Annotation of src/usr.bin/mandoc/roff.c, Revision 1.139
1.139 ! schwarze 1: /* $OpenBSD: roff.c,v 1.138 2015/04/19 13:50:10 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 */
1.109 schwarze 317: const struct mchars *mchars; /* character table */
1.1 schwarze 318: struct roffnode *last; /* leaf of stack */
1.96 schwarze 319: int *rstack; /* stack of inverted `ie' values */
1.52 schwarze 320: struct roffreg *regtab; /* number registers */
1.42 schwarze 321: struct roffkv *strtab; /* user-defined strings & macros */
322: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
323: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.16 schwarze 324: const char *current_string; /* value of last called user macro */
1.27 schwarze 325: struct tbl_node *first_tbl; /* first table parsed */
326: struct tbl_node *last_tbl; /* last table parsed */
327: struct tbl_node *tbl; /* current table being parsed */
1.32 schwarze 328: struct eqn_node *last_eqn; /* last equation parsed */
329: struct eqn_node *first_eqn; /* first equation parsed */
330: struct eqn_node *eqn; /* current equation being parsed */
1.102 schwarze 331: int eqn_inline; /* current equation is inline */
1.96 schwarze 332: int options; /* parse options */
333: int rstacksz; /* current size limit of rstack */
334: int rstackpos; /* position in rstack */
1.99 schwarze 335: int format; /* current file in mdoc or man format */
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);
1.106 schwarze 398: static int roff_evalcond(struct roff *r, int,
399: const char *, int *);
400: static int roff_evalnum(struct roff *, int,
401: const char *, int *, int *, int);
402: static int roff_evalpar(struct roff *, int,
1.133 schwarze 403: const char *, int *, int *, int);
1.71 schwarze 404: static int roff_evalstrcond(const char *, int *);
1.42 schwarze 405: static void roff_free1(struct roff *);
1.52 schwarze 406: static void roff_freereg(struct roffreg *);
1.42 schwarze 407: static void roff_freestr(struct roffkv *);
1.85 schwarze 408: static size_t roff_getname(struct roff *, char **, int, int);
1.133 schwarze 409: static int roff_getnum(const char *, int *, int *, int);
1.56 schwarze 410: static int roff_getop(const char *, int *, char *);
1.53 schwarze 411: static int roff_getregn(const struct roff *,
412: const char *, size_t);
1.65 schwarze 413: static int roff_getregro(const char *name);
1.80 schwarze 414: static const char *roff_getstrn(const struct roff *,
1.8 schwarze 415: const char *, size_t);
1.123 schwarze 416: static enum rofferr roff_insec(ROFF_ARGS);
1.51 schwarze 417: static enum rofferr roff_it(ROFF_ARGS);
1.21 schwarze 418: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.137 schwarze 419: static void roff_man_alloc1(struct roff_man *);
420: static void roff_man_free1(struct roff_man *);
1.6 schwarze 421: static enum rofferr roff_nr(ROFF_ARGS);
1.87 schwarze 422: static enum rofft roff_parse(struct roff *, char *, int *,
423: int, int);
1.110 schwarze 424: static enum rofferr roff_parsetext(struct buf *, int, int *);
425: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.29 schwarze 426: static enum rofferr roff_rm(ROFF_ARGS);
1.76 schwarze 427: static enum rofferr roff_rr(ROFF_ARGS);
1.8 schwarze 428: static void roff_setstr(struct roff *,
1.16 schwarze 429: const char *, const char *, int);
1.80 schwarze 430: static void roff_setstrn(struct roffkv **, const char *,
1.42 schwarze 431: size_t, const char *, size_t, int);
1.14 schwarze 432: static enum rofferr roff_so(ROFF_ARGS);
1.42 schwarze 433: static enum rofferr roff_tr(ROFF_ARGS);
1.47 schwarze 434: static enum rofferr roff_Dd(ROFF_ARGS);
435: static enum rofferr roff_TH(ROFF_ARGS);
1.27 schwarze 436: static enum rofferr roff_TE(ROFF_ARGS);
437: static enum rofferr roff_TS(ROFF_ARGS);
1.32 schwarze 438: static enum rofferr roff_EQ(ROFF_ARGS);
439: static enum rofferr roff_EN(ROFF_ARGS);
1.27 schwarze 440: static enum rofferr roff_T_(ROFF_ARGS);
1.123 schwarze 441: static enum rofferr roff_unsupp(ROFF_ARGS);
1.16 schwarze 442: static enum rofferr roff_userdef(ROFF_ARGS);
1.1 schwarze 443:
1.138 schwarze 444: /* --- constant data ------------------------------------------------------ */
445:
1.42 schwarze 446: /* See roffhash_find() */
1.3 schwarze 447:
448: #define ASCII_HI 126
449: #define ASCII_LO 33
450: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
451:
1.133 schwarze 452: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
453: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
454:
1.3 schwarze 455: static struct roffmac *hash[HASHWIDTH];
456:
457: static struct roffmac roffs[ROFF_MAX] = {
1.123 schwarze 458: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.21 schwarze 459: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 460: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
461: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
462: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 463: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 464: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.3 schwarze 465: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 466: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.66 schwarze 467: { "as", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 468: { "as1", roff_ds, NULL, NULL, 0, NULL },
469: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
470: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
471: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
472: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
473: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
474: { "box", roff_unsupp, NULL, NULL, 0, NULL },
475: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
476: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
477: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
478: { "break", roff_unsupp, NULL, NULL, 0, NULL },
479: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
480: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
481: { "brp", roff_brp, NULL, NULL, 0, NULL },
482: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
483: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.48 schwarze 484: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.67 schwarze 485: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 486: { "cf", roff_insec, NULL, NULL, 0, NULL },
487: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
488: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
489: { "char", roff_unsupp, NULL, NULL, 0, NULL },
490: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
491: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
492: { "close", roff_insec, NULL, NULL, 0, NULL },
493: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
494: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
495: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
496: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
497: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
498: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
499: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
500: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
501: { "da", roff_unsupp, NULL, NULL, 0, NULL },
502: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
503: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.3 schwarze 504: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 505: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
506: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 507: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 508: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
509: { "device", roff_unsupp, NULL, NULL, 0, NULL },
510: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
511: { "di", roff_unsupp, NULL, NULL, 0, NULL },
512: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.7 schwarze 513: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 514: { "ds1", roff_ds, NULL, NULL, 0, NULL },
515: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
516: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
517: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
518: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
519: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 520: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.123 schwarze 521: { "em", roff_unsupp, NULL, NULL, 0, NULL },
522: { "EN", roff_EN, NULL, NULL, 0, NULL },
523: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
524: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
525: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
526: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
527: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
528: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
529: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
530: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.58 schwarze 531: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 532: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
533: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
534: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
535: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
536: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
538: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
543: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
544: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
545: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
551: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
552: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
553: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
554: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
555: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.59 schwarze 556: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 557: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 558: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
559: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
560: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
561: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
562: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 563: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
564: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
565: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 566: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.51 schwarze 567: { "it", roff_it, NULL, NULL, 0, NULL },
1.123 schwarze 568: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
569: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
570: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
571: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
572: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
573: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
574: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
575: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
576: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
577: { "length", roff_unsupp, NULL, NULL, 0, NULL },
578: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
579: { "lf", roff_insec, NULL, NULL, 0, NULL },
580: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
581: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
582: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
583: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
584: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
585: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
586: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
587: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
588: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
589: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
590: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
591: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
592: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
593: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.128 schwarze 594: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 595: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
596: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 597: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
598: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
599: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
600: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.14 schwarze 601: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.123 schwarze 602: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
603: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 604: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 605: { "nx", roff_insec, NULL, NULL, 0, NULL },
606: { "open", roff_insec, NULL, NULL, 0, NULL },
607: { "opena", roff_insec, NULL, NULL, 0, NULL },
608: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
609: { "output", roff_unsupp, NULL, NULL, 0, NULL },
610: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
611: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
612: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "pi", roff_insec, NULL, NULL, 0, NULL },
615: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.101 schwarze 616: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 617: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
618: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
619: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
620: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 621: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 622: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.127 schwarze 623: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 624: { "pso", roff_insec, NULL, NULL, 0, NULL },
625: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
626: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
627: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
628: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
629: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
630: { "return", roff_unsupp, NULL, NULL, 0, NULL },
631: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
632: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
633: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.29 schwarze 634: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.123 schwarze 635: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
636: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.76 schwarze 637: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.123 schwarze 638: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
639: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
640: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
641: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
642: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
643: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
644: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.14 schwarze 645: { "so", roff_so, NULL, NULL, 0, NULL },
1.123 schwarze 646: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
647: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
648: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
649: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
650: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
651: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
652: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
653: { "sy", roff_insec, NULL, NULL, 0, NULL },
654: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.127 schwarze 655: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 656: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
657: { "TE", roff_TE, NULL, NULL, 0, NULL },
658: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.127 schwarze 659: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 660: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
661: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
662: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
663: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
664: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.42 schwarze 665: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.123 schwarze 666: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
667: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
668: { "trf", roff_insec, NULL, NULL, 0, NULL },
669: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
670: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
671: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
672: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.27 schwarze 673: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.123 schwarze 674: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
675: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
676: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
677: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
678: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
679: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
680: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
681: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
682: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
683: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
684: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
685: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
686: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
687: { "while", roff_unsupp, NULL, NULL, 0, NULL },
688: { "write", roff_insec, NULL, NULL, 0, NULL },
689: { "writec", roff_insec, NULL, NULL, 0, NULL },
690: { "writem", roff_insec, NULL, NULL, 0, NULL },
691: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 692: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.16 schwarze 693: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1 schwarze 694: };
695:
1.73 schwarze 696: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47 schwarze 697: const char *const __mdoc_reserved[] = {
698: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
699: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
700: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
701: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73 schwarze 702: "Dt", "Dv", "Dx", "D1",
703: "Ec", "Ed", "Ef", "Ek", "El", "Em",
704: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47 schwarze 705: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73 schwarze 706: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
707: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47 schwarze 708: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73 schwarze 709: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
710: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
711: "Sc", "Sh", "Sm", "So", "Sq",
1.47 schwarze 712: "Ss", "St", "Sx", "Sy",
713: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73 schwarze 714: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47 schwarze 715: "%P", "%Q", "%R", "%T", "%U", "%V",
716: NULL
717: };
718:
1.73 schwarze 719: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47 schwarze 720: const char *const __man_reserved[] = {
1.73 schwarze 721: "AT", "B", "BI", "BR", "DT",
722: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
723: "LP", "OP", "P", "PD", "PP",
724: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
725: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47 schwarze 726: NULL
727: };
728:
1.37 schwarze 729: /* Array of injected predefined strings. */
730: #define PREDEFS_MAX 38
731: static const struct predef predefs[PREDEFS_MAX] = {
732: #include "predefs.in"
733: };
734:
1.42 schwarze 735: /* See roffhash_find() */
1.3 schwarze 736: #define ROFF_HASH(p) (p[0] - ASCII_LO)
737:
1.51 schwarze 738: static int roffit_lines; /* number of lines to delay */
739: static char *roffit_macro; /* nil-terminated macro line */
740:
1.80 schwarze 741:
1.138 schwarze 742: /* --- request table ------------------------------------------------------ */
743:
1.3 schwarze 744: static void
1.42 schwarze 745: roffhash_init(void)
1.3 schwarze 746: {
747: struct roffmac *n;
748: int buc, i;
749:
1.16 schwarze 750: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3 schwarze 751: assert(roffs[i].name[0] >= ASCII_LO);
752: assert(roffs[i].name[0] <= ASCII_HI);
753:
754: buc = ROFF_HASH(roffs[i].name);
755:
756: if (NULL != (n = hash[buc])) {
757: for ( ; n->next; n = n->next)
758: /* Do nothing. */ ;
759: n->next = &roffs[i];
760: } else
761: hash[buc] = &roffs[i];
762: }
763: }
764:
1.1 schwarze 765: /*
766: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
767: * the nil-terminated string name could be found.
768: */
769: static enum rofft
1.42 schwarze 770: roffhash_find(const char *p, size_t s)
1.1 schwarze 771: {
1.3 schwarze 772: int buc;
773: struct roffmac *n;
1.1 schwarze 774:
1.3 schwarze 775: /*
776: * libroff has an extremely simple hashtable, for the time
777: * being, which simply keys on the first character, which must
778: * be printable, then walks a chain. It works well enough until
779: * optimised.
