Annotation of src/usr.bin/mandoc/roff.c, Revision 1.151
1.151 ! schwarze 1: /* $OpenBSD: roff.c,v 1.150 2015/10/12 00:07:27 schwarze Exp $ */
1.1 schwarze 2: /*
1.139 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.119 schwarze 4: * Copyright (c) 2010-2015 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>
1.51 schwarze 23: #include <stdio.h>
1.1 schwarze 24: #include <stdlib.h>
25: #include <string.h>
26:
27: #include "mandoc.h"
1.74 schwarze 28: #include "mandoc_aux.h"
1.137 schwarze 29: #include "roff.h"
1.98 daniel 30: #include "libmandoc.h"
1.138 schwarze 31: #include "roff_int.h"
1.27 schwarze 32: #include "libroff.h"
1.1 schwarze 33:
1.37 schwarze 34: /* Maximum number of nested if-else conditionals. */
1.2 schwarze 35: #define RSTACK_MAX 128
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:
1.1 schwarze 42: enum rofft {
1.123 schwarze 43: ROFF_ab,
1.20 schwarze 44: ROFF_ad,
1.123 schwarze 45: ROFF_af,
46: ROFF_aln,
47: ROFF_als,
1.2 schwarze 48: ROFF_am,
1.123 schwarze 49: ROFF_am1,
1.2 schwarze 50: ROFF_ami,
1.123 schwarze 51: ROFF_ami1,
1.66 schwarze 52: ROFF_as,
1.123 schwarze 53: ROFF_as1,
54: ROFF_asciify,
55: ROFF_backtrace,
56: ROFF_bd,
57: ROFF_bleedat,
58: ROFF_blm,
59: ROFF_box,
60: ROFF_boxa,
61: ROFF_bp,
62: ROFF_BP,
63: /* MAN_br, MDOC_br */
64: ROFF_break,
65: ROFF_breakchar,
66: ROFF_brnl,
67: ROFF_brp,
68: ROFF_brpnl,
69: ROFF_c2,
1.48 schwarze 70: ROFF_cc,
1.67 schwarze 71: ROFF_ce,
1.123 schwarze 72: ROFF_cf,
73: ROFF_cflags,
74: ROFF_ch,
75: ROFF_char,
76: ROFF_chop,
77: ROFF_class,
78: ROFF_close,
79: ROFF_CL,
80: ROFF_color,
81: ROFF_composite,
82: ROFF_continue,
83: ROFF_cp,
84: ROFF_cropat,
85: ROFF_cs,
86: ROFF_cu,
87: ROFF_da,
88: ROFF_dch,
89: ROFF_Dd,
1.1 schwarze 90: ROFF_de,
1.123 schwarze 91: ROFF_de1,
92: ROFF_defcolor,
1.1 schwarze 93: ROFF_dei,
1.123 schwarze 94: ROFF_dei1,
95: ROFF_device,
96: ROFF_devicem,
97: ROFF_di,
98: ROFF_do,
1.2 schwarze 99: ROFF_ds,
1.123 schwarze 100: ROFF_ds1,
101: ROFF_dwh,
102: ROFF_dt,
103: ROFF_ec,
104: ROFF_ecr,
105: ROFF_ecs,
1.2 schwarze 106: ROFF_el,
1.123 schwarze 107: ROFF_em,
108: ROFF_EN,
109: ROFF_eo,
110: ROFF_EP,
111: ROFF_EQ,
112: ROFF_errprint,
113: ROFF_ev,
114: ROFF_evc,
115: ROFF_ex,
116: ROFF_fallback,
1.58 schwarze 117: ROFF_fam,
1.123 schwarze 118: ROFF_fc,
119: ROFF_fchar,
120: ROFF_fcolor,
121: ROFF_fdeferlig,
122: ROFF_feature,
123: /* MAN_fi; ignored in mdoc(7) */
124: ROFF_fkern,
125: ROFF_fl,
126: ROFF_flig,
127: ROFF_fp,
128: ROFF_fps,
129: ROFF_fschar,
130: ROFF_fspacewidth,
131: ROFF_fspecial,
132: /* MAN_ft; ignored in mdoc(7) */
133: ROFF_ftr,
134: ROFF_fzoom,
135: ROFF_gcolor,
136: ROFF_hc,
137: ROFF_hcode,
138: ROFF_hidechar,
139: ROFF_hla,
140: ROFF_hlm,
141: ROFF_hpf,
142: ROFF_hpfa,
143: ROFF_hpfcode,
1.59 schwarze 144: ROFF_hw,
1.20 schwarze 145: ROFF_hy,
1.123 schwarze 146: ROFF_hylang,
147: ROFF_hylen,
148: ROFF_hym,
149: ROFF_hypp,
150: ROFF_hys,
1.2 schwarze 151: ROFF_ie,
152: ROFF_if,
1.1 schwarze 153: ROFF_ig,
1.123 schwarze 154: /* MAN_in; ignored in mdoc(7) */
155: ROFF_index,
1.30 schwarze 156: ROFF_it,
1.123 schwarze 157: ROFF_itc,
158: ROFF_IX,
159: ROFF_kern,
160: ROFF_kernafter,
161: ROFF_kernbefore,
162: ROFF_kernpair,
163: ROFF_lc,
164: ROFF_lc_ctype,
165: ROFF_lds,
166: ROFF_length,
167: ROFF_letadj,
168: ROFF_lf,
169: ROFF_lg,
170: ROFF_lhang,
171: ROFF_linetabs,
172: /* MAN_ll, MDOC_ll */
173: ROFF_lnr,
174: ROFF_lnrf,
175: ROFF_lpfx,
176: ROFF_ls,
177: ROFF_lsm,
178: ROFF_lt,
179: ROFF_mc,
180: ROFF_mediasize,
181: ROFF_minss,
182: ROFF_mk,
183: ROFF_mso,
1.128 schwarze 184: ROFF_na,
1.20 schwarze 185: ROFF_ne,
1.123 schwarze 186: /* MAN_nf; ignored in mdoc(7) */
1.20 schwarze 187: ROFF_nh,
1.123 schwarze 188: ROFF_nhychar,
189: ROFF_nm,
190: ROFF_nn,
191: ROFF_nop,
1.14 schwarze 192: ROFF_nr,
1.123 schwarze 193: ROFF_nrf,
194: ROFF_nroff,
1.31 schwarze 195: ROFF_ns,
1.123 schwarze 196: ROFF_nx,
197: ROFF_open,
198: ROFF_opena,
199: ROFF_os,
200: ROFF_output,
201: ROFF_padj,
202: ROFF_papersize,
203: ROFF_pc,
204: ROFF_pev,
205: ROFF_pi,
206: ROFF_PI,
1.101 schwarze 207: ROFF_pl,
1.123 schwarze 208: ROFF_pm,
209: ROFF_pn,
210: ROFF_pnr,
211: ROFF_po,
1.31 schwarze 212: ROFF_ps,
1.123 schwarze 213: ROFF_psbb,
214: ROFF_pshape,
215: ROFF_pso,
216: ROFF_ptr,
217: ROFF_pvs,
218: ROFF_rchar,
219: ROFF_rd,
220: ROFF_recursionlimit,
221: ROFF_return,
222: ROFF_rfschar,
223: ROFF_rhang,
224: ROFF_rj,
1.2 schwarze 225: ROFF_rm,
1.123 schwarze 226: ROFF_rn,
227: ROFF_rnn,
1.76 schwarze 228: ROFF_rr,
1.123 schwarze 229: ROFF_rs,
230: ROFF_rt,
231: ROFF_schar,
232: ROFF_sentchar,
233: ROFF_shc,
234: ROFF_shift,
235: ROFF_sizes,
1.14 schwarze 236: ROFF_so,
1.123 schwarze 237: /* MAN_sp, MDOC_sp */
238: ROFF_spacewidth,
239: ROFF_special,
240: ROFF_spreadwarn,
241: ROFF_ss,
242: ROFF_sty,
243: ROFF_substring,
244: ROFF_sv,
245: ROFF_sy,
246: ROFF_T_,
1.31 schwarze 247: ROFF_ta,
1.123 schwarze 248: ROFF_tc,
249: ROFF_TE,
250: ROFF_TH,
251: ROFF_ti,
252: ROFF_tkf,
253: ROFF_tl,
254: ROFF_tm,
255: ROFF_tm1,
256: ROFF_tmc,
1.2 schwarze 257: ROFF_tr,
1.123 schwarze 258: ROFF_track,
259: ROFF_transchar,
260: ROFF_trf,
261: ROFF_trimat,
262: ROFF_trin,
263: ROFF_trnt,
264: ROFF_troff,
1.27 schwarze 265: ROFF_TS,
1.123 schwarze 266: ROFF_uf,
267: ROFF_ul,
268: ROFF_unformat,
269: ROFF_unwatch,
270: ROFF_unwatchn,
271: ROFF_vpt,
272: ROFF_vs,
273: ROFF_warn,
274: ROFF_warnscale,
275: ROFF_watch,
276: ROFF_watchlength,
277: ROFF_watchn,
278: ROFF_wh,
279: ROFF_while,
280: ROFF_write,
281: ROFF_writec,
282: ROFF_writem,
283: ROFF_xflag,
1.2 schwarze 284: ROFF_cblock,
1.16 schwarze 285: ROFF_USERDEF,
1.1 schwarze 286: ROFF_MAX
287: };
288:
1.41 schwarze 289: /*
1.42 schwarze 290: * An incredibly-simple string buffer.
291: */
1.8 schwarze 292: struct roffstr {
1.42 schwarze 293: char *p; /* nil-terminated buffer */
294: size_t sz; /* saved strlen(p) */
295: };
296:
297: /*
298: * A key-value roffstr pair as part of a singly-linked list.
299: */
300: struct roffkv {
301: struct roffstr key;
302: struct roffstr val;
303: struct roffkv *next; /* next in list */
1.8 schwarze 304: };
305:
1.52 schwarze 306: /*
307: * A single number register as part of a singly-linked list.
308: */
309: struct roffreg {
310: struct roffstr key;
1.53 schwarze 311: int val;
1.52 schwarze 312: struct roffreg *next;
313: };
314:
1.1 schwarze 315: struct roff {
1.35 schwarze 316: struct mparse *parse; /* parse point */
1.1 schwarze 317: struct roffnode *last; /* leaf of stack */
1.96 schwarze 318: int *rstack; /* stack of inverted `ie' values */
1.52 schwarze 319: struct roffreg *regtab; /* number registers */
1.42 schwarze 320: struct roffkv *strtab; /* user-defined strings & macros */
321: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
322: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.16 schwarze 323: const char *current_string; /* value of last called user macro */
1.27 schwarze 324: struct tbl_node *first_tbl; /* first table parsed */
325: struct tbl_node *last_tbl; /* last table parsed */
326: struct tbl_node *tbl; /* current table being parsed */
1.32 schwarze 327: struct eqn_node *last_eqn; /* last equation parsed */
328: struct eqn_node *first_eqn; /* first equation parsed */
329: struct eqn_node *eqn; /* current equation being parsed */
1.102 schwarze 330: int eqn_inline; /* current equation is inline */
1.96 schwarze 331: int options; /* parse options */
332: int rstacksz; /* current size limit of rstack */
333: int rstackpos; /* position in rstack */
1.99 schwarze 334: int format; /* current file in mdoc or man format */
1.145 schwarze 335: int argc; /* number of args of the last macro */
1.96 schwarze 336: char control; /* control character */
1.1 schwarze 337: };
338:
339: struct roffnode {
340: enum rofft tok; /* type of node */
341: struct roffnode *parent; /* up one in stack */
342: int line; /* parse line */
343: int col; /* parse col */
1.16 schwarze 344: char *name; /* node name, e.g. macro name */
1.2 schwarze 345: char *end; /* end-rules: custom token */
346: int endspan; /* end-rules: next-line or infty */
1.71 schwarze 347: int rule; /* current evaluation rule */
1.1 schwarze 348: };
349:
350: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
351: enum rofft tok, /* tok of macro */ \
1.110 schwarze 352: struct buf *buf, /* input buffer */ \
1.1 schwarze 353: int ln, /* parse line */ \
1.2 schwarze 354: int ppos, /* original pos in buffer */ \
355: int pos, /* current pos in buffer */ \
356: int *offs /* reset offset of buffer data */
1.1 schwarze 357:
358: typedef enum rofferr (*roffproc)(ROFF_ARGS);
359:
360: struct roffmac {
361: const char *name; /* macro name */
1.2 schwarze 362: roffproc proc; /* process new macro */
363: roffproc text; /* process as child text of macro */
364: roffproc sub; /* process as child of macro */
365: int flags;
366: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.3 schwarze 367: struct roffmac *next;
1.1 schwarze 368: };
369:
1.37 schwarze 370: struct predef {
371: const char *name; /* predefined input name */
372: const char *str; /* replacement symbol */
373: };
374:
375: #define PREDEF(__name, __str) \
376: { (__name), (__str) },
377:
1.138 schwarze 378: /* --- function prototypes ------------------------------------------------ */
379:
1.42 schwarze 380: static enum rofft roffhash_find(const char *, size_t);
381: static void roffhash_init(void);
382: static void roffnode_cleanscope(struct roff *);
383: static void roffnode_pop(struct roff *);
384: static void roffnode_push(struct roff *, enum rofft,
385: const char *, int, int);
1.2 schwarze 386: static enum rofferr roff_block(ROFF_ARGS);
387: static enum rofferr roff_block_text(ROFF_ARGS);
388: static enum rofferr roff_block_sub(ROFF_ARGS);
1.123 schwarze 389: static enum rofferr roff_brp(ROFF_ARGS);
1.2 schwarze 390: static enum rofferr roff_cblock(ROFF_ARGS);
1.48 schwarze 391: static enum rofferr roff_cc(ROFF_ARGS);
1.68 schwarze 392: static void roff_ccond(struct roff *, int, int);
1.2 schwarze 393: static enum rofferr roff_cond(ROFF_ARGS);
394: static enum rofferr roff_cond_text(ROFF_ARGS);
395: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.7 schwarze 396: static enum rofferr roff_ds(ROFF_ARGS);
1.110 schwarze 397: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.143 schwarze 398: static int roff_evalcond(struct roff *r, int, char *, int *);
1.106 schwarze 399: static int roff_evalnum(struct roff *, int,
400: const char *, int *, int *, int);
401: static int roff_evalpar(struct roff *, int,
1.133 schwarze 402: const char *, int *, int *, int);
1.71 schwarze 403: static int roff_evalstrcond(const char *, int *);
1.42 schwarze 404: static void roff_free1(struct roff *);
1.52 schwarze 405: static void roff_freereg(struct roffreg *);
1.42 schwarze 406: static void roff_freestr(struct roffkv *);
1.85 schwarze 407: static size_t roff_getname(struct roff *, char **, int, int);
1.133 schwarze 408: static int roff_getnum(const char *, int *, int *, int);
1.56 schwarze 409: static int roff_getop(const char *, int *, char *);
1.53 schwarze 410: static int roff_getregn(const struct roff *,
411: const char *, size_t);
1.145 schwarze 412: static int roff_getregro(const struct roff *,
413: const char *name);
1.80 schwarze 414: static const char *roff_getstrn(const struct roff *,
1.8 schwarze 415: const char *, size_t);
1.143 schwarze 416: static int roff_hasregn(const struct roff *,
417: const char *, size_t);
1.123 schwarze 418: static enum rofferr roff_insec(ROFF_ARGS);
1.51 schwarze 419: static enum rofferr roff_it(ROFF_ARGS);
1.21 schwarze 420: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.137 schwarze 421: static void roff_man_alloc1(struct roff_man *);
422: static void roff_man_free1(struct roff_man *);
1.6 schwarze 423: static enum rofferr roff_nr(ROFF_ARGS);
1.87 schwarze 424: static enum rofft roff_parse(struct roff *, char *, int *,
425: int, int);
1.110 schwarze 426: static enum rofferr roff_parsetext(struct buf *, int, int *);
427: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.29 schwarze 428: static enum rofferr roff_rm(ROFF_ARGS);
1.76 schwarze 429: static enum rofferr roff_rr(ROFF_ARGS);
1.8 schwarze 430: static void roff_setstr(struct roff *,
1.16 schwarze 431: const char *, const char *, int);
1.80 schwarze 432: static void roff_setstrn(struct roffkv **, const char *,
1.42 schwarze 433: size_t, const char *, size_t, int);
1.14 schwarze 434: static enum rofferr roff_so(ROFF_ARGS);
1.42 schwarze 435: static enum rofferr roff_tr(ROFF_ARGS);
1.47 schwarze 436: static enum rofferr roff_Dd(ROFF_ARGS);
437: static enum rofferr roff_TH(ROFF_ARGS);
1.27 schwarze 438: static enum rofferr roff_TE(ROFF_ARGS);
439: static enum rofferr roff_TS(ROFF_ARGS);
1.32 schwarze 440: static enum rofferr roff_EQ(ROFF_ARGS);
441: static enum rofferr roff_EN(ROFF_ARGS);
1.27 schwarze 442: static enum rofferr roff_T_(ROFF_ARGS);
1.123 schwarze 443: static enum rofferr roff_unsupp(ROFF_ARGS);
1.16 schwarze 444: static enum rofferr roff_userdef(ROFF_ARGS);
1.1 schwarze 445:
1.138 schwarze 446: /* --- constant data ------------------------------------------------------ */
447:
1.42 schwarze 448: /* See roffhash_find() */
1.3 schwarze 449:
450: #define ASCII_HI 126
451: #define ASCII_LO 33
452: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
453:
1.133 schwarze 454: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
455: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
456:
1.3 schwarze 457: static struct roffmac *hash[HASHWIDTH];
458:
459: static struct roffmac roffs[ROFF_MAX] = {
1.123 schwarze 460: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.21 schwarze 461: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 462: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
463: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
464: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 465: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 466: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.3 schwarze 467: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 468: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.66 schwarze 469: { "as", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 470: { "as1", roff_ds, NULL, NULL, 0, NULL },
471: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
472: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
473: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
474: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
475: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
476: { "box", roff_unsupp, NULL, NULL, 0, NULL },
477: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
478: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
479: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
480: { "break", roff_unsupp, NULL, NULL, 0, NULL },
481: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
482: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
483: { "brp", roff_brp, NULL, NULL, 0, NULL },
484: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
485: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.48 schwarze 486: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.67 schwarze 487: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 488: { "cf", roff_insec, NULL, NULL, 0, NULL },
489: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
490: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
491: { "char", roff_unsupp, NULL, NULL, 0, NULL },
492: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
493: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
494: { "close", roff_insec, NULL, NULL, 0, NULL },
495: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
496: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
497: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
498: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
499: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
500: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
501: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
502: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
503: { "da", roff_unsupp, NULL, NULL, 0, NULL },
504: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
505: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.3 schwarze 506: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 507: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
508: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 509: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 510: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
511: { "device", roff_unsupp, NULL, NULL, 0, NULL },
512: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
513: { "di", roff_unsupp, NULL, NULL, 0, NULL },
514: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.7 schwarze 515: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 516: { "ds1", roff_ds, NULL, NULL, 0, NULL },
517: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
518: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
519: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
520: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
521: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 522: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.123 schwarze 523: { "em", roff_unsupp, NULL, NULL, 0, NULL },
524: { "EN", roff_EN, NULL, NULL, 0, NULL },
525: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
526: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
527: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
528: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
529: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
530: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
531: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
532: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.58 schwarze 533: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 534: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
535: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
536: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
538: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
543: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
