Annotation of src/usr.bin/mandoc/roff.c, Revision 1.247
1.247 ! schwarze 1: /* $OpenBSD: roff.c,v 1.246 2020/04/08 11:54:14 schwarze Exp $ */
1.1 schwarze 2: /*
1.244 schwarze 3: * Copyright (c) 2010-2015, 2017-2020 Ingo Schwarze <schwarze@openbsd.org>
1.139 schwarze 4: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
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.
1.244 schwarze 17: *
18: * Implementation of the roff(7) parser for mandoc(1).
1.1 schwarze 19: */
1.99 schwarze 20: #include <sys/types.h>
21:
1.1 schwarze 22: #include <assert.h>
1.3 schwarze 23: #include <ctype.h>
1.117 schwarze 24: #include <limits.h>
1.167 schwarze 25: #include <stddef.h>
26: #include <stdint.h>
1.51 schwarze 27: #include <stdio.h>
1.1 schwarze 28: #include <stdlib.h>
29: #include <string.h>
30:
1.74 schwarze 31: #include "mandoc_aux.h"
1.167 schwarze 32: #include "mandoc_ohash.h"
1.217 schwarze 33: #include "mandoc.h"
1.137 schwarze 34: #include "roff.h"
1.221 schwarze 35: #include "mandoc_parse.h"
1.98 daniel 36: #include "libmandoc.h"
1.138 schwarze 37: #include "roff_int.h"
1.218 schwarze 38: #include "tbl_parse.h"
1.219 schwarze 39: #include "eqn_parse.h"
1.2 schwarze 40:
1.227 schwarze 41: /*
42: * ASCII_ESC is used to signal from roff_getarg() to roff_expand()
43: * that an escape sequence resulted from copy-in processing and
44: * needs to be checked or interpolated. As it is used nowhere
45: * else, it is defined here rather than in a header file.
46: */
47: #define ASCII_ESC 27
48:
1.43 schwarze 49: /* Maximum number of string expansions per line, to break infinite loops. */
50: #define EXPAND_LIMIT 1000
51:
1.187 schwarze 52: /* Types of definitions of macros and strings. */
53: #define ROFFDEF_USER (1 << 1) /* User-defined. */
54: #define ROFFDEF_PRE (1 << 2) /* Predefined. */
55: #define ROFFDEF_REN (1 << 3) /* Renamed standard macro. */
56: #define ROFFDEF_STD (1 << 4) /* mdoc(7) or man(7) macro. */
57: #define ROFFDEF_ANY (ROFFDEF_USER | ROFFDEF_PRE | \
58: ROFFDEF_REN | ROFFDEF_STD)
1.197 schwarze 59: #define ROFFDEF_UNDEF (1 << 5) /* Completely undefined. */
1.187 schwarze 60:
1.138 schwarze 61: /* --- data types --------------------------------------------------------- */
62:
1.41 schwarze 63: /*
1.42 schwarze 64: * An incredibly-simple string buffer.
65: */
1.8 schwarze 66: struct roffstr {
1.42 schwarze 67: char *p; /* nil-terminated buffer */
68: size_t sz; /* saved strlen(p) */
69: };
70:
71: /*
72: * A key-value roffstr pair as part of a singly-linked list.
73: */
74: struct roffkv {
75: struct roffstr key;
76: struct roffstr val;
77: struct roffkv *next; /* next in list */
1.8 schwarze 78: };
79:
1.52 schwarze 80: /*
81: * A single number register as part of a singly-linked list.
82: */
83: struct roffreg {
84: struct roffstr key;
1.53 schwarze 85: int val;
1.199 schwarze 86: int step;
1.52 schwarze 87: struct roffreg *next;
88: };
89:
1.167 schwarze 90: /*
91: * Association of request and macro names with token IDs.
92: */
93: struct roffreq {
94: enum roff_tok tok;
95: char name[];
96: };
97:
1.211 schwarze 98: /*
99: * A macro processing context.
100: * More than one is needed when macro calls are nested.
101: */
102: struct mctx {
103: char **argv;
104: int argc;
105: int argsz;
106: };
107:
1.1 schwarze 108: struct roff {
1.168 schwarze 109: struct roff_man *man; /* mdoc or man parser */
1.1 schwarze 110: struct roffnode *last; /* leaf of stack */
1.211 schwarze 111: struct mctx *mstack; /* stack of macro contexts */
1.96 schwarze 112: int *rstack; /* stack of inverted `ie' values */
1.167 schwarze 113: struct ohash *reqtab; /* request lookup table */
1.52 schwarze 114: struct roffreg *regtab; /* number registers */
1.42 schwarze 115: struct roffkv *strtab; /* user-defined strings & macros */
1.178 schwarze 116: struct roffkv *rentab; /* renamed strings & macros */
1.42 schwarze 117: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
118: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.16 schwarze 119: const char *current_string; /* value of last called user macro */
1.27 schwarze 120: struct tbl_node *first_tbl; /* first table parsed */
121: struct tbl_node *last_tbl; /* last table parsed */
122: struct tbl_node *tbl; /* current table being parsed */
1.191 schwarze 123: struct eqn_node *last_eqn; /* equation parser */
124: struct eqn_node *eqn; /* active equation parser */
1.102 schwarze 125: int eqn_inline; /* current equation is inline */
1.96 schwarze 126: int options; /* parse options */
1.211 schwarze 127: int mstacksz; /* current size of mstack */
128: int mstackpos; /* position in mstack */
1.96 schwarze 129: int rstacksz; /* current size limit of rstack */
130: int rstackpos; /* position in rstack */
1.99 schwarze 131: int format; /* current file in mdoc or man format */
1.96 schwarze 132: char control; /* control character */
1.175 schwarze 133: char escape; /* escape character */
1.1 schwarze 134: };
135:
1.236 schwarze 136: /*
137: * A macro definition, condition, or ignored block.
138: */
1.1 schwarze 139: struct roffnode {
1.166 schwarze 140: enum roff_tok tok; /* type of node */
1.1 schwarze 141: struct roffnode *parent; /* up one in stack */
142: int line; /* parse line */
143: int col; /* parse col */
1.16 schwarze 144: char *name; /* node name, e.g. macro name */
1.236 schwarze 145: char *end; /* custom end macro of the block */
146: int endspan; /* scope to: 1=eol 2=next line -1=\} */
147: int rule; /* content is: 1=evaluated 0=skipped */
1.1 schwarze 148: };
149:
150: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.166 schwarze 151: enum roff_tok tok, /* tok of macro */ \
1.110 schwarze 152: struct buf *buf, /* input buffer */ \
1.1 schwarze 153: int ln, /* parse line */ \
1.2 schwarze 154: int ppos, /* original pos in buffer */ \
155: int pos, /* current pos in buffer */ \
156: int *offs /* reset offset of buffer data */
1.1 schwarze 157:
1.212 schwarze 158: typedef int (*roffproc)(ROFF_ARGS);
1.1 schwarze 159:
160: struct roffmac {
1.2 schwarze 161: roffproc proc; /* process new macro */
162: roffproc text; /* process as child text of macro */
163: roffproc sub; /* process as child of macro */
164: int flags;
165: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.1 schwarze 166: };
167:
1.37 schwarze 168: struct predef {
169: const char *name; /* predefined input name */
170: const char *str; /* replacement symbol */
171: };
172:
173: #define PREDEF(__name, __str) \
174: { (__name), (__str) },
175:
1.138 schwarze 176: /* --- function prototypes ------------------------------------------------ */
177:
1.212 schwarze 178: static int roffnode_cleanscope(struct roff *);
179: static int roffnode_pop(struct roff *);
1.166 schwarze 180: static void roffnode_push(struct roff *, enum roff_tok,
1.42 schwarze 181: const char *, int, int);
1.218 schwarze 182: static void roff_addtbl(struct roff_man *, int, struct tbl_node *);
1.212 schwarze 183: static int roff_als(ROFF_ARGS);
184: static int roff_block(ROFF_ARGS);
185: static int roff_block_text(ROFF_ARGS);
186: static int roff_block_sub(ROFF_ARGS);
1.236 schwarze 187: static int roff_break(ROFF_ARGS);
1.212 schwarze 188: static int roff_cblock(ROFF_ARGS);
189: static int roff_cc(ROFF_ARGS);
190: static int roff_ccond(struct roff *, int, int);
1.213 schwarze 191: static int roff_char(ROFF_ARGS);
1.212 schwarze 192: static int roff_cond(ROFF_ARGS);
1.247 ! schwarze 193: static int roff_cond_checkend(ROFF_ARGS);
1.212 schwarze 194: static int roff_cond_text(ROFF_ARGS);
195: static int roff_cond_sub(ROFF_ARGS);
196: static int roff_ds(ROFF_ARGS);
197: static int roff_ec(ROFF_ARGS);
198: static int roff_eo(ROFF_ARGS);
199: static int roff_eqndelim(struct roff *, struct buf *, int);
1.244 schwarze 200: static int roff_evalcond(struct roff *, int, char *, int *);
1.106 schwarze 201: static int roff_evalnum(struct roff *, int,
202: const char *, int *, int *, int);
203: static int roff_evalpar(struct roff *, int,
1.133 schwarze 204: const char *, int *, int *, int);
1.71 schwarze 205: static int roff_evalstrcond(const char *, int *);
1.227 schwarze 206: static int roff_expand(struct roff *, struct buf *,
207: int, int, char);
1.42 schwarze 208: static void roff_free1(struct roff *);
1.52 schwarze 209: static void roff_freereg(struct roffreg *);
1.42 schwarze 210: static void roff_freestr(struct roffkv *);
1.85 schwarze 211: static size_t roff_getname(struct roff *, char **, int, int);
1.133 schwarze 212: static int roff_getnum(const char *, int *, int *, int);
1.56 schwarze 213: static int roff_getop(const char *, int *, char *);
1.199 schwarze 214: static int roff_getregn(struct roff *,
215: const char *, size_t, char);
1.145 schwarze 216: static int roff_getregro(const struct roff *,
217: const char *name);
1.197 schwarze 218: static const char *roff_getstrn(struct roff *,
1.187 schwarze 219: const char *, size_t, int *);
1.143 schwarze 220: static int roff_hasregn(const struct roff *,
221: const char *, size_t);
1.212 schwarze 222: static int roff_insec(ROFF_ARGS);
223: static int roff_it(ROFF_ARGS);
224: static int roff_line_ignore(ROFF_ARGS);
1.137 schwarze 225: static void roff_man_alloc1(struct roff_man *);
226: static void roff_man_free1(struct roff_man *);
1.212 schwarze 227: static int roff_manyarg(ROFF_ARGS);
1.230 schwarze 228: static int roff_noarg(ROFF_ARGS);
1.212 schwarze 229: static int roff_nop(ROFF_ARGS);
230: static int roff_nr(ROFF_ARGS);
231: static int roff_onearg(ROFF_ARGS);
1.166 schwarze 232: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.87 schwarze 233: int, int);
1.212 schwarze 234: static int roff_parsetext(struct roff *, struct buf *,
1.177 schwarze 235: int, int *);
1.212 schwarze 236: static int roff_renamed(ROFF_ARGS);
237: static int roff_return(ROFF_ARGS);
238: static int roff_rm(ROFF_ARGS);
239: static int roff_rn(ROFF_ARGS);
240: static int roff_rr(ROFF_ARGS);
1.198 schwarze 241: static void roff_setregn(struct roff *, const char *,
1.199 schwarze 242: size_t, int, char, int);
1.8 schwarze 243: static void roff_setstr(struct roff *,
1.16 schwarze 244: const char *, const char *, int);
1.80 schwarze 245: static void roff_setstrn(struct roffkv **, const char *,
1.42 schwarze 246: size_t, const char *, size_t, int);
1.212 schwarze 247: static int roff_shift(ROFF_ARGS);
248: static int roff_so(ROFF_ARGS);
249: static int roff_tr(ROFF_ARGS);
250: static int roff_Dd(ROFF_ARGS);
251: static int roff_TE(ROFF_ARGS);
252: static int roff_TS(ROFF_ARGS);
253: static int roff_EQ(ROFF_ARGS);
254: static int roff_EN(ROFF_ARGS);
255: static int roff_T_(ROFF_ARGS);
256: static int roff_unsupp(ROFF_ARGS);
257: static int roff_userdef(ROFF_ARGS);
1.1 schwarze 258:
1.138 schwarze 259: /* --- constant data ------------------------------------------------------ */
260:
1.133 schwarze 261: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
262: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
263:
1.166 schwarze 264: const char *__roff_name[MAN_MAX + 1] = {
1.230 schwarze 265: "br", "ce", "fi", "ft",
266: "ll", "mc", "nf",
267: "po", "rj", "sp",
1.184 schwarze 268: "ta", "ti", NULL,
1.166 schwarze 269: "ab", "ad", "af", "aln",
270: "als", "am", "am1", "ami",
271: "ami1", "as", "as1", "asciify",
272: "backtrace", "bd", "bleedat", "blm",
273: "box", "boxa", "bp", "BP",
274: "break", "breakchar", "brnl", "brp",
1.177 schwarze 275: "brpnl", "c2", "cc",
1.166 schwarze 276: "cf", "cflags", "ch", "char",
277: "chop", "class", "close", "CL",
278: "color", "composite", "continue", "cp",
279: "cropat", "cs", "cu", "da",
280: "dch", "Dd", "de", "de1",
281: "defcolor", "dei", "dei1", "device",
282: "devicem", "di", "do", "ds",
283: "ds1", "dwh", "dt", "ec",
284: "ecr", "ecs", "el", "em",
285: "EN", "eo", "EP", "EQ",
286: "errprint", "ev", "evc", "ex",
287: "fallback", "fam", "fc", "fchar",
288: "fcolor", "fdeferlig", "feature", "fkern",
289: "fl", "flig", "fp", "fps",
290: "fschar", "fspacewidth", "fspecial", "ftr",
291: "fzoom", "gcolor", "hc", "hcode",
292: "hidechar", "hla", "hlm", "hpf",
293: "hpfa", "hpfcode", "hw", "hy",
294: "hylang", "hylen", "hym", "hypp",
295: "hys", "ie", "if", "ig",
296: "index", "it", "itc", "IX",
297: "kern", "kernafter", "kernbefore", "kernpair",
298: "lc", "lc_ctype", "lds", "length",
299: "letadj", "lf", "lg", "lhang",
300: "linetabs", "lnr", "lnrf", "lpfx",
1.176 schwarze 301: "ls", "lsm", "lt",
1.166 schwarze 302: "mediasize", "minss", "mk", "mso",
303: "na", "ne", "nh", "nhychar",
304: "nm", "nn", "nop", "nr",
305: "nrf", "nroff", "ns", "nx",
306: "open", "opena", "os", "output",
307: "padj", "papersize", "pc", "pev",
308: "pi", "PI", "pl", "pm",
1.184 schwarze 309: "pn", "pnr", "ps",
1.166 schwarze 310: "psbb", "pshape", "pso", "ptr",
311: "pvs", "rchar", "rd", "recursionlimit",
1.181 schwarze 312: "return", "rfschar", "rhang",
1.166 schwarze 313: "rm", "rn", "rnn", "rr",
314: "rs", "rt", "schar", "sentchar",
315: "shc", "shift", "sizes", "so",
316: "spacewidth", "special", "spreadwarn", "ss",
317: "sty", "substring", "sv", "sy",
1.172 schwarze 318: "T&", "tc", "TE",
1.173 schwarze 319: "TH", "tkf", "tl",
1.166 schwarze 320: "tm", "tm1", "tmc", "tr",
321: "track", "transchar", "trf", "trimat",
322: "trin", "trnt", "troff", "TS",
323: "uf", "ul", "unformat", "unwatch",
324: "unwatchn", "vpt", "vs", "warn",
325: "warnscale", "watch", "watchlength", "watchn",
326: "wh", "while", "write", "writec",
327: "writem", "xflag", ".", NULL,
1.178 schwarze 328: NULL, "text",
1.166 schwarze 329: "Dd", "Dt", "Os", "Sh",
330: "Ss", "Pp", "D1", "Dl",
331: "Bd", "Ed", "Bl", "El",
332: "It", "Ad", "An", "Ap",
333: "Ar", "Cd", "Cm", "Dv",
334: "Er", "Ev", "Ex", "Fa",
335: "Fd", "Fl", "Fn", "Ft",
336: "Ic", "In", "Li", "Nd",
337: "Nm", "Op", "Ot", "Pa",
338: "Rv", "St", "Va", "Vt",
339: "Xr", "%A", "%B", "%D",
340: "%I", "%J", "%N", "%O",
341: "%P", "%R", "%T", "%V",
342: "Ac", "Ao", "Aq", "At",
343: "Bc", "Bf", "Bo", "Bq",
344: "Bsx", "Bx", "Db", "Dc",
345: "Do", "Dq", "Ec", "Ef",
346: "Em", "Eo", "Fx", "Ms",
347: "No", "Ns", "Nx", "Ox",
348: "Pc", "Pf", "Po", "Pq",
349: "Qc", "Ql", "Qo", "Qq",
350: "Re", "Rs", "Sc", "So",
351: "Sq", "Sm", "Sx", "Sy",
352: "Tn", "Ux", "Xc", "Xo",
353: "Fo", "Fc", "Oo", "Oc",
354: "Bk", "Ek", "Bt", "Hf",
355: "Fr", "Ud", "Lb", "Lp",
356: "Lk", "Mt", "Brq", "Bro",
357: "Brc", "%C", "Es", "En",
1.171 schwarze 358: "Dx", "%Q", "%U", "Ta",
1.241 schwarze 359: "Tg", NULL,
1.166 schwarze 360: "TH", "SH", "SS", "TP",
1.204 schwarze 361: "TQ",
1.166 schwarze 362: "LP", "PP", "P", "IP",
363: "HP", "SM", "SB", "BI",
364: "IB", "BR", "RB", "R",
365: "B", "I", "IR", "RI",
366: "RE", "RS", "DT", "UC",
1.169 schwarze 367: "PD", "AT", "in",
1.205 schwarze 368: "SY", "YS", "OP",
369: "EX", "EE", "UR",
1.189 bentley 370: "UE", "MT", "ME", NULL
1.166 schwarze 371: };
372: const char *const *roff_name = __roff_name;
373:
374: static struct roffmac roffs[TOKEN_NONE] = {
1.230 schwarze 375: { roff_noarg, NULL, NULL, 0 }, /* br */
1.177 schwarze 376: { roff_onearg, NULL, NULL, 0 }, /* ce */
1.230 schwarze 377: { roff_noarg, NULL, NULL, 0 }, /* fi */
1.169 schwarze 378: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.170 schwarze 379: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.176 schwarze 380: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.230 schwarze 381: { roff_noarg, NULL, NULL, 0 }, /* nf */
1.184 schwarze 382: { roff_onearg, NULL, NULL, 0 }, /* po */
1.181 schwarze 383: { roff_onearg, NULL, NULL, 0 }, /* rj */
1.171 schwarze 384: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.172 schwarze 385: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.173 schwarze 386: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.168 schwarze 387: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.167 schwarze 388: { roff_unsupp, NULL, NULL, 0 }, /* ab */
389: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
390: { roff_line_ignore, NULL, NULL, 0 }, /* af */
391: { roff_unsupp, NULL, NULL, 0 }, /* aln */
1.183 schwarze 392: { roff_als, NULL, NULL, 0 }, /* als */
1.167 schwarze 393: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
394: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
395: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
396: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
397: { roff_ds, NULL, NULL, 0 }, /* as */
398: { roff_ds, NULL, NULL, 0 }, /* as1 */
399: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
400: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
401: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
402: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
403: { roff_unsupp, NULL, NULL, 0 }, /* blm */
404: { roff_unsupp, NULL, NULL, 0 }, /* box */
405: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
406: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
407: { roff_unsupp, NULL, NULL, 0 }, /* BP */
1.236 schwarze 408: { roff_break, NULL, NULL, 0 }, /* break */
1.167 schwarze 409: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
410: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.230 schwarze 411: { roff_noarg, NULL, NULL, 0 }, /* brp */
1.167 schwarze 412: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
413: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
414: { roff_cc, NULL, NULL, 0 }, /* cc */
415: { roff_insec, NULL, NULL, 0 }, /* cf */
416: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
417: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
1.213 schwarze 418: { roff_char, NULL, NULL, 0 }, /* char */
1.167 schwarze 419: { roff_unsupp, NULL, NULL, 0 }, /* chop */
420: { roff_line_ignore, NULL, NULL, 0 }, /* class */
421: { roff_insec, NULL, NULL, 0 }, /* close */
422: { roff_unsupp, NULL, NULL, 0 }, /* CL */
423: { roff_line_ignore, NULL, NULL, 0 }, /* color */
424: { roff_unsupp, NULL, NULL, 0 }, /* composite */
425: { roff_unsupp, NULL, NULL, 0 }, /* continue */
426: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
427: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
428: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
429: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
430: { roff_unsupp, NULL, NULL, 0 }, /* da */
431: { roff_unsupp, NULL, NULL, 0 }, /* dch */
432: { roff_Dd, NULL, NULL, 0 }, /* Dd */
433: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
434: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
435: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
436: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
437: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
438: { roff_unsupp, NULL, NULL, 0 }, /* device */
439: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
440: { roff_unsupp, NULL, NULL, 0 }, /* di */
441: { roff_unsupp, NULL, NULL, 0 }, /* do */
442: { roff_ds, NULL, NULL, 0 }, /* ds */
443: { roff_ds, NULL, NULL, 0 }, /* ds1 */
444: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
445: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.175 schwarze 446: { roff_ec, NULL, NULL, 0 }, /* ec */
1.167 schwarze 447: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
448: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
449: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
450: { roff_unsupp, NULL, NULL, 0 }, /* em */
451: { roff_EN, NULL, NULL, 0 }, /* EN */
1.175 schwarze 452: { roff_eo, NULL, NULL, 0 }, /* eo */
