schwarze 2068:        while (r->last != NULL && r->last->endspan > 0) {
1.173     schwarze 2069:                if (--r->last->endspan != 0)
1.76      kristaps 2070:                        break;
1.340     schwarze 2071:                inloop += roffnode_pop(r);
1.76      kristaps 2072:        }
1.340     schwarze 2073:        return inloop;
1.67      kristaps 2074: }
                   2075:
1.364     schwarze 2076: /*
1.375     schwarze 2077:  * Handle the closing "\}" of a conditional block.
1.364     schwarze 2078:  * Apart from generating warnings, this only pops nodes.
                   2079:  * Return the number of loops ended.
                   2080:  */
1.340     schwarze 2081: static int
1.195     schwarze 2082: roff_ccond(struct roff *r, int ln, int ppos)
1.74      kristaps 2083: {
1.76      kristaps 2084:        if (NULL == r->last) {
1.350     schwarze 2085:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2086:                return 0;
1.76      kristaps 2087:        }
                   2088:
1.82      kristaps 2089:        switch (r->last->tok) {
1.207     schwarze 2090:        case ROFF_el:
                   2091:        case ROFF_ie:
                   2092:        case ROFF_if:
1.340     schwarze 2093:        case ROFF_while:
1.82      kristaps 2094:                break;
                   2095:        default:
1.350     schwarze 2096:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2097:                return 0;
1.75      kristaps 2098:        }
                   2099:
1.76      kristaps 2100:        if (r->last->endspan > -1) {
1.350     schwarze 2101:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.340     schwarze 2102:                return 0;
1.76      kristaps 2103:        }
                   2104:
1.340     schwarze 2105:        return roffnode_pop(r) + roffnode_cleanscope(r);
1.76      kristaps 2106: }
                   2107:
1.340     schwarze 2108: static int
1.80      kristaps 2109: roff_block(ROFF_ARGS)
1.76      kristaps 2110: {
1.315     schwarze 2111:        const char      *name, *value;
                   2112:        char            *call, *cp, *iname, *rname;
                   2113:        size_t           csz, namesz, rsz;
                   2114:        int              deftype;
1.106     kristaps 2115:
1.220     schwarze 2116:        /* Ignore groff compatibility mode for now. */
1.76      kristaps 2117:
1.238     schwarze 2118:        if (tok == ROFF_de1)
1.220     schwarze 2119:                tok = ROFF_de;
1.251     schwarze 2120:        else if (tok == ROFF_dei1)
                   2121:                tok = ROFF_dei;
1.238     schwarze 2122:        else if (tok == ROFF_am1)
1.220     schwarze 2123:                tok = ROFF_am;
1.251     schwarze 2124:        else if (tok == ROFF_ami1)
                   2125:                tok = ROFF_ami;
1.220     schwarze 2126:
                   2127:        /* Parse the macro name argument. */
                   2128:
1.238     schwarze 2129:        cp = buf->buf + pos;
                   2130:        if (tok == ROFF_ig) {
1.220     schwarze 2131:                iname = NULL;
                   2132:                namesz = 0;
                   2133:        } else {
                   2134:                iname = cp;
                   2135:                namesz = roff_getname(r, &cp, ln, ppos);
                   2136:                iname[namesz] = '\0';
                   2137:        }
1.107     kristaps 2138:
1.220     schwarze 2139:        /* Resolve the macro name argument if it is indirect. */
1.107     kristaps 2140:
1.238     schwarze 2141:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2142:                deftype = ROFFDEF_USER;
                   2143:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2144:                if (name == NULL) {
1.350     schwarze 2145:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2146:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2147:                            "%.*s", (int)namesz, iname);
                   2148:                        namesz = 0;
                   2149:                } else
                   2150:                        namesz = strlen(name);
                   2151:        } else
                   2152:                name = iname;
1.107     kristaps 2153:
1.238     schwarze 2154:        if (namesz == 0 && tok != ROFF_ig) {
1.350     schwarze 2155:                mandoc_msg(MANDOCERR_REQ_EMPTY,
                   2156:                    ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2157:                return ROFF_IGN;
1.220     schwarze 2158:        }
1.80      kristaps 2159:
1.106     kristaps 2160:        roffnode_push(r, tok, name, ln, ppos);
                   2161:
                   2162:        /*
                   2163:         * At the beginning of a `de' macro, clear the existing string
                   2164:         * with the same name, if there is one.  New content will be
1.193     schwarze 2165:         * appended from roff_block_text() in multiline mode.
1.106     kristaps 2166:         */
1.107     kristaps 2167:
1.311     schwarze 2168:        if (tok == ROFF_de || tok == ROFF_dei) {
1.213     schwarze 2169:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311     schwarze 2170:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315     schwarze 2171:        } else if (tok == ROFF_am || tok == ROFF_ami) {
                   2172:                deftype = ROFFDEF_ANY;
                   2173:                value = roff_getstrn(r, iname, namesz, &deftype);
                   2174:                switch (deftype) {  /* Before appending, ... */
                   2175:                case ROFFDEF_PRE: /* copy predefined to user-defined. */
                   2176:                        roff_setstrn(&r->strtab, name, namesz,
                   2177:                            value, strlen(value), 0);
                   2178:                        break;
                   2179:                case ROFFDEF_REN: /* call original standard macro. */
                   2180:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2181:                            (int)strlen(value), value);
                   2182:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2183:                        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
                   2184:                        free(call);
                   2185:                        break;
                   2186:                case ROFFDEF_STD:  /* rename and call standard macro. */
                   2187:                        rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
                   2188:                        roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
                   2189:                        csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
                   2190:                            (int)rsz, rname);
                   2191:                        roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
                   2192:                        free(call);
                   2193:                        free(rname);
                   2194:                        break;
                   2195:                default:
                   2196:                        break;
                   2197:                }
1.311     schwarze 2198:        }
1.76      kristaps 2199:
1.238     schwarze 2200:        if (*cp == '\0')
1.277     schwarze 2201:                return ROFF_IGN;
1.78      kristaps 2202:
1.220     schwarze 2203:        /* Get the custom end marker. */
1.107     kristaps 2204:
1.220     schwarze 2205:        iname = cp;
1.213     schwarze 2206:        namesz = roff_getname(r, &cp, ln, ppos);
1.220     schwarze 2207:
                   2208:        /* Resolve the end marker if it is indirect. */
                   2209:
1.238     schwarze 2210:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315     schwarze 2211:                deftype = ROFFDEF_USER;
                   2212:                name = roff_getstrn(r, iname, namesz, &deftype);
                   2213:                if (name == NULL) {
1.350     schwarze 2214:                        mandoc_msg(MANDOCERR_STR_UNDEF,
                   2215:                            ln, (int)(iname - buf->buf),
1.220     schwarze 2216:                            "%.*s", (int)namesz, iname);
                   2217:                        namesz = 0;
                   2218:                } else
                   2219:                        namesz = strlen(name);
                   2220:        } else
                   2221:                name = iname;
                   2222:
1.213     schwarze 2223:        if (namesz)
                   2224:                r->last->end = mandoc_strndup(name, namesz);
1.78      kristaps 2225:
1.238     schwarze 2226:        if (*cp != '\0')
1.350     schwarze 2227:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.295     schwarze 2228:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74      kristaps 2229:
1.277     schwarze 2230:        return ROFF_IGN;
1.78      kristaps 2231: }
                   2232:
1.340     schwarze 2233: static int
1.80      kristaps 2234: roff_block_sub(ROFF_ARGS)
1.79      kristaps 2235: {
1.294     schwarze 2236:        enum roff_tok   t;
1.79      kristaps 2237:        int             i, j;
                   2238:
                   2239:        /*
1.385     schwarze 2240:         * If a custom end marker is a user-defined or predefined macro
                   2241:         * or a request, interpret it.
1.79      kristaps 2242:         */
                   2243:
                   2244:        if (r->last->end) {
1.130     kristaps 2245:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238     schwarze 2246:                        if (buf->buf[i] != r->last->end[j])
1.79      kristaps 2247:                                break;
                   2248:
1.238     schwarze 2249:                if (r->last->end[j] == '\0' &&
                   2250:                    (buf->buf[i] == '\0' ||
                   2251:                     buf->buf[i] == ' ' ||
                   2252:                     buf->buf[i] == '\t')) {
1.79      kristaps 2253:                        roffnode_pop(r);
                   2254:                        roffnode_cleanscope(r);
                   2255:
1.238     schwarze 2256:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130     kristaps 2257:                                i++;
                   2258:
                   2259:                        pos = i;
1.238     schwarze 2260:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294     schwarze 2261:                            TOKEN_NONE)
1.277     schwarze 2262:                                return ROFF_RERUN;
                   2263:                        return ROFF_IGN;
1.79      kristaps 2264:                }
                   2265:        }
                   2266:
1.385     schwarze 2267:        /* Handle the standard end marker. */
1.79      kristaps 2268:
1.238     schwarze 2269:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.385     schwarze 2270:        if (t == ROFF_cblock)
                   2271:                return roff_cblock(r, t, buf, ln, ppos, pos, offs);
1.79      kristaps 2272:
1.385     schwarze 2273:        /* Not an end marker, so append the line to the block. */
1.79      kristaps 2274:
1.385     schwarze 2275:        if (tok != ROFF_ig)
                   2276:                roff_setstr(r, r->last->name, buf->buf + ppos, 2);
                   2277:        return ROFF_IGN;
1.79      kristaps 2278: }
                   2279:
1.340     schwarze 2280: static int
1.80      kristaps 2281: roff_block_text(ROFF_ARGS)
1.78      kristaps 2282: {
                   2283:
1.238     schwarze 2284:        if (tok != ROFF_ig)
                   2285:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106     kristaps 2286:
1.277     schwarze 2287:        return ROFF_IGN;
1.78      kristaps 2288: }
                   2289:
1.375     schwarze 2290: /*
                   2291:  * Check for a closing "\}" and handle it.
                   2292:  * In this function, the final "int *offs" argument is used for
                   2293:  * different purposes than elsewhere:
                   2294:  * Input: *offs == 0: caller wants to discard arguments following \}
                   2295:  *        *offs == 1: caller wants to preserve text following \}
                   2296:  * Output: *offs = 0: tell caller to discard input line
                   2297:  *         *offs = 1: tell caller to use input line
                   2298:  */
1.340     schwarze 2299: static int
1.375     schwarze 2300: roff_cond_checkend(ROFF_ARGS)
1.82      kristaps 2301: {
1.340     schwarze 2302:        char            *ep;
                   2303:        int              endloop, irc, rr;
1.82      kristaps 2304:
1.340     schwarze 2305:        irc = ROFF_IGN;
1.82      kristaps 2306:        rr = r->last->rule;
1.340     schwarze 2307:        endloop = tok != ROFF_while ? ROFF_IGN :
                   2308:            rr ? ROFF_LOOPCONT : ROFF_LOOPEXIT;
                   2309:        if (roffnode_cleanscope(r))
                   2310:                irc |= endloop;
1.144     kristaps 2311:
1.196     schwarze 2312:        /*
1.375     schwarze 2313:         * If "\}" occurs on a macro line without a preceding macro or
                   2314:         * a text line contains nothing else, drop the line completely.
1.196     schwarze 2315:         */
                   2316:
1.238     schwarze 2317:        ep = buf->buf + pos;
1.375     schwarze 2318:        if (ep[0] == '\\' && ep[1] == '}' && (ep[2] == '\0' || *offs == 0))
1.198     schwarze 2319:                rr = 0;
1.196     schwarze 2320:
1.343     schwarze 2321:        /*
1.375     schwarze 2322:         * The closing delimiter "\}" rewinds the conditional scope
1.343     schwarze 2323:         * but is otherwise ignored when interpreting the line.
                   2324:         */
1.144     kristaps 2325:
1.238     schwarze 2326:        while ((ep = strchr(ep, '\\')) != NULL) {
1.318     schwarze 2327:                switch (ep[1]) {
                   2328:                case '}':
1.376     schwarze 2329:                        if (ep[2] == '\0')
                   2330:                                ep[0] = '\0';
                   2331:                        else if (rr)
1.375     schwarze 2332:                                ep[1] = '&';
                   2333:                        else
                   2334:                                memmove(ep, ep + 2, strlen(ep + 2) + 1);
1.340     schwarze 2335:                        if (roff_ccond(r, ln, ep - buf->buf))
                   2336:                                irc |= endloop;
1.318     schwarze 2337:                        break;
                   2338:                case '\0':
                   2339:                        ++ep;
                   2340:                        break;
                   2341:                default:
                   2342:                        ep += 2;
                   2343:                        break;
1.197     schwarze 2344:                }
1.177     schwarze 2345:        }
1.375     schwarze 2346:        *offs = rr;
                   2347:        return irc;
                   2348: }
1.318     schwarze 2349:
1.375     schwarze 2350: /*
                   2351:  * Parse and process a request or macro line in conditional scope.