544: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
545: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
551: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
552: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
553: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
554: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
555: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
556: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
557: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.59 schwarze 558: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 559: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 560: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
561: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
562: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
563: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
564: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 565: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
566: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
567: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 568: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.51 schwarze 569: { "it", roff_it, NULL, NULL, 0, NULL },
1.123 schwarze 570: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
571: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
572: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
573: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
574: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
575: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
576: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
577: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
578: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
579: { "length", roff_unsupp, NULL, NULL, 0, NULL },
580: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
581: { "lf", roff_insec, NULL, NULL, 0, NULL },
582: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
583: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
584: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
585: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
586: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
587: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
588: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
589: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
590: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
591: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
592: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
593: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
594: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
595: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.128 schwarze 596: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 597: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
598: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 599: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
600: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
601: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
602: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.14 schwarze 603: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.123 schwarze 604: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
605: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 606: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 607: { "nx", roff_insec, NULL, NULL, 0, NULL },
608: { "open", roff_insec, NULL, NULL, 0, NULL },
609: { "opena", roff_insec, NULL, NULL, 0, NULL },
610: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
611: { "output", roff_unsupp, NULL, NULL, 0, NULL },
612: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
615: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
616: { "pi", roff_insec, NULL, NULL, 0, NULL },
617: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.101 schwarze 618: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 619: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
620: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
621: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
622: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 623: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 624: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.127 schwarze 625: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 626: { "pso", roff_insec, NULL, NULL, 0, NULL },
627: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
628: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
629: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
630: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
631: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
632: { "return", roff_unsupp, NULL, NULL, 0, NULL },
633: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
634: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
635: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.29 schwarze 636: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.123 schwarze 637: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
638: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.76 schwarze 639: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.123 schwarze 640: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
641: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
642: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
643: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
644: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
645: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
646: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.14 schwarze 647: { "so", roff_so, NULL, NULL, 0, NULL },
1.123 schwarze 648: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
649: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
650: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
651: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
652: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
653: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
654: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
655: { "sy", roff_insec, NULL, NULL, 0, NULL },
656: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.127 schwarze 657: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 658: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
659: { "TE", roff_TE, NULL, NULL, 0, NULL },
660: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.127 schwarze 661: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 662: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
663: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
664: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
665: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
666: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.42 schwarze 667: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.123 schwarze 668: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
670: { "trf", roff_insec, NULL, NULL, 0, NULL },
671: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
672: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
673: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
674: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.27 schwarze 675: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.123 schwarze 676: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
677: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
678: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
679: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
680: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
681: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
682: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
683: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
684: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
685: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
686: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
687: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
688: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
689: { "while", roff_unsupp, NULL, NULL, 0, NULL },
690: { "write", roff_insec, NULL, NULL, 0, NULL },
691: { "writec", roff_insec, NULL, NULL, 0, NULL },
692: { "writem", roff_insec, NULL, NULL, 0, NULL },
693: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 694: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.16 schwarze 695: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1 schwarze 696: };
697:
1.73 schwarze 698: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47 schwarze 699: const char *const __mdoc_reserved[] = {
700: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
701: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
702: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
703: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73 schwarze 704: "Dt", "Dv", "Dx", "D1",
705: "Ec", "Ed", "Ef", "Ek", "El", "Em",
706: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47 schwarze 707: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73 schwarze 708: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
709: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47 schwarze 710: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73 schwarze 711: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
712: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
713: "Sc", "Sh", "Sm", "So", "Sq",
1.47 schwarze 714: "Ss", "St", "Sx", "Sy",
715: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73 schwarze 716: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47 schwarze 717: "%P", "%Q", "%R", "%T", "%U", "%V",
718: NULL
719: };
720:
1.73 schwarze 721: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47 schwarze 722: const char *const __man_reserved[] = {
1.73 schwarze 723: "AT", "B", "BI", "BR", "DT",
724: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
725: "LP", "OP", "P", "PD", "PP",
726: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
727: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47 schwarze 728: NULL
729: };
730:
1.37 schwarze 731: /* Array of injected predefined strings. */
732: #define PREDEFS_MAX 38
733: static const struct predef predefs[PREDEFS_MAX] = {
734: #include "predefs.in"
735: };
736:
1.42 schwarze 737: /* See roffhash_find() */
1.3 schwarze 738: #define ROFF_HASH(p) (p[0] - ASCII_LO)
739:
1.51 schwarze 740: static int roffit_lines; /* number of lines to delay */
741: static char *roffit_macro; /* nil-terminated macro line */
742:
1.80 schwarze 743:
1.138 schwarze 744: /* --- request table ------------------------------------------------------ */
745:
1.3 schwarze 746: static void
1.42 schwarze 747: roffhash_init(void)
1.3 schwarze 748: {
749: struct roffmac *n;
750: int buc, i;
751:
1.16 schwarze 752: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3 schwarze 753: assert(roffs[i].name[0] >= ASCII_LO);
754: assert(roffs[i].name[0] <= ASCII_HI);
755:
756: buc = ROFF_HASH(roffs[i].name);
757:
758: if (NULL != (n = hash[buc])) {
759: for ( ; n->next; n = n->next)
760: /* Do nothing. */ ;
761: n->next = &roffs[i];
762: } else
763: hash[buc] = &roffs[i];
764: }
765: }
766:
1.1 schwarze 767: /*
768: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
769: * the nil-terminated string name could be found.
770: */
771: static enum rofft
1.42 schwarze 772: roffhash_find(const char *p, size_t s)
1.1 schwarze 773: {
1.3 schwarze 774: int buc;
775: struct roffmac *n;
1.1 schwarze 776:
1.3 schwarze 777: /*
778: * libroff has an extremely simple hashtable, for the time
779: * being, which simply keys on the first character, which must
780: * be printable, then walks a chain. It works well enough until
781: * optimised.
782: */
783:
784: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
1.149 schwarze 785: return ROFF_MAX;
1.3 schwarze 786:
787: buc = ROFF_HASH(p);
788:
789: if (NULL == (n = hash[buc]))
1.149 schwarze 790: return ROFF_MAX;
1.3 schwarze 791: for ( ; n; n = n->next)
1.16 schwarze 792: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.149 schwarze 793: return (enum rofft)(n - roffs);
1.1 schwarze 794:
1.149 schwarze 795: return ROFF_MAX;
1.1 schwarze 796: }
797:
1.138 schwarze 798: /* --- stack of request blocks -------------------------------------------- */
799:
1.1 schwarze 800: /*
801: * Pop the current node off of the stack of roff instructions currently
802: * pending.
803: */
804: static void
805: roffnode_pop(struct roff *r)
806: {
807: struct roffnode *p;
808:
1.2 schwarze 809: assert(r->last);
1.80 schwarze 810: p = r->last;
1.2 schwarze 811:
812: r->last = r->last->parent;
1.16 schwarze 813: free(p->name);
814: free(p->end);
1.1 schwarze 815: free(p);
816: }
817:
818: /*
819: * Push a roff node onto the instruction stack. This must later be
820: * removed with roffnode_pop().
821: */
1.11 schwarze 822: static void
1.16 schwarze 823: roffnode_push(struct roff *r, enum rofft tok, const char *name,
824: int line, int col)
1.1 schwarze 825: {
826: struct roffnode *p;
827:
1.11 schwarze 828: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 829: p->tok = tok;
1.16 schwarze 830: if (name)
831: p->name = mandoc_strdup(name);
1.1 schwarze 832: p->parent = r->last;
833: p->line = line;
834: p->col = col;
1.71 schwarze 835: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 836:
837: r->last = p;
838: }
839:
1.138 schwarze 840: /* --- roff parser state data management ---------------------------------- */
841:
1.1 schwarze 842: static void
843: roff_free1(struct roff *r)
844: {
1.49 schwarze 845: struct tbl_node *tbl;
1.32 schwarze 846: struct eqn_node *e;
1.42 schwarze 847: int i;
1.27 schwarze 848:
1.49 schwarze 849: while (NULL != (tbl = r->first_tbl)) {
850: r->first_tbl = tbl->next;
851: tbl_free(tbl);
1.27 schwarze 852: }
853: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 854:
1.32 schwarze 855: while (NULL != (e = r->first_eqn)) {
856: r->first_eqn = e->next;
857: eqn_free(e);
858: }
859: r->first_eqn = r->last_eqn = r->eqn = NULL;
860:
1.1 schwarze 861: while (r->last)
862: roffnode_pop(r);
1.27 schwarze 863:
1.96 schwarze 864: free (r->rstack);
865: r->rstack = NULL;
866: r->rstacksz = 0;
867: r->rstackpos = -1;
868:
869: roff_freereg(r->regtab);
870: r->regtab = NULL;
871:
1.42 schwarze 872: roff_freestr(r->strtab);
873: roff_freestr(r->xmbtab);
874: r->strtab = r->xmbtab = NULL;
875:
876: if (r->xtab)
877: for (i = 0; i < 128; i++)
878: free(r->xtab[i].p);
879: free(r->xtab);
880: r->xtab = NULL;
1.1 schwarze 881: }
882:
883: void
884: roff_reset(struct roff *r)
885: {
886:
887: roff_free1(r);
1.99 schwarze 888: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48 schwarze 889: r->control = 0;
1.1 schwarze 890: }
891:
892: void
893: roff_free(struct roff *r)
894: {
895:
896: roff_free1(r);
897: free(r);
898: }
899:
900: struct roff *
1.151 ! schwarze 901: roff_alloc(struct mparse *parse, int options)
1.1 schwarze 902: {
903: struct roff *r;
904:
1.11 schwarze 905: r = mandoc_calloc(1, sizeof(struct roff));
1.35 schwarze 906: r->parse = parse;
1.72 schwarze 907: r->options = options;
1.99 schwarze 908: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2 schwarze 909: r->rstackpos = -1;
1.80 schwarze 910:
1.42 schwarze 911: roffhash_init();
1.37 schwarze 912:
1.149 schwarze 913: return r;
1.137 schwarze 914: }
915:
1.138 schwarze 916: /* --- syntax tree state data management ---------------------------------- */
917:
1.137 schwarze 918: static void
919: roff_man_free1(struct roff_man *man)
920: {
921:
1.138 schwarze 922: if (man->first != NULL)
923: roff_node_delete(man, man->first);
1.137 schwarze 924: free(man->meta.msec);
925: free(man->meta.vol);
926: free(man->meta.os);
927: free(man->meta.arch);
928: free(man->meta.title);
929: free(man->meta.name);
930: free(man->meta.date);
931: }
932:
933: static void
934: roff_man_alloc1(struct roff_man *man)
935: {
936:
937: memset(&man->meta, 0, sizeof(man->meta));
938: man->first = mandoc_calloc(1, sizeof(*man->first));
939: man->first->type = ROFFT_ROOT;
940: man->last = man->first;
941: man->last_es = NULL;
942: man->flags = 0;
943: man->macroset = MACROSET_NONE;
944: man->lastsec = man->lastnamed = SEC_NONE;
945: man->next = ROFF_NEXT_CHILD;
946: }
947:
948: void
949: roff_man_reset(struct roff_man *man)
950: {
951:
952: roff_man_free1(man);
953: roff_man_alloc1(man);
954: }
955:
956: void
957: roff_man_free(struct roff_man *man)
958: {
959:
960: roff_man_free1(man);
961: free(man);
962: }
963:
964: struct roff_man *
965: roff_man_alloc(struct roff *roff, struct mparse *parse,
966: const char *defos, int quick)
967: {
968: struct roff_man *man;
969:
970: man = mandoc_calloc(1, sizeof(*man));
971: man->parse = parse;
972: man->roff = roff;
973: man->defos = defos;
974: man->quick = quick;
975: roff_man_alloc1(man);
1.149 schwarze 976: return man;
1.1 schwarze 977: }
978:
1.138 schwarze 979: /* --- syntax tree handling ----------------------------------------------- */
980:
981: struct roff_node *
982: roff_node_alloc(struct roff_man *man, int line, int pos,
983: enum roff_type type, int tok)
984: {
985: struct roff_node *n;
986:
987: n = mandoc_calloc(1, sizeof(*n));
988: n->line = line;
989: n->pos = pos;
990: n->tok = tok;
991: n->type = type;
992: n->sec = man->lastsec;
993:
994: if (man->flags & MDOC_SYNOPSIS)
995: n->flags |= MDOC_SYNPRETTY;
996: else
997: n->flags &= ~MDOC_SYNPRETTY;
998: if (man->flags & MDOC_NEWLINE)
999: n->flags |= MDOC_LINE;
1000: man->flags &= ~MDOC_NEWLINE;
1001:
1.149 schwarze 1002: return n;
1.138 schwarze 1003: }
1004:
1005: void
1006: roff_node_append(struct roff_man *man, struct roff_node *n)
1007: {
1008:
1009: switch (man->next) {
1010: case ROFF_NEXT_SIBLING:
1011: man->last->next = n;
1012: n->prev = man->last;
1013: n->parent = man->last->parent;
1014: break;
1015: case ROFF_NEXT_CHILD:
1016: man->last->child = n;
1017: n->parent = man->last;
1018: break;
1019: default:
1020: abort();
1021: }
1022: n->parent->nchild++;
1.142 schwarze 1023: n->parent->last = n;
1.138 schwarze 1024:
1025: /*
1026: * Copy over the normalised-data pointer of our parent. Not
1027: * everybody has one, but copying a null pointer is fine.
