[BACK]Return to roff.c CVS log [TXT][DIR] Up to [local] / src / usr.bin / mandoc

Annotation of src/usr.bin/mandoc/roff.c, Revision 1.168

1.168   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.167 2017/04/29 12:43:55 schwarze Exp $ */
1.1       schwarze    2: /*
1.139     schwarze    3:  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.159     schwarze    4:  * Copyright (c) 2010-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
1.1       schwarze    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
                      7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
                      9:  *
1.16      schwarze   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1       schwarze   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16      schwarze   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1       schwarze   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.
                     17:  */
1.99      schwarze   18: #include <sys/types.h>
                     19:
1.1       schwarze   20: #include <assert.h>
1.3       schwarze   21: #include <ctype.h>
1.117     schwarze   22: #include <limits.h>
1.167     schwarze   23: #include <stddef.h>
                     24: #include <stdint.h>
1.51      schwarze   25: #include <stdio.h>
1.1       schwarze   26: #include <stdlib.h>
                     27: #include <string.h>
                     28:
                     29: #include "mandoc.h"
1.74      schwarze   30: #include "mandoc_aux.h"
1.167     schwarze   31: #include "mandoc_ohash.h"
1.137     schwarze   32: #include "roff.h"
1.98      daniel     33: #include "libmandoc.h"
1.138     schwarze   34: #include "roff_int.h"
1.27      schwarze   35: #include "libroff.h"
1.2       schwarze   36:
1.43      schwarze   37: /* Maximum number of string expansions per line, to break infinite loops. */
                     38: #define        EXPAND_LIMIT    1000
                     39:
1.138     schwarze   40: /* --- data types --------------------------------------------------------- */
                     41:
1.41      schwarze   42: /*
1.42      schwarze   43:  * An incredibly-simple string buffer.
                     44:  */
1.8       schwarze   45: struct roffstr {
1.42      schwarze   46:        char            *p; /* nil-terminated buffer */
                     47:        size_t           sz; /* saved strlen(p) */
                     48: };
                     49:
                     50: /*
                     51:  * A key-value roffstr pair as part of a singly-linked list.
                     52:  */
                     53: struct roffkv {
                     54:        struct roffstr   key;
                     55:        struct roffstr   val;
                     56:        struct roffkv   *next; /* next in list */
1.8       schwarze   57: };
                     58:
1.52      schwarze   59: /*
                     60:  * A single number register as part of a singly-linked list.
                     61:  */
                     62: struct roffreg {
                     63:        struct roffstr   key;
1.53      schwarze   64:        int              val;
1.52      schwarze   65:        struct roffreg  *next;
                     66: };
                     67:
1.167     schwarze   68: /*
                     69:  * Association of request and macro names with token IDs.
                     70:  */
                     71: struct roffreq {
                     72:        enum roff_tok    tok;
                     73:        char             name[];
                     74: };
                     75:
1.1       schwarze   76: struct roff {
1.35      schwarze   77:        struct mparse   *parse; /* parse point */
1.168   ! schwarze   78:        struct roff_man *man; /* mdoc or man parser */
1.1       schwarze   79:        struct roffnode *last; /* leaf of stack */
1.96      schwarze   80:        int             *rstack; /* stack of inverted `ie' values */
1.167     schwarze   81:        struct ohash    *reqtab; /* request lookup table */
1.52      schwarze   82:        struct roffreg  *regtab; /* number registers */
1.42      schwarze   83:        struct roffkv   *strtab; /* user-defined strings & macros */
                     84:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                     85:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze   86:        const char      *current_string; /* value of last called user macro */
1.27      schwarze   87:        struct tbl_node *first_tbl; /* first table parsed */
                     88:        struct tbl_node *last_tbl; /* last table parsed */
                     89:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze   90:        struct eqn_node *last_eqn; /* last equation parsed */
                     91:        struct eqn_node *first_eqn; /* first equation parsed */
                     92:        struct eqn_node *eqn; /* current equation being parsed */
1.102     schwarze   93:        int              eqn_inline; /* current equation is inline */
1.96      schwarze   94:        int              options; /* parse options */
                     95:        int              rstacksz; /* current size limit of rstack */
                     96:        int              rstackpos; /* position in rstack */
1.99      schwarze   97:        int              format; /* current file in mdoc or man format */
1.145     schwarze   98:        int              argc; /* number of args of the last macro */
1.96      schwarze   99:        char             control; /* control character */
1.1       schwarze  100: };
                    101:
                    102: struct roffnode {
1.166     schwarze  103:        enum roff_tok    tok; /* type of node */
1.1       schwarze  104:        struct roffnode *parent; /* up one in stack */
                    105:        int              line; /* parse line */
                    106:        int              col; /* parse col */
1.16      schwarze  107:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  108:        char            *end; /* end-rules: custom token */
                    109:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  110:        int              rule; /* current evaluation rule */
1.1       schwarze  111: };
                    112:
                    113: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
1.166     schwarze  114:                         enum roff_tok tok, /* tok of macro */ \
1.110     schwarze  115:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  116:                         int ln, /* parse line */ \
1.2       schwarze  117:                         int ppos, /* original pos in buffer */ \
                    118:                         int pos, /* current pos in buffer */ \
                    119:                         int *offs /* reset offset of buffer data */
1.1       schwarze  120:
                    121: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    122:
                    123: struct roffmac {
1.2       schwarze  124:        roffproc         proc; /* process new macro */
                    125:        roffproc         text; /* process as child text of macro */
                    126:        roffproc         sub; /* process as child of macro */
                    127:        int              flags;
                    128: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.1       schwarze  129: };
                    130:
1.37      schwarze  131: struct predef {
                    132:        const char      *name; /* predefined input name */
                    133:        const char      *str; /* replacement symbol */
                    134: };
                    135:
                    136: #define        PREDEF(__name, __str) \
                    137:        { (__name), (__str) },
                    138:
1.138     schwarze  139: /* --- function prototypes ------------------------------------------------ */
                    140:
1.42      schwarze  141: static void             roffnode_cleanscope(struct roff *);
                    142: static void             roffnode_pop(struct roff *);
1.166     schwarze  143: static void             roffnode_push(struct roff *, enum roff_tok,
1.42      schwarze  144:                                const char *, int, int);
1.2       schwarze  145: static enum rofferr     roff_block(ROFF_ARGS);
                    146: static enum rofferr     roff_block_text(ROFF_ARGS);
                    147: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.168   ! schwarze  148: static enum rofferr     roff_br(ROFF_ARGS);
1.2       schwarze  149: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  150: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  151: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  152: static enum rofferr     roff_cond(ROFF_ARGS);
                    153: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    154: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  155: static enum rofferr     roff_ds(ROFF_ARGS);
1.110     schwarze  156: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.143     schwarze  157: static int              roff_evalcond(struct roff *r, int, char *, int *);
1.106     schwarze  158: static int              roff_evalnum(struct roff *, int,
                    159:                                const char *, int *, int *, int);
                    160: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  161:                                const char *, int *, int *, int);
1.71      schwarze  162: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  163: static void             roff_free1(struct roff *);
1.52      schwarze  164: static void             roff_freereg(struct roffreg *);
1.42      schwarze  165: static void             roff_freestr(struct roffkv *);
1.85      schwarze  166: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  167: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  168: static int              roff_getop(const char *, int *, char *);
1.53      schwarze  169: static int              roff_getregn(const struct roff *,
                    170:                                const char *, size_t);
1.145     schwarze  171: static int              roff_getregro(const struct roff *,
                    172:                                const char *name);
1.80      schwarze  173: static const char      *roff_getstrn(const struct roff *,
1.8       schwarze  174:                                const char *, size_t);
1.143     schwarze  175: static int              roff_hasregn(const struct roff *,
                    176:                                const char *, size_t);
1.123     schwarze  177: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  178: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  179: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137     schwarze  180: static void             roff_man_alloc1(struct roff_man *);
                    181: static void             roff_man_free1(struct roff_man *);
1.6       schwarze  182: static enum rofferr     roff_nr(ROFF_ARGS);
1.166     schwarze  183: static enum roff_tok    roff_parse(struct roff *, char *, int *,
1.87      schwarze  184:                                int, int);
1.110     schwarze  185: static enum rofferr     roff_parsetext(struct buf *, int, int *);
                    186: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  187: static enum rofferr     roff_rm(ROFF_ARGS);
1.76      schwarze  188: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  189: static void             roff_setstr(struct roff *,
1.16      schwarze  190:                                const char *, const char *, int);
1.80      schwarze  191: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  192:                                size_t, const char *, size_t, int);
1.14      schwarze  193: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  194: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  195: static enum rofferr     roff_Dd(ROFF_ARGS);
                    196: static enum rofferr     roff_TH(ROFF_ARGS);
1.27      schwarze  197: static enum rofferr     roff_TE(ROFF_ARGS);
                    198: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  199: static enum rofferr     roff_EQ(ROFF_ARGS);
                    200: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  201: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  202: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  203: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  204:
1.138     schwarze  205: /* --- constant data ------------------------------------------------------ */
                    206:
1.133     schwarze  207: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    208: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    209:
1.166     schwarze  210: const char *__roff_name[MAN_MAX + 1] = {
1.168   ! schwarze  211:        "br",           NULL,
1.166     schwarze  212:        "ab",           "ad",           "af",           "aln",
                    213:        "als",          "am",           "am1",          "ami",
                    214:        "ami1",         "as",           "as1",          "asciify",
                    215:        "backtrace",    "bd",           "bleedat",      "blm",
                    216:         "box",         "boxa",         "bp",           "BP",
                    217:        "break",        "breakchar",    "brnl",         "brp",
                    218:        "brpnl",        "c2",           "cc",           "ce",
                    219:        "cf",           "cflags",       "ch",           "char",
                    220:        "chop",         "class",        "close",        "CL",
                    221:        "color",        "composite",    "continue",     "cp",
                    222:        "cropat",       "cs",           "cu",           "da",
                    223:        "dch",          "Dd",           "de",           "de1",
                    224:        "defcolor",     "dei",          "dei1",         "device",
                    225:        "devicem",      "di",           "do",           "ds",
                    226:        "ds1",          "dwh",          "dt",           "ec",
                    227:        "ecr",          "ecs",          "el",           "em",
                    228:        "EN",           "eo",           "EP",           "EQ",
                    229:        "errprint",     "ev",           "evc",          "ex",
                    230:        "fallback",     "fam",          "fc",           "fchar",
                    231:        "fcolor",       "fdeferlig",    "feature",      "fkern",
                    232:        "fl",           "flig",         "fp",           "fps",
                    233:        "fschar",       "fspacewidth",  "fspecial",     "ftr",
                    234:        "fzoom",        "gcolor",       "hc",           "hcode",
                    235:        "hidechar",     "hla",          "hlm",          "hpf",
                    236:        "hpfa",         "hpfcode",      "hw",           "hy",
                    237:        "hylang",       "hylen",        "hym",          "hypp",
                    238:        "hys",          "ie",           "if",           "ig",
                    239:        "index",        "it",           "itc",          "IX",
                    240:        "kern",         "kernafter",    "kernbefore",   "kernpair",
                    241:        "lc",           "lc_ctype",     "lds",          "length",
                    242:        "letadj",       "lf",           "lg",           "lhang",
                    243:        "linetabs",     "lnr",          "lnrf",         "lpfx",
                    244:        "ls",           "lsm",          "lt",           "mc",
                    245:        "mediasize",    "minss",        "mk",           "mso",
                    246:        "na",           "ne",           "nh",           "nhychar",
                    247:        "nm",           "nn",           "nop",          "nr",
                    248:        "nrf",          "nroff",        "ns",           "nx",
                    249:        "open",         "opena",        "os",           "output",
                    250:        "padj",         "papersize",    "pc",           "pev",
                    251:        "pi",           "PI",           "pl",           "pm",
                    252:        "pn",           "pnr",          "po",           "ps",
                    253:        "psbb",         "pshape",       "pso",          "ptr",
                    254:        "pvs",          "rchar",        "rd",           "recursionlimit",
                    255:        "return",       "rfschar",      "rhang",        "rj",
                    256:        "rm",           "rn",           "rnn",          "rr",
                    257:        "rs",           "rt",           "schar",        "sentchar",
                    258:        "shc",          "shift",        "sizes",        "so",
                    259:        "spacewidth",   "special",      "spreadwarn",   "ss",
                    260:        "sty",          "substring",    "sv",           "sy",
                    261:        "T&",           "ta",           "tc",           "TE",
                    262:        "TH",           "ti",           "tkf",          "tl",
                    263:        "tm",           "tm1",          "tmc",          "tr",
                    264:        "track",        "transchar",    "trf",          "trimat",
                    265:        "trin",         "trnt",         "troff",        "TS",
                    266:        "uf",           "ul",           "unformat",     "unwatch",
                    267:        "unwatchn",     "vpt",          "vs",           "warn",
                    268:        "warnscale",    "watch",        "watchlength",  "watchn",
                    269:        "wh",           "while",        "write",        "writec",
                    270:        "writem",       "xflag",        ".",            NULL,
                    271:        "text",
                    272:        "Dd",           "Dt",           "Os",           "Sh",
                    273:        "Ss",           "Pp",           "D1",           "Dl",
                    274:        "Bd",           "Ed",           "Bl",           "El",
                    275:        "It",           "Ad",           "An",           "Ap",
                    276:        "Ar",           "Cd",           "Cm",           "Dv",
                    277:        "Er",           "Ev",           "Ex",           "Fa",
                    278:        "Fd",           "Fl",           "Fn",           "Ft",
                    279:        "Ic",           "In",           "Li",           "Nd",
                    280:        "Nm",           "Op",           "Ot",           "Pa",
                    281:        "Rv",           "St",           "Va",           "Vt",
                    282:        "Xr",           "%A",           "%B",           "%D",
                    283:        "%I",           "%J",           "%N",           "%O",
                    284:        "%P",           "%R",           "%T",           "%V",
                    285:        "Ac",           "Ao",           "Aq",           "At",
                    286:        "Bc",           "Bf",           "Bo",           "Bq",
                    287:        "Bsx",          "Bx",           "Db",           "Dc",
                    288:        "Do",           "Dq",           "Ec",           "Ef",
                    289:        "Em",           "Eo",           "Fx",           "Ms",
                    290:        "No",           "Ns",           "Nx",           "Ox",
                    291:        "Pc",           "Pf",           "Po",           "Pq",
                    292:        "Qc",           "Ql",           "Qo",           "Qq",
                    293:        "Re",           "Rs",           "Sc",           "So",
                    294:        "Sq",           "Sm",           "Sx",           "Sy",
                    295:        "Tn",           "Ux",           "Xc",           "Xo",
                    296:        "Fo",           "Fc",           "Oo",           "Oc",
                    297:        "Bk",           "Ek",           "Bt",           "Hf",
                    298:        "Fr",           "Ud",           "Lb",           "Lp",
                    299:        "Lk",           "Mt",           "Brq",          "Bro",
                    300:        "Brc",          "%C",           "Es",           "En",
1.168   ! schwarze  301:        "Dx",           "%Q",           "sp",
1.166     schwarze  302:        "%U",           "Ta",           "ll",           NULL,
                    303:        "TH",           "SH",           "SS",           "TP",
                    304:        "LP",           "PP",           "P",            "IP",
                    305:        "HP",           "SM",           "SB",           "BI",
                    306:        "IB",           "BR",           "RB",           "R",
                    307:        "B",            "I",            "IR",           "RI",
1.168   ! schwarze  308:        "sp",           "nf",           "fi",
1.166     schwarze  309:        "RE",           "RS",           "DT",           "UC",
                    310:        "PD",           "AT",           "in",           "ft",
                    311:        "OP",           "EX",           "EE",           "UR",
                    312:        "UE",           "ll",           NULL
                    313: };
                    314: const  char *const *roff_name = __roff_name;
                    315:
                    316: static struct roffmac   roffs[TOKEN_NONE] = {
1.168   ! schwarze  317:        { roff_br, NULL, NULL, 0 },  /* br */
        !           318:        { NULL, NULL, NULL, 0 },  /* ROFF_MAX */
1.167     schwarze  319:        { roff_unsupp, NULL, NULL, 0 },  /* ab */
                    320:        { roff_line_ignore, NULL, NULL, 0 },  /* ad */
                    321:        { roff_line_ignore, NULL, NULL, 0 },  /* af */
                    322:        { roff_unsupp, NULL, NULL, 0 },  /* aln */
                    323:        { roff_unsupp, NULL, NULL, 0 },  /* als */
                    324:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am */
                    325:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* am1 */
                    326:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami */
                    327:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ami1 */
                    328:        { roff_ds, NULL, NULL, 0 },  /* as */
                    329:        { roff_ds, NULL, NULL, 0 },  /* as1 */
                    330:        { roff_unsupp, NULL, NULL, 0 },  /* asciify */
                    331:        { roff_line_ignore, NULL, NULL, 0 },  /* backtrace */
                    332:        { roff_line_ignore, NULL, NULL, 0 },  /* bd */
                    333:        { roff_line_ignore, NULL, NULL, 0 },  /* bleedat */