                   2352:  */
                   2353: static int
                   2354: roff_cond_sub(ROFF_ARGS)
                   2355: {
                   2356:        struct roffnode *bl;
1.386     schwarze 2357:        int              irc, rr, spos;
1.375     schwarze 2358:        enum roff_tok    t;
                   2359:
                   2360:        rr = 0;  /* If arguments follow "\}", skip them. */
                   2361:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.386     schwarze 2362:        spos = pos;
1.368     schwarze 2363:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   2364:
1.318     schwarze 2365:        /*
1.386     schwarze 2366:         * Handle requests and macros if the conditional evaluated
                   2367:         * to true or if they are structurally required.
                   2368:         * The .break request is always handled specially.
1.318     schwarze 2369:         */
                   2370:
1.364     schwarze 2371:        if (t == ROFF_break) {
                   2372:                if (irc & ROFF_LOOPMASK)
                   2373:                        irc = ROFF_IGN | ROFF_LOOPEXIT;
                   2374:                else if (rr) {
                   2375:                        for (bl = r->last; bl != NULL; bl = bl->parent) {
                   2376:                                bl->rule = 0;
                   2377:                                if (bl->tok == ROFF_while)
                   2378:                                        break;
                   2379:                        }
                   2380:                }
1.386     schwarze 2381:        } else if (rr || (t < TOKEN_NONE && roffs[t].flags & ROFFMAC_STRUCT)) {
                   2382:                irc |= roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
1.383     schwarze 2383:                if (irc & ROFF_WHILE)
                   2384:                        irc &= ~(ROFF_LOOPCONT | ROFF_LOOPEXIT);
1.386     schwarze 2385:        }
1.340     schwarze 2386:        return irc;
1.82      kristaps 2387: }
                   2388:
1.375     schwarze 2389: /*
                   2390:  * Parse and process a text line in conditional scope.
                   2391:  */
1.340     schwarze 2392: static int
1.82      kristaps 2393: roff_cond_text(ROFF_ARGS)
1.78      kristaps 2394: {
1.375     schwarze 2395:        int      irc, rr;
1.343     schwarze 2396:
1.375     schwarze 2397:        rr = 1;  /* If arguments follow "\}", preserve them. */
                   2398:        irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.340     schwarze 2399:        if (rr)
                   2400:                irc |= ROFF_CONT;
                   2401:        return irc;
1.74      kristaps 2402: }
                   2403:
1.266     schwarze 2404: /* --- handling of numeric and conditional expressions -------------------- */
                   2405:
1.204     schwarze 2406: /*
                   2407:  * Parse a single signed integer number.  Stop at the first non-digit.
                   2408:  * If there is at least one digit, return success and advance the
                   2409:  * parse point, else return failure and let the parse point unchanged.
                   2410:  * Ignore overflows, treat them just like the C language.
                   2411:  */
1.184     schwarze 2412: static int
1.261     schwarze 2413: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184     schwarze 2414: {
1.261     schwarze 2415:        int      myres, scaled, n, p;
1.206     schwarze 2416:
                   2417:        if (NULL == res)
                   2418:                res = &myres;
1.184     schwarze 2419:
                   2420:        p = *pos;
                   2421:        n = v[p] == '-';
1.261     schwarze 2422:        if (n || v[p] == '+')
1.184     schwarze 2423:                p++;
                   2424:
1.261     schwarze 2425:        if (flags & ROFFNUM_WHITE)
                   2426:                while (isspace((unsigned char)v[p]))
                   2427:                        p++;
                   2428:
1.184     schwarze 2429:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204     schwarze 2430:                *res = 10 * *res + v[p] - '0';
1.184     schwarze 2431:        if (p == *pos + n)
                   2432:                return 0;
                   2433:
                   2434:        if (n)
                   2435:                *res = -*res;
                   2436:
1.254     schwarze 2437:        /* Each number may be followed by one optional scaling unit. */
                   2438:
                   2439:        switch (v[p]) {
                   2440:        case 'f':
1.261     schwarze 2441:                scaled = *res * 65536;
1.254     schwarze 2442:                break;
                   2443:        case 'i':
1.261     schwarze 2444:                scaled = *res * 240;
1.254     schwarze 2445:                break;
                   2446:        case 'c':
1.261     schwarze 2447:                scaled = *res * 240 / 2.54;
1.254     schwarze 2448:                break;
                   2449:        case 'v':
                   2450:        case 'P':
1.261     schwarze 2451:                scaled = *res * 40;
1.254     schwarze 2452:                break;
                   2453:        case 'm':
                   2454:        case 'n':
1.261     schwarze 2455:                scaled = *res * 24;
1.254     schwarze 2456:                break;
                   2457:        case 'p':
1.261     schwarze 2458:                scaled = *res * 10 / 3;
1.254     schwarze 2459:                break;
                   2460:        case 'u':
1.261     schwarze 2461:                scaled = *res;
1.254     schwarze 2462:                break;
                   2463:        case 'M':
1.261     schwarze 2464:                scaled = *res * 6 / 25;
1.254     schwarze 2465:                break;
                   2466:        default:
1.261     schwarze 2467:                scaled = *res;
1.254     schwarze 2468:                p--;
                   2469:                break;
                   2470:        }
1.261     schwarze 2471:        if (flags & ROFFNUM_SCALE)
                   2472:                *res = scaled;
1.254     schwarze 2473:
                   2474:        *pos = p + 1;
1.277     schwarze 2475:        return 1;
1.184     schwarze 2476: }
                   2477:
1.198     schwarze 2478: /*
                   2479:  * Evaluate a string comparison condition.
                   2480:  * The first character is the delimiter.
                   2481:  * Succeed if the string up to its second occurrence
                   2482:  * matches the string up to its third occurence.
                   2483:  * Advance the cursor after the third occurrence
                   2484:  * or lacking that, to the end of the line.
                   2485:  */
                   2486: static int
                   2487: roff_evalstrcond(const char *v, int *pos)
                   2488: {
                   2489:        const char      *s1, *s2, *s3;
                   2490:        int              match;
                   2491:
                   2492:        match = 0;
                   2493:        s1 = v + *pos;          /* initial delimiter */
                   2494:        s2 = s1 + 1;            /* for scanning the first string */
                   2495:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   2496:
                   2497:        if (NULL == s3)         /* found no middle delimiter */
                   2498:                goto out;
                   2499:
                   2500:        while ('\0' != *++s3) {
                   2501:                if (*s2 != *s3) {  /* mismatch */
                   2502:                        s3 = strchr(s3, *s1);
                   2503:                        break;
                   2504:                }
                   2505:                if (*s3 == *s1) {  /* found the final delimiter */
                   2506:                        match = 1;
                   2507:                        break;
                   2508:                }
                   2509:                s2++;
                   2510:        }
                   2511:
                   2512: out:
                   2513:        if (NULL == s3)
                   2514:                s3 = strchr(s2, '\0');
1.242     schwarze 2515:        else if (*s3 != '\0')
1.198     schwarze 2516:                s3++;
                   2517:        *pos = s3 - v;
1.277     schwarze 2518:        return match;
1.198     schwarze 2519: }
                   2520:
1.204     schwarze 2521: /*
                   2522:  * Evaluate an optionally negated single character, numerical,
                   2523:  * or string condition.
                   2524:  */
1.198     schwarze 2525: static int
1.271     schwarze 2526: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88      kristaps 2527: {
1.336     schwarze 2528:        const char      *start, *end;
                   2529:        char            *cp, *name;
                   2530:        size_t           sz;
                   2531:        int              deftype, len, number, savepos, istrue, wanttrue;
1.88      kristaps 2532:
1.198     schwarze 2533:        if ('!' == v[*pos]) {
                   2534:                wanttrue = 0;
                   2535:                (*pos)++;
                   2536:        } else
                   2537:                wanttrue = 1;
                   2538:
1.88      kristaps 2539:        switch (v[*pos]) {
1.240     schwarze 2540:        case '\0':
1.277     schwarze 2541:                return 0;
1.207     schwarze 2542:        case 'n':
                   2543:        case 'o':
1.88      kristaps 2544:                (*pos)++;
1.277     schwarze 2545:                return wanttrue;
1.207     schwarze 2546:        case 'e':
                   2547:        case 't':
1.239     schwarze 2548:        case 'v':
1.88      kristaps 2549:                (*pos)++;
1.277     schwarze 2550:                return !wanttrue;
1.336     schwarze 2551:        case 'c':
                   2552:                do {
                   2553:                        (*pos)++;
                   2554:                } while (v[*pos] == ' ');
                   2555:
                   2556:                /*
                   2557:                 * Quirk for groff compatibility:
                   2558:                 * The horizontal tab is neither available nor unavailable.
                   2559:                 */
                   2560:
                   2561:                if (v[*pos] == '\t') {
                   2562:                        (*pos)++;
                   2563:                        return 0;
                   2564:                }
                   2565:
                   2566:                /* Printable ASCII characters are available. */
                   2567:
                   2568:                if (v[*pos] != '\\') {
                   2569:                        (*pos)++;
                   2570:                        return wanttrue;
                   2571:                }
                   2572:
                   2573:                end = v + ++*pos;
                   2574:                switch (mandoc_escape(&end, &start, &len)) {
                   2575:                case ESCAPE_SPECIAL:
                   2576:                        istrue = mchars_spec2cp(start, len) != -1;
                   2577:                        break;
                   2578:                case ESCAPE_UNICODE:
                   2579:                        istrue = 1;
                   2580:                        break;
                   2581:                case ESCAPE_NUMBERED:
                   2582:                        istrue = mchars_num2char(start, len) != -1;
                   2583:                        break;
                   2584:                default:
                   2585:                        istrue = !wanttrue;
                   2586:                        break;
                   2587:                }
                   2588:                *pos = end - v;
                   2589:                return istrue == wanttrue;
1.310     schwarze 2590:        case 'd':
1.271     schwarze 2591:        case 'r':
1.310     schwarze 2592:                cp = v + *pos + 1;
                   2593:                while (*cp == ' ')
                   2594:                        cp++;
                   2595:                name = cp;
                   2596:                sz = roff_getname(r, &cp, ln, cp - v);
1.315     schwarze 2597:                if (sz == 0)
                   2598:                        istrue = 0;
                   2599:                else if (v[*pos] == 'r')
                   2600:                        istrue = roff_hasregn(r, name, sz);
                   2601:                else {
                   2602:                        deftype = ROFFDEF_ANY;
                   2603:                        roff_getstrn(r, name, sz, &deftype);
                   2604:                        istrue = !!deftype;
                   2605:                }
1.363     schwarze 2606:                *pos = (name + sz) - v;
1.310     schwarze 2607:                return istrue == wanttrue;
1.88      kristaps 2608:        default:
                   2609:                break;
                   2610:        }
                   2611:
1.241     schwarze 2612:        savepos = *pos;
1.261     schwarze 2613:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277     schwarze 2614:                return (number > 0) == wanttrue;
1.241     schwarze 2615:        else if (*pos == savepos)
1.277     schwarze 2616:                return roff_evalstrcond(v, pos) == wanttrue;
1.204     schwarze 2617:        else
1.277     schwarze 2618:                return 0;
1.88      kristaps 2619: }
                   2620:
1.340     schwarze 2621: static int
1.103     kristaps 2622: roff_line_ignore(ROFF_ARGS)
1.89      kristaps 2623: {
1.123     schwarze 2624:
1.277     schwarze 2625:        return ROFF_IGN;
1.89      kristaps 2626: }
                   2627:
1.340     schwarze 2628: static int
1.251     schwarze 2629: roff_insec(ROFF_ARGS)
                   2630: {
                   2631:
1.350     schwarze 2632:        mandoc_msg(MANDOCERR_REQ_INSEC, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2633:        return ROFF_IGN;
1.251     schwarze 2634: }
                   2635:
1.340     schwarze 2636: static int
1.251     schwarze 2637: roff_unsupp(ROFF_ARGS)
                   2638: {
                   2639:
1.350     schwarze 2640:        mandoc_msg(MANDOCERR_REQ_UNSUPP, ln, ppos, "%s", roff_name[tok]);
1.277     schwarze 2641:        return ROFF_IGN;
1.251     schwarze 2642: }
                   2643:
1.340     schwarze 2644: static int
1.82      kristaps 2645: roff_cond(ROFF_ARGS)
1.74      kristaps 2646: {
1.340     schwarze 2647:        int      irc;
1.173     schwarze 2648:
                   2649:        roffnode_push(r, tok, NULL, ln, ppos);
1.74      kristaps 2650:
1.207     schwarze 2651:        /*
1.134     kristaps 2652:         * An `.el' has no conditional body: it will consume the value
                   2653:         * of the current rstack entry set in prior `ie' calls or
1.207     schwarze 2654:         * defaults to DENY.