1028: */
1029:
1030: switch (n->type) {
1031: case ROFFT_BODY:
1032: if (n->end != ENDBODY_NOT)
1033: break;
1034: /* FALLTHROUGH */
1035: case ROFFT_TAIL:
1036: case ROFFT_HEAD:
1037: n->norm = n->parent->norm;
1038: break;
1039: default:
1040: break;
1041: }
1042:
1043: if (man->macroset == MACROSET_MDOC)
1044: mdoc_valid_pre(man, n);
1045:
1046: switch (n->type) {
1047: case ROFFT_HEAD:
1048: assert(n->parent->type == ROFFT_BLOCK);
1049: n->parent->head = n;
1050: break;
1051: case ROFFT_BODY:
1052: if (n->end)
1053: break;
1054: assert(n->parent->type == ROFFT_BLOCK);
1055: n->parent->body = n;
1056: break;
1057: case ROFFT_TAIL:
1058: assert(n->parent->type == ROFFT_BLOCK);
1059: n->parent->tail = n;
1060: break;
1061: default:
1062: break;
1063: }
1064: man->last = n;
1065: }
1066:
1.139 schwarze 1067: void
1068: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
1069: {
1070: struct roff_node *n;
1071:
1072: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
1073: n->string = roff_strdup(man->roff, word);
1074: roff_node_append(man, n);
1075: if (man->macroset == MACROSET_MDOC)
1076: mdoc_valid_post(man);
1077: else
1078: man_valid_post(man);
1079: man->next = ROFF_NEXT_SIBLING;
1080: }
1081:
1082: void
1083: roff_word_append(struct roff_man *man, const char *word)
1084: {
1085: struct roff_node *n;
1086: char *addstr, *newstr;
1087:
1088: n = man->last;
1089: addstr = roff_strdup(man->roff, word);
1090: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1091: free(addstr);
1092: free(n->string);
1093: n->string = newstr;
1094: man->next = ROFF_NEXT_SIBLING;
1.140 schwarze 1095: }
1096:
1097: void
1098: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
1099: {
1100: struct roff_node *n;
1101:
1102: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
1103: roff_node_append(man, n);
1104: man->next = ROFF_NEXT_CHILD;
1105: }
1106:
1107: struct roff_node *
1108: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
1109: {
1110: struct roff_node *n;
1111:
1112: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
1113: roff_node_append(man, n);
1114: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 1115: return n;
1.139 schwarze 1116: }
1117:
1.138 schwarze 1118: struct roff_node *
1119: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
1120: {
1121: struct roff_node *n;
1122:
1123: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
1124: roff_node_append(man, n);
1125: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 1126: return n;
1.138 schwarze 1127: }
1128:
1129: struct roff_node *
1130: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1131: {
1132: struct roff_node *n;
1133:
1134: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1135: roff_node_append(man, n);
1136: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 1137: return n;
1.139 schwarze 1138: }
1139:
1140: void
1141: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
1142: {
1143: struct roff_node *n;
1144:
1145: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
1146: n->eqn = eqn;
1147: if (eqn->ln > man->last->line)
1148: n->flags |= MDOC_LINE;
1149: roff_node_append(man, n);
1150: man->next = ROFF_NEXT_SIBLING;
1151: }
1152:
1153: void
1154: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
1155: {
1156: struct roff_node *n;
1157:
1158: if (man->macroset == MACROSET_MAN)
1159: man_breakscope(man, TOKEN_NONE);
1160: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
1161: n->span = tbl;
1162: roff_node_append(man, n);
1163: if (man->macroset == MACROSET_MDOC)
1164: mdoc_valid_post(man);
1165: else
1166: man_valid_post(man);
1167: man->next = ROFF_NEXT_SIBLING;
1.138 schwarze 1168: }
1169:
1170: void
1171: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1172: {
1173:
1174: /* Adjust siblings. */
1175:
1176: if (n->prev)
1177: n->prev->next = n->next;
1178: if (n->next)
1179: n->next->prev = n->prev;
1180:
1181: /* Adjust parent. */
1182:
1183: if (n->parent != NULL) {
1184: n->parent->nchild--;
1185: if (n->parent->child == n)
1186: n->parent->child = n->next;
1187: if (n->parent->last == n)
1188: n->parent->last = n->prev;
1189: }
1190:
1191: /* Adjust parse point. */
1192:
1193: if (man == NULL)
1194: return;
1195: if (man->last == n) {
1196: if (n->prev == NULL) {
1197: man->last = n->parent;
1198: man->next = ROFF_NEXT_CHILD;
1199: } else {
1200: man->last = n->prev;
1201: man->next = ROFF_NEXT_SIBLING;
1202: }
1203: }
1204: if (man->first == n)
1205: man->first = NULL;
1206: }
1207:
1208: void
1209: roff_node_free(struct roff_node *n)
1210: {
1211:
1212: if (n->args != NULL)
1213: mdoc_argv_free(n->args);
1214: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1215: free(n->norm);
1216: free(n->string);
1217: free(n);
1218: }
1219:
1220: void
1221: roff_node_delete(struct roff_man *man, struct roff_node *n)
1222: {
1223:
1224: while (n->child != NULL)
1225: roff_node_delete(man, n->child);
1226: assert(n->nchild == 0);
1227: roff_node_unlink(man, n);
1228: roff_node_free(n);
1.141 schwarze 1229: }
1230:
1231: void
1232: deroff(char **dest, const struct roff_node *n)
1233: {
1234: char *cp;
1235: size_t sz;
1236:
1237: if (n->type != ROFFT_TEXT) {
1238: for (n = n->child; n != NULL; n = n->next)
1239: deroff(dest, n);
1240: return;
1241: }
1242:
1243: /* Skip leading whitespace and escape sequences. */
1244:
1245: cp = n->string;
1246: while (*cp != '\0') {
1247: if ('\\' == *cp) {
1248: cp++;
1249: mandoc_escape((const char **)&cp, NULL, NULL);
1250: } else if (isspace((unsigned char)*cp))
1251: cp++;
1252: else
1253: break;
1254: }
1255:
1256: /* Skip trailing whitespace. */
1257:
1258: for (sz = strlen(cp); sz; sz--)
1259: if ( ! isspace((unsigned char)cp[sz-1]))
1260: break;
1261:
1262: /* Skip empty strings. */
1263:
1264: if (sz == 0)
1265: return;
1266:
1267: if (*dest == NULL) {
1268: *dest = mandoc_strndup(cp, sz);
1269: return;
1270: }
1271:
1272: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1273: free(*dest);
1274: *dest = cp;
1.138 schwarze 1275: }
1276:
1277: /* --- main functions of the roff parser ---------------------------------- */
1278:
1.8 schwarze 1279: /*
1.79 schwarze 1280: * In the current line, expand escape sequences that tend to get
1281: * used in numerical expressions and conditional requests.
1282: * Also check the syntax of the remaining escape sequences.
1.8 schwarze 1283: */
1.45 schwarze 1284: static enum rofferr
1.110 schwarze 1285: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8 schwarze 1286: {
1.81 schwarze 1287: char ubuf[24]; /* buffer to print the number */
1.78 schwarze 1288: const char *start; /* start of the string to process */
1.82 schwarze 1289: char *stesc; /* start of an escape sequence ('\\') */
1.23 schwarze 1290: const char *stnam; /* start of the name, after "[(*" */
1291: const char *cp; /* end of the name, e.g. before ']' */
1292: const char *res; /* the string to be substituted */
1.110 schwarze 1293: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 1294: size_t maxl; /* expected length of the escape name */
1295: size_t naml; /* actual length of the escape name */
1.109 schwarze 1296: enum mandoc_esc esc; /* type of the escape sequence */
1297: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 1298: int expand_count; /* to avoid infinite loops */
1.79 schwarze 1299: int npos; /* position in numeric expression */
1.91 schwarze 1300: int arg_complete; /* argument not interrupted by eol */
1.79 schwarze 1301: char term; /* character terminating the escape */
1.8 schwarze 1302:
1.43 schwarze 1303: expand_count = 0;
1.110 schwarze 1304: start = buf->buf + pos;
1.78 schwarze 1305: stesc = strchr(start, '\0') - 1;
1306: while (stesc-- > start) {
1.43 schwarze 1307:
1.78 schwarze 1308: /* Search backwards for the next backslash. */
1309:
1.110 schwarze 1310: if (*stesc != '\\')
1.78 schwarze 1311: continue;
1312:
1313: /* If it is escaped, skip it. */
1314:
1315: for (cp = stesc - 1; cp >= start; cp--)
1.110 schwarze 1316: if (*cp != '\\')
1.78 schwarze 1317: break;
1318:
1.110 schwarze 1319: if ((stesc - cp) % 2 == 0) {
1.82 schwarze 1320: stesc = (char *)cp;
1.78 schwarze 1321: continue;
1322: }
1.23 schwarze 1323:
1.79 schwarze 1324: /* Decide whether to expand or to check only. */
1.23 schwarze 1325:
1.79 schwarze 1326: term = '\0';
1.78 schwarze 1327: cp = stesc + 1;
1.53 schwarze 1328: switch (*cp) {
1.80 schwarze 1329: case '*':
1.53 schwarze 1330: res = NULL;
1331: break;
1.80 schwarze 1332: case 'B':
1333: case 'w':
1.79 schwarze 1334: term = cp[1];
1335: /* FALLTHROUGH */
1.80 schwarze 1336: case 'n':
1.53 schwarze 1337: res = ubuf;
1338: break;
1339: default:
1.109 schwarze 1340: esc = mandoc_escape(&cp, &stnam, &inaml);
1341: if (esc == ESCAPE_ERROR ||
1342: (esc == ESCAPE_SPECIAL &&
1.151 ! schwarze 1343: mchars_spec2cp(stnam, inaml) < 0))
1.92 schwarze 1344: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110 schwarze 1345: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1346: "%.*s", (int)(cp - stesc), stesc);
1.78 schwarze 1347: continue;
1.42 schwarze 1348: }
1349:
1.78 schwarze 1350: if (EXPAND_LIMIT < ++expand_count) {
1351: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110 schwarze 1352: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1353: return ROFF_IGN;
1.78 schwarze 1354: }
1.23 schwarze 1355:
1356: /*
1357: * The third character decides the length
1.53 schwarze 1358: * of the name of the string or register.
1.23 schwarze 1359: * Save a pointer to the name.
1360: */
1361:
1.110 schwarze 1362: if (term == '\0') {
1.79 schwarze 1363: switch (*++cp) {
1.80 schwarze 1364: case '\0':
1.79 schwarze 1365: maxl = 0;
1366: break;
1.80 schwarze 1367: case '(':
1.79 schwarze 1368: cp++;
1369: maxl = 2;
1370: break;
1.80 schwarze 1371: case '[':
1.79 schwarze 1372: cp++;
1373: term = ']';
1374: maxl = 0;
1375: break;
1376: default:
1377: maxl = 1;
1378: break;
1379: }
1380: } else {
1381: cp += 2;
1.8 schwarze 1382: maxl = 0;
1383: }
1.23 schwarze 1384: stnam = cp;
1.8 schwarze 1385:
1.23 schwarze 1386: /* Advance to the end of the name. */
1.8 schwarze 1387:
1.125 schwarze 1388: naml = 0;
1.91 schwarze 1389: arg_complete = 1;
1.125 schwarze 1390: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1391: if (*cp == '\0') {
1.92 schwarze 1392: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 1393: ln, (int)(stesc - buf->buf), stesc);
1.91 schwarze 1394: arg_complete = 0;
1.79 schwarze 1395: break;
1.42 schwarze 1396: }
1.110 schwarze 1397: if (maxl == 0 && *cp == term) {
1.79 schwarze 1398: cp++;
1.125 schwarze 1399: break;
1400: }
1401: if (*cp++ != '\\' || stesc[1] != 'w') {
1402: naml++;
1403: continue;
1404: }
1405: switch (mandoc_escape(&cp, NULL, NULL)) {
1406: case ESCAPE_SPECIAL:
1407: case ESCAPE_UNICODE:
1408: case ESCAPE_NUMBERED:
1409: case ESCAPE_OVERSTRIKE:
1410: naml++;
1411: break;
1412: default:
1.8 schwarze 1413: break;
1.79 schwarze 1414: }
1.8 schwarze 1415: }
1416:
1.23 schwarze 1417: /*
1418: * Retrieve the replacement string; if it is
1419: * undefined, resume searching for escapes.
