Annotation of src/usr.bin/mandoc/roff.c, Revision 1.125
1.125 ! schwarze 1: /* $OpenBSD: roff.c,v 1.124 2015/01/21 02:16:11 schwarze Exp $ */
1.1 schwarze 2: /*
1.119 schwarze 3: * Copyright (c) 2010, 2011, 2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
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"
1.98 daniel 29: #include "libmandoc.h"
1.27 schwarze 30: #include "libroff.h"
1.1 schwarze 31:
1.37 schwarze 32: /* Maximum number of nested if-else conditionals. */
1.2 schwarze 33: #define RSTACK_MAX 128
34:
1.43 schwarze 35: /* Maximum number of string expansions per line, to break infinite loops. */
36: #define EXPAND_LIMIT 1000
37:
1.1 schwarze 38: enum rofft {
1.123 schwarze 39: ROFF_ab,
1.20 schwarze 40: ROFF_ad,
1.123 schwarze 41: ROFF_af,
42: ROFF_aln,
43: ROFF_als,
1.2 schwarze 44: ROFF_am,
1.123 schwarze 45: ROFF_am1,
1.2 schwarze 46: ROFF_ami,
1.123 schwarze 47: ROFF_ami1,
1.66 schwarze 48: ROFF_as,
1.123 schwarze 49: ROFF_as1,
50: ROFF_asciify,
51: ROFF_backtrace,
52: ROFF_bd,
53: ROFF_bleedat,
54: ROFF_blm,
55: ROFF_box,
56: ROFF_boxa,
57: ROFF_bp,
58: ROFF_BP,
59: /* MAN_br, MDOC_br */
60: ROFF_break,
61: ROFF_breakchar,
62: ROFF_brnl,
63: ROFF_brp,
64: ROFF_brpnl,
65: ROFF_c2,
1.48 schwarze 66: ROFF_cc,
1.67 schwarze 67: ROFF_ce,
1.123 schwarze 68: ROFF_cf,
69: ROFF_cflags,
70: ROFF_ch,
71: ROFF_char,
72: ROFF_chop,
73: ROFF_class,
74: ROFF_close,
75: ROFF_CL,
76: ROFF_color,
77: ROFF_composite,
78: ROFF_continue,
79: ROFF_cp,
80: ROFF_cropat,
81: ROFF_cs,
82: ROFF_cu,
83: ROFF_da,
84: ROFF_dch,
85: ROFF_Dd,
1.1 schwarze 86: ROFF_de,
1.123 schwarze 87: ROFF_de1,
88: ROFF_defcolor,
1.1 schwarze 89: ROFF_dei,
1.123 schwarze 90: ROFF_dei1,
91: ROFF_device,
92: ROFF_devicem,
93: ROFF_di,
94: ROFF_do,
1.2 schwarze 95: ROFF_ds,
1.123 schwarze 96: ROFF_ds1,
97: ROFF_dwh,
98: ROFF_dt,
99: ROFF_ec,
100: ROFF_ecr,
101: ROFF_ecs,
1.2 schwarze 102: ROFF_el,
1.123 schwarze 103: ROFF_em,
104: ROFF_EN,
105: ROFF_eo,
106: ROFF_EP,
107: ROFF_EQ,
108: ROFF_errprint,
109: ROFF_ev,
110: ROFF_evc,
111: ROFF_ex,
112: ROFF_fallback,
1.58 schwarze 113: ROFF_fam,
1.123 schwarze 114: ROFF_fc,
115: ROFF_fchar,
116: ROFF_fcolor,
117: ROFF_fdeferlig,
118: ROFF_feature,
119: /* MAN_fi; ignored in mdoc(7) */
120: ROFF_fkern,
121: ROFF_fl,
122: ROFF_flig,
123: ROFF_fp,
124: ROFF_fps,
125: ROFF_fschar,
126: ROFF_fspacewidth,
127: ROFF_fspecial,
128: /* MAN_ft; ignored in mdoc(7) */
129: ROFF_ftr,
130: ROFF_fzoom,
131: ROFF_gcolor,
132: ROFF_hc,
133: ROFF_hcode,
134: ROFF_hidechar,
135: ROFF_hla,
136: ROFF_hlm,
137: ROFF_hpf,
138: ROFF_hpfa,
139: ROFF_hpfcode,
1.59 schwarze 140: ROFF_hw,
1.20 schwarze 141: ROFF_hy,
1.123 schwarze 142: ROFF_hylang,
143: ROFF_hylen,
144: ROFF_hym,
145: ROFF_hypp,
146: ROFF_hys,
1.2 schwarze 147: ROFF_ie,
148: ROFF_if,
1.1 schwarze 149: ROFF_ig,
1.123 schwarze 150: /* MAN_in; ignored in mdoc(7) */
151: ROFF_index,
1.30 schwarze 152: ROFF_it,
1.123 schwarze 153: ROFF_itc,
154: ROFF_IX,
155: ROFF_kern,
156: ROFF_kernafter,
157: ROFF_kernbefore,
158: ROFF_kernpair,
159: ROFF_lc,
160: ROFF_lc_ctype,
161: ROFF_lds,
162: ROFF_length,
163: ROFF_letadj,
164: ROFF_lf,
165: ROFF_lg,
166: ROFF_lhang,
167: ROFF_linetabs,
168: /* MAN_ll, MDOC_ll */
169: ROFF_lnr,
170: ROFF_lnrf,
171: ROFF_lpfx,
172: ROFF_ls,
173: ROFF_lsm,
174: ROFF_lt,
175: ROFF_mc,
176: ROFF_mediasize,
177: ROFF_minss,
178: ROFF_mk,
179: ROFF_mso,
180: /* MAN_na; ignored in mdoc(7) */
1.20 schwarze 181: ROFF_ne,
1.123 schwarze 182: /* MAN_nf; ignored in mdoc(7) */
1.20 schwarze 183: ROFF_nh,
1.123 schwarze 184: ROFF_nhychar,
185: ROFF_nm,
186: ROFF_nn,
187: ROFF_nop,
1.14 schwarze 188: ROFF_nr,
1.123 schwarze 189: ROFF_nrf,
190: ROFF_nroff,
1.31 schwarze 191: ROFF_ns,
1.123 schwarze 192: ROFF_nx,
193: ROFF_open,
194: ROFF_opena,
195: ROFF_os,
196: ROFF_output,
197: ROFF_padj,
198: ROFF_papersize,
199: ROFF_pc,
200: ROFF_pev,
201: ROFF_pi,
202: ROFF_PI,
1.101 schwarze 203: ROFF_pl,
1.123 schwarze 204: ROFF_pm,
205: ROFF_pn,
206: ROFF_pnr,
207: ROFF_po,
1.31 schwarze 208: ROFF_ps,
1.123 schwarze 209: ROFF_psbb,
210: ROFF_pshape,
211: ROFF_pso,
212: ROFF_ptr,
213: ROFF_pvs,
214: ROFF_rchar,
215: ROFF_rd,
216: ROFF_recursionlimit,
217: ROFF_return,
218: ROFF_rfschar,
219: ROFF_rhang,
220: ROFF_rj,
1.2 schwarze 221: ROFF_rm,
1.123 schwarze 222: ROFF_rn,
223: ROFF_rnn,
1.76 schwarze 224: ROFF_rr,
1.123 schwarze 225: ROFF_rs,
226: ROFF_rt,
227: ROFF_schar,
228: ROFF_sentchar,
229: ROFF_shc,
230: ROFF_shift,
231: ROFF_sizes,
1.14 schwarze 232: ROFF_so,
1.123 schwarze 233: /* MAN_sp, MDOC_sp */
234: ROFF_spacewidth,
235: ROFF_special,
236: ROFF_spreadwarn,
237: ROFF_ss,
238: ROFF_sty,
239: ROFF_substring,
240: ROFF_sv,
241: ROFF_sy,
242: ROFF_T_,
1.31 schwarze 243: ROFF_ta,
1.123 schwarze 244: ROFF_tc,