1.167 schwarze 453: { roff_unsupp, NULL, NULL, 0 }, /* EP */
454: { roff_EQ, NULL, NULL, 0 }, /* EQ */
455: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
456: { roff_unsupp, NULL, NULL, 0 }, /* ev */
457: { roff_unsupp, NULL, NULL, 0 }, /* evc */
458: { roff_unsupp, NULL, NULL, 0 }, /* ex */
459: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
460: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
461: { roff_unsupp, NULL, NULL, 0 }, /* fc */
462: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
463: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
464: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
465: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
466: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
467: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
468: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
469: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
470: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
471: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
472: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
473: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
474: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
475: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
476: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
477: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
478: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
479: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
480: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
481: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
482: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
483: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
484: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
485: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
486: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
487: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
488: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
489: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
490: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
491: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
492: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
493: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
494: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
495: { roff_unsupp, NULL, NULL, 0 }, /* index */
496: { roff_it, NULL, NULL, 0 }, /* it */
497: { roff_unsupp, NULL, NULL, 0 }, /* itc */
498: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
499: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
500: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
501: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
502: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
503: { roff_unsupp, NULL, NULL, 0 }, /* lc */
504: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
505: { roff_unsupp, NULL, NULL, 0 }, /* lds */
506: { roff_unsupp, NULL, NULL, 0 }, /* length */
507: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
508: { roff_insec, NULL, NULL, 0 }, /* lf */
509: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
510: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
511: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
512: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
513: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
514: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
515: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
516: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
517: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
518: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
519: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
520: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
521: { roff_insec, NULL, NULL, 0 }, /* mso */
522: { roff_line_ignore, NULL, NULL, 0 }, /* na */
523: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
524: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
525: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
526: { roff_unsupp, NULL, NULL, 0 }, /* nm */
527: { roff_unsupp, NULL, NULL, 0 }, /* nn */
1.202 schwarze 528: { roff_nop, NULL, NULL, 0 }, /* nop */
1.167 schwarze 529: { roff_nr, NULL, NULL, 0 }, /* nr */
530: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
531: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
532: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
533: { roff_insec, NULL, NULL, 0 }, /* nx */
534: { roff_insec, NULL, NULL, 0 }, /* open */
535: { roff_insec, NULL, NULL, 0 }, /* opena */
536: { roff_line_ignore, NULL, NULL, 0 }, /* os */
537: { roff_unsupp, NULL, NULL, 0 }, /* output */
538: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
539: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
540: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
541: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
542: { roff_insec, NULL, NULL, 0 }, /* pi */
543: { roff_unsupp, NULL, NULL, 0 }, /* PI */
544: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
545: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
546: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
547: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
548: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
549: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
550: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
551: { roff_insec, NULL, NULL, 0 }, /* pso */
552: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
553: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
554: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
555: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
556: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
1.211 schwarze 557: { roff_return, NULL, NULL, 0 }, /* return */
1.167 schwarze 558: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
559: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
560: { roff_rm, NULL, NULL, 0 }, /* rm */
1.178 schwarze 561: { roff_rn, NULL, NULL, 0 }, /* rn */
1.167 schwarze 562: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
563: { roff_rr, NULL, NULL, 0 }, /* rr */
564: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
565: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
566: { roff_unsupp, NULL, NULL, 0 }, /* schar */
567: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
568: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
1.211 schwarze 569: { roff_shift, NULL, NULL, 0 }, /* shift */
1.167 schwarze 570: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
571: { roff_so, NULL, NULL, 0 }, /* so */
572: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
573: { roff_line_ignore, NULL, NULL, 0 }, /* special */
574: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
575: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
576: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
577: { roff_unsupp, NULL, NULL, 0 }, /* substring */
578: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
579: { roff_insec, NULL, NULL, 0 }, /* sy */
580: { roff_T_, NULL, NULL, 0 }, /* T& */
581: { roff_unsupp, NULL, NULL, 0 }, /* tc */
582: { roff_TE, NULL, NULL, 0 }, /* TE */
1.187 schwarze 583: { roff_Dd, NULL, NULL, 0 }, /* TH */
1.167 schwarze 584: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
585: { roff_unsupp, NULL, NULL, 0 }, /* tl */
586: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
587: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
588: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
589: { roff_tr, NULL, NULL, 0 }, /* tr */
590: { roff_line_ignore, NULL, NULL, 0 }, /* track */
591: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
592: { roff_insec, NULL, NULL, 0 }, /* trf */
593: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
594: { roff_unsupp, NULL, NULL, 0 }, /* trin */
595: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
596: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
597: { roff_TS, NULL, NULL, 0 }, /* TS */
598: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
599: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
600: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
601: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
602: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
603: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
604: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
605: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
606: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
607: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
608: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
609: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
610: { roff_unsupp, NULL, NULL, 0 }, /* wh */
1.212 schwarze 611: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /*while*/
1.167 schwarze 612: { roff_insec, NULL, NULL, 0 }, /* write */
613: { roff_insec, NULL, NULL, 0 }, /* writec */
614: { roff_insec, NULL, NULL, 0 }, /* writem */
615: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
616: { roff_cblock, NULL, NULL, 0 }, /* . */
1.178 schwarze 617: { roff_renamed, NULL, NULL, 0 },
1.167 schwarze 618: { roff_userdef, NULL, NULL, 0 }
1.1 schwarze 619: };
620:
1.37 schwarze 621: /* Array of injected predefined strings. */
622: #define PREDEFS_MAX 38
623: static const struct predef predefs[PREDEFS_MAX] = {
624: #include "predefs.in"
625: };
626:
1.177 schwarze 627: static int roffce_lines; /* number of input lines to center */
628: static struct roff_node *roffce_node; /* active request */
1.51 schwarze 629: static int roffit_lines; /* number of lines to delay */
630: static char *roffit_macro; /* nil-terminated macro line */
631:
1.80 schwarze 632:
1.138 schwarze 633: /* --- request table ------------------------------------------------------ */
634:
1.167 schwarze 635: struct ohash *
636: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3 schwarze 637: {
1.167 schwarze 638: struct ohash *htab;
639: struct roffreq *req;
640: enum roff_tok tok;
641: size_t sz;
642: unsigned int slot;
1.3 schwarze 643:
1.167 schwarze 644: htab = mandoc_malloc(sizeof(*htab));
645: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3 schwarze 646:
1.167 schwarze 647: for (tok = mintok; tok < maxtok; tok++) {
1.168 schwarze 648: if (roff_name[tok] == NULL)
649: continue;
1.167 schwarze 650: sz = strlen(roff_name[tok]);
651: req = mandoc_malloc(sizeof(*req) + sz + 1);
652: req->tok = tok;
653: memcpy(req->name, roff_name[tok], sz + 1);
654: slot = ohash_qlookup(htab, req->name);
655: ohash_insert(htab, slot, req);
1.3 schwarze 656: }
1.167 schwarze 657: return htab;
1.3 schwarze 658: }
659:
1.167 schwarze 660: void
661: roffhash_free(struct ohash *htab)
1.1 schwarze 662: {
1.167 schwarze 663: struct roffreq *req;
664: unsigned int slot;
1.1 schwarze 665:
1.167 schwarze 666: if (htab == NULL)
667: return;
668: for (req = ohash_first(htab, &slot); req != NULL;
669: req = ohash_next(htab, &slot))
670: free(req);
671: ohash_delete(htab);
672: free(htab);
673: }
674:
675: enum roff_tok
676: roffhash_find(struct ohash *htab, const char *name, size_t sz)
677: {
678: struct roffreq *req;
679: const char *end;
680:
681: if (sz) {
682: end = name + sz;
683: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
684: } else
685: req = ohash_find(htab, ohash_qlookup(htab, name));
686: return req == NULL ? TOKEN_NONE : req->tok;
1.1 schwarze 687: }
688:
1.138 schwarze 689: /* --- stack of request blocks -------------------------------------------- */
690:
1.1 schwarze 691: /*
692: * Pop the current node off of the stack of roff instructions currently
1.236 schwarze 693: * pending. Return 1 if it is a loop or 0 otherwise.
1.1 schwarze 694: */
1.212 schwarze 695: static int
1.1 schwarze 696: roffnode_pop(struct roff *r)
697: {
698: struct roffnode *p;
1.212 schwarze 699: int inloop;
1.1 schwarze 700:
1.80 schwarze 701: p = r->last;
1.212 schwarze 702: inloop = p->tok == ROFF_while;
703: r->last = p->parent;
1.16 schwarze 704: free(p->name);
705: free(p->end);
1.1 schwarze 706: free(p);
1.212 schwarze 707: return inloop;
1.1 schwarze 708: }
709:
710: /*
711: * Push a roff node onto the instruction stack. This must later be
712: * removed with roffnode_pop().
713: */
1.11 schwarze 714: static void
1.166 schwarze 715: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16 schwarze 716: int line, int col)
1.1 schwarze 717: {
718: struct roffnode *p;
719:
1.11 schwarze 720: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 721: p->tok = tok;
1.16 schwarze 722: if (name)
723: p->name = mandoc_strdup(name);
1.1 schwarze 724: p->parent = r->last;
725: p->line = line;
726: p->col = col;
1.71 schwarze 727: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 728:
729: r->last = p;
730: }
731:
1.138 schwarze 732: /* --- roff parser state data management ---------------------------------- */
733:
1.1 schwarze 734: static void
735: roff_free1(struct roff *r)
736: {
1.42 schwarze 737: int i;
1.27 schwarze 738:
1.218 schwarze 739: tbl_free(r->first_tbl);
1.27 schwarze 740: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 741:
1.219 schwarze 742: eqn_free(r->last_eqn);
1.191 schwarze 743: r->last_eqn = r->eqn = NULL;
1.32 schwarze 744:
1.211 schwarze 745: while (r->mstackpos >= 0)
746: roff_userret(r);
747:
1.1 schwarze 748: while (r->last)
749: roffnode_pop(r);
1.27 schwarze 750:
1.96 schwarze 751: free (r->rstack);
752: r->rstack = NULL;
753: r->rstacksz = 0;
754: r->rstackpos = -1;
755:
756: roff_freereg(r->regtab);
757: r->regtab = NULL;
758:
1.42 schwarze 759: roff_freestr(r->strtab);
1.178 schwarze 760: roff_freestr(r->rentab);
1.42 schwarze 761: roff_freestr(r->xmbtab);
1.178 schwarze 762: r->strtab = r->rentab = r->xmbtab = NULL;
1.42 schwarze 763:
764: if (r->xtab)
765: for (i = 0; i < 128; i++)
766: free(r->xtab[i].p);
767: free(r->xtab);
768: r->xtab = NULL;
1.1 schwarze 769: }
770:
771: void
772: roff_reset(struct roff *r)
773: {
774: roff_free1(r);
1.239 schwarze 775: r->options |= MPARSE_COMMENT;
1.99 schwarze 776: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175 schwarze 777: r->control = '\0';
778: r->escape = '\\';
1.179 schwarze 779: roffce_lines = 0;
780: roffce_node = NULL;
781: roffit_lines = 0;
782: roffit_macro = NULL;
1.1 schwarze 783: }
784:
785: void
786: roff_free(struct roff *r)
787: {
1.238 schwarze 788: int i;
1.211 schwarze 789:
1.1 schwarze 790: roff_free1(r);
1.211 schwarze 791: for (i = 0; i < r->mstacksz; i++)
792: free(r->mstack[i].argv);
793: free(r->mstack);
1.167 schwarze 794: roffhash_free(r->reqtab);
1.1 schwarze 795: free(r);
796: }
797:
798: struct roff *
1.223 schwarze 799: roff_alloc(int options)
1.1 schwarze 800: {
801: struct roff *r;
802:
1.11 schwarze 803: r = mandoc_calloc(1, sizeof(struct roff));
1.200 schwarze 804: r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.239 schwarze 805: r->options = options | MPARSE_COMMENT;
1.99 schwarze 806: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.211 schwarze 807: r->mstackpos = -1;
1.2 schwarze 808: r->rstackpos = -1;
1.175 schwarze 809: r->escape = '\\';
1.149 schwarze 810: return r;
1.137 schwarze 811: }
812:
1.138 schwarze 813: /* --- syntax tree state data management ---------------------------------- */
814:
1.137 schwarze 815: static void
816: roff_man_free1(struct roff_man *man)
817: {
1.228 schwarze 818: if (man->meta.first != NULL)
819: roff_node_delete(man, man->meta.first);
1.137 schwarze 820: free(man->meta.msec);
821: free(man->meta.vol);
822: free(man->meta.os);
823: free(man->meta.arch);
824: free(man->meta.title);
825: free(man->meta.name);
826: free(man->meta.date);
1.228 schwarze 827: free(man->meta.sodest);
1.137 schwarze 828: }
829:
1.229 schwarze 830: void
831: roff_state_reset(struct roff_man *man)
832: {
833: man->last = man->meta.first;
834: man->last_es = NULL;
835: man->flags = 0;
836: man->lastsec = man->lastnamed = SEC_NONE;
837: man->next = ROFF_NEXT_CHILD;
838: roff_setreg(man->roff, "nS", 0, '=');
839: }
840:
1.137 schwarze 841: static void
842: roff_man_alloc1(struct roff_man *man)
843: {
844: memset(&man->meta, 0, sizeof(man->meta));
1.228 schwarze 845: man->meta.first = mandoc_calloc(1, sizeof(*man->meta.first));
846: man->meta.first->type = ROFFT_ROOT;
847: man->meta.macroset = MACROSET_NONE;
1.229 schwarze 848: roff_state_reset(man);
1.137 schwarze 849: }
850:
851: void
852: roff_man_reset(struct roff_man *man)
853: {
854: roff_man_free1(man);
855: roff_man_alloc1(man);
856: }
857:
858: void
859: roff_man_free(struct roff_man *man)
860: {
861: roff_man_free1(man);
862: free(man);
863: }
864:
865: struct roff_man *
1.223 schwarze 866: roff_man_alloc(struct roff *roff, const char *os_s, int quick)
1.137 schwarze 867: {
868: struct roff_man *man;
869:
870: man = mandoc_calloc(1, sizeof(*man));
871: man->roff = roff;
1.188 schwarze 872: man->os_s = os_s;
1.137 schwarze 873: man->quick = quick;
874: roff_man_alloc1(man);
1.168 schwarze 875: roff->man = man;
1.149 schwarze 876: return man;
1.1 schwarze 877: }
878:
1.138 schwarze 879: /* --- syntax tree handling ----------------------------------------------- */
880:
881: struct roff_node *
882: roff_node_alloc(struct roff_man *man, int line, int pos,
883: enum roff_type type, int tok)
884: {
885: struct roff_node *n;
886:
887: n = mandoc_calloc(1, sizeof(*n));
888: n->line = line;
889: n->pos = pos;
890: n->tok = tok;
891: n->type = type;
892: n->sec = man->lastsec;
893:
894: if (man->flags & MDOC_SYNOPSIS)
1.157 schwarze 895: n->flags |= NODE_SYNPRETTY;
1.138 schwarze 896: else
1.157 schwarze 897: n->flags &= ~NODE_SYNPRETTY;
1.232 schwarze 898: if ((man->flags & (ROFF_NOFILL | ROFF_NONOFILL)) == ROFF_NOFILL)
1.231 schwarze 899: n->flags |= NODE_NOFILL;
900: else
901: n->flags &= ~NODE_NOFILL;
1.138 schwarze 902: if (man->flags & MDOC_NEWLINE)
1.157 schwarze 903: n->flags |= NODE_LINE;
1.138 schwarze 904: man->flags &= ~MDOC_NEWLINE;
905:
1.149 schwarze 906: return n;
1.138 schwarze 907: }
908:
909: void
910: roff_node_append(struct roff_man *man, struct roff_node *n)
911: {
912:
913: switch (man->next) {
914: case ROFF_NEXT_SIBLING:
1.153 schwarze 915: if (man->last->next != NULL) {
916: n->next = man->last->next;
917: man->last->next->prev = n;
918: } else
919: man->last->parent->last = n;
1.138 schwarze 920: man->last->next = n;
921: n->prev = man->last;
922: n->parent = man->last->parent;
923: break;
924: case ROFF_NEXT_CHILD:
1.159 schwarze 925: if (man->last->child != NULL) {
926: n->next = man->last->child;
927: man->last->child->prev = n;
928: } else
929: man->last->last = n;
1.138 schwarze 930: man->last->child = n;
931: n->parent = man->last;
932: break;
933: default:
934: abort();
935: }
1.154 schwarze 936: man->last = n;
1.138 schwarze 937:
938: switch (n->type) {
939: case ROFFT_HEAD:
940: n->parent->head = n;
941: break;
942: case ROFFT_BODY:
1.154 schwarze 943: if (n->end != ENDBODY_NOT)
944: return;
1.138 schwarze 945: n->parent->body = n;
946: break;
947: case ROFFT_TAIL:
948: n->parent->tail = n;
949: break;
950: default:
1.154 schwarze 951: return;
1.138 schwarze 952: }
1.154 schwarze 953:
954: /*
955: * Copy over the normalised-data pointer of our parent. Not
956: * everybody has one, but copying a null pointer is fine.
