Annotation of src/usr.bin/mandoc/roff.c, Revision 1.174
1.174 ! schwarze 1: /* $OpenBSD: roff.c,v 1.173 2017/05/08 15:33:43 schwarze Exp $ */
1.1 schwarze 2: /*
1.139 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.159 schwarze 4: * Copyright (c) 2010-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
1.1 schwarze 5: *
6: * Permission to use, copy, modify, and distribute this software for any
7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
9: *
1.16 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1 schwarze 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1 schwarze 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17: */
1.99 schwarze 18: #include <sys/types.h>
19:
1.1 schwarze 20: #include <assert.h>
1.3 schwarze 21: #include <ctype.h>
1.117 schwarze 22: #include <limits.h>
1.167 schwarze 23: #include <stddef.h>
24: #include <stdint.h>
1.51 schwarze 25: #include <stdio.h>
1.1 schwarze 26: #include <stdlib.h>
27: #include <string.h>
28:
29: #include "mandoc.h"
1.74 schwarze 30: #include "mandoc_aux.h"
1.167 schwarze 31: #include "mandoc_ohash.h"
1.137 schwarze 32: #include "roff.h"
1.98 daniel 33: #include "libmandoc.h"
1.138 schwarze 34: #include "roff_int.h"
1.27 schwarze 35: #include "libroff.h"
1.2 schwarze 36:
1.43 schwarze 37: /* Maximum number of string expansions per line, to break infinite loops. */
38: #define EXPAND_LIMIT 1000
39:
1.138 schwarze 40: /* --- data types --------------------------------------------------------- */
41:
1.41 schwarze 42: /*
1.42 schwarze 43: * An incredibly-simple string buffer.
44: */
1.8 schwarze 45: struct roffstr {
1.42 schwarze 46: char *p; /* nil-terminated buffer */
47: size_t sz; /* saved strlen(p) */
48: };
49:
50: /*
51: * A key-value roffstr pair as part of a singly-linked list.
52: */
53: struct roffkv {
54: struct roffstr key;
55: struct roffstr val;
56: struct roffkv *next; /* next in list */
1.8 schwarze 57: };
58:
1.52 schwarze 59: /*
60: * A single number register as part of a singly-linked list.
61: */
62: struct roffreg {
63: struct roffstr key;
1.53 schwarze 64: int val;
1.52 schwarze 65: struct roffreg *next;
66: };
67:
1.167 schwarze 68: /*
69: * Association of request and macro names with token IDs.
70: */
71: struct roffreq {
72: enum roff_tok tok;
73: char name[];
74: };
75:
1.1 schwarze 76: struct roff {
1.35 schwarze 77: struct mparse *parse; /* parse point */
1.168 schwarze 78: struct roff_man *man; /* mdoc or man parser */
1.1 schwarze 79: struct roffnode *last; /* leaf of stack */
1.96 schwarze 80: int *rstack; /* stack of inverted `ie' values */
1.167 schwarze 81: struct ohash *reqtab; /* request lookup table */
1.52 schwarze 82: struct roffreg *regtab; /* number registers */
1.42 schwarze 83: struct roffkv *strtab; /* user-defined strings & macros */
84: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
85: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.16 schwarze 86: const char *current_string; /* value of last called user macro */
1.27 schwarze 87: struct tbl_node *first_tbl; /* first table parsed */
88: struct tbl_node *last_tbl; /* last table parsed */
89: struct tbl_node *tbl; /* current table being parsed */
1.32 schwarze 90: struct eqn_node *last_eqn; /* last equation parsed */
91: struct eqn_node *first_eqn; /* first equation parsed */
92: struct eqn_node *eqn; /* current equation being parsed */
1.102 schwarze 93: int eqn_inline; /* current equation is inline */
1.96 schwarze 94: int options; /* parse options */
95: int rstacksz; /* current size limit of rstack */
96: int rstackpos; /* position in rstack */
1.99 schwarze 97: int format; /* current file in mdoc or man format */
1.145 schwarze 98: int argc; /* number of args of the last macro */
1.96 schwarze 99: char control; /* control character */
1.1 schwarze 100: };
101:
102: struct roffnode {
1.166 schwarze 103: enum roff_tok tok; /* type of node */
1.1 schwarze 104: struct roffnode *parent; /* up one in stack */
105: int line; /* parse line */
106: int col; /* parse col */
1.16 schwarze 107: char *name; /* node name, e.g. macro name */
1.2 schwarze 108: char *end; /* end-rules: custom token */
109: int endspan; /* end-rules: next-line or infty */
1.71 schwarze 110: int rule; /* current evaluation rule */
1.1 schwarze 111: };
112:
113: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.166 schwarze 114: enum roff_tok tok, /* tok of macro */ \
1.110 schwarze 115: struct buf *buf, /* input buffer */ \
1.1 schwarze 116: int ln, /* parse line */ \
1.2 schwarze 117: int ppos, /* original pos in buffer */ \
118: int pos, /* current pos in buffer */ \
119: int *offs /* reset offset of buffer data */
1.1 schwarze 120:
121: typedef enum rofferr (*roffproc)(ROFF_ARGS);
122:
123: struct roffmac {
1.2 schwarze 124: roffproc proc; /* process new macro */
125: roffproc text; /* process as child text of macro */
126: roffproc sub; /* process as child of macro */
127: int flags;
128: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.1 schwarze 129: };
130:
1.37 schwarze 131: struct predef {
132: const char *name; /* predefined input name */
133: const char *str; /* replacement symbol */
134: };
135:
136: #define PREDEF(__name, __str) \
137: { (__name), (__str) },
138:
1.138 schwarze 139: /* --- function prototypes ------------------------------------------------ */
140:
1.42 schwarze 141: static void roffnode_cleanscope(struct roff *);
142: static void roffnode_pop(struct roff *);
1.166 schwarze 143: static void roffnode_push(struct roff *, enum roff_tok,
1.42 schwarze 144: const char *, int, int);
1.2 schwarze 145: static enum rofferr roff_block(ROFF_ARGS);
146: static enum rofferr roff_block_text(ROFF_ARGS);
147: static enum rofferr roff_block_sub(ROFF_ARGS);
1.168 schwarze 148: static enum rofferr roff_br(ROFF_ARGS);
1.2 schwarze 149: static enum rofferr roff_cblock(ROFF_ARGS);
1.48 schwarze 150: static enum rofferr roff_cc(ROFF_ARGS);
1.68 schwarze 151: static void roff_ccond(struct roff *, int, int);
1.2 schwarze 152: static enum rofferr roff_cond(ROFF_ARGS);
153: static enum rofferr roff_cond_text(ROFF_ARGS);
154: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.7 schwarze 155: static enum rofferr roff_ds(ROFF_ARGS);
1.110 schwarze 156: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.143 schwarze 157: static int roff_evalcond(struct roff *r, int, char *, int *);
1.106 schwarze 158: static int roff_evalnum(struct roff *, int,
159: const char *, int *, int *, int);
160: static int roff_evalpar(struct roff *, int,
1.133 schwarze 161: const char *, int *, int *, int);
1.71 schwarze 162: static int roff_evalstrcond(const char *, int *);
1.42 schwarze 163: static void roff_free1(struct roff *);
1.52 schwarze 164: static void roff_freereg(struct roffreg *);
1.42 schwarze 165: static void roff_freestr(struct roffkv *);
1.85 schwarze 166: static size_t roff_getname(struct roff *, char **, int, int);
1.133 schwarze 167: static int roff_getnum(const char *, int *, int *, int);
1.56 schwarze 168: static int roff_getop(const char *, int *, char *);
1.53 schwarze 169: static int roff_getregn(const struct roff *,
170: const char *, size_t);
1.145 schwarze 171: static int roff_getregro(const struct roff *,
172: const char *name);
1.80 schwarze 173: static const char *roff_getstrn(const struct roff *,
1.8 schwarze 174: const char *, size_t);
1.143 schwarze 175: static int roff_hasregn(const struct roff *,
176: const char *, size_t);
1.123 schwarze 177: static enum rofferr roff_insec(ROFF_ARGS);
1.51 schwarze 178: static enum rofferr roff_it(ROFF_ARGS);
1.21 schwarze 179: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.137 schwarze 180: static void roff_man_alloc1(struct roff_man *);
181: static void roff_man_free1(struct roff_man *);
1.172 schwarze 182: static enum rofferr roff_manyarg(ROFF_ARGS);
1.6 schwarze 183: static enum rofferr roff_nr(ROFF_ARGS);
1.169 schwarze 184: static enum rofferr roff_onearg(ROFF_ARGS);
1.166 schwarze 185: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.87 schwarze 186: int, int);
1.110 schwarze 187: static enum rofferr roff_parsetext(struct buf *, int, int *);
188: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.29 schwarze 189: static enum rofferr roff_rm(ROFF_ARGS);
1.76 schwarze 190: static enum rofferr roff_rr(ROFF_ARGS);
1.8 schwarze 191: static void roff_setstr(struct roff *,
1.16 schwarze 192: const char *, const char *, int);
1.80 schwarze 193: static void roff_setstrn(struct roffkv **, const char *,
1.42 schwarze 194: size_t, const char *, size_t, int);
1.14 schwarze 195: static enum rofferr roff_so(ROFF_ARGS);
1.42 schwarze 196: static enum rofferr roff_tr(ROFF_ARGS);
1.47 schwarze 197: static enum rofferr roff_Dd(ROFF_ARGS);
198: static enum rofferr roff_TH(ROFF_ARGS);
1.27 schwarze 199: static enum rofferr roff_TE(ROFF_ARGS);
200: static enum rofferr roff_TS(ROFF_ARGS);
1.32 schwarze 201: static enum rofferr roff_EQ(ROFF_ARGS);
202: static enum rofferr roff_EN(ROFF_ARGS);
1.27 schwarze 203: static enum rofferr roff_T_(ROFF_ARGS);
1.123 schwarze 204: static enum rofferr roff_unsupp(ROFF_ARGS);
1.16 schwarze 205: static enum rofferr roff_userdef(ROFF_ARGS);
1.1 schwarze 206:
1.138 schwarze 207: /* --- constant data ------------------------------------------------------ */
208:
1.133 schwarze 209: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
210: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
211:
1.166 schwarze 212: const char *__roff_name[MAN_MAX + 1] = {
1.171 schwarze 213: "br", "ft", "ll", "sp",
1.173 schwarze 214: "ta", "ti", NULL,
1.166 schwarze 215: "ab", "ad", "af", "aln",
216: "als", "am", "am1", "ami",
217: "ami1", "as", "as1", "asciify",
218: "backtrace", "bd", "bleedat", "blm",
219: "box", "boxa", "bp", "BP",
220: "break", "breakchar", "brnl", "brp",
221: "brpnl", "c2", "cc", "ce",
222: "cf", "cflags", "ch", "char",
223: "chop", "class", "close", "CL",
224: "color", "composite", "continue", "cp",
225: "cropat", "cs", "cu", "da",
226: "dch", "Dd", "de", "de1",
227: "defcolor", "dei", "dei1", "device",
228: "devicem", "di", "do", "ds",
229: "ds1", "dwh", "dt", "ec",
230: "ecr", "ecs", "el", "em",
231: "EN", "eo", "EP", "EQ",
232: "errprint", "ev", "evc", "ex",
233: "fallback", "fam", "fc", "fchar",
234: "fcolor", "fdeferlig", "feature", "fkern",
235: "fl", "flig", "fp", "fps",
236: "fschar", "fspacewidth", "fspecial", "ftr",
237: "fzoom", "gcolor", "hc", "hcode",
238: "hidechar", "hla", "hlm", "hpf",
239: "hpfa", "hpfcode", "hw", "hy",
240: "hylang", "hylen", "hym", "hypp",
241: "hys", "ie", "if", "ig",
242: "index", "it", "itc", "IX",
243: "kern", "kernafter", "kernbefore", "kernpair",
244: "lc", "lc_ctype", "lds", "length",
245: "letadj", "lf", "lg", "lhang",
246: "linetabs", "lnr", "lnrf", "lpfx",
247: "ls", "lsm", "lt", "mc",
248: "mediasize", "minss", "mk", "mso",
249: "na", "ne", "nh", "nhychar",
250: "nm", "nn", "nop", "nr",
251: "nrf", "nroff", "ns", "nx",
252: "open", "opena", "os", "output",
253: "padj", "papersize", "pc", "pev",
254: "pi", "PI", "pl", "pm",
255: "pn", "pnr", "po", "ps",
256: "psbb", "pshape", "pso", "ptr",
257: "pvs", "rchar", "rd", "recursionlimit",
258: "return", "rfschar", "rhang", "rj",
259: "rm", "rn", "rnn", "rr",
260: "rs", "rt", "schar", "sentchar",
261: "shc", "shift", "sizes", "so",
262: "spacewidth", "special", "spreadwarn", "ss",
263: "sty", "substring", "sv", "sy",
1.172 schwarze 264: "T&", "tc", "TE",
1.173 schwarze 265: "TH", "tkf", "tl",
1.166 schwarze 266: "tm", "tm1", "tmc", "tr",
267: "track", "transchar", "trf", "trimat",
268: "trin", "trnt", "troff", "TS",
269: "uf", "ul", "unformat", "unwatch",
270: "unwatchn", "vpt", "vs", "warn",
271: "warnscale", "watch", "watchlength", "watchn",
272: "wh", "while", "write", "writec",
273: "writem", "xflag", ".", NULL,
274: "text",
275: "Dd", "Dt", "Os", "Sh",
276: "Ss", "Pp", "D1", "Dl",
277: "Bd", "Ed", "Bl", "El",
278: "It", "Ad", "An", "Ap",
279: "Ar", "Cd", "Cm", "Dv",
280: "Er", "Ev", "Ex", "Fa",
281: "Fd", "Fl", "Fn", "Ft",
282: "Ic", "In", "Li", "Nd",
283: "Nm", "Op", "Ot", "Pa",
284: "Rv", "St", "Va", "Vt",
285: "Xr", "%A", "%B", "%D",
286: "%I", "%J", "%N", "%O",
287: "%P", "%R", "%T", "%V",
288: "Ac", "Ao", "Aq", "At",
289: "Bc", "Bf", "Bo", "Bq",
290: "Bsx", "Bx", "Db", "Dc",
291: "Do", "Dq", "Ec", "Ef",
292: "Em", "Eo", "Fx", "Ms",
293: "No", "Ns", "Nx", "Ox",
294: "Pc", "Pf", "Po", "Pq",
295: "Qc", "Ql", "Qo", "Qq",
296: "Re", "Rs", "Sc", "So",
297: "Sq", "Sm", "Sx", "Sy",
298: "Tn", "Ux", "Xc", "Xo",
299: "Fo", "Fc", "Oo", "Oc",
300: "Bk", "Ek", "Bt", "Hf",
301: "Fr", "Ud", "Lb", "Lp",
302: "Lk", "Mt", "Brq", "Bro",
303: "Brc", "%C", "Es", "En",
1.171 schwarze 304: "Dx", "%Q", "%U", "Ta",
305: NULL,
1.166 schwarze 306: "TH", "SH", "SS", "TP",
307: "LP", "PP", "P", "IP",
308: "HP", "SM", "SB", "BI",
309: "IB", "BR", "RB", "R",
310: "B", "I", "IR", "RI",
1.171 schwarze 311: "nf", "fi",
1.166 schwarze 312: "RE", "RS", "DT", "UC",
1.169 schwarze 313: "PD", "AT", "in",
1.166 schwarze 314: "OP", "EX", "EE", "UR",
1.170 schwarze 315: "UE", NULL
1.166 schwarze 316: };
317: const char *const *roff_name = __roff_name;
318:
319: static struct roffmac roffs[TOKEN_NONE] = {
1.168 schwarze 320: { roff_br, NULL, NULL, 0 }, /* br */
1.169 schwarze 321: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.170 schwarze 322: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.171 schwarze 323: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.172 schwarze 324: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.173 schwarze 325: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.168 schwarze 326: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.167 schwarze 327: { roff_unsupp, NULL, NULL, 0 }, /* ab */
328: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
329: { roff_line_ignore, NULL, NULL, 0 }, /* af */
330: { roff_unsupp, NULL, NULL, 0 }, /* aln */
331: { roff_unsupp, NULL, NULL, 0 }, /* als */
332: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
333: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
334: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
335: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
336: { roff_ds, NULL, NULL, 0 }, /* as */
337: { roff_ds, NULL, NULL, 0 }, /* as1 */
338: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
339: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
340: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
341: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
342: { roff_unsupp, NULL, NULL, 0 }, /* blm */
343: { roff_unsupp, NULL, NULL, 0 }, /* box */
344: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
345: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
346: { roff_unsupp, NULL, NULL, 0 }, /* BP */
347: { roff_unsupp, NULL, NULL, 0 }, /* break */
348: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
349: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.168 schwarze 350: { roff_br, NULL, NULL, 0 }, /* brp */
1.167 schwarze 351: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
352: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
353: { roff_cc, NULL, NULL, 0 }, /* cc */
354: { roff_line_ignore, NULL, NULL, 0 }, /* ce */
355: { roff_insec, NULL, NULL, 0 }, /* cf */
356: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