245: ROFF_TE,
246: ROFF_TH,
247: ROFF_ti,
248: ROFF_tkf,
249: ROFF_tl,
250: ROFF_tm,
251: ROFF_tm1,
252: ROFF_tmc,
1.2 schwarze 253: ROFF_tr,
1.123 schwarze 254: ROFF_track,
255: ROFF_transchar,
256: ROFF_trf,
257: ROFF_trimat,
258: ROFF_trin,
259: ROFF_trnt,
260: ROFF_troff,
1.27 schwarze 261: ROFF_TS,
1.123 schwarze 262: ROFF_uf,
263: ROFF_ul,
264: ROFF_unformat,
265: ROFF_unwatch,
266: ROFF_unwatchn,
267: ROFF_vpt,
268: ROFF_vs,
269: ROFF_warn,
270: ROFF_warnscale,
271: ROFF_watch,
272: ROFF_watchlength,
273: ROFF_watchn,
274: ROFF_wh,
275: ROFF_while,
276: ROFF_write,
277: ROFF_writec,
278: ROFF_writem,
279: ROFF_xflag,
1.2 schwarze 280: ROFF_cblock,
1.16 schwarze 281: ROFF_USERDEF,
1.1 schwarze 282: ROFF_MAX
283: };
284:
1.41 schwarze 285: /*
1.42 schwarze 286: * An incredibly-simple string buffer.
287: */
1.8 schwarze 288: struct roffstr {
1.42 schwarze 289: char *p; /* nil-terminated buffer */
290: size_t sz; /* saved strlen(p) */
291: };
292:
293: /*
294: * A key-value roffstr pair as part of a singly-linked list.
295: */
296: struct roffkv {
297: struct roffstr key;
298: struct roffstr val;
299: struct roffkv *next; /* next in list */
1.8 schwarze 300: };
301:
1.52 schwarze 302: /*
303: * A single number register as part of a singly-linked list.
304: */
305: struct roffreg {
306: struct roffstr key;
1.53 schwarze 307: int val;
1.52 schwarze 308: struct roffreg *next;
309: };
310:
1.1 schwarze 311: struct roff {
1.35 schwarze 312: struct mparse *parse; /* parse point */
1.109 schwarze 313: const struct mchars *mchars; /* character table */
1.1 schwarze 314: struct roffnode *last; /* leaf of stack */
1.96 schwarze 315: int *rstack; /* stack of inverted `ie' values */
1.52 schwarze 316: struct roffreg *regtab; /* number registers */
1.42 schwarze 317: struct roffkv *strtab; /* user-defined strings & macros */
318: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
319: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.16 schwarze 320: const char *current_string; /* value of last called user macro */
1.27 schwarze 321: struct tbl_node *first_tbl; /* first table parsed */
322: struct tbl_node *last_tbl; /* last table parsed */
323: struct tbl_node *tbl; /* current table being parsed */
1.32 schwarze 324: struct eqn_node *last_eqn; /* last equation parsed */
325: struct eqn_node *first_eqn; /* first equation parsed */
326: struct eqn_node *eqn; /* current equation being parsed */
1.102 schwarze 327: int eqn_inline; /* current equation is inline */
1.96 schwarze 328: int options; /* parse options */
329: int rstacksz; /* current size limit of rstack */
330: int rstackpos; /* position in rstack */
1.99 schwarze 331: int format; /* current file in mdoc or man format */
1.96 schwarze 332: char control; /* control character */
1.1 schwarze 333: };
334:
335: struct roffnode {
336: enum rofft tok; /* type of node */
337: struct roffnode *parent; /* up one in stack */
338: int line; /* parse line */
339: int col; /* parse col */
1.16 schwarze 340: char *name; /* node name, e.g. macro name */
1.2 schwarze 341: char *end; /* end-rules: custom token */
342: int endspan; /* end-rules: next-line or infty */
1.71 schwarze 343: int rule; /* current evaluation rule */
1.1 schwarze 344: };
345:
346: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
347: enum rofft tok, /* tok of macro */ \
1.110 schwarze 348: struct buf *buf, /* input buffer */ \
1.1 schwarze 349: int ln, /* parse line */ \
1.2 schwarze 350: int ppos, /* original pos in buffer */ \
351: int pos, /* current pos in buffer */ \
352: int *offs /* reset offset of buffer data */
1.1 schwarze 353:
354: typedef enum rofferr (*roffproc)(ROFF_ARGS);
355:
356: struct roffmac {
357: const char *name; /* macro name */
1.2 schwarze 358: roffproc proc; /* process new macro */
359: roffproc text; /* process as child text of macro */
360: roffproc sub; /* process as child of macro */
361: int flags;
362: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.3 schwarze 363: struct roffmac *next;
1.1 schwarze 364: };
365:
1.37 schwarze 366: struct predef {
367: const char *name; /* predefined input name */
368: const char *str; /* replacement symbol */
369: };
370:
371: #define PREDEF(__name, __str) \
372: { (__name), (__str) },
373:
1.42 schwarze 374: static enum rofft roffhash_find(const char *, size_t);
375: static void roffhash_init(void);
376: static void roffnode_cleanscope(struct roff *);
377: static void roffnode_pop(struct roff *);
378: static void roffnode_push(struct roff *, enum rofft,
379: const char *, int, int);
1.2 schwarze 380: static enum rofferr roff_block(ROFF_ARGS);
381: static enum rofferr roff_block_text(ROFF_ARGS);
382: static enum rofferr roff_block_sub(ROFF_ARGS);
1.123 schwarze 383: static enum rofferr roff_brp(ROFF_ARGS);