780: */
781:
782: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
783: return(ROFF_MAX);
784:
785: buc = ROFF_HASH(p);
786:
787: if (NULL == (n = hash[buc]))
788: return(ROFF_MAX);
789: for ( ; n; n = n->next)
1.16 schwarze 790: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3 schwarze 791: return((enum rofft)(n - roffs));
1.1 schwarze 792:
793: return(ROFF_MAX);
794: }
795:
1.138 schwarze 796: /* --- stack of request blocks -------------------------------------------- */
797:
1.1 schwarze 798: /*
799: * Pop the current node off of the stack of roff instructions currently
800: * pending.
801: */
802: static void
803: roffnode_pop(struct roff *r)
804: {
805: struct roffnode *p;
806:
1.2 schwarze 807: assert(r->last);
1.80 schwarze 808: p = r->last;
1.2 schwarze 809:
810: r->last = r->last->parent;
1.16 schwarze 811: free(p->name);
812: free(p->end);
1.1 schwarze 813: free(p);
814: }
815:
816: /*
817: * Push a roff node onto the instruction stack. This must later be
818: * removed with roffnode_pop().
819: */
1.11 schwarze 820: static void
1.16 schwarze 821: roffnode_push(struct roff *r, enum rofft tok, const char *name,
822: int line, int col)
1.1 schwarze 823: {
824: struct roffnode *p;
825:
1.11 schwarze 826: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 827: p->tok = tok;
1.16 schwarze 828: if (name)
829: p->name = mandoc_strdup(name);
1.1 schwarze 830: p->parent = r->last;
831: p->line = line;
832: p->col = col;
1.71 schwarze 833: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 834:
835: r->last = p;
836: }
837:
1.138 schwarze 838: /* --- roff parser state data management ---------------------------------- */
839:
1.1 schwarze 840: static void
841: roff_free1(struct roff *r)
842: {
1.49 schwarze 843: struct tbl_node *tbl;
1.32 schwarze 844: struct eqn_node *e;
1.42 schwarze 845: int i;
1.27 schwarze 846:
1.49 schwarze 847: while (NULL != (tbl = r->first_tbl)) {
848: r->first_tbl = tbl->next;
849: tbl_free(tbl);
1.27 schwarze 850: }
851: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 852:
1.32 schwarze 853: while (NULL != (e = r->first_eqn)) {
854: r->first_eqn = e->next;
855: eqn_free(e);
856: }
857: r->first_eqn = r->last_eqn = r->eqn = NULL;
858:
1.1 schwarze 859: while (r->last)
860: roffnode_pop(r);
1.27 schwarze 861:
1.96 schwarze 862: free (r->rstack);
863: r->rstack = NULL;
864: r->rstacksz = 0;
865: r->rstackpos = -1;
866:
867: roff_freereg(r->regtab);
868: r->regtab = NULL;
869:
1.42 schwarze 870: roff_freestr(r->strtab);
871: roff_freestr(r->xmbtab);
872: r->strtab = r->xmbtab = NULL;
873:
874: if (r->xtab)
875: for (i = 0; i < 128; i++)
876: free(r->xtab[i].p);
877: free(r->xtab);
878: r->xtab = NULL;
1.1 schwarze 879: }
880:
881: void
882: roff_reset(struct roff *r)
883: {
884:
885: roff_free1(r);
1.99 schwarze 886: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48 schwarze 887: r->control = 0;
1.1 schwarze 888: }
889:
890: void
891: roff_free(struct roff *r)
892: {
893:
894: roff_free1(r);
895: free(r);
896: }
897:
898: struct roff *
1.109 schwarze 899: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.1 schwarze 900: {
901: struct roff *r;
902:
1.11 schwarze 903: r = mandoc_calloc(1, sizeof(struct roff));
1.35 schwarze 904: r->parse = parse;
1.109 schwarze 905: r->mchars = mchars;
1.72 schwarze 906: r->options = options;
1.99 schwarze 907: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2 schwarze 908: r->rstackpos = -1;
1.80 schwarze 909:
1.42 schwarze 910: roffhash_init();
1.37 schwarze 911:
1.1 schwarze 912: return(r);
1.137 schwarze 913: }
914:
1.138 schwarze 915: /* --- syntax tree state data management ---------------------------------- */
916:
1.137 schwarze 917: static void
918: roff_man_free1(struct roff_man *man)
919: {
920:
1.138 schwarze 921: if (man->first != NULL)
922: roff_node_delete(man, man->first);
1.137 schwarze 923: free(man->meta.msec);
924: free(man->meta.vol);
925: free(man->meta.os);
926: free(man->meta.arch);
927: free(man->meta.title);
928: free(man->meta.name);
929: free(man->meta.date);
930: }
931:
932: static void
933: roff_man_alloc1(struct roff_man *man)
934: {
935:
936: memset(&man->meta, 0, sizeof(man->meta));
937: man->first = mandoc_calloc(1, sizeof(*man->first));
938: man->first->type = ROFFT_ROOT;
939: man->last = man->first;
940: man->last_es = NULL;
941: man->flags = 0;
942: man->macroset = MACROSET_NONE;
943: man->lastsec = man->lastnamed = SEC_NONE;
944: man->next = ROFF_NEXT_CHILD;
945: }
946:
947: void
948: roff_man_reset(struct roff_man *man)
949: {
950:
951: roff_man_free1(man);
952: roff_man_alloc1(man);
953: }
954:
955: void
956: roff_man_free(struct roff_man *man)
957: {
958:
959: roff_man_free1(man);
960: free(man);
961: }
962:
963: struct roff_man *
964: roff_man_alloc(struct roff *roff, struct mparse *parse,
965: const char *defos, int quick)
966: {
967: struct roff_man *man;
968:
969: man = mandoc_calloc(1, sizeof(*man));
970: man->parse = parse;
971: man->roff = roff;
972: man->defos = defos;
973: man->quick = quick;
974: roff_man_alloc1(man);
975: return(man);
1.1 schwarze 976: }
977:
1.138 schwarze 978: /* --- syntax tree handling ----------------------------------------------- */
979:
980: struct roff_node *
981: roff_node_alloc(struct roff_man *man, int line, int pos,
982: enum roff_type type, int tok)
983: {
984: struct roff_node *n;
985:
986: n = mandoc_calloc(1, sizeof(*n));
987: n->line = line;
988: n->pos = pos;
989: n->tok = tok;
990: n->type = type;
991: n->sec = man->lastsec;
992:
993: if (man->flags & MDOC_SYNOPSIS)
994: n->flags |= MDOC_SYNPRETTY;
995: else
996: n->flags &= ~MDOC_SYNPRETTY;
997: if (man->flags & MDOC_NEWLINE)
998: n->flags |= MDOC_LINE;
999: man->flags &= ~MDOC_NEWLINE;
1000:
1001: return(n);
1002: }
1003:
1004: void
1005: roff_node_append(struct roff_man *man, struct roff_node *n)
1006: {
1007:
1008: switch (man->next) {
1009: case ROFF_NEXT_SIBLING:
1010: man->last->next = n;
1011: n->prev = man->last;
1012: n->parent = man->last->parent;
1013: break;
1014: case ROFF_NEXT_CHILD:
1015: man->last->child = n;
1016: n->parent = man->last;
1017: break;
1018: default:
1019: abort();
1020: /* NOTREACHED */
1021: }
1022: n->parent->nchild++;
1023:
1024: /*
1025: * Copy over the normalised-data pointer of our parent. Not
1026: * everybody has one, but copying a null pointer is fine.
1027: */
1028:
1029: switch (n->type) {
1030: case ROFFT_BODY:
1031: if (n->end != ENDBODY_NOT)
1032: break;
1033: /* FALLTHROUGH */
1034: case ROFFT_TAIL:
1035: /* FALLTHROUGH */
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;
! 1095: }
! 1096:
1.138 schwarze 1097: struct roff_node *
1098: roff_head_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_HEAD, tok);
1103: roff_node_append(man, n);
1104: man->next = ROFF_NEXT_CHILD;
1105: return(n);
1106: }
1107:
1108: struct roff_node *
1109: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1110: {
1111: struct roff_node *n;
1112:
1113: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1114: roff_node_append(man, n);
1115: man->next = ROFF_NEXT_CHILD;
1116: return(n);
1.139 ! schwarze 1117: }
! 1118:
! 1119: void
! 1120: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
! 1121: {
! 1122: struct roff_node *n;
! 1123:
! 1124: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
! 1125: n->eqn = eqn;
! 1126: if (eqn->ln > man->last->line)
! 1127: n->flags |= MDOC_LINE;
! 1128: roff_node_append(man, n);
! 1129: man->next = ROFF_NEXT_SIBLING;
! 1130: }
! 1131:
! 1132: void
! 1133: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
! 1134: {
! 1135: struct roff_node *n;
! 1136:
! 1137: if (man->macroset == MACROSET_MAN)
! 1138: man_breakscope(man, TOKEN_NONE);
! 1139: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
! 1140: n->span = tbl;
! 1141: roff_node_append(man, n);
! 1142: if (man->macroset == MACROSET_MDOC)
! 1143: mdoc_valid_post(man);
! 1144: else
! 1145: man_valid_post(man);
! 1146: man->next = ROFF_NEXT_SIBLING;
1.138 schwarze 1147: }
1148:
1149: void
1150: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1151: {
1152:
1153: /* Adjust siblings. */
1154:
1155: if (n->prev)
1156: n->prev->next = n->next;
1157: if (n->next)
1158: n->next->prev = n->prev;
1159:
1160: /* Adjust parent. */
1161:
1162: if (n->parent != NULL) {
1163: n->parent->nchild--;
1164: if (n->parent->child == n)
1165: n->parent->child = n->next;
1166: if (n->parent->last == n)
1167: n->parent->last = n->prev;
1168: }
1169:
1170: /* Adjust parse point. */
1171:
1172: if (man == NULL)
1173: return;
1174: if (man->last == n) {
1175: if (n->prev == NULL) {
1176: man->last = n->parent;
1177: man->next = ROFF_NEXT_CHILD;
1178: } else {
1179: man->last = n->prev;
1180: man->next = ROFF_NEXT_SIBLING;
1181: }
1182: }
1183: if (man->first == n)
1184: man->first = NULL;
1185: }
1186:
1187: void
1188: roff_node_free(struct roff_node *n)
1189: {
1190:
1191: if (n->args != NULL)
1192: mdoc_argv_free(n->args);
1193: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1194: free(n->norm);
1195: free(n->string);
1196: free(n);
1197: }
1198:
1199: void
1200: roff_node_delete(struct roff_man *man, struct roff_node *n)
1201: {
1202:
1203: while (n->child != NULL)
1204: roff_node_delete(man, n->child);
1205: assert(n->nchild == 0);
1206: roff_node_unlink(man, n);
1207: roff_node_free(n);
1208: }
1209:
1210: /* --- main functions of the roff parser ---------------------------------- */
1211:
1.8 schwarze 1212: /*
1.79 schwarze 1213: * In the current line, expand escape sequences that tend to get
1214: * used in numerical expressions and conditional requests.
1215: * Also check the syntax of the remaining escape sequences.
1.8 schwarze 1216: */
1.45 schwarze 1217: static enum rofferr
1.110 schwarze 1218: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8 schwarze 1219: {
1.81 schwarze 1220: char ubuf[24]; /* buffer to print the number */
1.78 schwarze 1221: const char *start; /* start of the string to process */
1.82 schwarze 1222: char *stesc; /* start of an escape sequence ('\\') */
1.23 schwarze 1223: const char *stnam; /* start of the name, after "[(*" */
1224: const char *cp; /* end of the name, e.g. before ']' */
1225: const char *res; /* the string to be substituted */
1.110 schwarze 1226: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 1227: size_t maxl; /* expected length of the escape name */
1228: size_t naml; /* actual length of the escape name */
1.109 schwarze 1229: enum mandoc_esc esc; /* type of the escape sequence */
1230: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 1231: int expand_count; /* to avoid infinite loops */
1.79 schwarze 1232: int npos; /* position in numeric expression */
1.91 schwarze 1233: int arg_complete; /* argument not interrupted by eol */
1.79 schwarze 1234: char term; /* character terminating the escape */
1.8 schwarze 1235:
1.43 schwarze 1236: expand_count = 0;
1.110 schwarze 1237: start = buf->buf + pos;
1.78 schwarze 1238: stesc = strchr(start, '\0') - 1;
1239: while (stesc-- > start) {
1.43 schwarze 1240:
1.78 schwarze 1241: /* Search backwards for the next backslash. */
1242:
1.110 schwarze 1243: if (*stesc != '\\')
1.78 schwarze 1244: continue;
1245:
1246: /* If it is escaped, skip it. */
1247:
1248: for (cp = stesc - 1; cp >= start; cp--)
1.110 schwarze 1249: if (*cp != '\\')
1.78 schwarze 1250: break;
1251:
1.110 schwarze 1252: if ((stesc - cp) % 2 == 0) {
1.82 schwarze 1253: stesc = (char *)cp;
1.78 schwarze 1254: continue;
1255: }
1.23 schwarze 1256:
1.79 schwarze 1257: /* Decide whether to expand or to check only. */
1.23 schwarze 1258:
1.79 schwarze 1259: term = '\0';
1.78 schwarze 1260: cp = stesc + 1;
1.53 schwarze 1261: switch (*cp) {
1.80 schwarze 1262: case '*':
1.53 schwarze 1263: res = NULL;
1264: break;
1.80 schwarze 1265: case 'B':
1.79 schwarze 1266: /* FALLTHROUGH */
1.80 schwarze 1267: case 'w':
1.79 schwarze 1268: term = cp[1];
1269: /* FALLTHROUGH */
1.80 schwarze 1270: case 'n':
1.53 schwarze 1271: res = ubuf;
1272: break;
1273: default:
1.109 schwarze 1274: esc = mandoc_escape(&cp, &stnam, &inaml);
1275: if (esc == ESCAPE_ERROR ||
1276: (esc == ESCAPE_SPECIAL &&
1277: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.92 schwarze 1278: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110 schwarze 1279: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1280: "%.*s", (int)(cp - stesc), stesc);
1.78 schwarze 1281: continue;
1.42 schwarze 1282: }
1283:
1.78 schwarze 1284: if (EXPAND_LIMIT < ++expand_count) {
1285: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110 schwarze 1286: ln, (int)(stesc - buf->buf), NULL);
1.78 schwarze 1287: return(ROFF_IGN);
1288: }
1.23 schwarze 1289:
1290: /*
1291: * The third character decides the length
1.53 schwarze 1292: * of the name of the string or register.