1.134     kristaps 2655:         *
                   2656:         * If we're not an `el', however, then evaluate the conditional.
                   2657:         */
1.133     kristaps 2658:
1.238     schwarze 2659:        r->last->rule = tok == ROFF_el ?
1.207     schwarze 2660:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238     schwarze 2661:            roff_evalcond(r, ln, buf->buf, &pos);
1.77      kristaps 2662:
1.134     kristaps 2663:        /*
                   2664:         * An if-else will put the NEGATION of the current evaluated
                   2665:         * conditional into the stack of rules.
                   2666:         */
                   2667:
1.238     schwarze 2668:        if (tok == ROFF_ie) {
1.223     schwarze 2669:                if (r->rstackpos + 1 == r->rstacksz) {
                   2670:                        r->rstacksz += 16;
                   2671:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2672:                            r->rstacksz, sizeof(int));
1.134     kristaps 2673:                }
1.198     schwarze 2674:                r->rstack[++r->rstackpos] = !r->last->rule;
1.82      kristaps 2675:        }
1.88      kristaps 2676:
                   2677:        /* If the parent has false as its rule, then so do we. */
                   2678:
1.198     schwarze 2679:        if (r->last->parent && !r->last->parent->rule)
                   2680:                r->last->rule = 0;
1.88      kristaps 2681:
                   2682:        /*
1.173     schwarze 2683:         * Determine scope.
                   2684:         * If there is nothing on the line after the conditional,
                   2685:         * not even whitespace, use next-line scope.
1.340     schwarze 2686:         * Except that .while does not support next-line scope.
1.88      kristaps 2687:         */
1.74      kristaps 2688:
1.340     schwarze 2689:        if (buf->buf[pos] == '\0' && tok != ROFF_while) {
1.173     schwarze 2690:                r->last->endspan = 2;
                   2691:                goto out;
                   2692:        }
                   2693:
1.238     schwarze 2694:        while (buf->buf[pos] == ' ')
1.173     schwarze 2695:                pos++;
                   2696:
                   2697:        /* An opening brace requests multiline scope. */
1.75      kristaps 2698:
1.238     schwarze 2699:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75      kristaps 2700:                r->last->endspan = -1;
                   2701:                pos += 2;
1.272     schwarze 2702:                while (buf->buf[pos] == ' ')
                   2703:                        pos++;
1.173     schwarze 2704:                goto out;
1.207     schwarze 2705:        }
1.74      kristaps 2706:
1.77      kristaps 2707:        /*
1.173     schwarze 2708:         * Anything else following the conditional causes
                   2709:         * single-line scope.  Warn if the scope contains
                   2710:         * nothing but trailing whitespace.
1.77      kristaps 2711:         */
                   2712:
1.238     schwarze 2713:        if (buf->buf[pos] == '\0')
1.350     schwarze 2714:                mandoc_msg(MANDOCERR_COND_EMPTY,
                   2715:                    ln, ppos, "%s", roff_name[tok]);
1.77      kristaps 2716:
1.173     schwarze 2717:        r->last->endspan = 1;
1.74      kristaps 2718:
1.173     schwarze 2719: out:
1.75      kristaps 2720:        *offs = pos;
1.340     schwarze 2721:        irc = ROFF_RERUN;
                   2722:        if (tok == ROFF_while)
                   2723:                irc |= ROFF_WHILE;
                   2724:        return irc;
1.83      schwarze 2725: }
                   2726:
1.340     schwarze 2727: static int
1.92      schwarze 2728: roff_ds(ROFF_ARGS)
                   2729: {
1.212     schwarze 2730:        char            *string;
                   2731:        const char      *name;
                   2732:        size_t           namesz;
1.96      kristaps 2733:
1.251     schwarze 2734:        /* Ignore groff compatibility mode for now. */
                   2735:
                   2736:        if (tok == ROFF_ds1)
                   2737:                tok = ROFF_ds;
                   2738:        else if (tok == ROFF_as1)
                   2739:                tok = ROFF_as;
                   2740:
1.96      kristaps 2741:        /*
1.212     schwarze 2742:         * The first word is the name of the string.
                   2743:         * If it is empty or terminated by an escape sequence,
                   2744:         * abort the `ds' request without defining anything.
1.96      kristaps 2745:         */
1.92      schwarze 2746:
1.238     schwarze 2747:        name = string = buf->buf + pos;
                   2748:        if (*name == '\0')
1.277     schwarze 2749:                return ROFF_IGN;
1.92      schwarze 2750:
1.212     schwarze 2751:        namesz = roff_getname(r, &string, ln, pos);
1.363     schwarze 2752:        switch (name[namesz]) {
                   2753:        case '\\':
1.277     schwarze 2754:                return ROFF_IGN;
1.363     schwarze 2755:        case '\t':
                   2756:                string = buf->buf + pos + namesz;
                   2757:                break;
                   2758:        default:
                   2759:                break;
                   2760:        }
1.212     schwarze 2761:
                   2762:        /* Read past the initial double-quote, if any. */
1.238     schwarze 2763:        if (*string == '"')
1.92      schwarze 2764:                string++;
                   2765:
1.96      kristaps 2766:        /* The rest is the value. */
1.212     schwarze 2767:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2768:            ROFF_as == tok);
1.311     schwarze 2769:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277     schwarze 2770:        return ROFF_IGN;
1.92      schwarze 2771: }
                   2772:
1.204     schwarze 2773: /*
                   2774:  * Parse a single operator, one or two characters long.
                   2775:  * If the operator is recognized, return success and advance the
                   2776:  * parse point, else return failure and let the parse point unchanged.
                   2777:  */
                   2778: static int
                   2779: roff_getop(const char *v, int *pos, char *res)
                   2780: {
                   2781:
                   2782:        *res = v[*pos];
                   2783:
                   2784:        switch (*res) {
1.207     schwarze 2785:        case '+':
                   2786:        case '-':
                   2787:        case '*':
                   2788:        case '/':
                   2789:        case '%':
                   2790:        case '&':
                   2791:        case ':':
1.204     schwarze 2792:                break;
                   2793:        case '<':
                   2794:                switch (v[*pos + 1]) {
1.207     schwarze 2795:                case '=':
1.204     schwarze 2796:                        *res = 'l';
                   2797:                        (*pos)++;
                   2798:                        break;
1.207     schwarze 2799:                case '>':
1.204     schwarze 2800:                        *res = '!';
                   2801:                        (*pos)++;
                   2802:                        break;
1.207     schwarze 2803:                case '?':
1.204     schwarze 2804:                        *res = 'i';
                   2805:                        (*pos)++;
                   2806:                        break;
                   2807:                default:
                   2808:                        break;
                   2809:                }
                   2810:                break;
                   2811:        case '>':
                   2812:                switch (v[*pos + 1]) {
1.207     schwarze 2813:                case '=':
1.204     schwarze 2814:                        *res = 'g';
                   2815:                        (*pos)++;
                   2816:                        break;
1.207     schwarze 2817:                case '?':
1.204     schwarze 2818:                        *res = 'a';
                   2819:                        (*pos)++;
                   2820:                        break;
                   2821:                default:
                   2822:                        break;
                   2823:                }
                   2824:                break;
                   2825:        case '=':
                   2826:                if ('=' == v[*pos + 1])
                   2827:                        (*pos)++;
                   2828:                break;
                   2829:        default:
1.277     schwarze 2830:                return 0;
1.204     schwarze 2831:        }
                   2832:        (*pos)++;
                   2833:
1.277     schwarze 2834:        return *res;
1.204     schwarze 2835: }
                   2836:
                   2837: /*
                   2838:  * Evaluate either a parenthesized numeric expression
                   2839:  * or a single signed integer number.
                   2840:  */
                   2841: static int
1.235     schwarze 2842: roff_evalpar(struct roff *r, int ln,
1.261     schwarze 2843:        const char *v, int *pos, int *res, int flags)
1.204     schwarze 2844: {
                   2845:
                   2846:        if ('(' != v[*pos])
1.277     schwarze 2847:                return roff_getnum(v, pos, res, flags);
1.204     schwarze 2848:
                   2849:        (*pos)++;
1.261     schwarze 2850:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277     schwarze 2851:                return 0;
1.204     schwarze 2852:
1.206     schwarze 2853:        /*
                   2854:         * Omission of the closing parenthesis
                   2855:         * is an error in validation mode,
                   2856:         * but ignored in evaluation mode.
                   2857:         */
                   2858:
1.204     schwarze 2859:        if (')' == v[*pos])
                   2860:                (*pos)++;
1.206     schwarze 2861:        else if (NULL == res)
1.277     schwarze 2862:                return 0;
1.204     schwarze 2863:
1.277     schwarze 2864:        return 1;
1.204     schwarze 2865: }
                   2866:
                   2867: /*
                   2868:  * Evaluate a complete numeric expression.
                   2869:  * Proceed left to right, there is no concept of precedence.
                   2870:  */
                   2871: static int
1.235     schwarze 2872: roff_evalnum(struct roff *r, int ln, const char *v,
1.261     schwarze 2873:        int *pos, int *res, int flags)
1.204     schwarze 2874: {
                   2875:        int              mypos, operand2;
                   2876:        char             operator;
                   2877:
                   2878:        if (NULL == pos) {
                   2879:                mypos = 0;
                   2880:                pos = &mypos;
                   2881:        }
                   2882:
1.261     schwarze 2883:        if (flags & ROFFNUM_WHITE)
1.204     schwarze 2884:                while (isspace((unsigned char)v[*pos]))
                   2885:                        (*pos)++;
                   2886:
1.261     schwarze 2887:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277     schwarze 2888:                return 0;
1.204     schwarze 2889:
                   2890:        while (1) {
1.261     schwarze 2891:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2892:                        while (isspace((unsigned char)v[*pos]))
                   2893:                                (*pos)++;
                   2894:
                   2895:                if ( ! roff_getop(v, pos, &operator))
                   2896:                        break;
                   2897:
1.261     schwarze 2898:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2899:                        while (isspace((unsigned char)v[*pos]))
                   2900:                                (*pos)++;
                   2901:
1.261     schwarze 2902:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277     schwarze 2903:                        return 0;
1.204     schwarze 2904:
1.261     schwarze 2905:                if (flags & ROFFNUM_WHITE)
1.204     schwarze 2906:                        while (isspace((unsigned char)v[*pos]))
                   2907:                                (*pos)++;
1.206     schwarze 2908:
                   2909:                if (NULL == res)
                   2910:                        continue;
1.204     schwarze 2911:
                   2912:                switch (operator) {
1.207     schwarze 2913:                case '+':
1.204     schwarze 2914:                        *res += operand2;
                   2915:                        break;
1.207     schwarze 2916:                case '-':
1.204     schwarze 2917:                        *res -= operand2;
                   2918:                        break;
1.207     schwarze 2919:                case '*':
1.204     schwarze 2920:                        *res *= operand2;
                   2921:                        break;
1.207     schwarze 2922:                case '/':
1.244     schwarze 2923:                        if (operand2 == 0) {
1.235     schwarze 2924:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2925:                                        ln, *pos, "%s", v);
1.234     kristaps 2926:                                *res = 0;
                   2927:                                break;
                   2928:                        }
1.204     schwarze 2929:                        *res /= operand2;
                   2930:                        break;
1.207     schwarze 2931:                case '%':
1.244     schwarze 2932:                        if (operand2 == 0) {
                   2933:                                mandoc_msg(MANDOCERR_DIVZERO,
1.350     schwarze 2934:                                        ln, *pos, "%s", v);
1.244     schwarze 2935:                                *res = 0;
                   2936:                                break;
                   2937:                        }
1.204     schwarze 2938:                        *res %= operand2;
                   2939:                        break;
1.207     schwarze 2940:                case '<':
1.204     schwarze 2941:                        *res = *res < operand2;
                   2942:                        break;
1.207     schwarze 2943:                case '>':
1.204     schwarze 2944:                        *res = *res > operand2;
                   2945:                        break;
1.207     schwarze 2946:                case 'l':
1.204     schwarze 2947:                        *res = *res <= operand2;
                   2948:                        break;
1.207     schwarze 2949:                case 'g':
1.204     schwarze 2950:                        *res = *res >= operand2;
                   2951:                        break;
1.207     schwarze 2952:                case '=':
1.204     schwarze 2953:                        *res = *res == operand2;
                   2954:                        break;
1.207     schwarze 2955:                case '!':
1.204     schwarze 2956:                        *res = *res != operand2;
                   2957:                        break;
1.207     schwarze 2958:                case '&':
1.204     schwarze 2959:                        *res = *res && operand2;
                   2960:                        break;
1.207     schwarze 2961:                case ':':
1.204     schwarze 2962:                        *res = *res || operand2;
                   2963:                        break;
1.207     schwarze 2964:                case 'i':
1.204     schwarze 2965:                        if (operand2 < *res)
                   2966:                                *res = operand2;
                   2967:                        break;
1.207     schwarze 2968:                case 'a':
1.204     schwarze 2969:                        if (operand2 > *res)
                   2970:                                *res = operand2;
                   2971:                        break;
                   2972:                default:
                   2973:                        abort();
                   2974:                }
                   2975:        }
1.277     schwarze 2976:        return 1;
1.204     schwarze 2977: }
                   2978:
1.266     schwarze 2979: /* --- register management ------------------------------------------------ */
                   2980:
1.180     schwarze 2981: void
1.187     schwarze 2982: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147     kristaps 2983: {
1.327     schwarze 2984:        roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.326     schwarze 2985: }
                   2986:
                   2987: static void
                   2988: roff_setregn(struct roff *r, const char *name, size_t len,
1.327     schwarze 2989:     int val, char sign, int step)
1.326     schwarze 2990: {
1.180     schwarze 2991:        struct roffreg  *reg;
                   2992:
                   2993:        /* Search for an existing register with the same name. */
                   2994:        reg = r->regtab;
                   2995:
1.326     schwarze 2996:        while (reg != NULL && (reg->key.sz != len ||
                   2997:            strncmp(reg->key.p, name, len) != 0))
1.180     schwarze 2998:                reg = reg->next;
1.147     kristaps 2999:
1.180     schwarze 3000:        if (NULL == reg) {
                   3001:                /* Create a new register. */
                   3002:                reg = mandoc_malloc(sizeof(struct roffreg));
1.326     schwarze 3003:                reg->key.p = mandoc_strndup(name, len);
                   3004:                reg->key.sz = len;
1.187     schwarze 3005:                reg->val = 0;
1.327     schwarze 3006:                reg->step = 0;
1.180     schwarze 3007:                reg->next = r->regtab;
                   3008:                r->regtab = reg;
                   3009:        }
                   3010:
1.187     schwarze 3011:        if ('+' == sign)
                   3012:                reg->val += val;
                   3013:        else if ('-' == sign)
                   3014:                reg->val -= val;
                   3015:        else
                   3016:                reg->val = val;
1.327     schwarze 3017:        if (step != INT_MIN)
                   3018:                reg->step = step;
1.147     kristaps 3019: }
                   3020:
1.192     schwarze 3021: /*
                   3022:  * Handle some predefined read-only number registers.