1420: */
1421:
1.79 schwarze 1422: switch (stesc[1]) {
1.80 schwarze 1423: case '*':
1.91 schwarze 1424: if (arg_complete)
1425: res = roff_getstrn(r, stnam, naml);
1.79 schwarze 1426: break;
1.80 schwarze 1427: case 'B':
1.79 schwarze 1428: npos = 0;
1.91 schwarze 1429: ubuf[0] = arg_complete &&
1.133 schwarze 1430: roff_evalnum(r, ln, stnam, &npos,
1431: NULL, ROFFNUM_SCALE) &&
1.91 schwarze 1432: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1433: ubuf[1] = '\0';
1434: break;
1.80 schwarze 1435: case 'n':
1.91 schwarze 1436: if (arg_complete)
1437: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1438: roff_getregn(r, stnam, naml));
1439: else
1440: ubuf[0] = '\0';
1.79 schwarze 1441: break;
1.80 schwarze 1442: case 'w':
1.91 schwarze 1443: /* use even incomplete args */
1.81 schwarze 1444: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1445: 24 * (int)naml);
1446: break;
1447: }
1.8 schwarze 1448:
1.110 schwarze 1449: if (res == NULL) {
1.92 schwarze 1450: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1451: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1452: "%.*s", (int)naml, stnam);
1.37 schwarze 1453: res = "";
1.118 schwarze 1454: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1455: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1456: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1457: return ROFF_IGN;
1.8 schwarze 1458: }
1459:
1.23 schwarze 1460: /* Replace the escape sequence by the string. */
1461:
1.82 schwarze 1462: *stesc = '\0';
1.110 schwarze 1463: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1464: buf->buf, res, cp) + 1;
1.8 schwarze 1465:
1.78 schwarze 1466: /* Prepare for the next replacement. */
1.8 schwarze 1467:
1.78 schwarze 1468: start = nbuf + pos;
1.110 schwarze 1469: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1470: free(buf->buf);
1471: buf->buf = nbuf;
1.42 schwarze 1472: }
1.149 schwarze 1473: return ROFF_CONT;
1.42 schwarze 1474: }
1475:
1476: /*
1.147 schwarze 1477: * Process text streams.
1.42 schwarze 1478: */
1479: static enum rofferr
1.110 schwarze 1480: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42 schwarze 1481: {
1482: size_t sz;
1483: const char *start;
1.51 schwarze 1484: char *p;
1485: int isz;
1.42 schwarze 1486: enum mandoc_esc esc;
1487:
1.147 schwarze 1488: /* Spring the input line trap. */
1489:
1490: if (roffit_lines == 1) {
1491: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1492: free(buf->buf);
1493: buf->buf = p;
1494: buf->sz = isz + 1;
1495: *offs = 0;
1496: free(roffit_macro);
1497: roffit_lines = 0;
1.149 schwarze 1498: return ROFF_REPARSE;
1.147 schwarze 1499: } else if (roffit_lines > 1)
1500: --roffit_lines;
1501:
1502: /* Convert all breakable hyphens into ASCII_HYPH. */
1503:
1.110 schwarze 1504: start = p = buf->buf + pos;
1.42 schwarze 1505:
1.110 schwarze 1506: while (*p != '\0') {
1.42 schwarze 1507: sz = strcspn(p, "-\\");
1508: p += sz;
1509:
1.110 schwarze 1510: if (*p == '\0')
1.42 schwarze 1511: break;
1512:
1.110 schwarze 1513: if (*p == '\\') {
1.42 schwarze 1514: /* Skip over escapes. */
1515: p++;
1.62 schwarze 1516: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1517: if (esc == ESCAPE_ERROR)
1.42 schwarze 1518: break;
1.136 schwarze 1519: while (*p == '-')
1520: p++;
1.42 schwarze 1521: continue;
1522: } else if (p == start) {
1523: p++;
1524: continue;
1525: }
1526:
1.44 schwarze 1527: if (isalpha((unsigned char)p[-1]) &&
1528: isalpha((unsigned char)p[1]))
1.42 schwarze 1529: *p = ASCII_HYPH;
1530: p++;
1.8 schwarze 1531: }
1.149 schwarze 1532: return ROFF_CONT;
1.8 schwarze 1533: }
1534:
1.1 schwarze 1535: enum rofferr
1.110 schwarze 1536: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1537: {
1538: enum rofft t;
1.27 schwarze 1539: enum rofferr e;
1.110 schwarze 1540: int pos; /* parse point */
1.115 schwarze 1541: int spos; /* saved parse point for messages */
1.110 schwarze 1542: int ppos; /* original offset in buf->buf */
1543: int ctl; /* macro line (boolean) */
1544:
1545: ppos = pos = *offs;
1.1 schwarze 1546:
1.102 schwarze 1547: /* Handle in-line equation delimiters. */
1548:
1.108 schwarze 1549: if (r->tbl == NULL &&
1550: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1551: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1552: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1553: if (e == ROFF_REPARSE)
1.149 schwarze 1554: return e;
1.102 schwarze 1555: assert(e == ROFF_CONT);
1556: }
1557:
1558: /* Expand some escape sequences. */
1.8 schwarze 1559:
1.110 schwarze 1560: e = roff_res(r, buf, ln, pos);
1561: if (e == ROFF_IGN)
1.149 schwarze 1562: return e;
1.110 schwarze 1563: assert(e == ROFF_CONT);
1.8 schwarze 1564:
1.110 schwarze 1565: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1566:
1.8 schwarze 1567: /*
1.2 schwarze 1568: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1569: * text through the macro's filter.
1570: * Equations process all content themselves.
1571: * Tables process almost all content themselves, but we want
1572: * to warn about macros before passing it there.
1.2 schwarze 1573: */
1574:
1.124 schwarze 1575: if (r->last != NULL && ! ctl) {
1.2 schwarze 1576: t = r->last->tok;
1577: assert(roffs[t].text);
1.110 schwarze 1578: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1579: assert(e == ROFF_IGN || e == ROFF_CONT);
1580: if (e != ROFF_CONT)
1.149 schwarze 1581: return e;
1.54 schwarze 1582: }
1.124 schwarze 1583: if (r->eqn != NULL)
1.149 schwarze 1584: return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.124 schwarze 1585: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.149 schwarze 1586: return tbl_read(r->tbl, ln, buf->buf, ppos);
1.124 schwarze 1587: if ( ! ctl)
1.149 schwarze 1588: return roff_parsetext(buf, pos, offs);
1.100 schwarze 1589:
1590: /* Skip empty request lines. */
1591:
1.110 schwarze 1592: if (buf->buf[pos] == '"') {
1.100 schwarze 1593: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1594: ln, pos, NULL);
1.149 schwarze 1595: return ROFF_IGN;
1.110 schwarze 1596: } else if (buf->buf[pos] == '\0')
1.149 schwarze 1597: return ROFF_IGN;
1.2 schwarze 1598:
1599: /*
1600: * If a scope is open, go to the child handler for that macro,
1601: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1602: * Don't do so if an equation is open.
1.2 schwarze 1603: */
1604:
1605: if (r->last) {
1.1 schwarze 1606: t = r->last->tok;
1607: assert(roffs[t].sub);
1.149 schwarze 1608: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1609: }
1610:
1.115 schwarze 1611: /* No scope is open. This is a new request or macro. */
1612:
1613: spos = pos;
1614: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1615:
1616: /* Tables ignore most macros. */
1617:
1618: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1619: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1620: ln, pos, buf->buf + spos);
1.129 schwarze 1621: if (t == ROFF_TS)
1.149 schwarze 1622: return ROFF_IGN;
1.129 schwarze 1623: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1624: pos++;
1625: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1626: pos++;
1.149 schwarze 1627: return tbl_read(r->tbl, ln, buf->buf, pos);
1.115 schwarze 1628: }
1629:
1.2 schwarze 1630: /*
1.115 schwarze 1631: * This is neither a roff request nor a user-defined macro.
1632: * Let the standard macro set parsers handle it.
1.2 schwarze 1633: */
1634:
1.115 schwarze 1635: if (t == ROFF_MAX)
1.149 schwarze 1636: return ROFF_CONT;
1.115 schwarze 1637:
1638: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1639:
1.2 schwarze 1640: assert(roffs[t].proc);
1.149 schwarze 1641: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1642: }
1643:
1.27 schwarze 1644: void
1.2 schwarze 1645: roff_endparse(struct roff *r)
1646: {
1.1 schwarze 1647:
1.27 schwarze 1648: if (r->last)
1.94 schwarze 1649: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1650: r->last->line, r->last->col,
1651: roffs[r->last->tok].name);
1.27 schwarze 1652:
1.32 schwarze 1653: if (r->eqn) {
1.94 schwarze 1654: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1655: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41 schwarze 1656: eqn_end(&r->eqn);
1.32 schwarze 1657: }
1658:
1.27 schwarze 1659: if (r->tbl) {
1.94 schwarze 1660: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1661: r->tbl->line, r->tbl->pos, "TS");
1.41 schwarze 1662: tbl_end(&r->tbl);
1.27 schwarze 1663: }
1.1 schwarze 1664: }
1665:
1666: /*
1667: * Parse a roff node's type from the input buffer. This must be in the
1668: * form of ".foo xxx" in the usual way.
1669: */
1670: static enum rofft
1.87 schwarze 1671: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1672: {
1.87 schwarze 1673: char *cp;
1.16 schwarze 1674: const char *mac;
1675: size_t maclen;
1.1 schwarze 1676: enum rofft t;
1677:
1.87 schwarze 1678: cp = buf + *pos;
1679:
1680: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.149 schwarze 1681: return ROFF_MAX;
1.1 schwarze 1682:
1.87 schwarze 1683: mac = cp;
1684: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 1685:
1.16 schwarze 1686: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42 schwarze 1687: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1 schwarze 1688:
1.87 schwarze 1689: if (ROFF_MAX != t)
1690: *pos = cp - buf;
1.1 schwarze 1691:
1.149 schwarze 1692: return t;
1.1 schwarze 1693: }
1694:
1.138 schwarze 1695: /* --- handling of request blocks ----------------------------------------- */
1696:
1.1 schwarze 1697: static enum rofferr
1.2 schwarze 1698: roff_cblock(ROFF_ARGS)
1.1 schwarze 1699: {
1700:
1.2 schwarze 1701: /*
1702: * A block-close `..' should only be invoked as a child of an
1703: * ignore macro, otherwise raise a warning and just ignore it.
1704: */
1705:
1.110 schwarze 1706: if (r->last == NULL) {
1.94 schwarze 1707: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1708: ln, ppos, "..");
1.149 schwarze 1709: return ROFF_IGN;
1.2 schwarze 1710: }
1.1 schwarze 1711:
1.2 schwarze 1712: switch (r->last->tok) {
1.80 schwarze 1713: case ROFF_am:
1.93 schwarze 1714: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80 schwarze 1715: case ROFF_ami:
1716: case ROFF_de:
1.23 schwarze 1717: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80 schwarze 1718: case ROFF_dei:
1719: case ROFF_ig:
1.2 schwarze 1720: break;
1721: default:
1.94 schwarze 1722: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1723: ln, ppos, "..");
1.149 schwarze 1724: return ROFF_IGN;
1.2 schwarze 1725: }
1726:
1.110 schwarze 1727: if (buf->buf[pos] != '\0')
1.90 schwarze 1728: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 1729: ".. %s", buf->buf + pos);
1.2 schwarze 1730:
1731: roffnode_pop(r);
1732: roffnode_cleanscope(r);
1.149 schwarze 1733: return ROFF_IGN;
1.2 schwarze 1734:
1735: }
1.1 schwarze 1736:
1.2 schwarze 1737: static void
1738: roffnode_cleanscope(struct roff *r)
1739: {
1.1 schwarze 1740:
1.2 schwarze 1741: while (r->last) {
1.46 schwarze 1742: if (--r->last->endspan != 0)
1.2 schwarze 1743: break;
1744: roffnode_pop(r);
1745: }
1746: }
1.1 schwarze 1747:
1.68 schwarze 1748: static void
1749: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 1750: {
1.1 schwarze 1751:
1.2 schwarze 1752: if (NULL == r->last) {
1.94 schwarze 1753: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1754: ln, ppos, "\\}");
1.68 schwarze 1755: return;
1.2 schwarze 1756: }
1.1 schwarze 1757:
1.2 schwarze 1758: switch (r->last->tok) {
1.80 schwarze 1759: case ROFF_el:
1760: case ROFF_ie:
1761: case ROFF_if:
1.2 schwarze 1762: break;
1763: default:
1.94 schwarze 1764: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1765: ln, ppos, "\\}");
1.68 schwarze 1766: return;
1.2 schwarze 1767: }
1.1 schwarze 1768:
1.2 schwarze 1769: if (r->last->endspan > -1) {
1.94 schwarze 1770: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1771: ln, ppos, "\\}");
1.68 schwarze 1772: return;
1.2 schwarze 1773: }
1774:
1775: roffnode_pop(r);
1776: roffnode_cleanscope(r);
1.68 schwarze 1777: return;
1.1 schwarze 1778: }
1779:
1780: static enum rofferr
1.2 schwarze 1781: roff_block(ROFF_ARGS)
1.1 schwarze 1782: {
1.93 schwarze 1783: const char *name;
1784: char *iname, *cp;
1.86 schwarze 1785: size_t namesz;
1.16 schwarze 1786:
1.93 schwarze 1787: /* Ignore groff compatibility mode for now. */
1.2 schwarze 1788:
1.110 schwarze 1789: if (tok == ROFF_de1)
1.93 schwarze 1790: tok = ROFF_de;
1.123 schwarze 1791: else if (tok == ROFF_dei1)
1792: tok = ROFF_dei;
1.110 schwarze 1793: else if (tok == ROFF_am1)
1.93 schwarze 1794: tok = ROFF_am;
1.123 schwarze 1795: else if (tok == ROFF_ami1)
1796: tok = ROFF_ami;
1.93 schwarze 1797:
1798: /* Parse the macro name argument. */
1799:
1.110 schwarze 1800: cp = buf->buf + pos;
1801: if (tok == ROFF_ig) {
1.93 schwarze 1802: iname = NULL;
1803: namesz = 0;
1804: } else {
1805: iname = cp;
1806: namesz = roff_getname(r, &cp, ln, ppos);
1807: iname[namesz] = '\0';
1808: }
1.22 schwarze 1809:
1.93 schwarze 1810: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 1811:
1.110 schwarze 1812: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1813: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1814: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1815: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1816: "%.*s", (int)namesz, iname);
1817: namesz = 0;
1818: } else
1819: namesz = strlen(name);
1820: } else
1821: name = iname;
1.22 schwarze 1822:
1.110 schwarze 1823: if (namesz == 0 && tok != ROFF_ig) {
1.93 schwarze 1824: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1825: ln, ppos, roffs[tok].name);
1.149 schwarze 1826: return ROFF_IGN;
1.93 schwarze 1827: }
1.2 schwarze 1828:
1.16 schwarze 1829: roffnode_push(r, tok, name, ln, ppos);
1830:
1831: /*
1832: * At the beginning of a `de' macro, clear the existing string
1833: * with the same name, if there is one. New content will be
1.66 schwarze 1834: * appended from roff_block_text() in multiline mode.