1.23 schwarze 1293: * Save a pointer to the name.
1294: */
1295:
1.110 schwarze 1296: if (term == '\0') {
1.79 schwarze 1297: switch (*++cp) {
1.80 schwarze 1298: case '\0':
1.79 schwarze 1299: maxl = 0;
1300: break;
1.80 schwarze 1301: case '(':
1.79 schwarze 1302: cp++;
1303: maxl = 2;
1304: break;
1.80 schwarze 1305: case '[':
1.79 schwarze 1306: cp++;
1307: term = ']';
1308: maxl = 0;
1309: break;
1310: default:
1311: maxl = 1;
1312: break;
1313: }
1314: } else {
1315: cp += 2;
1.8 schwarze 1316: maxl = 0;
1317: }
1.23 schwarze 1318: stnam = cp;
1.8 schwarze 1319:
1.23 schwarze 1320: /* Advance to the end of the name. */
1.8 schwarze 1321:
1.125 schwarze 1322: naml = 0;
1.91 schwarze 1323: arg_complete = 1;
1.125 schwarze 1324: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1325: if (*cp == '\0') {
1.92 schwarze 1326: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 1327: ln, (int)(stesc - buf->buf), stesc);
1.91 schwarze 1328: arg_complete = 0;
1.79 schwarze 1329: break;
1.42 schwarze 1330: }
1.110 schwarze 1331: if (maxl == 0 && *cp == term) {
1.79 schwarze 1332: cp++;
1.125 schwarze 1333: break;
1334: }
1335: if (*cp++ != '\\' || stesc[1] != 'w') {
1336: naml++;
1337: continue;
1338: }
1339: switch (mandoc_escape(&cp, NULL, NULL)) {
1340: case ESCAPE_SPECIAL:
1341: /* FALLTHROUGH */
1342: case ESCAPE_UNICODE:
1343: /* FALLTHROUGH */
1344: case ESCAPE_NUMBERED:
1345: /* FALLTHROUGH */
1346: case ESCAPE_OVERSTRIKE:
1347: naml++;
1348: break;
1349: default:
1.8 schwarze 1350: break;
1.79 schwarze 1351: }
1.8 schwarze 1352: }
1353:
1.23 schwarze 1354: /*
1355: * Retrieve the replacement string; if it is
1356: * undefined, resume searching for escapes.
1357: */
1358:
1.79 schwarze 1359: switch (stesc[1]) {
1.80 schwarze 1360: case '*':
1.91 schwarze 1361: if (arg_complete)
1362: res = roff_getstrn(r, stnam, naml);
1.79 schwarze 1363: break;
1.80 schwarze 1364: case 'B':
1.79 schwarze 1365: npos = 0;
1.91 schwarze 1366: ubuf[0] = arg_complete &&
1.133 schwarze 1367: roff_evalnum(r, ln, stnam, &npos,
1368: NULL, ROFFNUM_SCALE) &&
1.91 schwarze 1369: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1370: ubuf[1] = '\0';
1371: break;
1.80 schwarze 1372: case 'n':
1.91 schwarze 1373: if (arg_complete)
1374: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1375: roff_getregn(r, stnam, naml));
1376: else
1377: ubuf[0] = '\0';
1.79 schwarze 1378: break;
1.80 schwarze 1379: case 'w':
1.91 schwarze 1380: /* use even incomplete args */
1.81 schwarze 1381: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1382: 24 * (int)naml);
1383: break;
1384: }
1.8 schwarze 1385:
1.110 schwarze 1386: if (res == NULL) {
1.92 schwarze 1387: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1388: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1389: "%.*s", (int)naml, stnam);
1.37 schwarze 1390: res = "";
1.118 schwarze 1391: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1392: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1393: ln, (int)(stesc - buf->buf), NULL);
1394: return(ROFF_IGN);
1.8 schwarze 1395: }
1396:
1.23 schwarze 1397: /* Replace the escape sequence by the string. */
1398:
1.82 schwarze 1399: *stesc = '\0';
1.110 schwarze 1400: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1401: buf->buf, res, cp) + 1;
1.8 schwarze 1402:
1.78 schwarze 1403: /* Prepare for the next replacement. */
1.8 schwarze 1404:
1.78 schwarze 1405: start = nbuf + pos;
1.110 schwarze 1406: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1407: free(buf->buf);
1408: buf->buf = nbuf;
1.42 schwarze 1409: }
1.45 schwarze 1410: return(ROFF_CONT);
1.42 schwarze 1411: }
1412:
1413: /*
1.51 schwarze 1414: * Process text streams:
1415: * Convert all breakable hyphens into ASCII_HYPH.
1416: * Decrement and spring input line trap.
1.42 schwarze 1417: */
1418: static enum rofferr
1.110 schwarze 1419: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42 schwarze 1420: {
1421: size_t sz;
1422: const char *start;
1.51 schwarze 1423: char *p;
1424: int isz;
1.42 schwarze 1425: enum mandoc_esc esc;
1426:
1.110 schwarze 1427: start = p = buf->buf + pos;
1.42 schwarze 1428:
1.110 schwarze 1429: while (*p != '\0') {
1.42 schwarze 1430: sz = strcspn(p, "-\\");
1431: p += sz;
1432:
1.110 schwarze 1433: if (*p == '\0')
1.42 schwarze 1434: break;
1435:
1.110 schwarze 1436: if (*p == '\\') {
1.42 schwarze 1437: /* Skip over escapes. */
1438: p++;
1.62 schwarze 1439: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1440: if (esc == ESCAPE_ERROR)
1.42 schwarze 1441: break;
1.136 schwarze 1442: while (*p == '-')
1443: p++;
1.42 schwarze 1444: continue;
1445: } else if (p == start) {
1446: p++;
1447: continue;
1448: }
1449:
1.44 schwarze 1450: if (isalpha((unsigned char)p[-1]) &&
1451: isalpha((unsigned char)p[1]))
1.42 schwarze 1452: *p = ASCII_HYPH;
1453: p++;
1.8 schwarze 1454: }
1455:
1.51 schwarze 1456: /* Spring the input line trap. */
1.110 schwarze 1457: if (roffit_lines == 1) {
1458: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1459: free(buf->buf);
1460: buf->buf = p;
1461: buf->sz = isz + 1;
1.51 schwarze 1462: *offs = 0;
1463: free(roffit_macro);
1464: roffit_lines = 0;
1465: return(ROFF_REPARSE);
1.110 schwarze 1466: } else if (roffit_lines > 1)
1.51 schwarze 1467: --roffit_lines;
1.42 schwarze 1468: return(ROFF_CONT);
1.8 schwarze 1469: }
1470:
1.1 schwarze 1471: enum rofferr
1.110 schwarze 1472: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1473: {
1474: enum rofft t;
1.27 schwarze 1475: enum rofferr e;
1.110 schwarze 1476: int pos; /* parse point */
1.115 schwarze 1477: int spos; /* saved parse point for messages */
1.110 schwarze 1478: int ppos; /* original offset in buf->buf */
1479: int ctl; /* macro line (boolean) */
1480:
1481: ppos = pos = *offs;
1.1 schwarze 1482:
1.102 schwarze 1483: /* Handle in-line equation delimiters. */
1484:
1.108 schwarze 1485: if (r->tbl == NULL &&
1486: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1487: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1488: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1489: if (e == ROFF_REPARSE)
1490: return(e);
1491: assert(e == ROFF_CONT);
1492: }
1493:
1494: /* Expand some escape sequences. */
1.8 schwarze 1495:
1.110 schwarze 1496: e = roff_res(r, buf, ln, pos);
1497: if (e == ROFF_IGN)
1.45 schwarze 1498: return(e);
1.110 schwarze 1499: assert(e == ROFF_CONT);
1.8 schwarze 1500:
1.110 schwarze 1501: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1502:
1.8 schwarze 1503: /*
1.2 schwarze 1504: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1505: * text through the macro's filter.
1506: * Equations process all content themselves.
1507: * Tables process almost all content themselves, but we want
1508: * to warn about macros before passing it there.
1.2 schwarze 1509: */
1510:
1.124 schwarze 1511: if (r->last != NULL && ! ctl) {
1.2 schwarze 1512: t = r->last->tok;
1513: assert(roffs[t].text);
1.110 schwarze 1514: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1515: assert(e == ROFF_IGN || e == ROFF_CONT);
1516: if (e != ROFF_CONT)
1.32 schwarze 1517: return(e);
1.54 schwarze 1518: }
1.124 schwarze 1519: if (r->eqn != NULL)
1.110 schwarze 1520: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.124 schwarze 1521: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.131 schwarze 1522: return(tbl_read(r->tbl, ln, buf->buf, ppos));
1.124 schwarze 1523: if ( ! ctl)
1.110 schwarze 1524: return(roff_parsetext(buf, pos, offs));
1.100 schwarze 1525:
1526: /* Skip empty request lines. */
1527:
1.110 schwarze 1528: if (buf->buf[pos] == '"') {
1.100 schwarze 1529: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1530: ln, pos, NULL);
1531: return(ROFF_IGN);
1.110 schwarze 1532: } else if (buf->buf[pos] == '\0')
1.100 schwarze 1533: return(ROFF_IGN);
1.2 schwarze 1534:
1535: /*
1536: * If a scope is open, go to the child handler for that macro,
1537: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1538: * Don't do so if an equation is open.
1.2 schwarze 1539: */
1540:
1541: if (r->last) {
1.1 schwarze 1542: t = r->last->tok;
1543: assert(roffs[t].sub);
1.110 schwarze 1544: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1545: }
1546:
1.115 schwarze 1547: /* No scope is open. This is a new request or macro. */
1548:
1549: spos = pos;
1550: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1551:
1552: /* Tables ignore most macros. */
1553:
1554: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1555: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1556: ln, pos, buf->buf + spos);
1.129 schwarze 1557: if (t == ROFF_TS)
1558: return(ROFF_IGN);
1559: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1560: pos++;
1561: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1562: pos++;
1563: return(tbl_read(r->tbl, ln, buf->buf, pos));
1.115 schwarze 1564: }
1565:
1.2 schwarze 1566: /*
1.115 schwarze 1567: * This is neither a roff request nor a user-defined macro.
1568: * Let the standard macro set parsers handle it.
1.2 schwarze 1569: */
1570:
1.115 schwarze 1571: if (t == ROFF_MAX)
1.1 schwarze 1572: return(ROFF_CONT);
1.115 schwarze 1573:
1574: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1575:
1.2 schwarze 1576: assert(roffs[t].proc);
1.110 schwarze 1577: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1578: }
1579:
1.27 schwarze 1580: void
1.2 schwarze 1581: roff_endparse(struct roff *r)
1582: {
1.1 schwarze 1583:
1.27 schwarze 1584: if (r->last)
1.94 schwarze 1585: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1586: r->last->line, r->last->col,
1587: roffs[r->last->tok].name);
1.27 schwarze 1588:
1.32 schwarze 1589: if (r->eqn) {
1.94 schwarze 1590: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1591: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41 schwarze 1592: eqn_end(&r->eqn);
1.32 schwarze 1593: }
1594:
1.27 schwarze 1595: if (r->tbl) {
1.94 schwarze 1596: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1597: r->tbl->line, r->tbl->pos, "TS");
1.41 schwarze 1598: tbl_end(&r->tbl);
1.27 schwarze 1599: }
1.1 schwarze 1600: }
1601:
1602: /*
1603: * Parse a roff node's type from the input buffer. This must be in the
1604: * form of ".foo xxx" in the usual way.