                    334:        { roff_unsupp, NULL, NULL, 0 },  /* blm */
                    335:        { roff_unsupp, NULL, NULL, 0 },  /* box */
                    336:        { roff_unsupp, NULL, NULL, 0 },  /* boxa */
                    337:        { roff_line_ignore, NULL, NULL, 0 },  /* bp */
                    338:        { roff_unsupp, NULL, NULL, 0 },  /* BP */
                    339:        { roff_unsupp, NULL, NULL, 0 },  /* break */
                    340:        { roff_line_ignore, NULL, NULL, 0 },  /* breakchar */
                    341:        { roff_line_ignore, NULL, NULL, 0 },  /* brnl */
1.168   ! schwarze  342:        { roff_br, NULL, NULL, 0 },  /* brp */
1.167     schwarze  343:        { roff_line_ignore, NULL, NULL, 0 },  /* brpnl */
                    344:        { roff_unsupp, NULL, NULL, 0 },  /* c2 */
                    345:        { roff_cc, NULL, NULL, 0 },  /* cc */
                    346:        { roff_line_ignore, NULL, NULL, 0 },  /* ce */
                    347:        { roff_insec, NULL, NULL, 0 },  /* cf */
                    348:        { roff_line_ignore, NULL, NULL, 0 },  /* cflags */
                    349:        { roff_line_ignore, NULL, NULL, 0 },  /* ch */
                    350:        { roff_unsupp, NULL, NULL, 0 },  /* char */
                    351:        { roff_unsupp, NULL, NULL, 0 },  /* chop */
                    352:        { roff_line_ignore, NULL, NULL, 0 },  /* class */
                    353:        { roff_insec, NULL, NULL, 0 },  /* close */
                    354:        { roff_unsupp, NULL, NULL, 0 },  /* CL */
                    355:        { roff_line_ignore, NULL, NULL, 0 },  /* color */
                    356:        { roff_unsupp, NULL, NULL, 0 },  /* composite */
                    357:        { roff_unsupp, NULL, NULL, 0 },  /* continue */
                    358:        { roff_line_ignore, NULL, NULL, 0 },  /* cp */
                    359:        { roff_line_ignore, NULL, NULL, 0 },  /* cropat */
                    360:        { roff_line_ignore, NULL, NULL, 0 },  /* cs */
                    361:        { roff_line_ignore, NULL, NULL, 0 },  /* cu */
                    362:        { roff_unsupp, NULL, NULL, 0 },  /* da */
                    363:        { roff_unsupp, NULL, NULL, 0 },  /* dch */
                    364:        { roff_Dd, NULL, NULL, 0 },  /* Dd */
                    365:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de */
                    366:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* de1 */
                    367:        { roff_line_ignore, NULL, NULL, 0 },  /* defcolor */
                    368:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei */
                    369:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* dei1 */
                    370:        { roff_unsupp, NULL, NULL, 0 },  /* device */
                    371:        { roff_unsupp, NULL, NULL, 0 },  /* devicem */
                    372:        { roff_unsupp, NULL, NULL, 0 },  /* di */
                    373:        { roff_unsupp, NULL, NULL, 0 },  /* do */
                    374:        { roff_ds, NULL, NULL, 0 },  /* ds */
                    375:        { roff_ds, NULL, NULL, 0 },  /* ds1 */
                    376:        { roff_unsupp, NULL, NULL, 0 },  /* dwh */
                    377:        { roff_unsupp, NULL, NULL, 0 },  /* dt */
                    378:        { roff_unsupp, NULL, NULL, 0 },  /* ec */
                    379:        { roff_unsupp, NULL, NULL, 0 },  /* ecr */
                    380:        { roff_unsupp, NULL, NULL, 0 },  /* ecs */
                    381:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* el */
                    382:        { roff_unsupp, NULL, NULL, 0 },  /* em */
                    383:        { roff_EN, NULL, NULL, 0 },  /* EN */
                    384:        { roff_unsupp, NULL, NULL, 0 },  /* eo */
                    385:        { roff_unsupp, NULL, NULL, 0 },  /* EP */
                    386:        { roff_EQ, NULL, NULL, 0 },  /* EQ */
                    387:        { roff_line_ignore, NULL, NULL, 0 },  /* errprint */
                    388:        { roff_unsupp, NULL, NULL, 0 },  /* ev */
                    389:        { roff_unsupp, NULL, NULL, 0 },  /* evc */
                    390:        { roff_unsupp, NULL, NULL, 0 },  /* ex */
                    391:        { roff_line_ignore, NULL, NULL, 0 },  /* fallback */
                    392:        { roff_line_ignore, NULL, NULL, 0 },  /* fam */
                    393:        { roff_unsupp, NULL, NULL, 0 },  /* fc */
                    394:        { roff_unsupp, NULL, NULL, 0 },  /* fchar */
                    395:        { roff_line_ignore, NULL, NULL, 0 },  /* fcolor */
                    396:        { roff_line_ignore, NULL, NULL, 0 },  /* fdeferlig */
                    397:        { roff_line_ignore, NULL, NULL, 0 },  /* feature */
                    398:        { roff_line_ignore, NULL, NULL, 0 },  /* fkern */
                    399:        { roff_line_ignore, NULL, NULL, 0 },  /* fl */
                    400:        { roff_line_ignore, NULL, NULL, 0 },  /* flig */
                    401:        { roff_line_ignore, NULL, NULL, 0 },  /* fp */
                    402:        { roff_line_ignore, NULL, NULL, 0 },  /* fps */
                    403:        { roff_unsupp, NULL, NULL, 0 },  /* fschar */
                    404:        { roff_line_ignore, NULL, NULL, 0 },  /* fspacewidth */
                    405:        { roff_line_ignore, NULL, NULL, 0 },  /* fspecial */
                    406:        { roff_line_ignore, NULL, NULL, 0 },  /* ftr */
                    407:        { roff_line_ignore, NULL, NULL, 0 },  /* fzoom */
                    408:        { roff_line_ignore, NULL, NULL, 0 },  /* gcolor */
                    409:        { roff_line_ignore, NULL, NULL, 0 },  /* hc */
                    410:        { roff_line_ignore, NULL, NULL, 0 },  /* hcode */
                    411:        { roff_line_ignore, NULL, NULL, 0 },  /* hidechar */
                    412:        { roff_line_ignore, NULL, NULL, 0 },  /* hla */
                    413:        { roff_line_ignore, NULL, NULL, 0 },  /* hlm */
                    414:        { roff_line_ignore, NULL, NULL, 0 },  /* hpf */
                    415:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfa */
                    416:        { roff_line_ignore, NULL, NULL, 0 },  /* hpfcode */
                    417:        { roff_line_ignore, NULL, NULL, 0 },  /* hw */
                    418:        { roff_line_ignore, NULL, NULL, 0 },  /* hy */
                    419:        { roff_line_ignore, NULL, NULL, 0 },  /* hylang */
                    420:        { roff_line_ignore, NULL, NULL, 0 },  /* hylen */
                    421:        { roff_line_ignore, NULL, NULL, 0 },  /* hym */
                    422:        { roff_line_ignore, NULL, NULL, 0 },  /* hypp */
                    423:        { roff_line_ignore, NULL, NULL, 0 },  /* hys */
                    424:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* ie */
                    425:        { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT },  /* if */
                    426:        { roff_block, roff_block_text, roff_block_sub, 0 },  /* ig */
                    427:        { roff_unsupp, NULL, NULL, 0 },  /* index */
                    428:        { roff_it, NULL, NULL, 0 },  /* it */
                    429:        { roff_unsupp, NULL, NULL, 0 },  /* itc */
                    430:        { roff_line_ignore, NULL, NULL, 0 },  /* IX */
                    431:        { roff_line_ignore, NULL, NULL, 0 },  /* kern */
                    432:        { roff_line_ignore, NULL, NULL, 0 },  /* kernafter */
                    433:        { roff_line_ignore, NULL, NULL, 0 },  /* kernbefore */
                    434:        { roff_line_ignore, NULL, NULL, 0 },  /* kernpair */
                    435:        { roff_unsupp, NULL, NULL, 0 },  /* lc */
                    436:        { roff_unsupp, NULL, NULL, 0 },  /* lc_ctype */
                    437:        { roff_unsupp, NULL, NULL, 0 },  /* lds */
                    438:        { roff_unsupp, NULL, NULL, 0 },  /* length */
                    439:        { roff_line_ignore, NULL, NULL, 0 },  /* letadj */
                    440:        { roff_insec, NULL, NULL, 0 },  /* lf */
                    441:        { roff_line_ignore, NULL, NULL, 0 },  /* lg */
                    442:        { roff_line_ignore, NULL, NULL, 0 },  /* lhang */
                    443:        { roff_unsupp, NULL, NULL, 0 },  /* linetabs */
                    444:        { roff_unsupp, NULL, NULL, 0 },  /* lnr */
                    445:        { roff_unsupp, NULL, NULL, 0 },  /* lnrf */
                    446:        { roff_unsupp, NULL, NULL, 0 },  /* lpfx */
                    447:        { roff_line_ignore, NULL, NULL, 0 },  /* ls */
                    448:        { roff_unsupp, NULL, NULL, 0 },  /* lsm */
                    449:        { roff_line_ignore, NULL, NULL, 0 },  /* lt */
                    450:        { roff_line_ignore, NULL, NULL, 0 },  /* mc */
                    451:        { roff_line_ignore, NULL, NULL, 0 },  /* mediasize */
                    452:        { roff_line_ignore, NULL, NULL, 0 },  /* minss */
                    453:        { roff_line_ignore, NULL, NULL, 0 },  /* mk */
                    454:        { roff_insec, NULL, NULL, 0 },  /* mso */
                    455:        { roff_line_ignore, NULL, NULL, 0 },  /* na */
                    456:        { roff_line_ignore, NULL, NULL, 0 },  /* ne */
                    457:        { roff_line_ignore, NULL, NULL, 0 },  /* nh */
                    458:        { roff_line_ignore, NULL, NULL, 0 },  /* nhychar */
                    459:        { roff_unsupp, NULL, NULL, 0 },  /* nm */
                    460:        { roff_unsupp, NULL, NULL, 0 },  /* nn */
                    461:        { roff_unsupp, NULL, NULL, 0 },  /* nop */
                    462:        { roff_nr, NULL, NULL, 0 },  /* nr */
                    463:        { roff_unsupp, NULL, NULL, 0 },  /* nrf */
                    464:        { roff_line_ignore, NULL, NULL, 0 },  /* nroff */
                    465:        { roff_line_ignore, NULL, NULL, 0 },  /* ns */
                    466:        { roff_insec, NULL, NULL, 0 },  /* nx */
                    467:        { roff_insec, NULL, NULL, 0 },  /* open */
                    468:        { roff_insec, NULL, NULL, 0 },  /* opena */
                    469:        { roff_line_ignore, NULL, NULL, 0 },  /* os */
                    470:        { roff_unsupp, NULL, NULL, 0 },  /* output */
                    471:        { roff_line_ignore, NULL, NULL, 0 },  /* padj */
                    472:        { roff_line_ignore, NULL, NULL, 0 },  /* papersize */
                    473:        { roff_line_ignore, NULL, NULL, 0 },  /* pc */
                    474:        { roff_line_ignore, NULL, NULL, 0 },  /* pev */
                    475:        { roff_insec, NULL, NULL, 0 },  /* pi */
                    476:        { roff_unsupp, NULL, NULL, 0 },  /* PI */
                    477:        { roff_line_ignore, NULL, NULL, 0 },  /* pl */
                    478:        { roff_line_ignore, NULL, NULL, 0 },  /* pm */
                    479:        { roff_line_ignore, NULL, NULL, 0 },  /* pn */
                    480:        { roff_line_ignore, NULL, NULL, 0 },  /* pnr */
                    481:        { roff_line_ignore, NULL, NULL, 0 },  /* po */
                    482:        { roff_line_ignore, NULL, NULL, 0 },  /* ps */
                    483:        { roff_unsupp, NULL, NULL, 0 },  /* psbb */
                    484:        { roff_unsupp, NULL, NULL, 0 },  /* pshape */
                    485:        { roff_insec, NULL, NULL, 0 },  /* pso */
                    486:        { roff_line_ignore, NULL, NULL, 0 },  /* ptr */
                    487:        { roff_line_ignore, NULL, NULL, 0 },  /* pvs */
                    488:        { roff_unsupp, NULL, NULL, 0 },  /* rchar */
                    489:        { roff_line_ignore, NULL, NULL, 0 },  /* rd */
                    490:        { roff_line_ignore, NULL, NULL, 0 },  /* recursionlimit */
                    491:        { roff_unsupp, NULL, NULL, 0 },  /* return */
                    492:        { roff_unsupp, NULL, NULL, 0 },  /* rfschar */
                    493:        { roff_line_ignore, NULL, NULL, 0 },  /* rhang */
                    494:        { roff_line_ignore, NULL, NULL, 0 },  /* rj */
                    495:        { roff_rm, NULL, NULL, 0 },  /* rm */
                    496:        { roff_unsupp, NULL, NULL, 0 },  /* rn */
                    497:        { roff_unsupp, NULL, NULL, 0 },  /* rnn */
                    498:        { roff_rr, NULL, NULL, 0 },  /* rr */
                    499:        { roff_line_ignore, NULL, NULL, 0 },  /* rs */
                    500:        { roff_line_ignore, NULL, NULL, 0 },  /* rt */
                    501:        { roff_unsupp, NULL, NULL, 0 },  /* schar */
                    502:        { roff_line_ignore, NULL, NULL, 0 },  /* sentchar */
                    503:        { roff_line_ignore, NULL, NULL, 0 },  /* shc */
                    504:        { roff_unsupp, NULL, NULL, 0 },  /* shift */
                    505:        { roff_line_ignore, NULL, NULL, 0 },  /* sizes */
                    506:        { roff_so, NULL, NULL, 0 },  /* so */
                    507:        { roff_line_ignore, NULL, NULL, 0 },  /* spacewidth */
                    508:        { roff_line_ignore, NULL, NULL, 0 },  /* special */
                    509:        { roff_line_ignore, NULL, NULL, 0 },  /* spreadwarn */
                    510:        { roff_line_ignore, NULL, NULL, 0 },  /* ss */
                    511:        { roff_line_ignore, NULL, NULL, 0 },  /* sty */
                    512:        { roff_unsupp, NULL, NULL, 0 },  /* substring */
                    513:        { roff_line_ignore, NULL, NULL, 0 },  /* sv */
                    514:        { roff_insec, NULL, NULL, 0 },  /* sy */
                    515:        { roff_T_, NULL, NULL, 0 },  /* T& */
                    516:        { roff_unsupp, NULL, NULL, 0 },  /* ta */
                    517:        { roff_unsupp, NULL, NULL, 0 },  /* tc */
                    518:        { roff_TE, NULL, NULL, 0 },  /* TE */
                    519:        { roff_TH, NULL, NULL, 0 },  /* TH */
                    520:        { roff_unsupp, NULL, NULL, 0 },  /* ti */
                    521:        { roff_line_ignore, NULL, NULL, 0 },  /* tkf */
                    522:        { roff_unsupp, NULL, NULL, 0 },  /* tl */
                    523:        { roff_line_ignore, NULL, NULL, 0 },  /* tm */
                    524:        { roff_line_ignore, NULL, NULL, 0 },  /* tm1 */
                    525:        { roff_line_ignore, NULL, NULL, 0 },  /* tmc */
                    526:        { roff_tr, NULL, NULL, 0 },  /* tr */
                    527:        { roff_line_ignore, NULL, NULL, 0 },  /* track */
                    528:        { roff_line_ignore, NULL, NULL, 0 },  /* transchar */
                    529:        { roff_insec, NULL, NULL, 0 },  /* trf */
                    530:        { roff_line_ignore, NULL, NULL, 0 },  /* trimat */
                    531:        { roff_unsupp, NULL, NULL, 0 },  /* trin */
                    532:        { roff_unsupp, NULL, NULL, 0 },  /* trnt */
                    533:        { roff_line_ignore, NULL, NULL, 0 },  /* troff */
                    534:        { roff_TS, NULL, NULL, 0 },  /* TS */
                    535:        { roff_line_ignore, NULL, NULL, 0 },  /* uf */
                    536:        { roff_line_ignore, NULL, NULL, 0 },  /* ul */
                    537:        { roff_unsupp, NULL, NULL, 0 },  /* unformat */
                    538:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatch */
                    539:        { roff_line_ignore, NULL, NULL, 0 },  /* unwatchn */
                    540:        { roff_line_ignore, NULL, NULL, 0 },  /* vpt */
                    541:        { roff_line_ignore, NULL, NULL, 0 },  /* vs */
                    542:        { roff_line_ignore, NULL, NULL, 0 },  /* warn */
                    543:        { roff_line_ignore, NULL, NULL, 0 },  /* warnscale */
                    544:        { roff_line_ignore, NULL, NULL, 0 },  /* watch */
                    545:        { roff_line_ignore, NULL, NULL, 0 },  /* watchlength */
                    546:        { roff_line_ignore, NULL, NULL, 0 },  /* watchn */
                    547:        { roff_unsupp, NULL, NULL, 0 },  /* wh */
                    548:        { roff_unsupp, NULL, NULL, 0 },  /* while */
                    549:        { roff_insec, NULL, NULL, 0 },  /* write */
                    550:        { roff_insec, NULL, NULL, 0 },  /* writec */
                    551:        { roff_insec, NULL, NULL, 0 },  /* writem */
                    552:        { roff_line_ignore, NULL, NULL, 0 },  /* xflag */
                    553:        { roff_cblock, NULL, NULL, 0 },  /* . */
                    554:        { roff_userdef, NULL, NULL, 0 }
1.1       schwarze  555: };
                    556:
1.73      schwarze  557: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47      schwarze  558: const  char *const __mdoc_reserved[] = {
                    559:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    560:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    561:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    562:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73      schwarze  563:        "Dt", "Dv", "Dx", "D1",
                    564:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    565:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47      schwarze  566:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73      schwarze  567:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    568:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47      schwarze  569:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73      schwarze  570:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    571:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    572:        "Sc", "Sh", "Sm", "So", "Sq",
1.47      schwarze  573:        "Ss", "St", "Sx", "Sy",
                    574:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73      schwarze  575:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47      schwarze  576:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    577:        NULL
                    578: };
                    579:
1.73      schwarze  580: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47      schwarze  581: const  char *const __man_reserved[] = {
1.73      schwarze  582:        "AT", "B", "BI", "BR", "DT",
                    583:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    584:        "LP", "OP", "P", "PD", "PP",
                    585:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    586:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47      schwarze  587:        NULL
                    588: };
                    589:
1.37      schwarze  590: /* Array of injected predefined strings. */
                    591: #define        PREDEFS_MAX      38
                    592: static const struct predef predefs[PREDEFS_MAX] = {
                    593: #include "predefs.in"
                    594: };
                    595:
1.51      schwarze  596: static int      roffit_lines;  /* number of lines to delay */
                    597: static char    *roffit_macro;  /* nil-terminated macro line */
                    598:
1.80      schwarze  599:
1.138     schwarze  600: /* --- request table ------------------------------------------------------ */
                    601:
1.167     schwarze  602: struct ohash *
                    603: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3       schwarze  604: {
1.167     schwarze  605:        struct ohash    *htab;
                    606:        struct roffreq  *req;
                    607:        enum roff_tok    tok;
                    608:        size_t           sz;
                    609:        unsigned int     slot;
1.3       schwarze  610:
1.167     schwarze  611:        htab = mandoc_malloc(sizeof(*htab));
                    612:        mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3       schwarze  613:
1.167     schwarze  614:        for (tok = mintok; tok < maxtok; tok++) {
1.168   ! schwarze  615:                if (roff_name[tok] == NULL)
        !           616:                        continue;
1.167     schwarze  617:                sz = strlen(roff_name[tok]);
                    618:                req = mandoc_malloc(sizeof(*req) + sz + 1);
                    619:                req->tok = tok;
                    620:                memcpy(req->name, roff_name[tok], sz + 1);
                    621:                slot = ohash_qlookup(htab, req->name);
                    622:                ohash_insert(htab, slot, req);
1.3       schwarze  623:        }
1.167     schwarze  624:        return htab;
1.3       schwarze  625: }
                    626:
1.167     schwarze  627: void
                    628: roffhash_free(struct ohash *htab)
1.1       schwarze  629: {
1.167     schwarze  630:        struct roffreq  *req;
                    631:        unsigned int     slot;
1.1       schwarze  632:
1.167     schwarze  633:        if (htab == NULL)
                    634:                return;
                    635:        for (req = ohash_first(htab, &slot); req != NULL;
                    636:             req = ohash_next(htab, &slot))
                    637:                free(req);
                    638:        ohash_delete(htab);
                    639:        free(htab);
                    640: }
                    641:
                    642: enum roff_tok
                    643: roffhash_find(struct ohash *htab, const char *name, size_t sz)
                    644: {
                    645:        struct roffreq  *req;
                    646:        const char      *end;
                    647:
                    648:        if (sz) {
                    649:                end = name + sz;
                    650:                req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
                    651:        } else
                    652:                req = ohash_find(htab, ohash_qlookup(htab, name));
                    653:        return req == NULL ? TOKEN_NONE : req->tok;
1.1       schwarze  654: }
                    655:
1.138     schwarze  656: /* --- stack of request blocks -------------------------------------------- */
                    657:
1.1       schwarze  658: /*
                    659:  * Pop the current node off of the stack of roff instructions currently
                    660:  * pending.