1.2 schwarze 384: static enum rofferr roff_cblock(ROFF_ARGS);
1.48 schwarze 385: static enum rofferr roff_cc(ROFF_ARGS);
1.68 schwarze 386: static void roff_ccond(struct roff *, int, int);
1.2 schwarze 387: static enum rofferr roff_cond(ROFF_ARGS);
388: static enum rofferr roff_cond_text(ROFF_ARGS);
389: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.7 schwarze 390: static enum rofferr roff_ds(ROFF_ARGS);
1.110 schwarze 391: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.106 schwarze 392: static int roff_evalcond(struct roff *r, int,
393: const char *, int *);
394: static int roff_evalnum(struct roff *, int,
395: const char *, int *, int *, int);
396: static int roff_evalpar(struct roff *, int,
397: const char *, int *, int *);
1.71 schwarze 398: static int roff_evalstrcond(const char *, int *);
1.42 schwarze 399: static void roff_free1(struct roff *);
1.52 schwarze 400: static void roff_freereg(struct roffreg *);
1.42 schwarze 401: static void roff_freestr(struct roffkv *);
1.85 schwarze 402: static size_t roff_getname(struct roff *, char **, int, int);
1.56 schwarze 403: static int roff_getnum(const char *, int *, int *);
404: static int roff_getop(const char *, int *, char *);
1.53 schwarze 405: static int roff_getregn(const struct roff *,
406: const char *, size_t);
1.65 schwarze 407: static int roff_getregro(const char *name);
1.80 schwarze 408: static const char *roff_getstrn(const struct roff *,
1.8 schwarze 409: const char *, size_t);
1.123 schwarze 410: static enum rofferr roff_insec(ROFF_ARGS);
1.51 schwarze 411: static enum rofferr roff_it(ROFF_ARGS);
1.21 schwarze 412: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.6 schwarze 413: static enum rofferr roff_nr(ROFF_ARGS);
1.87 schwarze 414: static enum rofft roff_parse(struct roff *, char *, int *,
415: int, int);
1.110 schwarze 416: static enum rofferr roff_parsetext(struct buf *, int, int *);
417: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.29 schwarze 418: static enum rofferr roff_rm(ROFF_ARGS);
1.76 schwarze 419: static enum rofferr roff_rr(ROFF_ARGS);
1.8 schwarze 420: static void roff_setstr(struct roff *,
1.16 schwarze 421: const char *, const char *, int);
1.80 schwarze 422: static void roff_setstrn(struct roffkv **, const char *,
1.42 schwarze 423: size_t, const char *, size_t, int);
1.14 schwarze 424: static enum rofferr roff_so(ROFF_ARGS);
1.42 schwarze 425: static enum rofferr roff_tr(ROFF_ARGS);
1.47 schwarze 426: static enum rofferr roff_Dd(ROFF_ARGS);
427: static enum rofferr roff_TH(ROFF_ARGS);
1.27 schwarze 428: static enum rofferr roff_TE(ROFF_ARGS);
429: static enum rofferr roff_TS(ROFF_ARGS);
1.32 schwarze 430: static enum rofferr roff_EQ(ROFF_ARGS);
431: static enum rofferr roff_EN(ROFF_ARGS);
1.27 schwarze 432: static enum rofferr roff_T_(ROFF_ARGS);
1.123 schwarze 433: static enum rofferr roff_unsupp(ROFF_ARGS);
1.16 schwarze 434: static enum rofferr roff_userdef(ROFF_ARGS);
1.1 schwarze 435:
1.42 schwarze 436: /* See roffhash_find() */
1.3 schwarze 437:
438: #define ASCII_HI 126
439: #define ASCII_LO 33
440: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
441:
442: static struct roffmac *hash[HASHWIDTH];
443:
444: static struct roffmac roffs[ROFF_MAX] = {
1.123 schwarze 445: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.21 schwarze 446: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 447: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
448: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
449: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 450: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 451: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.3 schwarze 452: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 453: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.66 schwarze 454: { "as", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 455: { "as1", roff_ds, NULL, NULL, 0, NULL },
456: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
457: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
458: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
459: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
460: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
461: { "box", roff_unsupp, NULL, NULL, 0, NULL },
462: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
463: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
464: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
465: { "break", roff_unsupp, NULL, NULL, 0, NULL },
466: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
467: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
468: { "brp", roff_brp, NULL, NULL, 0, NULL },
469: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