1605: */
1606: static enum rofft
1.87 schwarze 1607: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1608: {
1.87 schwarze 1609: char *cp;
1.16 schwarze 1610: const char *mac;
1611: size_t maclen;
1.1 schwarze 1612: enum rofft t;
1613:
1.87 schwarze 1614: cp = buf + *pos;
1615:
1616: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.1 schwarze 1617: return(ROFF_MAX);
1618:
1.87 schwarze 1619: mac = cp;
1620: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 1621:
1.16 schwarze 1622: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42 schwarze 1623: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1 schwarze 1624:
1.87 schwarze 1625: if (ROFF_MAX != t)
1626: *pos = cp - buf;
1.1 schwarze 1627:
1628: return(t);
1629: }
1630:
1.138 schwarze 1631: /* --- handling of request blocks ----------------------------------------- */
1632:
1.1 schwarze 1633: static enum rofferr
1.2 schwarze 1634: roff_cblock(ROFF_ARGS)
1.1 schwarze 1635: {
1636:
1.2 schwarze 1637: /*
1638: * A block-close `..' should only be invoked as a child of an
1639: * ignore macro, otherwise raise a warning and just ignore it.
1640: */
1641:
1.110 schwarze 1642: if (r->last == NULL) {
1.94 schwarze 1643: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1644: ln, ppos, "..");
1.2 schwarze 1645: return(ROFF_IGN);
1646: }
1.1 schwarze 1647:
1.2 schwarze 1648: switch (r->last->tok) {
1.80 schwarze 1649: case ROFF_am:
1.93 schwarze 1650: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.2 schwarze 1651: /* FALLTHROUGH */
1.80 schwarze 1652: case ROFF_ami:
1.2 schwarze 1653: /* FALLTHROUGH */
1.80 schwarze 1654: case ROFF_de:
1.23 schwarze 1655: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2 schwarze 1656: /* FALLTHROUGH */
1.80 schwarze 1657: case ROFF_dei:
1.2 schwarze 1658: /* FALLTHROUGH */
1.80 schwarze 1659: case ROFF_ig:
1.2 schwarze 1660: break;
1661: default:
1.94 schwarze 1662: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1663: ln, ppos, "..");
1.1 schwarze 1664: return(ROFF_IGN);
1.2 schwarze 1665: }
1666:
1.110 schwarze 1667: if (buf->buf[pos] != '\0')
1.90 schwarze 1668: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 1669: ".. %s", buf->buf + pos);
1.2 schwarze 1670:
1671: roffnode_pop(r);
1672: roffnode_cleanscope(r);
1673: return(ROFF_IGN);
1674:
1675: }
1.1 schwarze 1676:
1.2 schwarze 1677: static void
1678: roffnode_cleanscope(struct roff *r)
1679: {
1.1 schwarze 1680:
1.2 schwarze 1681: while (r->last) {
1.46 schwarze 1682: if (--r->last->endspan != 0)
1.2 schwarze 1683: break;
1684: roffnode_pop(r);
1685: }
1686: }
1.1 schwarze 1687:
1.68 schwarze 1688: static void
1689: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 1690: {
1.1 schwarze 1691:
1.2 schwarze 1692: if (NULL == r->last) {
1.94 schwarze 1693: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1694: ln, ppos, "\\}");
1.68 schwarze 1695: return;
1.2 schwarze 1696: }
1.1 schwarze 1697:
1.2 schwarze 1698: switch (r->last->tok) {
1.80 schwarze 1699: case ROFF_el:
1.2 schwarze 1700: /* FALLTHROUGH */
1.80 schwarze 1701: case ROFF_ie:
1.2 schwarze 1702: /* FALLTHROUGH */
1.80 schwarze 1703: case ROFF_if:
1.2 schwarze 1704: break;
1705: default:
1.94 schwarze 1706: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1707: ln, ppos, "\\}");
1.68 schwarze 1708: return;
1.2 schwarze 1709: }
1.1 schwarze 1710:
1.2 schwarze 1711: if (r->last->endspan > -1) {
1.94 schwarze 1712: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1713: ln, ppos, "\\}");
1.68 schwarze 1714: return;
1.2 schwarze 1715: }
1716:
1717: roffnode_pop(r);
1718: roffnode_cleanscope(r);
1.68 schwarze 1719: return;
1.1 schwarze 1720: }
1721:
1722: static enum rofferr
1.2 schwarze 1723: roff_block(ROFF_ARGS)
1.1 schwarze 1724: {
1.93 schwarze 1725: const char *name;
1726: char *iname, *cp;
1.86 schwarze 1727: size_t namesz;
1.16 schwarze 1728:
1.93 schwarze 1729: /* Ignore groff compatibility mode for now. */
1.2 schwarze 1730:
1.110 schwarze 1731: if (tok == ROFF_de1)
1.93 schwarze 1732: tok = ROFF_de;
1.123 schwarze 1733: else if (tok == ROFF_dei1)
1734: tok = ROFF_dei;
1.110 schwarze 1735: else if (tok == ROFF_am1)
1.93 schwarze 1736: tok = ROFF_am;
1.123 schwarze 1737: else if (tok == ROFF_ami1)
1738: tok = ROFF_ami;
1.93 schwarze 1739:
1740: /* Parse the macro name argument. */
1741:
1.110 schwarze 1742: cp = buf->buf + pos;
1743: if (tok == ROFF_ig) {
1.93 schwarze 1744: iname = NULL;
1745: namesz = 0;
1746: } else {
1747: iname = cp;
1748: namesz = roff_getname(r, &cp, ln, ppos);
1749: iname[namesz] = '\0';
1750: }
1.22 schwarze 1751:
1.93 schwarze 1752: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 1753:
1.110 schwarze 1754: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1755: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1756: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1757: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1758: "%.*s", (int)namesz, iname);
1759: namesz = 0;
1760: } else
1761: namesz = strlen(name);
1762: } else
1763: name = iname;
1.22 schwarze 1764:
1.110 schwarze 1765: if (namesz == 0 && tok != ROFF_ig) {
1.93 schwarze 1766: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1767: ln, ppos, roffs[tok].name);
1768: return(ROFF_IGN);
1769: }
1.2 schwarze 1770:
1.16 schwarze 1771: roffnode_push(r, tok, name, ln, ppos);
1772:
1773: /*
1774: * At the beginning of a `de' macro, clear the existing string
1775: * with the same name, if there is one. New content will be
1.66 schwarze 1776: * appended from roff_block_text() in multiline mode.
1.16 schwarze 1777: */
1.22 schwarze 1778:
1.110 schwarze 1779: if (tok == ROFF_de || tok == ROFF_dei)
1.86 schwarze 1780: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2 schwarze 1781:
1.110 schwarze 1782: if (*cp == '\0')
1.2 schwarze 1783: return(ROFF_IGN);
1.1 schwarze 1784:
1.93 schwarze 1785: /* Get the custom end marker. */
1.22 schwarze 1786:
1.93 schwarze 1787: iname = cp;
1.86 schwarze 1788: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 1789:
1790: /* Resolve the end marker if it is indirect. */
1791:
1.110 schwarze 1792: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1793: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1794: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1795: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1796: "%.*s", (int)namesz, iname);
1797: namesz = 0;
1798: } else
1799: namesz = strlen(name);
1800: } else
1801: name = iname;
1802:
1.86 schwarze 1803: if (namesz)
1804: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 1805:
1.110 schwarze 1806: if (*cp != '\0')
1.90 schwarze 1807: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1808: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.1 schwarze 1809:
1810: return(ROFF_IGN);
1811: }
1812:
1813: static enum rofferr
1.2 schwarze 1814: roff_block_sub(ROFF_ARGS)
1.1 schwarze 1815: {
1.2 schwarze 1816: enum rofft t;
1817: int i, j;
1818:
1819: /*
1820: * First check whether a custom macro exists at this level. If
1821: * it does, then check against it. This is some of groff's
1822: * stranger behaviours. If we encountered a custom end-scope
1823: * tag and that tag also happens to be a "real" macro, then we
1824: * need to try interpreting it again as a real macro. If it's
1825: * not, then return ignore. Else continue.
1826: */
1827:
1828: if (r->last->end) {
1.35 schwarze 1829: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 1830: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 1831: break;
1.1 schwarze 1832:
1.110 schwarze 1833: if (r->last->end[j] == '\0' &&
1834: (buf->buf[i] == '\0' ||
1835: buf->buf[i] == ' ' ||
1836: buf->buf[i] == '\t')) {
1.2 schwarze 1837: roffnode_pop(r);
1838: roffnode_cleanscope(r);
1.1 schwarze 1839:
1.110 schwarze 1840: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 1841: i++;
1842:
1843: pos = i;
1.110 schwarze 1844: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1845: ROFF_MAX)
1.2 schwarze 1846: return(ROFF_RERUN);
1847: return(ROFF_IGN);
1848: }
1.1 schwarze 1849: }
1850:
1.2 schwarze 1851: /*
1852: * If we have no custom end-query or lookup failed, then try
1853: * pulling it out of the hashtable.
1854: */
1.1 schwarze 1855:
1.110 schwarze 1856: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 1857:
1.110 schwarze 1858: if (t != ROFF_cblock) {
1859: if (tok != ROFF_ig)
1860: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.1 schwarze 1861: return(ROFF_IGN);
1.16 schwarze 1862: }
1.1 schwarze 1863:
1.2 schwarze 1864: assert(roffs[t].proc);
1.110 schwarze 1865: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1866: }
1867:
1868: static enum rofferr
1869: roff_block_text(ROFF_ARGS)
1870: {
1871:
1.110 schwarze 1872: if (tok != ROFF_ig)
1873: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 1874:
1.2 schwarze 1875: return(ROFF_IGN);
1876: }
1877:
1878: static enum rofferr
1879: roff_cond_sub(ROFF_ARGS)
1880: {
1881: enum rofft t;
1.37 schwarze 1882: char *ep;
1.71 schwarze 1883: int rr;
1.2 schwarze 1884:
1885: rr = r->last->rule;
1.37 schwarze 1886: roffnode_cleanscope(r);
1.110 schwarze 1887: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2 schwarze 1888:
1.37 schwarze 1889: /*
1.50 schwarze 1890: * Fully handle known macros when they are structurally
1891: * required or when the conditional evaluated to true.
1.5 schwarze 1892: */
1893:
1.110 schwarze 1894: if ((t != ROFF_MAX) &&
1895: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.50 schwarze 1896: assert(roffs[t].proc);
1.110 schwarze 1897: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.50 schwarze 1898: }
1.39 schwarze 1899:
1.69 schwarze 1900: /*
1901: * If `\}' occurs on a macro line without a preceding macro,
1902: * drop the line completely.
1903: */
1904:
1.110 schwarze 1905: ep = buf->buf + pos;
1906: if (ep[0] == '\\' && ep[1] == '}')
1.71 schwarze 1907: rr = 0;
1.69 schwarze 1908:
1.50 schwarze 1909: /* Always check for the closing delimiter `\}'. */
1.39 schwarze 1910:
1.110 schwarze 1911: while ((ep = strchr(ep, '\\')) != NULL) {
1912: if (*(++ep) == '}') {
1.70 schwarze 1913: *ep = '&';
1.110 schwarze 1914: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1915: }
1.119 schwarze 1916: if (*ep != '\0')
1917: ++ep;
1.50 schwarze 1918: }
1.71 schwarze 1919: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1920: }
1921:
1922: static enum rofferr
1923: roff_cond_text(ROFF_ARGS)
1924: {
1.37 schwarze 1925: char *ep;
1.71 schwarze 1926: int rr;
1.2 schwarze 1927:
1928: rr = r->last->rule;
1.37 schwarze 1929: roffnode_cleanscope(r);
1.1 schwarze 1930:
1.110 schwarze 1931: ep = buf->buf + pos;
1932: while ((ep = strchr(ep, '\\')) != NULL) {
1933: if (*(++ep) == '}') {
1.70 schwarze 1934: *ep = '&';
1.110 schwarze 1935: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1936: }
1.119 schwarze 1937: if (*ep != '\0')
1938: ++ep;
1.2 schwarze 1939: }
1.71 schwarze 1940: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1941: }
1942:
1.138 schwarze 1943: /* --- handling of numeric and conditional expressions -------------------- */
1944:
1.77 schwarze 1945: /*
1946: * Parse a single signed integer number. Stop at the first non-digit.