357: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
358: { roff_unsupp, NULL, NULL, 0 }, /* char */
359: { roff_unsupp, NULL, NULL, 0 }, /* chop */
360: { roff_line_ignore, NULL, NULL, 0 }, /* class */
361: { roff_insec, NULL, NULL, 0 }, /* close */
362: { roff_unsupp, NULL, NULL, 0 }, /* CL */
363: { roff_line_ignore, NULL, NULL, 0 }, /* color */
364: { roff_unsupp, NULL, NULL, 0 }, /* composite */
365: { roff_unsupp, NULL, NULL, 0 }, /* continue */
366: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
367: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
368: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
369: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
370: { roff_unsupp, NULL, NULL, 0 }, /* da */
371: { roff_unsupp, NULL, NULL, 0 }, /* dch */
372: { roff_Dd, NULL, NULL, 0 }, /* Dd */
373: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
374: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
375: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
376: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
377: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
378: { roff_unsupp, NULL, NULL, 0 }, /* device */
379: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
380: { roff_unsupp, NULL, NULL, 0 }, /* di */
381: { roff_unsupp, NULL, NULL, 0 }, /* do */
382: { roff_ds, NULL, NULL, 0 }, /* ds */
383: { roff_ds, NULL, NULL, 0 }, /* ds1 */
384: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
385: { roff_unsupp, NULL, NULL, 0 }, /* dt */
386: { roff_unsupp, NULL, NULL, 0 }, /* ec */
387: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
388: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
389: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
390: { roff_unsupp, NULL, NULL, 0 }, /* em */
391: { roff_EN, NULL, NULL, 0 }, /* EN */
392: { roff_unsupp, NULL, NULL, 0 }, /* eo */
393: { roff_unsupp, NULL, NULL, 0 }, /* EP */
394: { roff_EQ, NULL, NULL, 0 }, /* EQ */
395: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
396: { roff_unsupp, NULL, NULL, 0 }, /* ev */
397: { roff_unsupp, NULL, NULL, 0 }, /* evc */
398: { roff_unsupp, NULL, NULL, 0 }, /* ex */
399: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
400: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
401: { roff_unsupp, NULL, NULL, 0 }, /* fc */
402: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
403: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
404: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
405: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
406: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
407: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
408: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
409: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
410: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
411: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
412: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
413: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
414: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
415: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
416: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
417: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
418: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
419: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
420: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
421: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
422: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
423: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
424: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
425: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
426: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
427: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
428: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
429: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
430: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
431: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
432: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
433: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
434: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
435: { roff_unsupp, NULL, NULL, 0 }, /* index */
436: { roff_it, NULL, NULL, 0 }, /* it */
437: { roff_unsupp, NULL, NULL, 0 }, /* itc */
438: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
439: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
440: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
441: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
442: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
443: { roff_unsupp, NULL, NULL, 0 }, /* lc */
444: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
445: { roff_unsupp, NULL, NULL, 0 }, /* lds */
446: { roff_unsupp, NULL, NULL, 0 }, /* length */
447: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
448: { roff_insec, NULL, NULL, 0 }, /* lf */
449: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
450: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
451: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
452: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
453: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
454: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
455: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
456: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
457: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
458: { roff_line_ignore, NULL, NULL, 0 }, /* mc */
459: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
460: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
461: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
462: { roff_insec, NULL, NULL, 0 }, /* mso */
463: { roff_line_ignore, NULL, NULL, 0 }, /* na */
464: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
465: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
466: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
467: { roff_unsupp, NULL, NULL, 0 }, /* nm */
468: { roff_unsupp, NULL, NULL, 0 }, /* nn */
469: { roff_unsupp, NULL, NULL, 0 }, /* nop */
470: { roff_nr, NULL, NULL, 0 }, /* nr */
471: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
472: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
473: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
474: { roff_insec, NULL, NULL, 0 }, /* nx */
475: { roff_insec, NULL, NULL, 0 }, /* open */
476: { roff_insec, NULL, NULL, 0 }, /* opena */
477: { roff_line_ignore, NULL, NULL, 0 }, /* os */
478: { roff_unsupp, NULL, NULL, 0 }, /* output */
479: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
480: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
481: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
482: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
483: { roff_insec, NULL, NULL, 0 }, /* pi */
484: { roff_unsupp, NULL, NULL, 0 }, /* PI */
485: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
486: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
487: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
488: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
489: { roff_line_ignore, NULL, NULL, 0 }, /* po */
490: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
491: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
492: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
493: { roff_insec, NULL, NULL, 0 }, /* pso */
494: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
495: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
496: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
497: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
498: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
499: { roff_unsupp, NULL, NULL, 0 }, /* return */
500: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
501: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
502: { roff_line_ignore, NULL, NULL, 0 }, /* rj */
503: { roff_rm, NULL, NULL, 0 }, /* rm */
504: { roff_unsupp, NULL, NULL, 0 }, /* rn */
505: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
506: { roff_rr, NULL, NULL, 0 }, /* rr */
507: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
508: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
509: { roff_unsupp, NULL, NULL, 0 }, /* schar */
510: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
511: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
512: { roff_unsupp, NULL, NULL, 0 }, /* shift */
513: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
514: { roff_so, NULL, NULL, 0 }, /* so */
515: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
516: { roff_line_ignore, NULL, NULL, 0 }, /* special */
517: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
518: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
519: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
520: { roff_unsupp, NULL, NULL, 0 }, /* substring */
521: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
522: { roff_insec, NULL, NULL, 0 }, /* sy */
523: { roff_T_, NULL, NULL, 0 }, /* T& */
524: { roff_unsupp, NULL, NULL, 0 }, /* tc */
525: { roff_TE, NULL, NULL, 0 }, /* TE */
526: { roff_TH, NULL, NULL, 0 }, /* TH */
527: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
528: { roff_unsupp, NULL, NULL, 0 }, /* tl */
529: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
530: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
531: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
532: { roff_tr, NULL, NULL, 0 }, /* tr */
533: { roff_line_ignore, NULL, NULL, 0 }, /* track */
534: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
535: { roff_insec, NULL, NULL, 0 }, /* trf */
536: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
537: { roff_unsupp, NULL, NULL, 0 }, /* trin */
538: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
539: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
540: { roff_TS, NULL, NULL, 0 }, /* TS */
541: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
542: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
543: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
544: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
545: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
546: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
547: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
548: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
549: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
550: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
551: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
552: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
553: { roff_unsupp, NULL, NULL, 0 }, /* wh */
554: { roff_unsupp, NULL, NULL, 0 }, /* while */
555: { roff_insec, NULL, NULL, 0 }, /* write */
556: { roff_insec, NULL, NULL, 0 }, /* writec */
557: { roff_insec, NULL, NULL, 0 }, /* writem */
558: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
559: { roff_cblock, NULL, NULL, 0 }, /* . */
560: { roff_userdef, NULL, NULL, 0 }
1.1 schwarze 561: };
562:
1.73 schwarze 563: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47 schwarze 564: const char *const __mdoc_reserved[] = {
565: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
566: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
567: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
568: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73 schwarze 569: "Dt", "Dv", "Dx", "D1",
570: "Ec", "Ed", "Ef", "Ek", "El", "Em",
571: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47 schwarze 572: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73 schwarze 573: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
574: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47 schwarze 575: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73 schwarze 576: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
577: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
578: "Sc", "Sh", "Sm", "So", "Sq",
1.47 schwarze 579: "Ss", "St", "Sx", "Sy",
580: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73 schwarze 581: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47 schwarze 582: "%P", "%Q", "%R", "%T", "%U", "%V",
583: NULL
584: };
585:
1.73 schwarze 586: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47 schwarze 587: const char *const __man_reserved[] = {
1.73 schwarze 588: "AT", "B", "BI", "BR", "DT",
589: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
590: "LP", "OP", "P", "PD", "PP",
591: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
592: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47 schwarze 593: NULL
594: };
595:
1.37 schwarze 596: /* Array of injected predefined strings. */
597: #define PREDEFS_MAX 38
598: static const struct predef predefs[PREDEFS_MAX] = {
599: #include "predefs.in"
600: };
601:
1.51 schwarze 602: static int roffit_lines; /* number of lines to delay */
603: static char *roffit_macro; /* nil-terminated macro line */
604:
1.80 schwarze 605:
1.138 schwarze 606: /* --- request table ------------------------------------------------------ */
607:
1.167 schwarze 608: struct ohash *
609: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3 schwarze 610: {
1.167 schwarze 611: struct ohash *htab;
612: struct roffreq *req;
613: enum roff_tok tok;
614: size_t sz;
615: unsigned int slot;
1.3 schwarze 616:
1.167 schwarze 617: htab = mandoc_malloc(sizeof(*htab));
618: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3 schwarze 619:
1.167 schwarze 620: for (tok = mintok; tok < maxtok; tok++) {
1.168 schwarze 621: if (roff_name[tok] == NULL)
622: continue;
1.167 schwarze 623: sz = strlen(roff_name[tok]);
624: req = mandoc_malloc(sizeof(*req) + sz + 1);
625: req->tok = tok;
626: memcpy(req->name, roff_name[tok], sz + 1);
627: slot = ohash_qlookup(htab, req->name);
628: ohash_insert(htab, slot, req);
1.3 schwarze 629: }
1.167 schwarze 630: return htab;
1.3 schwarze 631: }
632:
1.167 schwarze 633: void
634: roffhash_free(struct ohash *htab)
1.1 schwarze 635: {
1.167 schwarze 636: struct roffreq *req;
637: unsigned int slot;
1.1 schwarze 638:
1.167 schwarze 639: if (htab == NULL)
640: return;
641: for (req = ohash_first(htab, &slot); req != NULL;
642: req = ohash_next(htab, &slot))
643: free(req);
644: ohash_delete(htab);
645: free(htab);
646: }
647:
648: enum roff_tok
649: roffhash_find(struct ohash *htab, const char *name, size_t sz)
650: {
651: struct roffreq *req;
652: const char *end;
653:
654: if (sz) {
655: end = name + sz;
656: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
657: } else
658: req = ohash_find(htab, ohash_qlookup(htab, name));
659: return req == NULL ? TOKEN_NONE : req->tok;
1.1 schwarze 660: }
661:
1.138 schwarze 662: /* --- stack of request blocks -------------------------------------------- */
663:
1.1 schwarze 664: /*
665: * Pop the current node off of the stack of roff instructions currently
666: * pending.
667: */
668: static void
669: roffnode_pop(struct roff *r)
670: {
671: struct roffnode *p;
672:
1.2 schwarze 673: assert(r->last);
1.80 schwarze 674: p = r->last;
1.2 schwarze 675:
676: r->last = r->last->parent;
1.16 schwarze 677: free(p->name);
678: free(p->end);
1.1 schwarze 679: free(p);
680: }
681:
682: /*
683: * Push a roff node onto the instruction stack. This must later be
684: * removed with roffnode_pop().