957: */
958:
959: n->norm = n->parent->norm;
960: assert(n->parent->type == ROFFT_BLOCK);
1.138 schwarze 961: }
962:
1.139 schwarze 963: void
964: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
965: {
966: struct roff_node *n;
967:
968: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
969: n->string = roff_strdup(man->roff, word);
970: roff_node_append(man, n);
1.158 schwarze 971: n->flags |= NODE_VALID | NODE_ENDED;
1.139 schwarze 972: man->next = ROFF_NEXT_SIBLING;
973: }
974:
975: void
976: roff_word_append(struct roff_man *man, const char *word)
977: {
978: struct roff_node *n;
979: char *addstr, *newstr;
980:
981: n = man->last;
982: addstr = roff_strdup(man->roff, word);
983: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
984: free(addstr);
985: free(n->string);
986: n->string = newstr;
987: man->next = ROFF_NEXT_SIBLING;
1.140 schwarze 988: }
989:
990: void
991: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
992: {
993: struct roff_node *n;
994:
995: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
996: roff_node_append(man, n);
997: man->next = ROFF_NEXT_CHILD;
998: }
999:
1000: struct roff_node *
1001: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
1002: {
1003: struct roff_node *n;
1004:
1005: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
1006: roff_node_append(man, n);
1007: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 1008: return n;
1.139 schwarze 1009: }
1010:
1.138 schwarze 1011: struct roff_node *
1012: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
1013: {
1014: struct roff_node *n;
1015:
1016: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
1017: roff_node_append(man, n);
1018: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 1019: return n;
1.138 schwarze 1020: }
1021:
1022: struct roff_node *
1023: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1024: {
1025: struct roff_node *n;
1026:
1027: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1028: roff_node_append(man, n);
1029: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 1030: return n;
1.139 schwarze 1031: }
1032:
1.193 schwarze 1033: static void
1.218 schwarze 1034: roff_addtbl(struct roff_man *man, int line, struct tbl_node *tbl)
1.139 schwarze 1035: {
1036: struct roff_node *n;
1.218 schwarze 1037: struct tbl_span *span;
1.139 schwarze 1038:
1.228 schwarze 1039: if (man->meta.macroset == MACROSET_MAN)
1.174 schwarze 1040: man_breakscope(man, ROFF_TS);
1.193 schwarze 1041: while ((span = tbl_span(tbl)) != NULL) {
1.218 schwarze 1042: n = roff_node_alloc(man, line, 0, ROFFT_TBL, TOKEN_NONE);
1.193 schwarze 1043: n->span = span;
1044: roff_node_append(man, n);
1045: n->flags |= NODE_VALID | NODE_ENDED;
1046: man->next = ROFF_NEXT_SIBLING;
1047: }
1.138 schwarze 1048: }
1049:
1050: void
1051: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1052: {
1053:
1054: /* Adjust siblings. */
1055:
1056: if (n->prev)
1057: n->prev->next = n->next;
1058: if (n->next)
1059: n->next->prev = n->prev;
1060:
1061: /* Adjust parent. */
1062:
1063: if (n->parent != NULL) {
1064: if (n->parent->child == n)
1065: n->parent->child = n->next;
1066: if (n->parent->last == n)
1067: n->parent->last = n->prev;
1068: }
1069:
1070: /* Adjust parse point. */
1071:
1072: if (man == NULL)
1073: return;
1074: if (man->last == n) {
1075: if (n->prev == NULL) {
1076: man->last = n->parent;
1077: man->next = ROFF_NEXT_CHILD;
1078: } else {
1079: man->last = n->prev;
1080: man->next = ROFF_NEXT_SIBLING;
1081: }
1082: }
1.228 schwarze 1083: if (man->meta.first == n)
1084: man->meta.first = NULL;
1.216 schwarze 1085: }
1086:
1087: void
1088: roff_node_relink(struct roff_man *man, struct roff_node *n)
1089: {
1090: roff_node_unlink(man, n);
1091: n->prev = n->next = NULL;
1092: roff_node_append(man, n);
1.138 schwarze 1093: }
1094:
1095: void
1096: roff_node_free(struct roff_node *n)
1097: {
1098:
1099: if (n->args != NULL)
1100: mdoc_argv_free(n->args);
1101: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1102: free(n->norm);
1.219 schwarze 1103: eqn_box_free(n->eqn);
1.138 schwarze 1104: free(n->string);
1.246 schwarze 1105: free(n->tag);
1.138 schwarze 1106: free(n);
1107: }
1108:
1109: void
1110: roff_node_delete(struct roff_man *man, struct roff_node *n)
1111: {
1112:
1113: while (n->child != NULL)
1114: roff_node_delete(man, n->child);
1115: roff_node_unlink(man, n);
1116: roff_node_free(n);
1.242 schwarze 1117: }
1118:
1119: int
1120: roff_node_transparent(struct roff_node *n)
1121: {
1122: if (n == NULL)
1123: return 0;
1124: if (n->type == ROFFT_COMMENT || n->flags & NODE_NOPRT)
1125: return 1;
1.245 schwarze 1126: return roff_tok_transparent(n->tok);
1127: }
1128:
1129: int
1130: roff_tok_transparent(enum roff_tok tok)
1131: {
1132: switch (tok) {
1.242 schwarze 1133: case ROFF_ft:
1134: case ROFF_ll:
1135: case ROFF_mc:
1136: case ROFF_po:
1137: case ROFF_ta:
1138: case MDOC_Db:
1139: case MDOC_Es:
1140: case MDOC_Sm:
1141: case MDOC_Tg:
1142: case MAN_DT:
1143: case MAN_UC:
1144: case MAN_PD:
1145: case MAN_AT:
1146: return 1;
1147: default:
1148: return 0;
1149: }
1150: }
1151:
1152: struct roff_node *
1153: roff_node_child(struct roff_node *n)
1154: {
1155: for (n = n->child; roff_node_transparent(n); n = n->next)
1156: continue;
1157: return n;
1158: }
1159:
1160: struct roff_node *
1161: roff_node_prev(struct roff_node *n)
1162: {
1163: do {
1164: n = n->prev;
1165: } while (roff_node_transparent(n));
1166: return n;
1167: }
1168:
1169: struct roff_node *
1170: roff_node_next(struct roff_node *n)
1171: {
1172: do {
1173: n = n->next;
1174: } while (roff_node_transparent(n));
1175: return n;
1.141 schwarze 1176: }
1177:
1178: void
1179: deroff(char **dest, const struct roff_node *n)
1180: {
1181: char *cp;
1182: size_t sz;
1183:
1.243 schwarze 1184: if (n->string == NULL) {
1.141 schwarze 1185: for (n = n->child; n != NULL; n = n->next)
1186: deroff(dest, n);
1187: return;
1188: }
1189:
1.160 schwarze 1190: /* Skip leading whitespace. */
1.141 schwarze 1191:
1.160 schwarze 1192: for (cp = n->string; *cp != '\0'; cp++) {
1.161 schwarze 1193: if (cp[0] == '\\' && cp[1] != '\0' &&
1194: strchr(" %&0^|~", cp[1]) != NULL)
1.141 schwarze 1195: cp++;
1.160 schwarze 1196: else if ( ! isspace((unsigned char)*cp))
1.141 schwarze 1197: break;
1198: }
1199:
1.161 schwarze 1200: /* Skip trailing backslash. */
1201:
1202: sz = strlen(cp);
1.162 schwarze 1203: if (sz > 0 && cp[sz - 1] == '\\')
1.161 schwarze 1204: sz--;
1205:
1.141 schwarze 1206: /* Skip trailing whitespace. */
1207:
1.161 schwarze 1208: for (; sz; sz--)
1.141 schwarze 1209: if ( ! isspace((unsigned char)cp[sz-1]))
1210: break;
1211:
1212: /* Skip empty strings. */
1213:
1214: if (sz == 0)
1215: return;
1216:
1217: if (*dest == NULL) {
1218: *dest = mandoc_strndup(cp, sz);
1219: return;
1220: }
1221:
1222: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1223: free(*dest);
1224: *dest = cp;
1.138 schwarze 1225: }
1226:
1227: /* --- main functions of the roff parser ---------------------------------- */
1228:
1.8 schwarze 1229: /*
1.227 schwarze 1230: * In the current line, expand escape sequences that produce parsable
1231: * input text. Also check the syntax of the remaining escape sequences,
1232: * which typically produce output glyphs or change formatter state.
1.8 schwarze 1233: */
1.212 schwarze 1234: static int
1.227 schwarze 1235: roff_expand(struct roff *r, struct buf *buf, int ln, int pos, char newesc)
1.8 schwarze 1236: {
1.211 schwarze 1237: struct mctx *ctx; /* current macro call context */
1.81 schwarze 1238: char ubuf[24]; /* buffer to print the number */
1.200 schwarze 1239: struct roff_node *n; /* used for header comments */
1.78 schwarze 1240: const char *start; /* start of the string to process */
1.82 schwarze 1241: char *stesc; /* start of an escape sequence ('\\') */
1.224 schwarze 1242: const char *esct; /* type of esccape sequence */
1.200 schwarze 1243: char *ep; /* end of comment string */
1.23 schwarze 1244: const char *stnam; /* start of the name, after "[(*" */
1245: const char *cp; /* end of the name, e.g. before ']' */
1246: const char *res; /* the string to be substituted */
1.110 schwarze 1247: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 1248: size_t maxl; /* expected length of the escape name */
1249: size_t naml; /* actual length of the escape name */
1.211 schwarze 1250: size_t asz; /* length of the replacement */
1251: size_t rsz; /* length of the rest of the string */
1.109 schwarze 1252: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 1253: int expand_count; /* to avoid infinite loops */
1.79 schwarze 1254: int npos; /* position in numeric expression */
1.91 schwarze 1255: int arg_complete; /* argument not interrupted by eol */
1.211 schwarze 1256: int quote_args; /* true for \\$@, false for \\$* */
1.175 schwarze 1257: int done; /* no more input available */
1.187 schwarze 1258: int deftype; /* type of definition to paste */
1.195 schwarze 1259: int rcsid; /* kind of RCS id seen */
1.224 schwarze 1260: enum mandocerr err; /* for escape sequence problems */
1.199 schwarze 1261: char sign; /* increment number register */
1.79 schwarze 1262: char term; /* character terminating the escape */
1.8 schwarze 1263:
1.175 schwarze 1264: /* Search forward for comments. */
1265:
1266: done = 0;
1267: start = buf->buf + pos;
1268: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1.227 schwarze 1269: if (stesc[0] != newesc || stesc[1] == '\0')
1.175 schwarze 1270: continue;
1271: stesc++;
1272: if (*stesc != '"' && *stesc != '#')
1273: continue;
1.185 schwarze 1274:
1275: /* Comment found, look for RCS id. */
1276:
1.195 schwarze 1277: rcsid = 0;
1.185 schwarze 1278: if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.195 schwarze 1279: rcsid = 1 << MANDOC_OS_OPENBSD;
1.185 schwarze 1280: cp += 8;
1281: } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.195 schwarze 1282: rcsid = 1 << MANDOC_OS_NETBSD;
1.185 schwarze 1283: cp += 7;
1284: }
1285: if (cp != NULL &&
1286: isalnum((unsigned char)*cp) == 0 &&
1.186 schwarze 1287: strchr(cp, '$') != NULL) {
1.195 schwarze 1288: if (r->man->meta.rcsids & rcsid)
1.222 schwarze 1289: mandoc_msg(MANDOCERR_RCS_REP, ln,
1290: (int)(stesc - buf->buf) + 1,
1291: "%s", stesc + 1);
1.195 schwarze 1292: r->man->meta.rcsids |= rcsid;
1.186 schwarze 1293: }
1.185 schwarze 1294:
1295: /* Handle trailing whitespace. */
1296:
1.200 schwarze 1297: ep = strchr(stesc--, '\0') - 1;
1298: if (*ep == '\n') {
1.175 schwarze 1299: done = 1;
1.200 schwarze 1300: ep--;
1.175 schwarze 1301: }
1.200 schwarze 1302: if (*ep == ' ' || *ep == '\t')
1.222 schwarze 1303: mandoc_msg(MANDOCERR_SPACE_EOL,
1304: ln, (int)(ep - buf->buf), NULL);
1.200 schwarze 1305:
1306: /*
1307: * Save comments preceding the title macro
1308: * in the syntax tree.
1309: */
1310:
1.239 schwarze 1311: if (newesc != ASCII_ESC && r->options & MPARSE_COMMENT) {
1.200 schwarze 1312: while (*ep == ' ' || *ep == '\t')
1313: ep--;
1314: ep[1] = '\0';
1315: n = roff_node_alloc(r->man,
1316: ln, stesc + 1 - buf->buf,
1317: ROFFT_COMMENT, TOKEN_NONE);
1318: n->string = mandoc_strdup(stesc + 2);
1319: roff_node_append(r->man, n);
1320: n->flags |= NODE_VALID | NODE_ENDED;
1321: r->man->next = ROFF_NEXT_SIBLING;
1322: }
1323:
1.206 schwarze 1324: /* Line continuation with comment. */
1325:
1326: if (stesc[1] == '#') {
1327: *stesc = '\0';
1.212 schwarze 1328: return ROFF_IGN | ROFF_APPEND;
1.206 schwarze 1329: }
1330:
1331: /* Discard normal comments. */
1.200 schwarze 1332:
1.207 schwarze 1333: while (stesc > start && stesc[-1] == ' ' &&
1334: (stesc == start + 1 || stesc[-2] != '\\'))
1.175 schwarze 1335: stesc--;
1336: *stesc = '\0';
1337: break;
1338: }
1339: if (stesc == start)
1340: return ROFF_CONT;
1341: stesc--;
1342:
1343: /* Notice the end of the input. */
1344:
1345: if (*stesc == '\n') {
1346: *stesc-- = '\0';
1347: done = 1;
1348: }
1349:
1.43 schwarze 1350: expand_count = 0;
1.175 schwarze 1351: while (stesc >= start) {
1.227 schwarze 1352: if (*stesc != newesc) {
1.43 schwarze 1353:
1.227 schwarze 1354: /*
1355: * If we have a non-standard escape character,
1356: * escape literal backslashes because all
1357: * processing in subsequent functions uses
1358: * the standard escaping rules.
1359: */
1.78 schwarze 1360:
1.227 schwarze 1361: if (newesc != ASCII_ESC && *stesc == '\\') {
1.175 schwarze 1362: *stesc = '\0';
1363: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1364: buf->buf, stesc + 1) + 1;
1365: start = nbuf + pos;
1366: stesc = nbuf + (stesc - buf->buf);
1367: free(buf->buf);
1368: buf->buf = nbuf;
1369: }
1.227 schwarze 1370:
1371: /* Search backwards for the next escape. */
1372:
1.175 schwarze 1373: stesc--;
1.78 schwarze 1374: continue;
1.175 schwarze 1375: }
1.78 schwarze 1376:
1377: /* If it is escaped, skip it. */
1378:
1379: for (cp = stesc - 1; cp >= start; cp--)
1.175 schwarze 1380: if (*cp != r->escape)
1.78 schwarze 1381: break;
1382:
1.110 schwarze 1383: if ((stesc - cp) % 2 == 0) {
1.175 schwarze 1384: while (stesc > cp)
1385: *stesc-- = '\\';
1.78 schwarze 1386: continue;
1.175 schwarze 1387: } else if (stesc[1] != '\0') {
1388: *stesc = '\\';
1389: } else {
1390: *stesc-- = '\0';
1391: if (done)
1392: continue;
1393: else
1.212 schwarze 1394: return ROFF_IGN | ROFF_APPEND;
1.78 schwarze 1395: }
1.23 schwarze 1396:
1.79 schwarze 1397: /* Decide whether to expand or to check only. */
1.23 schwarze 1398:
1.79 schwarze 1399: term = '\0';
1.78 schwarze 1400: cp = stesc + 1;
1.224 schwarze 1401: if (*cp == 'E')
1402: cp++;
1403: esct = cp;
1404: switch (*esct) {
1.80 schwarze 1405: case '*':
1.211 schwarze 1406: case '$':
1.53 schwarze 1407: res = NULL;
1408: break;
1.80 schwarze 1409: case 'B':
1410: case 'w':
1.79 schwarze 1411: term = cp[1];
1412: /* FALLTHROUGH */
1.80 schwarze 1413: case 'n':
1.199 schwarze 1414: sign = cp[1];
1415: if (sign == '+' || sign == '-')
1416: cp++;
1.53 schwarze 1417: res = ubuf;
1418: break;
1419: default:
1.224 schwarze 1420: err = MANDOCERR_OK;
1421: switch(mandoc_escape(&cp, &stnam, &inaml)) {
1422: case ESCAPE_SPECIAL:
1423: if (mchars_spec2cp(stnam, inaml) >= 0)
1424: break;
1425: /* FALLTHROUGH */
1426: case ESCAPE_ERROR:
1427: err = MANDOCERR_ESC_BAD;
1428: break;
1429: case ESCAPE_UNDEF:
1430: err = MANDOCERR_ESC_UNDEF;
1431: break;
1432: case ESCAPE_UNSUPP:
1433: err = MANDOCERR_ESC_UNSUPP;
1434: break;
1435: default:
1436: break;
1437: }
1438: if (err != MANDOCERR_OK)
1439: mandoc_msg(err, ln, (int)(stesc - buf->buf),
1.92 schwarze 1440: "%.*s", (int)(cp - stesc), stesc);
1.175 schwarze 1441: stesc--;
1.78 schwarze 1442: continue;
1.42 schwarze 1443: }
1444:
1.78 schwarze 1445: if (EXPAND_LIMIT < ++expand_count) {
1.222 schwarze 1446: mandoc_msg(MANDOCERR_ROFFLOOP,
1.110 schwarze 1447: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1448: return ROFF_IGN;
1.78 schwarze 1449: }
1.23 schwarze 1450:
1451: /*
1452: * The third character decides the length
1.53 schwarze 1453: * of the name of the string or register.
1.23 schwarze 1454: * Save a pointer to the name.
1455: */
1456:
1.110 schwarze 1457: if (term == '\0') {
1.79 schwarze 1458: switch (*++cp) {
1.80 schwarze 1459: case '\0':
1.79 schwarze 1460: maxl = 0;
1461: break;
1.80 schwarze 1462: case '(':
1.79 schwarze 1463: cp++;
1464: maxl = 2;
1465: break;
1.80 schwarze 1466: case '[':
1.79 schwarze 1467: cp++;
1468: term = ']';
1469: maxl = 0;
1470: break;
1471: default:
1472: maxl = 1;
1473: break;
1474: }
1475: } else {
1476: cp += 2;
1.8 schwarze 1477: maxl = 0;
1478: }
1.23 schwarze 1479: stnam = cp;
1.8 schwarze 1480:
1.23 schwarze 1481: /* Advance to the end of the name. */
1.8 schwarze 1482:
1.125 schwarze 1483: naml = 0;
1.91 schwarze 1484: arg_complete = 1;
1.125 schwarze 1485: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1486: if (*cp == '\0') {
1.222 schwarze 1487: mandoc_msg(MANDOCERR_ESC_BAD, ln,
1488: (int)(stesc - buf->buf), "%s", stesc);
1.91 schwarze 1489: arg_complete = 0;
1.79 schwarze 1490: break;
1.42 schwarze 1491: }
1.110 schwarze 1492: if (maxl == 0 && *cp == term) {
1.79 schwarze 1493: cp++;
1.125 schwarze 1494: break;
1495: }
1.224 schwarze 1496: if (*cp++ != '\\' || *esct != 'w') {
1.125 schwarze 1497: naml++;
1498: continue;
1499: }
1500: switch (mandoc_escape(&cp, NULL, NULL)) {
1501: case ESCAPE_SPECIAL:
1502: case ESCAPE_UNICODE:
1503: case ESCAPE_NUMBERED:
1.224 schwarze 1504: case ESCAPE_UNDEF:
1.125 schwarze 1505: case ESCAPE_OVERSTRIKE:
1506: naml++;
1507: break;
1508: default:
1.8 schwarze 1509: break;
1.79 schwarze 1510: }
1.8 schwarze 1511: }
1512:
1.23 schwarze 1513: /*
1514: * Retrieve the replacement string; if it is
1515: * undefined, resume searching for escapes.
1516: */
1517:
1.224 schwarze 1518: switch (*esct) {
1.80 schwarze 1519: case '*':
1.187 schwarze 1520: if (arg_complete) {
1521: deftype = ROFFDEF_USER | ROFFDEF_PRE;
1522: res = roff_getstrn(r, stnam, naml, &deftype);
1.203 schwarze 1523:
1524: /*
1525: * If not overriden, let \*(.T
1526: * through to the formatters.
1527: */
1528:
1529: if (res == NULL && naml == 2 &&
1530: stnam[0] == '.' && stnam[1] == 'T') {
1531: roff_setstrn(&r->strtab,
1532: ".T", 2, NULL, 0, 0);
1533: stesc--;
1534: continue;
1535: }
1.187 schwarze 1536: }
1.79 schwarze 1537: break;
1.211 schwarze 1538: case '$':
1539: if (r->mstackpos < 0) {
1.222 schwarze 1540: mandoc_msg(MANDOCERR_ARG_UNDEF, ln,
1541: (int)(stesc - buf->buf), "%.3s", stesc);
1.211 schwarze 1542: break;
1543: }
1544: ctx = r->mstack + r->mstackpos;
1.224 schwarze 1545: npos = esct[1] - '1';
1.211 schwarze 1546: if (npos >= 0 && npos <= 8) {
1547: res = npos < ctx->argc ?
1548: ctx->argv[npos] : "";
1549: break;
1550: }
1.224 schwarze 1551: if (esct[1] == '*')
1.211 schwarze 1552: quote_args = 0;
1.224 schwarze 1553: else if (esct[1] == '@')
1.211 schwarze 1554: quote_args = 1;
1555: else {
1.222 schwarze 1556: mandoc_msg(MANDOCERR_ARG_NONUM, ln,
1557: (int)(stesc - buf->buf), "%.3s", stesc);
1.211 schwarze 1558: break;
1559: }
1560: asz = 0;
1561: for (npos = 0; npos < ctx->argc; npos++) {
1562: if (npos)
1563: asz++; /* blank */
1564: if (quote_args)
1565: asz += 2; /* quotes */
1566: asz += strlen(ctx->argv[npos]);
1567: }
1568: if (asz != 3) {
1569: rsz = buf->sz - (stesc - buf->buf) - 3;
1570: if (asz < 3)
1571: memmove(stesc + asz, stesc + 3, rsz);
1572: buf->sz += asz - 3;
1573: nbuf = mandoc_realloc(buf->buf, buf->sz);
1574: start = nbuf + pos;
1575: stesc = nbuf + (stesc - buf->buf);
1576: buf->buf = nbuf;
1577: if (asz > 3)
1578: memmove(stesc + asz, stesc + 3, rsz);
1579: }
1580: for (npos = 0; npos < ctx->argc; npos++) {
1581: if (npos)
1582: *stesc++ = ' ';
1583: if (quote_args)
1584: *stesc++ = '"';
1585: cp = ctx->argv[npos];
1586: while (*cp != '\0')
1587: *stesc++ = *cp++;
1588: if (quote_args)
1589: *stesc++ = '"';
1590: }
1591: continue;
1.80 schwarze 1592: case 'B':
1.79 schwarze 1593: npos = 0;
1.91 schwarze 1594: ubuf[0] = arg_complete &&
1.133 schwarze 1595: roff_evalnum(r, ln, stnam, &npos,
1596: NULL, ROFFNUM_SCALE) &&
1.91 schwarze 1597: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1598: ubuf[1] = '\0';
1599: break;
1.80 schwarze 1600: case 'n':
1.91 schwarze 1601: if (arg_complete)
1602: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.199 schwarze 1603: roff_getregn(r, stnam, naml, sign));
1.91 schwarze 1604: else
1605: ubuf[0] = '\0';
1.79 schwarze 1606: break;
1.80 schwarze 1607: case 'w':
1.91 schwarze 1608: /* use even incomplete args */
1.81 schwarze 1609: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1610: 24 * (int)naml);
1611: break;
1612: }
1.8 schwarze 1613:
1.110 schwarze 1614: if (res == NULL) {
1.224 schwarze 1615: if (*esct == '*')
1.222 schwarze 1616: mandoc_msg(MANDOCERR_STR_UNDEF,
1617: ln, (int)(stesc - buf->buf),
1.211 schwarze 1618: "%.*s", (int)naml, stnam);
1.37 schwarze 1619: res = "";
1.118 schwarze 1620: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1.222 schwarze 1621: mandoc_msg(MANDOCERR_ROFFLOOP,
1.118 schwarze 1622: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1623: return ROFF_IGN;
1.8 schwarze 1624: }
1625:
1.23 schwarze 1626: /* Replace the escape sequence by the string. */
1627:
1.82 schwarze 1628: *stesc = '\0';
1.110 schwarze 1629: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1630: buf->buf, res, cp) + 1;
1.8 schwarze 1631:
1.78 schwarze 1632: /* Prepare for the next replacement. */
1.8 schwarze 1633:
1.78 schwarze 1634: start = nbuf + pos;
1.110 schwarze 1635: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1636: free(buf->buf);
1637: buf->buf = nbuf;
1.42 schwarze 1638: }
1.149 schwarze 1639: return ROFF_CONT;
1.42 schwarze 1640: }
1.225 schwarze 1641:
1642: /*
1643: * Parse a quoted or unquoted roff-style request or macro argument.