                   3023:  * For now, return -1 if the requested register is not predefined;
                   3024:  * in case a predefined read-only register having the value -1
                   3025:  * were to turn up, another special value would have to be chosen.
                   3026:  */
                   3027: static int
1.273     schwarze 3028: roff_getregro(const struct roff *r, const char *name)
1.192     schwarze 3029: {
                   3030:
                   3031:        switch (*name) {
1.273     schwarze 3032:        case '$':  /* Number of arguments of the last macro evaluated. */
1.339     schwarze 3033:                return r->mstackpos < 0 ? 0 : r->mstack[r->mstackpos].argc;
1.207     schwarze 3034:        case 'A':  /* ASCII approximation mode is always off. */
1.277     schwarze 3035:                return 0;
1.207     schwarze 3036:        case 'g':  /* Groff compatibility mode is always on. */
1.277     schwarze 3037:                return 1;
1.207     schwarze 3038:        case 'H':  /* Fixed horizontal resolution. */
1.277     schwarze 3039:                return 24;
1.207     schwarze 3040:        case 'j':  /* Always adjust left margin only. */
1.277     schwarze 3041:                return 0;
1.207     schwarze 3042:        case 'T':  /* Some output device is always defined. */
1.277     schwarze 3043:                return 1;
1.207     schwarze 3044:        case 'V':  /* Fixed vertical resolution. */
1.277     schwarze 3045:                return 40;
1.192     schwarze 3046:        default:
1.277     schwarze 3047:                return -1;
1.192     schwarze 3048:        }
                   3049: }
                   3050:
1.181     schwarze 3051: int
1.326     schwarze 3052: roff_getreg(struct roff *r, const char *name)
1.147     kristaps 3053: {
1.327     schwarze 3054:        return roff_getregn(r, name, strlen(name), '\0');
1.181     schwarze 3055: }
                   3056:
                   3057: static int
1.327     schwarze 3058: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.181     schwarze 3059: {
                   3060:        struct roffreg  *reg;
1.192     schwarze 3061:        int              val;
                   3062:
                   3063:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3064:                val = roff_getregro(r, name + 1);
1.192     schwarze 3065:                if (-1 != val)
1.277     schwarze 3066:                        return val;
1.192     schwarze 3067:        }
1.181     schwarze 3068:
1.327     schwarze 3069:        for (reg = r->regtab; reg; reg = reg->next) {
1.181     schwarze 3070:                if (len == reg->key.sz &&
1.327     schwarze 3071:                    0 == strncmp(name, reg->key.p, len)) {
                   3072:                        switch (sign) {
                   3073:                        case '+':
                   3074:                                reg->val += reg->step;
                   3075:                                break;
                   3076:                        case '-':
                   3077:                                reg->val -= reg->step;
                   3078:                                break;
                   3079:                        default:
                   3080:                                break;
                   3081:                        }
1.277     schwarze 3082:                        return reg->val;
1.327     schwarze 3083:                }
                   3084:        }
1.271     schwarze 3085:
1.327     schwarze 3086:        roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.277     schwarze 3087:        return 0;
1.271     schwarze 3088: }
                   3089:
                   3090: static int
                   3091: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   3092: {
                   3093:        struct roffreg  *reg;
                   3094:        int              val;
                   3095:
                   3096:        if ('.' == name[0] && 2 == len) {
1.273     schwarze 3097:                val = roff_getregro(r, name + 1);
1.271     schwarze 3098:                if (-1 != val)
1.277     schwarze 3099:                        return 1;
1.271     schwarze 3100:        }
                   3101:
                   3102:        for (reg = r->regtab; reg; reg = reg->next)
                   3103:                if (len == reg->key.sz &&
                   3104:                    0 == strncmp(name, reg->key.p, len))
1.277     schwarze 3105:                        return 1;
1.147     kristaps 3106:
1.277     schwarze 3107:        return 0;
1.147     kristaps 3108: }
                   3109:
1.180     schwarze 3110: static void
                   3111: roff_freereg(struct roffreg *reg)
1.147     kristaps 3112: {
1.180     schwarze 3113:        struct roffreg  *old_reg;
1.147     kristaps 3114:
1.180     schwarze 3115:        while (NULL != reg) {
                   3116:                free(reg->key.p);
                   3117:                old_reg = reg;
                   3118:                reg = reg->next;
                   3119:                free(old_reg);
                   3120:        }
1.147     kristaps 3121: }
1.92      schwarze 3122:
1.340     schwarze 3123: static int
1.89      kristaps 3124: roff_nr(ROFF_ARGS)
1.83      schwarze 3125: {
1.327     schwarze 3126:        char            *key, *val, *step;
1.212     schwarze 3127:        size_t           keysz;
1.327     schwarze 3128:        int              iv, is, len;
1.187     schwarze 3129:        char             sign;
1.89      kristaps 3130:
1.238     schwarze 3131:        key = val = buf->buf + pos;
                   3132:        if (*key == '\0')
1.277     schwarze 3133:                return ROFF_IGN;
1.212     schwarze 3134:
                   3135:        keysz = roff_getname(r, &val, ln, pos);
1.363     schwarze 3136:        if (key[keysz] == '\\' || key[keysz] == '\t')
1.277     schwarze 3137:                return ROFF_IGN;
1.89      kristaps 3138:
1.187     schwarze 3139:        sign = *val;
1.238     schwarze 3140:        if (sign == '+' || sign == '-')
1.187     schwarze 3141:                val++;
                   3142:
1.327     schwarze 3143:        len = 0;
                   3144:        if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
                   3145:                return ROFF_IGN;
                   3146:
                   3147:        step = val + len;
                   3148:        while (isspace((unsigned char)*step))
                   3149:                step++;
                   3150:        if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
                   3151:                is = INT_MIN;
1.109     kristaps 3152:
1.327     schwarze 3153:        roff_setregn(r, key, keysz, iv, sign, is);
1.277     schwarze 3154:        return ROFF_IGN;
1.203     schwarze 3155: }
                   3156:
1.340     schwarze 3157: static int
1.203     schwarze 3158: roff_rr(ROFF_ARGS)
                   3159: {
                   3160:        struct roffreg  *reg, **prev;
1.212     schwarze 3161:        char            *name, *cp;
                   3162:        size_t           namesz;
1.203     schwarze 3163:
1.238     schwarze 3164:        name = cp = buf->buf + pos;
                   3165:        if (*name == '\0')
1.277     schwarze 3166:                return ROFF_IGN;
1.212     schwarze 3167:        namesz = roff_getname(r, &cp, ln, pos);
                   3168:        name[namesz] = '\0';
1.203     schwarze 3169:
                   3170:        prev = &r->regtab;
                   3171:        while (1) {
                   3172:                reg = *prev;
1.238     schwarze 3173:                if (reg == NULL || !strcmp(name, reg->key.p))
1.203     schwarze 3174:                        break;
                   3175:                prev = &reg->next;
                   3176:        }
1.238     schwarze 3177:        if (reg != NULL) {
1.203     schwarze 3178:                *prev = reg->next;
                   3179:                free(reg->key.p);
                   3180:                free(reg);
                   3181:        }
1.277     schwarze 3182:        return ROFF_IGN;
1.122     schwarze 3183: }
                   3184:
1.266     schwarze 3185: /* --- handler functions for roff requests -------------------------------- */
                   3186:
1.340     schwarze 3187: static int
1.122     schwarze 3188: roff_rm(ROFF_ARGS)
                   3189: {
                   3190:        const char       *name;
                   3191:        char             *cp;
1.212     schwarze 3192:        size_t            namesz;
1.122     schwarze 3193:
1.238     schwarze 3194:        cp = buf->buf + pos;
                   3195:        while (*cp != '\0') {
1.212     schwarze 3196:                name = cp;
1.238     schwarze 3197:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212     schwarze 3198:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311     schwarze 3199:                roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.363     schwarze 3200:                if (name[namesz] == '\\' || name[namesz] == '\t')
1.212     schwarze 3201:                        break;
1.122     schwarze 3202:        }
1.277     schwarze 3203:        return ROFF_IGN;
1.178     schwarze 3204: }
                   3205:
1.340     schwarze 3206: static int
1.178     schwarze 3207: roff_it(ROFF_ARGS)
                   3208: {
                   3209:        int              iv;
                   3210:
                   3211:        /* Parse the number of lines. */
1.261     schwarze 3212:
                   3213:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.350     schwarze 3214:                mandoc_msg(MANDOCERR_IT_NONUM,
                   3215:                    ln, ppos, "%s", buf->buf + 1);
1.277     schwarze 3216:                return ROFF_IGN;
1.178     schwarze 3217:        }
                   3218:
1.262     schwarze 3219:        while (isspace((unsigned char)buf->buf[pos]))
                   3220:                pos++;
                   3221:
                   3222:        /*
                   3223:         * Arm the input line trap.
                   3224:         * Special-casing "an-trap" is an ugly workaround to cope
                   3225:         * with DocBook stupidly fiddling with man(7) internals.
                   3226:         */
1.261     schwarze 3227:
1.178     schwarze 3228:        roffit_lines = iv;
1.262     schwarze 3229:        roffit_macro = mandoc_strdup(iv != 1 ||
                   3230:            strcmp(buf->buf + pos, "an-trap") ?