685: */
1.11 schwarze 686: static void
1.166 schwarze 687: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16 schwarze 688: int line, int col)
1.1 schwarze 689: {
690: struct roffnode *p;
691:
1.11 schwarze 692: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 693: p->tok = tok;
1.16 schwarze 694: if (name)
695: p->name = mandoc_strdup(name);
1.1 schwarze 696: p->parent = r->last;
697: p->line = line;
698: p->col = col;
1.71 schwarze 699: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 700:
701: r->last = p;
702: }
703:
1.138 schwarze 704: /* --- roff parser state data management ---------------------------------- */
705:
1.1 schwarze 706: static void
707: roff_free1(struct roff *r)
708: {
1.49 schwarze 709: struct tbl_node *tbl;
1.32 schwarze 710: struct eqn_node *e;
1.42 schwarze 711: int i;
1.27 schwarze 712:
1.49 schwarze 713: while (NULL != (tbl = r->first_tbl)) {
714: r->first_tbl = tbl->next;
715: tbl_free(tbl);
1.27 schwarze 716: }
717: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 718:
1.32 schwarze 719: while (NULL != (e = r->first_eqn)) {
720: r->first_eqn = e->next;
721: eqn_free(e);
722: }
723: r->first_eqn = r->last_eqn = r->eqn = NULL;
724:
1.1 schwarze 725: while (r->last)
726: roffnode_pop(r);
1.27 schwarze 727:
1.96 schwarze 728: free (r->rstack);
729: r->rstack = NULL;
730: r->rstacksz = 0;
731: r->rstackpos = -1;
732:
733: roff_freereg(r->regtab);
734: r->regtab = NULL;
735:
1.42 schwarze 736: roff_freestr(r->strtab);
737: roff_freestr(r->xmbtab);
738: r->strtab = r->xmbtab = NULL;
739:
740: if (r->xtab)
741: for (i = 0; i < 128; i++)
742: free(r->xtab[i].p);
743: free(r->xtab);
744: r->xtab = NULL;
1.1 schwarze 745: }
746:
747: void
748: roff_reset(struct roff *r)
749: {
750: roff_free1(r);
1.99 schwarze 751: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48 schwarze 752: r->control = 0;
1.1 schwarze 753: }
754:
755: void
756: roff_free(struct roff *r)
757: {
758: roff_free1(r);
1.167 schwarze 759: roffhash_free(r->reqtab);
1.1 schwarze 760: free(r);
761: }
762:
763: struct roff *
1.151 schwarze 764: roff_alloc(struct mparse *parse, int options)
1.1 schwarze 765: {
766: struct roff *r;
767:
1.11 schwarze 768: r = mandoc_calloc(1, sizeof(struct roff));
1.35 schwarze 769: r->parse = parse;
1.167 schwarze 770: r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.72 schwarze 771: r->options = options;
1.99 schwarze 772: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2 schwarze 773: r->rstackpos = -1;
1.149 schwarze 774: return r;
1.137 schwarze 775: }
776:
1.138 schwarze 777: /* --- syntax tree state data management ---------------------------------- */
778:
1.137 schwarze 779: static void
780: roff_man_free1(struct roff_man *man)
781: {
782:
1.138 schwarze 783: if (man->first != NULL)
784: roff_node_delete(man, man->first);
1.137 schwarze 785: free(man->meta.msec);
786: free(man->meta.vol);
787: free(man->meta.os);
788: free(man->meta.arch);
789: free(man->meta.title);
790: free(man->meta.name);
791: free(man->meta.date);
792: }
793:
794: static void
795: roff_man_alloc1(struct roff_man *man)
796: {
797:
798: memset(&man->meta, 0, sizeof(man->meta));
799: man->first = mandoc_calloc(1, sizeof(*man->first));
800: man->first->type = ROFFT_ROOT;
801: man->last = man->first;
802: man->last_es = NULL;
803: man->flags = 0;
804: man->macroset = MACROSET_NONE;
805: man->lastsec = man->lastnamed = SEC_NONE;
806: man->next = ROFF_NEXT_CHILD;
807: }
808:
809: void
810: roff_man_reset(struct roff_man *man)
811: {
812:
813: roff_man_free1(man);
814: roff_man_alloc1(man);
815: }
816:
817: void
818: roff_man_free(struct roff_man *man)
819: {
820:
821: roff_man_free1(man);
822: free(man);
823: }
824:
825: struct roff_man *
826: roff_man_alloc(struct roff *roff, struct mparse *parse,
827: const char *defos, int quick)
828: {
829: struct roff_man *man;
830:
831: man = mandoc_calloc(1, sizeof(*man));
832: man->parse = parse;
833: man->roff = roff;
834: man->defos = defos;
835: man->quick = quick;
836: roff_man_alloc1(man);
1.168 schwarze 837: roff->man = man;
1.149 schwarze 838: return man;
1.1 schwarze 839: }
840:
1.138 schwarze 841: /* --- syntax tree handling ----------------------------------------------- */
842:
843: struct roff_node *
844: roff_node_alloc(struct roff_man *man, int line, int pos,
845: enum roff_type type, int tok)
846: {
847: struct roff_node *n;
848:
849: n = mandoc_calloc(1, sizeof(*n));
850: n->line = line;
851: n->pos = pos;
852: n->tok = tok;
853: n->type = type;
854: n->sec = man->lastsec;
855:
856: if (man->flags & MDOC_SYNOPSIS)
1.157 schwarze 857: n->flags |= NODE_SYNPRETTY;
1.138 schwarze 858: else
1.157 schwarze 859: n->flags &= ~NODE_SYNPRETTY;
1.138 schwarze 860: if (man->flags & MDOC_NEWLINE)
1.157 schwarze 861: n->flags |= NODE_LINE;
1.138 schwarze 862: man->flags &= ~MDOC_NEWLINE;
863:
1.149 schwarze 864: return n;
1.138 schwarze 865: }
866:
867: void
868: roff_node_append(struct roff_man *man, struct roff_node *n)
869: {
870:
871: switch (man->next) {
872: case ROFF_NEXT_SIBLING:
1.153 schwarze 873: if (man->last->next != NULL) {
874: n->next = man->last->next;
875: man->last->next->prev = n;
876: } else
877: man->last->parent->last = n;
1.138 schwarze 878: man->last->next = n;
879: n->prev = man->last;
880: n->parent = man->last->parent;
881: break;
882: case ROFF_NEXT_CHILD:
1.159 schwarze 883: if (man->last->child != NULL) {
884: n->next = man->last->child;
885: man->last->child->prev = n;
886: } else
887: man->last->last = n;
1.138 schwarze 888: man->last->child = n;
889: n->parent = man->last;
890: break;
891: default:
892: abort();
893: }
1.154 schwarze 894: man->last = n;
1.138 schwarze 895:
896: switch (n->type) {
897: case ROFFT_HEAD:
898: n->parent->head = n;
899: break;
900: case ROFFT_BODY:
1.154 schwarze 901: if (n->end != ENDBODY_NOT)
902: return;
1.138 schwarze 903: n->parent->body = n;
904: break;
905: case ROFFT_TAIL:
906: n->parent->tail = n;
907: break;
908: default:
1.154 schwarze 909: return;
1.138 schwarze 910: }
1.154 schwarze 911:
912: /*
913: * Copy over the normalised-data pointer of our parent. Not
914: * everybody has one, but copying a null pointer is fine.
915: */
916:
917: n->norm = n->parent->norm;
918: assert(n->parent->type == ROFFT_BLOCK);
1.138 schwarze 919: }
920:
1.139 schwarze 921: void
922: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
923: {
924: struct roff_node *n;
925:
926: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
927: n->string = roff_strdup(man->roff, word);
928: roff_node_append(man, n);
1.158 schwarze 929: n->flags |= NODE_VALID | NODE_ENDED;
1.139 schwarze 930: man->next = ROFF_NEXT_SIBLING;
931: }
932:
933: void
934: roff_word_append(struct roff_man *man, const char *word)
935: {
936: struct roff_node *n;
937: char *addstr, *newstr;
938:
939: n = man->last;
940: addstr = roff_strdup(man->roff, word);
941: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
942: free(addstr);
943: free(n->string);
944: n->string = newstr;
945: man->next = ROFF_NEXT_SIBLING;
1.140 schwarze 946: }
947:
948: void
949: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
950: {
951: struct roff_node *n;
952:
953: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
954: roff_node_append(man, n);
955: man->next = ROFF_NEXT_CHILD;
956: }
957:
958: struct roff_node *
959: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
960: {
961: struct roff_node *n;
962:
963: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
964: roff_node_append(man, n);
965: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 966: return n;
1.139 schwarze 967: }
968:
1.138 schwarze 969: struct roff_node *
970: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
971: {
972: struct roff_node *n;
973:
974: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
975: roff_node_append(man, n);
976: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 977: return n;
1.138 schwarze 978: }
979:
980: struct roff_node *
981: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
982: {
983: struct roff_node *n;
984:
985: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
986: roff_node_append(man, n);
987: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 988: return n;
1.139 schwarze 989: }
990:
991: void
992: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
993: {
994: struct roff_node *n;
995:
996: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
997: n->eqn = eqn;
998: if (eqn->ln > man->last->line)
1.157 schwarze 999: n->flags |= NODE_LINE;
1.139 schwarze 1000: roff_node_append(man, n);
1001: man->next = ROFF_NEXT_SIBLING;
1002: }
1003:
1004: void
1005: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
1006: {
1007: struct roff_node *n;
1008:
1009: if (man->macroset == MACROSET_MAN)
1.174 ! schwarze 1010: man_breakscope(man, ROFF_TS);
1.139 schwarze 1011: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
1012: n->span = tbl;
1013: roff_node_append(man, n);
1.158 schwarze 1014: n->flags |= NODE_VALID | NODE_ENDED;
1.139 schwarze 1015: man->next = ROFF_NEXT_SIBLING;
1.138 schwarze 1016: }
1017:
1018: void
1019: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1020: {
1021:
1022: /* Adjust siblings. */
1023:
1024: if (n->prev)
1025: n->prev->next = n->next;
1026: if (n->next)
1027: n->next->prev = n->prev;
1028:
1029: /* Adjust parent. */
1030:
1031: if (n->parent != NULL) {
1032: if (n->parent->child == n)
1033: n->parent->child = n->next;
1034: if (n->parent->last == n)
1035: n->parent->last = n->prev;
1036: }
1037:
1038: /* Adjust parse point. */
1039:
1040: if (man == NULL)
1041: return;
1042: if (man->last == n) {
1043: if (n->prev == NULL) {
1044: man->last = n->parent;
1045: man->next = ROFF_NEXT_CHILD;
1046: } else {
1047: man->last = n->prev;
1048: man->next = ROFF_NEXT_SIBLING;
1049: }
1050: }
1051: if (man->first == n)
1052: man->first = NULL;
1053: }
1054:
1055: void
1056: roff_node_free(struct roff_node *n)
1057: {
1058:
1059: if (n->args != NULL)
1060: mdoc_argv_free(n->args);
1061: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1062: free(n->norm);
1063: free(n->string);
1064: free(n);
1065: }
1066:
1067: void
1068: roff_node_delete(struct roff_man *man, struct roff_node *n)
1069: {
1070:
1071: while (n->child != NULL)
1072: roff_node_delete(man, n->child);
1073: roff_node_unlink(man, n);
1074: roff_node_free(n);
1.141 schwarze 1075: }
1076:
1077: void
1078: deroff(char **dest, const struct roff_node *n)
1079: {
1080: char *cp;
1081: size_t sz;
1082:
1083: if (n->type != ROFFT_TEXT) {
1084: for (n = n->child; n != NULL; n = n->next)
1085: deroff(dest, n);
1086: return;
1087: }
1088:
1.160 schwarze 1089: /* Skip leading whitespace. */
1.141 schwarze 1090:
1.160 schwarze 1091: for (cp = n->string; *cp != '\0'; cp++) {
1.161 schwarze 1092: if (cp[0] == '\\' && cp[1] != '\0' &&
1093: strchr(" %&0^|~", cp[1]) != NULL)
1.141 schwarze 1094: cp++;
1.160 schwarze 1095: else if ( ! isspace((unsigned char)*cp))
1.141 schwarze 1096: break;
1097: }
1098:
1.161 schwarze 1099: /* Skip trailing backslash. */
1100:
1101: sz = strlen(cp);
1.162 schwarze 1102: if (sz > 0 && cp[sz - 1] == '\\')
1.161 schwarze 1103: sz--;
1104:
1.141 schwarze 1105: /* Skip trailing whitespace. */
1106:
1.161 schwarze 1107: for (; sz; sz--)
1.141 schwarze 1108: if ( ! isspace((unsigned char)cp[sz-1]))
1109: break;
1110:
1111: /* Skip empty strings. */
1112:
1113: if (sz == 0)
1114: return;
1115:
1116: if (*dest == NULL) {
1117: *dest = mandoc_strndup(cp, sz);
1118: return;
1119: }
1120:
1121: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1122: free(*dest);
1123: *dest = cp;
1.138 schwarze 1124: }
1125:
1126: /* --- main functions of the roff parser ---------------------------------- */
1127:
1.8 schwarze 1128: /*
1.79 schwarze 1129: * In the current line, expand escape sequences that tend to get
1130: * used in numerical expressions and conditional requests.
1131: * Also check the syntax of the remaining escape sequences.
1.8 schwarze 1132: */
1.45 schwarze 1133: static enum rofferr
1.110 schwarze 1134: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8 schwarze 1135: {
1.81 schwarze 1136: char ubuf[24]; /* buffer to print the number */
1.78 schwarze 1137: const char *start; /* start of the string to process */
1.82 schwarze 1138: char *stesc; /* start of an escape sequence ('\\') */
1.23 schwarze 1139: const char *stnam; /* start of the name, after "[(*" */
1140: const char *cp; /* end of the name, e.g. before ']' */
1141: const char *res; /* the string to be substituted */
1.110 schwarze 1142: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 1143: size_t maxl; /* expected length of the escape name */
1144: size_t naml; /* actual length of the escape name */
1.109 schwarze 1145: enum mandoc_esc esc; /* type of the escape sequence */
1146: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 1147: int expand_count; /* to avoid infinite loops */
1.79 schwarze 1148: int npos; /* position in numeric expression */
1.91 schwarze 1149: int arg_complete; /* argument not interrupted by eol */
1.79 schwarze 1150: char term; /* character terminating the escape */
1.8 schwarze 1151:
1.43 schwarze 1152: expand_count = 0;
1.110 schwarze 1153: start = buf->buf + pos;
1.78 schwarze 1154: stesc = strchr(start, '\0') - 1;
1155: while (stesc-- > start) {
1.43 schwarze 1156:
1.78 schwarze 1157: /* Search backwards for the next backslash. */
1158:
1.110 schwarze 1159: if (*stesc != '\\')
1.78 schwarze 1160: continue;
1161:
1162: /* If it is escaped, skip it. */
1163:
1164: for (cp = stesc - 1; cp >= start; cp--)
1.110 schwarze 1165: if (*cp != '\\')
1.78 schwarze 1166: break;
1167:
1.110 schwarze 1168: if ((stesc - cp) % 2 == 0) {
1.82 schwarze 1169: stesc = (char *)cp;
1.78 schwarze 1170: continue;
1171: }
1.23 schwarze 1172:
1.79 schwarze 1173: /* Decide whether to expand or to check only. */
1.23 schwarze 1174:
1.79 schwarze 1175: term = '\0';
1.78 schwarze 1176: cp = stesc + 1;
1.53 schwarze 1177: switch (*cp) {
1.80 schwarze 1178: case '*':
1.53 schwarze 1179: res = NULL;
1180: break;
1.80 schwarze 1181: case 'B':
1182: case 'w':
1.79 schwarze 1183: term = cp[1];
1184: /* FALLTHROUGH */
1.80 schwarze 1185: case 'n':
1.53 schwarze 1186: res = ubuf;
1187: break;
1188: default:
1.109 schwarze 1189: esc = mandoc_escape(&cp, &stnam, &inaml);
1190: if (esc == ESCAPE_ERROR ||
1191: (esc == ESCAPE_SPECIAL &&
1.151 schwarze 1192: mchars_spec2cp(stnam, inaml) < 0))
1.92 schwarze 1193: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110 schwarze 1194: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1195: "%.*s", (int)(cp - stesc), stesc);
1.78 schwarze 1196: continue;
1.42 schwarze 1197: }
1198:
1.78 schwarze 1199: if (EXPAND_LIMIT < ++expand_count) {
1200: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110 schwarze 1201: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1202: return ROFF_IGN;
1.78 schwarze 1203: }
1.23 schwarze 1204:
1205: /*
1206: * The third character decides the length
1.53 schwarze 1207: * of the name of the string or register.
1.23 schwarze 1208: * Save a pointer to the name.
1209: */
1210:
1.110 schwarze 1211: if (term == '\0') {
1.79 schwarze 1212: switch (*++cp) {
1.80 schwarze 1213: case '\0':
1.79 schwarze 1214: maxl = 0;
1215: break;
1.80 schwarze 1216: case '(':
1.79 schwarze 1217: cp++;
1218: maxl = 2;
1219: break;
1.80 schwarze 1220: case '[':
1.79 schwarze 1221: cp++;
1222: term = ']';
1223: maxl = 0;
1224: break;
1225: default:
1226: maxl = 1;
1227: break;
1228: }
1229: } else {
1230: cp += 2;
1.8 schwarze 1231: maxl = 0;
1232: }
1.23 schwarze 1233: stnam = cp;
1.8 schwarze 1234:
1.23 schwarze 1235: /* Advance to the end of the name. */
1.8 schwarze 1236:
1.125 schwarze 1237: naml = 0;
1.91 schwarze 1238: arg_complete = 1;
1.125 schwarze 1239: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1240: if (*cp == '\0') {
1.92 schwarze 1241: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 1242: ln, (int)(stesc - buf->buf), stesc);
1.91 schwarze 1243: arg_complete = 0;
1.79 schwarze 1244: break;
1.42 schwarze 1245: }
1.110 schwarze 1246: if (maxl == 0 && *cp == term) {
1.79 schwarze 1247: cp++;
1.125 schwarze 1248: break;
1249: }
1250: if (*cp++ != '\\' || stesc[1] != 'w') {
1251: naml++;
1252: continue;
1253: }
1254: switch (mandoc_escape(&cp, NULL, NULL)) {
1255: case ESCAPE_SPECIAL:
1256: case ESCAPE_UNICODE:
1257: case ESCAPE_NUMBERED:
1258: case ESCAPE_OVERSTRIKE:
1259: naml++;
1260: break;
1261: default:
1.8 schwarze 1262: break;
1.79 schwarze 1263: }
1.8 schwarze 1264: }
1265:
1.23 schwarze 1266: /*
1267: * Retrieve the replacement string; if it is
1268: * undefined, resume searching for escapes.