1644: * Return a pointer to the parsed argument, which is either the original
1645: * pointer or advanced by one byte in case the argument is quoted.
1646: * NUL-terminate the argument in place.
1647: * Collapse pairs of quotes inside quoted arguments.
1648: * Advance the argument pointer to the next argument,
1649: * or to the NUL byte terminating the argument line.
1650: */
1651: char *
1.227 schwarze 1652: roff_getarg(struct roff *r, char **cpp, int ln, int *pos)
1.225 schwarze 1653: {
1.227 schwarze 1654: struct buf buf;
1.238 schwarze 1655: char *cp, *start;
1.227 schwarze 1656: int newesc, pairs, quoted, white;
1.225 schwarze 1657:
1658: /* Quoting can only start with a new word. */
1659: start = *cpp;
1660: quoted = 0;
1661: if ('"' == *start) {
1662: quoted = 1;
1663: start++;
1664: }
1665:
1.227 schwarze 1666: newesc = pairs = white = 0;
1.225 schwarze 1667: for (cp = start; '\0' != *cp; cp++) {
1668:
1669: /*
1670: * Move the following text left
1671: * after quoted quotes and after "\\" and "\t".
1672: */
1673: if (pairs)
1674: cp[-pairs] = cp[0];
1675:
1676: if ('\\' == cp[0]) {
1677: /*
1678: * In copy mode, translate double to single
1679: * backslashes and backslash-t to literal tabs.
1680: */
1681: switch (cp[1]) {
1682: case 'a':
1683: case 't':
1.226 schwarze 1684: cp[-pairs] = '\t';
1.227 schwarze 1685: pairs++;
1686: cp++;
1687: break;
1.225 schwarze 1688: case '\\':
1.227 schwarze 1689: newesc = 1;
1690: cp[-pairs] = ASCII_ESC;
1.225 schwarze 1691: pairs++;
1692: cp++;
1693: break;
1694: case ' ':
1695: /* Skip escaped blanks. */
1696: if (0 == quoted)
1697: cp++;
1698: break;
1699: default:
1700: break;
1701: }
1702: } else if (0 == quoted) {
1703: if (' ' == cp[0]) {
1704: /* Unescaped blanks end unquoted args. */
1705: white = 1;
1706: break;
1707: }
1708: } else if ('"' == cp[0]) {
1709: if ('"' == cp[1]) {
1710: /* Quoted quotes collapse. */
1711: pairs++;
1712: cp++;
1713: } else {
1714: /* Unquoted quotes end quoted args. */
1715: quoted = 2;
1716: break;
1717: }
1718: }
1719: }
1720:
1721: /* Quoted argument without a closing quote. */
1722: if (1 == quoted)
1723: mandoc_msg(MANDOCERR_ARG_QUOTE, ln, *pos, NULL);
1724:
1725: /* NUL-terminate this argument and move to the next one. */
1726: if (pairs)
1727: cp[-pairs] = '\0';
1728: if ('\0' != *cp) {
1729: *cp++ = '\0';
1730: while (' ' == *cp)
1731: cp++;
1732: }
1733: *pos += (int)(cp - start) + (quoted ? 1 : 0);
1734: *cpp = cp;
1735:
1736: if ('\0' == *cp && (white || ' ' == cp[-1]))
1737: mandoc_msg(MANDOCERR_SPACE_EOL, ln, *pos, NULL);
1738:
1.227 schwarze 1739: start = mandoc_strdup(start);
1740: if (newesc == 0)
1741: return start;
1742:
1743: buf.buf = start;
1744: buf.sz = strlen(start) + 1;
1745: buf.next = NULL;
1746: if (roff_expand(r, &buf, ln, 0, ASCII_ESC) & ROFF_IGN) {
1747: free(buf.buf);
1748: buf.buf = mandoc_strdup("");
1749: }
1750: return buf.buf;
1.225 schwarze 1751: }
1752:
1.42 schwarze 1753:
1754: /*
1.147 schwarze 1755: * Process text streams.
1.42 schwarze 1756: */
1.212 schwarze 1757: static int
1.177 schwarze 1758: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42 schwarze 1759: {
1760: size_t sz;
1761: const char *start;
1.51 schwarze 1762: char *p;
1763: int isz;
1.42 schwarze 1764: enum mandoc_esc esc;
1765:
1.147 schwarze 1766: /* Spring the input line trap. */
1767:
1768: if (roffit_lines == 1) {
1769: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1770: free(buf->buf);
1771: buf->buf = p;
1772: buf->sz = isz + 1;
1773: *offs = 0;
1774: free(roffit_macro);
1775: roffit_lines = 0;
1.149 schwarze 1776: return ROFF_REPARSE;
1.147 schwarze 1777: } else if (roffit_lines > 1)
1778: --roffit_lines;
1779:
1.177 schwarze 1780: if (roffce_node != NULL && buf->buf[pos] != '\0') {
1781: if (roffce_lines < 1) {
1782: r->man->last = roffce_node;
1783: r->man->next = ROFF_NEXT_SIBLING;
1784: roffce_lines = 0;
1785: roffce_node = NULL;
1786: } else
1787: roffce_lines--;
1788: }
1789:
1.147 schwarze 1790: /* Convert all breakable hyphens into ASCII_HYPH. */
1791:
1.110 schwarze 1792: start = p = buf->buf + pos;
1.42 schwarze 1793:
1.110 schwarze 1794: while (*p != '\0') {
1.42 schwarze 1795: sz = strcspn(p, "-\\");
1796: p += sz;
1797:
1.110 schwarze 1798: if (*p == '\0')
1.42 schwarze 1799: break;
1800:
1.110 schwarze 1801: if (*p == '\\') {
1.42 schwarze 1802: /* Skip over escapes. */
1803: p++;
1.62 schwarze 1804: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1805: if (esc == ESCAPE_ERROR)
1.42 schwarze 1806: break;
1.136 schwarze 1807: while (*p == '-')
1808: p++;
1.42 schwarze 1809: continue;
1810: } else if (p == start) {
1811: p++;
1812: continue;
1813: }
1814:
1.44 schwarze 1815: if (isalpha((unsigned char)p[-1]) &&
1816: isalpha((unsigned char)p[1]))
1.42 schwarze 1817: *p = ASCII_HYPH;
1818: p++;
1.8 schwarze 1819: }
1.149 schwarze 1820: return ROFF_CONT;
1.8 schwarze 1821: }
1822:
1.212 schwarze 1823: int
1.110 schwarze 1824: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1825: {
1.166 schwarze 1826: enum roff_tok t;
1.212 schwarze 1827: int e;
1.110 schwarze 1828: int pos; /* parse point */
1.115 schwarze 1829: int spos; /* saved parse point for messages */
1.110 schwarze 1830: int ppos; /* original offset in buf->buf */
1831: int ctl; /* macro line (boolean) */
1832:
1833: ppos = pos = *offs;
1.1 schwarze 1834:
1.102 schwarze 1835: /* Handle in-line equation delimiters. */
1836:
1.108 schwarze 1837: if (r->tbl == NULL &&
1838: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1839: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1840: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1841: if (e == ROFF_REPARSE)
1.149 schwarze 1842: return e;
1.102 schwarze 1843: assert(e == ROFF_CONT);
1844: }
1845:
1846: /* Expand some escape sequences. */
1.8 schwarze 1847:
1.227 schwarze 1848: e = roff_expand(r, buf, ln, pos, r->escape);
1.212 schwarze 1849: if ((e & ROFF_MASK) == ROFF_IGN)
1.149 schwarze 1850: return e;
1.110 schwarze 1851: assert(e == ROFF_CONT);
1.8 schwarze 1852:
1.110 schwarze 1853: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1854:
1.8 schwarze 1855: /*
1.2 schwarze 1856: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1857: * text through the macro's filter.
1858: * Equations process all content themselves.
1859: * Tables process almost all content themselves, but we want
1860: * to warn about macros before passing it there.
1.2 schwarze 1861: */
1862:
1.124 schwarze 1863: if (r->last != NULL && ! ctl) {
1.2 schwarze 1864: t = r->last->tok;
1.110 schwarze 1865: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.212 schwarze 1866: if ((e & ROFF_MASK) == ROFF_IGN)
1.149 schwarze 1867: return e;
1.212 schwarze 1868: e &= ~ROFF_MASK;
1869: } else
1870: e = ROFF_IGN;
1.191 schwarze 1871: if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
1872: eqn_read(r->eqn, buf->buf + ppos);
1.212 schwarze 1873: return e;
1.191 schwarze 1874: }
1.193 schwarze 1875: if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
1876: tbl_read(r->tbl, ln, buf->buf, ppos);
1.218 schwarze 1877: roff_addtbl(r->man, ln, r->tbl);
1.212 schwarze 1878: return e;
1.193 schwarze 1879: }
1.239 schwarze 1880: if ( ! ctl) {
1881: r->options &= ~MPARSE_COMMENT;
1.212 schwarze 1882: return roff_parsetext(r, buf, pos, offs) | e;
1.239 schwarze 1883: }
1.100 schwarze 1884:
1885: /* Skip empty request lines. */
1886:
1.110 schwarze 1887: if (buf->buf[pos] == '"') {
1.222 schwarze 1888: mandoc_msg(MANDOCERR_COMMENT_BAD, ln, pos, NULL);
1.149 schwarze 1889: return ROFF_IGN;
1.110 schwarze 1890: } else if (buf->buf[pos] == '\0')
1.149 schwarze 1891: return ROFF_IGN;
1.2 schwarze 1892:
1893: /*
1894: * If a scope is open, go to the child handler for that macro,
1895: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1896: * Don't do so if an equation is open.
1.2 schwarze 1897: */
1898:
1899: if (r->last) {
1.1 schwarze 1900: t = r->last->tok;
1.149 schwarze 1901: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1902: }
1903:
1.115 schwarze 1904: /* No scope is open. This is a new request or macro. */
1905:
1.239 schwarze 1906: r->options &= ~MPARSE_COMMENT;
1.115 schwarze 1907: spos = pos;
1908: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1909:
1910: /* Tables ignore most macros. */
1911:
1.180 schwarze 1912: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181 schwarze 1913: t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.222 schwarze 1914: mandoc_msg(MANDOCERR_TBLMACRO,
1915: ln, pos, "%s", buf->buf + spos);
1.180 schwarze 1916: if (t != TOKEN_NONE)
1.149 schwarze 1917: return ROFF_IGN;
1.129 schwarze 1918: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1919: pos++;
1.163 schwarze 1920: while (buf->buf[pos] == ' ')
1.129 schwarze 1921: pos++;
1.193 schwarze 1922: tbl_read(r->tbl, ln, buf->buf, pos);
1.218 schwarze 1923: roff_addtbl(r->man, ln, r->tbl);
1.193 schwarze 1924: return ROFF_IGN;
1.115 schwarze 1925: }
1926:
1.177 schwarze 1927: /* For now, let high level macros abort .ce mode. */
1928:
1929: if (ctl && roffce_node != NULL &&
1.196 schwarze 1930: (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
1931: t == ROFF_TH || t == ROFF_TS)) {
1.177 schwarze 1932: r->man->last = roffce_node;
1933: r->man->next = ROFF_NEXT_SIBLING;
1934: roffce_lines = 0;
1935: roffce_node = NULL;
1936: }
1937:
1.2 schwarze 1938: /*
1.115 schwarze 1939: * This is neither a roff request nor a user-defined macro.
1940: * Let the standard macro set parsers handle it.
1.2 schwarze 1941: */
1942:
1.166 schwarze 1943: if (t == TOKEN_NONE)
1.149 schwarze 1944: return ROFF_CONT;
1.115 schwarze 1945:
1946: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1947:
1.168 schwarze 1948: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2 schwarze 1949: }
1950:
1.211 schwarze 1951: /*
1952: * Internal interface function to tell the roff parser that execution
1953: * of the current macro ended. This is required because macro
1954: * definitions usually do not end with a .return request.
1955: */
1956: void
1957: roff_userret(struct roff *r)
1958: {
1959: struct mctx *ctx;
1960: int i;
1961:
1962: assert(r->mstackpos >= 0);
1963: ctx = r->mstack + r->mstackpos;
1964: for (i = 0; i < ctx->argc; i++)
1965: free(ctx->argv[i]);
1966: ctx->argc = 0;
1967: r->mstackpos--;
1968: }
1969:
1.27 schwarze 1970: void
1.2 schwarze 1971: roff_endparse(struct roff *r)
1972: {
1.193 schwarze 1973: if (r->last != NULL)
1.222 schwarze 1974: mandoc_msg(MANDOCERR_BLK_NOEND, r->last->line,
1975: r->last->col, "%s", roff_name[r->last->tok]);
1.27 schwarze 1976:
1.193 schwarze 1977: if (r->eqn != NULL) {
1.222 schwarze 1978: mandoc_msg(MANDOCERR_BLK_NOEND,
1.191 schwarze 1979: r->eqn->node->line, r->eqn->node->pos, "EQ");
1980: eqn_parse(r->eqn);
1981: r->eqn = NULL;
1.32 schwarze 1982: }
1983:
1.193 schwarze 1984: if (r->tbl != NULL) {
1.218 schwarze 1985: tbl_end(r->tbl, 1);
1.193 schwarze 1986: r->tbl = NULL;
1.27 schwarze 1987: }
1.1 schwarze 1988: }
1989:
1990: /*
1991: * Parse a roff node's type from the input buffer. This must be in the
1992: * form of ".foo xxx" in the usual way.
1993: */
1.166 schwarze 1994: static enum roff_tok
1.87 schwarze 1995: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1996: {
1.87 schwarze 1997: char *cp;
1.16 schwarze 1998: const char *mac;
1999: size_t maclen;
1.187 schwarze 2000: int deftype;
1.166 schwarze 2001: enum roff_tok t;
1.1 schwarze 2002:
1.87 schwarze 2003: cp = buf + *pos;
2004:
2005: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166 schwarze 2006: return TOKEN_NONE;
1.1 schwarze 2007:
1.87 schwarze 2008: mac = cp;
2009: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 2010:
1.187 schwarze 2011: deftype = ROFFDEF_USER | ROFFDEF_REN;
2012: r->current_string = roff_getstrn(r, mac, maclen, &deftype);
2013: switch (deftype) {
2014: case ROFFDEF_USER:
2015: t = ROFF_USERDEF;
2016: break;
2017: case ROFFDEF_REN:
2018: t = ROFF_RENAMED;
2019: break;
2020: default:
2021: t = roffhash_find(r->reqtab, mac, maclen);
2022: break;
2023: }
1.166 schwarze 2024: if (t != TOKEN_NONE)
1.87 schwarze 2025: *pos = cp - buf;
1.197 schwarze 2026: else if (deftype == ROFFDEF_UNDEF) {
2027: /* Using an undefined macro defines it to be empty. */
2028: roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
2029: roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
2030: }
1.149 schwarze 2031: return t;
1.1 schwarze 2032: }
2033:
1.138 schwarze 2034: /* --- handling of request blocks ----------------------------------------- */
2035:
1.247 ! schwarze 2036: /*
! 2037: * Close a macro definition block or an "ignore" block.
! 2038: */
1.212 schwarze 2039: static int
1.2 schwarze 2040: roff_cblock(ROFF_ARGS)
1.1 schwarze 2041: {
1.247 ! schwarze 2042: int rr;
1.2 schwarze 2043:
1.110 schwarze 2044: if (r->last == NULL) {
1.222 schwarze 2045: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.149 schwarze 2046: return ROFF_IGN;
1.2 schwarze 2047: }
1.1 schwarze 2048:
1.2 schwarze 2049: switch (r->last->tok) {
1.80 schwarze 2050: case ROFF_am:
2051: case ROFF_ami:
2052: case ROFF_de:
2053: case ROFF_dei:
2054: case ROFF_ig:
1.2 schwarze 2055: break;
1.247 ! schwarze 2056: case ROFF_am1:
! 2057: case ROFF_de1:
! 2058: /* Remapped in roff_block(). */
! 2059: abort();
1.2 schwarze 2060: default:
1.222 schwarze 2061: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "..");
1.149 schwarze 2062: return ROFF_IGN;
1.2 schwarze 2063: }
2064:
1.247 ! schwarze 2065: roffnode_pop(r);
! 2066: roffnode_cleanscope(r);
! 2067:
! 2068: /*
! 2069: * If a conditional block with braces is still open,
! 2070: * check for "\}" block end markers.
! 2071: */
! 2072:
! 2073: if (r->last != NULL && r->last->endspan < 0) {
! 2074: rr = 1; /* If arguments follow "\}", warn about them. */
! 2075: roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
! 2076: }
! 2077:
1.110 schwarze 2078: if (buf->buf[pos] != '\0')
1.222 schwarze 2079: mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.110 schwarze 2080: ".. %s", buf->buf + pos);
1.2 schwarze 2081:
1.149 schwarze 2082: return ROFF_IGN;
1.2 schwarze 2083: }
1.1 schwarze 2084:
1.236 schwarze 2085: /*
2086: * Pop all nodes ending at the end of the current input line.
2087: * Return the number of loops ended.
2088: */
1.212 schwarze 2089: static int
1.2 schwarze 2090: roffnode_cleanscope(struct roff *r)
2091: {
1.212 schwarze 2092: int inloop;
1.1 schwarze 2093:
1.212 schwarze 2094: inloop = 0;
1.247 ! schwarze 2095: while (r->last != NULL && r->last->endspan > 0) {
1.46 schwarze 2096: if (--r->last->endspan != 0)
1.2 schwarze 2097: break;
1.212 schwarze 2098: inloop += roffnode_pop(r);
1.2 schwarze 2099: }
1.212 schwarze 2100: return inloop;
1.2 schwarze 2101: }
1.1 schwarze 2102:
1.236 schwarze 2103: /*
1.247 ! schwarze 2104: * Handle the closing "\}" of a conditional block.
1.236 schwarze 2105: * Apart from generating warnings, this only pops nodes.
2106: * Return the number of loops ended.
2107: */
1.212 schwarze 2108: static int
1.68 schwarze 2109: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 2110: {
2111: if (NULL == r->last) {
1.222 schwarze 2112: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.212 schwarze 2113: return 0;
1.2 schwarze 2114: }
1.1 schwarze 2115:
1.2 schwarze 2116: switch (r->last->tok) {
1.80 schwarze 2117: case ROFF_el:
2118: case ROFF_ie:
2119: case ROFF_if:
1.212 schwarze 2120: case ROFF_while:
1.2 schwarze 2121: break;
2122: default:
1.222 schwarze 2123: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.212 schwarze 2124: return 0;
1.2 schwarze 2125: }
1.1 schwarze 2126:
1.2 schwarze 2127: if (r->last->endspan > -1) {
1.222 schwarze 2128: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "\\}");
1.212 schwarze 2129: return 0;
1.2 schwarze 2130: }
2131:
1.212 schwarze 2132: return roffnode_pop(r) + roffnode_cleanscope(r);
1.1 schwarze 2133: }
2134:
1.212 schwarze 2135: static int
1.2 schwarze 2136: roff_block(ROFF_ARGS)
1.1 schwarze 2137: {
1.187 schwarze 2138: const char *name, *value;
2139: char *call, *cp, *iname, *rname;
2140: size_t csz, namesz, rsz;
2141: int deftype;
1.16 schwarze 2142:
1.93 schwarze 2143: /* Ignore groff compatibility mode for now. */
1.2 schwarze 2144:
1.110 schwarze 2145: if (tok == ROFF_de1)
1.93 schwarze 2146: tok = ROFF_de;
1.123 schwarze 2147: else if (tok == ROFF_dei1)
2148: tok = ROFF_dei;
1.110 schwarze 2149: else if (tok == ROFF_am1)
1.93 schwarze 2150: tok = ROFF_am;
1.123 schwarze 2151: else if (tok == ROFF_ami1)
2152: tok = ROFF_ami;
1.93 schwarze 2153:
2154: /* Parse the macro name argument. */
2155:
1.110 schwarze 2156: cp = buf->buf + pos;
2157: if (tok == ROFF_ig) {
1.93 schwarze 2158: iname = NULL;
2159: namesz = 0;
2160: } else {
2161: iname = cp;
2162: namesz = roff_getname(r, &cp, ln, ppos);
2163: iname[namesz] = '\0';
2164: }
1.22 schwarze 2165:
1.93 schwarze 2166: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 2167:
1.110 schwarze 2168: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187 schwarze 2169: deftype = ROFFDEF_USER;
2170: name = roff_getstrn(r, iname, namesz, &deftype);
2171: if (name == NULL) {
1.222 schwarze 2172: mandoc_msg(MANDOCERR_STR_UNDEF,
2173: ln, (int)(iname - buf->buf),
1.93 schwarze 2174: "%.*s", (int)namesz, iname);
2175: namesz = 0;
2176: } else
2177: namesz = strlen(name);
2178: } else
2179: name = iname;
1.22 schwarze 2180:
1.110 schwarze 2181: if (namesz == 0 && tok != ROFF_ig) {
1.222 schwarze 2182: mandoc_msg(MANDOCERR_REQ_EMPTY,
2183: ln, ppos, "%s", roff_name[tok]);
1.149 schwarze 2184: return ROFF_IGN;
1.93 schwarze 2185: }
1.2 schwarze 2186:
1.16 schwarze 2187: roffnode_push(r, tok, name, ln, ppos);
2188:
2189: /*
2190: * At the beginning of a `de' macro, clear the existing string
2191: * with the same name, if there is one. New content will be
1.66 schwarze 2192: * appended from roff_block_text() in multiline mode.