470: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.48 schwarze 471: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.67 schwarze 472: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 473: { "cf", roff_insec, NULL, NULL, 0, NULL },
474: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
475: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
476: { "char", roff_unsupp, NULL, NULL, 0, NULL },
477: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
478: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
479: { "close", roff_insec, NULL, NULL, 0, NULL },
480: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
481: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
482: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
483: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
484: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
485: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
486: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
487: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
488: { "da", roff_unsupp, NULL, NULL, 0, NULL },
489: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
490: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.3 schwarze 491: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 492: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
493: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 494: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 495: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
496: { "device", roff_unsupp, NULL, NULL, 0, NULL },
497: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
498: { "di", roff_unsupp, NULL, NULL, 0, NULL },
499: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.7 schwarze 500: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 501: { "ds1", roff_ds, NULL, NULL, 0, NULL },
502: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
503: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
504: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
505: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
506: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 507: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.123 schwarze 508: { "em", roff_unsupp, NULL, NULL, 0, NULL },
509: { "EN", roff_EN, NULL, NULL, 0, NULL },
510: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
511: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
512: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
513: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
514: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
515: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
516: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
517: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.58 schwarze 518: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 519: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
520: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
521: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
522: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
523: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
524: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
525: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
526: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
527: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
528: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
529: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
530: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
531: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
532: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
533: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
534: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
535: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
536: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
538: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.59 schwarze 543: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 544: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 545: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 550: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
551: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
552: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 553: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.51 schwarze 554: { "it", roff_it, NULL, NULL, 0, NULL },
1.123 schwarze 555: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
556: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
557: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
558: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
559: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
560: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
561: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