1269: */
1270:
1.79 schwarze 1271: switch (stesc[1]) {
1.80 schwarze 1272: case '*':
1.91 schwarze 1273: if (arg_complete)
1274: res = roff_getstrn(r, stnam, naml);
1.79 schwarze 1275: break;
1.80 schwarze 1276: case 'B':
1.79 schwarze 1277: npos = 0;
1.91 schwarze 1278: ubuf[0] = arg_complete &&
1.133 schwarze 1279: roff_evalnum(r, ln, stnam, &npos,
1280: NULL, ROFFNUM_SCALE) &&
1.91 schwarze 1281: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1282: ubuf[1] = '\0';
1283: break;
1.80 schwarze 1284: case 'n':
1.91 schwarze 1285: if (arg_complete)
1286: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1287: roff_getregn(r, stnam, naml));
1288: else
1289: ubuf[0] = '\0';
1.79 schwarze 1290: break;
1.80 schwarze 1291: case 'w':
1.91 schwarze 1292: /* use even incomplete args */
1.81 schwarze 1293: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1294: 24 * (int)naml);
1295: break;
1296: }
1.8 schwarze 1297:
1.110 schwarze 1298: if (res == NULL) {
1.92 schwarze 1299: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1300: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1301: "%.*s", (int)naml, stnam);
1.37 schwarze 1302: res = "";
1.118 schwarze 1303: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1304: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1305: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1306: return ROFF_IGN;
1.8 schwarze 1307: }
1308:
1.23 schwarze 1309: /* Replace the escape sequence by the string. */
1310:
1.82 schwarze 1311: *stesc = '\0';
1.110 schwarze 1312: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1313: buf->buf, res, cp) + 1;
1.8 schwarze 1314:
1.78 schwarze 1315: /* Prepare for the next replacement. */
1.8 schwarze 1316:
1.78 schwarze 1317: start = nbuf + pos;
1.110 schwarze 1318: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1319: free(buf->buf);
1320: buf->buf = nbuf;
1.42 schwarze 1321: }
1.149 schwarze 1322: return ROFF_CONT;
1.42 schwarze 1323: }
1324:
1325: /*
1.147 schwarze 1326: * Process text streams.
1.42 schwarze 1327: */
1328: static enum rofferr
1.110 schwarze 1329: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42 schwarze 1330: {
1331: size_t sz;
1332: const char *start;
1.51 schwarze 1333: char *p;
1334: int isz;
1.42 schwarze 1335: enum mandoc_esc esc;
1336:
1.147 schwarze 1337: /* Spring the input line trap. */
1338:
1339: if (roffit_lines == 1) {
1340: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1341: free(buf->buf);
1342: buf->buf = p;
1343: buf->sz = isz + 1;
1344: *offs = 0;
1345: free(roffit_macro);
1346: roffit_lines = 0;
1.149 schwarze 1347: return ROFF_REPARSE;
1.147 schwarze 1348: } else if (roffit_lines > 1)
1349: --roffit_lines;
1350:
1351: /* Convert all breakable hyphens into ASCII_HYPH. */
1352:
1.110 schwarze 1353: start = p = buf->buf + pos;
1.42 schwarze 1354:
1.110 schwarze 1355: while (*p != '\0') {
1.42 schwarze 1356: sz = strcspn(p, "-\\");
1357: p += sz;
1358:
1.110 schwarze 1359: if (*p == '\0')
1.42 schwarze 1360: break;
1361:
1.110 schwarze 1362: if (*p == '\\') {
1.42 schwarze 1363: /* Skip over escapes. */
1364: p++;
1.62 schwarze 1365: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1366: if (esc == ESCAPE_ERROR)
1.42 schwarze 1367: break;
1.136 schwarze 1368: while (*p == '-')
1369: p++;
1.42 schwarze 1370: continue;
1371: } else if (p == start) {
1372: p++;
1373: continue;
1374: }
1375:
1.44 schwarze 1376: if (isalpha((unsigned char)p[-1]) &&
1377: isalpha((unsigned char)p[1]))
1.42 schwarze 1378: *p = ASCII_HYPH;
1379: p++;
1.8 schwarze 1380: }
1.149 schwarze 1381: return ROFF_CONT;
1.8 schwarze 1382: }
1383:
1.1 schwarze 1384: enum rofferr
1.110 schwarze 1385: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1386: {
1.166 schwarze 1387: enum roff_tok t;
1.27 schwarze 1388: enum rofferr e;
1.110 schwarze 1389: int pos; /* parse point */
1.115 schwarze 1390: int spos; /* saved parse point for messages */
1.110 schwarze 1391: int ppos; /* original offset in buf->buf */
1392: int ctl; /* macro line (boolean) */
1393:
1394: ppos = pos = *offs;
1.1 schwarze 1395:
1.102 schwarze 1396: /* Handle in-line equation delimiters. */
1397:
1.108 schwarze 1398: if (r->tbl == NULL &&
1399: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1400: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1401: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1402: if (e == ROFF_REPARSE)
1.149 schwarze 1403: return e;
1.102 schwarze 1404: assert(e == ROFF_CONT);
1405: }
1406:
1407: /* Expand some escape sequences. */
1.8 schwarze 1408:
1.110 schwarze 1409: e = roff_res(r, buf, ln, pos);
1410: if (e == ROFF_IGN)
1.149 schwarze 1411: return e;
1.110 schwarze 1412: assert(e == ROFF_CONT);
1.8 schwarze 1413:
1.110 schwarze 1414: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1415:
1.8 schwarze 1416: /*
1.2 schwarze 1417: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1418: * text through the macro's filter.
1419: * Equations process all content themselves.
1420: * Tables process almost all content themselves, but we want
1421: * to warn about macros before passing it there.
1.2 schwarze 1422: */
1423:
1.124 schwarze 1424: if (r->last != NULL && ! ctl) {
1.2 schwarze 1425: t = r->last->tok;
1.110 schwarze 1426: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166 schwarze 1427: if (e == ROFF_IGN)
1.149 schwarze 1428: return e;
1.166 schwarze 1429: assert(e == ROFF_CONT);
1.54 schwarze 1430: }
1.124 schwarze 1431: if (r->eqn != NULL)
1.149 schwarze 1432: return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.124 schwarze 1433: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.149 schwarze 1434: return tbl_read(r->tbl, ln, buf->buf, ppos);
1.124 schwarze 1435: if ( ! ctl)
1.149 schwarze 1436: return roff_parsetext(buf, pos, offs);
1.100 schwarze 1437:
1438: /* Skip empty request lines. */
1439:
1.110 schwarze 1440: if (buf->buf[pos] == '"') {
1.100 schwarze 1441: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1442: ln, pos, NULL);
1.149 schwarze 1443: return ROFF_IGN;
1.110 schwarze 1444: } else if (buf->buf[pos] == '\0')
1.149 schwarze 1445: return ROFF_IGN;
1.2 schwarze 1446:
1447: /*
1448: * If a scope is open, go to the child handler for that macro,
1449: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1450: * Don't do so if an equation is open.
1.2 schwarze 1451: */
1452:
1453: if (r->last) {
1.1 schwarze 1454: t = r->last->tok;
1.149 schwarze 1455: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1456: }
1457:
1.115 schwarze 1458: /* No scope is open. This is a new request or macro. */
1459:
1460: spos = pos;
1461: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1462:
1463: /* Tables ignore most macros. */
1464:
1.166 schwarze 1465: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS)) {
1.115 schwarze 1466: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1467: ln, pos, buf->buf + spos);
1.129 schwarze 1468: if (t == ROFF_TS)
1.149 schwarze 1469: return ROFF_IGN;
1.129 schwarze 1470: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1471: pos++;
1.163 schwarze 1472: while (buf->buf[pos] == ' ')
1.129 schwarze 1473: pos++;
1.149 schwarze 1474: return tbl_read(r->tbl, ln, buf->buf, pos);
1.115 schwarze 1475: }
1476:
1.2 schwarze 1477: /*
1.115 schwarze 1478: * This is neither a roff request nor a user-defined macro.
1479: * Let the standard macro set parsers handle it.
1.2 schwarze 1480: */
1481:
1.166 schwarze 1482: if (t == TOKEN_NONE)
1.149 schwarze 1483: return ROFF_CONT;
1.115 schwarze 1484:
1485: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1486:
1.168 schwarze 1487: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2 schwarze 1488: }
1489:
1.27 schwarze 1490: void
1.2 schwarze 1491: roff_endparse(struct roff *r)
1492: {
1.1 schwarze 1493:
1.27 schwarze 1494: if (r->last)
1.94 schwarze 1495: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1496: r->last->line, r->last->col,
1.167 schwarze 1497: roff_name[r->last->tok]);
1.27 schwarze 1498:
1.32 schwarze 1499: if (r->eqn) {
1.94 schwarze 1500: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1501: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41 schwarze 1502: eqn_end(&r->eqn);
1.32 schwarze 1503: }
1504:
1.27 schwarze 1505: if (r->tbl) {
1.94 schwarze 1506: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1507: r->tbl->line, r->tbl->pos, "TS");
1.41 schwarze 1508: tbl_end(&r->tbl);
1.27 schwarze 1509: }
1.1 schwarze 1510: }
1511:
1512: /*
1513: * Parse a roff node's type from the input buffer. This must be in the
1514: * form of ".foo xxx" in the usual way.
1515: */
1.166 schwarze 1516: static enum roff_tok
1.87 schwarze 1517: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1518: {
1.87 schwarze 1519: char *cp;
1.16 schwarze 1520: const char *mac;
1521: size_t maclen;
1.166 schwarze 1522: enum roff_tok t;
1.1 schwarze 1523:
1.87 schwarze 1524: cp = buf + *pos;
1525:
1526: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166 schwarze 1527: return TOKEN_NONE;
1.1 schwarze 1528:
1.87 schwarze 1529: mac = cp;
1530: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 1531:
1.16 schwarze 1532: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.167 schwarze 1533: ? ROFF_USERDEF : roffhash_find(r->reqtab, mac, maclen);
1.1 schwarze 1534:
1.166 schwarze 1535: if (t != TOKEN_NONE)
1.87 schwarze 1536: *pos = cp - buf;
1.1 schwarze 1537:
1.149 schwarze 1538: return t;
1.1 schwarze 1539: }
1540:
1.138 schwarze 1541: /* --- handling of request blocks ----------------------------------------- */
1542:
1.1 schwarze 1543: static enum rofferr
1.2 schwarze 1544: roff_cblock(ROFF_ARGS)
1.1 schwarze 1545: {
1546:
1.2 schwarze 1547: /*
1548: * A block-close `..' should only be invoked as a child of an
1549: * ignore macro, otherwise raise a warning and just ignore it.
1550: */
1551:
1.110 schwarze 1552: if (r->last == NULL) {
1.94 schwarze 1553: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1554: ln, ppos, "..");
1.149 schwarze 1555: return ROFF_IGN;
1.2 schwarze 1556: }
1.1 schwarze 1557:
1.2 schwarze 1558: switch (r->last->tok) {
1.80 schwarze 1559: case ROFF_am:
1.93 schwarze 1560: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80 schwarze 1561: case ROFF_ami:
1562: case ROFF_de:
1.23 schwarze 1563: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80 schwarze 1564: case ROFF_dei:
1565: case ROFF_ig:
1.2 schwarze 1566: break;
1567: default:
1.94 schwarze 1568: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1569: ln, ppos, "..");
1.149 schwarze 1570: return ROFF_IGN;
1.2 schwarze 1571: }
1572:
1.110 schwarze 1573: if (buf->buf[pos] != '\0')
1.90 schwarze 1574: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 1575: ".. %s", buf->buf + pos);
1.2 schwarze 1576:
1577: roffnode_pop(r);
1578: roffnode_cleanscope(r);
1.149 schwarze 1579: return ROFF_IGN;
1.2 schwarze 1580:
1581: }
1.1 schwarze 1582:
1.2 schwarze 1583: static void
1584: roffnode_cleanscope(struct roff *r)
1585: {
1.1 schwarze 1586:
1.2 schwarze 1587: while (r->last) {
1.46 schwarze 1588: if (--r->last->endspan != 0)
1.2 schwarze 1589: break;
1590: roffnode_pop(r);
1591: }
1592: }
1.1 schwarze 1593:
1.68 schwarze 1594: static void
1595: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 1596: {
1.1 schwarze 1597:
1.2 schwarze 1598: if (NULL == r->last) {
1.94 schwarze 1599: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1600: ln, ppos, "\\}");
1.68 schwarze 1601: return;
1.2 schwarze 1602: }
1.1 schwarze 1603:
1.2 schwarze 1604: switch (r->last->tok) {
1.80 schwarze 1605: case ROFF_el:
1606: case ROFF_ie:
1607: case ROFF_if:
1.2 schwarze 1608: break;
1609: default:
1.94 schwarze 1610: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1611: ln, ppos, "\\}");
1.68 schwarze 1612: return;
1.2 schwarze 1613: }
1.1 schwarze 1614:
1.2 schwarze 1615: if (r->last->endspan > -1) {
1.94 schwarze 1616: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1617: ln, ppos, "\\}");
1.68 schwarze 1618: return;
1.2 schwarze 1619: }
1620:
1621: roffnode_pop(r);
1622: roffnode_cleanscope(r);
1.68 schwarze 1623: return;
1.1 schwarze 1624: }
1625:
1626: static enum rofferr
1.2 schwarze 1627: roff_block(ROFF_ARGS)
1.1 schwarze 1628: {
1.93 schwarze 1629: const char *name;
1630: char *iname, *cp;
1.86 schwarze 1631: size_t namesz;
1.16 schwarze 1632:
1.93 schwarze 1633: /* Ignore groff compatibility mode for now. */
1.2 schwarze 1634:
1.110 schwarze 1635: if (tok == ROFF_de1)
1.93 schwarze 1636: tok = ROFF_de;
1.123 schwarze 1637: else if (tok == ROFF_dei1)
1638: tok = ROFF_dei;
1.110 schwarze 1639: else if (tok == ROFF_am1)
1.93 schwarze 1640: tok = ROFF_am;
1.123 schwarze 1641: else if (tok == ROFF_ami1)
1642: tok = ROFF_ami;
1.93 schwarze 1643:
1644: /* Parse the macro name argument. */
1645:
1.110 schwarze 1646: cp = buf->buf + pos;
1647: if (tok == ROFF_ig) {
1.93 schwarze 1648: iname = NULL;
1649: namesz = 0;
1650: } else {
1651: iname = cp;
1652: namesz = roff_getname(r, &cp, ln, ppos);
1653: iname[namesz] = '\0';
1654: }
1.22 schwarze 1655:
1.93 schwarze 1656: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 1657:
1.110 schwarze 1658: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1659: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1660: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1661: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1662: "%.*s", (int)namesz, iname);
1663: namesz = 0;
1664: } else
1665: namesz = strlen(name);
1666: } else
1667: name = iname;
1.22 schwarze 1668:
1.110 schwarze 1669: if (namesz == 0 && tok != ROFF_ig) {
1.93 schwarze 1670: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167 schwarze 1671: ln, ppos, roff_name[tok]);
1.149 schwarze 1672: return ROFF_IGN;
1.93 schwarze 1673: }
1.2 schwarze 1674:
1.16 schwarze 1675: roffnode_push(r, tok, name, ln, ppos);
1676:
1677: /*
1678: * At the beginning of a `de' macro, clear the existing string
1679: * with the same name, if there is one. New content will be
1.66 schwarze 1680: * appended from roff_block_text() in multiline mode.