1.16 schwarze 2193: */
1.22 schwarze 2194:
1.183 schwarze 2195: if (tok == ROFF_de || tok == ROFF_dei) {
1.86 schwarze 2196: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183 schwarze 2197: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.187 schwarze 2198: } else if (tok == ROFF_am || tok == ROFF_ami) {
2199: deftype = ROFFDEF_ANY;
2200: value = roff_getstrn(r, iname, namesz, &deftype);
2201: switch (deftype) { /* Before appending, ... */
2202: case ROFFDEF_PRE: /* copy predefined to user-defined. */
2203: roff_setstrn(&r->strtab, name, namesz,
2204: value, strlen(value), 0);
2205: break;
2206: case ROFFDEF_REN: /* call original standard macro. */
2207: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
2208: (int)strlen(value), value);
2209: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
2210: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
2211: free(call);
2212: break;
2213: case ROFFDEF_STD: /* rename and call standard macro. */
2214: rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
2215: roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
2216: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
2217: (int)rsz, rname);
2218: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
2219: free(call);
2220: free(rname);
2221: break;
2222: default:
2223: break;
2224: }
1.183 schwarze 2225: }
1.2 schwarze 2226:
1.110 schwarze 2227: if (*cp == '\0')
1.149 schwarze 2228: return ROFF_IGN;
1.1 schwarze 2229:
1.93 schwarze 2230: /* Get the custom end marker. */
1.22 schwarze 2231:
1.93 schwarze 2232: iname = cp;
1.86 schwarze 2233: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 2234:
2235: /* Resolve the end marker if it is indirect. */
2236:
1.110 schwarze 2237: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187 schwarze 2238: deftype = ROFFDEF_USER;
2239: name = roff_getstrn(r, iname, namesz, &deftype);
2240: if (name == NULL) {
1.222 schwarze 2241: mandoc_msg(MANDOCERR_STR_UNDEF,
2242: ln, (int)(iname - buf->buf),
1.93 schwarze 2243: "%.*s", (int)namesz, iname);
2244: namesz = 0;
2245: } else
2246: namesz = strlen(name);
2247: } else
2248: name = iname;
2249:
1.86 schwarze 2250: if (namesz)
2251: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 2252:
1.110 schwarze 2253: if (*cp != '\0')
1.222 schwarze 2254: mandoc_msg(MANDOCERR_ARG_EXCESS,
1.167 schwarze 2255: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1 schwarze 2256:
1.149 schwarze 2257: return ROFF_IGN;
1.1 schwarze 2258: }
2259:
1.212 schwarze 2260: static int
1.2 schwarze 2261: roff_block_sub(ROFF_ARGS)
1.1 schwarze 2262: {
1.166 schwarze 2263: enum roff_tok t;
1.2 schwarze 2264: int i, j;
2265:
2266: /*
2267: * First check whether a custom macro exists at this level. If
2268: * it does, then check against it. This is some of groff's
2269: * stranger behaviours. If we encountered a custom end-scope
2270: * tag and that tag also happens to be a "real" macro, then we
2271: * need to try interpreting it again as a real macro. If it's
2272: * not, then return ignore. Else continue.
2273: */
2274:
2275: if (r->last->end) {
1.35 schwarze 2276: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 2277: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 2278: break;
1.1 schwarze 2279:
1.110 schwarze 2280: if (r->last->end[j] == '\0' &&
2281: (buf->buf[i] == '\0' ||
2282: buf->buf[i] == ' ' ||
2283: buf->buf[i] == '\t')) {
1.2 schwarze 2284: roffnode_pop(r);
2285: roffnode_cleanscope(r);
1.1 schwarze 2286:
1.110 schwarze 2287: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 2288: i++;
2289:
2290: pos = i;
1.110 schwarze 2291: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166 schwarze 2292: TOKEN_NONE)
1.149 schwarze 2293: return ROFF_RERUN;
2294: return ROFF_IGN;
1.2 schwarze 2295: }
1.1 schwarze 2296: }
2297:
1.2 schwarze 2298: /*
2299: * If we have no custom end-query or lookup failed, then try
2300: * pulling it out of the hashtable.
2301: */
1.1 schwarze 2302:
1.110 schwarze 2303: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 2304:
1.110 schwarze 2305: if (t != ROFF_cblock) {
2306: if (tok != ROFF_ig)
2307: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149 schwarze 2308: return ROFF_IGN;
1.16 schwarze 2309: }
1.1 schwarze 2310:
1.149 schwarze 2311: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 2312: }
2313:
1.212 schwarze 2314: static int
1.2 schwarze 2315: roff_block_text(ROFF_ARGS)
2316: {
2317:
1.110 schwarze 2318: if (tok != ROFF_ig)
2319: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 2320:
1.149 schwarze 2321: return ROFF_IGN;
1.2 schwarze 2322: }
2323:
1.247 ! schwarze 2324: /*
! 2325: * Check for a closing "\}" and handle it.
! 2326: * In this function, the final "int *offs" argument is used for
! 2327: * different purposes than elsewhere:
! 2328: * Input: *offs == 0: caller wants to discard arguments following \}
! 2329: * *offs == 1: caller wants to preserve text following \}
! 2330: * Output: *offs = 0: tell caller to discard input line
! 2331: * *offs = 1: tell caller to use input line
! 2332: */
1.212 schwarze 2333: static int
1.247 ! schwarze 2334: roff_cond_checkend(ROFF_ARGS)
1.2 schwarze 2335: {
1.212 schwarze 2336: char *ep;
2337: int endloop, irc, rr;
1.2 schwarze 2338:
1.212 schwarze 2339: irc = ROFF_IGN;
1.2 schwarze 2340: rr = r->last->rule;
1.212 schwarze 2341: endloop = tok != ROFF_while ? ROFF_IGN :
2342: rr ? ROFF_LOOPCONT : ROFF_LOOPEXIT;
2343: if (roffnode_cleanscope(r))
2344: irc |= endloop;
1.39 schwarze 2345:
1.69 schwarze 2346: /*
1.247 ! schwarze 2347: * If "\}" occurs on a macro line without a preceding macro or
! 2348: * a text line contains nothing else, drop the line completely.
1.69 schwarze 2349: */
2350:
1.110 schwarze 2351: ep = buf->buf + pos;
1.247 ! schwarze 2352: if (ep[0] == '\\' && ep[1] == '}' && (ep[2] == '\0' || *offs == 0))
1.71 schwarze 2353: rr = 0;
1.69 schwarze 2354:
1.215 schwarze 2355: /*
1.247 ! schwarze 2356: * The closing delimiter "\}" rewinds the conditional scope
1.215 schwarze 2357: * but is otherwise ignored when interpreting the line.
2358: */
1.39 schwarze 2359:
1.110 schwarze 2360: while ((ep = strchr(ep, '\\')) != NULL) {
1.190 schwarze 2361: switch (ep[1]) {
2362: case '}':
1.247 ! schwarze 2363: if (rr)
! 2364: ep[1] = '&';
! 2365: else
! 2366: memmove(ep, ep + 2, strlen(ep + 2) + 1);
1.212 schwarze 2367: if (roff_ccond(r, ln, ep - buf->buf))
2368: irc |= endloop;
1.190 schwarze 2369: break;
2370: case '\0':
2371: ++ep;
2372: break;
2373: default:
2374: ep += 2;
2375: break;
1.70 schwarze 2376: }
1.50 schwarze 2377: }
1.247 ! schwarze 2378: *offs = rr;
! 2379: return irc;
! 2380: }
1.190 schwarze 2381:
1.247 ! schwarze 2382: /*
! 2383: * Parse and process a request or macro line in conditional scope.
! 2384: */
! 2385: static int
! 2386: roff_cond_sub(ROFF_ARGS)
! 2387: {
! 2388: struct roffnode *bl;
! 2389: int irc, rr;
! 2390: enum roff_tok t;
! 2391:
! 2392: rr = 0; /* If arguments follow "\}", skip them. */
! 2393: irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.240 schwarze 2394: t = roff_parse(r, buf->buf, &pos, ln, ppos);
2395:
2396: /* For now, let high level macros abort .ce mode. */
2397:
2398: if (roffce_node != NULL &&
2399: (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
2400: t == ROFF_TH || t == ROFF_TS)) {
2401: r->man->last = roffce_node;
2402: r->man->next = ROFF_NEXT_SIBLING;
2403: roffce_lines = 0;
2404: roffce_node = NULL;
2405: }
2406:
1.190 schwarze 2407: /*
2408: * Fully handle known macros when they are structurally
2409: * required or when the conditional evaluated to true.
2410: */
2411:
1.236 schwarze 2412: if (t == ROFF_break) {
2413: if (irc & ROFF_LOOPMASK)
2414: irc = ROFF_IGN | ROFF_LOOPEXIT;
2415: else if (rr) {
2416: for (bl = r->last; bl != NULL; bl = bl->parent) {
2417: bl->rule = 0;
2418: if (bl->tok == ROFF_while)
2419: break;
2420: }
2421: }
2422: } else if (t != TOKEN_NONE &&
2423: (rr || roffs[t].flags & ROFFMAC_STRUCT))
2424: irc |= (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
2425: else
2426: irc |= rr ? ROFF_CONT : ROFF_IGN;
1.212 schwarze 2427: return irc;
1.2 schwarze 2428: }
2429:
1.247 ! schwarze 2430: /*
! 2431: * Parse and process a text line in conditional scope.
! 2432: */
1.212 schwarze 2433: static int
1.2 schwarze 2434: roff_cond_text(ROFF_ARGS)
2435: {
1.247 ! schwarze 2436: int irc, rr;
1.215 schwarze 2437:
1.247 ! schwarze 2438: rr = 1; /* If arguments follow "\}", preserve them. */
! 2439: irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
1.212 schwarze 2440: if (rr)
2441: irc |= ROFF_CONT;
2442: return irc;
1.2 schwarze 2443: }
2444:
1.138 schwarze 2445: /* --- handling of numeric and conditional expressions -------------------- */
2446:
1.77 schwarze 2447: /*
2448: * Parse a single signed integer number. Stop at the first non-digit.
2449: * If there is at least one digit, return success and advance the
2450: * parse point, else return failure and let the parse point unchanged.
2451: * Ignore overflows, treat them just like the C language.
2452: */
1.56 schwarze 2453: static int
1.133 schwarze 2454: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56 schwarze 2455: {
1.133 schwarze 2456: int myres, scaled, n, p;
1.79 schwarze 2457:
2458: if (NULL == res)
2459: res = &myres;
1.56 schwarze 2460:
2461: p = *pos;
2462: n = v[p] == '-';
1.133 schwarze 2463: if (n || v[p] == '+')
1.56 schwarze 2464: p++;
2465:
1.133 schwarze 2466: if (flags & ROFFNUM_WHITE)
2467: while (isspace((unsigned char)v[p]))
2468: p++;
2469:
1.56 schwarze 2470: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 2471: *res = 10 * *res + v[p] - '0';
1.56 schwarze 2472: if (p == *pos + n)
2473: return 0;
2474:
2475: if (n)
2476: *res = -*res;
2477:
1.126 schwarze 2478: /* Each number may be followed by one optional scaling unit. */
2479:
2480: switch (v[p]) {
2481: case 'f':
1.133 schwarze 2482: scaled = *res * 65536;
1.126 schwarze 2483: break;
2484: case 'i':
1.133 schwarze 2485: scaled = *res * 240;
1.126 schwarze 2486: break;
2487: case 'c':
1.133 schwarze 2488: scaled = *res * 240 / 2.54;
1.126 schwarze 2489: break;
2490: case 'v':
2491: case 'P':
1.133 schwarze 2492: scaled = *res * 40;
1.126 schwarze 2493: break;
2494: case 'm':
2495: case 'n':
1.133 schwarze 2496: scaled = *res * 24;
1.126 schwarze 2497: break;
2498: case 'p':
1.133 schwarze 2499: scaled = *res * 10 / 3;
1.126 schwarze 2500: break;
2501: case 'u':
1.133 schwarze 2502: scaled = *res;
1.126 schwarze 2503: break;
2504: case 'M':
1.133 schwarze 2505: scaled = *res * 6 / 25;
1.126 schwarze 2506: break;
2507: default:
1.133 schwarze 2508: scaled = *res;
1.126 schwarze 2509: p--;
2510: break;
2511: }
1.133 schwarze 2512: if (flags & ROFFNUM_SCALE)
2513: *res = scaled;
1.126 schwarze 2514:
2515: *pos = p + 1;
1.149 schwarze 2516: return 1;
1.56 schwarze 2517: }
2518:
1.71 schwarze 2519: /*
2520: * Evaluate a string comparison condition.
2521: * The first character is the delimiter.
2522: * Succeed if the string up to its second occurrence
2523: * matches the string up to its third occurence.
2524: * Advance the cursor after the third occurrence
2525: * or lacking that, to the end of the line.
2526: */
2527: static int
2528: roff_evalstrcond(const char *v, int *pos)
2529: {
2530: const char *s1, *s2, *s3;
2531: int match;
2532:
2533: match = 0;
2534: s1 = v + *pos; /* initial delimiter */
2535: s2 = s1 + 1; /* for scanning the first string */
2536: s3 = strchr(s2, *s1); /* for scanning the second string */
2537:
2538: if (NULL == s3) /* found no middle delimiter */
2539: goto out;
2540:
2541: while ('\0' != *++s3) {
2542: if (*s2 != *s3) { /* mismatch */
2543: s3 = strchr(s3, *s1);
2544: break;
2545: }
2546: if (*s3 == *s1) { /* found the final delimiter */
2547: match = 1;
2548: break;
2549: }
2550: s2++;
2551: }
2552:
2553: out:
2554: if (NULL == s3)
2555: s3 = strchr(s2, '\0');
1.114 schwarze 2556: else if (*s3 != '\0')
1.71 schwarze 2557: s3++;
2558: *pos = s3 - v;
1.149 schwarze 2559: return match;
1.71 schwarze 2560: }
2561:
1.77 schwarze 2562: /*
2563: * Evaluate an optionally negated single character, numerical,
2564: * or string condition.
2565: */
1.71 schwarze 2566: static int
1.143 schwarze 2567: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5 schwarze 2568: {
1.208 schwarze 2569: const char *start, *end;
2570: char *cp, *name;
2571: size_t sz;
2572: int deftype, len, number, savepos, istrue, wanttrue;
1.5 schwarze 2573:
1.71 schwarze 2574: if ('!' == v[*pos]) {
2575: wanttrue = 0;
2576: (*pos)++;
2577: } else
2578: wanttrue = 1;
2579:
1.5 schwarze 2580: switch (v[*pos]) {
1.112 schwarze 2581: case '\0':
1.149 schwarze 2582: return 0;
1.80 schwarze 2583: case 'n':
2584: case 'o':
1.5 schwarze 2585: (*pos)++;
1.149 schwarze 2586: return wanttrue;
1.80 schwarze 2587: case 'e':
2588: case 't':
1.111 schwarze 2589: case 'v':
1.5 schwarze 2590: (*pos)++;
1.149 schwarze 2591: return !wanttrue;
1.208 schwarze 2592: case 'c':
2593: do {
2594: (*pos)++;
2595: } while (v[*pos] == ' ');
2596:
2597: /*
2598: * Quirk for groff compatibility:
2599: * The horizontal tab is neither available nor unavailable.
2600: */
2601:
2602: if (v[*pos] == '\t') {
2603: (*pos)++;
2604: return 0;
2605: }
2606:
2607: /* Printable ASCII characters are available. */
2608:
2609: if (v[*pos] != '\\') {
2610: (*pos)++;
2611: return wanttrue;
2612: }
2613:
2614: end = v + ++*pos;
2615: switch (mandoc_escape(&end, &start, &len)) {
2616: case ESCAPE_SPECIAL:
2617: istrue = mchars_spec2cp(start, len) != -1;
2618: break;
2619: case ESCAPE_UNICODE:
2620: istrue = 1;
2621: break;
2622: case ESCAPE_NUMBERED:
2623: istrue = mchars_num2char(start, len) != -1;
2624: break;
2625: default:
2626: istrue = !wanttrue;
2627: break;
2628: }
2629: *pos = end - v;
2630: return istrue == wanttrue;
1.182 schwarze 2631: case 'd':
1.143 schwarze 2632: case 'r':
1.182 schwarze 2633: cp = v + *pos + 1;
2634: while (*cp == ' ')
2635: cp++;
2636: name = cp;
2637: sz = roff_getname(r, &cp, ln, cp - v);
1.187 schwarze 2638: if (sz == 0)
2639: istrue = 0;
2640: else if (v[*pos] == 'r')
2641: istrue = roff_hasregn(r, name, sz);
2642: else {
2643: deftype = ROFFDEF_ANY;
2644: roff_getstrn(r, name, sz, &deftype);
2645: istrue = !!deftype;
2646: }
1.235 schwarze 2647: *pos = (name + sz) - v;
1.182 schwarze 2648: return istrue == wanttrue;
1.5 schwarze 2649: default:
2650: break;
2651: }
2652:
1.113 schwarze 2653: savepos = *pos;
1.133 schwarze 2654: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149 schwarze 2655: return (number > 0) == wanttrue;
1.113 schwarze 2656: else if (*pos == savepos)
1.149 schwarze 2657: return roff_evalstrcond(v, pos) == wanttrue;
1.77 schwarze 2658: else
1.149 schwarze 2659: return 0;
1.5 schwarze 2660: }
2661:
1.212 schwarze 2662: static int
1.21 schwarze 2663: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 2664: {
1.30 schwarze 2665:
1.149 schwarze 2666: return ROFF_IGN;
1.21 schwarze 2667: }
2668:
1.212 schwarze 2669: static int
1.123 schwarze 2670: roff_insec(ROFF_ARGS)
2671: {
2672:
1.222 schwarze 2673: mandoc_msg(MANDOCERR_REQ_INSEC, ln, ppos, "%s", roff_name[tok]);
1.149 schwarze 2674: return ROFF_IGN;
1.123 schwarze 2675: }
2676:
1.212 schwarze 2677: static int
1.123 schwarze 2678: roff_unsupp(ROFF_ARGS)
2679: {
2680:
1.222 schwarze 2681: mandoc_msg(MANDOCERR_REQ_UNSUPP, ln, ppos, "%s", roff_name[tok]);
1.149 schwarze 2682: return ROFF_IGN;
1.123 schwarze 2683: }
2684:
1.212 schwarze 2685: static int
1.2 schwarze 2686: roff_cond(ROFF_ARGS)
2687: {
1.212 schwarze 2688: int irc;
1.46 schwarze 2689:
2690: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 2691:
1.80 schwarze 2692: /*
1.35 schwarze 2693: * An `.el' has no conditional body: it will consume the value
2694: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 2695: * defaults to DENY.
1.35 schwarze 2696: *
2697: * If we're not an `el', however, then evaluate the conditional.
2698: */
1.1 schwarze 2699:
1.110 schwarze 2700: r->last->rule = tok == ROFF_el ?
1.80 schwarze 2701: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 2702: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 2703:
1.35 schwarze 2704: /*
2705: * An if-else will put the NEGATION of the current evaluated
2706: * conditional into the stack of rules.
2707: */
2708:
1.110 schwarze 2709: if (tok == ROFF_ie) {
1.96 schwarze 2710: if (r->rstackpos + 1 == r->rstacksz) {
2711: r->rstacksz += 16;
2712: r->rstack = mandoc_reallocarray(r->rstack,
2713: r->rstacksz, sizeof(int));
1.35 schwarze 2714: }
1.71 schwarze 2715: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 2716: }
1.5 schwarze 2717:
2718: /* If the parent has false as its rule, then so do we. */
2719:
1.71 schwarze 2720: if (r->last->parent && !r->last->parent->rule)
2721: r->last->rule = 0;
1.5 schwarze 2722:
2723: /*
1.46 schwarze 2724: * Determine scope.