1.16 schwarze 1835: */
1.22 schwarze 1836:
1.110 schwarze 1837: if (tok == ROFF_de || tok == ROFF_dei)
1.86 schwarze 1838: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2 schwarze 1839:
1.110 schwarze 1840: if (*cp == '\0')
1.149 schwarze 1841: return ROFF_IGN;
1.1 schwarze 1842:
1.93 schwarze 1843: /* Get the custom end marker. */
1.22 schwarze 1844:
1.93 schwarze 1845: iname = cp;
1.86 schwarze 1846: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 1847:
1848: /* Resolve the end marker if it is indirect. */
1849:
1.110 schwarze 1850: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1851: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1852: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1853: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1854: "%.*s", (int)namesz, iname);
1855: namesz = 0;
1856: } else
1857: namesz = strlen(name);
1858: } else
1859: name = iname;
1860:
1.86 schwarze 1861: if (namesz)
1862: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 1863:
1.110 schwarze 1864: if (*cp != '\0')
1.90 schwarze 1865: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1866: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.1 schwarze 1867:
1.149 schwarze 1868: return ROFF_IGN;
1.1 schwarze 1869: }
1870:
1871: static enum rofferr
1.2 schwarze 1872: roff_block_sub(ROFF_ARGS)
1.1 schwarze 1873: {
1.2 schwarze 1874: enum rofft t;
1875: int i, j;
1876:
1877: /*
1878: * First check whether a custom macro exists at this level. If
1879: * it does, then check against it. This is some of groff's
1880: * stranger behaviours. If we encountered a custom end-scope
1881: * tag and that tag also happens to be a "real" macro, then we
1882: * need to try interpreting it again as a real macro. If it's
1883: * not, then return ignore. Else continue.
1884: */
1885:
1886: if (r->last->end) {
1.35 schwarze 1887: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 1888: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 1889: break;
1.1 schwarze 1890:
1.110 schwarze 1891: if (r->last->end[j] == '\0' &&
1892: (buf->buf[i] == '\0' ||
1893: buf->buf[i] == ' ' ||
1894: buf->buf[i] == '\t')) {
1.2 schwarze 1895: roffnode_pop(r);
1896: roffnode_cleanscope(r);
1.1 schwarze 1897:
1.110 schwarze 1898: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 1899: i++;
1900:
1901: pos = i;
1.110 schwarze 1902: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1903: ROFF_MAX)
1.149 schwarze 1904: return ROFF_RERUN;
1905: return ROFF_IGN;
1.2 schwarze 1906: }
1.1 schwarze 1907: }
1908:
1.2 schwarze 1909: /*
1910: * If we have no custom end-query or lookup failed, then try
1911: * pulling it out of the hashtable.
1912: */
1.1 schwarze 1913:
1.110 schwarze 1914: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 1915:
1.110 schwarze 1916: if (t != ROFF_cblock) {
1917: if (tok != ROFF_ig)
1918: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149 schwarze 1919: return ROFF_IGN;
1.16 schwarze 1920: }
1.1 schwarze 1921:
1.2 schwarze 1922: assert(roffs[t].proc);
1.149 schwarze 1923: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1924: }
1925:
1926: static enum rofferr
1927: roff_block_text(ROFF_ARGS)
1928: {
1929:
1.110 schwarze 1930: if (tok != ROFF_ig)
1931: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 1932:
1.149 schwarze 1933: return ROFF_IGN;
1.2 schwarze 1934: }
1935:
1936: static enum rofferr
1937: roff_cond_sub(ROFF_ARGS)
1938: {
1939: enum rofft t;
1.37 schwarze 1940: char *ep;
1.71 schwarze 1941: int rr;
1.2 schwarze 1942:
1943: rr = r->last->rule;
1.37 schwarze 1944: roffnode_cleanscope(r);
1.110 schwarze 1945: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2 schwarze 1946:
1.37 schwarze 1947: /*
1.50 schwarze 1948: * Fully handle known macros when they are structurally
1949: * required or when the conditional evaluated to true.
1.5 schwarze 1950: */
1951:
1.110 schwarze 1952: if ((t != ROFF_MAX) &&
1953: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.50 schwarze 1954: assert(roffs[t].proc);
1.149 schwarze 1955: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.50 schwarze 1956: }
1.39 schwarze 1957:
1.69 schwarze 1958: /*
1959: * If `\}' occurs on a macro line without a preceding macro,
1960: * drop the line completely.
1961: */
1962:
1.110 schwarze 1963: ep = buf->buf + pos;
1964: if (ep[0] == '\\' && ep[1] == '}')
1.71 schwarze 1965: rr = 0;
1.69 schwarze 1966:
1.50 schwarze 1967: /* Always check for the closing delimiter `\}'. */
1.39 schwarze 1968:
1.110 schwarze 1969: while ((ep = strchr(ep, '\\')) != NULL) {
1970: if (*(++ep) == '}') {
1.70 schwarze 1971: *ep = '&';
1.110 schwarze 1972: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1973: }
1.119 schwarze 1974: if (*ep != '\0')
1975: ++ep;
1.50 schwarze 1976: }
1.149 schwarze 1977: return rr ? ROFF_CONT : ROFF_IGN;
1.2 schwarze 1978: }
1979:
1980: static enum rofferr
1981: roff_cond_text(ROFF_ARGS)
1982: {
1.37 schwarze 1983: char *ep;
1.71 schwarze 1984: int rr;
1.2 schwarze 1985:
1986: rr = r->last->rule;
1.37 schwarze 1987: roffnode_cleanscope(r);
1.1 schwarze 1988:
1.110 schwarze 1989: ep = buf->buf + pos;
1990: while ((ep = strchr(ep, '\\')) != NULL) {
1991: if (*(++ep) == '}') {
1.70 schwarze 1992: *ep = '&';
1.110 schwarze 1993: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1994: }
1.119 schwarze 1995: if (*ep != '\0')
1996: ++ep;
1.2 schwarze 1997: }
1.149 schwarze 1998: return rr ? ROFF_CONT : ROFF_IGN;
1.2 schwarze 1999: }
2000:
1.138 schwarze 2001: /* --- handling of numeric and conditional expressions -------------------- */
2002:
1.77 schwarze 2003: /*
2004: * Parse a single signed integer number. Stop at the first non-digit.
2005: * If there is at least one digit, return success and advance the
2006: * parse point, else return failure and let the parse point unchanged.
2007: * Ignore overflows, treat them just like the C language.
2008: */
1.56 schwarze 2009: static int
1.133 schwarze 2010: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56 schwarze 2011: {
1.133 schwarze 2012: int myres, scaled, n, p;
1.79 schwarze 2013:
2014: if (NULL == res)
2015: res = &myres;
1.56 schwarze 2016:
2017: p = *pos;
2018: n = v[p] == '-';
1.133 schwarze 2019: if (n || v[p] == '+')
1.56 schwarze 2020: p++;
2021:
1.133 schwarze 2022: if (flags & ROFFNUM_WHITE)
2023: while (isspace((unsigned char)v[p]))
2024: p++;
2025:
1.56 schwarze 2026: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 2027: *res = 10 * *res + v[p] - '0';
1.56 schwarze 2028: if (p == *pos + n)
2029: return 0;
2030:
2031: if (n)
2032: *res = -*res;
2033:
1.126 schwarze 2034: /* Each number may be followed by one optional scaling unit. */
2035:
2036: switch (v[p]) {
2037: case 'f':
1.133 schwarze 2038: scaled = *res * 65536;
1.126 schwarze 2039: break;
2040: case 'i':
1.133 schwarze 2041: scaled = *res * 240;
1.126 schwarze 2042: break;
2043: case 'c':
1.133 schwarze 2044: scaled = *res * 240 / 2.54;
1.126 schwarze 2045: break;
2046: case 'v':
2047: case 'P':
1.133 schwarze 2048: scaled = *res * 40;
1.126 schwarze 2049: break;
2050: case 'm':
2051: case 'n':
1.133 schwarze 2052: scaled = *res * 24;
1.126 schwarze 2053: break;
2054: case 'p':
1.133 schwarze 2055: scaled = *res * 10 / 3;
1.126 schwarze 2056: break;
2057: case 'u':
1.133 schwarze 2058: scaled = *res;
1.126 schwarze 2059: break;
2060: case 'M':
1.133 schwarze 2061: scaled = *res * 6 / 25;
1.126 schwarze 2062: break;
2063: default:
1.133 schwarze 2064: scaled = *res;
1.126 schwarze 2065: p--;
2066: break;
2067: }
1.133 schwarze 2068: if (flags & ROFFNUM_SCALE)
2069: *res = scaled;
1.126 schwarze 2070:
2071: *pos = p + 1;
1.149 schwarze 2072: return 1;
1.56 schwarze 2073: }
2074:
1.71 schwarze 2075: /*
2076: * Evaluate a string comparison condition.
2077: * The first character is the delimiter.
2078: * Succeed if the string up to its second occurrence
2079: * matches the string up to its third occurence.
2080: * Advance the cursor after the third occurrence
2081: * or lacking that, to the end of the line.
2082: */
2083: static int
2084: roff_evalstrcond(const char *v, int *pos)
2085: {
2086: const char *s1, *s2, *s3;
2087: int match;
2088:
2089: match = 0;
2090: s1 = v + *pos; /* initial delimiter */
2091: s2 = s1 + 1; /* for scanning the first string */
2092: s3 = strchr(s2, *s1); /* for scanning the second string */
2093:
2094: if (NULL == s3) /* found no middle delimiter */
2095: goto out;
2096:
2097: while ('\0' != *++s3) {
2098: if (*s2 != *s3) { /* mismatch */
2099: s3 = strchr(s3, *s1);
2100: break;
2101: }
2102: if (*s3 == *s1) { /* found the final delimiter */
2103: match = 1;
2104: break;
2105: }
2106: s2++;
2107: }
2108:
2109: out:
2110: if (NULL == s3)
2111: s3 = strchr(s2, '\0');
1.114 schwarze 2112: else if (*s3 != '\0')
1.71 schwarze 2113: s3++;
2114: *pos = s3 - v;
1.149 schwarze 2115: return match;
1.71 schwarze 2116: }
2117:
1.77 schwarze 2118: /*
2119: * Evaluate an optionally negated single character, numerical,
2120: * or string condition.
2121: */
1.71 schwarze 2122: static int
1.143 schwarze 2123: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5 schwarze 2124: {
1.143 schwarze 2125: char *cp, *name;
2126: size_t sz;
1.113 schwarze 2127: int number, savepos, wanttrue;
1.5 schwarze 2128:
1.71 schwarze 2129: if ('!' == v[*pos]) {
2130: wanttrue = 0;
2131: (*pos)++;
2132: } else
2133: wanttrue = 1;
2134:
1.5 schwarze 2135: switch (v[*pos]) {
1.112 schwarze 2136: case '\0':
1.149 schwarze 2137: return 0;
1.80 schwarze 2138: case 'n':
2139: case 'o':
1.5 schwarze 2140: (*pos)++;
1.149 schwarze 2141: return wanttrue;
1.80 schwarze 2142: case 'c':
2143: case 'd':
2144: case 'e':
2145: case 't':
1.111 schwarze 2146: case 'v':
1.5 schwarze 2147: (*pos)++;
1.149 schwarze 2148: return !wanttrue;
1.143 schwarze 2149: case 'r':
2150: cp = name = v + ++*pos;
2151: sz = roff_getname(r, &cp, ln, *pos);
2152: *pos = cp - v;
1.149 schwarze 2153: return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.5 schwarze 2154: default:
2155: break;
2156: }
2157:
1.113 schwarze 2158: savepos = *pos;
1.133 schwarze 2159: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149 schwarze 2160: return (number > 0) == wanttrue;
1.113 schwarze 2161: else if (*pos == savepos)
1.149 schwarze 2162: return roff_evalstrcond(v, pos) == wanttrue;
1.77 schwarze 2163: else
1.149 schwarze 2164: return 0;
1.5 schwarze 2165: }
2166:
1.2 schwarze 2167: static enum rofferr
1.21 schwarze 2168: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 2169: {
1.30 schwarze 2170:
1.149 schwarze 2171: return ROFF_IGN;
1.21 schwarze 2172: }
2173:
2174: static enum rofferr
1.123 schwarze 2175: roff_insec(ROFF_ARGS)
2176: {
2177:
2178: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
2179: ln, ppos, roffs[tok].name);
1.149 schwarze 2180: return ROFF_IGN;
1.123 schwarze 2181: }
2182:
2183: static enum rofferr
2184: roff_unsupp(ROFF_ARGS)
2185: {
2186:
2187: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
2188: ln, ppos, roffs[tok].name);
1.149 schwarze 2189: return ROFF_IGN;
1.123 schwarze 2190: }
2191:
2192: static enum rofferr
1.2 schwarze 2193: roff_cond(ROFF_ARGS)
2194: {
1.46 schwarze 2195:
2196: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 2197:
1.80 schwarze 2198: /*
1.35 schwarze 2199: * An `.el' has no conditional body: it will consume the value
2200: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 2201: * defaults to DENY.
1.35 schwarze 2202: *
2203: * If we're not an `el', however, then evaluate the conditional.
2204: */
1.1 schwarze 2205:
1.110 schwarze 2206: r->last->rule = tok == ROFF_el ?
1.80 schwarze 2207: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 2208: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 2209:
1.35 schwarze 2210: /*
2211: * An if-else will put the NEGATION of the current evaluated
2212: * conditional into the stack of rules.
2213: */
2214:
1.110 schwarze 2215: if (tok == ROFF_ie) {
1.96 schwarze 2216: if (r->rstackpos + 1 == r->rstacksz) {
2217: r->rstacksz += 16;
2218: r->rstack = mandoc_reallocarray(r->rstack,
2219: r->rstacksz, sizeof(int));
1.35 schwarze 2220: }
1.71 schwarze 2221: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 2222: }
1.5 schwarze 2223:
2224: /* If the parent has false as its rule, then so do we. */
2225:
1.71 schwarze 2226: if (r->last->parent && !r->last->parent->rule)
2227: r->last->rule = 0;
1.5 schwarze 2228:
2229: /*
1.46 schwarze 2230: * Determine scope.
2231: * If there is nothing on the line after the conditional,
2232: * not even whitespace, use next-line scope.