1.16 schwarze 1681: */
1.22 schwarze 1682:
1.110 schwarze 1683: if (tok == ROFF_de || tok == ROFF_dei)
1.86 schwarze 1684: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2 schwarze 1685:
1.110 schwarze 1686: if (*cp == '\0')
1.149 schwarze 1687: return ROFF_IGN;
1.1 schwarze 1688:
1.93 schwarze 1689: /* Get the custom end marker. */
1.22 schwarze 1690:
1.93 schwarze 1691: iname = cp;
1.86 schwarze 1692: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 1693:
1694: /* Resolve the end marker if it is indirect. */
1695:
1.110 schwarze 1696: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1697: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1698: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1699: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1700: "%.*s", (int)namesz, iname);
1701: namesz = 0;
1702: } else
1703: namesz = strlen(name);
1704: } else
1705: name = iname;
1706:
1.86 schwarze 1707: if (namesz)
1708: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 1709:
1.110 schwarze 1710: if (*cp != '\0')
1.90 schwarze 1711: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167 schwarze 1712: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1 schwarze 1713:
1.149 schwarze 1714: return ROFF_IGN;
1.1 schwarze 1715: }
1716:
1717: static enum rofferr
1.2 schwarze 1718: roff_block_sub(ROFF_ARGS)
1.1 schwarze 1719: {
1.166 schwarze 1720: enum roff_tok t;
1.2 schwarze 1721: int i, j;
1722:
1723: /*
1724: * First check whether a custom macro exists at this level. If
1725: * it does, then check against it. This is some of groff's
1726: * stranger behaviours. If we encountered a custom end-scope
1727: * tag and that tag also happens to be a "real" macro, then we
1728: * need to try interpreting it again as a real macro. If it's
1729: * not, then return ignore. Else continue.
1730: */
1731:
1732: if (r->last->end) {
1.35 schwarze 1733: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 1734: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 1735: break;
1.1 schwarze 1736:
1.110 schwarze 1737: if (r->last->end[j] == '\0' &&
1738: (buf->buf[i] == '\0' ||
1739: buf->buf[i] == ' ' ||
1740: buf->buf[i] == '\t')) {
1.2 schwarze 1741: roffnode_pop(r);
1742: roffnode_cleanscope(r);
1.1 schwarze 1743:
1.110 schwarze 1744: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 1745: i++;
1746:
1747: pos = i;
1.110 schwarze 1748: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166 schwarze 1749: TOKEN_NONE)
1.149 schwarze 1750: return ROFF_RERUN;
1751: return ROFF_IGN;
1.2 schwarze 1752: }
1.1 schwarze 1753: }
1754:
1.2 schwarze 1755: /*
1756: * If we have no custom end-query or lookup failed, then try
1757: * pulling it out of the hashtable.
1758: */
1.1 schwarze 1759:
1.110 schwarze 1760: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 1761:
1.110 schwarze 1762: if (t != ROFF_cblock) {
1763: if (tok != ROFF_ig)
1764: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149 schwarze 1765: return ROFF_IGN;
1.16 schwarze 1766: }
1.1 schwarze 1767:
1.149 schwarze 1768: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1769: }
1770:
1771: static enum rofferr
1772: roff_block_text(ROFF_ARGS)
1773: {
1774:
1.110 schwarze 1775: if (tok != ROFF_ig)
1776: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 1777:
1.149 schwarze 1778: return ROFF_IGN;
1.2 schwarze 1779: }
1780:
1781: static enum rofferr
1782: roff_cond_sub(ROFF_ARGS)
1783: {
1.166 schwarze 1784: enum roff_tok t;
1.37 schwarze 1785: char *ep;
1.71 schwarze 1786: int rr;
1.2 schwarze 1787:
1788: rr = r->last->rule;
1.37 schwarze 1789: roffnode_cleanscope(r);
1.110 schwarze 1790: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2 schwarze 1791:
1.37 schwarze 1792: /*
1.50 schwarze 1793: * Fully handle known macros when they are structurally
1794: * required or when the conditional evaluated to true.
1.5 schwarze 1795: */
1796:
1.166 schwarze 1797: if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.149 schwarze 1798: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.39 schwarze 1799:
1.69 schwarze 1800: /*
1801: * If `\}' occurs on a macro line without a preceding macro,
1802: * drop the line completely.
1803: */
1804:
1.110 schwarze 1805: ep = buf->buf + pos;
1806: if (ep[0] == '\\' && ep[1] == '}')
1.71 schwarze 1807: rr = 0;
1.69 schwarze 1808:
1.50 schwarze 1809: /* Always check for the closing delimiter `\}'. */
1.39 schwarze 1810:
1.110 schwarze 1811: while ((ep = strchr(ep, '\\')) != NULL) {
1812: if (*(++ep) == '}') {
1.70 schwarze 1813: *ep = '&';
1.110 schwarze 1814: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1815: }
1.119 schwarze 1816: if (*ep != '\0')
1817: ++ep;
1.50 schwarze 1818: }
1.149 schwarze 1819: return rr ? ROFF_CONT : ROFF_IGN;
1.2 schwarze 1820: }
1821:
1822: static enum rofferr
1823: roff_cond_text(ROFF_ARGS)
1824: {
1.37 schwarze 1825: char *ep;
1.71 schwarze 1826: int rr;
1.2 schwarze 1827:
1828: rr = r->last->rule;
1.37 schwarze 1829: roffnode_cleanscope(r);
1.1 schwarze 1830:
1.110 schwarze 1831: ep = buf->buf + pos;
1832: while ((ep = strchr(ep, '\\')) != NULL) {
1833: if (*(++ep) == '}') {
1.70 schwarze 1834: *ep = '&';
1.110 schwarze 1835: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1836: }
1.119 schwarze 1837: if (*ep != '\0')
1838: ++ep;
1.2 schwarze 1839: }
1.149 schwarze 1840: return rr ? ROFF_CONT : ROFF_IGN;
1.2 schwarze 1841: }
1842:
1.138 schwarze 1843: /* --- handling of numeric and conditional expressions -------------------- */
1844:
1.77 schwarze 1845: /*
1846: * Parse a single signed integer number. Stop at the first non-digit.
1847: * If there is at least one digit, return success and advance the
1848: * parse point, else return failure and let the parse point unchanged.
1849: * Ignore overflows, treat them just like the C language.
1850: */
1.56 schwarze 1851: static int
1.133 schwarze 1852: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56 schwarze 1853: {
1.133 schwarze 1854: int myres, scaled, n, p;
1.79 schwarze 1855:
1856: if (NULL == res)
1857: res = &myres;
1.56 schwarze 1858:
1859: p = *pos;
1860: n = v[p] == '-';
1.133 schwarze 1861: if (n || v[p] == '+')
1.56 schwarze 1862: p++;
1863:
1.133 schwarze 1864: if (flags & ROFFNUM_WHITE)
1865: while (isspace((unsigned char)v[p]))
1866: p++;
1867:
1.56 schwarze 1868: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 1869: *res = 10 * *res + v[p] - '0';
1.56 schwarze 1870: if (p == *pos + n)
1871: return 0;
1872:
1873: if (n)
1874: *res = -*res;
1875:
1.126 schwarze 1876: /* Each number may be followed by one optional scaling unit. */
1877:
1878: switch (v[p]) {
1879: case 'f':
1.133 schwarze 1880: scaled = *res * 65536;
1.126 schwarze 1881: break;
1882: case 'i':
1.133 schwarze 1883: scaled = *res * 240;
1.126 schwarze 1884: break;
1885: case 'c':
1.133 schwarze 1886: scaled = *res * 240 / 2.54;
1.126 schwarze 1887: break;
1888: case 'v':
1889: case 'P':
1.133 schwarze 1890: scaled = *res * 40;
1.126 schwarze 1891: break;
1892: case 'm':
1893: case 'n':
1.133 schwarze 1894: scaled = *res * 24;
1.126 schwarze 1895: break;
1896: case 'p':
1.133 schwarze 1897: scaled = *res * 10 / 3;
1.126 schwarze 1898: break;
1899: case 'u':
1.133 schwarze 1900: scaled = *res;
1.126 schwarze 1901: break;
1902: case 'M':
1.133 schwarze 1903: scaled = *res * 6 / 25;
1.126 schwarze 1904: break;
1905: default:
1.133 schwarze 1906: scaled = *res;
1.126 schwarze 1907: p--;
1908: break;
1909: }
1.133 schwarze 1910: if (flags & ROFFNUM_SCALE)
1911: *res = scaled;
1.126 schwarze 1912:
1913: *pos = p + 1;
1.149 schwarze 1914: return 1;
1.56 schwarze 1915: }
1916:
1.71 schwarze 1917: /*
1918: * Evaluate a string comparison condition.
1919: * The first character is the delimiter.
1920: * Succeed if the string up to its second occurrence
1921: * matches the string up to its third occurence.
1922: * Advance the cursor after the third occurrence
1923: * or lacking that, to the end of the line.
1924: */
1925: static int
1926: roff_evalstrcond(const char *v, int *pos)
1927: {
1928: const char *s1, *s2, *s3;
1929: int match;
1930:
1931: match = 0;
1932: s1 = v + *pos; /* initial delimiter */
1933: s2 = s1 + 1; /* for scanning the first string */
1934: s3 = strchr(s2, *s1); /* for scanning the second string */
1935:
1936: if (NULL == s3) /* found no middle delimiter */
1937: goto out;
1938:
1939: while ('\0' != *++s3) {
1940: if (*s2 != *s3) { /* mismatch */
1941: s3 = strchr(s3, *s1);
1942: break;
1943: }
1944: if (*s3 == *s1) { /* found the final delimiter */
1945: match = 1;
1946: break;
1947: }
1948: s2++;
1949: }
1950:
1951: out:
1952: if (NULL == s3)
1953: s3 = strchr(s2, '\0');
1.114 schwarze 1954: else if (*s3 != '\0')
1.71 schwarze 1955: s3++;
1956: *pos = s3 - v;
1.149 schwarze 1957: return match;
1.71 schwarze 1958: }
1959:
1.77 schwarze 1960: /*
1961: * Evaluate an optionally negated single character, numerical,
1962: * or string condition.
1963: */
1.71 schwarze 1964: static int
1.143 schwarze 1965: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5 schwarze 1966: {
1.143 schwarze 1967: char *cp, *name;
1968: size_t sz;
1.113 schwarze 1969: int number, savepos, wanttrue;
1.5 schwarze 1970:
1.71 schwarze 1971: if ('!' == v[*pos]) {
1972: wanttrue = 0;
1973: (*pos)++;
1974: } else
1975: wanttrue = 1;
1976:
1.5 schwarze 1977: switch (v[*pos]) {
1.112 schwarze 1978: case '\0':
1.149 schwarze 1979: return 0;
1.80 schwarze 1980: case 'n':
1981: case 'o':
1.5 schwarze 1982: (*pos)++;
1.149 schwarze 1983: return wanttrue;
1.80 schwarze 1984: case 'c':
1985: case 'd':
1986: case 'e':
1987: case 't':
1.111 schwarze 1988: case 'v':
1.5 schwarze 1989: (*pos)++;
1.149 schwarze 1990: return !wanttrue;
1.143 schwarze 1991: case 'r':
1992: cp = name = v + ++*pos;
1993: sz = roff_getname(r, &cp, ln, *pos);
1994: *pos = cp - v;
1.149 schwarze 1995: return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.5 schwarze 1996: default:
1997: break;
1998: }
1999:
1.113 schwarze 2000: savepos = *pos;
1.133 schwarze 2001: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149 schwarze 2002: return (number > 0) == wanttrue;
1.113 schwarze 2003: else if (*pos == savepos)
1.149 schwarze 2004: return roff_evalstrcond(v, pos) == wanttrue;
1.77 schwarze 2005: else
1.149 schwarze 2006: return 0;
1.5 schwarze 2007: }
2008:
1.2 schwarze 2009: static enum rofferr
1.21 schwarze 2010: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 2011: {
1.30 schwarze 2012:
1.149 schwarze 2013: return ROFF_IGN;
1.21 schwarze 2014: }
2015:
2016: static enum rofferr
1.123 schwarze 2017: roff_insec(ROFF_ARGS)
2018: {
2019:
2020: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167 schwarze 2021: ln, ppos, roff_name[tok]);
1.149 schwarze 2022: return ROFF_IGN;
1.123 schwarze 2023: }
2024:
2025: static enum rofferr
2026: roff_unsupp(ROFF_ARGS)
2027: {
2028:
2029: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167 schwarze 2030: ln, ppos, roff_name[tok]);
1.149 schwarze 2031: return ROFF_IGN;
1.123 schwarze 2032: }
2033:
2034: static enum rofferr
1.2 schwarze 2035: roff_cond(ROFF_ARGS)
2036: {
1.46 schwarze 2037:
2038: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 2039:
1.80 schwarze 2040: /*
1.35 schwarze 2041: * An `.el' has no conditional body: it will consume the value
2042: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 2043: * defaults to DENY.
1.35 schwarze 2044: *
2045: * If we're not an `el', however, then evaluate the conditional.
2046: */
1.1 schwarze 2047:
1.110 schwarze 2048: r->last->rule = tok == ROFF_el ?
1.80 schwarze 2049: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 2050: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 2051:
1.35 schwarze 2052: /*
2053: * An if-else will put the NEGATION of the current evaluated
2054: * conditional into the stack of rules.
2055: */
2056:
1.110 schwarze 2057: if (tok == ROFF_ie) {
1.96 schwarze 2058: if (r->rstackpos + 1 == r->rstacksz) {
2059: r->rstacksz += 16;
2060: r->rstack = mandoc_reallocarray(r->rstack,
2061: r->rstacksz, sizeof(int));
1.35 schwarze 2062: }
1.71 schwarze 2063: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 2064: }
1.5 schwarze 2065:
2066: /* If the parent has false as its rule, then so do we. */
2067:
1.71 schwarze 2068: if (r->last->parent && !r->last->parent->rule)
2069: r->last->rule = 0;
1.5 schwarze 2070:
2071: /*
1.46 schwarze 2072: * Determine scope.
2073: * If there is nothing on the line after the conditional,
2074: * not even whitespace, use next-line scope.
1.5 schwarze 2075: */
1.2 schwarze 2076:
1.110 schwarze 2077: if (buf->buf[pos] == '\0') {
1.46 schwarze 2078: r->last->endspan = 2;
2079: goto out;
2080: }
2081:
1.110 schwarze 2082: while (buf->buf[pos] == ' ')
1.46 schwarze 2083: pos++;
2084:
2085: /* An opening brace requests multiline scope. */
1.2 schwarze 2086:
1.110 schwarze 2087: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 2088: r->last->endspan = -1;
2089: pos += 2;
1.144 schwarze 2090: while (buf->buf[pos] == ' ')
2091: pos++;
1.46 schwarze 2092: goto out;
1.80 schwarze 2093: }
1.2 schwarze 2094:
2095: /*
1.46 schwarze 2096: * Anything else following the conditional causes
2097: * single-line scope. Warn if the scope contains
2098: * nothing but trailing whitespace.