1947: * If there is at least one digit, return success and advance the
1948: * parse point, else return failure and let the parse point unchanged.
1949: * Ignore overflows, treat them just like the C language.
1950: */
1.56 schwarze 1951: static int
1.133 schwarze 1952: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56 schwarze 1953: {
1.133 schwarze 1954: int myres, scaled, n, p;
1.79 schwarze 1955:
1956: if (NULL == res)
1957: res = &myres;
1.56 schwarze 1958:
1959: p = *pos;
1960: n = v[p] == '-';
1.133 schwarze 1961: if (n || v[p] == '+')
1.56 schwarze 1962: p++;
1963:
1.133 schwarze 1964: if (flags & ROFFNUM_WHITE)
1965: while (isspace((unsigned char)v[p]))
1966: p++;
1967:
1.56 schwarze 1968: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 1969: *res = 10 * *res + v[p] - '0';
1.56 schwarze 1970: if (p == *pos + n)
1971: return 0;
1972:
1973: if (n)
1974: *res = -*res;
1975:
1.126 schwarze 1976: /* Each number may be followed by one optional scaling unit. */
1977:
1978: switch (v[p]) {
1979: case 'f':
1.133 schwarze 1980: scaled = *res * 65536;
1.126 schwarze 1981: break;
1982: case 'i':
1.133 schwarze 1983: scaled = *res * 240;
1.126 schwarze 1984: break;
1985: case 'c':
1.133 schwarze 1986: scaled = *res * 240 / 2.54;
1.126 schwarze 1987: break;
1988: case 'v':
1989: /* FALLTROUGH */
1990: case 'P':
1.133 schwarze 1991: scaled = *res * 40;
1.126 schwarze 1992: break;
1993: case 'm':
1994: /* FALLTROUGH */
1995: case 'n':
1.133 schwarze 1996: scaled = *res * 24;
1.126 schwarze 1997: break;
1998: case 'p':
1.133 schwarze 1999: scaled = *res * 10 / 3;
1.126 schwarze 2000: break;
2001: case 'u':
1.133 schwarze 2002: scaled = *res;
1.126 schwarze 2003: break;
2004: case 'M':
1.133 schwarze 2005: scaled = *res * 6 / 25;
1.126 schwarze 2006: break;
2007: default:
1.133 schwarze 2008: scaled = *res;
1.126 schwarze 2009: p--;
2010: break;
2011: }
1.133 schwarze 2012: if (flags & ROFFNUM_SCALE)
2013: *res = scaled;
1.126 schwarze 2014:
2015: *pos = p + 1;
2016: return(1);
1.56 schwarze 2017: }
2018:
1.71 schwarze 2019: /*
2020: * Evaluate a string comparison condition.
2021: * The first character is the delimiter.
2022: * Succeed if the string up to its second occurrence
2023: * matches the string up to its third occurence.
2024: * Advance the cursor after the third occurrence
2025: * or lacking that, to the end of the line.
2026: */
2027: static int
2028: roff_evalstrcond(const char *v, int *pos)
2029: {
2030: const char *s1, *s2, *s3;
2031: int match;
2032:
2033: match = 0;
2034: s1 = v + *pos; /* initial delimiter */
2035: s2 = s1 + 1; /* for scanning the first string */
2036: s3 = strchr(s2, *s1); /* for scanning the second string */
2037:
2038: if (NULL == s3) /* found no middle delimiter */
2039: goto out;
2040:
2041: while ('\0' != *++s3) {
2042: if (*s2 != *s3) { /* mismatch */
2043: s3 = strchr(s3, *s1);
2044: break;
2045: }
2046: if (*s3 == *s1) { /* found the final delimiter */
2047: match = 1;
2048: break;
2049: }
2050: s2++;
2051: }
2052:
2053: out:
2054: if (NULL == s3)
2055: s3 = strchr(s2, '\0');
1.114 schwarze 2056: else if (*s3 != '\0')
1.71 schwarze 2057: s3++;
2058: *pos = s3 - v;
2059: return(match);
2060: }
2061:
1.77 schwarze 2062: /*
2063: * Evaluate an optionally negated single character, numerical,
2064: * or string condition.
2065: */
1.71 schwarze 2066: static int
1.106 schwarze 2067: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.5 schwarze 2068: {
1.113 schwarze 2069: int number, savepos, wanttrue;
1.5 schwarze 2070:
1.71 schwarze 2071: if ('!' == v[*pos]) {
2072: wanttrue = 0;
2073: (*pos)++;
2074: } else
2075: wanttrue = 1;
2076:
1.5 schwarze 2077: switch (v[*pos]) {
1.112 schwarze 2078: case '\0':
2079: return(0);
1.80 schwarze 2080: case 'n':
1.71 schwarze 2081: /* FALLTHROUGH */
1.80 schwarze 2082: case 'o':
1.5 schwarze 2083: (*pos)++;
1.71 schwarze 2084: return(wanttrue);
1.80 schwarze 2085: case 'c':
1.71 schwarze 2086: /* FALLTHROUGH */
1.80 schwarze 2087: case 'd':
1.71 schwarze 2088: /* FALLTHROUGH */
1.80 schwarze 2089: case 'e':
1.5 schwarze 2090: /* FALLTHROUGH */
1.80 schwarze 2091: case 'r':
1.5 schwarze 2092: /* FALLTHROUGH */
1.80 schwarze 2093: case 't':
1.111 schwarze 2094: /* FALLTHROUGH */
2095: case 'v':
1.5 schwarze 2096: (*pos)++;
1.71 schwarze 2097: return(!wanttrue);
1.5 schwarze 2098: default:
2099: break;
2100: }
2101:
1.113 schwarze 2102: savepos = *pos;
1.133 schwarze 2103: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.77 schwarze 2104: return((number > 0) == wanttrue);
1.113 schwarze 2105: else if (*pos == savepos)
2106: return(roff_evalstrcond(v, pos) == wanttrue);
1.77 schwarze 2107: else
1.113 schwarze 2108: return (0);
1.5 schwarze 2109: }
2110:
1.2 schwarze 2111: static enum rofferr
1.21 schwarze 2112: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 2113: {
1.30 schwarze 2114:
1.21 schwarze 2115: return(ROFF_IGN);
2116: }
2117:
2118: static enum rofferr
1.123 schwarze 2119: roff_insec(ROFF_ARGS)
2120: {
2121:
2122: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
2123: ln, ppos, roffs[tok].name);
2124: return(ROFF_IGN);
2125: }
2126:
2127: static enum rofferr
2128: roff_unsupp(ROFF_ARGS)
2129: {
2130:
2131: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
2132: ln, ppos, roffs[tok].name);
2133: return(ROFF_IGN);
2134: }
2135:
2136: static enum rofferr
1.2 schwarze 2137: roff_cond(ROFF_ARGS)
2138: {
1.46 schwarze 2139:
2140: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 2141:
1.80 schwarze 2142: /*
1.35 schwarze 2143: * An `.el' has no conditional body: it will consume the value
2144: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 2145: * defaults to DENY.
1.35 schwarze 2146: *
2147: * If we're not an `el', however, then evaluate the conditional.
2148: */
1.1 schwarze 2149:
1.110 schwarze 2150: r->last->rule = tok == ROFF_el ?
1.80 schwarze 2151: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 2152: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 2153:
1.35 schwarze 2154: /*
2155: * An if-else will put the NEGATION of the current evaluated
2156: * conditional into the stack of rules.
2157: */
2158:
1.110 schwarze 2159: if (tok == ROFF_ie) {
1.96 schwarze 2160: if (r->rstackpos + 1 == r->rstacksz) {
2161: r->rstacksz += 16;
2162: r->rstack = mandoc_reallocarray(r->rstack,
2163: r->rstacksz, sizeof(int));
1.35 schwarze 2164: }
1.71 schwarze 2165: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 2166: }
1.5 schwarze 2167:
2168: /* If the parent has false as its rule, then so do we. */
2169:
1.71 schwarze 2170: if (r->last->parent && !r->last->parent->rule)
2171: r->last->rule = 0;
1.5 schwarze 2172:
2173: /*
1.46 schwarze 2174: * Determine scope.
2175: * If there is nothing on the line after the conditional,
2176: * not even whitespace, use next-line scope.
1.5 schwarze 2177: */
1.2 schwarze 2178:
1.110 schwarze 2179: if (buf->buf[pos] == '\0') {
1.46 schwarze 2180: r->last->endspan = 2;
2181: goto out;
2182: }
2183:
1.110 schwarze 2184: while (buf->buf[pos] == ' ')
1.46 schwarze 2185: pos++;
2186:
2187: /* An opening brace requests multiline scope. */
1.2 schwarze 2188:
1.110 schwarze 2189: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 2190: r->last->endspan = -1;
2191: pos += 2;
1.46 schwarze 2192: goto out;
1.80 schwarze 2193: }
1.2 schwarze 2194:
2195: /*
1.46 schwarze 2196: * Anything else following the conditional causes
2197: * single-line scope. Warn if the scope contains
2198: * nothing but trailing whitespace.
1.2 schwarze 2199: */
2200:
1.110 schwarze 2201: if (buf->buf[pos] == '\0')
1.89 schwarze 2202: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
2203: ln, ppos, roffs[tok].name);
1.2 schwarze 2204:
1.46 schwarze 2205: r->last->endspan = 1;
1.1 schwarze 2206:
1.46 schwarze 2207: out:
1.2 schwarze 2208: *offs = pos;
2209: return(ROFF_RERUN);
1.1 schwarze 2210: }
2211:
1.2 schwarze 2212: static enum rofferr
1.7 schwarze 2213: roff_ds(ROFF_ARGS)
2214: {
1.85 schwarze 2215: char *string;
2216: const char *name;
2217: size_t namesz;
1.10 schwarze 2218:
1.123 schwarze 2219: /* Ignore groff compatibility mode for now. */
2220:
2221: if (tok == ROFF_ds1)
2222: tok = ROFF_ds;
2223: else if (tok == ROFF_as1)
2224: tok = ROFF_as;
2225:
1.10 schwarze 2226: /*
1.85 schwarze 2227: * The first word is the name of the string.
2228: * If it is empty or terminated by an escape sequence,
2229: * abort the `ds' request without defining anything.
1.10 schwarze 2230: */
1.7 schwarze 2231:
1.110 schwarze 2232: name = string = buf->buf + pos;
2233: if (*name == '\0')
1.7 schwarze 2234: return(ROFF_IGN);
2235:
1.85 schwarze 2236: namesz = roff_getname(r, &string, ln, pos);
1.110 schwarze 2237: if (name[namesz] == '\\')
1.85 schwarze 2238: return(ROFF_IGN);
2239:
2240: /* Read past the initial double-quote, if any. */
1.110 schwarze 2241: if (*string == '"')
1.7 schwarze 2242: string++;
2243:
1.10 schwarze 2244: /* The rest is the value. */
1.85 schwarze 2245: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2246: ROFF_as == tok);
1.7 schwarze 2247: return(ROFF_IGN);
2248: }
2249:
1.77 schwarze 2250: /*
2251: * Parse a single operator, one or two characters long.
2252: * If the operator is recognized, return success and advance the
2253: * parse point, else return failure and let the parse point unchanged.
2254: */
2255: static int
2256: roff_getop(const char *v, int *pos, char *res)
2257: {
2258:
2259: *res = v[*pos];
2260:
2261: switch (*res) {
1.80 schwarze 2262: case '+':
1.77 schwarze 2263: /* FALLTHROUGH */
1.80 schwarze 2264: case '-':
1.77 schwarze 2265: /* FALLTHROUGH */
1.80 schwarze 2266: case '*':
1.77 schwarze 2267: /* FALLTHROUGH */
1.80 schwarze 2268: case '/':
1.77 schwarze 2269: /* FALLTHROUGH */
1.80 schwarze 2270: case '%':
1.77 schwarze 2271: /* FALLTHROUGH */
1.80 schwarze 2272: case '&':
1.77 schwarze 2273: /* FALLTHROUGH */
1.80 schwarze 2274: case ':':
1.77 schwarze 2275: break;
2276: case '<':
2277: switch (v[*pos + 1]) {
1.80 schwarze 2278: case '=':
1.77 schwarze 2279: *res = 'l';
2280: (*pos)++;
2281: break;
1.80 schwarze 2282: case '>':
1.77 schwarze 2283: *res = '!';
2284: (*pos)++;
2285: break;
1.80 schwarze 2286: case '?':
1.77 schwarze 2287: *res = 'i';
2288: (*pos)++;
2289: break;
2290: default:
2291: break;
2292: }
2293: break;
2294: case '>':
2295: switch (v[*pos + 1]) {
1.80 schwarze 2296: case '=':
1.77 schwarze 2297: *res = 'g';
2298: (*pos)++;
2299: break;
1.80 schwarze 2300: case '?':
1.77 schwarze 2301: *res = 'a';
2302: (*pos)++;
2303: break;
2304: default:
2305: break;
2306: }
2307: break;
2308: case '=':
2309: if ('=' == v[*pos + 1])
2310: (*pos)++;
2311: break;
2312: default:
2313: return(0);
2314: }
2315: (*pos)++;
2316:
2317: return(*res);
2318: }
2319:
2320: /*
2321: * Evaluate either a parenthesized numeric expression
2322: * or a single signed integer number.