562: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
563: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
564: { "length", roff_unsupp, NULL, NULL, 0, NULL },
565: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
566: { "lf", roff_insec, NULL, NULL, 0, NULL },
567: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
568: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
569: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
570: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
571: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
572: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
573: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
574: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
575: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
576: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
577: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
578: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
579: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
580: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.21 schwarze 581: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
582: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 583: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
584: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
585: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
586: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.14 schwarze 587: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.123 schwarze 588: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
589: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 590: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 591: { "nx", roff_insec, NULL, NULL, 0, NULL },
592: { "open", roff_insec, NULL, NULL, 0, NULL },
593: { "opena", roff_insec, NULL, NULL, 0, NULL },
594: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
595: { "output", roff_unsupp, NULL, NULL, 0, NULL },
596: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
597: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
598: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
599: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
600: { "pi", roff_insec, NULL, NULL, 0, NULL },
601: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.101 schwarze 602: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 603: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
604: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
605: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
606: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 607: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 608: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
609: { "pshape", roff_line_ignore, NULL, NULL, 0, NULL },
610: { "pso", roff_insec, NULL, NULL, 0, NULL },
611: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
612: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
614: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
615: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
616: { "return", roff_unsupp, NULL, NULL, 0, NULL },
617: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
618: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
619: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.29 schwarze 620: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.123 schwarze 621: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
622: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.76 schwarze 623: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.123 schwarze 624: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
625: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
626: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
627: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
628: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
629: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
630: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.14 schwarze 631: { "so", roff_so, NULL, NULL, 0, NULL },
1.123 schwarze 632: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
633: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
634: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
635: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
636: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
637: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
638: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
639: { "sy", roff_insec, NULL, NULL, 0, NULL },
640: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.31 schwarze 641: { "ta", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 642: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
643: { "TE", roff_TE, NULL, NULL, 0, NULL },