                   3231:            buf->buf + pos : "br");
1.277     schwarze 3232:        return ROFF_IGN;
1.175     schwarze 3233: }
                   3234:
1.340     schwarze 3235: static int
1.175     schwarze 3236: roff_Dd(ROFF_ARGS)
                   3237: {
1.315     schwarze 3238:        int              mask;
                   3239:        enum roff_tok    t, te;
1.227     schwarze 3240:
1.315     schwarze 3241:        switch (tok) {
                   3242:        case ROFF_Dd:
                   3243:                tok = MDOC_Dd;
                   3244:                te = MDOC_MAX;
                   3245:                if (r->format == 0)
                   3246:                        r->format = MPARSE_MDOC;
                   3247:                mask = MPARSE_MDOC | MPARSE_QUICK;
                   3248:                break;
                   3249:        case ROFF_TH:
                   3250:                tok = MAN_TH;
                   3251:                te = MAN_MAX;
                   3252:                if (r->format == 0)
                   3253:                        r->format = MPARSE_MAN;
                   3254:                mask = MPARSE_QUICK;
                   3255:                break;
                   3256:        default:
                   3257:                abort();
                   3258:        }
                   3259:        if ((r->options & mask) == 0)
                   3260:                for (t = tok; t < te; t++)
                   3261:                        roff_setstr(r, roff_name[t], NULL, 0);
1.277     schwarze 3262:        return ROFF_CONT;
1.109     kristaps 3263: }
                   3264:
1.340     schwarze 3265: static int
1.109     kristaps 3266: roff_TE(ROFF_ARGS)
                   3267: {
1.361     schwarze 3268:        r->man->flags &= ~ROFF_NONOFILL;
1.321     schwarze 3269:        if (r->tbl == NULL) {
1.350     schwarze 3270:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "TE");
1.321     schwarze 3271:                return ROFF_IGN;
                   3272:        }
1.346     schwarze 3273:        if (tbl_end(r->tbl, 0) == 0) {
1.321     schwarze 3274:                r->tbl = NULL;
1.258     schwarze 3275:                free(buf->buf);
                   3276:                buf->buf = mandoc_strdup(".sp");
                   3277:                buf->sz = 4;
1.329     schwarze 3278:                *offs = 0;
1.277     schwarze 3279:                return ROFF_REPARSE;
1.258     schwarze 3280:        }
1.321     schwarze 3281:        r->tbl = NULL;
1.277     schwarze 3282:        return ROFF_IGN;
1.112     kristaps 3283: }
                   3284:
1.340     schwarze 3285: static int
1.112     kristaps 3286: roff_T_(ROFF_ARGS)
                   3287: {
                   3288:
                   3289:        if (NULL == r->tbl)
1.350     schwarze 3290:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "T&");
1.112     kristaps 3291:        else
1.296     schwarze 3292:                tbl_restart(ln, ppos, r->tbl);
1.112     kristaps 3293:
1.277     schwarze 3294:        return ROFF_IGN;
1.109     kristaps 3295: }
                   3296:
1.230     schwarze 3297: /*
                   3298:  * Handle in-line equation delimiters.
                   3299:  */
1.340     schwarze 3300: static int
1.238     schwarze 3301: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151     kristaps 3302: {
1.233     schwarze 3303:        char            *cp1, *cp2;
                   3304:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151     kristaps 3305:
1.230     schwarze 3306:        /*
                   3307:         * Outside equations, look for an opening delimiter.
                   3308:         * If we are inside an equation, we already know it is
                   3309:         * in-line, or this function wouldn't have been called;
                   3310:         * so look for a closing delimiter.
                   3311:         */
                   3312:
1.238     schwarze 3313:        cp1 = buf->buf + pos;
1.230     schwarze 3314:        cp2 = strchr(cp1, r->eqn == NULL ?
                   3315:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   3316:        if (cp2 == NULL)
1.277     schwarze 3317:                return ROFF_CONT;
1.230     schwarze 3318:
1.233     schwarze 3319:        *cp2++ = '\0';
                   3320:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   3321:
                   3322:        /* Handle preceding text, protecting whitespace. */
                   3323:
1.238     schwarze 3324:        if (*buf->buf != '\0') {
1.233     schwarze 3325:                if (r->eqn == NULL)
                   3326:                        bef_pr = "\\&";
                   3327:                bef_nl = "\n";
                   3328:        }
                   3329:
                   3330:        /*
                   3331:         * Prepare replacing the delimiter with an equation macro
                   3332:         * and drop leading white space from the equation.
                   3333:         */
1.230     schwarze 3334:
1.233     schwarze 3335:        if (r->eqn == NULL) {
                   3336:                while (*cp2 == ' ')
                   3337:                        cp2++;
                   3338:                mac = ".EQ";
                   3339:        } else
                   3340:                mac = ".EN";
                   3341:
                   3342:        /* Handle following text, protecting whitespace. */
                   3343:
                   3344:        if (*cp2 != '\0') {
                   3345:                aft_nl = "\n";
                   3346:                if (r->eqn != NULL)
                   3347:                        aft_pr = "\\&";
                   3348:        }
                   3349:
                   3350:        /* Do the actual replacement. */
                   3351:
1.238     schwarze 3352:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233     schwarze 3353:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238     schwarze 3354:        free(buf->buf);
                   3355:        buf->buf = cp1;
1.230     schwarze 3356:
                   3357:        /* Toggle the in-line state of the eqn subsystem. */
                   3358:
                   3359:        r->eqn_inline = r->eqn == NULL;
1.277     schwarze 3360:        return ROFF_REPARSE;
1.151     kristaps 3361: }
                   3362:
1.340     schwarze 3363: static int
1.235     schwarze 3364: roff_EQ(ROFF_ARGS)
1.125     kristaps 3365: {
1.319     schwarze 3366:        struct roff_node        *n;
                   3367:
1.356     schwarze 3368:        if (r->man->meta.macroset == MACROSET_MAN)
1.322     schwarze 3369:                man_breakscope(r->man, ROFF_EQ);
1.319     schwarze 3370:        n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
                   3371:        if (ln > r->man->last->line)
                   3372:                n->flags |= NODE_LINE;
1.348     schwarze 3373:        n->eqn = eqn_box_new();
1.319     schwarze 3374:        roff_node_append(r->man, n);
                   3375:        r->man->next = ROFF_NEXT_SIBLING;
1.125     kristaps 3376:
1.238     schwarze 3377:        assert(r->eqn == NULL);
1.319     schwarze 3378:        if (r->last_eqn == NULL)
1.351     schwarze 3379:                r->last_eqn = eqn_alloc();
1.319     schwarze 3380:        else
                   3381:                eqn_reset(r->last_eqn);
                   3382:        r->eqn = r->last_eqn;
                   3383:        r->eqn->node = n;
1.151     kristaps 3384:
1.238     schwarze 3385:        if (buf->buf[pos] != '\0')
1.350     schwarze 3386:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.238     schwarze 3387:                    ".EQ %s", buf->buf + pos);
1.151     kristaps 3388:
1.277     schwarze 3389:        return ROFF_IGN;
1.125     kristaps 3390: }
                   3391:
1.340     schwarze 3392: static int
1.125     kristaps 3393: roff_EN(ROFF_ARGS)
                   3394: {
1.319     schwarze 3395:        if (r->eqn != NULL) {
                   3396:                eqn_parse(r->eqn);
                   3397:                r->eqn = NULL;
                   3398:        } else
1.350     schwarze 3399:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "EN");
1.319     schwarze 3400:        if (buf->buf[pos] != '\0')
1.350     schwarze 3401:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.319     schwarze 3402:                    "EN %s", buf->buf + pos);
1.277     schwarze 3403:        return ROFF_IGN;
1.125     kristaps 3404: }
                   3405:
1.340     schwarze 3406: static int
1.109     kristaps 3407: roff_TS(ROFF_ARGS)
                   3408: {
1.321     schwarze 3409:        if (r->tbl != NULL) {
1.350     schwarze 3410:                mandoc_msg(MANDOCERR_BLK_BROKEN, ln, ppos, "TS breaks TS");
1.346     schwarze 3411:                tbl_end(r->tbl, 0);
1.115     kristaps 3412:        }
1.361     schwarze 3413:        r->man->flags |= ROFF_NONOFILL;
1.351     schwarze 3414:        r->tbl = tbl_alloc(ppos, ln, r->last_tbl);
1.346     schwarze 3415:        if (r->last_tbl == NULL)
1.321     schwarze 3416:                r->first_tbl = r->tbl;
                   3417:        r->last_tbl = r->tbl;
1.297     schwarze 3418:        return ROFF_IGN;
                   3419: }
                   3420:
1.340     schwarze 3421: static int
1.358     schwarze 3422: roff_noarg(ROFF_ARGS)
                   3423: {
                   3424:        if (r->man->flags & (MAN_BLINE | MAN_ELINE))
                   3425:                man_breakscope(r->man, tok);
                   3426:        if (tok == ROFF_brp)
                   3427:                tok = ROFF_br;
                   3428:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3429:        if (buf->buf[pos] != '\0')
                   3430:                mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
                   3431:                   "%s %s", roff_name[tok], buf->buf + pos);
                   3432:        if (tok == ROFF_nf)
                   3433:                r->man->flags |= ROFF_NOFILL;
                   3434:        else if (tok == ROFF_fi)
                   3435:                r->man->flags &= ~ROFF_NOFILL;
                   3436:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3437:        r->man->next = ROFF_NEXT_SIBLING;
                   3438:        return ROFF_IGN;
                   3439: }
                   3440:
                   3441: static int
1.297     schwarze 3442: roff_onearg(ROFF_ARGS)
                   3443: {
                   3444:        struct roff_node        *n;
                   3445:        char                    *cp;
1.305     schwarze 3446:        int                      npos;
1.297     schwarze 3447:
1.302     schwarze 3448:        if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320     schwarze 3449:            (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
                   3450:             tok == ROFF_ti))
1.302     schwarze 3451:                man_breakscope(r->man, tok);
                   3452:
1.309     schwarze 3453:        if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305     schwarze 3454:                r->man->last = roffce_node;
                   3455:                r->man->next = ROFF_NEXT_SIBLING;
                   3456:        }
                   3457:
1.297     schwarze 3458:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3459:        n = r->man->last;
                   3460:
                   3461:        cp = buf->buf + pos;
                   3462:        if (*cp != '\0') {
                   3463:                while (*cp != '\0' && *cp != ' ')
                   3464:                        cp++;
                   3465:                while (*cp == ' ')
                   3466:                        *cp++ = '\0';
                   3467:                if (*cp != '\0')
1.350     schwarze 3468:                        mandoc_msg(MANDOCERR_ARG_EXCESS,
                   3469:                            ln, (int)(cp - buf->buf),
1.297     schwarze 3470:                            "%s ... %s", roff_name[tok], cp);
                   3471:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300     schwarze 3472:        }
                   3473:
1.309     schwarze 3474:        if (tok == ROFF_ce || tok == ROFF_rj) {
                   3475:                if (r->man->last->type == ROFFT_ELEM) {
1.305     schwarze 3476:                        roff_word_alloc(r->man, ln, pos, "1");
                   3477:                        r->man->last->flags |= NODE_NOSRC;
                   3478:                }
                   3479:                npos = 0;
                   3480:                if (roff_evalnum(r, ln, r->man->last->string, &npos,
                   3481:                    &roffce_lines, 0) == 0) {
1.350     schwarze 3482:                        mandoc_msg(MANDOCERR_CE_NONUM,
                   3483:                            ln, pos, "ce %s", buf->buf + pos);
1.305     schwarze 3484:                        roffce_lines = 1;
                   3485:                }
                   3486:                if (roffce_lines < 1) {
                   3487:                        r->man->last = r->man->last->parent;
                   3488:                        roffce_node = NULL;
                   3489:                        roffce_lines = 0;
                   3490:                } else
                   3491:                        roffce_node = r->man->last->parent;
                   3492:        } else {
                   3493:                n->flags |= NODE_VALID | NODE_ENDED;
                   3494:                r->man->last = n;
                   3495:        }
                   3496:        n->flags |= NODE_LINE;
1.300     schwarze 3497:        r->man->next = ROFF_NEXT_SIBLING;
                   3498:        return ROFF_IGN;
                   3499: }
                   3500:
1.340     schwarze 3501: static int
1.300     schwarze 3502: roff_manyarg(ROFF_ARGS)
                   3503: {
                   3504:        struct roff_node        *n;
                   3505:        char                    *sp, *ep;
                   3506:
                   3507:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3508:        n = r->man->last;
                   3509:
                   3510:        for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
                   3511:                while (*ep != '\0' && *ep != ' ')
                   3512:                        ep++;
                   3513:                while (*ep == ' ')
                   3514:                        *ep++ = '\0';
                   3515:                roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297     schwarze 3516:        }
                   3517:
                   3518:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3519:        r->man->last = n;
                   3520:        r->man->next = ROFF_NEXT_SIBLING;
1.277     schwarze 3521:        return ROFF_IGN;
1.251     schwarze 3522: }
                   3523:
1.340     schwarze 3524: static int
1.311     schwarze 3525: roff_als(ROFF_ARGS)
                   3526: {
                   3527:        char            *oldn, *newn, *end, *value;
                   3528:        size_t           oldsz, newsz, valsz;
                   3529:
                   3530:        newn = oldn = buf->buf + pos;
                   3531:        if (*newn == '\0')
                   3532:                return ROFF_IGN;
                   3533:
                   3534:        newsz = roff_getname(r, &oldn, ln, pos);
1.363     schwarze 3535:        if (newn[newsz] == '\\' || newn[newsz] == '\t' || *oldn == '\0')
1.311     schwarze 3536:                return ROFF_IGN;
                   3537:
                   3538:        end = oldn;
                   3539:        oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
                   3540:        if (oldsz == 0)
                   3541:                return ROFF_IGN;
                   3542:
1.338     schwarze 3543:        valsz = mandoc_asprintf(&value, ".%.*s \\$@\\\"\n",
1.315     schwarze 3544:            (int)oldsz, oldn);
1.311     schwarze 3545:        roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
                   3546:        roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3547:        free(value);
1.364     schwarze 3548:        return ROFF_IGN;
                   3549: }
                   3550:
                   3551: /*
                   3552:  * The .break request only makes sense inside conditionals,
                   3553:  * and that case is already handled in roff_cond_sub().