2725: * If there is nothing on the line after the conditional,
2726: * not even whitespace, use next-line scope.
1.212 schwarze 2727: * Except that .while does not support next-line scope.
1.5 schwarze 2728: */
1.2 schwarze 2729:
1.212 schwarze 2730: if (buf->buf[pos] == '\0' && tok != ROFF_while) {
1.46 schwarze 2731: r->last->endspan = 2;
2732: goto out;
2733: }
2734:
1.110 schwarze 2735: while (buf->buf[pos] == ' ')
1.46 schwarze 2736: pos++;
2737:
2738: /* An opening brace requests multiline scope. */
1.2 schwarze 2739:
1.110 schwarze 2740: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 2741: r->last->endspan = -1;
2742: pos += 2;
1.144 schwarze 2743: while (buf->buf[pos] == ' ')
2744: pos++;
1.46 schwarze 2745: goto out;
1.80 schwarze 2746: }
1.2 schwarze 2747:
2748: /*
1.46 schwarze 2749: * Anything else following the conditional causes
2750: * single-line scope. Warn if the scope contains
2751: * nothing but trailing whitespace.
1.2 schwarze 2752: */
2753:
1.110 schwarze 2754: if (buf->buf[pos] == '\0')
1.222 schwarze 2755: mandoc_msg(MANDOCERR_COND_EMPTY,
2756: ln, ppos, "%s", roff_name[tok]);
1.2 schwarze 2757:
1.46 schwarze 2758: r->last->endspan = 1;
1.1 schwarze 2759:
1.46 schwarze 2760: out:
1.2 schwarze 2761: *offs = pos;
1.212 schwarze 2762: irc = ROFF_RERUN;
2763: if (tok == ROFF_while)
2764: irc |= ROFF_WHILE;
2765: return irc;
1.1 schwarze 2766: }
2767:
1.212 schwarze 2768: static int
1.7 schwarze 2769: roff_ds(ROFF_ARGS)
2770: {
1.85 schwarze 2771: char *string;
2772: const char *name;
2773: size_t namesz;
1.10 schwarze 2774:
1.123 schwarze 2775: /* Ignore groff compatibility mode for now. */
2776:
2777: if (tok == ROFF_ds1)
2778: tok = ROFF_ds;
2779: else if (tok == ROFF_as1)
2780: tok = ROFF_as;
2781:
1.10 schwarze 2782: /*
1.85 schwarze 2783: * The first word is the name of the string.
2784: * If it is empty or terminated by an escape sequence,
2785: * abort the `ds' request without defining anything.
1.10 schwarze 2786: */
1.7 schwarze 2787:
1.110 schwarze 2788: name = string = buf->buf + pos;
2789: if (*name == '\0')
1.149 schwarze 2790: return ROFF_IGN;
1.7 schwarze 2791:
1.85 schwarze 2792: namesz = roff_getname(r, &string, ln, pos);
1.235 schwarze 2793: switch (name[namesz]) {
2794: case '\\':
1.149 schwarze 2795: return ROFF_IGN;
1.235 schwarze 2796: case '\t':
2797: string = buf->buf + pos + namesz;
2798: break;
2799: default:
2800: break;
2801: }
1.85 schwarze 2802:
2803: /* Read past the initial double-quote, if any. */
1.110 schwarze 2804: if (*string == '"')
1.7 schwarze 2805: string++;
2806:
1.10 schwarze 2807: /* The rest is the value. */
1.85 schwarze 2808: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2809: ROFF_as == tok);
1.183 schwarze 2810: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149 schwarze 2811: return ROFF_IGN;
1.7 schwarze 2812: }
2813:
1.77 schwarze 2814: /*
2815: * Parse a single operator, one or two characters long.
2816: * If the operator is recognized, return success and advance the
2817: * parse point, else return failure and let the parse point unchanged.
2818: */
2819: static int
2820: roff_getop(const char *v, int *pos, char *res)
2821: {
2822:
2823: *res = v[*pos];
2824:
2825: switch (*res) {
1.80 schwarze 2826: case '+':
2827: case '-':
2828: case '*':
2829: case '/':
2830: case '%':
2831: case '&':
2832: case ':':
1.77 schwarze 2833: break;
2834: case '<':
2835: switch (v[*pos + 1]) {
1.80 schwarze 2836: case '=':
1.77 schwarze 2837: *res = 'l';
2838: (*pos)++;
2839: break;
1.80 schwarze 2840: case '>':
1.77 schwarze 2841: *res = '!';
2842: (*pos)++;
2843: break;
1.80 schwarze 2844: case '?':
1.77 schwarze 2845: *res = 'i';
2846: (*pos)++;
2847: break;
2848: default:
2849: break;
2850: }
2851: break;
2852: case '>':
2853: switch (v[*pos + 1]) {
1.80 schwarze 2854: case '=':
1.77 schwarze 2855: *res = 'g';
2856: (*pos)++;
2857: break;
1.80 schwarze 2858: case '?':
1.77 schwarze 2859: *res = 'a';
2860: (*pos)++;
2861: break;
2862: default:
2863: break;
2864: }
2865: break;
2866: case '=':
2867: if ('=' == v[*pos + 1])
2868: (*pos)++;
2869: break;
2870: default:
1.149 schwarze 2871: return 0;
1.77 schwarze 2872: }
2873: (*pos)++;
2874:
1.149 schwarze 2875: return *res;
1.77 schwarze 2876: }
2877:
2878: /*
2879: * Evaluate either a parenthesized numeric expression
2880: * or a single signed integer number.
2881: */
2882: static int
1.106 schwarze 2883: roff_evalpar(struct roff *r, int ln,
1.133 schwarze 2884: const char *v, int *pos, int *res, int flags)
1.77 schwarze 2885: {
2886:
2887: if ('(' != v[*pos])
1.149 schwarze 2888: return roff_getnum(v, pos, res, flags);
1.77 schwarze 2889:
2890: (*pos)++;
1.133 schwarze 2891: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149 schwarze 2892: return 0;
1.77 schwarze 2893:
1.79 schwarze 2894: /*
2895: * Omission of the closing parenthesis
2896: * is an error in validation mode,
2897: * but ignored in evaluation mode.
2898: */
2899:
1.77 schwarze 2900: if (')' == v[*pos])
2901: (*pos)++;
1.79 schwarze 2902: else if (NULL == res)
1.149 schwarze 2903: return 0;
1.77 schwarze 2904:
1.149 schwarze 2905: return 1;
1.77 schwarze 2906: }
2907:
2908: /*
2909: * Evaluate a complete numeric expression.
2910: * Proceed left to right, there is no concept of precedence.
2911: */
2912: static int
1.106 schwarze 2913: roff_evalnum(struct roff *r, int ln, const char *v,
1.133 schwarze 2914: int *pos, int *res, int flags)
1.77 schwarze 2915: {
2916: int mypos, operand2;
2917: char operator;
2918:
2919: if (NULL == pos) {
2920: mypos = 0;
2921: pos = &mypos;
2922: }
2923:
1.133 schwarze 2924: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2925: while (isspace((unsigned char)v[*pos]))
2926: (*pos)++;
2927:
1.133 schwarze 2928: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149 schwarze 2929: return 0;
1.77 schwarze 2930:
2931: while (1) {
1.133 schwarze 2932: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2933: while (isspace((unsigned char)v[*pos]))
2934: (*pos)++;
2935:
2936: if ( ! roff_getop(v, pos, &operator))
2937: break;
2938:
1.133 schwarze 2939: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2940: while (isspace((unsigned char)v[*pos]))
2941: (*pos)++;
2942:
1.133 schwarze 2943: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149 schwarze 2944: return 0;
1.77 schwarze 2945:
1.133 schwarze 2946: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2947: while (isspace((unsigned char)v[*pos]))
2948: (*pos)++;
1.79 schwarze 2949:
2950: if (NULL == res)
2951: continue;
1.77 schwarze 2952:
2953: switch (operator) {
1.80 schwarze 2954: case '+':
1.77 schwarze 2955: *res += operand2;
2956: break;
1.80 schwarze 2957: case '-':
1.77 schwarze 2958: *res -= operand2;
2959: break;
1.80 schwarze 2960: case '*':
1.77 schwarze 2961: *res *= operand2;
2962: break;
1.80 schwarze 2963: case '/':
1.116 schwarze 2964: if (operand2 == 0) {
1.106 schwarze 2965: mandoc_msg(MANDOCERR_DIVZERO,
1.222 schwarze 2966: ln, *pos, "%s", v);
1.106 schwarze 2967: *res = 0;
2968: break;
2969: }
1.77 schwarze 2970: *res /= operand2;
2971: break;
1.80 schwarze 2972: case '%':
1.116 schwarze 2973: if (operand2 == 0) {
2974: mandoc_msg(MANDOCERR_DIVZERO,
1.222 schwarze 2975: ln, *pos, "%s", v);
1.116 schwarze 2976: *res = 0;
2977: break;
2978: }
1.77 schwarze 2979: *res %= operand2;
2980: break;
1.80 schwarze 2981: case '<':
1.77 schwarze 2982: *res = *res < operand2;
2983: break;
1.80 schwarze 2984: case '>':
1.77 schwarze 2985: *res = *res > operand2;
2986: break;
1.80 schwarze 2987: case 'l':
1.77 schwarze 2988: *res = *res <= operand2;
2989: break;
1.80 schwarze 2990: case 'g':
1.77 schwarze 2991: *res = *res >= operand2;
2992: break;
1.80 schwarze 2993: case '=':
1.77 schwarze 2994: *res = *res == operand2;
2995: break;
1.80 schwarze 2996: case '!':
1.77 schwarze 2997: *res = *res != operand2;
2998: break;
1.80 schwarze 2999: case '&':
1.77 schwarze 3000: *res = *res && operand2;
3001: break;
1.80 schwarze 3002: case ':':
1.77 schwarze 3003: *res = *res || operand2;
3004: break;
1.80 schwarze 3005: case 'i':
1.77 schwarze 3006: if (operand2 < *res)
3007: *res = operand2;
3008: break;
1.80 schwarze 3009: case 'a':
1.77 schwarze 3010: if (operand2 > *res)
3011: *res = operand2;
3012: break;
3013: default:
3014: abort();
3015: }
3016: }
1.149 schwarze 3017: return 1;
1.77 schwarze 3018: }
3019:
1.138 schwarze 3020: /* --- register management ------------------------------------------------ */
3021:
1.52 schwarze 3022: void
1.60 schwarze 3023: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 3024: {
1.199 schwarze 3025: roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.198 schwarze 3026: }
3027:
3028: static void
3029: roff_setregn(struct roff *r, const char *name, size_t len,
1.199 schwarze 3030: int val, char sign, int step)
1.198 schwarze 3031: {
1.52 schwarze 3032: struct roffreg *reg;
3033:
3034: /* Search for an existing register with the same name. */
3035: reg = r->regtab;
3036:
1.198 schwarze 3037: while (reg != NULL && (reg->key.sz != len ||
3038: strncmp(reg->key.p, name, len) != 0))
1.52 schwarze 3039: reg = reg->next;
1.41 schwarze 3040:
1.52 schwarze 3041: if (NULL == reg) {
3042: /* Create a new register. */
3043: reg = mandoc_malloc(sizeof(struct roffreg));
1.198 schwarze 3044: reg->key.p = mandoc_strndup(name, len);
3045: reg->key.sz = len;
1.60 schwarze 3046: reg->val = 0;
1.199 schwarze 3047: reg->step = 0;
1.52 schwarze 3048: reg->next = r->regtab;
3049: r->regtab = reg;
3050: }
3051:
1.60 schwarze 3052: if ('+' == sign)
3053: reg->val += val;
3054: else if ('-' == sign)
3055: reg->val -= val;
3056: else
3057: reg->val = val;
1.199 schwarze 3058: if (step != INT_MIN)
3059: reg->step = step;
1.41 schwarze 3060: }
3061:
1.65 schwarze 3062: /*
3063: * Handle some predefined read-only number registers.
3064: * For now, return -1 if the requested register is not predefined;
3065: * in case a predefined read-only register having the value -1
3066: * were to turn up, another special value would have to be chosen.
3067: */
3068: static int
1.145 schwarze 3069: roff_getregro(const struct roff *r, const char *name)
1.65 schwarze 3070: {
3071:
3072: switch (*name) {
1.145 schwarze 3073: case '$': /* Number of arguments of the last macro evaluated. */
1.211 schwarze 3074: return r->mstackpos < 0 ? 0 : r->mstack[r->mstackpos].argc;
1.80 schwarze 3075: case 'A': /* ASCII approximation mode is always off. */
1.149 schwarze 3076: return 0;
1.80 schwarze 3077: case 'g': /* Groff compatibility mode is always on. */
1.149 schwarze 3078: return 1;
1.80 schwarze 3079: case 'H': /* Fixed horizontal resolution. */
1.149 schwarze 3080: return 24;
1.80 schwarze 3081: case 'j': /* Always adjust left margin only. */
1.149 schwarze 3082: return 0;
1.80 schwarze 3083: case 'T': /* Some output device is always defined. */
1.149 schwarze 3084: return 1;
1.80 schwarze 3085: case 'V': /* Fixed vertical resolution. */
1.149 schwarze 3086: return 40;
1.65 schwarze 3087: default:
1.149 schwarze 3088: return -1;
1.65 schwarze 3089: }
3090: }
3091:
1.53 schwarze 3092: int
1.198 schwarze 3093: roff_getreg(struct roff *r, const char *name)
1.41 schwarze 3094: {
1.199 schwarze 3095: return roff_getregn(r, name, strlen(name), '\0');
1.53 schwarze 3096: }
3097:
3098: static int
1.199 schwarze 3099: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.53 schwarze 3100: {
3101: struct roffreg *reg;
1.65 schwarze 3102: int val;
3103:
3104: if ('.' == name[0] && 2 == len) {
1.145 schwarze 3105: val = roff_getregro(r, name + 1);
1.65 schwarze 3106: if (-1 != val)
1.149 schwarze 3107: return val;
1.65 schwarze 3108: }
1.53 schwarze 3109:
1.199 schwarze 3110: for (reg = r->regtab; reg; reg = reg->next) {
1.53 schwarze 3111: if (len == reg->key.sz &&
1.199 schwarze 3112: 0 == strncmp(name, reg->key.p, len)) {
3113: switch (sign) {
3114: case '+':
3115: reg->val += reg->step;
3116: break;
3117: case '-':
3118: reg->val -= reg->step;
3119: break;
3120: default:
3121: break;
3122: }
1.149 schwarze 3123: return reg->val;
1.199 schwarze 3124: }
3125: }
1.143 schwarze 3126:
1.199 schwarze 3127: roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.149 schwarze 3128: return 0;
1.143 schwarze 3129: }
3130:
3131: static int
3132: roff_hasregn(const struct roff *r, const char *name, size_t len)
3133: {
3134: struct roffreg *reg;
3135: int val;
3136:
3137: if ('.' == name[0] && 2 == len) {
1.145 schwarze 3138: val = roff_getregro(r, name + 1);
1.143 schwarze 3139: if (-1 != val)
1.149 schwarze 3140: return 1;
1.143 schwarze 3141: }
3142:
3143: for (reg = r->regtab; reg; reg = reg->next)
3144: if (len == reg->key.sz &&
3145: 0 == strncmp(name, reg->key.p, len))
1.149 schwarze 3146: return 1;
1.41 schwarze 3147:
1.149 schwarze 3148: return 0;
1.41 schwarze 3149: }
3150:
1.52 schwarze 3151: static void
3152: roff_freereg(struct roffreg *reg)
1.41 schwarze 3153: {
1.52 schwarze 3154: struct roffreg *old_reg;
1.41 schwarze 3155:
1.52 schwarze 3156: while (NULL != reg) {
3157: free(reg->key.p);
3158: old_reg = reg;
3159: reg = reg->next;
3160: free(old_reg);
3161: }
1.41 schwarze 3162: }
1.7 schwarze 3163:
1.212 schwarze 3164: static int
1.6 schwarze 3165: roff_nr(ROFF_ARGS)
1.1 schwarze 3166: {
1.199 schwarze 3167: char *key, *val, *step;
1.85 schwarze 3168: size_t keysz;
1.199 schwarze 3169: int iv, is, len;
1.60 schwarze 3170: char sign;
1.6 schwarze 3171:
1.110 schwarze 3172: key = val = buf->buf + pos;
3173: if (*key == '\0')
1.149 schwarze 3174: return ROFF_IGN;
1.85 schwarze 3175:
3176: keysz = roff_getname(r, &val, ln, pos);
1.235 schwarze 3177: if (key[keysz] == '\\' || key[keysz] == '\t')
1.149 schwarze 3178: return ROFF_IGN;
1.6 schwarze 3179:
1.60 schwarze 3180: sign = *val;
1.110 schwarze 3181: if (sign == '+' || sign == '-')
1.60 schwarze 3182: val++;
3183:
1.199 schwarze 3184: len = 0;
3185: if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
3186: return ROFF_IGN;
3187:
3188: step = val + len;
3189: while (isspace((unsigned char)*step))
3190: step++;
3191: if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
3192: is = INT_MIN;
1.1 schwarze 3193:
1.199 schwarze 3194: roff_setregn(r, key, keysz, iv, sign, is);
1.149 schwarze 3195: return ROFF_IGN;
1.76 schwarze 3196: }
3197:
1.212 schwarze 3198: static int
1.76 schwarze 3199: roff_rr(ROFF_ARGS)
3200: {
3201: struct roffreg *reg, **prev;
1.85 schwarze 3202: char *name, *cp;
3203: size_t namesz;
1.76 schwarze 3204:
1.110 schwarze 3205: name = cp = buf->buf + pos;
3206: if (*name == '\0')
1.149 schwarze 3207: return ROFF_IGN;
1.85 schwarze 3208: namesz = roff_getname(r, &cp, ln, pos);
3209: name[namesz] = '\0';
1.76 schwarze 3210:
3211: prev = &r->regtab;
3212: while (1) {
3213: reg = *prev;
1.110 schwarze 3214: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 3215: break;
3216: prev = ®->next;
3217: }
1.110 schwarze 3218: if (reg != NULL) {
1.76 schwarze 3219: *prev = reg->next;
3220: free(reg->key.p);
3221: free(reg);
3222: }
1.149 schwarze 3223: return ROFF_IGN;
1.29 schwarze 3224: }
3225:
1.138 schwarze 3226: /* --- handler functions for roff requests -------------------------------- */
3227:
1.212 schwarze 3228: static int
1.29 schwarze 3229: roff_rm(ROFF_ARGS)
3230: {
3231: const char *name;
3232: char *cp;
1.85 schwarze 3233: size_t namesz;
1.29 schwarze 3234:
1.110 schwarze 3235: cp = buf->buf + pos;
3236: while (*cp != '\0') {
1.85 schwarze 3237: name = cp;
1.110 schwarze 3238: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 3239: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183 schwarze 3240: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.235 schwarze 3241: if (name[namesz] == '\\' || name[namesz] == '\t')
1.85 schwarze 3242: break;
1.29 schwarze 3243: }
1.149 schwarze 3244: return ROFF_IGN;
1.51 schwarze 3245: }
3246:
1.212 schwarze 3247: static int
1.51 schwarze 3248: roff_it(ROFF_ARGS)
3249: {
3250: int iv;
3251:
3252: /* Parse the number of lines. */
1.133 schwarze 3253:
3254: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 3255: mandoc_msg(MANDOCERR_IT_NONUM,
3256: ln, ppos, "%s", buf->buf + 1);
1.149 schwarze 3257: return ROFF_IGN;
1.51 schwarze 3258: }
3259:
1.134 schwarze 3260: while (isspace((unsigned char)buf->buf[pos]))
3261: pos++;
3262:
3263: /*
3264: * Arm the input line trap.
3265: * Special-casing "an-trap" is an ugly workaround to cope
3266: * with DocBook stupidly fiddling with man(7) internals.
3267: */
1.133 schwarze 3268:
1.51 schwarze 3269: roffit_lines = iv;
1.134 schwarze 3270: roffit_macro = mandoc_strdup(iv != 1 ||
3271: strcmp(buf->buf + pos, "an-trap") ?