2323: */
2324: static int
1.106 schwarze 2325: roff_evalpar(struct roff *r, int ln,
1.133 schwarze 2326: const char *v, int *pos, int *res, int flags)
1.77 schwarze 2327: {
2328:
2329: if ('(' != v[*pos])
1.133 schwarze 2330: return(roff_getnum(v, pos, res, flags));
1.77 schwarze 2331:
2332: (*pos)++;
1.133 schwarze 2333: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.77 schwarze 2334: return(0);
2335:
1.79 schwarze 2336: /*
2337: * Omission of the closing parenthesis
2338: * is an error in validation mode,
2339: * but ignored in evaluation mode.
2340: */
2341:
1.77 schwarze 2342: if (')' == v[*pos])
2343: (*pos)++;
1.79 schwarze 2344: else if (NULL == res)
2345: return(0);
1.77 schwarze 2346:
2347: return(1);
2348: }
2349:
2350: /*
2351: * Evaluate a complete numeric expression.
2352: * Proceed left to right, there is no concept of precedence.
2353: */
2354: static int
1.106 schwarze 2355: roff_evalnum(struct roff *r, int ln, const char *v,
1.133 schwarze 2356: int *pos, int *res, int flags)
1.77 schwarze 2357: {
2358: int mypos, operand2;
2359: char operator;
2360:
2361: if (NULL == pos) {
2362: mypos = 0;
2363: pos = &mypos;
2364: }
2365:
1.133 schwarze 2366: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2367: while (isspace((unsigned char)v[*pos]))
2368: (*pos)++;
2369:
1.133 schwarze 2370: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.77 schwarze 2371: return(0);
2372:
2373: while (1) {
1.133 schwarze 2374: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2375: while (isspace((unsigned char)v[*pos]))
2376: (*pos)++;
2377:
2378: if ( ! roff_getop(v, pos, &operator))
2379: break;
2380:
1.133 schwarze 2381: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2382: while (isspace((unsigned char)v[*pos]))
2383: (*pos)++;
2384:
1.133 schwarze 2385: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.77 schwarze 2386: return(0);
2387:
1.133 schwarze 2388: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2389: while (isspace((unsigned char)v[*pos]))
2390: (*pos)++;
1.79 schwarze 2391:
2392: if (NULL == res)
2393: continue;
1.77 schwarze 2394:
2395: switch (operator) {
1.80 schwarze 2396: case '+':
1.77 schwarze 2397: *res += operand2;
2398: break;
1.80 schwarze 2399: case '-':
1.77 schwarze 2400: *res -= operand2;
2401: break;
1.80 schwarze 2402: case '*':
1.77 schwarze 2403: *res *= operand2;
2404: break;
1.80 schwarze 2405: case '/':
1.116 schwarze 2406: if (operand2 == 0) {
1.106 schwarze 2407: mandoc_msg(MANDOCERR_DIVZERO,
2408: r->parse, ln, *pos, v);
2409: *res = 0;
2410: break;
2411: }
1.77 schwarze 2412: *res /= operand2;
2413: break;
1.80 schwarze 2414: case '%':
1.116 schwarze 2415: if (operand2 == 0) {
2416: mandoc_msg(MANDOCERR_DIVZERO,
2417: r->parse, ln, *pos, v);
2418: *res = 0;
2419: break;
2420: }
1.77 schwarze 2421: *res %= operand2;
2422: break;
1.80 schwarze 2423: case '<':
1.77 schwarze 2424: *res = *res < operand2;
2425: break;
1.80 schwarze 2426: case '>':
1.77 schwarze 2427: *res = *res > operand2;
2428: break;
1.80 schwarze 2429: case 'l':
1.77 schwarze 2430: *res = *res <= operand2;
2431: break;
1.80 schwarze 2432: case 'g':
1.77 schwarze 2433: *res = *res >= operand2;
2434: break;
1.80 schwarze 2435: case '=':
1.77 schwarze 2436: *res = *res == operand2;
2437: break;
1.80 schwarze 2438: case '!':
1.77 schwarze 2439: *res = *res != operand2;
2440: break;
1.80 schwarze 2441: case '&':
1.77 schwarze 2442: *res = *res && operand2;
2443: break;
1.80 schwarze 2444: case ':':
1.77 schwarze 2445: *res = *res || operand2;
2446: break;
1.80 schwarze 2447: case 'i':
1.77 schwarze 2448: if (operand2 < *res)
2449: *res = operand2;
2450: break;
1.80 schwarze 2451: case 'a':
1.77 schwarze 2452: if (operand2 > *res)
2453: *res = operand2;
2454: break;
2455: default:
2456: abort();
2457: }
2458: }
2459: return(1);
2460: }
2461:
1.138 schwarze 2462: /* --- register management ------------------------------------------------ */
2463:
1.52 schwarze 2464: void
1.60 schwarze 2465: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 2466: {
1.52 schwarze 2467: struct roffreg *reg;
2468:
2469: /* Search for an existing register with the same name. */
2470: reg = r->regtab;
2471:
2472: while (reg && strcmp(name, reg->key.p))
2473: reg = reg->next;
1.41 schwarze 2474:
1.52 schwarze 2475: if (NULL == reg) {
2476: /* Create a new register. */
2477: reg = mandoc_malloc(sizeof(struct roffreg));
2478: reg->key.p = mandoc_strdup(name);
2479: reg->key.sz = strlen(name);
1.60 schwarze 2480: reg->val = 0;
1.52 schwarze 2481: reg->next = r->regtab;
2482: r->regtab = reg;
2483: }
2484:
1.60 schwarze 2485: if ('+' == sign)
2486: reg->val += val;
2487: else if ('-' == sign)
2488: reg->val -= val;
2489: else
2490: reg->val = val;
1.41 schwarze 2491: }
2492:
1.65 schwarze 2493: /*
2494: * Handle some predefined read-only number registers.
2495: * For now, return -1 if the requested register is not predefined;
2496: * in case a predefined read-only register having the value -1
2497: * were to turn up, another special value would have to be chosen.
2498: */
2499: static int
2500: roff_getregro(const char *name)
2501: {
2502:
2503: switch (*name) {
1.80 schwarze 2504: case 'A': /* ASCII approximation mode is always off. */
1.65 schwarze 2505: return(0);
1.80 schwarze 2506: case 'g': /* Groff compatibility mode is always on. */
1.65 schwarze 2507: return(1);
1.80 schwarze 2508: case 'H': /* Fixed horizontal resolution. */
1.65 schwarze 2509: return (24);
1.80 schwarze 2510: case 'j': /* Always adjust left margin only. */
1.65 schwarze 2511: return(0);
1.80 schwarze 2512: case 'T': /* Some output device is always defined. */
1.65 schwarze 2513: return(1);
1.80 schwarze 2514: case 'V': /* Fixed vertical resolution. */
1.65 schwarze 2515: return (40);
2516: default:
2517: return (-1);
2518: }
2519: }
2520:
1.53 schwarze 2521: int
1.52 schwarze 2522: roff_getreg(const struct roff *r, const char *name)
1.41 schwarze 2523: {
1.52 schwarze 2524: struct roffreg *reg;
1.65 schwarze 2525: int val;
2526:
2527: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
2528: val = roff_getregro(name + 1);
2529: if (-1 != val)
2530: return (val);
2531: }
1.52 schwarze 2532:
2533: for (reg = r->regtab; reg; reg = reg->next)
2534: if (0 == strcmp(name, reg->key.p))
1.53 schwarze 2535: return(reg->val);
2536:
2537: return(0);
2538: }
2539:
2540: static int
2541: roff_getregn(const struct roff *r, const char *name, size_t len)
2542: {
2543: struct roffreg *reg;
1.65 schwarze 2544: int val;
2545:
2546: if ('.' == name[0] && 2 == len) {
2547: val = roff_getregro(name + 1);
2548: if (-1 != val)
2549: return (val);
2550: }
1.53 schwarze 2551:
2552: for (reg = r->regtab; reg; reg = reg->next)
2553: if (len == reg->key.sz &&
2554: 0 == strncmp(name, reg->key.p, len))
2555: return(reg->val);
1.41 schwarze 2556:
1.52 schwarze 2557: return(0);
1.41 schwarze 2558: }
2559:
1.52 schwarze 2560: static void
2561: roff_freereg(struct roffreg *reg)
1.41 schwarze 2562: {
1.52 schwarze 2563: struct roffreg *old_reg;
1.41 schwarze 2564:
1.52 schwarze 2565: while (NULL != reg) {
2566: free(reg->key.p);
2567: old_reg = reg;
2568: reg = reg->next;
2569: free(old_reg);
2570: }
1.41 schwarze 2571: }
1.7 schwarze 2572:
2573: static enum rofferr
1.6 schwarze 2574: roff_nr(ROFF_ARGS)
1.1 schwarze 2575: {
1.85 schwarze 2576: char *key, *val;
2577: size_t keysz;
1.37 schwarze 2578: int iv;
1.60 schwarze 2579: char sign;
1.6 schwarze 2580:
1.110 schwarze 2581: key = val = buf->buf + pos;
2582: if (*key == '\0')
1.85 schwarze 2583: return(ROFF_IGN);
2584:
2585: keysz = roff_getname(r, &val, ln, pos);
1.110 schwarze 2586: if (key[keysz] == '\\')
1.85 schwarze 2587: return(ROFF_IGN);
2588: key[keysz] = '\0';
1.6 schwarze 2589:
1.60 schwarze 2590: sign = *val;
1.110 schwarze 2591: if (sign == '+' || sign == '-')
1.60 schwarze 2592: val++;
2593:
1.133 schwarze 2594: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77 schwarze 2595: roff_setreg(r, key, iv, sign);
1.1 schwarze 2596:
1.76 schwarze 2597: return(ROFF_IGN);
2598: }
2599:
2600: static enum rofferr
2601: roff_rr(ROFF_ARGS)
2602: {
2603: struct roffreg *reg, **prev;
1.85 schwarze 2604: char *name, *cp;
2605: size_t namesz;
1.76 schwarze 2606:
1.110 schwarze 2607: name = cp = buf->buf + pos;
2608: if (*name == '\0')
1.85 schwarze 2609: return(ROFF_IGN);
2610: namesz = roff_getname(r, &cp, ln, pos);
2611: name[namesz] = '\0';
1.76 schwarze 2612:
2613: prev = &r->regtab;
2614: while (1) {
2615: reg = *prev;
1.110 schwarze 2616: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 2617: break;
2618: prev = ®->next;
2619: }
1.110 schwarze 2620: if (reg != NULL) {
1.76 schwarze 2621: *prev = reg->next;
2622: free(reg->key.p);
2623: free(reg);
2624: }
1.29 schwarze 2625: return(ROFF_IGN);
2626: }
2627:
1.138 schwarze 2628: /* --- handler functions for roff requests -------------------------------- */
2629:
1.29 schwarze 2630: static enum rofferr
2631: roff_rm(ROFF_ARGS)
2632: {
2633: const char *name;
2634: char *cp;
1.85 schwarze 2635: size_t namesz;
1.29 schwarze 2636:
1.110 schwarze 2637: cp = buf->buf + pos;
2638: while (*cp != '\0') {
1.85 schwarze 2639: name = cp;
1.110 schwarze 2640: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 2641: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110 schwarze 2642: if (name[namesz] == '\\')
1.85 schwarze 2643: break;
1.29 schwarze 2644: }
1.51 schwarze 2645: return(ROFF_IGN);
2646: }
2647:
2648: static enum rofferr
2649: roff_it(ROFF_ARGS)
2650: {
2651: int iv;
2652:
2653: /* Parse the number of lines. */
1.133 schwarze 2654:
2655: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95 schwarze 2656: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110 schwarze 2657: ln, ppos, buf->buf + 1);
1.51 schwarze 2658: return(ROFF_IGN);
2659: }
2660:
1.134 schwarze 2661: while (isspace((unsigned char)buf->buf[pos]))
2662: pos++;
2663:
2664: /*
2665: * Arm the input line trap.
2666: * Special-casing "an-trap" is an ugly workaround to cope
2667: * with DocBook stupidly fiddling with man(7) internals.