                    661:  */
                    662: static void
                    663: roffnode_pop(struct roff *r)
                    664: {
                    665:        struct roffnode *p;
                    666:
1.2       schwarze  667:        assert(r->last);
1.80      schwarze  668:        p = r->last;
1.2       schwarze  669:
                    670:        r->last = r->last->parent;
1.16      schwarze  671:        free(p->name);
                    672:        free(p->end);
1.1       schwarze  673:        free(p);
                    674: }
                    675:
                    676: /*
                    677:  * Push a roff node onto the instruction stack.  This must later be
                    678:  * removed with roffnode_pop().
                    679:  */
1.11      schwarze  680: static void
1.166     schwarze  681: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16      schwarze  682:                int line, int col)
1.1       schwarze  683: {
                    684:        struct roffnode *p;
                    685:
1.11      schwarze  686:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  687:        p->tok = tok;
1.16      schwarze  688:        if (name)
                    689:                p->name = mandoc_strdup(name);
1.1       schwarze  690:        p->parent = r->last;
                    691:        p->line = line;
                    692:        p->col = col;
1.71      schwarze  693:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  694:
                    695:        r->last = p;
                    696: }
                    697:
1.138     schwarze  698: /* --- roff parser state data management ---------------------------------- */
                    699:
1.1       schwarze  700: static void
                    701: roff_free1(struct roff *r)
                    702: {
1.49      schwarze  703:        struct tbl_node *tbl;
1.32      schwarze  704:        struct eqn_node *e;
1.42      schwarze  705:        int              i;
1.27      schwarze  706:
1.49      schwarze  707:        while (NULL != (tbl = r->first_tbl)) {
                    708:                r->first_tbl = tbl->next;
                    709:                tbl_free(tbl);
1.27      schwarze  710:        }
                    711:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  712:
1.32      schwarze  713:        while (NULL != (e = r->first_eqn)) {
                    714:                r->first_eqn = e->next;
                    715:                eqn_free(e);
                    716:        }
                    717:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    718:
1.1       schwarze  719:        while (r->last)
                    720:                roffnode_pop(r);
1.27      schwarze  721:
1.96      schwarze  722:        free (r->rstack);
                    723:        r->rstack = NULL;
                    724:        r->rstacksz = 0;
                    725:        r->rstackpos = -1;
                    726:
                    727:        roff_freereg(r->regtab);
                    728:        r->regtab = NULL;
                    729:
1.42      schwarze  730:        roff_freestr(r->strtab);
                    731:        roff_freestr(r->xmbtab);
                    732:        r->strtab = r->xmbtab = NULL;
                    733:
                    734:        if (r->xtab)
                    735:                for (i = 0; i < 128; i++)
                    736:                        free(r->xtab[i].p);
                    737:        free(r->xtab);
                    738:        r->xtab = NULL;
1.1       schwarze  739: }
                    740:
                    741: void
                    742: roff_reset(struct roff *r)
                    743: {
                    744:        roff_free1(r);
1.99      schwarze  745:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48      schwarze  746:        r->control = 0;
1.1       schwarze  747: }
                    748:
                    749: void
                    750: roff_free(struct roff *r)
                    751: {
                    752:        roff_free1(r);
1.167     schwarze  753:        roffhash_free(r->reqtab);
1.1       schwarze  754:        free(r);
                    755: }
                    756:
                    757: struct roff *
1.151     schwarze  758: roff_alloc(struct mparse *parse, int options)
1.1       schwarze  759: {
                    760:        struct roff     *r;
                    761:
1.11      schwarze  762:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  763:        r->parse = parse;
1.167     schwarze  764:        r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72      schwarze  765:        r->options = options;
1.99      schwarze  766:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  767:        r->rstackpos = -1;
1.149     schwarze  768:        return r;
1.137     schwarze  769: }
                    770:
1.138     schwarze  771: /* --- syntax tree state data management ---------------------------------- */
                    772:
1.137     schwarze  773: static void
                    774: roff_man_free1(struct roff_man *man)
                    775: {
                    776:
1.138     schwarze  777:        if (man->first != NULL)
                    778:                roff_node_delete(man, man->first);
1.137     schwarze  779:        free(man->meta.msec);
                    780:        free(man->meta.vol);
                    781:        free(man->meta.os);
                    782:        free(man->meta.arch);
                    783:        free(man->meta.title);
                    784:        free(man->meta.name);
                    785:        free(man->meta.date);
                    786: }
                    787:
                    788: static void
                    789: roff_man_alloc1(struct roff_man *man)
                    790: {
                    791:
                    792:        memset(&man->meta, 0, sizeof(man->meta));
                    793:        man->first = mandoc_calloc(1, sizeof(*man->first));
                    794:        man->first->type = ROFFT_ROOT;
                    795:        man->last = man->first;
                    796:        man->last_es = NULL;
                    797:        man->flags = 0;
                    798:        man->macroset = MACROSET_NONE;
                    799:        man->lastsec = man->lastnamed = SEC_NONE;
                    800:        man->next = ROFF_NEXT_CHILD;
                    801: }
                    802:
                    803: void
                    804: roff_man_reset(struct roff_man *man)
                    805: {
                    806:
                    807:        roff_man_free1(man);
                    808:        roff_man_alloc1(man);
                    809: }
                    810:
                    811: void
                    812: roff_man_free(struct roff_man *man)
                    813: {
                    814:
                    815:        roff_man_free1(man);
                    816:        free(man);
                    817: }
                    818:
                    819: struct roff_man *
                    820: roff_man_alloc(struct roff *roff, struct mparse *parse,
                    821:        const char *defos, int quick)
                    822: {
                    823:        struct roff_man *man;
                    824:
                    825:        man = mandoc_calloc(1, sizeof(*man));
                    826:        man->parse = parse;
                    827:        man->roff = roff;
                    828:        man->defos = defos;
                    829:        man->quick = quick;
                    830:        roff_man_alloc1(man);
1.168   ! schwarze  831:        roff->man = man;
1.149     schwarze  832:        return man;
1.1       schwarze  833: }
                    834:
1.138     schwarze  835: /* --- syntax tree handling ----------------------------------------------- */
                    836:
                    837: struct roff_node *
                    838: roff_node_alloc(struct roff_man *man, int line, int pos,
                    839:        enum roff_type type, int tok)
                    840: {
                    841:        struct roff_node        *n;
                    842:
                    843:        n = mandoc_calloc(1, sizeof(*n));
                    844:        n->line = line;
                    845:        n->pos = pos;
                    846:        n->tok = tok;
                    847:        n->type = type;
                    848:        n->sec = man->lastsec;
                    849:
                    850:        if (man->flags & MDOC_SYNOPSIS)
1.157     schwarze  851:                n->flags |= NODE_SYNPRETTY;
1.138     schwarze  852:        else
1.157     schwarze  853:                n->flags &= ~NODE_SYNPRETTY;
1.138     schwarze  854:        if (man->flags & MDOC_NEWLINE)
1.157     schwarze  855:                n->flags |= NODE_LINE;
1.138     schwarze  856:        man->flags &= ~MDOC_NEWLINE;
                    857:
1.149     schwarze  858:        return n;
1.138     schwarze  859: }
                    860:
                    861: void
                    862: roff_node_append(struct roff_man *man, struct roff_node *n)
                    863: {
                    864:
                    865:        switch (man->next) {
                    866:        case ROFF_NEXT_SIBLING:
1.153     schwarze  867:                if (man->last->next != NULL) {
                    868:                        n->next = man->last->next;
                    869:                        man->last->next->prev = n;
                    870:                } else
                    871:                        man->last->parent->last = n;
1.138     schwarze  872:                man->last->next = n;
                    873:                n->prev = man->last;
                    874:                n->parent = man->last->parent;
                    875:                break;
                    876:        case ROFF_NEXT_CHILD:
1.159     schwarze  877:                if (man->last->child != NULL) {
                    878:                        n->next = man->last->child;
                    879:                        man->last->child->prev = n;
                    880:                } else
                    881:                        man->last->last = n;
1.138     schwarze  882:                man->last->child = n;
                    883:                n->parent = man->last;
                    884:                break;
                    885:        default:
                    886:                abort();
                    887:        }
1.154     schwarze  888:        man->last = n;
1.138     schwarze  889:
                    890:        switch (n->type) {
                    891:        case ROFFT_HEAD:
                    892:                n->parent->head = n;
                    893:                break;
                    894:        case ROFFT_BODY:
1.154     schwarze  895:                if (n->end != ENDBODY_NOT)
                    896:                        return;
1.138     schwarze  897:                n->parent->body = n;
                    898:                break;
                    899:        case ROFFT_TAIL:
                    900:                n->parent->tail = n;
                    901:                break;
                    902:        default:
1.154     schwarze  903:                return;
1.138     schwarze  904:        }
1.154     schwarze  905:
                    906:        /*
                    907:         * Copy over the normalised-data pointer of our parent.  Not
                    908:         * everybody has one, but copying a null pointer is fine.
                    909:         */
                    910:
                    911:        n->norm = n->parent->norm;
                    912:        assert(n->parent->type == ROFFT_BLOCK);
1.138     schwarze  913: }
                    914:
1.139     schwarze  915: void
                    916: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
                    917: {
                    918:        struct roff_node        *n;
                    919:
                    920:        n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
                    921:        n->string = roff_strdup(man->roff, word);
                    922:        roff_node_append(man, n);
1.158     schwarze  923:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze  924:        man->next = ROFF_NEXT_SIBLING;
                    925: }
                    926:
                    927: void
                    928: roff_word_append(struct roff_man *man, const char *word)
                    929: {
                    930:        struct roff_node        *n;
                    931:        char                    *addstr, *newstr;
                    932:
                    933:        n = man->last;
                    934:        addstr = roff_strdup(man->roff, word);
                    935:        mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
                    936:        free(addstr);
                    937:        free(n->string);
                    938:        n->string = newstr;
                    939:        man->next = ROFF_NEXT_SIBLING;
1.140     schwarze  940: }
                    941:
                    942: void
                    943: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
                    944: {
                    945:        struct roff_node        *n;
                    946:
                    947:        n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
                    948:        roff_node_append(man, n);
                    949:        man->next = ROFF_NEXT_CHILD;
                    950: }
                    951:
                    952: struct roff_node *
                    953: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
                    954: {
                    955:        struct roff_node        *n;
                    956:
                    957:        n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
                    958:        roff_node_append(man, n);
                    959:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  960:        return n;
1.139     schwarze  961: }
                    962:
1.138     schwarze  963: struct roff_node *
                    964: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
                    965: {
                    966:        struct roff_node        *n;
                    967:
                    968:        n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
                    969:        roff_node_append(man, n);
                    970:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  971:        return n;
1.138     schwarze  972: }
                    973:
                    974: struct roff_node *
                    975: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
                    976: {
                    977:        struct roff_node        *n;
                    978:
                    979:        n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
                    980:        roff_node_append(man, n);
                    981:        man->next = ROFF_NEXT_CHILD;
1.149     schwarze  982:        return n;
1.139     schwarze  983: }
                    984:
                    985: void
                    986: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
                    987: {
                    988:        struct roff_node        *n;
                    989:
                    990:        n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
                    991:        n->eqn = eqn;
                    992:        if (eqn->ln > man->last->line)
1.157     schwarze  993:                n->flags |= NODE_LINE;
1.139     schwarze  994:        roff_node_append(man, n);
                    995:        man->next = ROFF_NEXT_SIBLING;
                    996: }
                    997:
                    998: void
                    999: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
                   1000: {
                   1001:        struct roff_node        *n;
                   1002:
                   1003:        if (man->macroset == MACROSET_MAN)
                   1004:                man_breakscope(man, TOKEN_NONE);
                   1005:        n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
                   1006:        n->span = tbl;
                   1007:        roff_node_append(man, n);
1.158     schwarze 1008:        n->flags |= NODE_VALID | NODE_ENDED;
1.139     schwarze 1009:        man->next = ROFF_NEXT_SIBLING;
1.138     schwarze 1010: }
                   1011:
                   1012: void
                   1013: roff_node_unlink(struct roff_man *man, struct roff_node *n)
                   1014: {
                   1015:
                   1016:        /* Adjust siblings. */
                   1017:
                   1018:        if (n->prev)
                   1019:                n->prev->next = n->next;
                   1020:        if (n->next)
                   1021:                n->next->prev = n->prev;
                   1022:
                   1023:        /* Adjust parent. */
                   1024:
                   1025:        if (n->parent != NULL) {
                   1026:                if (n->parent->child == n)
                   1027:                        n->parent->child = n->next;
                   1028:                if (n->parent->last == n)
                   1029:                        n->parent->last = n->prev;
                   1030:        }
                   1031:
                   1032:        /* Adjust parse point. */
                   1033:
                   1034:        if (man == NULL)
                   1035:                return;
                   1036:        if (man->last == n) {
                   1037:                if (n->prev == NULL) {
                   1038:                        man->last = n->parent;
                   1039:                        man->next = ROFF_NEXT_CHILD;
                   1040:                } else {
                   1041:                        man->last = n->prev;
                   1042:                        man->next = ROFF_NEXT_SIBLING;
                   1043:                }
                   1044:        }
                   1045:        if (man->first == n)
                   1046:                man->first = NULL;
                   1047: }
                   1048:
                   1049: void
                   1050: roff_node_free(struct roff_node *n)
                   1051: {
                   1052:
                   1053:        if (n->args != NULL)
                   1054:                mdoc_argv_free(n->args);
                   1055:        if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
                   1056:                free(n->norm);
                   1057:        free(n->string);
                   1058:        free(n);
                   1059: }
                   1060:
                   1061: void
                   1062: roff_node_delete(struct roff_man *man, struct roff_node *n)
                   1063: {
                   1064:
                   1065:        while (n->child != NULL)
                   1066:                roff_node_delete(man, n->child);
                   1067:        roff_node_unlink(man, n);
                   1068:        roff_node_free(n);
1.141     schwarze 1069: }
                   1070:
                   1071: void
                   1072: deroff(char **dest, const struct roff_node *n)
                   1073: {
                   1074:        char    *cp;
                   1075:        size_t   sz;
                   1076:
                   1077:        if (n->type != ROFFT_TEXT) {
                   1078:                for (n = n->child; n != NULL; n = n->next)
                   1079:                        deroff(dest, n);
                   1080:                return;
                   1081:        }
                   1082:
1.160     schwarze 1083:        /* Skip leading whitespace. */
1.141     schwarze 1084:
1.160     schwarze 1085:        for (cp = n->string; *cp != '\0'; cp++) {
1.161     schwarze 1086:                if (cp[0] == '\\' && cp[1] != '\0' &&
                   1087:                    strchr(" %&0^|~", cp[1]) != NULL)
1.141     schwarze 1088:                        cp++;
1.160     schwarze 1089:                else if ( ! isspace((unsigned char)*cp))
1.141     schwarze 1090:                        break;
                   1091:        }
                   1092:
1.161     schwarze 1093:        /* Skip trailing backslash. */
                   1094:
                   1095:        sz = strlen(cp);
1.162     schwarze 1096:        if (sz > 0 && cp[sz - 1] == '\\')
1.161     schwarze 1097:                sz--;
                   1098:
1.141     schwarze 1099:        /* Skip trailing whitespace. */
                   1100:
1.161     schwarze 1101:        for (; sz; sz--)
1.141     schwarze 1102:                if ( ! isspace((unsigned char)cp[sz-1]))
                   1103:                        break;
                   1104:
                   1105:        /* Skip empty strings. */
                   1106:
                   1107:        if (sz == 0)
                   1108:                return;
                   1109:
                   1110:        if (*dest == NULL) {
                   1111:                *dest = mandoc_strndup(cp, sz);
                   1112:                return;
                   1113:        }
                   1114:
                   1115:        mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
                   1116:        free(*dest);
                   1117:        *dest = cp;
1.138     schwarze 1118: }
                   1119:
                   1120: /* --- main functions of the roff parser ---------------------------------- */
                   1121:
1.8       schwarze 1122: /*
1.79      schwarze 1123:  * In the current line, expand escape sequences that tend to get
                   1124:  * used in numerical expressions and conditional requests.