644: { "TH", roff_TH, NULL, NULL, 0, NULL },
645: { "ti", roff_line_ignore, NULL, NULL, 0, NULL },
646: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
647: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
648: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
649: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
650: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.42 schwarze 651: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.123 schwarze 652: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
653: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
654: { "trf", roff_insec, NULL, NULL, 0, NULL },
655: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
656: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
657: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
658: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.27 schwarze 659: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.123 schwarze 660: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
661: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
662: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
663: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
664: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
665: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
666: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
667: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
668: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
670: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
671: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
672: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
673: { "while", roff_unsupp, NULL, NULL, 0, NULL },
674: { "write", roff_insec, NULL, NULL, 0, NULL },
675: { "writec", roff_insec, NULL, NULL, 0, NULL },
676: { "writem", roff_insec, NULL, NULL, 0, NULL },
677: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 678: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.16 schwarze 679: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1 schwarze 680: };
681:
1.73 schwarze 682: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47 schwarze 683: const char *const __mdoc_reserved[] = {
684: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
685: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
686: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
687: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73 schwarze 688: "Dt", "Dv", "Dx", "D1",
689: "Ec", "Ed", "Ef", "Ek", "El", "Em",
690: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47 schwarze 691: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73 schwarze 692: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
693: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47 schwarze 694: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73 schwarze 695: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
696: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
697: "Sc", "Sh", "Sm", "So", "Sq",
1.47 schwarze 698: "Ss", "St", "Sx", "Sy",
699: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73 schwarze 700: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47 schwarze 701: "%P", "%Q", "%R", "%T", "%U", "%V",
702: NULL
703: };
704:
1.73 schwarze 705: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47 schwarze 706: const char *const __man_reserved[] = {
1.73 schwarze 707: "AT", "B", "BI", "BR", "DT",
708: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
709: "LP", "OP", "P", "PD", "PP",
710: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
711: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47 schwarze 712: NULL
713: };
714:
1.37 schwarze 715: /* Array of injected predefined strings. */
716: #define PREDEFS_MAX 38
717: static const struct predef predefs[PREDEFS_MAX] = {
718: #include "predefs.in"
719: };
720:
1.42 schwarze 721: /* See roffhash_find() */
1.3 schwarze 722: #define ROFF_HASH(p) (p[0] - ASCII_LO)
723:
1.51 schwarze 724: static int roffit_lines; /* number of lines to delay */
725: static char *roffit_macro; /* nil-terminated macro line */
726:
1.80 schwarze 727:
1.3 schwarze 728: static void
1.42 schwarze 729: roffhash_init(void)
1.3 schwarze 730: {
731: struct roffmac *n;
732: int buc, i;
733:
1.16 schwarze 734: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3 schwarze 735: assert(roffs[i].name[0] >= ASCII_LO);
736: assert(roffs[i].name[0] <= ASCII_HI);
737:
738: buc = ROFF_HASH(roffs[i].name);
739:
740: if (NULL != (n = hash[buc])) {
741: for ( ; n->next; n = n->next)
742: /* Do nothing. */ ;
743: n->next = &roffs[i];
744: } else
745: hash[buc] = &roffs[i];
746: }
747: }
748:
1.1 schwarze 749: /*
750: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
751: * the nil-terminated string name could be found.