                   3554:  */
                   3555: static int
                   3556: roff_break(ROFF_ARGS)
                   3557: {
                   3558:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, pos, "break");
1.296     schwarze 3559:        return ROFF_IGN;
1.92      schwarze 3560: }
                   3561:
1.340     schwarze 3562: static int
1.174     kristaps 3563: roff_cc(ROFF_ARGS)
                   3564: {
                   3565:        const char      *p;
                   3566:
1.238     schwarze 3567:        p = buf->buf + pos;
1.174     kristaps 3568:
1.238     schwarze 3569:        if (*p == '\0' || (r->control = *p++) == '.')
1.303     schwarze 3570:                r->control = '\0';
1.174     kristaps 3571:
1.238     schwarze 3572:        if (*p != '\0')
1.350     schwarze 3573:                mandoc_msg(MANDOCERR_ARG_EXCESS,
1.260     schwarze 3574:                    ln, p - buf->buf, "cc ... %s", p);
1.174     kristaps 3575:
1.341     schwarze 3576:        return ROFF_IGN;
                   3577: }
                   3578:
                   3579: static int
                   3580: roff_char(ROFF_ARGS)
                   3581: {
                   3582:        const char      *p, *kp, *vp;
                   3583:        size_t           ksz, vsz;
                   3584:        int              font;
                   3585:
                   3586:        /* Parse the character to be replaced. */
                   3587:
                   3588:        kp = buf->buf + pos;
                   3589:        p = kp + 1;
                   3590:        if (*kp == '\0' || (*kp == '\\' &&
                   3591:             mandoc_escape(&p, NULL, NULL) != ESCAPE_SPECIAL) ||
                   3592:            (*p != ' ' && *p != '\0')) {
1.350     schwarze 3593:                mandoc_msg(MANDOCERR_CHAR_ARG, ln, pos, "char %s", kp);
1.341     schwarze 3594:                return ROFF_IGN;
                   3595:        }
                   3596:        ksz = p - kp;
                   3597:        while (*p == ' ')
                   3598:                p++;
                   3599:
                   3600:        /*
                   3601:         * If the replacement string contains a font escape sequence,
                   3602:         * we have to restore the font at the end.
                   3603:         */
                   3604:
                   3605:        vp = p;
                   3606:        vsz = strlen(p);
                   3607:        font = 0;
                   3608:        while (*p != '\0') {
                   3609:                if (*p++ != '\\')
                   3610:                        continue;
                   3611:                switch (mandoc_escape(&p, NULL, NULL)) {
                   3612:                case ESCAPE_FONT:
                   3613:                case ESCAPE_FONTROMAN:
                   3614:                case ESCAPE_FONTITALIC:
                   3615:                case ESCAPE_FONTBOLD:
                   3616:                case ESCAPE_FONTBI:
1.378     schwarze 3617:                case ESCAPE_FONTCR:
                   3618:                case ESCAPE_FONTCB:
                   3619:                case ESCAPE_FONTCI:
1.341     schwarze 3620:                case ESCAPE_FONTPREV:
                   3621:                        font++;
                   3622:                        break;
                   3623:                default:
                   3624:                        break;
                   3625:                }
                   3626:        }
                   3627:        if (font > 1)
1.350     schwarze 3628:                mandoc_msg(MANDOCERR_CHAR_FONT,
                   3629:                    ln, (int)(vp - buf->buf), "%s", vp);
1.341     schwarze 3630:
                   3631:        /*
                   3632:         * Approximate the effect of .char using the .tr tables.
                   3633:         * XXX In groff, .char and .tr interact differently.
                   3634:         */
                   3635:
                   3636:        if (ksz == 1) {
                   3637:                if (r->xtab == NULL)
                   3638:                        r->xtab = mandoc_calloc(128, sizeof(*r->xtab));
                   3639:                assert((unsigned int)*kp < 128);
                   3640:                free(r->xtab[(int)*kp].p);
                   3641:                r->xtab[(int)*kp].sz = mandoc_asprintf(&r->xtab[(int)*kp].p,
                   3642:                    "%s%s", vp, font ? "\fP" : "");
                   3643:        } else {
                   3644:                roff_setstrn(&r->xmbtab, kp, ksz, vp, vsz, 0);
                   3645:                if (font)
                   3646:                        roff_setstrn(&r->xmbtab, kp, ksz, "\\fP", 3, 1);
                   3647:        }
1.277     schwarze 3648:        return ROFF_IGN;
1.174     kristaps 3649: }
                   3650:
1.340     schwarze 3651: static int
1.303     schwarze 3652: roff_ec(ROFF_ARGS)
                   3653: {
                   3654:        const char      *p;
                   3655:
                   3656:        p = buf->buf + pos;
                   3657:        if (*p == '\0')
                   3658:                r->escape = '\\';
                   3659:        else {
                   3660:                r->escape = *p;
                   3661:                if (*++p != '\0')
1.350     schwarze 3662:                        mandoc_msg(MANDOCERR_ARG_EXCESS, ln,
                   3663:                            (int)(p - buf->buf), "ec ... %s", p);
1.303     schwarze 3664:        }
                   3665:        return ROFF_IGN;
                   3666: }
                   3667:
1.340     schwarze 3668: static int
1.303     schwarze 3669: roff_eo(ROFF_ARGS)
                   3670: {
                   3671:        r->escape = '\0';
                   3672:        if (buf->buf[pos] != '\0')
1.350     schwarze 3673:                mandoc_msg(MANDOCERR_ARG_SKIP,
1.303     schwarze 3674:                    ln, pos, "eo %s", buf->buf + pos);
1.384     schwarze 3675:        return ROFF_IGN;
                   3676: }
                   3677:
                   3678: static int
                   3679: roff_mc(ROFF_ARGS)
                   3680: {
                   3681:        struct roff_node        *n;
                   3682:        char                    *cp;
                   3683:
                   3684:        /* Parse the first argument. */
                   3685:
                   3686:        cp = buf->buf + pos;
                   3687:        if (*cp != '\0')
                   3688:                cp++;
                   3689:        if (buf->buf[pos] == '\\') {
                   3690:                switch (mandoc_escape((const char **)&cp, NULL, NULL)) {
                   3691:                case ESCAPE_SPECIAL:
                   3692:                case ESCAPE_UNICODE:
                   3693:                case ESCAPE_NUMBERED:
                   3694:                        break;
                   3695:                default:
                   3696:                        *cp = '\0';
                   3697:                        mandoc_msg(MANDOCERR_MC_ESC, ln, pos,
                   3698:                            "mc %s", buf->buf + pos);
                   3699:                        buf->buf[pos] = '\0';
                   3700:                        break;
                   3701:                }
                   3702:        }
                   3703:
                   3704:        /* Ignore additional arguments. */
                   3705:
                   3706:        while (*cp == ' ')
                   3707:                *cp++ = '\0';
                   3708:        if (*cp != '\0') {
                   3709:                mandoc_msg(MANDOCERR_MC_DIST, ln, (int)(cp - buf->buf),
                   3710:                    "mc ... %s", cp);
                   3711:                *cp = '\0';
                   3712:        }
                   3713:
                   3714:        /* Create the .mc node. */
                   3715:
                   3716:        roff_elem_alloc(r->man, ln, ppos, tok);
                   3717:        n = r->man->last;
                   3718:        if (buf->buf[pos] != '\0')
                   3719:                roff_word_alloc(r->man, ln, pos, buf->buf + pos);
                   3720:        n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
                   3721:        r->man->last = n;
                   3722:        r->man->next = ROFF_NEXT_SIBLING;
1.303     schwarze 3723:        return ROFF_IGN;
1.330     schwarze 3724: }
                   3725:
1.340     schwarze 3726: static int
1.330     schwarze 3727: roff_nop(ROFF_ARGS)
                   3728: {
                   3729:        while (buf->buf[pos] == ' ')
                   3730:                pos++;
                   3731:        *offs = pos;
                   3732:        return ROFF_RERUN;
1.303     schwarze 3733: }
                   3734:
1.340     schwarze 3735: static int
1.164     kristaps 3736: roff_tr(ROFF_ARGS)
                   3737: {
                   3738:        const char      *p, *first, *second;
                   3739:        size_t           fsz, ssz;
                   3740:
1.238     schwarze 3741:        p = buf->buf + pos;
1.164     kristaps 3742:
1.238     schwarze 3743:        if (*p == '\0') {
1.350     schwarze 3744:                mandoc_msg(MANDOCERR_REQ_EMPTY, ln, ppos, "tr");
1.277     schwarze 3745:                return ROFF_IGN;
1.164     kristaps 3746:        }
                   3747:
1.238     schwarze 3748:        while (*p != '\0') {
1.164     kristaps 3749:                fsz = ssz = 1;
                   3750:
                   3751:                first = p++;
1.238     schwarze 3752:                if (*first == '\\') {
1.394     schwarze 3753:                        if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR)
1.277     schwarze 3754:                                return ROFF_IGN;
1.164     kristaps 3755:                        fsz = (size_t)(p - first);
                   3756:                }
                   3757:
                   3758:                second = p++;
1.238     schwarze 3759:                if (*second == '\\') {
1.394     schwarze 3760:                        if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR)
1.277     schwarze 3761:                                return ROFF_IGN;
1.164     kristaps 3762:                        ssz = (size_t)(p - second);
1.238     schwarze 3763:                } else if (*second == '\0') {
1.350     schwarze 3764:                        mandoc_msg(MANDOCERR_TR_ODD, ln,
                   3765:                            (int)(first - buf->buf), "tr %s", first);
1.164     kristaps 3766:                        second = " ";
1.165     kristaps 3767:                        p--;
1.164     kristaps 3768:                }
                   3769:
1.167     kristaps 3770:                if (fsz > 1) {
1.207     schwarze 3771:                        roff_setstrn(&r->xmbtab, first, fsz,
                   3772:                            second, ssz, 0);
1.167     kristaps 3773:                        continue;
                   3774:                }
                   3775:
1.238     schwarze 3776:                if (r->xtab == NULL)
1.207     schwarze 3777:                        r->xtab = mandoc_calloc(128,
                   3778:                            sizeof(struct roffstr));
1.167     kristaps 3779:
                   3780:                free(r->xtab[(int)*first].p);
                   3781:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   3782:                r->xtab[(int)*first].sz = ssz;
1.164     kristaps 3783:        }
                   3784:
1.277     schwarze 3785:        return ROFF_IGN;
1.164     kristaps 3786: }
                   3787:
1.339     schwarze 3788: /*
                   3789:  * Implementation of the .return request.
                   3790:  * There is no need to call roff_userret() from here.
                   3791:  * The read module will call that after rewinding the reader stack
                   3792:  * to the place from where the current macro was called.