                   1125:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze 1126:  */
1.45      schwarze 1127: static enum rofferr
1.110     schwarze 1128: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze 1129: {
1.81      schwarze 1130:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze 1131:        const char      *start; /* start of the string to process */
1.82      schwarze 1132:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze 1133:        const char      *stnam; /* start of the name, after "[(*" */
                   1134:        const char      *cp;    /* end of the name, e.g. before ']' */
                   1135:        const char      *res;   /* the string to be substituted */
1.110     schwarze 1136:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze 1137:        size_t           maxl;  /* expected length of the escape name */
                   1138:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze 1139:        enum mandoc_esc  esc;   /* type of the escape sequence */
                   1140:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze 1141:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze 1142:        int              npos;  /* position in numeric expression */
1.91      schwarze 1143:        int              arg_complete; /* argument not interrupted by eol */
1.79      schwarze 1144:        char             term;  /* character terminating the escape */
1.8       schwarze 1145:
1.43      schwarze 1146:        expand_count = 0;
1.110     schwarze 1147:        start = buf->buf + pos;
1.78      schwarze 1148:        stesc = strchr(start, '\0') - 1;
                   1149:        while (stesc-- > start) {
1.43      schwarze 1150:
1.78      schwarze 1151:                /* Search backwards for the next backslash. */
                   1152:
1.110     schwarze 1153:                if (*stesc != '\\')
1.78      schwarze 1154:                        continue;
                   1155:
                   1156:                /* If it is escaped, skip it. */
                   1157:
                   1158:                for (cp = stesc - 1; cp >= start; cp--)
1.110     schwarze 1159:                        if (*cp != '\\')
1.78      schwarze 1160:                                break;
                   1161:
1.110     schwarze 1162:                if ((stesc - cp) % 2 == 0) {
1.82      schwarze 1163:                        stesc = (char *)cp;
1.78      schwarze 1164:                        continue;
                   1165:                }
1.23      schwarze 1166:
1.79      schwarze 1167:                /* Decide whether to expand or to check only. */
1.23      schwarze 1168:
1.79      schwarze 1169:                term = '\0';
1.78      schwarze 1170:                cp = stesc + 1;
1.53      schwarze 1171:                switch (*cp) {
1.80      schwarze 1172:                case '*':
1.53      schwarze 1173:                        res = NULL;
                   1174:                        break;
1.80      schwarze 1175:                case 'B':
                   1176:                case 'w':
1.79      schwarze 1177:                        term = cp[1];
                   1178:                        /* FALLTHROUGH */
1.80      schwarze 1179:                case 'n':
1.53      schwarze 1180:                        res = ubuf;
                   1181:                        break;
                   1182:                default:
1.109     schwarze 1183:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1184:                        if (esc == ESCAPE_ERROR ||
                   1185:                            (esc == ESCAPE_SPECIAL &&
1.151     schwarze 1186:                             mchars_spec2cp(stnam, inaml) < 0))
1.92      schwarze 1187:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1188:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1189:                                    "%.*s", (int)(cp - stesc), stesc);
1.78      schwarze 1190:                        continue;
1.42      schwarze 1191:                }
                   1192:
1.78      schwarze 1193:                if (EXPAND_LIMIT < ++expand_count) {
                   1194:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1195:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1196:                        return ROFF_IGN;
1.78      schwarze 1197:                }
1.23      schwarze 1198:
                   1199:                /*
                   1200:                 * The third character decides the length
1.53      schwarze 1201:                 * of the name of the string or register.
1.23      schwarze 1202:                 * Save a pointer to the name.
                   1203:                 */
                   1204:
1.110     schwarze 1205:                if (term == '\0') {
1.79      schwarze 1206:                        switch (*++cp) {
1.80      schwarze 1207:                        case '\0':
1.79      schwarze 1208:                                maxl = 0;
                   1209:                                break;
1.80      schwarze 1210:                        case '(':
1.79      schwarze 1211:                                cp++;
                   1212:                                maxl = 2;
                   1213:                                break;
1.80      schwarze 1214:                        case '[':
1.79      schwarze 1215:                                cp++;
                   1216:                                term = ']';
                   1217:                                maxl = 0;
                   1218:                                break;
                   1219:                        default:
                   1220:                                maxl = 1;
                   1221:                                break;
                   1222:                        }
                   1223:                } else {
                   1224:                        cp += 2;
1.8       schwarze 1225:                        maxl = 0;
                   1226:                }
1.23      schwarze 1227:                stnam = cp;
1.8       schwarze 1228:
1.23      schwarze 1229:                /* Advance to the end of the name. */
1.8       schwarze 1230:
1.125     schwarze 1231:                naml = 0;
1.91      schwarze 1232:                arg_complete = 1;
1.125     schwarze 1233:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1234:                        if (*cp == '\0') {
1.92      schwarze 1235:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1236:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1237:                                arg_complete = 0;
1.79      schwarze 1238:                                break;
1.42      schwarze 1239:                        }
1.110     schwarze 1240:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1241:                                cp++;
1.125     schwarze 1242:                                break;
                   1243:                        }
                   1244:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1245:                                naml++;
                   1246:                                continue;
                   1247:                        }
                   1248:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1249:                        case ESCAPE_SPECIAL:
                   1250:                        case ESCAPE_UNICODE:
                   1251:                        case ESCAPE_NUMBERED:
                   1252:                        case ESCAPE_OVERSTRIKE:
                   1253:                                naml++;
                   1254:                                break;
                   1255:                        default:
1.8       schwarze 1256:                                break;
1.79      schwarze 1257:                        }
1.8       schwarze 1258:                }
                   1259:
1.23      schwarze 1260:                /*
                   1261:                 * Retrieve the replacement string; if it is
                   1262:                 * undefined, resume searching for escapes.
                   1263:                 */
                   1264:
1.79      schwarze 1265:                switch (stesc[1]) {
1.80      schwarze 1266:                case '*':
1.91      schwarze 1267:                        if (arg_complete)
                   1268:                                res = roff_getstrn(r, stnam, naml);
1.79      schwarze 1269:                        break;
1.80      schwarze 1270:                case 'B':
1.79      schwarze 1271:                        npos = 0;
1.91      schwarze 1272:                        ubuf[0] = arg_complete &&
1.133     schwarze 1273:                            roff_evalnum(r, ln, stnam, &npos,
                   1274:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1275:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1276:                        ubuf[1] = '\0';
                   1277:                        break;
1.80      schwarze 1278:                case 'n':
1.91      schwarze 1279:                        if (arg_complete)
                   1280:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1281:                                    roff_getregn(r, stnam, naml));
                   1282:                        else
                   1283:                                ubuf[0] = '\0';
1.79      schwarze 1284:                        break;
1.80      schwarze 1285:                case 'w':
1.91      schwarze 1286:                        /* use even incomplete args */
1.81      schwarze 1287:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1288:                            24 * (int)naml);
                   1289:                        break;
                   1290:                }
1.8       schwarze 1291:
1.110     schwarze 1292:                if (res == NULL) {
1.92      schwarze 1293:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1294:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1295:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1296:                        res = "";
1.118     schwarze 1297:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1298:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1299:                            ln, (int)(stesc - buf->buf), NULL);
1.149     schwarze 1300:                        return ROFF_IGN;
1.8       schwarze 1301:                }
                   1302:
1.23      schwarze 1303:                /* Replace the escape sequence by the string. */
                   1304:
1.82      schwarze 1305:                *stesc = '\0';
1.110     schwarze 1306:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1307:                    buf->buf, res, cp) + 1;
1.8       schwarze 1308:
1.78      schwarze 1309:                /* Prepare for the next replacement. */
1.8       schwarze 1310:
1.78      schwarze 1311:                start = nbuf + pos;
1.110     schwarze 1312:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1313:                free(buf->buf);
                   1314:                buf->buf = nbuf;
1.42      schwarze 1315:        }
1.149     schwarze 1316:        return ROFF_CONT;
1.42      schwarze 1317: }
                   1318:
                   1319: /*
1.147     schwarze 1320:  * Process text streams.
1.42      schwarze 1321:  */
                   1322: static enum rofferr
1.110     schwarze 1323: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42      schwarze 1324: {
                   1325:        size_t           sz;
                   1326:        const char      *start;
1.51      schwarze 1327:        char            *p;
                   1328:        int              isz;
1.42      schwarze 1329:        enum mandoc_esc  esc;
                   1330:
1.147     schwarze 1331:        /* Spring the input line trap. */
                   1332:
                   1333:        if (roffit_lines == 1) {
                   1334:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1335:                free(buf->buf);
                   1336:                buf->buf = p;
                   1337:                buf->sz = isz + 1;
                   1338:                *offs = 0;
                   1339:                free(roffit_macro);
                   1340:                roffit_lines = 0;
1.149     schwarze 1341:                return ROFF_REPARSE;
1.147     schwarze 1342:        } else if (roffit_lines > 1)
                   1343:                --roffit_lines;
                   1344:
                   1345:        /* Convert all breakable hyphens into ASCII_HYPH. */
                   1346:
1.110     schwarze 1347:        start = p = buf->buf + pos;
1.42      schwarze 1348:
1.110     schwarze 1349:        while (*p != '\0') {
1.42      schwarze 1350:                sz = strcspn(p, "-\\");
                   1351:                p += sz;
                   1352:
1.110     schwarze 1353:                if (*p == '\0')
1.42      schwarze 1354:                        break;
                   1355:
1.110     schwarze 1356:                if (*p == '\\') {
1.42      schwarze 1357:                        /* Skip over escapes. */
                   1358:                        p++;
1.62      schwarze 1359:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1360:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1361:                                break;
1.136     schwarze 1362:                        while (*p == '-')
                   1363:                                p++;
1.42      schwarze 1364:                        continue;
                   1365:                } else if (p == start) {
                   1366:                        p++;
                   1367:                        continue;
                   1368:                }
                   1369:
1.44      schwarze 1370:                if (isalpha((unsigned char)p[-1]) &&
                   1371:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1372:                        *p = ASCII_HYPH;
                   1373:                p++;
1.8       schwarze 1374:        }
1.149     schwarze 1375:        return ROFF_CONT;
1.8       schwarze 1376: }
                   1377:
1.1       schwarze 1378: enum rofferr
1.110     schwarze 1379: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1380: {
1.166     schwarze 1381:        enum roff_tok    t;
1.27      schwarze 1382:        enum rofferr     e;
1.110     schwarze 1383:        int              pos;   /* parse point */
1.115     schwarze 1384:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1385:        int              ppos;  /* original offset in buf->buf */
                   1386:        int              ctl;   /* macro line (boolean) */
                   1387:
                   1388:        ppos = pos = *offs;
1.1       schwarze 1389:
1.102     schwarze 1390:        /* Handle in-line equation delimiters. */
                   1391:
1.108     schwarze 1392:        if (r->tbl == NULL &&
                   1393:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1394:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1395:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1396:                if (e == ROFF_REPARSE)
1.149     schwarze 1397:                        return e;
1.102     schwarze 1398:                assert(e == ROFF_CONT);
                   1399:        }
                   1400:
                   1401:        /* Expand some escape sequences. */
1.8       schwarze 1402:
1.110     schwarze 1403:        e = roff_res(r, buf, ln, pos);
                   1404:        if (e == ROFF_IGN)
1.149     schwarze 1405:                return e;
1.110     schwarze 1406:        assert(e == ROFF_CONT);
1.8       schwarze 1407:
1.110     schwarze 1408:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1409:
1.8       schwarze 1410:        /*
1.2       schwarze 1411:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1412:         * text through the macro's filter.
                   1413:         * Equations process all content themselves.
                   1414:         * Tables process almost all content themselves, but we want
                   1415:         * to warn about macros before passing it there.
1.2       schwarze 1416:         */
                   1417:
1.124     schwarze 1418:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1419:                t = r->last->tok;
1.110     schwarze 1420:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166     schwarze 1421:                if (e == ROFF_IGN)
1.149     schwarze 1422:                        return e;
1.166     schwarze 1423:                assert(e == ROFF_CONT);
1.54      schwarze 1424:        }
1.124     schwarze 1425:        if (r->eqn != NULL)
1.149     schwarze 1426:                return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.124     schwarze 1427:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.149     schwarze 1428:                return tbl_read(r->tbl, ln, buf->buf, ppos);
1.124     schwarze 1429:        if ( ! ctl)
1.149     schwarze 1430:                return roff_parsetext(buf, pos, offs);
1.100     schwarze 1431:
                   1432:        /* Skip empty request lines. */
                   1433:
1.110     schwarze 1434:        if (buf->buf[pos] == '"') {
1.100     schwarze 1435:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1436:                    ln, pos, NULL);
1.149     schwarze 1437:                return ROFF_IGN;
1.110     schwarze 1438:        } else if (buf->buf[pos] == '\0')
1.149     schwarze 1439:                return ROFF_IGN;
1.2       schwarze 1440:
                   1441:        /*
                   1442:         * If a scope is open, go to the child handler for that macro,
                   1443:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1444:         * Don't do so if an equation is open.
1.2       schwarze 1445:         */
                   1446:
                   1447:        if (r->last) {
1.1       schwarze 1448:                t = r->last->tok;
1.149     schwarze 1449:                return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1450:        }
                   1451:
1.115     schwarze 1452:        /* No scope is open.  This is a new request or macro. */
                   1453:
                   1454:        spos = pos;
                   1455:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1456:
                   1457:        /* Tables ignore most macros. */
                   1458:
1.166     schwarze 1459:        if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS)) {
1.115     schwarze 1460:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1461:                    ln, pos, buf->buf + spos);
1.129     schwarze 1462:                if (t == ROFF_TS)
1.149     schwarze 1463:                        return ROFF_IGN;
1.129     schwarze 1464:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1465:                        pos++;
1.163     schwarze 1466:                while (buf->buf[pos] == ' ')
1.129     schwarze 1467:                        pos++;
1.149     schwarze 1468:                return tbl_read(r->tbl, ln, buf->buf, pos);
1.115     schwarze 1469:        }
                   1470:
1.2       schwarze 1471:        /*
1.115     schwarze 1472:         * This is neither a roff request nor a user-defined macro.
                   1473:         * Let the standard macro set parsers handle it.
1.2       schwarze 1474:         */
                   1475:
1.166     schwarze 1476:        if (t == TOKEN_NONE)
1.149     schwarze 1477:                return ROFF_CONT;
1.115     schwarze 1478:
                   1479:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1480:
1.168   ! schwarze 1481:        return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2       schwarze 1482: }
                   1483:
1.27      schwarze 1484: void
1.2       schwarze 1485: roff_endparse(struct roff *r)
                   1486: {
1.1       schwarze 1487:
1.27      schwarze 1488:        if (r->last)
1.94      schwarze 1489:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1490:                    r->last->line, r->last->col,
1.167     schwarze 1491:                    roff_name[r->last->tok]);
1.27      schwarze 1492:
1.32      schwarze 1493:        if (r->eqn) {
1.94      schwarze 1494:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1495:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41      schwarze 1496:                eqn_end(&r->eqn);
1.32      schwarze 1497:        }
                   1498:
1.27      schwarze 1499:        if (r->tbl) {
1.94      schwarze 1500:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1501:                    r->tbl->line, r->tbl->pos, "TS");
1.41      schwarze 1502:                tbl_end(&r->tbl);
1.27      schwarze 1503:        }
1.1       schwarze 1504: }
                   1505:
                   1506: /*
                   1507:  * Parse a roff node's type from the input buffer.  This must be in the
                   1508:  * form of ".foo xxx" in the usual way.