1.5 schwarze 2233: */
1.2 schwarze 2234:
1.110 schwarze 2235: if (buf->buf[pos] == '\0') {
1.46 schwarze 2236: r->last->endspan = 2;
2237: goto out;
2238: }
2239:
1.110 schwarze 2240: while (buf->buf[pos] == ' ')
1.46 schwarze 2241: pos++;
2242:
2243: /* An opening brace requests multiline scope. */
1.2 schwarze 2244:
1.110 schwarze 2245: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 2246: r->last->endspan = -1;
2247: pos += 2;
1.144 schwarze 2248: while (buf->buf[pos] == ' ')
2249: pos++;
1.46 schwarze 2250: goto out;
1.80 schwarze 2251: }
1.2 schwarze 2252:
2253: /*
1.46 schwarze 2254: * Anything else following the conditional causes
2255: * single-line scope. Warn if the scope contains
2256: * nothing but trailing whitespace.
1.2 schwarze 2257: */
2258:
1.110 schwarze 2259: if (buf->buf[pos] == '\0')
1.89 schwarze 2260: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
2261: ln, ppos, roffs[tok].name);
1.2 schwarze 2262:
1.46 schwarze 2263: r->last->endspan = 1;
1.1 schwarze 2264:
1.46 schwarze 2265: out:
1.2 schwarze 2266: *offs = pos;
1.149 schwarze 2267: return ROFF_RERUN;
1.1 schwarze 2268: }
2269:
1.2 schwarze 2270: static enum rofferr
1.7 schwarze 2271: roff_ds(ROFF_ARGS)
2272: {
1.85 schwarze 2273: char *string;
2274: const char *name;
2275: size_t namesz;
1.10 schwarze 2276:
1.123 schwarze 2277: /* Ignore groff compatibility mode for now. */
2278:
2279: if (tok == ROFF_ds1)
2280: tok = ROFF_ds;
2281: else if (tok == ROFF_as1)
2282: tok = ROFF_as;
2283:
1.10 schwarze 2284: /*
1.85 schwarze 2285: * The first word is the name of the string.
2286: * If it is empty or terminated by an escape sequence,
2287: * abort the `ds' request without defining anything.
1.10 schwarze 2288: */
1.7 schwarze 2289:
1.110 schwarze 2290: name = string = buf->buf + pos;
2291: if (*name == '\0')
1.149 schwarze 2292: return ROFF_IGN;
1.7 schwarze 2293:
1.85 schwarze 2294: namesz = roff_getname(r, &string, ln, pos);
1.110 schwarze 2295: if (name[namesz] == '\\')
1.149 schwarze 2296: return ROFF_IGN;
1.85 schwarze 2297:
2298: /* Read past the initial double-quote, if any. */
1.110 schwarze 2299: if (*string == '"')
1.7 schwarze 2300: string++;
2301:
1.10 schwarze 2302: /* The rest is the value. */
1.85 schwarze 2303: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2304: ROFF_as == tok);
1.149 schwarze 2305: return ROFF_IGN;
1.7 schwarze 2306: }
2307:
1.77 schwarze 2308: /*
2309: * Parse a single operator, one or two characters long.
2310: * If the operator is recognized, return success and advance the
2311: * parse point, else return failure and let the parse point unchanged.
2312: */
2313: static int
2314: roff_getop(const char *v, int *pos, char *res)
2315: {
2316:
2317: *res = v[*pos];
2318:
2319: switch (*res) {
1.80 schwarze 2320: case '+':
2321: case '-':
2322: case '*':
2323: case '/':
2324: case '%':
2325: case '&':
2326: case ':':
1.77 schwarze 2327: break;
2328: case '<':
2329: switch (v[*pos + 1]) {
1.80 schwarze 2330: case '=':
1.77 schwarze 2331: *res = 'l';
2332: (*pos)++;
2333: break;
1.80 schwarze 2334: case '>':
1.77 schwarze 2335: *res = '!';
2336: (*pos)++;
2337: break;
1.80 schwarze 2338: case '?':
1.77 schwarze 2339: *res = 'i';
2340: (*pos)++;
2341: break;
2342: default:
2343: break;
2344: }
2345: break;
2346: case '>':
2347: switch (v[*pos + 1]) {
1.80 schwarze 2348: case '=':
1.77 schwarze 2349: *res = 'g';
2350: (*pos)++;
2351: break;
1.80 schwarze 2352: case '?':
1.77 schwarze 2353: *res = 'a';
2354: (*pos)++;
2355: break;
2356: default:
2357: break;
2358: }
2359: break;
2360: case '=':
2361: if ('=' == v[*pos + 1])
2362: (*pos)++;
2363: break;
2364: default:
1.149 schwarze 2365: return 0;
1.77 schwarze 2366: }
2367: (*pos)++;
2368:
1.149 schwarze 2369: return *res;
1.77 schwarze 2370: }
2371:
2372: /*
2373: * Evaluate either a parenthesized numeric expression
2374: * or a single signed integer number.
2375: */
2376: static int
1.106 schwarze 2377: roff_evalpar(struct roff *r, int ln,
1.133 schwarze 2378: const char *v, int *pos, int *res, int flags)
1.77 schwarze 2379: {
2380:
2381: if ('(' != v[*pos])
1.149 schwarze 2382: return roff_getnum(v, pos, res, flags);
1.77 schwarze 2383:
2384: (*pos)++;
1.133 schwarze 2385: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149 schwarze 2386: return 0;
1.77 schwarze 2387:
1.79 schwarze 2388: /*
2389: * Omission of the closing parenthesis
2390: * is an error in validation mode,
2391: * but ignored in evaluation mode.
2392: */
2393:
1.77 schwarze 2394: if (')' == v[*pos])
2395: (*pos)++;
1.79 schwarze 2396: else if (NULL == res)
1.149 schwarze 2397: return 0;
1.77 schwarze 2398:
1.149 schwarze 2399: return 1;
1.77 schwarze 2400: }
2401:
2402: /*
2403: * Evaluate a complete numeric expression.
2404: * Proceed left to right, there is no concept of precedence.
2405: */
2406: static int
1.106 schwarze 2407: roff_evalnum(struct roff *r, int ln, const char *v,
1.133 schwarze 2408: int *pos, int *res, int flags)
1.77 schwarze 2409: {
2410: int mypos, operand2;
2411: char operator;
2412:
2413: if (NULL == pos) {
2414: mypos = 0;
2415: pos = &mypos;
2416: }
2417:
1.133 schwarze 2418: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2419: while (isspace((unsigned char)v[*pos]))
2420: (*pos)++;
2421:
1.133 schwarze 2422: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149 schwarze 2423: return 0;
1.77 schwarze 2424:
2425: while (1) {
1.133 schwarze 2426: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2427: while (isspace((unsigned char)v[*pos]))
2428: (*pos)++;
2429:
2430: if ( ! roff_getop(v, pos, &operator))
2431: break;
2432:
1.133 schwarze 2433: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2434: while (isspace((unsigned char)v[*pos]))
2435: (*pos)++;
2436:
1.133 schwarze 2437: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149 schwarze 2438: return 0;
1.77 schwarze 2439:
1.133 schwarze 2440: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2441: while (isspace((unsigned char)v[*pos]))
2442: (*pos)++;
1.79 schwarze 2443:
2444: if (NULL == res)
2445: continue;
1.77 schwarze 2446:
2447: switch (operator) {
1.80 schwarze 2448: case '+':
1.77 schwarze 2449: *res += operand2;
2450: break;
1.80 schwarze 2451: case '-':
1.77 schwarze 2452: *res -= operand2;
2453: break;
1.80 schwarze 2454: case '*':
1.77 schwarze 2455: *res *= operand2;
2456: break;
1.80 schwarze 2457: case '/':
1.116 schwarze 2458: if (operand2 == 0) {
1.106 schwarze 2459: mandoc_msg(MANDOCERR_DIVZERO,
2460: r->parse, ln, *pos, v);
2461: *res = 0;
2462: break;
2463: }
1.77 schwarze 2464: *res /= operand2;
2465: break;
1.80 schwarze 2466: case '%':
1.116 schwarze 2467: if (operand2 == 0) {
2468: mandoc_msg(MANDOCERR_DIVZERO,
2469: r->parse, ln, *pos, v);
2470: *res = 0;
2471: break;
2472: }
1.77 schwarze 2473: *res %= operand2;
2474: break;
1.80 schwarze 2475: case '<':
1.77 schwarze 2476: *res = *res < operand2;
2477: break;
1.80 schwarze 2478: case '>':
1.77 schwarze 2479: *res = *res > operand2;
2480: break;
1.80 schwarze 2481: case 'l':
1.77 schwarze 2482: *res = *res <= operand2;
2483: break;
1.80 schwarze 2484: case 'g':
1.77 schwarze 2485: *res = *res >= operand2;
2486: break;
1.80 schwarze 2487: case '=':
1.77 schwarze 2488: *res = *res == operand2;
2489: break;
1.80 schwarze 2490: case '!':
1.77 schwarze 2491: *res = *res != operand2;
2492: break;
1.80 schwarze 2493: case '&':
1.77 schwarze 2494: *res = *res && operand2;
2495: break;
1.80 schwarze 2496: case ':':
1.77 schwarze 2497: *res = *res || operand2;
2498: break;
1.80 schwarze 2499: case 'i':
1.77 schwarze 2500: if (operand2 < *res)
2501: *res = operand2;
2502: break;
1.80 schwarze 2503: case 'a':
1.77 schwarze 2504: if (operand2 > *res)
2505: *res = operand2;
2506: break;
2507: default:
2508: abort();
2509: }
2510: }
1.149 schwarze 2511: return 1;
1.77 schwarze 2512: }
2513:
1.138 schwarze 2514: /* --- register management ------------------------------------------------ */
2515:
1.52 schwarze 2516: void
1.60 schwarze 2517: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 2518: {
1.52 schwarze 2519: struct roffreg *reg;
2520:
2521: /* Search for an existing register with the same name. */
2522: reg = r->regtab;
2523:
2524: while (reg && strcmp(name, reg->key.p))
2525: reg = reg->next;
1.41 schwarze 2526:
1.52 schwarze 2527: if (NULL == reg) {
2528: /* Create a new register. */
2529: reg = mandoc_malloc(sizeof(struct roffreg));
2530: reg->key.p = mandoc_strdup(name);
2531: reg->key.sz = strlen(name);
1.60 schwarze 2532: reg->val = 0;
1.52 schwarze 2533: reg->next = r->regtab;
2534: r->regtab = reg;
2535: }
2536:
1.60 schwarze 2537: if ('+' == sign)
2538: reg->val += val;
2539: else if ('-' == sign)
2540: reg->val -= val;
2541: else
2542: reg->val = val;
1.41 schwarze 2543: }
2544:
1.65 schwarze 2545: /*
2546: * Handle some predefined read-only number registers.
2547: * For now, return -1 if the requested register is not predefined;
2548: * in case a predefined read-only register having the value -1
2549: * were to turn up, another special value would have to be chosen.
2550: */
2551: static int
1.145 schwarze 2552: roff_getregro(const struct roff *r, const char *name)
1.65 schwarze 2553: {
2554:
2555: switch (*name) {
1.145 schwarze 2556: case '$': /* Number of arguments of the last macro evaluated. */
1.149 schwarze 2557: return r->argc;
1.80 schwarze 2558: case 'A': /* ASCII approximation mode is always off. */
1.149 schwarze 2559: return 0;
1.80 schwarze 2560: case 'g': /* Groff compatibility mode is always on. */
1.149 schwarze 2561: return 1;
1.80 schwarze 2562: case 'H': /* Fixed horizontal resolution. */
1.149 schwarze 2563: return 24;
1.80 schwarze 2564: case 'j': /* Always adjust left margin only. */
1.149 schwarze 2565: return 0;
1.80 schwarze 2566: case 'T': /* Some output device is always defined. */
1.149 schwarze 2567: return 1;
1.80 schwarze 2568: case 'V': /* Fixed vertical resolution. */
1.149 schwarze 2569: return 40;
1.65 schwarze 2570: default:
1.149 schwarze 2571: return -1;
1.65 schwarze 2572: }
2573: }
2574:
1.53 schwarze 2575: int
1.52 schwarze 2576: roff_getreg(const struct roff *r, const char *name)
1.41 schwarze 2577: {
1.52 schwarze 2578: struct roffreg *reg;
1.65 schwarze 2579: int val;
2580:
2581: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.145 schwarze 2582: val = roff_getregro(r, name + 1);
1.65 schwarze 2583: if (-1 != val)
1.149 schwarze 2584: return val;
1.65 schwarze 2585: }
1.52 schwarze 2586:
2587: for (reg = r->regtab; reg; reg = reg->next)
2588: if (0 == strcmp(name, reg->key.p))
1.149 schwarze 2589: return reg->val;
1.53 schwarze 2590:
1.149 schwarze 2591: return 0;
1.53 schwarze 2592: }
2593:
2594: static int
2595: roff_getregn(const struct roff *r, const char *name, size_t len)
2596: {
2597: struct roffreg *reg;
1.65 schwarze 2598: int val;
2599:
2600: if ('.' == name[0] && 2 == len) {
1.145 schwarze 2601: val = roff_getregro(r, name + 1);
1.65 schwarze 2602: if (-1 != val)
1.149 schwarze 2603: return val;
1.65 schwarze 2604: }
1.53 schwarze 2605:
2606: for (reg = r->regtab; reg; reg = reg->next)
2607: if (len == reg->key.sz &&
2608: 0 == strncmp(name, reg->key.p, len))
1.149 schwarze 2609: return reg->val;
1.143 schwarze 2610:
1.149 schwarze 2611: return 0;
1.143 schwarze 2612: }
2613:
2614: static int
2615: roff_hasregn(const struct roff *r, const char *name, size_t len)
2616: {
2617: struct roffreg *reg;
2618: int val;
2619:
2620: if ('.' == name[0] && 2 == len) {
1.145 schwarze 2621: val = roff_getregro(r, name + 1);
1.143 schwarze 2622: if (-1 != val)
1.149 schwarze 2623: return 1;
1.143 schwarze 2624: }
2625:
2626: for (reg = r->regtab; reg; reg = reg->next)
2627: if (len == reg->key.sz &&
2628: 0 == strncmp(name, reg->key.p, len))
1.149 schwarze 2629: return 1;
1.41 schwarze 2630:
1.149 schwarze 2631: return 0;
1.41 schwarze 2632: }
2633:
1.52 schwarze 2634: static void
2635: roff_freereg(struct roffreg *reg)
1.41 schwarze 2636: {
1.52 schwarze 2637: struct roffreg *old_reg;
1.41 schwarze 2638:
1.52 schwarze 2639: while (NULL != reg) {
2640: free(reg->key.p);
2641: old_reg = reg;
2642: reg = reg->next;
2643: free(old_reg);
2644: }
1.41 schwarze 2645: }
1.7 schwarze 2646:
2647: static enum rofferr
1.6 schwarze 2648: roff_nr(ROFF_ARGS)
1.1 schwarze 2649: {
1.85 schwarze 2650: char *key, *val;
2651: size_t keysz;
1.37 schwarze 2652: int iv;
1.60 schwarze 2653: char sign;
1.6 schwarze 2654:
1.110 schwarze 2655: key = val = buf->buf + pos;
2656: if (*key == '\0')
1.149 schwarze 2657: return ROFF_IGN;
1.85 schwarze 2658:
2659: keysz = roff_getname(r, &val, ln, pos);
1.110 schwarze 2660: if (key[keysz] == '\\')
1.149 schwarze 2661: return ROFF_IGN;
1.85 schwarze 2662: key[keysz] = '\0';
1.6 schwarze 2663:
1.60 schwarze 2664: sign = *val;
1.110 schwarze 2665: if (sign == '+' || sign == '-')
1.60 schwarze 2666: val++;
2667:
1.133 schwarze 2668: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77 schwarze 2669: roff_setreg(r, key, iv, sign);
1.1 schwarze 2670:
1.149 schwarze 2671: return ROFF_IGN;
1.76 schwarze 2672: }
2673:
2674: static enum rofferr
2675: roff_rr(ROFF_ARGS)
2676: {
2677: struct roffreg *reg, **prev;
1.85 schwarze 2678: char *name, *cp;
2679: size_t namesz;
1.76 schwarze 2680:
1.110 schwarze 2681: name = cp = buf->buf + pos;
2682: if (*name == '\0')
1.149 schwarze 2683: return ROFF_IGN;
1.85 schwarze 2684: namesz = roff_getname(r, &cp, ln, pos);
2685: name[namesz] = '\0';
1.76 schwarze 2686:
2687: prev = &r->regtab;
2688: while (1) {
2689: reg = *prev;
1.110 schwarze 2690: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 2691: break;
2692: prev = ®->next;
2693: }
1.110 schwarze 2694: if (reg != NULL) {
1.76 schwarze 2695: *prev = reg->next;
2696: free(reg->key.p);
2697: free(reg);
2698: }
1.149 schwarze 2699: return ROFF_IGN;
1.29 schwarze 2700: }
2701:
1.138 schwarze 2702: /* --- handler functions for roff requests -------------------------------- */
2703:
1.29 schwarze 2704: static enum rofferr
2705: roff_rm(ROFF_ARGS)
2706: {
2707: const char *name;
2708: char *cp;
1.85 schwarze 2709: size_t namesz;
1.29 schwarze 2710:
1.110 schwarze 2711: cp = buf->buf + pos;
2712: while (*cp != '\0') {
1.85 schwarze 2713: name = cp;
1.110 schwarze 2714: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 2715: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110 schwarze 2716: if (name[namesz] == '\\')
1.85 schwarze 2717: break;
1.29 schwarze 2718: }
1.149 schwarze 2719: return ROFF_IGN;
1.51 schwarze 2720: }
2721:
2722: static enum rofferr
2723: roff_it(ROFF_ARGS)
2724: {
2725: int iv;
2726:
2727: /* Parse the number of lines. */
1.133 schwarze 2728:
2729: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95 schwarze 2730: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110 schwarze 2731: ln, ppos, buf->buf + 1);
1.149 schwarze 2732: return ROFF_IGN;
1.51 schwarze 2733: }
2734:
1.134 schwarze 2735: while (isspace((unsigned char)buf->buf[pos]))
2736: pos++;
2737:
2738: /*
2739: * Arm the input line trap.