3272: buf->buf + pos : "br");
1.149 schwarze 3273: return ROFF_IGN;
1.47 schwarze 3274: }
3275:
1.212 schwarze 3276: static int
1.47 schwarze 3277: roff_Dd(ROFF_ARGS)
3278: {
1.187 schwarze 3279: int mask;
3280: enum roff_tok t, te;
1.99 schwarze 3281:
1.187 schwarze 3282: switch (tok) {
3283: case ROFF_Dd:
3284: tok = MDOC_Dd;
3285: te = MDOC_MAX;
3286: if (r->format == 0)
3287: r->format = MPARSE_MDOC;
3288: mask = MPARSE_MDOC | MPARSE_QUICK;
3289: break;
3290: case ROFF_TH:
3291: tok = MAN_TH;
3292: te = MAN_MAX;
3293: if (r->format == 0)
3294: r->format = MPARSE_MAN;
3295: mask = MPARSE_QUICK;
3296: break;
3297: default:
3298: abort();
3299: }
3300: if ((r->options & mask) == 0)
3301: for (t = tok; t < te; t++)
3302: roff_setstr(r, roff_name[t], NULL, 0);
1.149 schwarze 3303: return ROFF_CONT;
1.14 schwarze 3304: }
3305:
1.212 schwarze 3306: static int
1.27 schwarze 3307: roff_TE(ROFF_ARGS)
3308: {
1.233 schwarze 3309: r->man->flags &= ~ROFF_NONOFILL;
1.193 schwarze 3310: if (r->tbl == NULL) {
1.222 schwarze 3311: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "TE");
1.193 schwarze 3312: return ROFF_IGN;
3313: }
1.218 schwarze 3314: if (tbl_end(r->tbl, 0) == 0) {
1.193 schwarze 3315: r->tbl = NULL;
1.130 schwarze 3316: free(buf->buf);
3317: buf->buf = mandoc_strdup(".sp");
3318: buf->sz = 4;
1.201 schwarze 3319: *offs = 0;
1.149 schwarze 3320: return ROFF_REPARSE;
1.130 schwarze 3321: }
1.193 schwarze 3322: r->tbl = NULL;
1.149 schwarze 3323: return ROFF_IGN;
1.27 schwarze 3324: }
3325:
1.212 schwarze 3326: static int
1.27 schwarze 3327: roff_T_(ROFF_ARGS)
3328: {
3329:
3330: if (NULL == r->tbl)
1.222 schwarze 3331: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "T&");
1.27 schwarze 3332: else
1.168 schwarze 3333: tbl_restart(ln, ppos, r->tbl);
1.27 schwarze 3334:
1.149 schwarze 3335: return ROFF_IGN;
1.27 schwarze 3336: }
3337:
1.102 schwarze 3338: /*
3339: * Handle in-line equation delimiters.
3340: */
1.212 schwarze 3341: static int
1.110 schwarze 3342: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 3343: {
1.105 schwarze 3344: char *cp1, *cp2;
3345: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 3346:
1.102 schwarze 3347: /*
3348: * Outside equations, look for an opening delimiter.
3349: * If we are inside an equation, we already know it is
3350: * in-line, or this function wouldn't have been called;
3351: * so look for a closing delimiter.
3352: */
3353:
1.110 schwarze 3354: cp1 = buf->buf + pos;
1.102 schwarze 3355: cp2 = strchr(cp1, r->eqn == NULL ?
3356: r->last_eqn->odelim : r->last_eqn->cdelim);
3357: if (cp2 == NULL)
1.149 schwarze 3358: return ROFF_CONT;
1.102 schwarze 3359:
1.105 schwarze 3360: *cp2++ = '\0';
3361: bef_pr = bef_nl = aft_nl = aft_pr = "";
3362:
3363: /* Handle preceding text, protecting whitespace. */
3364:
1.110 schwarze 3365: if (*buf->buf != '\0') {
1.105 schwarze 3366: if (r->eqn == NULL)
3367: bef_pr = "\\&";
3368: bef_nl = "\n";
3369: }
3370:
3371: /*
3372: * Prepare replacing the delimiter with an equation macro
3373: * and drop leading white space from the equation.
3374: */
1.102 schwarze 3375:
1.105 schwarze 3376: if (r->eqn == NULL) {
3377: while (*cp2 == ' ')
3378: cp2++;
3379: mac = ".EQ";
3380: } else
3381: mac = ".EN";
3382:
3383: /* Handle following text, protecting whitespace. */
3384:
3385: if (*cp2 != '\0') {
3386: aft_nl = "\n";
3387: if (r->eqn != NULL)
3388: aft_pr = "\\&";
3389: }
3390:
3391: /* Do the actual replacement. */
3392:
1.110 schwarze 3393: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 3394: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 3395: free(buf->buf);
3396: buf->buf = cp1;
1.102 schwarze 3397:
3398: /* Toggle the in-line state of the eqn subsystem. */
3399:
3400: r->eqn_inline = r->eqn == NULL;
1.149 schwarze 3401: return ROFF_REPARSE;
1.41 schwarze 3402: }
3403:
1.212 schwarze 3404: static int
1.107 schwarze 3405: roff_EQ(ROFF_ARGS)
1.32 schwarze 3406: {
1.191 schwarze 3407: struct roff_node *n;
3408:
1.228 schwarze 3409: if (r->man->meta.macroset == MACROSET_MAN)
1.194 schwarze 3410: man_breakscope(r->man, ROFF_EQ);
1.191 schwarze 3411: n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
3412: if (ln > r->man->last->line)
3413: n->flags |= NODE_LINE;
1.220 schwarze 3414: n->eqn = eqn_box_new();
1.191 schwarze 3415: roff_node_append(r->man, n);
3416: r->man->next = ROFF_NEXT_SIBLING;
1.32 schwarze 3417:
1.110 schwarze 3418: assert(r->eqn == NULL);
1.191 schwarze 3419: if (r->last_eqn == NULL)
1.223 schwarze 3420: r->last_eqn = eqn_alloc();
1.191 schwarze 3421: else
3422: eqn_reset(r->last_eqn);
3423: r->eqn = r->last_eqn;
3424: r->eqn->node = n;
1.41 schwarze 3425:
1.110 schwarze 3426: if (buf->buf[pos] != '\0')
1.222 schwarze 3427: mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.110 schwarze 3428: ".EQ %s", buf->buf + pos);
1.41 schwarze 3429:
1.149 schwarze 3430: return ROFF_IGN;
1.32 schwarze 3431: }
3432:
1.212 schwarze 3433: static int
1.32 schwarze 3434: roff_EN(ROFF_ARGS)
3435: {
1.191 schwarze 3436: if (r->eqn != NULL) {
3437: eqn_parse(r->eqn);
3438: r->eqn = NULL;
3439: } else
1.222 schwarze 3440: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, ppos, "EN");
1.191 schwarze 3441: if (buf->buf[pos] != '\0')
1.222 schwarze 3442: mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
1.191 schwarze 3443: "EN %s", buf->buf + pos);
1.149 schwarze 3444: return ROFF_IGN;
1.32 schwarze 3445: }
3446:
1.212 schwarze 3447: static int
1.27 schwarze 3448: roff_TS(ROFF_ARGS)
3449: {
1.193 schwarze 3450: if (r->tbl != NULL) {
1.222 schwarze 3451: mandoc_msg(MANDOCERR_BLK_BROKEN, ln, ppos, "TS breaks TS");
1.218 schwarze 3452: tbl_end(r->tbl, 0);
1.27 schwarze 3453: }
1.233 schwarze 3454: r->man->flags |= ROFF_NONOFILL;
1.223 schwarze 3455: r->tbl = tbl_alloc(ppos, ln, r->last_tbl);
1.218 schwarze 3456: if (r->last_tbl == NULL)
1.193 schwarze 3457: r->first_tbl = r->tbl;
3458: r->last_tbl = r->tbl;
1.169 schwarze 3459: return ROFF_IGN;
3460: }
3461:
1.212 schwarze 3462: static int
1.230 schwarze 3463: roff_noarg(ROFF_ARGS)
3464: {
3465: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
3466: man_breakscope(r->man, tok);
3467: if (tok == ROFF_brp)
3468: tok = ROFF_br;
3469: roff_elem_alloc(r->man, ln, ppos, tok);
3470: if (buf->buf[pos] != '\0')
3471: mandoc_msg(MANDOCERR_ARG_SKIP, ln, pos,
3472: "%s %s", roff_name[tok], buf->buf + pos);
3473: if (tok == ROFF_nf)
3474: r->man->flags |= ROFF_NOFILL;
3475: else if (tok == ROFF_fi)
3476: r->man->flags &= ~ROFF_NOFILL;
3477: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3478: r->man->next = ROFF_NEXT_SIBLING;
3479: return ROFF_IGN;
3480: }
3481:
3482: static int
1.169 schwarze 3483: roff_onearg(ROFF_ARGS)
3484: {
3485: struct roff_node *n;
3486: char *cp;
1.177 schwarze 3487: int npos;
1.169 schwarze 3488:
1.174 schwarze 3489: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.192 schwarze 3490: (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
3491: tok == ROFF_ti))
1.174 schwarze 3492: man_breakscope(r->man, tok);
3493:
1.181 schwarze 3494: if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177 schwarze 3495: r->man->last = roffce_node;
3496: r->man->next = ROFF_NEXT_SIBLING;
3497: }
3498:
1.169 schwarze 3499: roff_elem_alloc(r->man, ln, ppos, tok);
3500: n = r->man->last;
3501:
3502: cp = buf->buf + pos;
3503: if (*cp != '\0') {
3504: while (*cp != '\0' && *cp != ' ')
3505: cp++;
3506: while (*cp == ' ')
3507: *cp++ = '\0';
3508: if (*cp != '\0')
1.222 schwarze 3509: mandoc_msg(MANDOCERR_ARG_EXCESS,
3510: ln, (int)(cp - buf->buf),
1.169 schwarze 3511: "%s ... %s", roff_name[tok], cp);
3512: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172 schwarze 3513: }
3514:
1.181 schwarze 3515: if (tok == ROFF_ce || tok == ROFF_rj) {
3516: if (r->man->last->type == ROFFT_ELEM) {
1.177 schwarze 3517: roff_word_alloc(r->man, ln, pos, "1");
3518: r->man->last->flags |= NODE_NOSRC;
3519: }
3520: npos = 0;
3521: if (roff_evalnum(r, ln, r->man->last->string, &npos,
3522: &roffce_lines, 0) == 0) {
1.222 schwarze 3523: mandoc_msg(MANDOCERR_CE_NONUM,
3524: ln, pos, "ce %s", buf->buf + pos);
1.177 schwarze 3525: roffce_lines = 1;
3526: }
3527: if (roffce_lines < 1) {
3528: r->man->last = r->man->last->parent;
3529: roffce_node = NULL;
3530: roffce_lines = 0;
3531: } else
3532: roffce_node = r->man->last->parent;
3533: } else {
3534: n->flags |= NODE_VALID | NODE_ENDED;
3535: r->man->last = n;
3536: }
3537: n->flags |= NODE_LINE;
1.172 schwarze 3538: r->man->next = ROFF_NEXT_SIBLING;
3539: return ROFF_IGN;
3540: }
3541:
1.212 schwarze 3542: static int
1.172 schwarze 3543: roff_manyarg(ROFF_ARGS)
3544: {
3545: struct roff_node *n;
3546: char *sp, *ep;
3547:
3548: roff_elem_alloc(r->man, ln, ppos, tok);
3549: n = r->man->last;
3550:
3551: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
3552: while (*ep != '\0' && *ep != ' ')
3553: ep++;
3554: while (*ep == ' ')
3555: *ep++ = '\0';
3556: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169 schwarze 3557: }
3558:
3559: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3560: r->man->last = n;
3561: r->man->next = ROFF_NEXT_SIBLING;
1.149 schwarze 3562: return ROFF_IGN;
1.123 schwarze 3563: }
3564:
1.212 schwarze 3565: static int
1.183 schwarze 3566: roff_als(ROFF_ARGS)
3567: {
3568: char *oldn, *newn, *end, *value;
3569: size_t oldsz, newsz, valsz;
3570:
3571: newn = oldn = buf->buf + pos;
3572: if (*newn == '\0')
3573: return ROFF_IGN;
3574:
3575: newsz = roff_getname(r, &oldn, ln, pos);
1.235 schwarze 3576: if (newn[newsz] == '\\' || newn[newsz] == '\t' || *oldn == '\0')
1.183 schwarze 3577: return ROFF_IGN;
3578:
3579: end = oldn;
3580: oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
3581: if (oldsz == 0)
3582: return ROFF_IGN;
3583:
1.210 schwarze 3584: valsz = mandoc_asprintf(&value, ".%.*s \\$@\\\"\n",
1.187 schwarze 3585: (int)oldsz, oldn);
1.183 schwarze 3586: roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
3587: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3588: free(value);
1.236 schwarze 3589: return ROFF_IGN;
3590: }
3591:
3592: /*
3593: * The .break request only makes sense inside conditionals,
3594: * and that case is already handled in roff_cond_sub().
3595: */
3596: static int
3597: roff_break(ROFF_ARGS)
3598: {
3599: mandoc_msg(MANDOCERR_BLK_NOTOPEN, ln, pos, "break");
1.168 schwarze 3600: return ROFF_IGN;
1.27 schwarze 3601: }
3602:
1.212 schwarze 3603: static int
1.48 schwarze 3604: roff_cc(ROFF_ARGS)
3605: {
3606: const char *p;
3607:
1.110 schwarze 3608: p = buf->buf + pos;
1.48 schwarze 3609:
1.110 schwarze 3610: if (*p == '\0' || (r->control = *p++) == '.')
1.175 schwarze 3611: r->control = '\0';
1.48 schwarze 3612:
1.110 schwarze 3613: if (*p != '\0')
1.222 schwarze 3614: mandoc_msg(MANDOCERR_ARG_EXCESS,
1.132 schwarze 3615: ln, p - buf->buf, "cc ... %s", p);
1.48 schwarze 3616:
1.213 schwarze 3617: return ROFF_IGN;
3618: }
3619:
3620: static int
3621: roff_char(ROFF_ARGS)
3622: {
3623: const char *p, *kp, *vp;
3624: size_t ksz, vsz;
3625: int font;
3626:
3627: /* Parse the character to be replaced. */
3628:
3629: kp = buf->buf + pos;
3630: p = kp + 1;
3631: if (*kp == '\0' || (*kp == '\\' &&
3632: mandoc_escape(&p, NULL, NULL) != ESCAPE_SPECIAL) ||
3633: (*p != ' ' && *p != '\0')) {
1.222 schwarze 3634: mandoc_msg(MANDOCERR_CHAR_ARG, ln, pos, "char %s", kp);
1.213 schwarze 3635: return ROFF_IGN;
3636: }
3637: ksz = p - kp;
3638: while (*p == ' ')
3639: p++;
3640:
3641: /*
3642: * If the replacement string contains a font escape sequence,
3643: * we have to restore the font at the end.
3644: */
3645:
3646: vp = p;
3647: vsz = strlen(p);
3648: font = 0;
3649: while (*p != '\0') {
3650: if (*p++ != '\\')
3651: continue;
3652: switch (mandoc_escape(&p, NULL, NULL)) {
3653: case ESCAPE_FONT:
3654: case ESCAPE_FONTROMAN:
3655: case ESCAPE_FONTITALIC:
3656: case ESCAPE_FONTBOLD:
3657: case ESCAPE_FONTBI:
1.214 schwarze 3658: case ESCAPE_FONTCW:
1.213 schwarze 3659: case ESCAPE_FONTPREV:
3660: font++;
3661: break;
3662: default:
3663: break;
3664: }
3665: }
3666: if (font > 1)
1.222 schwarze 3667: mandoc_msg(MANDOCERR_CHAR_FONT,
3668: ln, (int)(vp - buf->buf), "%s", vp);
1.213 schwarze 3669:
3670: /*
3671: * Approximate the effect of .char using the .tr tables.
3672: * XXX In groff, .char and .tr interact differently.
3673: */
3674:
3675: if (ksz == 1) {
3676: if (r->xtab == NULL)
3677: r->xtab = mandoc_calloc(128, sizeof(*r->xtab));
3678: assert((unsigned int)*kp < 128);
3679: free(r->xtab[(int)*kp].p);
3680: r->xtab[(int)*kp].sz = mandoc_asprintf(&r->xtab[(int)*kp].p,
3681: "%s%s", vp, font ? "\fP" : "");
3682: } else {
3683: roff_setstrn(&r->xmbtab, kp, ksz, vp, vsz, 0);
3684: if (font)
3685: roff_setstrn(&r->xmbtab, kp, ksz, "\\fP", 3, 1);
3686: }
1.149 schwarze 3687: return ROFF_IGN;
1.48 schwarze 3688: }
3689:
1.212 schwarze 3690: static int
1.175 schwarze 3691: roff_ec(ROFF_ARGS)
3692: {
3693: const char *p;
3694:
3695: p = buf->buf + pos;
3696: if (*p == '\0')
3697: r->escape = '\\';
3698: else {
3699: r->escape = *p;
3700: if (*++p != '\0')
1.222 schwarze 3701: mandoc_msg(MANDOCERR_ARG_EXCESS, ln,
3702: (int)(p - buf->buf), "ec ... %s", p);
1.175 schwarze 3703: }
3704: return ROFF_IGN;
3705: }
3706:
1.212 schwarze 3707: static int
1.175 schwarze 3708: roff_eo(ROFF_ARGS)
3709: {
3710: r->escape = '\0';
3711: if (buf->buf[pos] != '\0')
1.222 schwarze 3712: mandoc_msg(MANDOCERR_ARG_SKIP,
1.175 schwarze 3713: ln, pos, "eo %s", buf->buf + pos);
3714: return ROFF_IGN;
1.202 schwarze 3715: }
3716:
1.212 schwarze 3717: static int
1.202 schwarze 3718: roff_nop(ROFF_ARGS)
3719: {
3720: while (buf->buf[pos] == ' ')
3721: pos++;
3722: *offs = pos;
3723: return ROFF_RERUN;
1.175 schwarze 3724: }
3725:
1.212 schwarze 3726: static int
1.42 schwarze 3727: roff_tr(ROFF_ARGS)
3728: {
3729: const char *p, *first, *second;
3730: size_t fsz, ssz;
3731: enum mandoc_esc esc;
3732:
1.110 schwarze 3733: p = buf->buf + pos;
1.42 schwarze 3734:
1.110 schwarze 3735: if (*p == '\0') {
1.222 schwarze 3736: mandoc_msg(MANDOCERR_REQ_EMPTY, ln, ppos, "tr");
1.149 schwarze 3737: return ROFF_IGN;
1.42 schwarze 3738: }
3739:
1.110 schwarze 3740: while (*p != '\0') {
1.42 schwarze 3741: fsz = ssz = 1;
3742:
3743: first = p++;
1.110 schwarze 3744: if (*first == '\\') {
1.42 schwarze 3745: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 3746: if (esc == ESCAPE_ERROR) {
1.222 schwarze 3747: mandoc_msg(MANDOCERR_ESC_BAD, ln,
3748: (int)(p - buf->buf), "%s", first);
1.149 schwarze 3749: return ROFF_IGN;
1.42 schwarze 3750: }
3751: fsz = (size_t)(p - first);
3752: }
3753:
3754: second = p++;
1.110 schwarze 3755: if (*second == '\\') {
1.42 schwarze 3756: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 3757: if (esc == ESCAPE_ERROR) {
1.222 schwarze 3758: mandoc_msg(MANDOCERR_ESC_BAD, ln,
3759: (int)(p - buf->buf), "%s", second);
1.149 schwarze 3760: return ROFF_IGN;
1.42 schwarze 3761: }
3762: ssz = (size_t)(p - second);
1.110 schwarze 3763: } else if (*second == '\0') {
1.222 schwarze 3764: mandoc_msg(MANDOCERR_TR_ODD, ln,
3765: (int)(first - buf->buf), "tr %s", first);
1.42 schwarze 3766: second = " ";
3767: p--;
3768: }
3769:
3770: if (fsz > 1) {
1.80 schwarze 3771: roff_setstrn(&r->xmbtab, first, fsz,
3772: second, ssz, 0);
1.42 schwarze 3773: continue;
3774: }
3775:
1.110 schwarze 3776: if (r->xtab == NULL)
1.80 schwarze 3777: r->xtab = mandoc_calloc(128,
3778: sizeof(struct roffstr));
1.42 schwarze 3779:
3780: free(r->xtab[(int)*first].p);
3781: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3782: r->xtab[(int)*first].sz = ssz;
3783: }
3784:
1.149 schwarze 3785: return ROFF_IGN;
1.42 schwarze 3786: }
3787:
1.211 schwarze 3788: /*
3789: * Implementation of the .return request.
3790: * There is no need to call roff_userret() from here.
3791: * The read module will call that after rewinding the reader stack
3792: * to the place from where the current macro was called.