1.2 schwarze 2099: */
2100:
1.110 schwarze 2101: if (buf->buf[pos] == '\0')
1.89 schwarze 2102: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167 schwarze 2103: ln, ppos, roff_name[tok]);
1.2 schwarze 2104:
1.46 schwarze 2105: r->last->endspan = 1;
1.1 schwarze 2106:
1.46 schwarze 2107: out:
1.2 schwarze 2108: *offs = pos;
1.149 schwarze 2109: return ROFF_RERUN;
1.1 schwarze 2110: }
2111:
1.2 schwarze 2112: static enum rofferr
1.7 schwarze 2113: roff_ds(ROFF_ARGS)
2114: {
1.85 schwarze 2115: char *string;
2116: const char *name;
2117: size_t namesz;
1.10 schwarze 2118:
1.123 schwarze 2119: /* Ignore groff compatibility mode for now. */
2120:
2121: if (tok == ROFF_ds1)
2122: tok = ROFF_ds;
2123: else if (tok == ROFF_as1)
2124: tok = ROFF_as;
2125:
1.10 schwarze 2126: /*
1.85 schwarze 2127: * The first word is the name of the string.
2128: * If it is empty or terminated by an escape sequence,
2129: * abort the `ds' request without defining anything.
1.10 schwarze 2130: */
1.7 schwarze 2131:
1.110 schwarze 2132: name = string = buf->buf + pos;
2133: if (*name == '\0')
1.149 schwarze 2134: return ROFF_IGN;
1.7 schwarze 2135:
1.85 schwarze 2136: namesz = roff_getname(r, &string, ln, pos);
1.110 schwarze 2137: if (name[namesz] == '\\')
1.149 schwarze 2138: return ROFF_IGN;
1.85 schwarze 2139:
2140: /* Read past the initial double-quote, if any. */
1.110 schwarze 2141: if (*string == '"')
1.7 schwarze 2142: string++;
2143:
1.10 schwarze 2144: /* The rest is the value. */
1.85 schwarze 2145: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2146: ROFF_as == tok);
1.149 schwarze 2147: return ROFF_IGN;
1.7 schwarze 2148: }
2149:
1.77 schwarze 2150: /*
2151: * Parse a single operator, one or two characters long.
2152: * If the operator is recognized, return success and advance the
2153: * parse point, else return failure and let the parse point unchanged.
2154: */
2155: static int
2156: roff_getop(const char *v, int *pos, char *res)
2157: {
2158:
2159: *res = v[*pos];
2160:
2161: switch (*res) {
1.80 schwarze 2162: case '+':
2163: case '-':
2164: case '*':
2165: case '/':
2166: case '%':
2167: case '&':
2168: case ':':
1.77 schwarze 2169: break;
2170: case '<':
2171: switch (v[*pos + 1]) {
1.80 schwarze 2172: case '=':
1.77 schwarze 2173: *res = 'l';
2174: (*pos)++;
2175: break;
1.80 schwarze 2176: case '>':
1.77 schwarze 2177: *res = '!';
2178: (*pos)++;
2179: break;
1.80 schwarze 2180: case '?':
1.77 schwarze 2181: *res = 'i';
2182: (*pos)++;
2183: break;
2184: default:
2185: break;
2186: }
2187: break;
2188: case '>':
2189: switch (v[*pos + 1]) {
1.80 schwarze 2190: case '=':
1.77 schwarze 2191: *res = 'g';
2192: (*pos)++;
2193: break;
1.80 schwarze 2194: case '?':
1.77 schwarze 2195: *res = 'a';
2196: (*pos)++;
2197: break;
2198: default:
2199: break;
2200: }
2201: break;
2202: case '=':
2203: if ('=' == v[*pos + 1])
2204: (*pos)++;
2205: break;
2206: default:
1.149 schwarze 2207: return 0;
1.77 schwarze 2208: }
2209: (*pos)++;
2210:
1.149 schwarze 2211: return *res;
1.77 schwarze 2212: }
2213:
2214: /*
2215: * Evaluate either a parenthesized numeric expression
2216: * or a single signed integer number.
2217: */
2218: static int
1.106 schwarze 2219: roff_evalpar(struct roff *r, int ln,
1.133 schwarze 2220: const char *v, int *pos, int *res, int flags)
1.77 schwarze 2221: {
2222:
2223: if ('(' != v[*pos])
1.149 schwarze 2224: return roff_getnum(v, pos, res, flags);
1.77 schwarze 2225:
2226: (*pos)++;
1.133 schwarze 2227: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149 schwarze 2228: return 0;
1.77 schwarze 2229:
1.79 schwarze 2230: /*
2231: * Omission of the closing parenthesis
2232: * is an error in validation mode,
2233: * but ignored in evaluation mode.
2234: */
2235:
1.77 schwarze 2236: if (')' == v[*pos])
2237: (*pos)++;
1.79 schwarze 2238: else if (NULL == res)
1.149 schwarze 2239: return 0;
1.77 schwarze 2240:
1.149 schwarze 2241: return 1;
1.77 schwarze 2242: }
2243:
2244: /*
2245: * Evaluate a complete numeric expression.
2246: * Proceed left to right, there is no concept of precedence.
2247: */
2248: static int
1.106 schwarze 2249: roff_evalnum(struct roff *r, int ln, const char *v,
1.133 schwarze 2250: int *pos, int *res, int flags)
1.77 schwarze 2251: {
2252: int mypos, operand2;
2253: char operator;
2254:
2255: if (NULL == pos) {
2256: mypos = 0;
2257: pos = &mypos;
2258: }
2259:
1.133 schwarze 2260: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2261: while (isspace((unsigned char)v[*pos]))
2262: (*pos)++;
2263:
1.133 schwarze 2264: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149 schwarze 2265: return 0;
1.77 schwarze 2266:
2267: while (1) {
1.133 schwarze 2268: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2269: while (isspace((unsigned char)v[*pos]))
2270: (*pos)++;
2271:
2272: if ( ! roff_getop(v, pos, &operator))
2273: break;
2274:
1.133 schwarze 2275: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2276: while (isspace((unsigned char)v[*pos]))
2277: (*pos)++;
2278:
1.133 schwarze 2279: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149 schwarze 2280: return 0;
1.77 schwarze 2281:
1.133 schwarze 2282: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2283: while (isspace((unsigned char)v[*pos]))
2284: (*pos)++;
1.79 schwarze 2285:
2286: if (NULL == res)
2287: continue;
1.77 schwarze 2288:
2289: switch (operator) {
1.80 schwarze 2290: case '+':
1.77 schwarze 2291: *res += operand2;
2292: break;
1.80 schwarze 2293: case '-':
1.77 schwarze 2294: *res -= operand2;
2295: break;
1.80 schwarze 2296: case '*':
1.77 schwarze 2297: *res *= operand2;
2298: break;
1.80 schwarze 2299: case '/':
1.116 schwarze 2300: if (operand2 == 0) {
1.106 schwarze 2301: mandoc_msg(MANDOCERR_DIVZERO,
2302: r->parse, ln, *pos, v);
2303: *res = 0;
2304: break;
2305: }
1.77 schwarze 2306: *res /= operand2;
2307: break;
1.80 schwarze 2308: case '%':
1.116 schwarze 2309: if (operand2 == 0) {
2310: mandoc_msg(MANDOCERR_DIVZERO,
2311: r->parse, ln, *pos, v);
2312: *res = 0;
2313: break;
2314: }
1.77 schwarze 2315: *res %= operand2;
2316: break;
1.80 schwarze 2317: case '<':
1.77 schwarze 2318: *res = *res < operand2;
2319: break;
1.80 schwarze 2320: case '>':
1.77 schwarze 2321: *res = *res > operand2;
2322: break;
1.80 schwarze 2323: case 'l':
1.77 schwarze 2324: *res = *res <= operand2;
2325: break;
1.80 schwarze 2326: case 'g':
1.77 schwarze 2327: *res = *res >= operand2;
2328: break;
1.80 schwarze 2329: case '=':
1.77 schwarze 2330: *res = *res == operand2;
2331: break;
1.80 schwarze 2332: case '!':
1.77 schwarze 2333: *res = *res != operand2;
2334: break;
1.80 schwarze 2335: case '&':
1.77 schwarze 2336: *res = *res && operand2;
2337: break;
1.80 schwarze 2338: case ':':
1.77 schwarze 2339: *res = *res || operand2;
2340: break;
1.80 schwarze 2341: case 'i':
1.77 schwarze 2342: if (operand2 < *res)
2343: *res = operand2;
2344: break;
1.80 schwarze 2345: case 'a':
1.77 schwarze 2346: if (operand2 > *res)
2347: *res = operand2;
2348: break;
2349: default:
2350: abort();
2351: }
2352: }
1.149 schwarze 2353: return 1;
1.77 schwarze 2354: }
2355:
1.138 schwarze 2356: /* --- register management ------------------------------------------------ */
2357:
1.52 schwarze 2358: void
1.60 schwarze 2359: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 2360: {
1.52 schwarze 2361: struct roffreg *reg;
2362:
2363: /* Search for an existing register with the same name. */
2364: reg = r->regtab;
2365:
2366: while (reg && strcmp(name, reg->key.p))
2367: reg = reg->next;
1.41 schwarze 2368:
1.52 schwarze 2369: if (NULL == reg) {
2370: /* Create a new register. */
2371: reg = mandoc_malloc(sizeof(struct roffreg));
2372: reg->key.p = mandoc_strdup(name);
2373: reg->key.sz = strlen(name);
1.60 schwarze 2374: reg->val = 0;
1.52 schwarze 2375: reg->next = r->regtab;
2376: r->regtab = reg;
2377: }
2378:
1.60 schwarze 2379: if ('+' == sign)
2380: reg->val += val;
2381: else if ('-' == sign)
2382: reg->val -= val;
2383: else
2384: reg->val = val;
1.41 schwarze 2385: }
2386:
1.65 schwarze 2387: /*
2388: * Handle some predefined read-only number registers.
2389: * For now, return -1 if the requested register is not predefined;
2390: * in case a predefined read-only register having the value -1
2391: * were to turn up, another special value would have to be chosen.
2392: */
2393: static int
1.145 schwarze 2394: roff_getregro(const struct roff *r, const char *name)
1.65 schwarze 2395: {
2396:
2397: switch (*name) {
1.145 schwarze 2398: case '$': /* Number of arguments of the last macro evaluated. */
1.149 schwarze 2399: return r->argc;
1.80 schwarze 2400: case 'A': /* ASCII approximation mode is always off. */
1.149 schwarze 2401: return 0;
1.80 schwarze 2402: case 'g': /* Groff compatibility mode is always on. */
1.149 schwarze 2403: return 1;
1.80 schwarze 2404: case 'H': /* Fixed horizontal resolution. */
1.149 schwarze 2405: return 24;
1.80 schwarze 2406: case 'j': /* Always adjust left margin only. */
1.149 schwarze 2407: return 0;
1.80 schwarze 2408: case 'T': /* Some output device is always defined. */
1.149 schwarze 2409: return 1;
1.80 schwarze 2410: case 'V': /* Fixed vertical resolution. */
1.149 schwarze 2411: return 40;
1.65 schwarze 2412: default:
1.149 schwarze 2413: return -1;
1.65 schwarze 2414: }
2415: }
2416:
1.53 schwarze 2417: int
1.52 schwarze 2418: roff_getreg(const struct roff *r, const char *name)
1.41 schwarze 2419: {
1.52 schwarze 2420: struct roffreg *reg;
1.65 schwarze 2421: int val;
2422:
2423: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145 schwarze 2424: val = roff_getregro(r, name + 1);
1.65 schwarze 2425: if (-1 != val)
1.149 schwarze 2426: return val;
1.65 schwarze 2427: }
1.52 schwarze 2428:
2429: for (reg = r->regtab; reg; reg = reg->next)
2430: if (0 == strcmp(name, reg->key.p))
1.149 schwarze 2431: return reg->val;
1.53 schwarze 2432:
1.149 schwarze 2433: return 0;
1.53 schwarze 2434: }
2435:
2436: static int
2437: roff_getregn(const struct roff *r, const char *name, size_t len)
2438: {
2439: struct roffreg *reg;
1.65 schwarze 2440: int val;
2441:
2442: if ('.' == name[0] && 2 == len) {
1.145 schwarze 2443: val = roff_getregro(r, name + 1);
1.65 schwarze 2444: if (-1 != val)
1.149 schwarze 2445: return val;
1.65 schwarze 2446: }
1.53 schwarze 2447:
2448: for (reg = r->regtab; reg; reg = reg->next)
2449: if (len == reg->key.sz &&
2450: 0 == strncmp(name, reg->key.p, len))
1.149 schwarze 2451: return reg->val;
1.143 schwarze 2452:
1.149 schwarze 2453: return 0;
1.143 schwarze 2454: }
2455:
2456: static int
2457: roff_hasregn(const struct roff *r, const char *name, size_t len)
2458: {
2459: struct roffreg *reg;
2460: int val;
2461:
2462: if ('.' == name[0] && 2 == len) {
1.145 schwarze 2463: val = roff_getregro(r, name + 1);
1.143 schwarze 2464: if (-1 != val)
1.149 schwarze 2465: return 1;
1.143 schwarze 2466: }
2467:
2468: for (reg = r->regtab; reg; reg = reg->next)
2469: if (len == reg->key.sz &&
2470: 0 == strncmp(name, reg->key.p, len))
1.149 schwarze 2471: return 1;
1.41 schwarze 2472:
1.149 schwarze 2473: return 0;
1.41 schwarze 2474: }
2475:
1.52 schwarze 2476: static void
2477: roff_freereg(struct roffreg *reg)
1.41 schwarze 2478: {
1.52 schwarze 2479: struct roffreg *old_reg;
1.41 schwarze 2480:
1.52 schwarze 2481: while (NULL != reg) {
2482: free(reg->key.p);
2483: old_reg = reg;
2484: reg = reg->next;
2485: free(old_reg);
2486: }
1.41 schwarze 2487: }
1.7 schwarze 2488:
2489: static enum rofferr
1.6 schwarze 2490: roff_nr(ROFF_ARGS)
1.1 schwarze 2491: {
1.85 schwarze 2492: char *key, *val;
2493: size_t keysz;
1.37 schwarze 2494: int iv;
1.60 schwarze 2495: char sign;
1.6 schwarze 2496:
1.110 schwarze 2497: key = val = buf->buf + pos;
2498: if (*key == '\0')
1.149 schwarze 2499: return ROFF_IGN;
1.85 schwarze 2500:
2501: keysz = roff_getname(r, &val, ln, pos);
1.110 schwarze 2502: if (key[keysz] == '\\')
1.149 schwarze 2503: return ROFF_IGN;
1.85 schwarze 2504: key[keysz] = '\0';
1.6 schwarze 2505:
1.60 schwarze 2506: sign = *val;
1.110 schwarze 2507: if (sign == '+' || sign == '-')
1.60 schwarze 2508: val++;
2509:
1.133 schwarze 2510: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77 schwarze 2511: roff_setreg(r, key, iv, sign);
1.1 schwarze 2512:
1.149 schwarze 2513: return ROFF_IGN;
1.76 schwarze 2514: }
2515:
2516: static enum rofferr
2517: roff_rr(ROFF_ARGS)
2518: {
2519: struct roffreg *reg, **prev;
1.85 schwarze 2520: char *name, *cp;
2521: size_t namesz;
1.76 schwarze 2522:
1.110 schwarze 2523: name = cp = buf->buf + pos;
2524: if (*name == '\0')
1.149 schwarze 2525: return ROFF_IGN;
1.85 schwarze 2526: namesz = roff_getname(r, &cp, ln, pos);
2527: name[namesz] = '\0';
1.76 schwarze 2528:
2529: prev = &r->regtab;
2530: while (1) {
2531: reg = *prev;
1.110 schwarze 2532: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 2533: break;
2534: prev = ®->next;
2535: }
1.110 schwarze 2536: if (reg != NULL) {
1.76 schwarze 2537: *prev = reg->next;
2538: free(reg->key.p);
2539: free(reg);
2540: }
1.149 schwarze 2541: return ROFF_IGN;
1.29 schwarze 2542: }
2543:
1.138 schwarze 2544: /* --- handler functions for roff requests -------------------------------- */
2545:
1.29 schwarze 2546: static enum rofferr
2547: roff_rm(ROFF_ARGS)
2548: {
2549: const char *name;
2550: char *cp;
1.85 schwarze 2551: size_t namesz;
1.29 schwarze 2552:
1.110 schwarze 2553: cp = buf->buf + pos;
2554: while (*cp != '\0') {
1.85 schwarze 2555: name = cp;
1.110 schwarze 2556: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 2557: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110 schwarze 2558: if (name[namesz] == '\\')
1.85 schwarze 2559: break;
1.29 schwarze 2560: }
1.149 schwarze 2561: return ROFF_IGN;
1.51 schwarze 2562: }
2563:
2564: static enum rofferr
2565: roff_it(ROFF_ARGS)
2566: {
2567: int iv;
2568:
2569: /* Parse the number of lines. */
1.133 schwarze 2570:
2571: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95 schwarze 2572: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110 schwarze 2573: ln, ppos, buf->buf + 1);
1.149 schwarze 2574: return ROFF_IGN;
1.51 schwarze 2575: }
2576:
1.134 schwarze 2577: while (isspace((unsigned char)buf->buf[pos]))
2578: pos++;
2579:
2580: /*
2581: * Arm the input line trap.