                   1509:  */
1.166     schwarze 1510: static enum roff_tok
1.87      schwarze 1511: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1512: {
1.87      schwarze 1513:        char            *cp;
1.16      schwarze 1514:        const char      *mac;
                   1515:        size_t           maclen;
1.166     schwarze 1516:        enum roff_tok    t;
1.1       schwarze 1517:
1.87      schwarze 1518:        cp = buf + *pos;
                   1519:
                   1520:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166     schwarze 1521:                return TOKEN_NONE;
1.1       schwarze 1522:
1.87      schwarze 1523:        mac = cp;
                   1524:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1525:
1.16      schwarze 1526:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.167     schwarze 1527:            ? ROFF_USERDEF : roffhash_find(r->reqtab, mac, maclen);
1.1       schwarze 1528:
1.166     schwarze 1529:        if (t != TOKEN_NONE)
1.87      schwarze 1530:                *pos = cp - buf;
1.1       schwarze 1531:
1.149     schwarze 1532:        return t;
1.1       schwarze 1533: }
                   1534:
1.138     schwarze 1535: /* --- handling of request blocks ----------------------------------------- */
                   1536:
1.1       schwarze 1537: static enum rofferr
1.2       schwarze 1538: roff_cblock(ROFF_ARGS)
1.1       schwarze 1539: {
                   1540:
1.2       schwarze 1541:        /*
                   1542:         * A block-close `..' should only be invoked as a child of an
                   1543:         * ignore macro, otherwise raise a warning and just ignore it.
                   1544:         */
                   1545:
1.110     schwarze 1546:        if (r->last == NULL) {
1.94      schwarze 1547:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1548:                    ln, ppos, "..");
1.149     schwarze 1549:                return ROFF_IGN;
1.2       schwarze 1550:        }
1.1       schwarze 1551:
1.2       schwarze 1552:        switch (r->last->tok) {
1.80      schwarze 1553:        case ROFF_am:
1.93      schwarze 1554:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80      schwarze 1555:        case ROFF_ami:
                   1556:        case ROFF_de:
1.23      schwarze 1557:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80      schwarze 1558:        case ROFF_dei:
                   1559:        case ROFF_ig:
1.2       schwarze 1560:                break;
                   1561:        default:
1.94      schwarze 1562:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1563:                    ln, ppos, "..");
1.149     schwarze 1564:                return ROFF_IGN;
1.2       schwarze 1565:        }
                   1566:
1.110     schwarze 1567:        if (buf->buf[pos] != '\0')
1.90      schwarze 1568:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1569:                    ".. %s", buf->buf + pos);
1.2       schwarze 1570:
                   1571:        roffnode_pop(r);
                   1572:        roffnode_cleanscope(r);
1.149     schwarze 1573:        return ROFF_IGN;
1.2       schwarze 1574:
                   1575: }
1.1       schwarze 1576:
1.2       schwarze 1577: static void
                   1578: roffnode_cleanscope(struct roff *r)
                   1579: {
1.1       schwarze 1580:
1.2       schwarze 1581:        while (r->last) {
1.46      schwarze 1582:                if (--r->last->endspan != 0)
1.2       schwarze 1583:                        break;
                   1584:                roffnode_pop(r);
                   1585:        }
                   1586: }
1.1       schwarze 1587:
1.68      schwarze 1588: static void
                   1589: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1590: {
1.1       schwarze 1591:
1.2       schwarze 1592:        if (NULL == r->last) {
1.94      schwarze 1593:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1594:                    ln, ppos, "\\}");
1.68      schwarze 1595:                return;
1.2       schwarze 1596:        }
1.1       schwarze 1597:
1.2       schwarze 1598:        switch (r->last->tok) {
1.80      schwarze 1599:        case ROFF_el:
                   1600:        case ROFF_ie:
                   1601:        case ROFF_if:
1.2       schwarze 1602:                break;
                   1603:        default:
1.94      schwarze 1604:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1605:                    ln, ppos, "\\}");
1.68      schwarze 1606:                return;
1.2       schwarze 1607:        }
1.1       schwarze 1608:
1.2       schwarze 1609:        if (r->last->endspan > -1) {
1.94      schwarze 1610:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1611:                    ln, ppos, "\\}");
1.68      schwarze 1612:                return;
1.2       schwarze 1613:        }
                   1614:
                   1615:        roffnode_pop(r);
                   1616:        roffnode_cleanscope(r);
1.68      schwarze 1617:        return;
1.1       schwarze 1618: }
                   1619:
                   1620: static enum rofferr
1.2       schwarze 1621: roff_block(ROFF_ARGS)
1.1       schwarze 1622: {
1.93      schwarze 1623:        const char      *name;
                   1624:        char            *iname, *cp;
1.86      schwarze 1625:        size_t           namesz;
1.16      schwarze 1626:
1.93      schwarze 1627:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1628:
1.110     schwarze 1629:        if (tok == ROFF_de1)
1.93      schwarze 1630:                tok = ROFF_de;
1.123     schwarze 1631:        else if (tok == ROFF_dei1)
                   1632:                tok = ROFF_dei;
1.110     schwarze 1633:        else if (tok == ROFF_am1)
1.93      schwarze 1634:                tok = ROFF_am;
1.123     schwarze 1635:        else if (tok == ROFF_ami1)
                   1636:                tok = ROFF_ami;
1.93      schwarze 1637:
                   1638:        /* Parse the macro name argument. */
                   1639:
1.110     schwarze 1640:        cp = buf->buf + pos;
                   1641:        if (tok == ROFF_ig) {
1.93      schwarze 1642:                iname = NULL;
                   1643:                namesz = 0;
                   1644:        } else {
                   1645:                iname = cp;
                   1646:                namesz = roff_getname(r, &cp, ln, ppos);
                   1647:                iname[namesz] = '\0';
                   1648:        }
1.22      schwarze 1649:
1.93      schwarze 1650:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1651:
1.110     schwarze 1652:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1653:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1654:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1655:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1656:                            "%.*s", (int)namesz, iname);
                   1657:                        namesz = 0;
                   1658:                } else
                   1659:                        namesz = strlen(name);
                   1660:        } else
                   1661:                name = iname;
1.22      schwarze 1662:
1.110     schwarze 1663:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1664:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167     schwarze 1665:                    ln, ppos, roff_name[tok]);
1.149     schwarze 1666:                return ROFF_IGN;
1.93      schwarze 1667:        }
1.2       schwarze 1668:
1.16      schwarze 1669:        roffnode_push(r, tok, name, ln, ppos);
                   1670:
                   1671:        /*
                   1672:         * At the beginning of a `de' macro, clear the existing string
                   1673:         * with the same name, if there is one.  New content will be
1.66      schwarze 1674:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1675:         */
1.22      schwarze 1676:
1.110     schwarze 1677:        if (tok == ROFF_de || tok == ROFF_dei)
1.86      schwarze 1678:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2       schwarze 1679:
1.110     schwarze 1680:        if (*cp == '\0')
1.149     schwarze 1681:                return ROFF_IGN;
1.1       schwarze 1682:
1.93      schwarze 1683:        /* Get the custom end marker. */
1.22      schwarze 1684:
1.93      schwarze 1685:        iname = cp;
1.86      schwarze 1686:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1687:
                   1688:        /* Resolve the end marker if it is indirect. */
                   1689:
1.110     schwarze 1690:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1691:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1692:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1693:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1694:                            "%.*s", (int)namesz, iname);
                   1695:                        namesz = 0;
                   1696:                } else
                   1697:                        namesz = strlen(name);
                   1698:        } else
                   1699:                name = iname;
                   1700:
1.86      schwarze 1701:        if (namesz)
                   1702:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1703:
1.110     schwarze 1704:        if (*cp != '\0')
1.90      schwarze 1705:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167     schwarze 1706:                    ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1       schwarze 1707:
1.149     schwarze 1708:        return ROFF_IGN;
1.1       schwarze 1709: }
                   1710:
                   1711: static enum rofferr
1.2       schwarze 1712: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1713: {
1.166     schwarze 1714:        enum roff_tok   t;
1.2       schwarze 1715:        int             i, j;
                   1716:
                   1717:        /*
                   1718:         * First check whether a custom macro exists at this level.  If
                   1719:         * it does, then check against it.  This is some of groff's
                   1720:         * stranger behaviours.  If we encountered a custom end-scope
                   1721:         * tag and that tag also happens to be a "real" macro, then we
                   1722:         * need to try interpreting it again as a real macro.  If it's
                   1723:         * not, then return ignore.  Else continue.
                   1724:         */
                   1725:
                   1726:        if (r->last->end) {
1.35      schwarze 1727:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1728:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1729:                                break;
1.1       schwarze 1730:
1.110     schwarze 1731:                if (r->last->end[j] == '\0' &&
                   1732:                    (buf->buf[i] == '\0' ||
                   1733:                     buf->buf[i] == ' ' ||
                   1734:                     buf->buf[i] == '\t')) {
1.2       schwarze 1735:                        roffnode_pop(r);
                   1736:                        roffnode_cleanscope(r);
1.1       schwarze 1737:
1.110     schwarze 1738:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1739:                                i++;
                   1740:
                   1741:                        pos = i;
1.110     schwarze 1742:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166     schwarze 1743:                            TOKEN_NONE)
1.149     schwarze 1744:                                return ROFF_RERUN;
                   1745:                        return ROFF_IGN;
1.2       schwarze 1746:                }
1.1       schwarze 1747:        }
                   1748:
1.2       schwarze 1749:        /*
                   1750:         * If we have no custom end-query or lookup failed, then try
                   1751:         * pulling it out of the hashtable.
                   1752:         */
1.1       schwarze 1753:
1.110     schwarze 1754:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1755:
1.110     schwarze 1756:        if (t != ROFF_cblock) {
                   1757:                if (tok != ROFF_ig)
                   1758:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149     schwarze 1759:                return ROFF_IGN;
1.16      schwarze 1760:        }
1.1       schwarze 1761:
1.149     schwarze 1762:        return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2       schwarze 1763: }
                   1764:
                   1765: static enum rofferr
                   1766: roff_block_text(ROFF_ARGS)
                   1767: {
                   1768:
1.110     schwarze 1769:        if (tok != ROFF_ig)
                   1770:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1771:
1.149     schwarze 1772:        return ROFF_IGN;
1.2       schwarze 1773: }
                   1774:
                   1775: static enum rofferr
                   1776: roff_cond_sub(ROFF_ARGS)
                   1777: {
1.166     schwarze 1778:        enum roff_tok    t;
1.37      schwarze 1779:        char            *ep;
1.71      schwarze 1780:        int              rr;
1.2       schwarze 1781:
                   1782:        rr = r->last->rule;
1.37      schwarze 1783:        roffnode_cleanscope(r);
1.110     schwarze 1784:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2       schwarze 1785:
1.37      schwarze 1786:        /*
1.50      schwarze 1787:         * Fully handle known macros when they are structurally
                   1788:         * required or when the conditional evaluated to true.
1.5       schwarze 1789:         */
                   1790:
1.166     schwarze 1791:        if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.149     schwarze 1792:                return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.39      schwarze 1793:
1.69      schwarze 1794:        /*
                   1795:         * If `\}' occurs on a macro line without a preceding macro,
                   1796:         * drop the line completely.
                   1797:         */
                   1798:
1.110     schwarze 1799:        ep = buf->buf + pos;
                   1800:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1801:                rr = 0;
1.69      schwarze 1802:
1.50      schwarze 1803:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1804:
1.110     schwarze 1805:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1806:                if (*(++ep) == '}') {
1.70      schwarze 1807:                        *ep = '&';
1.110     schwarze 1808:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1809:                }
1.119     schwarze 1810:                if (*ep != '\0')
                   1811:                        ++ep;
1.50      schwarze 1812:        }
1.149     schwarze 1813:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1814: }
                   1815:
                   1816: static enum rofferr
                   1817: roff_cond_text(ROFF_ARGS)
                   1818: {
1.37      schwarze 1819:        char            *ep;
1.71      schwarze 1820:        int              rr;
1.2       schwarze 1821:
                   1822:        rr = r->last->rule;
1.37      schwarze 1823:        roffnode_cleanscope(r);
1.1       schwarze 1824:
1.110     schwarze 1825:        ep = buf->buf + pos;
                   1826:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1827:                if (*(++ep) == '}') {
1.70      schwarze 1828:                        *ep = '&';
1.110     schwarze 1829:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1830:                }
1.119     schwarze 1831:                if (*ep != '\0')
                   1832:                        ++ep;
1.2       schwarze 1833:        }
1.149     schwarze 1834:        return rr ? ROFF_CONT : ROFF_IGN;
1.2       schwarze 1835: }
                   1836:
1.138     schwarze 1837: /* --- handling of numeric and conditional expressions -------------------- */
                   1838:
1.77      schwarze 1839: /*
                   1840:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1841:  * If there is at least one digit, return success and advance the
                   1842:  * parse point, else return failure and let the parse point unchanged.
                   1843:  * Ignore overflows, treat them just like the C language.
                   1844:  */
1.56      schwarze 1845: static int
1.133     schwarze 1846: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 1847: {
1.133     schwarze 1848:        int      myres, scaled, n, p;
1.79      schwarze 1849:
                   1850:        if (NULL == res)
                   1851:                res = &myres;
1.56      schwarze 1852:
                   1853:        p = *pos;
                   1854:        n = v[p] == '-';
1.133     schwarze 1855:        if (n || v[p] == '+')
1.56      schwarze 1856:                p++;
                   1857:
1.133     schwarze 1858:        if (flags & ROFFNUM_WHITE)
                   1859:                while (isspace((unsigned char)v[p]))
                   1860:                        p++;
                   1861:
1.56      schwarze 1862:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 1863:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 1864:        if (p == *pos + n)
                   1865:                return 0;
                   1866:
                   1867:        if (n)
                   1868:                *res = -*res;
                   1869:
1.126     schwarze 1870:        /* Each number may be followed by one optional scaling unit. */
                   1871:
                   1872:        switch (v[p]) {
                   1873:        case 'f':
1.133     schwarze 1874:                scaled = *res * 65536;
1.126     schwarze 1875:                break;
                   1876:        case 'i':
1.133     schwarze 1877:                scaled = *res * 240;
1.126     schwarze 1878:                break;
                   1879:        case 'c':
1.133     schwarze 1880:                scaled = *res * 240 / 2.54;
1.126     schwarze 1881:                break;
                   1882:        case 'v':
                   1883:        case 'P':
1.133     schwarze 1884:                scaled = *res * 40;
1.126     schwarze 1885:                break;
                   1886:        case 'm':
                   1887:        case 'n':
1.133     schwarze 1888:                scaled = *res * 24;
1.126     schwarze 1889:                break;
                   1890:        case 'p':
1.133     schwarze 1891:                scaled = *res * 10 / 3;
1.126     schwarze 1892:                break;
                   1893:        case 'u':
1.133     schwarze 1894:                scaled = *res;
1.126     schwarze 1895:                break;
                   1896:        case 'M':
1.133     schwarze 1897:                scaled = *res * 6 / 25;
1.126     schwarze 1898:                break;
                   1899:        default:
1.133     schwarze 1900:                scaled = *res;
1.126     schwarze 1901:                p--;
                   1902:                break;
                   1903:        }
1.133     schwarze 1904:        if (flags & ROFFNUM_SCALE)
                   1905:                *res = scaled;
1.126     schwarze 1906:
                   1907:        *pos = p + 1;
1.149     schwarze 1908:        return 1;
1.56      schwarze 1909: }
                   1910:
1.71      schwarze 1911: /*
                   1912:  * Evaluate a string comparison condition.
                   1913:  * The first character is the delimiter.
                   1914:  * Succeed if the string up to its second occurrence
                   1915:  * matches the string up to its third occurence.
                   1916:  * Advance the cursor after the third occurrence
                   1917:  * or lacking that, to the end of the line.
                   1918:  */
                   1919: static int
                   1920: roff_evalstrcond(const char *v, int *pos)
                   1921: {
                   1922:        const char      *s1, *s2, *s3;
                   1923:        int              match;
                   1924:
                   1925:        match = 0;
                   1926:        s1 = v + *pos;          /* initial delimiter */
                   1927:        s2 = s1 + 1;            /* for scanning the first string */
                   1928:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   1929:
                   1930:        if (NULL == s3)         /* found no middle delimiter */
                   1931:                goto out;
                   1932:
                   1933:        while ('\0' != *++s3) {
                   1934:                if (*s2 != *s3) {  /* mismatch */
                   1935:                        s3 = strchr(s3, *s1);
                   1936:                        break;
                   1937:                }
                   1938:                if (*s3 == *s1) {  /* found the final delimiter */
                   1939:                        match = 1;
                   1940:                        break;
                   1941:                }
                   1942:                s2++;
                   1943:        }
                   1944:
                   1945: out:
                   1946:        if (NULL == s3)
                   1947:                s3 = strchr(s2, '\0');
1.114     schwarze 1948:        else if (*s3 != '\0')
1.71      schwarze 1949:                s3++;
                   1950:        *pos = s3 - v;
1.149     schwarze 1951:        return match;
1.71      schwarze 1952: }
                   1953:
1.77      schwarze 1954: /*
                   1955:  * Evaluate an optionally negated single character, numerical,
                   1956:  * or string condition.