2740: * Special-casing "an-trap" is an ugly workaround to cope
2741: * with DocBook stupidly fiddling with man(7) internals.
2742: */
1.133 schwarze 2743:
1.51 schwarze 2744: roffit_lines = iv;
1.134 schwarze 2745: roffit_macro = mandoc_strdup(iv != 1 ||
2746: strcmp(buf->buf + pos, "an-trap") ?
2747: buf->buf + pos : "br");
1.149 schwarze 2748: return ROFF_IGN;
1.47 schwarze 2749: }
2750:
2751: static enum rofferr
2752: roff_Dd(ROFF_ARGS)
2753: {
2754: const char *const *cp;
2755:
1.99 schwarze 2756: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47 schwarze 2757: for (cp = __mdoc_reserved; *cp; cp++)
2758: roff_setstr(r, *cp, NULL, 0);
2759:
1.99 schwarze 2760: if (r->format == 0)
2761: r->format = MPARSE_MDOC;
2762:
1.149 schwarze 2763: return ROFF_CONT;
1.47 schwarze 2764: }
2765:
2766: static enum rofferr
2767: roff_TH(ROFF_ARGS)
2768: {
2769: const char *const *cp;
2770:
1.99 schwarze 2771: if ((r->options & MPARSE_QUICK) == 0)
1.47 schwarze 2772: for (cp = __man_reserved; *cp; cp++)
2773: roff_setstr(r, *cp, NULL, 0);
2774:
1.99 schwarze 2775: if (r->format == 0)
2776: r->format = MPARSE_MAN;
2777:
1.149 schwarze 2778: return ROFF_CONT;
1.14 schwarze 2779: }
2780:
2781: static enum rofferr
1.27 schwarze 2782: roff_TE(ROFF_ARGS)
2783: {
2784:
2785: if (NULL == r->tbl)
1.94 schwarze 2786: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2787: ln, ppos, "TE");
1.130 schwarze 2788: else if ( ! tbl_end(&r->tbl)) {
2789: free(buf->buf);
2790: buf->buf = mandoc_strdup(".sp");
2791: buf->sz = 4;
1.149 schwarze 2792: return ROFF_REPARSE;
1.130 schwarze 2793: }
1.149 schwarze 2794: return ROFF_IGN;
1.27 schwarze 2795: }
2796:
2797: static enum rofferr
2798: roff_T_(ROFF_ARGS)
2799: {
2800:
2801: if (NULL == r->tbl)
1.94 schwarze 2802: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2803: ln, ppos, "T&");
1.27 schwarze 2804: else
2805: tbl_restart(ppos, ln, r->tbl);
2806:
1.149 schwarze 2807: return ROFF_IGN;
1.27 schwarze 2808: }
2809:
1.102 schwarze 2810: /*
2811: * Handle in-line equation delimiters.
2812: */
2813: static enum rofferr
1.110 schwarze 2814: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 2815: {
1.105 schwarze 2816: char *cp1, *cp2;
2817: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 2818:
1.102 schwarze 2819: /*
2820: * Outside equations, look for an opening delimiter.
2821: * If we are inside an equation, we already know it is
2822: * in-line, or this function wouldn't have been called;
2823: * so look for a closing delimiter.
2824: */
2825:
1.110 schwarze 2826: cp1 = buf->buf + pos;
1.102 schwarze 2827: cp2 = strchr(cp1, r->eqn == NULL ?
2828: r->last_eqn->odelim : r->last_eqn->cdelim);
2829: if (cp2 == NULL)
1.149 schwarze 2830: return ROFF_CONT;
1.102 schwarze 2831:
1.105 schwarze 2832: *cp2++ = '\0';
2833: bef_pr = bef_nl = aft_nl = aft_pr = "";
2834:
2835: /* Handle preceding text, protecting whitespace. */
2836:
1.110 schwarze 2837: if (*buf->buf != '\0') {
1.105 schwarze 2838: if (r->eqn == NULL)
2839: bef_pr = "\\&";
2840: bef_nl = "\n";
2841: }
2842:
2843: /*
2844: * Prepare replacing the delimiter with an equation macro
2845: * and drop leading white space from the equation.
2846: */
1.102 schwarze 2847:
1.105 schwarze 2848: if (r->eqn == NULL) {
2849: while (*cp2 == ' ')
2850: cp2++;
2851: mac = ".EQ";
2852: } else
2853: mac = ".EN";
2854:
2855: /* Handle following text, protecting whitespace. */
2856:
2857: if (*cp2 != '\0') {
2858: aft_nl = "\n";
2859: if (r->eqn != NULL)
2860: aft_pr = "\\&";
2861: }
2862:
2863: /* Do the actual replacement. */
2864:
1.110 schwarze 2865: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 2866: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 2867: free(buf->buf);
2868: buf->buf = cp1;
1.102 schwarze 2869:
2870: /* Toggle the in-line state of the eqn subsystem. */
2871:
2872: r->eqn_inline = r->eqn == NULL;
1.149 schwarze 2873: return ROFF_REPARSE;
1.41 schwarze 2874: }
2875:
1.107 schwarze 2876: static enum rofferr
2877: roff_EQ(ROFF_ARGS)
1.32 schwarze 2878: {
1.41 schwarze 2879: struct eqn_node *e;
1.32 schwarze 2880:
1.110 schwarze 2881: assert(r->eqn == NULL);
1.107 schwarze 2882: e = eqn_alloc(ppos, ln, r->parse);
1.32 schwarze 2883:
1.102 schwarze 2884: if (r->last_eqn) {
1.32 schwarze 2885: r->last_eqn->next = e;
1.102 schwarze 2886: e->delim = r->last_eqn->delim;
2887: e->odelim = r->last_eqn->odelim;
2888: e->cdelim = r->last_eqn->cdelim;
2889: } else
1.32 schwarze 2890: r->first_eqn = r->last_eqn = e;
2891:
2892: r->eqn = r->last_eqn = e;
1.41 schwarze 2893:
1.110 schwarze 2894: if (buf->buf[pos] != '\0')
1.107 schwarze 2895: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 2896: ".EQ %s", buf->buf + pos);
1.41 schwarze 2897:
1.149 schwarze 2898: return ROFF_IGN;
1.32 schwarze 2899: }
2900:
2901: static enum rofferr
2902: roff_EN(ROFF_ARGS)
2903: {
2904:
1.94 schwarze 2905: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.149 schwarze 2906: return ROFF_IGN;
1.32 schwarze 2907: }
2908:
2909: static enum rofferr
1.27 schwarze 2910: roff_TS(ROFF_ARGS)
2911: {
1.49 schwarze 2912: struct tbl_node *tbl;
1.27 schwarze 2913:
2914: if (r->tbl) {
1.94 schwarze 2915: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2916: ln, ppos, "TS breaks TS");
1.41 schwarze 2917: tbl_end(&r->tbl);
1.27 schwarze 2918: }
2919:
1.49 schwarze 2920: tbl = tbl_alloc(ppos, ln, r->parse);
1.27 schwarze 2921:
2922: if (r->last_tbl)
1.49 schwarze 2923: r->last_tbl->next = tbl;
1.27 schwarze 2924: else
1.49 schwarze 2925: r->first_tbl = r->last_tbl = tbl;
1.27 schwarze 2926:
1.49 schwarze 2927: r->tbl = r->last_tbl = tbl;
1.149 schwarze 2928: return ROFF_IGN;
1.123 schwarze 2929: }
2930:
2931: static enum rofferr
2932: roff_brp(ROFF_ARGS)
2933: {
2934:
2935: buf->buf[pos - 1] = '\0';
1.149 schwarze 2936: return ROFF_CONT;
1.27 schwarze 2937: }
2938:
2939: static enum rofferr
1.48 schwarze 2940: roff_cc(ROFF_ARGS)
2941: {
2942: const char *p;
2943:
1.110 schwarze 2944: p = buf->buf + pos;
1.48 schwarze 2945:
1.110 schwarze 2946: if (*p == '\0' || (r->control = *p++) == '.')
1.48 schwarze 2947: r->control = 0;
2948:
1.110 schwarze 2949: if (*p != '\0')
1.132 schwarze 2950: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2951: ln, p - buf->buf, "cc ... %s", p);
1.48 schwarze 2952:
1.149 schwarze 2953: return ROFF_IGN;
1.48 schwarze 2954: }
2955:
2956: static enum rofferr
1.42 schwarze 2957: roff_tr(ROFF_ARGS)
2958: {
2959: const char *p, *first, *second;
2960: size_t fsz, ssz;
2961: enum mandoc_esc esc;
2962:
1.110 schwarze 2963: p = buf->buf + pos;
1.42 schwarze 2964:
1.110 schwarze 2965: if (*p == '\0') {
1.132 schwarze 2966: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149 schwarze 2967: return ROFF_IGN;
1.42 schwarze 2968: }
2969:
1.110 schwarze 2970: while (*p != '\0') {
1.42 schwarze 2971: fsz = ssz = 1;
2972:
2973: first = p++;
1.110 schwarze 2974: if (*first == '\\') {
1.42 schwarze 2975: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2976: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2977: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2978: ln, (int)(p - buf->buf), first);
1.149 schwarze 2979: return ROFF_IGN;
1.42 schwarze 2980: }
2981: fsz = (size_t)(p - first);
2982: }
2983:
2984: second = p++;
1.110 schwarze 2985: if (*second == '\\') {
1.42 schwarze 2986: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2987: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2988: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2989: ln, (int)(p - buf->buf), second);
1.149 schwarze 2990: return ROFF_IGN;
1.42 schwarze 2991: }
2992: ssz = (size_t)(p - second);
1.110 schwarze 2993: } else if (*second == '\0') {
1.132 schwarze 2994: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2995: ln, first - buf->buf, "tr %s", first);
1.42 schwarze 2996: second = " ";
2997: p--;
2998: }
2999:
3000: if (fsz > 1) {
1.80 schwarze 3001: roff_setstrn(&r->xmbtab, first, fsz,
3002: second, ssz, 0);
1.42 schwarze 3003: continue;
3004: }
3005:
1.110 schwarze 3006: if (r->xtab == NULL)
1.80 schwarze 3007: r->xtab = mandoc_calloc(128,
3008: sizeof(struct roffstr));
1.42 schwarze 3009:
3010: free(r->xtab[(int)*first].p);
3011: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3012: r->xtab[(int)*first].sz = ssz;
3013: }
3014:
1.149 schwarze 3015: return ROFF_IGN;
1.42 schwarze 3016: }
3017:
3018: static enum rofferr
1.14 schwarze 3019: roff_so(ROFF_ARGS)
3020: {
1.121 schwarze 3021: char *name, *cp;
1.15 schwarze 3022:
1.110 schwarze 3023: name = buf->buf + pos;
1.97 schwarze 3024: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14 schwarze 3025:
1.22 schwarze 3026: /*
3027: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3028: * opening anything that's not in our cwd or anything beneath
3029: * it. Thus, explicitly disallow traversing up the file-system
3030: * or using absolute paths.
3031: */
3032:
1.110 schwarze 3033: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83 schwarze 3034: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3035: ".so %s", name);
1.121 schwarze 3036: buf->sz = mandoc_asprintf(&cp,
3037: ".sp\nSee the file %s.\n.sp", name) + 1;
3038: free(buf->buf);
3039: buf->buf = cp;
3040: *offs = 0;
1.149 schwarze 3041: return ROFF_REPARSE;
1.14 schwarze 3042: }
3043:
3044: *offs = pos;
1.149 schwarze 3045: return ROFF_SO;
1.7 schwarze 3046: }
3047:
1.138 schwarze 3048: /* --- user defined strings and macros ------------------------------------ */
3049:
1.16 schwarze 3050: static enum rofferr
3051: roff_userdef(ROFF_ARGS)
1.12 schwarze 3052: {
1.135 schwarze 3053: const char *arg[9], *ap;
1.16 schwarze 3054: char *cp, *n1, *n2;
1.146 schwarze 3055: int i, ib, ie;
1.135 schwarze 3056: size_t asz, rsz;
1.12 schwarze 3057:
1.16 schwarze 3058: /*
3059: * Collect pointers to macro argument strings
1.61 schwarze 3060: * and NUL-terminate them.