3793: */
1.212 schwarze 3794: static int
1.211 schwarze 3795: roff_return(ROFF_ARGS)
3796: {
3797: if (r->mstackpos >= 0)
1.212 schwarze 3798: return ROFF_IGN | ROFF_USERRET;
1.211 schwarze 3799:
1.222 schwarze 3800: mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "return");
1.211 schwarze 3801: return ROFF_IGN;
3802: }
3803:
1.212 schwarze 3804: static int
1.178 schwarze 3805: roff_rn(ROFF_ARGS)
3806: {
3807: const char *value;
3808: char *oldn, *newn, *end;
3809: size_t oldsz, newsz;
1.187 schwarze 3810: int deftype;
1.178 schwarze 3811:
3812: oldn = newn = buf->buf + pos;
3813: if (*oldn == '\0')
3814: return ROFF_IGN;
3815:
3816: oldsz = roff_getname(r, &newn, ln, pos);
1.235 schwarze 3817: if (oldn[oldsz] == '\\' || oldn[oldsz] == '\t' || *newn == '\0')
1.178 schwarze 3818: return ROFF_IGN;
3819:
3820: end = newn;
3821: newsz = roff_getname(r, &end, ln, newn - buf->buf);
3822: if (newsz == 0)
3823: return ROFF_IGN;
3824:
1.187 schwarze 3825: deftype = ROFFDEF_ANY;
3826: value = roff_getstrn(r, oldn, oldsz, &deftype);
3827: switch (deftype) {
3828: case ROFFDEF_USER:
1.178 schwarze 3829: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3830: roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
3831: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.187 schwarze 3832: break;
3833: case ROFFDEF_PRE:
3834: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3835: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3836: break;
3837: case ROFFDEF_REN:
1.178 schwarze 3838: roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
3839: roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.187 schwarze 3840: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3841: break;
3842: case ROFFDEF_STD:
1.178 schwarze 3843: roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.187 schwarze 3844: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3845: break;
3846: default:
3847: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3848: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3849: break;
3850: }
1.178 schwarze 3851: return ROFF_IGN;
3852: }
3853:
1.212 schwarze 3854: static int
1.211 schwarze 3855: roff_shift(ROFF_ARGS)
3856: {
3857: struct mctx *ctx;
3858: int levels, i;
3859:
3860: levels = 1;
3861: if (buf->buf[pos] != '\0' &&
3862: roff_evalnum(r, ln, buf->buf, &pos, &levels, 0) == 0) {
1.222 schwarze 3863: mandoc_msg(MANDOCERR_CE_NONUM,
1.211 schwarze 3864: ln, pos, "shift %s", buf->buf + pos);
3865: levels = 1;
3866: }
3867: if (r->mstackpos < 0) {
1.222 schwarze 3868: mandoc_msg(MANDOCERR_REQ_NOMAC, ln, ppos, "shift");
1.211 schwarze 3869: return ROFF_IGN;
3870: }
3871: ctx = r->mstack + r->mstackpos;
3872: if (levels > ctx->argc) {
1.222 schwarze 3873: mandoc_msg(MANDOCERR_SHIFT,
1.211 schwarze 3874: ln, pos, "%d, but max is %d", levels, ctx->argc);
3875: levels = ctx->argc;
3876: }
3877: if (levels == 0)
3878: return ROFF_IGN;
3879: for (i = 0; i < levels; i++)
3880: free(ctx->argv[i]);
3881: ctx->argc -= levels;
3882: for (i = 0; i < ctx->argc; i++)
3883: ctx->argv[i] = ctx->argv[i + levels];
3884: return ROFF_IGN;
3885: }
3886:
1.212 schwarze 3887: static int
1.14 schwarze 3888: roff_so(ROFF_ARGS)
3889: {
1.121 schwarze 3890: char *name, *cp;
1.15 schwarze 3891:
1.110 schwarze 3892: name = buf->buf + pos;
1.222 schwarze 3893: mandoc_msg(MANDOCERR_SO, ln, ppos, "so %s", name);
1.14 schwarze 3894:
1.22 schwarze 3895: /*
3896: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3897: * opening anything that's not in our cwd or anything beneath
3898: * it. Thus, explicitly disallow traversing up the file-system
3899: * or using absolute paths.
3900: */
3901:
1.110 schwarze 3902: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.222 schwarze 3903: mandoc_msg(MANDOCERR_SO_PATH, ln, ppos, ".so %s", name);
1.121 schwarze 3904: buf->sz = mandoc_asprintf(&cp,
3905: ".sp\nSee the file %s.\n.sp", name) + 1;
3906: free(buf->buf);
3907: buf->buf = cp;
3908: *offs = 0;
1.149 schwarze 3909: return ROFF_REPARSE;
1.14 schwarze 3910: }
3911:
3912: *offs = pos;
1.149 schwarze 3913: return ROFF_SO;
1.7 schwarze 3914: }
3915:
1.138 schwarze 3916: /* --- user defined strings and macros ------------------------------------ */
3917:
1.212 schwarze 3918: static int
1.16 schwarze 3919: roff_userdef(ROFF_ARGS)
1.12 schwarze 3920: {
1.211 schwarze 3921: struct mctx *ctx;
3922: char *arg, *ap, *dst, *src;
3923: size_t sz;
3924:
1.237 schwarze 3925: /* If the macro is empty, ignore it altogether. */
3926:
3927: if (*r->current_string == '\0')
3928: return ROFF_IGN;
3929:
1.211 schwarze 3930: /* Initialize a new macro stack context. */
3931:
3932: if (++r->mstackpos == r->mstacksz) {
3933: r->mstack = mandoc_recallocarray(r->mstack,
3934: r->mstacksz, r->mstacksz + 8, sizeof(*r->mstack));
3935: r->mstacksz += 8;
3936: }
3937: ctx = r->mstack + r->mstackpos;
3938: ctx->argsz = 0;
3939: ctx->argc = 0;
3940: ctx->argv = NULL;
3941:
3942: /*
3943: * Collect pointers to macro argument strings,
3944: * NUL-terminating them and escaping quotes.
3945: */
3946:
3947: src = buf->buf + pos;
3948: while (*src != '\0') {
3949: if (ctx->argc == ctx->argsz) {
3950: ctx->argsz += 8;
3951: ctx->argv = mandoc_reallocarray(ctx->argv,
3952: ctx->argsz, sizeof(*ctx->argv));
3953: }
1.227 schwarze 3954: arg = roff_getarg(r, &src, ln, &pos);
1.211 schwarze 3955: sz = 1; /* For the terminating NUL. */
3956: for (ap = arg; *ap != '\0'; ap++)
3957: sz += *ap == '"' ? 4 : 1;
3958: ctx->argv[ctx->argc++] = dst = mandoc_malloc(sz);
3959: for (ap = arg; *ap != '\0'; ap++) {
3960: if (*ap == '"') {
3961: memcpy(dst, "\\(dq", 4);
3962: dst += 4;
3963: } else
3964: *dst++ = *ap;
1.16 schwarze 3965: }
1.211 schwarze 3966: *dst = '\0';
1.227 schwarze 3967: free(arg);
1.209 schwarze 3968: }
3969:
1.211 schwarze 3970: /* Replace the macro invocation by the macro definition. */
1.135 schwarze 3971:
1.110 schwarze 3972: free(buf->buf);
1.211 schwarze 3973: buf->buf = mandoc_strdup(r->current_string);
3974: buf->sz = strlen(buf->buf) + 1;
1.120 schwarze 3975: *offs = 0;
1.16 schwarze 3976:
1.237 schwarze 3977: return buf->buf[buf->sz - 2] == '\n' ?
1.212 schwarze 3978: ROFF_REPARSE | ROFF_USERCALL : ROFF_IGN | ROFF_APPEND;
1.12 schwarze 3979: }
1.28 schwarze 3980:
1.178 schwarze 3981: /*
3982: * Calling a high-level macro that was renamed with .rn.
3983: * r->current_string has already been set up by roff_parse().
3984: */
1.212 schwarze 3985: static int
1.178 schwarze 3986: roff_renamed(ROFF_ARGS)
3987: {
3988: char *nbuf;
3989:
1.187 schwarze 3990: buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
3991: buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.178 schwarze 3992: free(buf->buf);
3993: buf->buf = nbuf;
1.201 schwarze 3994: *offs = 0;
1.178 schwarze 3995: return ROFF_CONT;
3996: }
3997:
1.234 schwarze 3998: /*
3999: * Measure the length in bytes of the roff identifier at *cpp
4000: * and advance the pointer to the next word.
4001: */
1.85 schwarze 4002: static size_t
1.28 schwarze 4003: roff_getname(struct roff *r, char **cpp, int ln, int pos)
4004: {
4005: char *name, *cp;
1.85 schwarze 4006: size_t namesz;
1.28 schwarze 4007:
4008: name = *cpp;
1.234 schwarze 4009: if (*name == '\0')
1.149 schwarze 4010: return 0;
1.28 schwarze 4011:
1.234 schwarze 4012: /* Advance cp to the byte after the end of the name. */
4013:
1.85 schwarze 4014: for (cp = name; 1; cp++) {
1.234 schwarze 4015: namesz = cp - name;
1.235 schwarze 4016: if (*cp == '\0')
1.85 schwarze 4017: break;
1.235 schwarze 4018: if (*cp == ' ' || *cp == '\t') {
4019: cp++;
4020: break;
4021: }
1.234 schwarze 4022: if (*cp != '\\')
1.28 schwarze 4023: continue;
1.234 schwarze 4024: if (cp[1] == '{' || cp[1] == '}')
1.88 schwarze 4025: break;
1.234 schwarze 4026: if (*++cp == '\\')
1.28 schwarze 4027: continue;
1.222 schwarze 4028: mandoc_msg(MANDOCERR_NAMESC, ln, pos,
1.97 schwarze 4029: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 4030: mandoc_escape((const char **)&cp, NULL, NULL);
4031: break;
1.28 schwarze 4032: }
4033:
4034: /* Read past spaces. */
1.234 schwarze 4035:
4036: while (*cp == ' ')
1.28 schwarze 4037: cp++;
4038:
4039: *cpp = cp;
1.149 schwarze 4040: return namesz;
1.28 schwarze 4041: }
4042:
1.16 schwarze 4043: /*
4044: * Store *string into the user-defined string called *name.
4045: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 4046: * append == 0: replace mode
4047: * append == 1: single-line append mode
4048: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 4049: */
1.8 schwarze 4050: static void
1.16 schwarze 4051: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 4052: int append)
1.7 schwarze 4053: {
1.187 schwarze 4054: size_t namesz;
1.42 schwarze 4055:
1.187 schwarze 4056: namesz = strlen(name);
4057: roff_setstrn(&r->strtab, name, namesz, string,
1.80 schwarze 4058: string ? strlen(string) : 0, append);
1.187 schwarze 4059: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.42 schwarze 4060: }
4061:
4062: static void
4063: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 4064: const char *string, size_t stringsz, int append)
1.42 schwarze 4065: {
4066: struct roffkv *n;
4067: char *c;
4068: int i;
4069: size_t oldch, newch;
1.7 schwarze 4070:
1.16 schwarze 4071: /* Search for an existing string with the same name. */
1.42 schwarze 4072: n = *r;
4073:
1.84 schwarze 4074: while (n && (namesz != n->key.sz ||
4075: strncmp(n->key.p, name, namesz)))
1.7 schwarze 4076: n = n->next;
1.8 schwarze 4077:
4078: if (NULL == n) {
1.16 schwarze 4079: /* Create a new string table entry. */
1.42 schwarze 4080: n = mandoc_malloc(sizeof(struct roffkv));
4081: n->key.p = mandoc_strndup(name, namesz);
4082: n->key.sz = namesz;
4083: n->val.p = NULL;
4084: n->val.sz = 0;
4085: n->next = *r;
4086: *r = n;
1.66 schwarze 4087: } else if (0 == append) {
1.42 schwarze 4088: free(n->val.p);
4089: n->val.p = NULL;
4090: n->val.sz = 0;
1.16 schwarze 4091: }
4092:
4093: if (NULL == string)
4094: return;
4095:
4096: /*
4097: * One additional byte for the '\n' in multiline mode,
4098: * and one for the terminating '\0'.
4099: */
1.66 schwarze 4100: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 4101:
4102: if (NULL == n->val.p) {
4103: n->val.p = mandoc_malloc(newch);
4104: *n->val.p = '\0';
1.16 schwarze 4105: oldch = 0;
4106: } else {
1.42 schwarze 4107: oldch = n->val.sz;
4108: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 4109: }
4110:
4111: /* Skip existing content in the destination buffer. */
1.42 schwarze 4112: c = n->val.p + (int)oldch;
1.16 schwarze 4113:
4114: /* Append new content to the destination buffer. */
1.42 schwarze 4115: i = 0;
4116: while (i < (int)stringsz) {
1.16 schwarze 4117: /*
4118: * Rudimentary roff copy mode:
4119: * Handle escaped backslashes.
4120: */
1.42 schwarze 4121: if ('\\' == string[i] && '\\' == string[i + 1])
4122: i++;
4123: *c++ = string[i++];
1.16 schwarze 4124: }
1.8 schwarze 4125:
1.16 schwarze 4126: /* Append terminating bytes. */
1.66 schwarze 4127: if (1 < append)
1.16 schwarze 4128: *c++ = '\n';
1.42 schwarze 4129:
1.16 schwarze 4130: *c = '\0';
1.42 schwarze 4131: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 4132: }
4133:
1.8 schwarze 4134: static const char *
1.197 schwarze 4135: roff_getstrn(struct roff *r, const char *name, size_t len,
1.187 schwarze 4136: int *deftype)
1.7 schwarze 4137: {
1.187 schwarze 4138: const struct roffkv *n;
1.197 schwarze 4139: int found, i;
1.187 schwarze 4140: enum roff_tok tok;
4141:
1.197 schwarze 4142: found = 0;
4143: for (n = r->strtab; n != NULL; n = n->next) {
4144: if (strncmp(name, n->key.p, len) != 0 ||
4145: n->key.p[len] != '\0' || n->val.p == NULL)
4146: continue;
4147: if (*deftype & ROFFDEF_USER) {
4148: *deftype = ROFFDEF_USER;
4149: return n->val.p;
4150: } else {
4151: found = 1;
4152: break;
4153: }
4154: }
4155: for (n = r->rentab; n != NULL; n = n->next) {
4156: if (strncmp(name, n->key.p, len) != 0 ||
4157: n->key.p[len] != '\0' || n->val.p == NULL)
4158: continue;
4159: if (*deftype & ROFFDEF_REN) {
4160: *deftype = ROFFDEF_REN;
4161: return n->val.p;
4162: } else {
4163: found = 1;
4164: break;
1.187 schwarze 4165: }
4166: }
1.197 schwarze 4167: for (i = 0; i < PREDEFS_MAX; i++) {
4168: if (strncmp(name, predefs[i].name, len) != 0 ||
4169: predefs[i].name[len] != '\0')
4170: continue;
4171: if (*deftype & ROFFDEF_PRE) {
4172: *deftype = ROFFDEF_PRE;
4173: return predefs[i].str;
4174: } else {
4175: found = 1;
4176: break;
1.187 schwarze 4177: }
4178: }
1.228 schwarze 4179: if (r->man->meta.macroset != MACROSET_MAN) {
1.197 schwarze 4180: for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
4181: if (strncmp(name, roff_name[tok], len) != 0 ||
4182: roff_name[tok][len] != '\0')
4183: continue;
4184: if (*deftype & ROFFDEF_STD) {
4185: *deftype = ROFFDEF_STD;
4186: return NULL;
4187: } else {
4188: found = 1;
4189: break;
1.187 schwarze 4190: }
4191: }
4192: }
1.228 schwarze 4193: if (r->man->meta.macroset != MACROSET_MDOC) {
1.197 schwarze 4194: for (tok = MAN_TH; tok < MAN_MAX; tok++) {
4195: if (strncmp(name, roff_name[tok], len) != 0 ||
4196: roff_name[tok][len] != '\0')
4197: continue;
4198: if (*deftype & ROFFDEF_STD) {
4199: *deftype = ROFFDEF_STD;
4200: return NULL;
4201: } else {
4202: found = 1;
4203: break;
1.187 schwarze 4204: }
4205: }
1.197 schwarze 4206: }
4207:
4208: if (found == 0 && *deftype != ROFFDEF_ANY) {
4209: if (*deftype & ROFFDEF_REN) {
4210: /*
4211: * This might still be a request,
4212: * so do not treat it as undefined yet.
4213: */
4214: *deftype = ROFFDEF_UNDEF;
4215: return NULL;
1.187 schwarze 4216: }
1.197 schwarze 4217:
4218: /* Using an undefined string defines it to be empty. */
4219:
4220: roff_setstrn(&r->strtab, name, len, "", 0, 0);
4221: roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.187 schwarze 4222: }
1.197 schwarze 4223:
1.187 schwarze 4224: *deftype = 0;
1.149 schwarze 4225: return NULL;
1.7 schwarze 4226: }
4227:
1.8 schwarze 4228: static void
1.42 schwarze 4229: roff_freestr(struct roffkv *r)
1.7 schwarze 4230: {
1.42 schwarze 4231: struct roffkv *n, *nn;
1.7 schwarze 4232:
1.42 schwarze 4233: for (n = r; n; n = nn) {
4234: free(n->key.p);
4235: free(n->val.p);
1.7 schwarze 4236: nn = n->next;
4237: free(n);
4238: }
1.27 schwarze 4239: }
1.138 schwarze 4240:
4241: /* --- accessors and utility functions ------------------------------------ */
1.42 schwarze 4242:
4243: /*
4244: * Duplicate an input string, making the appropriate character
4245: * conversations (as stipulated by `tr') along the way.
4246: * Returns a heap-allocated string with all the replacements made.
4247: */
4248: char *
4249: roff_strdup(const struct roff *r, const char *p)
4250: {
4251: const struct roffkv *cp;
4252: char *res;
4253: const char *pp;
4254: size_t ssz, sz;
4255: enum mandoc_esc esc;
4256:
4257: if (NULL == r->xmbtab && NULL == r->xtab)
1.149 schwarze 4258: return mandoc_strdup(p);
1.42 schwarze 4259: else if ('\0' == *p)
1.149 schwarze 4260: return mandoc_strdup("");
1.42 schwarze 4261:
4262: /*
4263: * Step through each character looking for term matches
4264: * (remember that a `tr' can be invoked with an escape, which is
4265: * a glyph but the escape is multi-character).
4266: * We only do this if the character hash has been initialised
4267: * and the string is >0 length.
4268: */
4269:
4270: res = NULL;
4271: ssz = 0;
4272:
4273: while ('\0' != *p) {
1.161 schwarze 4274: assert((unsigned int)*p < 128);
4275: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42 schwarze 4276: sz = r->xtab[(int)*p].sz;
4277: res = mandoc_realloc(res, ssz + sz + 1);
4278: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
4279: ssz += sz;
4280: p++;
4281: continue;
4282: } else if ('\\' != *p) {
4283: res = mandoc_realloc(res, ssz + 2);
4284: res[ssz++] = *p++;
4285: continue;
4286: }
4287:
4288: /* Search for term matches. */
4289: for (cp = r->xmbtab; cp; cp = cp->next)
4290: if (0 == strncmp(p, cp->key.p, cp->key.sz))
4291: break;
4292:
4293: if (NULL != cp) {
4294: /*
4295: * A match has been found.
4296: * Append the match to the array and move
4297: * forward by its keysize.
4298: */
1.80 schwarze 4299: res = mandoc_realloc(res,
4300: ssz + cp->val.sz + 1);
1.42 schwarze 4301: memcpy(res + ssz, cp->val.p, cp->val.sz);
4302: ssz += cp->val.sz;
4303: p += (int)cp->key.sz;
4304: continue;
4305: }
4306:
4307: /*
4308: * Handle escapes carefully: we need to copy
4309: * over just the escape itself, or else we might
4310: * do replacements within the escape itself.
4311: * Make sure to pass along the bogus string.
4312: */
4313: pp = p++;
4314: esc = mandoc_escape(&p, NULL, NULL);
4315: if (ESCAPE_ERROR == esc) {
4316: sz = strlen(pp);
4317: res = mandoc_realloc(res, ssz + sz + 1);
4318: memcpy(res + ssz, pp, sz);
4319: break;
4320: }
1.80 schwarze 4321: /*
4322: * We bail out on bad escapes.
1.42 schwarze 4323: * No need to warn: we already did so when
1.227 schwarze 4324: * roff_expand() was called.
1.42 schwarze 4325: */
4326: sz = (int)(p - pp);
4327: res = mandoc_realloc(res, ssz + sz + 1);
4328: memcpy(res + ssz, pp, sz);
4329: ssz += sz;
4330: }
4331:
4332: res[(int)ssz] = '\0';
1.149 schwarze 4333: return res;
1.99 schwarze 4334: }
4335:
4336: int
4337: roff_getformat(const struct roff *r)
4338: {
4339:
1.149 schwarze 4340: return r->format;
1.48 schwarze 4341: }
4342:
4343: /*
1.80 schwarze 4344: * Find out whether a line is a macro line or not.
1.48 schwarze 4345: * If it is, adjust the current position and return one; if it isn't,
4346: * return zero and don't change the current position.
4347: * If the control character has been set with `.cc', then let that grain
4348: * precedence.
4349: * This is slighly contrary to groff, where using the non-breaking
4350: * control character when `cc' has been invoked will cause the
4351: * non-breaking macro contents to be printed verbatim.
4352: */
4353: int
4354: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
4355: {
4356: int pos;
4357:
4358: pos = *ppos;
4359:
1.175 schwarze 4360: if (r->control != '\0' && cp[pos] == r->control)
1.48 schwarze 4361: pos++;
1.175 schwarze 4362: else if (r->control != '\0')
1.149 schwarze 4363: return 0;
1.48 schwarze 4364: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
4365: pos += 2;
4366: else if ('.' == cp[pos] || '\'' == cp[pos])
4367: pos++;
4368: else
1.149 schwarze 4369: return 0;
1.48 schwarze 4370:
4371: while (' ' == cp[pos] || '\t' == cp[pos])
4372: pos++;
4373:
4374: *ppos = pos;
1.149 schwarze 4375: return 1;
1.1 schwarze 4376: }
![[BACK]](/icons/back.gif){kind=icon}
Return to roff.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / mandoc