                   1957:  */
1.71      schwarze 1958: static int
1.143     schwarze 1959: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5       schwarze 1960: {
1.143     schwarze 1961:        char    *cp, *name;
                   1962:        size_t   sz;
1.113     schwarze 1963:        int      number, savepos, wanttrue;
1.5       schwarze 1964:
1.71      schwarze 1965:        if ('!' == v[*pos]) {
                   1966:                wanttrue = 0;
                   1967:                (*pos)++;
                   1968:        } else
                   1969:                wanttrue = 1;
                   1970:
1.5       schwarze 1971:        switch (v[*pos]) {
1.112     schwarze 1972:        case '\0':
1.149     schwarze 1973:                return 0;
1.80      schwarze 1974:        case 'n':
                   1975:        case 'o':
1.5       schwarze 1976:                (*pos)++;
1.149     schwarze 1977:                return wanttrue;
1.80      schwarze 1978:        case 'c':
                   1979:        case 'd':
                   1980:        case 'e':
                   1981:        case 't':
1.111     schwarze 1982:        case 'v':
1.5       schwarze 1983:                (*pos)++;
1.149     schwarze 1984:                return !wanttrue;
1.143     schwarze 1985:        case 'r':
                   1986:                cp = name = v + ++*pos;
                   1987:                sz = roff_getname(r, &cp, ln, *pos);
                   1988:                *pos = cp - v;
1.149     schwarze 1989:                return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.5       schwarze 1990:        default:
                   1991:                break;
                   1992:        }
                   1993:
1.113     schwarze 1994:        savepos = *pos;
1.133     schwarze 1995:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149     schwarze 1996:                return (number > 0) == wanttrue;
1.113     schwarze 1997:        else if (*pos == savepos)
1.149     schwarze 1998:                return roff_evalstrcond(v, pos) == wanttrue;
1.77      schwarze 1999:        else
1.149     schwarze 2000:                return 0;
1.5       schwarze 2001: }
                   2002:
1.2       schwarze 2003: static enum rofferr
1.21      schwarze 2004: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 2005: {
1.30      schwarze 2006:
1.149     schwarze 2007:        return ROFF_IGN;
1.21      schwarze 2008: }
                   2009:
                   2010: static enum rofferr
1.123     schwarze 2011: roff_insec(ROFF_ARGS)
                   2012: {
                   2013:
                   2014:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167     schwarze 2015:            ln, ppos, roff_name[tok]);
1.149     schwarze 2016:        return ROFF_IGN;
1.123     schwarze 2017: }
                   2018:
                   2019: static enum rofferr
                   2020: roff_unsupp(ROFF_ARGS)
                   2021: {
                   2022:
                   2023:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167     schwarze 2024:            ln, ppos, roff_name[tok]);
1.149     schwarze 2025:        return ROFF_IGN;
1.123     schwarze 2026: }
                   2027:
                   2028: static enum rofferr
1.2       schwarze 2029: roff_cond(ROFF_ARGS)
                   2030: {
1.46      schwarze 2031:
                   2032:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 2033:
1.80      schwarze 2034:        /*
1.35      schwarze 2035:         * An `.el' has no conditional body: it will consume the value
                   2036:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 2037:         * defaults to DENY.
1.35      schwarze 2038:         *
                   2039:         * If we're not an `el', however, then evaluate the conditional.
                   2040:         */
1.1       schwarze 2041:
1.110     schwarze 2042:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 2043:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 2044:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 2045:
1.35      schwarze 2046:        /*
                   2047:         * An if-else will put the NEGATION of the current evaluated
                   2048:         * conditional into the stack of rules.
                   2049:         */
                   2050:
1.110     schwarze 2051:        if (tok == ROFF_ie) {
1.96      schwarze 2052:                if (r->rstackpos + 1 == r->rstacksz) {
                   2053:                        r->rstacksz += 16;
                   2054:                        r->rstack = mandoc_reallocarray(r->rstack,
                   2055:                            r->rstacksz, sizeof(int));
1.35      schwarze 2056:                }
1.71      schwarze 2057:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 2058:        }
1.5       schwarze 2059:
                   2060:        /* If the parent has false as its rule, then so do we. */
                   2061:
1.71      schwarze 2062:        if (r->last->parent && !r->last->parent->rule)
                   2063:                r->last->rule = 0;
1.5       schwarze 2064:
                   2065:        /*
1.46      schwarze 2066:         * Determine scope.
                   2067:         * If there is nothing on the line after the conditional,
                   2068:         * not even whitespace, use next-line scope.
1.5       schwarze 2069:         */
1.2       schwarze 2070:
1.110     schwarze 2071:        if (buf->buf[pos] == '\0') {
1.46      schwarze 2072:                r->last->endspan = 2;
                   2073:                goto out;
                   2074:        }
                   2075:
1.110     schwarze 2076:        while (buf->buf[pos] == ' ')
1.46      schwarze 2077:                pos++;
                   2078:
                   2079:        /* An opening brace requests multiline scope. */
1.2       schwarze 2080:
1.110     schwarze 2081:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 2082:                r->last->endspan = -1;
                   2083:                pos += 2;
1.144     schwarze 2084:                while (buf->buf[pos] == ' ')
                   2085:                        pos++;
1.46      schwarze 2086:                goto out;
1.80      schwarze 2087:        }
1.2       schwarze 2088:
                   2089:        /*
1.46      schwarze 2090:         * Anything else following the conditional causes
                   2091:         * single-line scope.  Warn if the scope contains
                   2092:         * nothing but trailing whitespace.
1.2       schwarze 2093:         */
                   2094:
1.110     schwarze 2095:        if (buf->buf[pos] == '\0')
1.89      schwarze 2096:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167     schwarze 2097:                    ln, ppos, roff_name[tok]);
1.2       schwarze 2098:
1.46      schwarze 2099:        r->last->endspan = 1;
1.1       schwarze 2100:
1.46      schwarze 2101: out:
1.2       schwarze 2102:        *offs = pos;
1.149     schwarze 2103:        return ROFF_RERUN;
1.1       schwarze 2104: }
                   2105:
1.2       schwarze 2106: static enum rofferr
1.7       schwarze 2107: roff_ds(ROFF_ARGS)
                   2108: {
1.85      schwarze 2109:        char            *string;
                   2110:        const char      *name;
                   2111:        size_t           namesz;
1.10      schwarze 2112:
1.123     schwarze 2113:        /* Ignore groff compatibility mode for now. */
                   2114:
                   2115:        if (tok == ROFF_ds1)
                   2116:                tok = ROFF_ds;
                   2117:        else if (tok == ROFF_as1)
                   2118:                tok = ROFF_as;
                   2119:
1.10      schwarze 2120:        /*
1.85      schwarze 2121:         * The first word is the name of the string.
                   2122:         * If it is empty or terminated by an escape sequence,
                   2123:         * abort the `ds' request without defining anything.
1.10      schwarze 2124:         */
1.7       schwarze 2125:
1.110     schwarze 2126:        name = string = buf->buf + pos;
                   2127:        if (*name == '\0')
1.149     schwarze 2128:                return ROFF_IGN;
1.7       schwarze 2129:
1.85      schwarze 2130:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 2131:        if (name[namesz] == '\\')
1.149     schwarze 2132:                return ROFF_IGN;
1.85      schwarze 2133:
                   2134:        /* Read past the initial double-quote, if any. */
1.110     schwarze 2135:        if (*string == '"')
1.7       schwarze 2136:                string++;
                   2137:
1.10      schwarze 2138:        /* The rest is the value. */
1.85      schwarze 2139:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   2140:            ROFF_as == tok);
1.149     schwarze 2141:        return ROFF_IGN;
1.7       schwarze 2142: }
                   2143:
1.77      schwarze 2144: /*
                   2145:  * Parse a single operator, one or two characters long.
                   2146:  * If the operator is recognized, return success and advance the
                   2147:  * parse point, else return failure and let the parse point unchanged.
                   2148:  */
                   2149: static int
                   2150: roff_getop(const char *v, int *pos, char *res)
                   2151: {
                   2152:
                   2153:        *res = v[*pos];
                   2154:
                   2155:        switch (*res) {
1.80      schwarze 2156:        case '+':
                   2157:        case '-':
                   2158:        case '*':
                   2159:        case '/':
                   2160:        case '%':
                   2161:        case '&':
                   2162:        case ':':
1.77      schwarze 2163:                break;
                   2164:        case '<':
                   2165:                switch (v[*pos + 1]) {
1.80      schwarze 2166:                case '=':
1.77      schwarze 2167:                        *res = 'l';
                   2168:                        (*pos)++;
                   2169:                        break;
1.80      schwarze 2170:                case '>':
1.77      schwarze 2171:                        *res = '!';
                   2172:                        (*pos)++;
                   2173:                        break;
1.80      schwarze 2174:                case '?':
1.77      schwarze 2175:                        *res = 'i';
                   2176:                        (*pos)++;
                   2177:                        break;
                   2178:                default:
                   2179:                        break;
                   2180:                }
                   2181:                break;
                   2182:        case '>':
                   2183:                switch (v[*pos + 1]) {
1.80      schwarze 2184:                case '=':
1.77      schwarze 2185:                        *res = 'g';
                   2186:                        (*pos)++;
                   2187:                        break;
1.80      schwarze 2188:                case '?':
1.77      schwarze 2189:                        *res = 'a';
                   2190:                        (*pos)++;
                   2191:                        break;
                   2192:                default:
                   2193:                        break;
                   2194:                }
                   2195:                break;
                   2196:        case '=':
                   2197:                if ('=' == v[*pos + 1])
                   2198:                        (*pos)++;
                   2199:                break;
                   2200:        default:
1.149     schwarze 2201:                return 0;
1.77      schwarze 2202:        }
                   2203:        (*pos)++;
                   2204:
1.149     schwarze 2205:        return *res;
1.77      schwarze 2206: }
                   2207:
                   2208: /*
                   2209:  * Evaluate either a parenthesized numeric expression
                   2210:  * or a single signed integer number.
                   2211:  */
                   2212: static int
1.106     schwarze 2213: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2214:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2215: {
                   2216:
                   2217:        if ('(' != v[*pos])
1.149     schwarze 2218:                return roff_getnum(v, pos, res, flags);
1.77      schwarze 2219:
                   2220:        (*pos)++;
1.133     schwarze 2221:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149     schwarze 2222:                return 0;
1.77      schwarze 2223:
1.79      schwarze 2224:        /*
                   2225:         * Omission of the closing parenthesis
                   2226:         * is an error in validation mode,
                   2227:         * but ignored in evaluation mode.
                   2228:         */
                   2229:
1.77      schwarze 2230:        if (')' == v[*pos])
                   2231:                (*pos)++;
1.79      schwarze 2232:        else if (NULL == res)
1.149     schwarze 2233:                return 0;
1.77      schwarze 2234:
1.149     schwarze 2235:        return 1;
1.77      schwarze 2236: }
                   2237:
                   2238: /*
                   2239:  * Evaluate a complete numeric expression.
                   2240:  * Proceed left to right, there is no concept of precedence.
                   2241:  */
                   2242: static int
1.106     schwarze 2243: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2244:        int *pos, int *res, int flags)
1.77      schwarze 2245: {
                   2246:        int              mypos, operand2;
                   2247:        char             operator;
                   2248:
                   2249:        if (NULL == pos) {
                   2250:                mypos = 0;
                   2251:                pos = &mypos;
                   2252:        }
                   2253:
1.133     schwarze 2254:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2255:                while (isspace((unsigned char)v[*pos]))
                   2256:                        (*pos)++;
                   2257:
1.133     schwarze 2258:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149     schwarze 2259:                return 0;
1.77      schwarze 2260:
                   2261:        while (1) {
1.133     schwarze 2262:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2263:                        while (isspace((unsigned char)v[*pos]))
                   2264:                                (*pos)++;
                   2265:
                   2266:                if ( ! roff_getop(v, pos, &operator))
                   2267:                        break;
                   2268:
1.133     schwarze 2269:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2270:                        while (isspace((unsigned char)v[*pos]))
                   2271:                                (*pos)++;
                   2272:
1.133     schwarze 2273:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149     schwarze 2274:                        return 0;
1.77      schwarze 2275:
1.133     schwarze 2276:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2277:                        while (isspace((unsigned char)v[*pos]))
                   2278:                                (*pos)++;
1.79      schwarze 2279:
                   2280:                if (NULL == res)
                   2281:                        continue;
1.77      schwarze 2282:
                   2283:                switch (operator) {
1.80      schwarze 2284:                case '+':
1.77      schwarze 2285:                        *res += operand2;
                   2286:                        break;
1.80      schwarze 2287:                case '-':
1.77      schwarze 2288:                        *res -= operand2;
                   2289:                        break;
1.80      schwarze 2290:                case '*':
1.77      schwarze 2291:                        *res *= operand2;
                   2292:                        break;
1.80      schwarze 2293:                case '/':
1.116     schwarze 2294:                        if (operand2 == 0) {
1.106     schwarze 2295:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2296:                                        r->parse, ln, *pos, v);
                   2297:                                *res = 0;
                   2298:                                break;
                   2299:                        }
1.77      schwarze 2300:                        *res /= operand2;
                   2301:                        break;
1.80      schwarze 2302:                case '%':
1.116     schwarze 2303:                        if (operand2 == 0) {
                   2304:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2305:                                        r->parse, ln, *pos, v);
                   2306:                                *res = 0;
                   2307:                                break;
                   2308:                        }
1.77      schwarze 2309:                        *res %= operand2;
                   2310:                        break;
1.80      schwarze 2311:                case '<':
1.77      schwarze 2312:                        *res = *res < operand2;
                   2313:                        break;
1.80      schwarze 2314:                case '>':
1.77      schwarze 2315:                        *res = *res > operand2;
                   2316:                        break;
1.80      schwarze 2317:                case 'l':
1.77      schwarze 2318:                        *res = *res <= operand2;
                   2319:                        break;
1.80      schwarze 2320:                case 'g':
1.77      schwarze 2321:                        *res = *res >= operand2;
                   2322:                        break;
1.80      schwarze 2323:                case '=':
1.77      schwarze 2324:                        *res = *res == operand2;
                   2325:                        break;
1.80      schwarze 2326:                case '!':
1.77      schwarze 2327:                        *res = *res != operand2;
                   2328:                        break;
1.80      schwarze 2329:                case '&':
1.77      schwarze 2330:                        *res = *res && operand2;
                   2331:                        break;
1.80      schwarze 2332:                case ':':
1.77      schwarze 2333:                        *res = *res || operand2;
                   2334:                        break;
1.80      schwarze 2335:                case 'i':
1.77      schwarze 2336:                        if (operand2 < *res)
                   2337:                                *res = operand2;
                   2338:                        break;
1.80      schwarze 2339:                case 'a':
1.77      schwarze 2340:                        if (operand2 > *res)
                   2341:                                *res = operand2;
                   2342:                        break;
                   2343:                default:
                   2344:                        abort();
                   2345:                }
                   2346:        }
1.149     schwarze 2347:        return 1;
1.77      schwarze 2348: }
                   2349:
1.138     schwarze 2350: /* --- register management ------------------------------------------------ */
                   2351:
1.52      schwarze 2352: void
1.60      schwarze 2353: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2354: {
1.52      schwarze 2355:        struct roffreg  *reg;
                   2356:
                   2357:        /* Search for an existing register with the same name. */
                   2358:        reg = r->regtab;
                   2359:
                   2360:        while (reg && strcmp(name, reg->key.p))
                   2361:                reg = reg->next;
1.41      schwarze 2362:
1.52      schwarze 2363:        if (NULL == reg) {
                   2364:                /* Create a new register. */
                   2365:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2366:                reg->key.p = mandoc_strdup(name);
                   2367:                reg->key.sz = strlen(name);
1.60      schwarze 2368:                reg->val = 0;
1.52      schwarze 2369:                reg->next = r->regtab;
                   2370:                r->regtab = reg;
                   2371:        }
                   2372:
1.60      schwarze 2373:        if ('+' == sign)
                   2374:                reg->val += val;
                   2375:        else if ('-' == sign)
                   2376:                reg->val -= val;
                   2377:        else
                   2378:                reg->val = val;
1.41      schwarze 2379: }
                   2380:
1.65      schwarze 2381: /*
                   2382:  * Handle some predefined read-only number registers.
                   2383:  * For now, return -1 if the requested register is not predefined;
                   2384:  * in case a predefined read-only register having the value -1
                   2385:  * were to turn up, another special value would have to be chosen.