1.16 schwarze 3061: */
1.135 schwarze 3062:
1.145 schwarze 3063: r->argc = 0;
1.110 schwarze 3064: cp = buf->buf + pos;
1.145 schwarze 3065: for (i = 0; i < 9; i++) {
3066: if (*cp == '\0')
3067: arg[i] = "";
3068: else {
3069: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3070: r->argc = i + 1;
3071: }
3072: }
1.16 schwarze 3073:
3074: /*
3075: * Expand macro arguments.
1.12 schwarze 3076: */
1.135 schwarze 3077:
3078: buf->sz = strlen(r->current_string) + 1;
3079: n1 = cp = mandoc_malloc(buf->sz);
3080: memcpy(n1, r->current_string, buf->sz);
3081: while (*cp != '\0') {
3082:
3083: /* Scan ahead for the next argument invocation. */
3084:
3085: if (*cp++ != '\\')
3086: continue;
3087: if (*cp++ != '$')
3088: continue;
1.146 schwarze 3089: if (*cp == '*') { /* \\$* inserts all arguments */
3090: ib = 0;
3091: ie = r->argc - 1;
3092: } else { /* \\$1 .. \\$9 insert one argument */
3093: ib = ie = *cp - '1';
3094: if (ib < 0 || ib > 8)
3095: continue;
3096: }
1.135 schwarze 3097: cp -= 2;
3098:
3099: /*
3100: * Determine the size of the expanded argument,
3101: * taking escaping of quotes into account.
3102: */
3103:
1.146 schwarze 3104: asz = ie > ib ? ie - ib : 0; /* for blanks */
3105: for (i = ib; i <= ie; i++) {
3106: for (ap = arg[i]; *ap != '\0'; ap++) {
3107: asz++;
3108: if (*ap == '"')
3109: asz += 3;
3110: }
1.135 schwarze 3111: }
3112: if (asz != 3) {
3113:
3114: /*
3115: * Determine the size of the rest of the
3116: * unexpanded macro, including the NUL.
3117: */
3118:
3119: rsz = buf->sz - (cp - n1) - 3;
3120:
3121: /*
3122: * When shrinking, move before
3123: * releasing the storage.
3124: */
3125:
3126: if (asz < 3)
3127: memmove(cp + asz, cp + 3, rsz);
3128:
3129: /*
3130: * Resize the storage for the macro
3131: * and readjust the parse pointer.
3132: */
3133:
3134: buf->sz += asz - 3;
3135: n2 = mandoc_realloc(n1, buf->sz);
3136: cp = n2 + (cp - n1);
3137: n1 = n2;
3138:
3139: /*
3140: * When growing, make room
3141: * for the expanded argument.
3142: */
3143:
3144: if (asz > 3)
3145: memmove(cp + asz, cp + 3, rsz);
3146: }
3147:
3148: /* Copy the expanded argument, escaping quotes. */
3149:
3150: n2 = cp;
1.146 schwarze 3151: for (i = ib; i <= ie; i++) {
3152: for (ap = arg[i]; *ap != '\0'; ap++) {
3153: if (*ap == '"') {
3154: memcpy(n2, "\\(dq", 4);
3155: n2 += 4;
3156: } else
3157: *n2++ = *ap;
3158: }
3159: if (i < ie)
3160: *n2++ = ' ';
1.16 schwarze 3161: }
1.12 schwarze 3162: }
3163:
1.16 schwarze 3164: /*
3165: * Replace the macro invocation
3166: * by the expanded macro.
3167: */
1.135 schwarze 3168:
1.110 schwarze 3169: free(buf->buf);
3170: buf->buf = n1;
1.120 schwarze 3171: *offs = 0;
1.16 schwarze 3172:
1.149 schwarze 3173: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3174: ROFF_REPARSE : ROFF_APPEND;
1.12 schwarze 3175: }
1.28 schwarze 3176:
1.85 schwarze 3177: static size_t
1.28 schwarze 3178: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3179: {
3180: char *name, *cp;
1.85 schwarze 3181: size_t namesz;
1.28 schwarze 3182:
3183: name = *cpp;
3184: if ('\0' == *name)
1.149 schwarze 3185: return 0;
1.28 schwarze 3186:
1.85 schwarze 3187: /* Read until end of name and terminate it with NUL. */
3188: for (cp = name; 1; cp++) {
3189: if ('\0' == *cp || ' ' == *cp) {
3190: namesz = cp - name;
3191: break;
3192: }
1.28 schwarze 3193: if ('\\' != *cp)
3194: continue;
1.88 schwarze 3195: namesz = cp - name;
3196: if ('{' == cp[1] || '}' == cp[1])
3197: break;
1.28 schwarze 3198: cp++;
3199: if ('\\' == *cp)
3200: continue;
1.97 schwarze 3201: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3202: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 3203: mandoc_escape((const char **)&cp, NULL, NULL);
3204: break;
1.28 schwarze 3205: }
3206:
3207: /* Read past spaces. */
3208: while (' ' == *cp)
3209: cp++;
3210:
3211: *cpp = cp;
1.149 schwarze 3212: return namesz;
1.28 schwarze 3213: }
3214:
1.16 schwarze 3215: /*
3216: * Store *string into the user-defined string called *name.
3217: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 3218: * append == 0: replace mode
3219: * append == 1: single-line append mode
3220: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 3221: */
1.8 schwarze 3222: static void
1.16 schwarze 3223: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 3224: int append)
1.7 schwarze 3225: {
1.42 schwarze 3226:
3227: roff_setstrn(&r->strtab, name, strlen(name), string,
1.80 schwarze 3228: string ? strlen(string) : 0, append);
1.42 schwarze 3229: }
3230:
3231: static void
3232: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 3233: const char *string, size_t stringsz, int append)
1.42 schwarze 3234: {
3235: struct roffkv *n;
3236: char *c;
3237: int i;
3238: size_t oldch, newch;
1.7 schwarze 3239:
1.16 schwarze 3240: /* Search for an existing string with the same name. */
1.42 schwarze 3241: n = *r;
3242:
1.84 schwarze 3243: while (n && (namesz != n->key.sz ||
3244: strncmp(n->key.p, name, namesz)))
1.7 schwarze 3245: n = n->next;
1.8 schwarze 3246:
3247: if (NULL == n) {
1.16 schwarze 3248: /* Create a new string table entry. */
1.42 schwarze 3249: n = mandoc_malloc(sizeof(struct roffkv));
3250: n->key.p = mandoc_strndup(name, namesz);
3251: n->key.sz = namesz;
3252: n->val.p = NULL;
3253: n->val.sz = 0;
3254: n->next = *r;
3255: *r = n;
1.66 schwarze 3256: } else if (0 == append) {
1.42 schwarze 3257: free(n->val.p);
3258: n->val.p = NULL;
3259: n->val.sz = 0;
1.16 schwarze 3260: }
3261:
3262: if (NULL == string)
3263: return;
3264:
3265: /*
3266: * One additional byte for the '\n' in multiline mode,
3267: * and one for the terminating '\0'.
3268: */
1.66 schwarze 3269: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 3270:
3271: if (NULL == n->val.p) {
3272: n->val.p = mandoc_malloc(newch);
3273: *n->val.p = '\0';
1.16 schwarze 3274: oldch = 0;
3275: } else {
1.42 schwarze 3276: oldch = n->val.sz;
3277: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 3278: }
3279:
3280: /* Skip existing content in the destination buffer. */
1.42 schwarze 3281: c = n->val.p + (int)oldch;
1.16 schwarze 3282:
3283: /* Append new content to the destination buffer. */
1.42 schwarze 3284: i = 0;
3285: while (i < (int)stringsz) {
1.16 schwarze 3286: /*
3287: * Rudimentary roff copy mode:
3288: * Handle escaped backslashes.
3289: */
1.42 schwarze 3290: if ('\\' == string[i] && '\\' == string[i + 1])
3291: i++;
3292: *c++ = string[i++];
1.16 schwarze 3293: }
1.8 schwarze 3294:
1.16 schwarze 3295: /* Append terminating bytes. */
1.66 schwarze 3296: if (1 < append)
1.16 schwarze 3297: *c++ = '\n';
1.42 schwarze 3298:
1.16 schwarze 3299: *c = '\0';
1.42 schwarze 3300: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 3301: }
3302:
1.8 schwarze 3303: static const char *
3304: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7 schwarze 3305: {
1.42 schwarze 3306: const struct roffkv *n;
1.64 schwarze 3307: int i;
1.7 schwarze 3308:
1.42 schwarze 3309: for (n = r->strtab; n; n = n->next)
1.80 schwarze 3310: if (0 == strncmp(name, n->key.p, len) &&
3311: '\0' == n->key.p[(int)len])
1.149 schwarze 3312: return n->val.p;
1.64 schwarze 3313:
3314: for (i = 0; i < PREDEFS_MAX; i++)
3315: if (0 == strncmp(name, predefs[i].name, len) &&
3316: '\0' == predefs[i].name[(int)len])
1.149 schwarze 3317: return predefs[i].str;
1.8 schwarze 3318:
1.149 schwarze 3319: return NULL;
1.7 schwarze 3320: }
3321:
1.8 schwarze 3322: static void
1.42 schwarze 3323: roff_freestr(struct roffkv *r)
1.7 schwarze 3324: {
1.42 schwarze 3325: struct roffkv *n, *nn;
1.7 schwarze 3326:
1.42 schwarze 3327: for (n = r; n; n = nn) {
3328: free(n->key.p);
3329: free(n->val.p);
1.7 schwarze 3330: nn = n->next;
3331: free(n);
3332: }
1.27 schwarze 3333: }
1.138 schwarze 3334:
3335: /* --- accessors and utility functions ------------------------------------ */
1.27 schwarze 3336:
3337: const struct tbl_span *
3338: roff_span(const struct roff *r)
3339: {
1.80 schwarze 3340:
1.149 schwarze 3341: return r->tbl ? tbl_span(r->tbl) : NULL;
1.32 schwarze 3342: }
3343:
3344: const struct eqn *
3345: roff_eqn(const struct roff *r)
3346: {
1.80 schwarze 3347:
1.149 schwarze 3348: return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.42 schwarze 3349: }
3350:
3351: /*
3352: * Duplicate an input string, making the appropriate character
3353: * conversations (as stipulated by `tr') along the way.
3354: * Returns a heap-allocated string with all the replacements made.
3355: */
3356: char *
3357: roff_strdup(const struct roff *r, const char *p)
3358: {
3359: const struct roffkv *cp;
3360: char *res;
3361: const char *pp;
3362: size_t ssz, sz;
3363: enum mandoc_esc esc;
3364:
3365: if (NULL == r->xmbtab && NULL == r->xtab)
1.149 schwarze 3366: return mandoc_strdup(p);
1.42 schwarze 3367: else if ('\0' == *p)
1.149 schwarze 3368: return mandoc_strdup("");
1.42 schwarze 3369:
3370: /*
3371: * Step through each character looking for term matches
3372: * (remember that a `tr' can be invoked with an escape, which is
3373: * a glyph but the escape is multi-character).
3374: * We only do this if the character hash has been initialised
3375: * and the string is >0 length.
3376: */
3377:
3378: res = NULL;
3379: ssz = 0;
3380:
3381: while ('\0' != *p) {
3382: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3383: sz = r->xtab[(int)*p].sz;
3384: res = mandoc_realloc(res, ssz + sz + 1);
3385: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3386: ssz += sz;
3387: p++;
3388: continue;
3389: } else if ('\\' != *p) {
3390: res = mandoc_realloc(res, ssz + 2);
3391: res[ssz++] = *p++;
3392: continue;
3393: }
3394:
3395: /* Search for term matches. */
3396: for (cp = r->xmbtab; cp; cp = cp->next)
3397: if (0 == strncmp(p, cp->key.p, cp->key.sz))
3398: break;
3399:
3400: if (NULL != cp) {
3401: /*
3402: * A match has been found.
3403: * Append the match to the array and move
3404: * forward by its keysize.
3405: */
1.80 schwarze 3406: res = mandoc_realloc(res,
3407: ssz + cp->val.sz + 1);
1.42 schwarze 3408: memcpy(res + ssz, cp->val.p, cp->val.sz);
3409: ssz += cp->val.sz;
3410: p += (int)cp->key.sz;
3411: continue;
3412: }
3413:
3414: /*
3415: * Handle escapes carefully: we need to copy
3416: * over just the escape itself, or else we might
3417: * do replacements within the escape itself.
3418: * Make sure to pass along the bogus string.
3419: */
3420: pp = p++;
3421: esc = mandoc_escape(&p, NULL, NULL);
3422: if (ESCAPE_ERROR == esc) {
3423: sz = strlen(pp);
3424: res = mandoc_realloc(res, ssz + sz + 1);
3425: memcpy(res + ssz, pp, sz);
3426: break;
3427: }
1.80 schwarze 3428: /*
3429: * We bail out on bad escapes.
1.42 schwarze 3430: * No need to warn: we already did so when
3431: * roff_res() was called.
3432: */
3433: sz = (int)(p - pp);
3434: res = mandoc_realloc(res, ssz + sz + 1);
3435: memcpy(res + ssz, pp, sz);
3436: ssz += sz;
3437: }
3438:
3439: res[(int)ssz] = '\0';
1.149 schwarze 3440: return res;
1.99 schwarze 3441: }
3442:
3443: int
3444: roff_getformat(const struct roff *r)
3445: {
3446:
1.149 schwarze 3447: return r->format;
1.48 schwarze 3448: }
3449:
3450: /*
1.80 schwarze 3451: * Find out whether a line is a macro line or not.
1.48 schwarze 3452: * If it is, adjust the current position and return one; if it isn't,
3453: * return zero and don't change the current position.
3454: * If the control character has been set with `.cc', then let that grain
3455: * precedence.
3456: * This is slighly contrary to groff, where using the non-breaking
3457: * control character when `cc' has been invoked will cause the
3458: * non-breaking macro contents to be printed verbatim.
3459: */
3460: int
3461: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3462: {
3463: int pos;
3464:
3465: pos = *ppos;
3466:
3467: if (0 != r->control && cp[pos] == r->control)
3468: pos++;
3469: else if (0 != r->control)
1.149 schwarze 3470: return 0;
1.48 schwarze 3471: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3472: pos += 2;
3473: else if ('.' == cp[pos] || '\'' == cp[pos])
3474: pos++;
3475: else
1.149 schwarze 3476: return 0;
1.48 schwarze 3477:
3478: while (' ' == cp[pos] || '\t' == cp[pos])
3479: pos++;
3480:
3481: *ppos = pos;
1.149 schwarze 3482: return 1;
1.1 schwarze 3483: }
![[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