2668: */
1.133 schwarze 2669:
1.51 schwarze 2670: roffit_lines = iv;
1.134 schwarze 2671: roffit_macro = mandoc_strdup(iv != 1 ||
2672: strcmp(buf->buf + pos, "an-trap") ?
2673: buf->buf + pos : "br");
1.2 schwarze 2674: return(ROFF_IGN);
1.47 schwarze 2675: }
2676:
2677: static enum rofferr
2678: roff_Dd(ROFF_ARGS)
2679: {
2680: const char *const *cp;
2681:
1.99 schwarze 2682: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47 schwarze 2683: for (cp = __mdoc_reserved; *cp; cp++)
2684: roff_setstr(r, *cp, NULL, 0);
2685:
1.99 schwarze 2686: if (r->format == 0)
2687: r->format = MPARSE_MDOC;
2688:
1.47 schwarze 2689: return(ROFF_CONT);
2690: }
2691:
2692: static enum rofferr
2693: roff_TH(ROFF_ARGS)
2694: {
2695: const char *const *cp;
2696:
1.99 schwarze 2697: if ((r->options & MPARSE_QUICK) == 0)
1.47 schwarze 2698: for (cp = __man_reserved; *cp; cp++)
2699: roff_setstr(r, *cp, NULL, 0);
2700:
1.99 schwarze 2701: if (r->format == 0)
2702: r->format = MPARSE_MAN;
2703:
1.47 schwarze 2704: return(ROFF_CONT);
1.14 schwarze 2705: }
2706:
2707: static enum rofferr
1.27 schwarze 2708: roff_TE(ROFF_ARGS)
2709: {
2710:
2711: if (NULL == r->tbl)
1.94 schwarze 2712: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2713: ln, ppos, "TE");
1.130 schwarze 2714: else if ( ! tbl_end(&r->tbl)) {
2715: free(buf->buf);
2716: buf->buf = mandoc_strdup(".sp");
2717: buf->sz = 4;
2718: return(ROFF_REPARSE);
2719: }
1.27 schwarze 2720: return(ROFF_IGN);
2721: }
2722:
2723: static enum rofferr
2724: roff_T_(ROFF_ARGS)
2725: {
2726:
2727: if (NULL == r->tbl)
1.94 schwarze 2728: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2729: ln, ppos, "T&");
1.27 schwarze 2730: else
2731: tbl_restart(ppos, ln, r->tbl);
2732:
2733: return(ROFF_IGN);
2734: }
2735:
1.102 schwarze 2736: /*
2737: * Handle in-line equation delimiters.
2738: */
2739: static enum rofferr
1.110 schwarze 2740: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 2741: {
1.105 schwarze 2742: char *cp1, *cp2;
2743: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 2744:
1.102 schwarze 2745: /*
2746: * Outside equations, look for an opening delimiter.
2747: * If we are inside an equation, we already know it is
2748: * in-line, or this function wouldn't have been called;
2749: * so look for a closing delimiter.
2750: */
2751:
1.110 schwarze 2752: cp1 = buf->buf + pos;
1.102 schwarze 2753: cp2 = strchr(cp1, r->eqn == NULL ?
2754: r->last_eqn->odelim : r->last_eqn->cdelim);
2755: if (cp2 == NULL)
2756: return(ROFF_CONT);
2757:
1.105 schwarze 2758: *cp2++ = '\0';
2759: bef_pr = bef_nl = aft_nl = aft_pr = "";
2760:
2761: /* Handle preceding text, protecting whitespace. */
2762:
1.110 schwarze 2763: if (*buf->buf != '\0') {
1.105 schwarze 2764: if (r->eqn == NULL)
2765: bef_pr = "\\&";
2766: bef_nl = "\n";
2767: }
2768:
2769: /*
2770: * Prepare replacing the delimiter with an equation macro
2771: * and drop leading white space from the equation.
2772: */
1.102 schwarze 2773:
1.105 schwarze 2774: if (r->eqn == NULL) {
2775: while (*cp2 == ' ')
2776: cp2++;
2777: mac = ".EQ";
2778: } else
2779: mac = ".EN";
2780:
2781: /* Handle following text, protecting whitespace. */
2782:
2783: if (*cp2 != '\0') {
2784: aft_nl = "\n";
2785: if (r->eqn != NULL)
2786: aft_pr = "\\&";
2787: }
2788:
2789: /* Do the actual replacement. */
2790:
1.110 schwarze 2791: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 2792: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 2793: free(buf->buf);
2794: buf->buf = cp1;
1.102 schwarze 2795:
2796: /* Toggle the in-line state of the eqn subsystem. */
2797:
2798: r->eqn_inline = r->eqn == NULL;
2799: return(ROFF_REPARSE);
1.41 schwarze 2800: }
2801:
1.107 schwarze 2802: static enum rofferr
2803: roff_EQ(ROFF_ARGS)
1.32 schwarze 2804: {
1.41 schwarze 2805: struct eqn_node *e;
1.32 schwarze 2806:
1.110 schwarze 2807: assert(r->eqn == NULL);
1.107 schwarze 2808: e = eqn_alloc(ppos, ln, r->parse);
1.32 schwarze 2809:
1.102 schwarze 2810: if (r->last_eqn) {
1.32 schwarze 2811: r->last_eqn->next = e;
1.102 schwarze 2812: e->delim = r->last_eqn->delim;
2813: e->odelim = r->last_eqn->odelim;
2814: e->cdelim = r->last_eqn->cdelim;
2815: } else
1.32 schwarze 2816: r->first_eqn = r->last_eqn = e;
2817:
2818: r->eqn = r->last_eqn = e;
1.41 schwarze 2819:
1.110 schwarze 2820: if (buf->buf[pos] != '\0')
1.107 schwarze 2821: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 2822: ".EQ %s", buf->buf + pos);
1.41 schwarze 2823:
1.32 schwarze 2824: return(ROFF_IGN);
2825: }
2826:
2827: static enum rofferr
2828: roff_EN(ROFF_ARGS)
2829: {
2830:
1.94 schwarze 2831: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.32 schwarze 2832: return(ROFF_IGN);
2833: }
2834:
2835: static enum rofferr
1.27 schwarze 2836: roff_TS(ROFF_ARGS)
2837: {
1.49 schwarze 2838: struct tbl_node *tbl;
1.27 schwarze 2839:
2840: if (r->tbl) {
1.94 schwarze 2841: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2842: ln, ppos, "TS breaks TS");
1.41 schwarze 2843: tbl_end(&r->tbl);
1.27 schwarze 2844: }
2845:
1.49 schwarze 2846: tbl = tbl_alloc(ppos, ln, r->parse);
1.27 schwarze 2847:
2848: if (r->last_tbl)
1.49 schwarze 2849: r->last_tbl->next = tbl;
1.27 schwarze 2850: else
1.49 schwarze 2851: r->first_tbl = r->last_tbl = tbl;
1.27 schwarze 2852:
1.49 schwarze 2853: r->tbl = r->last_tbl = tbl;
1.27 schwarze 2854: return(ROFF_IGN);
1.123 schwarze 2855: }
2856:
2857: static enum rofferr
2858: roff_brp(ROFF_ARGS)
2859: {
2860:
2861: buf->buf[pos - 1] = '\0';
2862: return(ROFF_CONT);
1.27 schwarze 2863: }
2864:
2865: static enum rofferr
1.48 schwarze 2866: roff_cc(ROFF_ARGS)
2867: {
2868: const char *p;
2869:
1.110 schwarze 2870: p = buf->buf + pos;
1.48 schwarze 2871:
1.110 schwarze 2872: if (*p == '\0' || (r->control = *p++) == '.')
1.48 schwarze 2873: r->control = 0;
2874:
1.110 schwarze 2875: if (*p != '\0')
1.132 schwarze 2876: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2877: ln, p - buf->buf, "cc ... %s", p);
1.48 schwarze 2878:
2879: return(ROFF_IGN);
2880: }
2881:
2882: static enum rofferr
1.42 schwarze 2883: roff_tr(ROFF_ARGS)
2884: {
2885: const char *p, *first, *second;
2886: size_t fsz, ssz;
2887: enum mandoc_esc esc;
2888:
1.110 schwarze 2889: p = buf->buf + pos;
1.42 schwarze 2890:
1.110 schwarze 2891: if (*p == '\0') {
1.132 schwarze 2892: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.42 schwarze 2893: return(ROFF_IGN);
2894: }
2895:
1.110 schwarze 2896: while (*p != '\0') {
1.42 schwarze 2897: fsz = ssz = 1;
2898:
2899: first = p++;
1.110 schwarze 2900: if (*first == '\\') {
1.42 schwarze 2901: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2902: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2903: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2904: ln, (int)(p - buf->buf), first);
1.42 schwarze 2905: return(ROFF_IGN);
2906: }
2907: fsz = (size_t)(p - first);
2908: }
2909:
2910: second = p++;
1.110 schwarze 2911: if (*second == '\\') {
1.42 schwarze 2912: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2913: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2914: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2915: ln, (int)(p - buf->buf), second);
1.42 schwarze 2916: return(ROFF_IGN);
2917: }
2918: ssz = (size_t)(p - second);
1.110 schwarze 2919: } else if (*second == '\0') {
1.132 schwarze 2920: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2921: ln, first - buf->buf, "tr %s", first);
1.42 schwarze 2922: second = " ";
2923: p--;
2924: }
2925:
2926: if (fsz > 1) {
1.80 schwarze 2927: roff_setstrn(&r->xmbtab, first, fsz,
2928: second, ssz, 0);
1.42 schwarze 2929: continue;
2930: }
2931:
1.110 schwarze 2932: if (r->xtab == NULL)
1.80 schwarze 2933: r->xtab = mandoc_calloc(128,
2934: sizeof(struct roffstr));
1.42 schwarze 2935:
2936: free(r->xtab[(int)*first].p);
2937: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2938: r->xtab[(int)*first].sz = ssz;
2939: }
2940:
2941: return(ROFF_IGN);
2942: }
2943:
2944: static enum rofferr
1.14 schwarze 2945: roff_so(ROFF_ARGS)
2946: {
1.121 schwarze 2947: char *name, *cp;
1.15 schwarze 2948:
1.110 schwarze 2949: name = buf->buf + pos;
1.97 schwarze 2950: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14 schwarze 2951:
1.22 schwarze 2952: /*
2953: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2954: * opening anything that's not in our cwd or anything beneath
2955: * it. Thus, explicitly disallow traversing up the file-system
2956: * or using absolute paths.
2957: */
2958:
1.110 schwarze 2959: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83 schwarze 2960: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2961: ".so %s", name);
1.121 schwarze 2962: buf->sz = mandoc_asprintf(&cp,
2963: ".sp\nSee the file %s.\n.sp", name) + 1;
2964: free(buf->buf);
2965: buf->buf = cp;
2966: *offs = 0;
2967: return(ROFF_REPARSE);
1.14 schwarze 2968: }
2969:
2970: *offs = pos;
2971: return(ROFF_SO);
1.7 schwarze 2972: }
2973:
1.138 schwarze 2974: /* --- user defined strings and macros ------------------------------------ */
2975:
1.16 schwarze 2976: static enum rofferr
2977: roff_userdef(ROFF_ARGS)
1.12 schwarze 2978: {
1.135 schwarze 2979: const char *arg[9], *ap;
1.16 schwarze 2980: char *cp, *n1, *n2;
1.25 schwarze 2981: int i;
1.135 schwarze 2982: size_t asz, rsz;
1.12 schwarze 2983:
1.16 schwarze 2984: /*
2985: * Collect pointers to macro argument strings
1.61 schwarze 2986: * and NUL-terminate them.
1.16 schwarze 2987: */
1.135 schwarze 2988:
1.110 schwarze 2989: cp = buf->buf + pos;
1.25 schwarze 2990: for (i = 0; i < 9; i++)
1.110 schwarze 2991: arg[i] = *cp == '\0' ? "" :
1.35 schwarze 2992: mandoc_getarg(r->parse, &cp, ln, &pos);
1.16 schwarze 2993:
2994: /*
2995: * Expand macro arguments.
1.12 schwarze 2996: */
1.135 schwarze 2997:
2998: buf->sz = strlen(r->current_string) + 1;
2999: n1 = cp = mandoc_malloc(buf->sz);
3000: memcpy(n1, r->current_string, buf->sz);
3001: while (*cp != '\0') {
3002:
3003: /* Scan ahead for the next argument invocation. */
3004:
3005: if (*cp++ != '\\')
3006: continue;
3007: if (*cp++ != '$')
3008: continue;
3009: i = *cp - '1';
3010: if (0 > i || 8 < i)
1.16 schwarze 3011: continue;
1.135 schwarze 3012: cp -= 2;
3013:
3014: /*
3015: * Determine the size of the expanded argument,
3016: * taking escaping of quotes into account.