                   2386:  */
                   2387: static int
1.145     schwarze 2388: roff_getregro(const struct roff *r, const char *name)
1.65      schwarze 2389: {
                   2390:
                   2391:        switch (*name) {
1.145     schwarze 2392:        case '$':  /* Number of arguments of the last macro evaluated. */
1.149     schwarze 2393:                return r->argc;
1.80      schwarze 2394:        case 'A':  /* ASCII approximation mode is always off. */
1.149     schwarze 2395:                return 0;
1.80      schwarze 2396:        case 'g':  /* Groff compatibility mode is always on. */
1.149     schwarze 2397:                return 1;
1.80      schwarze 2398:        case 'H':  /* Fixed horizontal resolution. */
1.149     schwarze 2399:                return 24;
1.80      schwarze 2400:        case 'j':  /* Always adjust left margin only. */
1.149     schwarze 2401:                return 0;
1.80      schwarze 2402:        case 'T':  /* Some output device is always defined. */
1.149     schwarze 2403:                return 1;
1.80      schwarze 2404:        case 'V':  /* Fixed vertical resolution. */
1.149     schwarze 2405:                return 40;
1.65      schwarze 2406:        default:
1.149     schwarze 2407:                return -1;
1.65      schwarze 2408:        }
                   2409: }
                   2410:
1.53      schwarze 2411: int
1.52      schwarze 2412: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2413: {
1.52      schwarze 2414:        struct roffreg  *reg;
1.65      schwarze 2415:        int              val;
                   2416:
                   2417:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145     schwarze 2418:                val = roff_getregro(r, name + 1);
1.65      schwarze 2419:                if (-1 != val)
1.149     schwarze 2420:                        return val;
1.65      schwarze 2421:        }
1.52      schwarze 2422:
                   2423:        for (reg = r->regtab; reg; reg = reg->next)
                   2424:                if (0 == strcmp(name, reg->key.p))
1.149     schwarze 2425:                        return reg->val;
1.53      schwarze 2426:
1.149     schwarze 2427:        return 0;
1.53      schwarze 2428: }
                   2429:
                   2430: static int
                   2431: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2432: {
                   2433:        struct roffreg  *reg;
1.65      schwarze 2434:        int              val;
                   2435:
                   2436:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2437:                val = roff_getregro(r, name + 1);
1.65      schwarze 2438:                if (-1 != val)
1.149     schwarze 2439:                        return val;
1.65      schwarze 2440:        }
1.53      schwarze 2441:
                   2442:        for (reg = r->regtab; reg; reg = reg->next)
                   2443:                if (len == reg->key.sz &&
                   2444:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2445:                        return reg->val;
1.143     schwarze 2446:
1.149     schwarze 2447:        return 0;
1.143     schwarze 2448: }
                   2449:
                   2450: static int
                   2451: roff_hasregn(const struct roff *r, const char *name, size_t len)
                   2452: {
                   2453:        struct roffreg  *reg;
                   2454:        int              val;
                   2455:
                   2456:        if ('.' == name[0] && 2 == len) {
1.145     schwarze 2457:                val = roff_getregro(r, name + 1);
1.143     schwarze 2458:                if (-1 != val)
1.149     schwarze 2459:                        return 1;
1.143     schwarze 2460:        }
                   2461:
                   2462:        for (reg = r->regtab; reg; reg = reg->next)
                   2463:                if (len == reg->key.sz &&
                   2464:                    0 == strncmp(name, reg->key.p, len))
1.149     schwarze 2465:                        return 1;
1.41      schwarze 2466:
1.149     schwarze 2467:        return 0;
1.41      schwarze 2468: }
                   2469:
1.52      schwarze 2470: static void
                   2471: roff_freereg(struct roffreg *reg)
1.41      schwarze 2472: {
1.52      schwarze 2473:        struct roffreg  *old_reg;
1.41      schwarze 2474:
1.52      schwarze 2475:        while (NULL != reg) {
                   2476:                free(reg->key.p);
                   2477:                old_reg = reg;
                   2478:                reg = reg->next;
                   2479:                free(old_reg);
                   2480:        }
1.41      schwarze 2481: }
1.7       schwarze 2482:
                   2483: static enum rofferr
1.6       schwarze 2484: roff_nr(ROFF_ARGS)
1.1       schwarze 2485: {
1.85      schwarze 2486:        char            *key, *val;
                   2487:        size_t           keysz;
1.37      schwarze 2488:        int              iv;
1.60      schwarze 2489:        char             sign;
1.6       schwarze 2490:
1.110     schwarze 2491:        key = val = buf->buf + pos;
                   2492:        if (*key == '\0')
1.149     schwarze 2493:                return ROFF_IGN;
1.85      schwarze 2494:
                   2495:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2496:        if (key[keysz] == '\\')
1.149     schwarze 2497:                return ROFF_IGN;
1.85      schwarze 2498:        key[keysz] = '\0';
1.6       schwarze 2499:
1.60      schwarze 2500:        sign = *val;
1.110     schwarze 2501:        if (sign == '+' || sign == '-')
1.60      schwarze 2502:                val++;
                   2503:
1.133     schwarze 2504:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77      schwarze 2505:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2506:
1.149     schwarze 2507:        return ROFF_IGN;
1.76      schwarze 2508: }
                   2509:
                   2510: static enum rofferr
                   2511: roff_rr(ROFF_ARGS)
                   2512: {
                   2513:        struct roffreg  *reg, **prev;
1.85      schwarze 2514:        char            *name, *cp;
                   2515:        size_t           namesz;
1.76      schwarze 2516:
1.110     schwarze 2517:        name = cp = buf->buf + pos;
                   2518:        if (*name == '\0')
1.149     schwarze 2519:                return ROFF_IGN;
1.85      schwarze 2520:        namesz = roff_getname(r, &cp, ln, pos);
                   2521:        name[namesz] = '\0';
1.76      schwarze 2522:
                   2523:        prev = &r->regtab;
                   2524:        while (1) {
                   2525:                reg = *prev;
1.110     schwarze 2526:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2527:                        break;
                   2528:                prev = &reg->next;
                   2529:        }
1.110     schwarze 2530:        if (reg != NULL) {
1.76      schwarze 2531:                *prev = reg->next;
                   2532:                free(reg->key.p);
                   2533:                free(reg);
                   2534:        }
1.149     schwarze 2535:        return ROFF_IGN;
1.29      schwarze 2536: }
                   2537:
1.138     schwarze 2538: /* --- handler functions for roff requests -------------------------------- */
                   2539:
1.29      schwarze 2540: static enum rofferr
                   2541: roff_rm(ROFF_ARGS)
                   2542: {
                   2543:        const char       *name;
                   2544:        char             *cp;
1.85      schwarze 2545:        size_t            namesz;
1.29      schwarze 2546:
1.110     schwarze 2547:        cp = buf->buf + pos;
                   2548:        while (*cp != '\0') {
1.85      schwarze 2549:                name = cp;
1.110     schwarze 2550:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2551:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110     schwarze 2552:                if (name[namesz] == '\\')
1.85      schwarze 2553:                        break;
1.29      schwarze 2554:        }
1.149     schwarze 2555:        return ROFF_IGN;
1.51      schwarze 2556: }
                   2557:
                   2558: static enum rofferr
                   2559: roff_it(ROFF_ARGS)
                   2560: {
                   2561:        int              iv;
                   2562:
                   2563:        /* Parse the number of lines. */
1.133     schwarze 2564:
                   2565:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2566:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2567:                    ln, ppos, buf->buf + 1);
1.149     schwarze 2568:                return ROFF_IGN;
1.51      schwarze 2569:        }
                   2570:
1.134     schwarze 2571:        while (isspace((unsigned char)buf->buf[pos]))
                   2572:                pos++;
                   2573:
                   2574:        /*
                   2575:         * Arm the input line trap.
                   2576:         * Special-casing "an-trap" is an ugly workaround to cope
                   2577:         * with DocBook stupidly fiddling with man(7) internals.
                   2578:         */
1.133     schwarze 2579:
1.51      schwarze 2580:        roffit_lines = iv;
1.134     schwarze 2581:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2582:            strcmp(buf->buf + pos, "an-trap") ?
                   2583:            buf->buf + pos : "br");
1.149     schwarze 2584:        return ROFF_IGN;
1.47      schwarze 2585: }
                   2586:
                   2587: static enum rofferr
                   2588: roff_Dd(ROFF_ARGS)
                   2589: {
                   2590:        const char *const       *cp;
                   2591:
1.99      schwarze 2592:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47      schwarze 2593:                for (cp = __mdoc_reserved; *cp; cp++)
                   2594:                        roff_setstr(r, *cp, NULL, 0);
                   2595:
1.99      schwarze 2596:        if (r->format == 0)
                   2597:                r->format = MPARSE_MDOC;
                   2598:
1.149     schwarze 2599:        return ROFF_CONT;
1.47      schwarze 2600: }
                   2601:
                   2602: static enum rofferr
                   2603: roff_TH(ROFF_ARGS)
                   2604: {
                   2605:        const char *const       *cp;
                   2606:
1.99      schwarze 2607:        if ((r->options & MPARSE_QUICK) == 0)
1.47      schwarze 2608:                for (cp = __man_reserved; *cp; cp++)
                   2609:                        roff_setstr(r, *cp, NULL, 0);
                   2610:
1.99      schwarze 2611:        if (r->format == 0)
                   2612:                r->format = MPARSE_MAN;
                   2613:
1.149     schwarze 2614:        return ROFF_CONT;
1.14      schwarze 2615: }
                   2616:
                   2617: static enum rofferr
1.27      schwarze 2618: roff_TE(ROFF_ARGS)
                   2619: {
                   2620:
                   2621:        if (NULL == r->tbl)
1.94      schwarze 2622:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2623:                    ln, ppos, "TE");
1.130     schwarze 2624:        else if ( ! tbl_end(&r->tbl)) {
                   2625:                free(buf->buf);
                   2626:                buf->buf = mandoc_strdup(".sp");
                   2627:                buf->sz = 4;
1.149     schwarze 2628:                return ROFF_REPARSE;
1.130     schwarze 2629:        }
1.149     schwarze 2630:        return ROFF_IGN;
1.27      schwarze 2631: }
                   2632:
                   2633: static enum rofferr
                   2634: roff_T_(ROFF_ARGS)
                   2635: {
                   2636:
                   2637:        if (NULL == r->tbl)
1.94      schwarze 2638:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2639:                    ln, ppos, "T&");
1.27      schwarze 2640:        else
1.168   ! schwarze 2641:                tbl_restart(ln, ppos, r->tbl);
1.27      schwarze 2642:
1.149     schwarze 2643:        return ROFF_IGN;
1.27      schwarze 2644: }
                   2645:
1.102     schwarze 2646: /*
                   2647:  * Handle in-line equation delimiters.
                   2648:  */
                   2649: static enum rofferr
1.110     schwarze 2650: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2651: {
1.105     schwarze 2652:        char            *cp1, *cp2;
                   2653:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2654:
1.102     schwarze 2655:        /*
                   2656:         * Outside equations, look for an opening delimiter.
                   2657:         * If we are inside an equation, we already know it is
                   2658:         * in-line, or this function wouldn't have been called;
                   2659:         * so look for a closing delimiter.
                   2660:         */
                   2661:
1.110     schwarze 2662:        cp1 = buf->buf + pos;
1.102     schwarze 2663:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2664:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2665:        if (cp2 == NULL)
1.149     schwarze 2666:                return ROFF_CONT;
1.102     schwarze 2667:
1.105     schwarze 2668:        *cp2++ = '\0';
                   2669:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2670:
                   2671:        /* Handle preceding text, protecting whitespace. */
                   2672:
1.110     schwarze 2673:        if (*buf->buf != '\0') {
1.105     schwarze 2674:                if (r->eqn == NULL)
                   2675:                        bef_pr = "\\&";
                   2676:                bef_nl = "\n";
                   2677:        }
                   2678:
                   2679:        /*
                   2680:         * Prepare replacing the delimiter with an equation macro
                   2681:         * and drop leading white space from the equation.
                   2682:         */
1.102     schwarze 2683:
1.105     schwarze 2684:        if (r->eqn == NULL) {
                   2685:                while (*cp2 == ' ')
                   2686:                        cp2++;
                   2687:                mac = ".EQ";
                   2688:        } else
                   2689:                mac = ".EN";
                   2690:
                   2691:        /* Handle following text, protecting whitespace. */
                   2692:
                   2693:        if (*cp2 != '\0') {
                   2694:                aft_nl = "\n";
                   2695:                if (r->eqn != NULL)
                   2696:                        aft_pr = "\\&";
                   2697:        }
                   2698:
                   2699:        /* Do the actual replacement. */
                   2700:
1.110     schwarze 2701:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2702:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2703:        free(buf->buf);
                   2704:        buf->buf = cp1;
1.102     schwarze 2705:
                   2706:        /* Toggle the in-line state of the eqn subsystem. */
                   2707:
                   2708:        r->eqn_inline = r->eqn == NULL;
1.149     schwarze 2709:        return ROFF_REPARSE;
1.41      schwarze 2710: }
                   2711:
1.107     schwarze 2712: static enum rofferr
                   2713: roff_EQ(ROFF_ARGS)
1.32      schwarze 2714: {
1.41      schwarze 2715:        struct eqn_node *e;
1.32      schwarze 2716:
1.110     schwarze 2717:        assert(r->eqn == NULL);
1.107     schwarze 2718:        e = eqn_alloc(ppos, ln, r->parse);
1.32      schwarze 2719:
1.102     schwarze 2720:        if (r->last_eqn) {
1.32      schwarze 2721:                r->last_eqn->next = e;
1.102     schwarze 2722:                e->delim = r->last_eqn->delim;
                   2723:                e->odelim = r->last_eqn->odelim;
                   2724:                e->cdelim = r->last_eqn->cdelim;
                   2725:        } else
1.32      schwarze 2726:                r->first_eqn = r->last_eqn = e;
                   2727:
                   2728:        r->eqn = r->last_eqn = e;
1.41      schwarze 2729:
1.110     schwarze 2730:        if (buf->buf[pos] != '\0')
1.107     schwarze 2731:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2732:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2733:
1.149     schwarze 2734:        return ROFF_IGN;
1.32      schwarze 2735: }
                   2736:
                   2737: static enum rofferr
                   2738: roff_EN(ROFF_ARGS)
                   2739: {
                   2740:
1.94      schwarze 2741:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.149     schwarze 2742:        return ROFF_IGN;
1.32      schwarze 2743: }
                   2744:
                   2745: static enum rofferr
1.27      schwarze 2746: roff_TS(ROFF_ARGS)
                   2747: {
1.49      schwarze 2748:        struct tbl_node *tbl;
1.27      schwarze 2749:
                   2750:        if (r->tbl) {
1.94      schwarze 2751:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2752:                    ln, ppos, "TS breaks TS");
1.41      schwarze 2753:                tbl_end(&r->tbl);
1.27      schwarze 2754:        }
                   2755:
1.49      schwarze 2756:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2757:
                   2758:        if (r->last_tbl)
1.49      schwarze 2759:                r->last_tbl->next = tbl;
1.27      schwarze 2760:        else
1.49      schwarze 2761:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 2762:
1.49      schwarze 2763:        r->tbl = r->last_tbl = tbl;
1.149     schwarze 2764:        return ROFF_IGN;
1.123     schwarze 2765: }
                   2766:
                   2767: static enum rofferr
1.168   ! schwarze 2768: roff_br(ROFF_ARGS)
1.123     schwarze 2769: {
1.168   ! schwarze 2770:        roff_elem_alloc(r->man, ln, ppos, ROFF_br);
        !          2771:        if (buf->buf[pos] != '\0')
        !          2772:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
        !          2773:                    "%s %s", roff_name[tok], buf->buf + pos);
        !          2774:        r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
        !          2775:        r->man->next = ROFF_NEXT_SIBLING;
        !          2776:        return ROFF_IGN;
1.27      schwarze 2777: }
                   2778:
                   2779: static enum rofferr
1.48      schwarze 2780: roff_cc(ROFF_ARGS)
                   2781: {
                   2782:        const char      *p;
                   2783:
1.110     schwarze 2784:        p = buf->buf + pos;
1.48      schwarze 2785:
1.110     schwarze 2786:        if (*p == '\0' || (r->control = *p++) == '.')
1.48      schwarze 2787:                r->control = 0;
                   2788:
1.110     schwarze 2789:        if (*p != '\0')
1.132     schwarze 2790:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   2791:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 2792:
1.149     schwarze 2793:        return ROFF_IGN;
1.48      schwarze 2794: }
                   2795:
                   2796: static enum rofferr
1.42      schwarze 2797: roff_tr(ROFF_ARGS)
                   2798: {
                   2799:        const char      *p, *first, *second;
                   2800:        size_t           fsz, ssz;
                   2801:        enum mandoc_esc  esc;
                   2802:
1.110     schwarze 2803:        p = buf->buf + pos;
1.42      schwarze 2804:
1.110     schwarze 2805:        if (*p == '\0') {
1.132     schwarze 2806:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149     schwarze 2807:                return ROFF_IGN;
1.42      schwarze 2808:        }
                   2809:
1.110     schwarze 2810:        while (*p != '\0') {
1.42      schwarze 2811:                fsz = ssz = 1;
                   2812:
                   2813:                first = p++;
1.110     schwarze 2814:                if (*first == '\\') {
1.42      schwarze 2815:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2816:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2817:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2818:                                    ln, (int)(p - buf->buf), first);
1.149     schwarze 2819:                                return ROFF_IGN;
1.42      schwarze 2820:                        }
                   2821:                        fsz = (size_t)(p - first);
                   2822:                }
                   2823:
                   2824:                second = p++;
1.110     schwarze 2825:                if (*second == '\\') {
1.42      schwarze 2826:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2827:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2828:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2829:                                    ln, (int)(p - buf->buf), second);
1.149     schwarze 2830:                                return ROFF_IGN;
1.42      schwarze 2831:                        }
                   2832:                        ssz = (size_t)(p - second);
1.110     schwarze 2833:                } else if (*second == '\0') {
1.132     schwarze 2834:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   2835:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 2836:                        second = " ";
                   2837:                        p--;
                   2838:                }
                   2839:
                   2840:                if (fsz > 1) {
1.80      schwarze 2841:                        roff_setstrn(&r->xmbtab, first, fsz,
                   2842:                            second, ssz, 0);
1.42      schwarze 2843:                        continue;
                   2844:                }
                   2845:
1.110     schwarze 2846:                if (r->xtab == NULL)
1.80      schwarze 2847:                        r->xtab = mandoc_calloc(128,
                   2848:                            sizeof(struct roffstr));
1.42      schwarze 2849:
                   2850:                free(r->xtab[(int)*first].p);
                   2851:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   2852:                r->xtab[(int)*first].sz = ssz;
                   2853:        }
                   2854:
1.149     schwarze 2855:        return ROFF_IGN;
1.42      schwarze 2856: }
                   2857:
                   2858: static enum rofferr
1.14      schwarze 2859: roff_so(ROFF_ARGS)
                   2860: {
1.121     schwarze 2861:        char *name, *cp;
1.15      schwarze 2862:
1.110     schwarze 2863:        name = buf->buf + pos;
1.97      schwarze 2864:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 2865:
1.22      schwarze 2866:        /*
                   2867:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   2868:         * opening anything that's not in our cwd or anything beneath
                   2869:         * it.  Thus, explicitly disallow traversing up the file-system
                   2870:         * or using absolute paths.