                   3793:  */
1.340     schwarze 3794: static int
1.339     schwarze 3795: roff_return(ROFF_ARGS)
                   3796: {
                   3797:        if (r->mstackpos >= 0)
1.340     schwarze 3798:                return ROFF_IGN | ROFF_USERRET;
1.339     schwarze 3799:
1.350     schwarze 3800:        mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "return");
1.339     schwarze 3801:        return ROFF_IGN;
                   3802: }
                   3803:
1.340     schwarze 3804: static int
1.306     schwarze 3805: roff_rn(ROFF_ARGS)
                   3806: {
                   3807:        const char      *value;
                   3808:        char            *oldn, *newn, *end;
                   3809:        size_t           oldsz, newsz;
1.315     schwarze 3810:        int              deftype;
1.306     schwarze 3811:
                   3812:        oldn = newn = buf->buf + pos;
                   3813:        if (*oldn == '\0')
                   3814:                return ROFF_IGN;
                   3815:
                   3816:        oldsz = roff_getname(r, &newn, ln, pos);
1.363     schwarze 3817:        if (oldn[oldsz] == '\\' || oldn[oldsz] == '\t' || *newn == '\0')
1.306     schwarze 3818:                return ROFF_IGN;
                   3819:
                   3820:        end = newn;
                   3821:        newsz = roff_getname(r, &end, ln, newn - buf->buf);
                   3822:        if (newsz == 0)
                   3823:                return ROFF_IGN;
                   3824:
1.315     schwarze 3825:        deftype = ROFFDEF_ANY;
                   3826:        value = roff_getstrn(r, oldn, oldsz, &deftype);
                   3827:        switch (deftype) {
                   3828:        case ROFFDEF_USER:
1.306     schwarze 3829:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3830:                roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
                   3831:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315     schwarze 3832:                break;
                   3833:        case ROFFDEF_PRE:
                   3834:                roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
                   3835:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3836:                break;
                   3837:        case ROFFDEF_REN:
1.306     schwarze 3838:                roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
                   3839:                roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315     schwarze 3840:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3841:                break;
                   3842:        case ROFFDEF_STD:
1.306     schwarze 3843:                roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315     schwarze 3844:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3845:                break;
                   3846:        default:
                   3847:                roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
                   3848:                roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
                   3849:                break;
                   3850:        }
1.306     schwarze 3851:        return ROFF_IGN;
                   3852: }
                   3853:
1.340     schwarze 3854: static int
1.339     schwarze 3855: roff_shift(ROFF_ARGS)
                   3856: {
                   3857:        struct mctx     *ctx;
1.382     schwarze 3858:        int              argpos, levels, i;
1.339     schwarze 3859:
1.382     schwarze 3860:        argpos = pos;
1.339     schwarze 3861:        levels = 1;
                   3862:        if (buf->buf[pos] != '\0' &&
                   3863:            roff_evalnum(r, ln, buf->buf, &pos, &levels, 0) == 0) {
1.350     schwarze 3864:                mandoc_msg(MANDOCERR_CE_NONUM,
1.339     schwarze 3865:                    ln, pos, "shift %s", buf->buf + pos);
                   3866:                levels = 1;
                   3867:        }
                   3868:        if (r->mstackpos < 0) {
1.350     schwarze 3869:                mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "shift");
1.339     schwarze 3870:                return ROFF_IGN;
                   3871:        }
                   3872:        ctx = r->mstack + r->mstackpos;
                   3873:        if (levels > ctx->argc) {
1.350     schwarze 3874:                mandoc_msg(MANDOCERR_SHIFT,
1.382     schwarze 3875:                    ln, argpos, "%d, but max is %d", levels, ctx->argc);
1.339     schwarze 3876:                levels = ctx->argc;
1.382     schwarze 3877:        }
                   3878:        if (levels < 0) {
                   3879:                mandoc_msg(MANDOCERR_ARG_NEG, ln, argpos, "shift %d", levels);
                   3880:                levels = 0;
1.339     schwarze 3881:        }
                   3882:        if (levels == 0)
                   3883:                return ROFF_IGN;
                   3884:        for (i = 0; i < levels; i++)
                   3885:                free(ctx->argv[i]);
                   3886:        ctx->argc -= levels;
                   3887:        for (i = 0; i < ctx->argc; i++)
                   3888:                ctx->argv[i] = ctx->argv[i + levels];
                   3889:        return ROFF_IGN;
                   3890: }
                   3891:
1.340     schwarze 3892: static int
1.105     kristaps 3893: roff_so(ROFF_ARGS)
                   3894: {
1.249     schwarze 3895:        char *name, *cp;
1.105     kristaps 3896:
1.238     schwarze 3897:        name = buf->buf + pos;
1.350     schwarze 3898:        mandoc_msg(MANDOCERR_SO, ln, ppos, "so %s", name);
1.105     kristaps 3899:
                   3900:        /*
                   3901:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   3902:         * opening anything that's not in our cwd or anything beneath
                   3903:         * it.  Thus, explicitly disallow traversing up the file-system
                   3904:         * or using absolute paths.
                   3905:         */
                   3906:
1.238     schwarze 3907:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.350     schwarze 3908:                mandoc_msg(MANDOCERR_SO_PATH, ln, ppos, ".so %s", name);
1.249     schwarze 3909:                buf->sz = mandoc_asprintf(&cp,
                   3910:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   3911:                free(buf->buf);
                   3912:                buf->buf = cp;
                   3913:                *offs = 0;
1.277     schwarze 3914:                return ROFF_REPARSE;
1.105     kristaps 3915:        }
                   3916:
                   3917:        *offs = pos;
1.277     schwarze 3918:        return ROFF_SO;
1.105     kristaps 3919: }
1.92      schwarze 3920:
1.266     schwarze 3921: /* --- user defined strings and macros ------------------------------------ */
                   3922:
1.340     schwarze 3923: static int
1.106     kristaps 3924: roff_userdef(ROFF_ARGS)
1.99      kristaps 3925: {
1.339     schwarze 3926:        struct mctx      *ctx;
                   3927:        char             *arg, *ap, *dst, *src;
                   3928:        size_t            sz;
                   3929:
1.365     schwarze 3930:        /* If the macro is empty, ignore it altogether. */
                   3931:
                   3932:        if (*r->current_string == '\0')
                   3933:                return ROFF_IGN;
                   3934:
1.339     schwarze 3935:        /* Initialize a new macro stack context. */
                   3936:
                   3937:        if (++r->mstackpos == r->mstacksz) {
                   3938:                r->mstack = mandoc_recallocarray(r->mstack,
                   3939:                    r->mstacksz, r->mstacksz + 8, sizeof(*r->mstack));
                   3940:                r->mstacksz += 8;
                   3941:        }
                   3942:        ctx = r->mstack + r->mstackpos;
                   3943:        ctx->argc = 0;
                   3944:
                   3945:        /*
                   3946:         * Collect pointers to macro argument strings,
                   3947:         * NUL-terminating them and escaping quotes.
                   3948:         */
                   3949:
                   3950:        src = buf->buf + pos;
                   3951:        while (*src != '\0') {
                   3952:                if (ctx->argc == ctx->argsz) {
                   3953:                        ctx->argsz += 8;
                   3954:                        ctx->argv = mandoc_reallocarray(ctx->argv,
                   3955:                            ctx->argsz, sizeof(*ctx->argv));
                   3956:                }
1.355     schwarze 3957:                arg = roff_getarg(r, &src, ln, &pos);
1.339     schwarze 3958:                sz = 1;  /* For the terminating NUL. */
                   3959:                for (ap = arg; *ap != '\0'; ap++)
                   3960:                        sz += *ap == '"' ? 4 : 1;
                   3961:                ctx->argv[ctx->argc++] = dst = mandoc_malloc(sz);
                   3962:                for (ap = arg; *ap != '\0'; ap++) {
                   3963:                        if (*ap == '"') {
                   3964:                                memcpy(dst, "\\(dq", 4);
                   3965:                                dst += 4;
                   3966:                        } else
                   3967:                                *dst++ = *ap;
1.106     kristaps 3968:                }
1.339     schwarze 3969:                *dst = '\0';
1.355     schwarze 3970:                free(arg);
1.337     schwarze 3971:        }
                   3972:
1.339     schwarze 3973:        /* Replace the macro invocation by the macro definition. */
1.263     schwarze 3974:
1.238     schwarze 3975:        free(buf->buf);
1.339     schwarze 3976:        buf->buf = mandoc_strdup(r->current_string);
                   3977:        buf->sz = strlen(buf->buf) + 1;
1.248     schwarze 3978:        *offs = 0;
1.106     kristaps 3979:
1.365     schwarze 3980:        return buf->buf[buf->sz - 2] == '\n' ?
1.340     schwarze 3981:            ROFF_REPARSE | ROFF_USERCALL : ROFF_IGN | ROFF_APPEND;
1.99      kristaps 3982: }
1.121     schwarze 3983:
1.306     schwarze 3984: /*
                   3985:  * Calling a high-level macro that was renamed with .rn.
                   3986:  * r->current_string has already been set up by roff_parse().
                   3987:  */
1.340     schwarze 3988: static int
1.306     schwarze 3989: roff_renamed(ROFF_ARGS)
                   3990: {
                   3991:        char    *nbuf;
                   3992:
1.315     schwarze 3993:        buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
                   3994:            buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306     schwarze 3995:        free(buf->buf);
                   3996:        buf->buf = nbuf;
1.329     schwarze 3997:        *offs = 0;
1.306     schwarze 3998:        return ROFF_CONT;
                   3999: }
                   4000:
1.362     schwarze 4001: /*
                   4002:  * Measure the length in bytes of the roff identifier at *cpp
                   4003:  * and advance the pointer to the next word.
                   4004:  */
1.212     schwarze 4005: static size_t
1.121     schwarze 4006: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   4007: {
                   4008:        char     *name, *cp;
1.393     schwarze 4009:        int       namesz, inam, iend;
1.121     schwarze 4010:
                   4011:        name = *cpp;
1.362     schwarze 4012:        if (*name == '\0')
1.277     schwarze 4013:                return 0;
1.121     schwarze 4014:
1.362     schwarze 4015:        /* Advance cp to the byte after the end of the name. */
                   4016:
1.393     schwarze 4017:        cp = name;
                   4018:        namesz = 0;
                   4019:        for (;;) {
1.363     schwarze 4020:                if (*cp == '\0')
1.212     schwarze 4021:                        break;
1.363     schwarze 4022:                if (*cp == ' ' || *cp == '\t') {
                   4023:                        cp++;
                   4024:                        break;
                   4025:                }
1.393     schwarze 4026:                if (*cp != '\\') {
                   4027:                        if (name + namesz < cp) {
                   4028:                                name[namesz] = *cp;
                   4029:                                *cp = ' ';
                   4030:                        }
                   4031:                        namesz++;
                   4032:                        cp++;
1.121     schwarze 4033:                        continue;
1.393     schwarze 4034:                }
1.362     schwarze 4035:                if (cp[1] == '{' || cp[1] == '}')
1.215     schwarze 4036:                        break;
1.393     schwarze 4037:                if (roff_escape(cp, 0, 0, NULL, &inam,
                   4038:                    NULL, NULL, &iend) != ESCAPE_UNDEF) {
                   4039:                        mandoc_msg(MANDOCERR_NAMESC, ln, pos,
                   4040:                            "%.*s%.*s", namesz, name, iend, cp);
                   4041:                        cp += iend;
                   4042:                        break;
                   4043:                }
                   4044:
                   4045:                /*
                   4046:                 * In an identifier, \\, \., \G and so on
                   4047:                 * are reduced to \, ., G and so on,
                   4048:                 * vaguely similar to copy mode.
                   4049:                 */
                   4050:
                   4051:                name[namesz++] = cp[inam];
                   4052:                while (iend--) {
                   4053:                        if (cp >= name + namesz)
                   4054:                                *cp = ' ';
                   4055:                        cp++;
                   4056:                }
1.121     schwarze 4057:        }
                   4058:
                   4059:        /* Read past spaces. */
1.362     schwarze 4060:
                   4061:        while (*cp == ' ')
1.121     schwarze 4062:                cp++;
                   4063:
                   4064:        *cpp = cp;
1.277     schwarze 4065:        return namesz;
1.121     schwarze 4066: }
                   4067:
1.106     kristaps 4068: /*
                   4069:  * Store *string into the user-defined string called *name.
                   4070:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193     schwarze 4071:  * append == 0: replace mode
                   4072:  * append == 1: single-line append mode
                   4073:  * append == 2: multiline append mode, append '\n' after each call
1.106     kristaps 4074:  */
1.94      kristaps 4075: static void
1.106     kristaps 4076: roff_setstr(struct roff *r, const char *name, const char *string,
1.193     schwarze 4077:        int append)
1.92      schwarze 4078: {
1.315     schwarze 4079:        size_t   namesz;
1.164     kristaps 4080:
1.315     schwarze 4081:        namesz = strlen(name);
                   4082:        roff_setstrn(&r->strtab, name, namesz, string,
1.207     schwarze 4083:            string ? strlen(string) : 0, append);
1.315     schwarze 4084:        roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164     kristaps 4085: }
                   4086:
                   4087: static void
1.166     kristaps 4088: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193     schwarze 4089:                const char *string, size_t stringsz, int append)
1.164     kristaps 4090: {
1.166     kristaps 4091:        struct roffkv   *n;
1.164     kristaps 4092:        char            *c;
                   4093:        int              i;
                   4094:        size_t           oldch, newch;
1.92      schwarze 4095:
1.106     kristaps 4096:        /* Search for an existing string with the same name. */
1.164     kristaps 4097:        n = *r;
                   4098:
1.211     schwarze 4099:        while (n && (namesz != n->key.sz ||
                   4100:                        strncmp(n->key.p, name, namesz)))
1.92      schwarze 4101:                n = n->next;
1.94      kristaps 4102:
                   4103:        if (NULL == n) {
1.106     kristaps 4104:                /* Create a new string table entry. */
1.166     kristaps 4105:                n = mandoc_malloc(sizeof(struct roffkv));
                   4106:                n->key.p = mandoc_strndup(name, namesz);
                   4107:                n->key.sz = namesz;
                   4108:                n->val.p = NULL;
                   4109:                n->val.sz = 0;
1.164     kristaps 4110:                n->next = *r;
                   4111:                *r = n;
1.193     schwarze 4112:        } else if (0 == append) {
1.166     kristaps 4113:                free(n->val.p);
                   4114:                n->val.p = NULL;
                   4115:                n->val.sz = 0;
1.106     kristaps 4116:        }
                   4117:
                   4118:        if (NULL == string)
                   4119:                return;
                   4120:
                   4121:        /*
                   4122:         * One additional byte for the '\n' in multiline mode,
                   4123:         * and one for the terminating '\0'.