3017: */
3018:
3019: asz = 0;
3020: for (ap = arg[i]; *ap != '\0'; ap++) {
3021: asz++;
3022: if (*ap == '"')
3023: asz += 3;
3024: }
3025: if (asz != 3) {
3026:
3027: /*
3028: * Determine the size of the rest of the
3029: * unexpanded macro, including the NUL.
3030: */
3031:
3032: rsz = buf->sz - (cp - n1) - 3;
3033:
3034: /*
3035: * When shrinking, move before
3036: * releasing the storage.
3037: */
3038:
3039: if (asz < 3)
3040: memmove(cp + asz, cp + 3, rsz);
3041:
3042: /*
3043: * Resize the storage for the macro
3044: * and readjust the parse pointer.
3045: */
3046:
3047: buf->sz += asz - 3;
3048: n2 = mandoc_realloc(n1, buf->sz);
3049: cp = n2 + (cp - n1);
3050: n1 = n2;
3051:
3052: /*
3053: * When growing, make room
3054: * for the expanded argument.
3055: */
3056:
3057: if (asz > 3)
3058: memmove(cp + asz, cp + 3, rsz);
3059: }
3060:
3061: /* Copy the expanded argument, escaping quotes. */
3062:
3063: n2 = cp;
3064: for (ap = arg[i]; *ap != '\0'; ap++) {
3065: if (*ap == '"') {
3066: memcpy(n2, "\\(dq", 4);
3067: n2 += 4;
3068: } else
3069: *n2++ = *ap;
1.16 schwarze 3070: }
1.12 schwarze 3071: }
3072:
1.16 schwarze 3073: /*
3074: * Replace the macro invocation
3075: * by the expanded macro.
3076: */
1.135 schwarze 3077:
1.110 schwarze 3078: free(buf->buf);
3079: buf->buf = n1;
1.120 schwarze 3080: *offs = 0;
1.16 schwarze 3081:
1.110 schwarze 3082: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.16 schwarze 3083: ROFF_REPARSE : ROFF_APPEND);
1.12 schwarze 3084: }
1.28 schwarze 3085:
1.85 schwarze 3086: static size_t
1.28 schwarze 3087: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3088: {
3089: char *name, *cp;
1.85 schwarze 3090: size_t namesz;
1.28 schwarze 3091:
3092: name = *cpp;
3093: if ('\0' == *name)
1.85 schwarze 3094: return(0);
1.28 schwarze 3095:
1.85 schwarze 3096: /* Read until end of name and terminate it with NUL. */
3097: for (cp = name; 1; cp++) {
3098: if ('\0' == *cp || ' ' == *cp) {
3099: namesz = cp - name;
3100: break;
3101: }
1.28 schwarze 3102: if ('\\' != *cp)
3103: continue;
1.88 schwarze 3104: namesz = cp - name;
3105: if ('{' == cp[1] || '}' == cp[1])
3106: break;
1.28 schwarze 3107: cp++;
3108: if ('\\' == *cp)
3109: continue;
1.97 schwarze 3110: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3111: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 3112: mandoc_escape((const char **)&cp, NULL, NULL);
3113: break;
1.28 schwarze 3114: }
3115:
3116: /* Read past spaces. */
3117: while (' ' == *cp)
3118: cp++;
3119:
3120: *cpp = cp;
1.85 schwarze 3121: return(namesz);
1.28 schwarze 3122: }
3123:
1.16 schwarze 3124: /*
3125: * Store *string into the user-defined string called *name.
3126: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 3127: * append == 0: replace mode
3128: * append == 1: single-line append mode
3129: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 3130: */
1.8 schwarze 3131: static void
1.16 schwarze 3132: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 3133: int append)
1.7 schwarze 3134: {
1.42 schwarze 3135:
3136: roff_setstrn(&r->strtab, name, strlen(name), string,
1.80 schwarze 3137: string ? strlen(string) : 0, append);
1.42 schwarze 3138: }
3139:
3140: static void
3141: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 3142: const char *string, size_t stringsz, int append)
1.42 schwarze 3143: {
3144: struct roffkv *n;
3145: char *c;
3146: int i;
3147: size_t oldch, newch;
1.7 schwarze 3148:
1.16 schwarze 3149: /* Search for an existing string with the same name. */
1.42 schwarze 3150: n = *r;
3151:
1.84 schwarze 3152: while (n && (namesz != n->key.sz ||
3153: strncmp(n->key.p, name, namesz)))
1.7 schwarze 3154: n = n->next;
1.8 schwarze 3155:
3156: if (NULL == n) {
1.16 schwarze 3157: /* Create a new string table entry. */
1.42 schwarze 3158: n = mandoc_malloc(sizeof(struct roffkv));
3159: n->key.p = mandoc_strndup(name, namesz);
3160: n->key.sz = namesz;
3161: n->val.p = NULL;
3162: n->val.sz = 0;
3163: n->next = *r;
3164: *r = n;
1.66 schwarze 3165: } else if (0 == append) {
1.42 schwarze 3166: free(n->val.p);
3167: n->val.p = NULL;
3168: n->val.sz = 0;
1.16 schwarze 3169: }
3170:
3171: if (NULL == string)
3172: return;
3173:
3174: /*
3175: * One additional byte for the '\n' in multiline mode,
3176: * and one for the terminating '\0'.
3177: */
1.66 schwarze 3178: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 3179:
3180: if (NULL == n->val.p) {
3181: n->val.p = mandoc_malloc(newch);
3182: *n->val.p = '\0';
1.16 schwarze 3183: oldch = 0;
3184: } else {
1.42 schwarze 3185: oldch = n->val.sz;
3186: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 3187: }
3188:
3189: /* Skip existing content in the destination buffer. */
1.42 schwarze 3190: c = n->val.p + (int)oldch;
1.16 schwarze 3191:
3192: /* Append new content to the destination buffer. */
1.42 schwarze 3193: i = 0;
3194: while (i < (int)stringsz) {
1.16 schwarze 3195: /*
3196: * Rudimentary roff copy mode:
3197: * Handle escaped backslashes.
3198: */
1.42 schwarze 3199: if ('\\' == string[i] && '\\' == string[i + 1])
3200: i++;
3201: *c++ = string[i++];
1.16 schwarze 3202: }
1.8 schwarze 3203:
1.16 schwarze 3204: /* Append terminating bytes. */
1.66 schwarze 3205: if (1 < append)
1.16 schwarze 3206: *c++ = '\n';
1.42 schwarze 3207:
1.16 schwarze 3208: *c = '\0';
1.42 schwarze 3209: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 3210: }
3211:
1.8 schwarze 3212: static const char *
3213: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7 schwarze 3214: {
1.42 schwarze 3215: const struct roffkv *n;
1.64 schwarze 3216: int i;
1.7 schwarze 3217:
1.42 schwarze 3218: for (n = r->strtab; n; n = n->next)
1.80 schwarze 3219: if (0 == strncmp(name, n->key.p, len) &&
3220: '\0' == n->key.p[(int)len])
1.42 schwarze 3221: return(n->val.p);
1.64 schwarze 3222:
3223: for (i = 0; i < PREDEFS_MAX; i++)
3224: if (0 == strncmp(name, predefs[i].name, len) &&
3225: '\0' == predefs[i].name[(int)len])
3226: return(predefs[i].str);
1.8 schwarze 3227:
1.42 schwarze 3228: return(NULL);
1.7 schwarze 3229: }
3230:
1.8 schwarze 3231: static void
1.42 schwarze 3232: roff_freestr(struct roffkv *r)
1.7 schwarze 3233: {
1.42 schwarze 3234: struct roffkv *n, *nn;
1.7 schwarze 3235:
1.42 schwarze 3236: for (n = r; n; n = nn) {
3237: free(n->key.p);
3238: free(n->val.p);
1.7 schwarze 3239: nn = n->next;
3240: free(n);
3241: }
1.27 schwarze 3242: }
1.138 schwarze 3243:
3244: /* --- accessors and utility functions ------------------------------------ */
1.27 schwarze 3245:
3246: const struct tbl_span *
3247: roff_span(const struct roff *r)
3248: {
1.80 schwarze 3249:
1.27 schwarze 3250: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32 schwarze 3251: }
3252:
3253: const struct eqn *
3254: roff_eqn(const struct roff *r)
3255: {
1.80 schwarze 3256:
1.32 schwarze 3257: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.42 schwarze 3258: }
3259:
3260: /*
3261: * Duplicate an input string, making the appropriate character
3262: * conversations (as stipulated by `tr') along the way.
3263: * Returns a heap-allocated string with all the replacements made.
3264: */
3265: char *
3266: roff_strdup(const struct roff *r, const char *p)
3267: {
3268: const struct roffkv *cp;
3269: char *res;
3270: const char *pp;
3271: size_t ssz, sz;
3272: enum mandoc_esc esc;
3273:
3274: if (NULL == r->xmbtab && NULL == r->xtab)
3275: return(mandoc_strdup(p));
3276: else if ('\0' == *p)
3277: return(mandoc_strdup(""));
3278:
3279: /*
3280: * Step through each character looking for term matches
3281: * (remember that a `tr' can be invoked with an escape, which is
3282: * a glyph but the escape is multi-character).
3283: * We only do this if the character hash has been initialised
3284: * and the string is >0 length.
3285: */
3286:
3287: res = NULL;
3288: ssz = 0;
3289:
3290: while ('\0' != *p) {
3291: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3292: sz = r->xtab[(int)*p].sz;
3293: res = mandoc_realloc(res, ssz + sz + 1);
3294: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3295: ssz += sz;
3296: p++;
3297: continue;
3298: } else if ('\\' != *p) {
3299: res = mandoc_realloc(res, ssz + 2);
3300: res[ssz++] = *p++;
3301: continue;
3302: }
3303:
3304: /* Search for term matches. */
3305: for (cp = r->xmbtab; cp; cp = cp->next)
3306: if (0 == strncmp(p, cp->key.p, cp->key.sz))
3307: break;
3308:
3309: if (NULL != cp) {
3310: /*
3311: * A match has been found.
3312: * Append the match to the array and move
3313: * forward by its keysize.
3314: */
1.80 schwarze 3315: res = mandoc_realloc(res,
3316: ssz + cp->val.sz + 1);
1.42 schwarze 3317: memcpy(res + ssz, cp->val.p, cp->val.sz);
3318: ssz += cp->val.sz;
3319: p += (int)cp->key.sz;
3320: continue;
3321: }
3322:
3323: /*
3324: * Handle escapes carefully: we need to copy
3325: * over just the escape itself, or else we might
3326: * do replacements within the escape itself.
3327: * Make sure to pass along the bogus string.
3328: */
3329: pp = p++;
3330: esc = mandoc_escape(&p, NULL, NULL);
3331: if (ESCAPE_ERROR == esc) {
3332: sz = strlen(pp);
3333: res = mandoc_realloc(res, ssz + sz + 1);
3334: memcpy(res + ssz, pp, sz);
3335: break;
3336: }
1.80 schwarze 3337: /*
3338: * We bail out on bad escapes.
1.42 schwarze 3339: * No need to warn: we already did so when
3340: * roff_res() was called.
3341: */
3342: sz = (int)(p - pp);
3343: res = mandoc_realloc(res, ssz + sz + 1);
3344: memcpy(res + ssz, pp, sz);
3345: ssz += sz;
3346: }
3347:
3348: res[(int)ssz] = '\0';
3349: return(res);
1.99 schwarze 3350: }
3351:
3352: int
3353: roff_getformat(const struct roff *r)
3354: {
3355:
3356: return(r->format);
1.48 schwarze 3357: }
3358:
3359: /*
1.80 schwarze 3360: * Find out whether a line is a macro line or not.
1.48 schwarze 3361: * If it is, adjust the current position and return one; if it isn't,
3362: * return zero and don't change the current position.
3363: * If the control character has been set with `.cc', then let that grain
3364: * precedence.
3365: * This is slighly contrary to groff, where using the non-breaking
3366: * control character when `cc' has been invoked will cause the
3367: * non-breaking macro contents to be printed verbatim.
3368: */
3369: int
3370: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3371: {
3372: int pos;
3373:
3374: pos = *ppos;
3375:
3376: if (0 != r->control && cp[pos] == r->control)
3377: pos++;
3378: else if (0 != r->control)
3379: return(0);
3380: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3381: pos += 2;
3382: else if ('.' == cp[pos] || '\'' == cp[pos])
3383: pos++;
3384: else
3385: return(0);
3386:
3387: while (' ' == cp[pos] || '\t' == cp[pos])
3388: pos++;
3389:
3390: *ppos = pos;
3391: return(1);
1.1 schwarze 3392: }
![[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