2582: * Special-casing "an-trap" is an ugly workaround to cope
2583: * with DocBook stupidly fiddling with man(7) internals.
2584: */
1.133 schwarze 2585:
1.51 schwarze 2586: roffit_lines = iv;
1.134 schwarze 2587: roffit_macro = mandoc_strdup(iv != 1 ||
2588: strcmp(buf->buf + pos, "an-trap") ?
2589: buf->buf + pos : "br");
1.149 schwarze 2590: return ROFF_IGN;
1.47 schwarze 2591: }
2592:
2593: static enum rofferr
2594: roff_Dd(ROFF_ARGS)
2595: {
2596: const char *const *cp;
2597:
1.99 schwarze 2598: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47 schwarze 2599: for (cp = __mdoc_reserved; *cp; cp++)
2600: roff_setstr(r, *cp, NULL, 0);
2601:
1.99 schwarze 2602: if (r->format == 0)
2603: r->format = MPARSE_MDOC;
2604:
1.149 schwarze 2605: return ROFF_CONT;
1.47 schwarze 2606: }
2607:
2608: static enum rofferr
2609: roff_TH(ROFF_ARGS)
2610: {
2611: const char *const *cp;
2612:
1.99 schwarze 2613: if ((r->options & MPARSE_QUICK) == 0)
1.47 schwarze 2614: for (cp = __man_reserved; *cp; cp++)
2615: roff_setstr(r, *cp, NULL, 0);
2616:
1.99 schwarze 2617: if (r->format == 0)
2618: r->format = MPARSE_MAN;
2619:
1.149 schwarze 2620: return ROFF_CONT;
1.14 schwarze 2621: }
2622:
2623: static enum rofferr
1.27 schwarze 2624: roff_TE(ROFF_ARGS)
2625: {
2626:
2627: if (NULL == r->tbl)
1.94 schwarze 2628: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2629: ln, ppos, "TE");
1.130 schwarze 2630: else if ( ! tbl_end(&r->tbl)) {
2631: free(buf->buf);
2632: buf->buf = mandoc_strdup(".sp");
2633: buf->sz = 4;
1.149 schwarze 2634: return ROFF_REPARSE;
1.130 schwarze 2635: }
1.149 schwarze 2636: return ROFF_IGN;
1.27 schwarze 2637: }
2638:
2639: static enum rofferr
2640: roff_T_(ROFF_ARGS)
2641: {
2642:
2643: if (NULL == r->tbl)
1.94 schwarze 2644: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2645: ln, ppos, "T&");
1.27 schwarze 2646: else
1.168 schwarze 2647: tbl_restart(ln, ppos, r->tbl);
1.27 schwarze 2648:
1.149 schwarze 2649: return ROFF_IGN;
1.27 schwarze 2650: }
2651:
1.102 schwarze 2652: /*
2653: * Handle in-line equation delimiters.
2654: */
2655: static enum rofferr
1.110 schwarze 2656: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 2657: {
1.105 schwarze 2658: char *cp1, *cp2;
2659: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 2660:
1.102 schwarze 2661: /*
2662: * Outside equations, look for an opening delimiter.
2663: * If we are inside an equation, we already know it is
2664: * in-line, or this function wouldn't have been called;
2665: * so look for a closing delimiter.
2666: */
2667:
1.110 schwarze 2668: cp1 = buf->buf + pos;
1.102 schwarze 2669: cp2 = strchr(cp1, r->eqn == NULL ?
2670: r->last_eqn->odelim : r->last_eqn->cdelim);
2671: if (cp2 == NULL)
1.149 schwarze 2672: return ROFF_CONT;
1.102 schwarze 2673:
1.105 schwarze 2674: *cp2++ = '\0';
2675: bef_pr = bef_nl = aft_nl = aft_pr = "";
2676:
2677: /* Handle preceding text, protecting whitespace. */
2678:
1.110 schwarze 2679: if (*buf->buf != '\0') {
1.105 schwarze 2680: if (r->eqn == NULL)
2681: bef_pr = "\\&";
2682: bef_nl = "\n";
2683: }
2684:
2685: /*
2686: * Prepare replacing the delimiter with an equation macro
2687: * and drop leading white space from the equation.
2688: */
1.102 schwarze 2689:
1.105 schwarze 2690: if (r->eqn == NULL) {
2691: while (*cp2 == ' ')
2692: cp2++;
2693: mac = ".EQ";
2694: } else
2695: mac = ".EN";
2696:
2697: /* Handle following text, protecting whitespace. */
2698:
2699: if (*cp2 != '\0') {
2700: aft_nl = "\n";
2701: if (r->eqn != NULL)
2702: aft_pr = "\\&";
2703: }
2704:
2705: /* Do the actual replacement. */
2706:
1.110 schwarze 2707: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 2708: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 2709: free(buf->buf);
2710: buf->buf = cp1;
1.102 schwarze 2711:
2712: /* Toggle the in-line state of the eqn subsystem. */
2713:
2714: r->eqn_inline = r->eqn == NULL;
1.149 schwarze 2715: return ROFF_REPARSE;
1.41 schwarze 2716: }
2717:
1.107 schwarze 2718: static enum rofferr
2719: roff_EQ(ROFF_ARGS)
1.32 schwarze 2720: {
1.41 schwarze 2721: struct eqn_node *e;
1.32 schwarze 2722:
1.110 schwarze 2723: assert(r->eqn == NULL);
1.107 schwarze 2724: e = eqn_alloc(ppos, ln, r->parse);
1.32 schwarze 2725:
1.102 schwarze 2726: if (r->last_eqn) {
1.32 schwarze 2727: r->last_eqn->next = e;
1.102 schwarze 2728: e->delim = r->last_eqn->delim;
2729: e->odelim = r->last_eqn->odelim;
2730: e->cdelim = r->last_eqn->cdelim;
2731: } else
1.32 schwarze 2732: r->first_eqn = r->last_eqn = e;
2733:
2734: r->eqn = r->last_eqn = e;
1.41 schwarze 2735:
1.110 schwarze 2736: if (buf->buf[pos] != '\0')
1.107 schwarze 2737: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 2738: ".EQ %s", buf->buf + pos);
1.41 schwarze 2739:
1.149 schwarze 2740: return ROFF_IGN;
1.32 schwarze 2741: }
2742:
2743: static enum rofferr
2744: roff_EN(ROFF_ARGS)
2745: {
2746:
1.94 schwarze 2747: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.149 schwarze 2748: return ROFF_IGN;
1.32 schwarze 2749: }
2750:
2751: static enum rofferr
1.27 schwarze 2752: roff_TS(ROFF_ARGS)
2753: {
1.49 schwarze 2754: struct tbl_node *tbl;
1.27 schwarze 2755:
2756: if (r->tbl) {
1.94 schwarze 2757: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2758: ln, ppos, "TS breaks TS");
1.41 schwarze 2759: tbl_end(&r->tbl);
1.27 schwarze 2760: }
2761:
1.49 schwarze 2762: tbl = tbl_alloc(ppos, ln, r->parse);
1.27 schwarze 2763:
2764: if (r->last_tbl)
1.49 schwarze 2765: r->last_tbl->next = tbl;
1.27 schwarze 2766: else
1.49 schwarze 2767: r->first_tbl = r->last_tbl = tbl;
1.27 schwarze 2768:
1.49 schwarze 2769: r->tbl = r->last_tbl = tbl;
1.169 schwarze 2770: return ROFF_IGN;
2771: }
2772:
2773: static enum rofferr
2774: roff_onearg(ROFF_ARGS)
2775: {
2776: struct roff_node *n;
2777: char *cp;
2778:
1.174 ! schwarze 2779: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
! 2780: (tok == ROFF_sp || tok == ROFF_ti))
! 2781: man_breakscope(r->man, tok);
! 2782:
1.169 schwarze 2783: roff_elem_alloc(r->man, ln, ppos, tok);
2784: n = r->man->last;
2785:
2786: cp = buf->buf + pos;
2787: if (*cp != '\0') {
2788: while (*cp != '\0' && *cp != ' ')
2789: cp++;
2790: while (*cp == ' ')
2791: *cp++ = '\0';
2792: if (*cp != '\0')
2793: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
2794: r->parse, ln, cp - buf->buf,
2795: "%s ... %s", roff_name[tok], cp);
2796: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172 schwarze 2797: }
2798:
2799: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
2800: r->man->last = n;
2801: r->man->next = ROFF_NEXT_SIBLING;
2802: return ROFF_IGN;
2803: }
2804:
2805: static enum rofferr
2806: roff_manyarg(ROFF_ARGS)
2807: {
2808: struct roff_node *n;
2809: char *sp, *ep;
2810:
2811: roff_elem_alloc(r->man, ln, ppos, tok);
2812: n = r->man->last;
2813:
2814: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
2815: while (*ep != '\0' && *ep != ' ')
2816: ep++;
2817: while (*ep == ' ')
2818: *ep++ = '\0';
2819: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169 schwarze 2820: }
2821:
2822: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
2823: r->man->last = n;
2824: r->man->next = ROFF_NEXT_SIBLING;
1.149 schwarze 2825: return ROFF_IGN;
1.123 schwarze 2826: }
2827:
2828: static enum rofferr
1.168 schwarze 2829: roff_br(ROFF_ARGS)
1.123 schwarze 2830: {
1.174 ! schwarze 2831: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
! 2832: man_breakscope(r->man, ROFF_br);
1.168 schwarze 2833: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
2834: if (buf->buf[pos] != '\0')
2835: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
2836: "%s %s", roff_name[tok], buf->buf + pos);
2837: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
2838: r->man->next = ROFF_NEXT_SIBLING;
2839: return ROFF_IGN;
1.27 schwarze 2840: }
2841:
2842: static enum rofferr
1.48 schwarze 2843: roff_cc(ROFF_ARGS)
2844: {
2845: const char *p;
2846:
1.110 schwarze 2847: p = buf->buf + pos;
1.48 schwarze 2848:
1.110 schwarze 2849: if (*p == '\0' || (r->control = *p++) == '.')
1.48 schwarze 2850: r->control = 0;
2851:
1.110 schwarze 2852: if (*p != '\0')
1.132 schwarze 2853: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2854: ln, p - buf->buf, "cc ... %s", p);
1.48 schwarze 2855:
1.149 schwarze 2856: return ROFF_IGN;
1.48 schwarze 2857: }
2858:
2859: static enum rofferr
1.42 schwarze 2860: roff_tr(ROFF_ARGS)
2861: {
2862: const char *p, *first, *second;
2863: size_t fsz, ssz;
2864: enum mandoc_esc esc;
2865:
1.110 schwarze 2866: p = buf->buf + pos;
1.42 schwarze 2867:
1.110 schwarze 2868: if (*p == '\0') {
1.132 schwarze 2869: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149 schwarze 2870: return ROFF_IGN;
1.42 schwarze 2871: }
2872:
1.110 schwarze 2873: while (*p != '\0') {
1.42 schwarze 2874: fsz = ssz = 1;
2875:
2876: first = p++;
1.110 schwarze 2877: if (*first == '\\') {
1.42 schwarze 2878: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2879: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2880: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2881: ln, (int)(p - buf->buf), first);
1.149 schwarze 2882: return ROFF_IGN;
1.42 schwarze 2883: }
2884: fsz = (size_t)(p - first);
2885: }
2886:
2887: second = p++;
1.110 schwarze 2888: if (*second == '\\') {
1.42 schwarze 2889: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2890: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2891: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2892: ln, (int)(p - buf->buf), second);
1.149 schwarze 2893: return ROFF_IGN;
1.42 schwarze 2894: }
2895: ssz = (size_t)(p - second);
1.110 schwarze 2896: } else if (*second == '\0') {
1.132 schwarze 2897: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2898: ln, first - buf->buf, "tr %s", first);
1.42 schwarze 2899: second = " ";
2900: p--;
2901: }
2902:
2903: if (fsz > 1) {
1.80 schwarze 2904: roff_setstrn(&r->xmbtab, first, fsz,
2905: second, ssz, 0);
1.42 schwarze 2906: continue;
2907: }
2908:
1.110 schwarze 2909: if (r->xtab == NULL)
1.80 schwarze 2910: r->xtab = mandoc_calloc(128,
2911: sizeof(struct roffstr));
1.42 schwarze 2912:
2913: free(r->xtab[(int)*first].p);
2914: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2915: r->xtab[(int)*first].sz = ssz;
2916: }
2917:
1.149 schwarze 2918: return ROFF_IGN;
1.42 schwarze 2919: }
2920:
2921: static enum rofferr
1.14 schwarze 2922: roff_so(ROFF_ARGS)
2923: {
1.121 schwarze 2924: char *name, *cp;
1.15 schwarze 2925:
1.110 schwarze 2926: name = buf->buf + pos;
1.97 schwarze 2927: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14 schwarze 2928:
1.22 schwarze 2929: /*
2930: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2931: * opening anything that's not in our cwd or anything beneath
2932: * it. Thus, explicitly disallow traversing up the file-system
2933: * or using absolute paths.
2934: */
2935:
1.110 schwarze 2936: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83 schwarze 2937: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2938: ".so %s", name);
1.121 schwarze 2939: buf->sz = mandoc_asprintf(&cp,
2940: ".sp\nSee the file %s.\n.sp", name) + 1;
2941: free(buf->buf);
2942: buf->buf = cp;
2943: *offs = 0;
1.149 schwarze 2944: return ROFF_REPARSE;
1.14 schwarze 2945: }
2946:
2947: *offs = pos;
1.149 schwarze 2948: return ROFF_SO;
1.7 schwarze 2949: }
2950:
1.138 schwarze 2951: /* --- user defined strings and macros ------------------------------------ */
2952:
1.16 schwarze 2953: static enum rofferr
2954: roff_userdef(ROFF_ARGS)
1.12 schwarze 2955: {
1.135 schwarze 2956: const char *arg[9], *ap;
1.16 schwarze 2957: char *cp, *n1, *n2;
1.164 schwarze 2958: int expand_count, i, ib, ie;
1.135 schwarze 2959: size_t asz, rsz;
1.12 schwarze 2960:
1.16 schwarze 2961: /*
2962: * Collect pointers to macro argument strings
1.61 schwarze 2963: * and NUL-terminate them.