                   4124:         */
1.193     schwarze 4125:        newch = stringsz + (1 < append ? 2u : 1u);
1.164     kristaps 4126:
1.166     kristaps 4127:        if (NULL == n->val.p) {
                   4128:                n->val.p = mandoc_malloc(newch);
                   4129:                *n->val.p = '\0';
1.106     kristaps 4130:                oldch = 0;
                   4131:        } else {
1.166     kristaps 4132:                oldch = n->val.sz;
                   4133:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106     kristaps 4134:        }
                   4135:
                   4136:        /* Skip existing content in the destination buffer. */
1.166     kristaps 4137:        c = n->val.p + (int)oldch;
1.106     kristaps 4138:
                   4139:        /* Append new content to the destination buffer. */
1.164     kristaps 4140:        i = 0;
                   4141:        while (i < (int)stringsz) {
1.106     kristaps 4142:                /*
                   4143:                 * Rudimentary roff copy mode:
                   4144:                 * Handle escaped backslashes.
                   4145:                 */
1.164     kristaps 4146:                if ('\\' == string[i] && '\\' == string[i + 1])
                   4147:                        i++;
                   4148:                *c++ = string[i++];
1.106     kristaps 4149:        }
1.94      kristaps 4150:
1.106     kristaps 4151:        /* Append terminating bytes. */
1.193     schwarze 4152:        if (1 < append)
1.106     kristaps 4153:                *c++ = '\n';
1.163     kristaps 4154:
1.106     kristaps 4155:        *c = '\0';
1.166     kristaps 4156:        n->val.sz = (int)(c - n->val.p);
1.92      schwarze 4157: }
                   4158:
1.94      kristaps 4159: static const char *
1.325     schwarze 4160: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315     schwarze 4161:     int *deftype)
1.92      schwarze 4162: {
1.315     schwarze 4163:        const struct roffkv     *n;
1.325     schwarze 4164:        int                      found, i;
1.315     schwarze 4165:        enum roff_tok            tok;
                   4166:
1.325     schwarze 4167:        found = 0;
                   4168:        for (n = r->strtab; n != NULL; n = n->next) {
                   4169:                if (strncmp(name, n->key.p, len) != 0 ||
                   4170:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4171:                        continue;
                   4172:                if (*deftype & ROFFDEF_USER) {
                   4173:                        *deftype = ROFFDEF_USER;
                   4174:                        return n->val.p;
                   4175:                } else {
                   4176:                        found = 1;
                   4177:                        break;
                   4178:                }
                   4179:        }
                   4180:        for (n = r->rentab; n != NULL; n = n->next) {
                   4181:                if (strncmp(name, n->key.p, len) != 0 ||
                   4182:                    n->key.p[len] != '\0' || n->val.p == NULL)
                   4183:                        continue;
                   4184:                if (*deftype & ROFFDEF_REN) {
                   4185:                        *deftype = ROFFDEF_REN;
                   4186:                        return n->val.p;
                   4187:                } else {
                   4188:                        found = 1;
                   4189:                        break;
1.315     schwarze 4190:                }
                   4191:        }
1.325     schwarze 4192:        for (i = 0; i < PREDEFS_MAX; i++) {
                   4193:                if (strncmp(name, predefs[i].name, len) != 0 ||
                   4194:                    predefs[i].name[len] != '\0')
                   4195:                        continue;
                   4196:                if (*deftype & ROFFDEF_PRE) {
                   4197:                        *deftype = ROFFDEF_PRE;
                   4198:                        return predefs[i].str;
                   4199:                } else {
                   4200:                        found = 1;
                   4201:                        break;
1.315     schwarze 4202:                }
                   4203:        }
1.356     schwarze 4204:        if (r->man->meta.macroset != MACROSET_MAN) {
1.325     schwarze 4205:                for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
                   4206:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4207:                            roff_name[tok][len] != '\0')
                   4208:                                continue;
                   4209:                        if (*deftype & ROFFDEF_STD) {
                   4210:                                *deftype = ROFFDEF_STD;
                   4211:                                return NULL;
                   4212:                        } else {
                   4213:                                found = 1;
                   4214:                                break;
1.315     schwarze 4215:                        }
                   4216:                }
                   4217:        }
1.356     schwarze 4218:        if (r->man->meta.macroset != MACROSET_MDOC) {
1.325     schwarze 4219:                for (tok = MAN_TH; tok < MAN_MAX; tok++) {
                   4220:                        if (strncmp(name, roff_name[tok], len) != 0 ||
                   4221:                            roff_name[tok][len] != '\0')
                   4222:                                continue;
                   4223:                        if (*deftype & ROFFDEF_STD) {
                   4224:                                *deftype = ROFFDEF_STD;
                   4225:                                return NULL;
                   4226:                        } else {
                   4227:                                found = 1;
                   4228:                                break;
1.315     schwarze 4229:                        }
                   4230:                }
1.325     schwarze 4231:        }
                   4232:
                   4233:        if (found == 0 && *deftype != ROFFDEF_ANY) {
                   4234:                if (*deftype & ROFFDEF_REN) {
                   4235:                        /*
                   4236:                         * This might still be a request,
                   4237:                         * so do not treat it as undefined yet.
                   4238:                         */
                   4239:                        *deftype = ROFFDEF_UNDEF;
                   4240:                        return NULL;
1.315     schwarze 4241:                }
1.325     schwarze 4242:
                   4243:                /* Using an undefined string defines it to be empty. */
                   4244:
                   4245:                roff_setstrn(&r->strtab, name, len, "", 0, 0);
                   4246:                roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315     schwarze 4247:        }
1.325     schwarze 4248:
1.315     schwarze 4249:        *deftype = 0;
1.277     schwarze 4250:        return NULL;
1.92      schwarze 4251: }
                   4252:
1.94      kristaps 4253: static void
1.167     kristaps 4254: roff_freestr(struct roffkv *r)
1.92      schwarze 4255: {
1.166     kristaps 4256:        struct roffkv    *n, *nn;
1.92      schwarze 4257:
1.167     kristaps 4258:        for (n = r; n; n = nn) {
1.166     kristaps 4259:                free(n->key.p);
                   4260:                free(n->val.p);
1.92      schwarze 4261:                nn = n->next;
                   4262:                free(n);
                   4263:        }
1.114     kristaps 4264: }
1.266     schwarze 4265:
                   4266: /* --- accessors and utility functions ------------------------------------ */
1.164     kristaps 4267:
                   4268: /*
                   4269:  * Duplicate an input string, making the appropriate character
                   4270:  * conversations (as stipulated by `tr') along the way.
                   4271:  * Returns a heap-allocated string with all the replacements made.
                   4272:  */
                   4273: char *
                   4274: roff_strdup(const struct roff *r, const char *p)
                   4275: {
1.166     kristaps 4276:        const struct roffkv *cp;
1.164     kristaps 4277:        char            *res;
                   4278:        const char      *pp;
                   4279:        size_t           ssz, sz;
                   4280:        enum mandoc_esc  esc;
                   4281:
1.167     kristaps 4282:        if (NULL == r->xmbtab && NULL == r->xtab)
1.277     schwarze 4283:                return mandoc_strdup(p);
1.164     kristaps 4284:        else if ('\0' == *p)
1.277     schwarze 4285:                return mandoc_strdup("");
1.164     kristaps 4286:
                   4287:        /*
                   4288:         * Step through each character looking for term matches
                   4289:         * (remember that a `tr' can be invoked with an escape, which is
                   4290:         * a glyph but the escape is multi-character).
                   4291:         * We only do this if the character hash has been initialised
                   4292:         * and the string is >0 length.
                   4293:         */
                   4294:
                   4295:        res = NULL;
                   4296:        ssz = 0;
                   4297:
                   4298:        while ('\0' != *p) {
1.289     schwarze 4299:                assert((unsigned int)*p < 128);
                   4300:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167     kristaps 4301:                        sz = r->xtab[(int)*p].sz;
                   4302:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4303:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   4304:                        ssz += sz;
                   4305:                        p++;
                   4306:                        continue;
                   4307:                } else if ('\\' != *p) {
                   4308:                        res = mandoc_realloc(res, ssz + 2);
                   4309:                        res[ssz++] = *p++;
                   4310:                        continue;
                   4311:                }
                   4312:
1.164     kristaps 4313:                /* Search for term matches. */
1.167     kristaps 4314:                for (cp = r->xmbtab; cp; cp = cp->next)
1.166     kristaps 4315:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164     kristaps 4316:                                break;
                   4317:
                   4318:                if (NULL != cp) {
                   4319:                        /*
                   4320:                         * A match has been found.
                   4321:                         * Append the match to the array and move
                   4322:                         * forward by its keysize.
                   4323:                         */
1.207     schwarze 4324:                        res = mandoc_realloc(res,
                   4325:                            ssz + cp->val.sz + 1);
1.166     kristaps 4326:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   4327:                        ssz += cp->val.sz;
                   4328:                        p += (int)cp->key.sz;
1.164     kristaps 4329:                        continue;
                   4330:                }
                   4331:
1.167     kristaps 4332:                /*
                   4333:                 * Handle escapes carefully: we need to copy
                   4334:                 * over just the escape itself, or else we might
                   4335:                 * do replacements within the escape itself.
                   4336:                 * Make sure to pass along the bogus string.
                   4337:                 */
                   4338:                pp = p++;
                   4339:                esc = mandoc_escape(&p, NULL, NULL);
                   4340:                if (ESCAPE_ERROR == esc) {
                   4341:                        sz = strlen(pp);
1.164     kristaps 4342:                        res = mandoc_realloc(res, ssz + sz + 1);
                   4343:                        memcpy(res + ssz, pp, sz);
1.167     kristaps 4344:                        break;
1.164     kristaps 4345:                }
1.207     schwarze 4346:                /*
                   4347:                 * We bail out on bad escapes.
1.167     kristaps 4348:                 * No need to warn: we already did so when
1.355     schwarze 4349:                 * roff_expand() was called.
1.167     kristaps 4350:                 */
                   4351:                sz = (int)(p - pp);
                   4352:                res = mandoc_realloc(res, ssz + sz + 1);
                   4353:                memcpy(res + ssz, pp, sz);
                   4354:                ssz += sz;
1.164     kristaps 4355:        }
                   4356:
                   4357:        res[(int)ssz] = '\0';
1.277     schwarze 4358:        return res;
1.227     schwarze 4359: }
                   4360:
                   4361: int
                   4362: roff_getformat(const struct roff *r)
                   4363: {
                   4364:
1.277     schwarze 4365:        return r->format;
1.174     kristaps 4366: }
                   4367:
                   4368: /*
1.207     schwarze 4369:  * Find out whether a line is a macro line or not.
1.174     kristaps 4370:  * If it is, adjust the current position and return one; if it isn't,
                   4371:  * return zero and don't change the current position.
                   4372:  * If the control character has been set with `.cc', then let that grain
                   4373:  * precedence.
                   4374:  * This is slighly contrary to groff, where using the non-breaking
                   4375:  * control character when `cc' has been invoked will cause the
                   4376:  * non-breaking macro contents to be printed verbatim.
                   4377:  */
                   4378: int
                   4379: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   4380: {
                   4381:        int             pos;
                   4382:
                   4383:        pos = *ppos;
                   4384:
1.303     schwarze 4385:        if (r->control != '\0' && cp[pos] == r->control)
1.174     kristaps 4386:                pos++;
1.303     schwarze 4387:        else if (r->control != '\0')
1.277     schwarze 4388:                return 0;
1.174     kristaps 4389:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   4390:                pos += 2;
                   4391:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   4392:                pos++;
                   4393:        else
1.277     schwarze 4394:                return 0;
1.174     kristaps 4395:
                   4396:        while (' ' == cp[pos] || '\t' == cp[pos])
                   4397:                pos++;
                   4398:
                   4399:        *ppos = pos;
1.277     schwarze 4400:        return 1;
1.74      kristaps 4401: }

CVSweb