                   2871:         */
                   2872:
1.110     schwarze 2873:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 2874:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   2875:                    ".so %s", name);
1.121     schwarze 2876:                buf->sz = mandoc_asprintf(&cp,
                   2877:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   2878:                free(buf->buf);
                   2879:                buf->buf = cp;
                   2880:                *offs = 0;
1.149     schwarze 2881:                return ROFF_REPARSE;
1.14      schwarze 2882:        }
                   2883:
                   2884:        *offs = pos;
1.149     schwarze 2885:        return ROFF_SO;
1.7       schwarze 2886: }
                   2887:
1.138     schwarze 2888: /* --- user defined strings and macros ------------------------------------ */
                   2889:
1.16      schwarze 2890: static enum rofferr
                   2891: roff_userdef(ROFF_ARGS)
1.12      schwarze 2892: {
1.135     schwarze 2893:        const char       *arg[9], *ap;
1.16      schwarze 2894:        char             *cp, *n1, *n2;
1.164     schwarze 2895:        int               expand_count, i, ib, ie;
1.135     schwarze 2896:        size_t            asz, rsz;
1.12      schwarze 2897:
1.16      schwarze 2898:        /*
                   2899:         * Collect pointers to macro argument strings
1.61      schwarze 2900:         * and NUL-terminate them.
1.16      schwarze 2901:         */
1.135     schwarze 2902:
1.145     schwarze 2903:        r->argc = 0;
1.110     schwarze 2904:        cp = buf->buf + pos;
1.145     schwarze 2905:        for (i = 0; i < 9; i++) {
                   2906:                if (*cp == '\0')
                   2907:                        arg[i] = "";
                   2908:                else {
                   2909:                        arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
                   2910:                        r->argc = i + 1;
                   2911:                }
                   2912:        }
1.16      schwarze 2913:
                   2914:        /*
                   2915:         * Expand macro arguments.
1.12      schwarze 2916:         */
1.135     schwarze 2917:
                   2918:        buf->sz = strlen(r->current_string) + 1;
1.164     schwarze 2919:        n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135     schwarze 2920:        memcpy(n1, r->current_string, buf->sz);
1.164     schwarze 2921:        expand_count = 0;
1.135     schwarze 2922:        while (*cp != '\0') {
                   2923:
                   2924:                /* Scan ahead for the next argument invocation. */
                   2925:
                   2926:                if (*cp++ != '\\')
                   2927:                        continue;
                   2928:                if (*cp++ != '$')
                   2929:                        continue;
1.146     schwarze 2930:                if (*cp == '*') {  /* \\$* inserts all arguments */
                   2931:                        ib = 0;
                   2932:                        ie = r->argc - 1;
                   2933:                } else {  /* \\$1 .. \\$9 insert one argument */
                   2934:                        ib = ie = *cp - '1';
                   2935:                        if (ib < 0 || ib > 8)
                   2936:                                continue;
                   2937:                }
1.135     schwarze 2938:                cp -= 2;
1.164     schwarze 2939:
                   2940:                /*
                   2941:                 * Prevent infinite recursion.
                   2942:                 */
                   2943:
                   2944:                if (cp >= n2)
                   2945:                        expand_count = 1;
                   2946:                else if (++expand_count > EXPAND_LIMIT) {
                   2947:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   2948:                            ln, (int)(cp - n1), NULL);
1.165     schwarze 2949:                        free(buf->buf);
                   2950:                        buf->buf = n1;
1.164     schwarze 2951:                        return ROFF_IGN;
                   2952:                }
1.135     schwarze 2953:
                   2954:                /*
                   2955:                 * Determine the size of the expanded argument,
                   2956:                 * taking escaping of quotes into account.
                   2957:                 */
                   2958:
1.146     schwarze 2959:                asz = ie > ib ? ie - ib : 0;  /* for blanks */
                   2960:                for (i = ib; i <= ie; i++) {
                   2961:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   2962:                                asz++;
                   2963:                                if (*ap == '"')
                   2964:                                        asz += 3;
                   2965:                        }
1.135     schwarze 2966:                }
                   2967:                if (asz != 3) {
                   2968:
                   2969:                        /*
                   2970:                         * Determine the size of the rest of the
                   2971:                         * unexpanded macro, including the NUL.
                   2972:                         */
                   2973:
                   2974:                        rsz = buf->sz - (cp - n1) - 3;
                   2975:
                   2976:                        /*
                   2977:                         * When shrinking, move before
                   2978:                         * releasing the storage.
                   2979:                         */
                   2980:
                   2981:                        if (asz < 3)
                   2982:                                memmove(cp + asz, cp + 3, rsz);
                   2983:
                   2984:                        /*
                   2985:                         * Resize the storage for the macro
                   2986:                         * and readjust the parse pointer.
                   2987:                         */
                   2988:
                   2989:                        buf->sz += asz - 3;
                   2990:                        n2 = mandoc_realloc(n1, buf->sz);
                   2991:                        cp = n2 + (cp - n1);
                   2992:                        n1 = n2;
                   2993:
                   2994:                        /*
                   2995:                         * When growing, make room
                   2996:                         * for the expanded argument.
                   2997:                         */
                   2998:
                   2999:                        if (asz > 3)
                   3000:                                memmove(cp + asz, cp + 3, rsz);
                   3001:                }
                   3002:
                   3003:                /* Copy the expanded argument, escaping quotes. */
                   3004:
                   3005:                n2 = cp;
1.146     schwarze 3006:                for (i = ib; i <= ie; i++) {
                   3007:                        for (ap = arg[i]; *ap != '\0'; ap++) {
                   3008:                                if (*ap == '"') {
                   3009:                                        memcpy(n2, "\\(dq", 4);
                   3010:                                        n2 += 4;
                   3011:                                } else
                   3012:                                        *n2++ = *ap;
                   3013:                        }
                   3014:                        if (i < ie)
                   3015:                                *n2++ = ' ';
1.16      schwarze 3016:                }
1.12      schwarze 3017:        }
                   3018:
1.16      schwarze 3019:        /*
                   3020:         * Replace the macro invocation
                   3021:         * by the expanded macro.
                   3022:         */
1.135     schwarze 3023:
1.110     schwarze 3024:        free(buf->buf);
                   3025:        buf->buf = n1;
1.120     schwarze 3026:        *offs = 0;
1.16      schwarze 3027:
1.149     schwarze 3028:        return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
                   3029:           ROFF_REPARSE : ROFF_APPEND;
1.12      schwarze 3030: }
1.28      schwarze 3031:
1.85      schwarze 3032: static size_t
1.28      schwarze 3033: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   3034: {
                   3035:        char     *name, *cp;
1.85      schwarze 3036:        size_t    namesz;
1.28      schwarze 3037:
                   3038:        name = *cpp;
                   3039:        if ('\0' == *name)
1.149     schwarze 3040:                return 0;
1.28      schwarze 3041:
1.85      schwarze 3042:        /* Read until end of name and terminate it with NUL. */
                   3043:        for (cp = name; 1; cp++) {
                   3044:                if ('\0' == *cp || ' ' == *cp) {
                   3045:                        namesz = cp - name;
                   3046:                        break;
                   3047:                }
1.28      schwarze 3048:                if ('\\' != *cp)
                   3049:                        continue;
1.88      schwarze 3050:                namesz = cp - name;
                   3051:                if ('{' == cp[1] || '}' == cp[1])
                   3052:                        break;
1.28      schwarze 3053:                cp++;
                   3054:                if ('\\' == *cp)
                   3055:                        continue;
1.97      schwarze 3056:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   3057:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 3058:                mandoc_escape((const char **)&cp, NULL, NULL);
                   3059:                break;
1.28      schwarze 3060:        }
                   3061:
                   3062:        /* Read past spaces. */
                   3063:        while (' ' == *cp)
                   3064:                cp++;
                   3065:
                   3066:        *cpp = cp;
1.149     schwarze 3067:        return namesz;
1.28      schwarze 3068: }
                   3069:
1.16      schwarze 3070: /*
                   3071:  * Store *string into the user-defined string called *name.
                   3072:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 3073:  * append == 0: replace mode
                   3074:  * append == 1: single-line append mode
                   3075:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 3076:  */
1.8       schwarze 3077: static void
1.16      schwarze 3078: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 3079:        int append)
1.7       schwarze 3080: {
1.42      schwarze 3081:
                   3082:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.80      schwarze 3083:            string ? strlen(string) : 0, append);
1.42      schwarze 3084: }
                   3085:
                   3086: static void
                   3087: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 3088:                const char *string, size_t stringsz, int append)
1.42      schwarze 3089: {
                   3090:        struct roffkv   *n;
                   3091:        char            *c;
                   3092:        int              i;
                   3093:        size_t           oldch, newch;
1.7       schwarze 3094:
1.16      schwarze 3095:        /* Search for an existing string with the same name. */
1.42      schwarze 3096:        n = *r;
                   3097:
1.84      schwarze 3098:        while (n && (namesz != n->key.sz ||
                   3099:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 3100:                n = n->next;
1.8       schwarze 3101:
                   3102:        if (NULL == n) {
1.16      schwarze 3103:                /* Create a new string table entry. */
1.42      schwarze 3104:                n = mandoc_malloc(sizeof(struct roffkv));
                   3105:                n->key.p = mandoc_strndup(name, namesz);
                   3106:                n->key.sz = namesz;
                   3107:                n->val.p = NULL;
                   3108:                n->val.sz = 0;
                   3109:                n->next = *r;
                   3110:                *r = n;
1.66      schwarze 3111:        } else if (0 == append) {
1.42      schwarze 3112:                free(n->val.p);
                   3113:                n->val.p = NULL;
                   3114:                n->val.sz = 0;
1.16      schwarze 3115:        }
                   3116:
                   3117:        if (NULL == string)
                   3118:                return;
                   3119:
                   3120:        /*
                   3121:         * One additional byte for the '\n' in multiline mode,
                   3122:         * and one for the terminating '\0'.
                   3123:         */
1.66      schwarze 3124:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 3125:
                   3126:        if (NULL == n->val.p) {
                   3127:                n->val.p = mandoc_malloc(newch);
                   3128:                *n->val.p = '\0';
1.16      schwarze 3129:                oldch = 0;
                   3130:        } else {
1.42      schwarze 3131:                oldch = n->val.sz;
                   3132:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 3133:        }
                   3134:
                   3135:        /* Skip existing content in the destination buffer. */
1.42      schwarze 3136:        c = n->val.p + (int)oldch;
1.16      schwarze 3137:
                   3138:        /* Append new content to the destination buffer. */
1.42      schwarze 3139:        i = 0;
                   3140:        while (i < (int)stringsz) {
1.16      schwarze 3141:                /*
                   3142:                 * Rudimentary roff copy mode:
                   3143:                 * Handle escaped backslashes.
                   3144:                 */
1.42      schwarze 3145:                if ('\\' == string[i] && '\\' == string[i + 1])
                   3146:                        i++;
                   3147:                *c++ = string[i++];
1.16      schwarze 3148:        }
1.8       schwarze 3149:
1.16      schwarze 3150:        /* Append terminating bytes. */
1.66      schwarze 3151:        if (1 < append)
1.16      schwarze 3152:                *c++ = '\n';
1.42      schwarze 3153:
1.16      schwarze 3154:        *c = '\0';
1.42      schwarze 3155:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 3156: }
                   3157:
1.8       schwarze 3158: static const char *
                   3159: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7       schwarze 3160: {
1.42      schwarze 3161:        const struct roffkv *n;
1.64      schwarze 3162:        int i;
1.7       schwarze 3163:
1.42      schwarze 3164:        for (n = r->strtab; n; n = n->next)
1.80      schwarze 3165:                if (0 == strncmp(name, n->key.p, len) &&
                   3166:                    '\0' == n->key.p[(int)len])
1.149     schwarze 3167:                        return n->val.p;
1.64      schwarze 3168:
                   3169:        for (i = 0; i < PREDEFS_MAX; i++)
                   3170:                if (0 == strncmp(name, predefs[i].name, len) &&
                   3171:                                '\0' == predefs[i].name[(int)len])
1.149     schwarze 3172:                        return predefs[i].str;
1.8       schwarze 3173:
1.149     schwarze 3174:        return NULL;
1.7       schwarze 3175: }
                   3176:
1.8       schwarze 3177: static void
1.42      schwarze 3178: roff_freestr(struct roffkv *r)
1.7       schwarze 3179: {
1.42      schwarze 3180:        struct roffkv    *n, *nn;
1.7       schwarze 3181:
1.42      schwarze 3182:        for (n = r; n; n = nn) {
                   3183:                free(n->key.p);
                   3184:                free(n->val.p);
1.7       schwarze 3185:                nn = n->next;
                   3186:                free(n);
                   3187:        }
1.27      schwarze 3188: }
1.138     schwarze 3189:
                   3190: /* --- accessors and utility functions ------------------------------------ */
1.27      schwarze 3191:
                   3192: const struct tbl_span *
                   3193: roff_span(const struct roff *r)
                   3194: {
1.80      schwarze 3195:
1.149     schwarze 3196:        return r->tbl ? tbl_span(r->tbl) : NULL;
1.32      schwarze 3197: }
                   3198:
                   3199: const struct eqn *
                   3200: roff_eqn(const struct roff *r)
                   3201: {
1.80      schwarze 3202:
1.149     schwarze 3203:        return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.42      schwarze 3204: }
                   3205:
                   3206: /*
                   3207:  * Duplicate an input string, making the appropriate character
                   3208:  * conversations (as stipulated by `tr') along the way.
                   3209:  * Returns a heap-allocated string with all the replacements made.
                   3210:  */
                   3211: char *
                   3212: roff_strdup(const struct roff *r, const char *p)
                   3213: {
                   3214:        const struct roffkv *cp;
                   3215:        char            *res;
                   3216:        const char      *pp;
                   3217:        size_t           ssz, sz;
                   3218:        enum mandoc_esc  esc;
                   3219:
                   3220:        if (NULL == r->xmbtab && NULL == r->xtab)
1.149     schwarze 3221:                return mandoc_strdup(p);
1.42      schwarze 3222:        else if ('\0' == *p)
1.149     schwarze 3223:                return mandoc_strdup("");
1.42      schwarze 3224:
                   3225:        /*
                   3226:         * Step through each character looking for term matches
                   3227:         * (remember that a `tr' can be invoked with an escape, which is
                   3228:         * a glyph but the escape is multi-character).
                   3229:         * We only do this if the character hash has been initialised
                   3230:         * and the string is >0 length.
                   3231:         */
                   3232:
                   3233:        res = NULL;
                   3234:        ssz = 0;
                   3235:
                   3236:        while ('\0' != *p) {
1.161     schwarze 3237:                assert((unsigned int)*p < 128);
                   3238:                if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42      schwarze 3239:                        sz = r->xtab[(int)*p].sz;
                   3240:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3241:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3242:                        ssz += sz;
                   3243:                        p++;
                   3244:                        continue;
                   3245:                } else if ('\\' != *p) {
                   3246:                        res = mandoc_realloc(res, ssz + 2);
                   3247:                        res[ssz++] = *p++;
                   3248:                        continue;
                   3249:                }
                   3250:
                   3251:                /* Search for term matches. */
                   3252:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3253:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3254:                                break;
                   3255:
                   3256:                if (NULL != cp) {
                   3257:                        /*
                   3258:                         * A match has been found.
                   3259:                         * Append the match to the array and move
                   3260:                         * forward by its keysize.
                   3261:                         */
1.80      schwarze 3262:                        res = mandoc_realloc(res,
                   3263:                            ssz + cp->val.sz + 1);
1.42      schwarze 3264:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3265:                        ssz += cp->val.sz;
                   3266:                        p += (int)cp->key.sz;
                   3267:                        continue;
                   3268:                }
                   3269:
                   3270:                /*
                   3271:                 * Handle escapes carefully: we need to copy
                   3272:                 * over just the escape itself, or else we might
                   3273:                 * do replacements within the escape itself.
                   3274:                 * Make sure to pass along the bogus string.
                   3275:                 */
                   3276:                pp = p++;
                   3277:                esc = mandoc_escape(&p, NULL, NULL);
                   3278:                if (ESCAPE_ERROR == esc) {
                   3279:                        sz = strlen(pp);
                   3280:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3281:                        memcpy(res + ssz, pp, sz);
                   3282:                        break;
                   3283:                }
1.80      schwarze 3284:                /*
                   3285:                 * We bail out on bad escapes.
1.42      schwarze 3286:                 * No need to warn: we already did so when
                   3287:                 * roff_res() was called.
                   3288:                 */
                   3289:                sz = (int)(p - pp);
                   3290:                res = mandoc_realloc(res, ssz + sz + 1);
                   3291:                memcpy(res + ssz, pp, sz);
                   3292:                ssz += sz;
                   3293:        }
                   3294:
                   3295:        res[(int)ssz] = '\0';
1.149     schwarze 3296:        return res;
1.99      schwarze 3297: }
                   3298:
                   3299: int
                   3300: roff_getformat(const struct roff *r)
                   3301: {
                   3302:
1.149     schwarze 3303:        return r->format;
1.48      schwarze 3304: }
                   3305:
                   3306: /*
1.80      schwarze 3307:  * Find out whether a line is a macro line or not.
1.48      schwarze 3308:  * If it is, adjust the current position and return one; if it isn't,
                   3309:  * return zero and don't change the current position.
                   3310:  * If the control character has been set with `.cc', then let that grain
                   3311:  * precedence.
                   3312:  * This is slighly contrary to groff, where using the non-breaking
                   3313:  * control character when `cc' has been invoked will cause the
                   3314:  * non-breaking macro contents to be printed verbatim.
                   3315:  */
                   3316: int
                   3317: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3318: {
                   3319:        int             pos;
                   3320:
                   3321:        pos = *ppos;
                   3322:
                   3323:        if (0 != r->control && cp[pos] == r->control)
                   3324:                pos++;
                   3325:        else if (0 != r->control)
1.149     schwarze 3326:                return 0;
1.48      schwarze 3327:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3328:                pos += 2;
                   3329:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3330:                pos++;
                   3331:        else
1.149     schwarze 3332:                return 0;
1.48      schwarze 3333:
                   3334:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3335:                pos++;
                   3336:
                   3337:        *ppos = pos;
1.149     schwarze 3338:        return 1;
1.1       schwarze 3339: }