1.16 schwarze 2964: */
1.135 schwarze 2965:
1.145 schwarze 2966: r->argc = 0;
1.110 schwarze 2967: cp = buf->buf + pos;
1.145 schwarze 2968: for (i = 0; i < 9; i++) {
2969: if (*cp == '\0')
2970: arg[i] = "";
2971: else {
2972: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
2973: r->argc = i + 1;
2974: }
2975: }
1.16 schwarze 2976:
2977: /*
2978: * Expand macro arguments.
1.12 schwarze 2979: */
1.135 schwarze 2980:
2981: buf->sz = strlen(r->current_string) + 1;
1.164 schwarze 2982: n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135 schwarze 2983: memcpy(n1, r->current_string, buf->sz);
1.164 schwarze 2984: expand_count = 0;
1.135 schwarze 2985: while (*cp != '\0') {
2986:
2987: /* Scan ahead for the next argument invocation. */
2988:
2989: if (*cp++ != '\\')
2990: continue;
2991: if (*cp++ != '$')
2992: continue;
1.146 schwarze 2993: if (*cp == '*') { /* \\$* inserts all arguments */
2994: ib = 0;
2995: ie = r->argc - 1;
2996: } else { /* \\$1 .. \\$9 insert one argument */
2997: ib = ie = *cp - '1';
2998: if (ib < 0 || ib > 8)
2999: continue;
3000: }
1.135 schwarze 3001: cp -= 2;
1.164 schwarze 3002:
3003: /*
3004: * Prevent infinite recursion.
3005: */
3006:
3007: if (cp >= n2)
3008: expand_count = 1;
3009: else if (++expand_count > EXPAND_LIMIT) {
3010: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
3011: ln, (int)(cp - n1), NULL);
1.165 schwarze 3012: free(buf->buf);
3013: buf->buf = n1;
1.164 schwarze 3014: return ROFF_IGN;
3015: }
1.135 schwarze 3016:
3017: /*
3018: * Determine the size of the expanded argument,
3019: * taking escaping of quotes into account.
3020: */
3021:
1.146 schwarze 3022: asz = ie > ib ? ie - ib : 0; /* for blanks */
3023: for (i = ib; i <= ie; i++) {
3024: for (ap = arg[i]; *ap != '\0'; ap++) {
3025: asz++;
3026: if (*ap == '"')
3027: asz += 3;
3028: }
1.135 schwarze 3029: }
3030: if (asz != 3) {
3031:
3032: /*
3033: * Determine the size of the rest of the
3034: * unexpanded macro, including the NUL.
3035: */
3036:
3037: rsz = buf->sz - (cp - n1) - 3;
3038:
3039: /*
3040: * When shrinking, move before
3041: * releasing the storage.
3042: */
3043:
3044: if (asz < 3)
3045: memmove(cp + asz, cp + 3, rsz);
3046:
3047: /*
3048: * Resize the storage for the macro
3049: * and readjust the parse pointer.
3050: */
3051:
3052: buf->sz += asz - 3;
3053: n2 = mandoc_realloc(n1, buf->sz);
3054: cp = n2 + (cp - n1);
3055: n1 = n2;
3056:
3057: /*
3058: * When growing, make room
3059: * for the expanded argument.
3060: */
3061:
3062: if (asz > 3)
3063: memmove(cp + asz, cp + 3, rsz);
3064: }
3065:
3066: /* Copy the expanded argument, escaping quotes. */
3067:
3068: n2 = cp;
1.146 schwarze 3069: for (i = ib; i <= ie; i++) {
3070: for (ap = arg[i]; *ap != '\0'; ap++) {
3071: if (*ap == '"') {
3072: memcpy(n2, "\\(dq", 4);
3073: n2 += 4;
3074: } else
3075: *n2++ = *ap;
3076: }
3077: if (i < ie)
3078: *n2++ = ' ';
1.16 schwarze 3079: }
1.12 schwarze 3080: }
3081:
1.16 schwarze 3082: /*
3083: * Replace the macro invocation
3084: * by the expanded macro.
3085: */
1.135 schwarze 3086:
1.110 schwarze 3087: free(buf->buf);
3088: buf->buf = n1;
1.120 schwarze 3089: *offs = 0;
1.16 schwarze 3090:
1.149 schwarze 3091: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3092: ROFF_REPARSE : ROFF_APPEND;
1.12 schwarze 3093: }
1.28 schwarze 3094:
1.85 schwarze 3095: static size_t
1.28 schwarze 3096: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3097: {
3098: char *name, *cp;
1.85 schwarze 3099: size_t namesz;
1.28 schwarze 3100:
3101: name = *cpp;
3102: if ('\0' == *name)
1.149 schwarze 3103: return 0;
1.28 schwarze 3104:
1.85 schwarze 3105: /* Read until end of name and terminate it with NUL. */
3106: for (cp = name; 1; cp++) {
3107: if ('\0' == *cp || ' ' == *cp) {
3108: namesz = cp - name;
3109: break;
3110: }
1.28 schwarze 3111: if ('\\' != *cp)
3112: continue;
1.88 schwarze 3113: namesz = cp - name;
3114: if ('{' == cp[1] || '}' == cp[1])
3115: break;
1.28 schwarze 3116: cp++;
3117: if ('\\' == *cp)
3118: continue;
1.97 schwarze 3119: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3120: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 3121: mandoc_escape((const char **)&cp, NULL, NULL);
3122: break;
1.28 schwarze 3123: }
3124:
3125: /* Read past spaces. */
3126: while (' ' == *cp)
3127: cp++;
3128:
3129: *cpp = cp;
1.149 schwarze 3130: return namesz;
1.28 schwarze 3131: }
3132:
1.16 schwarze 3133: /*
3134: * Store *string into the user-defined string called *name.
3135: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 3136: * append == 0: replace mode
3137: * append == 1: single-line append mode
3138: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 3139: */
1.8 schwarze 3140: static void
1.16 schwarze 3141: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 3142: int append)
1.7 schwarze 3143: {
1.42 schwarze 3144:
3145: roff_setstrn(&r->strtab, name, strlen(name), string,
1.80 schwarze 3146: string ? strlen(string) : 0, append);
1.42 schwarze 3147: }
3148:
3149: static void
3150: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 3151: const char *string, size_t stringsz, int append)
1.42 schwarze 3152: {
3153: struct roffkv *n;
3154: char *c;
3155: int i;
3156: size_t oldch, newch;
1.7 schwarze 3157:
1.16 schwarze 3158: /* Search for an existing string with the same name. */
1.42 schwarze 3159: n = *r;
3160:
1.84 schwarze 3161: while (n && (namesz != n->key.sz ||
3162: strncmp(n->key.p, name, namesz)))
1.7 schwarze 3163: n = n->next;
1.8 schwarze 3164:
3165: if (NULL == n) {
1.16 schwarze 3166: /* Create a new string table entry. */
1.42 schwarze 3167: n = mandoc_malloc(sizeof(struct roffkv));
3168: n->key.p = mandoc_strndup(name, namesz);
3169: n->key.sz = namesz;
3170: n->val.p = NULL;
3171: n->val.sz = 0;
3172: n->next = *r;
3173: *r = n;
1.66 schwarze 3174: } else if (0 == append) {
1.42 schwarze 3175: free(n->val.p);
3176: n->val.p = NULL;
3177: n->val.sz = 0;
1.16 schwarze 3178: }
3179:
3180: if (NULL == string)
3181: return;
3182:
3183: /*
3184: * One additional byte for the '\n' in multiline mode,
3185: * and one for the terminating '\0'.
3186: */
1.66 schwarze 3187: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 3188:
3189: if (NULL == n->val.p) {
3190: n->val.p = mandoc_malloc(newch);
3191: *n->val.p = '\0';
1.16 schwarze 3192: oldch = 0;
3193: } else {
1.42 schwarze 3194: oldch = n->val.sz;
3195: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 3196: }
3197:
3198: /* Skip existing content in the destination buffer. */
1.42 schwarze 3199: c = n->val.p + (int)oldch;
1.16 schwarze 3200:
3201: /* Append new content to the destination buffer. */
1.42 schwarze 3202: i = 0;
3203: while (i < (int)stringsz) {
1.16 schwarze 3204: /*
3205: * Rudimentary roff copy mode:
3206: * Handle escaped backslashes.
3207: */
1.42 schwarze 3208: if ('\\' == string[i] && '\\' == string[i + 1])
3209: i++;
3210: *c++ = string[i++];
1.16 schwarze 3211: }
1.8 schwarze 3212:
1.16 schwarze 3213: /* Append terminating bytes. */
1.66 schwarze 3214: if (1 < append)
1.16 schwarze 3215: *c++ = '\n';
1.42 schwarze 3216:
1.16 schwarze 3217: *c = '\0';
1.42 schwarze 3218: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 3219: }
3220:
1.8 schwarze 3221: static const char *
3222: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7 schwarze 3223: {
1.42 schwarze 3224: const struct roffkv *n;
1.64 schwarze 3225: int i;
1.7 schwarze 3226:
1.42 schwarze 3227: for (n = r->strtab; n; n = n->next)
1.80 schwarze 3228: if (0 == strncmp(name, n->key.p, len) &&
3229: '\0' == n->key.p[(int)len])
1.149 schwarze 3230: return n->val.p;
1.64 schwarze 3231:
3232: for (i = 0; i < PREDEFS_MAX; i++)
3233: if (0 == strncmp(name, predefs[i].name, len) &&
3234: '\0' == predefs[i].name[(int)len])
1.149 schwarze 3235: return predefs[i].str;
1.8 schwarze 3236:
1.149 schwarze 3237: return NULL;
1.7 schwarze 3238: }
3239:
1.8 schwarze 3240: static void
1.42 schwarze 3241: roff_freestr(struct roffkv *r)
1.7 schwarze 3242: {
1.42 schwarze 3243: struct roffkv *n, *nn;
1.7 schwarze 3244:
1.42 schwarze 3245: for (n = r; n; n = nn) {
3246: free(n->key.p);
3247: free(n->val.p);
1.7 schwarze 3248: nn = n->next;
3249: free(n);
3250: }
1.27 schwarze 3251: }
1.138 schwarze 3252:
3253: /* --- accessors and utility functions ------------------------------------ */
1.27 schwarze 3254:
3255: const struct tbl_span *
3256: roff_span(const struct roff *r)
3257: {
1.80 schwarze 3258:
1.149 schwarze 3259: return r->tbl ? tbl_span(r->tbl) : NULL;
1.32 schwarze 3260: }
3261:
3262: const struct eqn *
3263: roff_eqn(const struct roff *r)
3264: {
1.80 schwarze 3265:
1.149 schwarze 3266: return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.42 schwarze 3267: }
3268:
3269: /*
3270: * Duplicate an input string, making the appropriate character
3271: * conversations (as stipulated by `tr') along the way.
3272: * Returns a heap-allocated string with all the replacements made.
3273: */
3274: char *
3275: roff_strdup(const struct roff *r, const char *p)
3276: {
3277: const struct roffkv *cp;
3278: char *res;
3279: const char *pp;
3280: size_t ssz, sz;
3281: enum mandoc_esc esc;
3282:
3283: if (NULL == r->xmbtab && NULL == r->xtab)
1.149 schwarze 3284: return mandoc_strdup(p);
1.42 schwarze 3285: else if ('\0' == *p)
1.149 schwarze 3286: return mandoc_strdup("");
1.42 schwarze 3287:
3288: /*
3289: * Step through each character looking for term matches
3290: * (remember that a `tr' can be invoked with an escape, which is
3291: * a glyph but the escape is multi-character).
3292: * We only do this if the character hash has been initialised
3293: * and the string is >0 length.
3294: */
3295:
3296: res = NULL;
3297: ssz = 0;
3298:
3299: while ('\0' != *p) {
1.161 schwarze 3300: assert((unsigned int)*p < 128);
3301: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42 schwarze 3302: sz = r->xtab[(int)*p].sz;
3303: res = mandoc_realloc(res, ssz + sz + 1);
3304: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3305: ssz += sz;
3306: p++;
3307: continue;
3308: } else if ('\\' != *p) {
3309: res = mandoc_realloc(res, ssz + 2);
3310: res[ssz++] = *p++;
3311: continue;
3312: }
3313:
3314: /* Search for term matches. */
3315: for (cp = r->xmbtab; cp; cp = cp->next)
3316: if (0 == strncmp(p, cp->key.p, cp->key.sz))
3317: break;
3318:
3319: if (NULL != cp) {
3320: /*
3321: * A match has been found.
3322: * Append the match to the array and move
3323: * forward by its keysize.
3324: */
1.80 schwarze 3325: res = mandoc_realloc(res,
3326: ssz + cp->val.sz + 1);
1.42 schwarze 3327: memcpy(res + ssz, cp->val.p, cp->val.sz);
3328: ssz += cp->val.sz;
3329: p += (int)cp->key.sz;
3330: continue;
3331: }
3332:
3333: /*
3334: * Handle escapes carefully: we need to copy
3335: * over just the escape itself, or else we might
3336: * do replacements within the escape itself.
3337: * Make sure to pass along the bogus string.
3338: */
3339: pp = p++;
3340: esc = mandoc_escape(&p, NULL, NULL);
3341: if (ESCAPE_ERROR == esc) {
3342: sz = strlen(pp);
3343: res = mandoc_realloc(res, ssz + sz + 1);
3344: memcpy(res + ssz, pp, sz);
3345: break;
3346: }
1.80 schwarze 3347: /*
3348: * We bail out on bad escapes.
1.42 schwarze 3349: * No need to warn: we already did so when
3350: * roff_res() was called.
3351: */
3352: sz = (int)(p - pp);
3353: res = mandoc_realloc(res, ssz + sz + 1);
3354: memcpy(res + ssz, pp, sz);
3355: ssz += sz;
3356: }
3357:
3358: res[(int)ssz] = '\0';
1.149 schwarze 3359: return res;
1.99 schwarze 3360: }
3361:
3362: int
3363: roff_getformat(const struct roff *r)
3364: {
3365:
1.149 schwarze 3366: return r->format;
1.48 schwarze 3367: }
3368:
3369: /*
1.80 schwarze 3370: * Find out whether a line is a macro line or not.
1.48 schwarze 3371: * If it is, adjust the current position and return one; if it isn't,
3372: * return zero and don't change the current position.
3373: * If the control character has been set with `.cc', then let that grain
3374: * precedence.
3375: * This is slighly contrary to groff, where using the non-breaking
3376: * control character when `cc' has been invoked will cause the
3377: * non-breaking macro contents to be printed verbatim.
3378: */
3379: int
3380: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3381: {
3382: int pos;
3383:
3384: pos = *ppos;
3385:
3386: if (0 != r->control && cp[pos] == r->control)
3387: pos++;
3388: else if (0 != r->control)
1.149 schwarze 3389: return 0;
1.48 schwarze 3390: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3391: pos += 2;
3392: else if ('.' == cp[pos] || '\'' == cp[pos])
3393: pos++;
3394: else
1.149 schwarze 3395: return 0;
1.48 schwarze 3396:
3397: while (' ' == cp[pos] || '\t' == cp[pos])
3398: pos++;
3399:
3400: *ppos = pos;
1.149 schwarze 3401: return 1;
1.1 schwarze 3402: }
![[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