752: */
753: static enum rofft
1.42 schwarze 754: roffhash_find(const char *p, size_t s)
1.1 schwarze 755: {
1.3 schwarze 756: int buc;
757: struct roffmac *n;
1.1 schwarze 758:
1.3 schwarze 759: /*
760: * libroff has an extremely simple hashtable, for the time
761: * being, which simply keys on the first character, which must
762: * be printable, then walks a chain. It works well enough until
763: * optimised.
764: */
765:
766: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
767: return(ROFF_MAX);
768:
769: buc = ROFF_HASH(p);
770:
771: if (NULL == (n = hash[buc]))
772: return(ROFF_MAX);
773: for ( ; n; n = n->next)
1.16 schwarze 774: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3 schwarze 775: return((enum rofft)(n - roffs));
1.1 schwarze 776:
777: return(ROFF_MAX);
778: }
779:
780: /*
781: * Pop the current node off of the stack of roff instructions currently
782: * pending.
783: */
784: static void
785: roffnode_pop(struct roff *r)
786: {
787: struct roffnode *p;
788:
1.2 schwarze 789: assert(r->last);
1.80 schwarze 790: p = r->last;
1.2 schwarze 791:
792: r->last = r->last->parent;
1.16 schwarze 793: free(p->name);
794: free(p->end);
1.1 schwarze 795: free(p);
796: }
797:
798: /*
799: * Push a roff node onto the instruction stack. This must later be
800: * removed with roffnode_pop().
801: */
1.11 schwarze 802: static void
1.16 schwarze 803: roffnode_push(struct roff *r, enum rofft tok, const char *name,
804: int line, int col)
1.1 schwarze 805: {
806: struct roffnode *p;
807:
1.11 schwarze 808: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 809: p->tok = tok;
1.16 schwarze 810: if (name)
811: p->name = mandoc_strdup(name);
1.1 schwarze 812: p->parent = r->last;
813: p->line = line;
814: p->col = col;
1.71 schwarze 815: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 816:
817: r->last = p;
818: }
819:
820: static void
821: roff_free1(struct roff *r)
822: {
1.49 schwarze 823: struct tbl_node *tbl;
1.32 schwarze 824: struct eqn_node *e;
1.42 schwarze 825: int i;
1.27 schwarze 826:
1.49 schwarze 827: while (NULL != (tbl = r->first_tbl)) {
828: r->first_tbl = tbl->next;
829: tbl_free(tbl);
1.27 schwarze 830: }
831: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 832:
1.32 schwarze 833: while (NULL != (e = r->first_eqn)) {
834: r->first_eqn = e->next;
835: eqn_free(e);
836: }
837: r->first_eqn = r->last_eqn = r->eqn = NULL;
838:
1.1 schwarze 839: while (r->last)
840: roffnode_pop(r);
1.27 schwarze 841:
1.96 schwarze 842: free (r->rstack);
843: r->rstack = NULL;
844: r->rstacksz = 0;
845: r->rstackpos = -1;
846:
847: roff_freereg(r->regtab);
848: r->regtab = NULL;
849:
1.42 schwarze 850: roff_freestr(r->strtab);
851: roff_freestr(r->xmbtab);
852: r->strtab = r->xmbtab = NULL;
853:
854: if (r->xtab)
855: for (i = 0; i < 128; i++)
856: free(r->xtab[i].p);
857: free(r->xtab);
858: r->xtab = NULL;
1.1 schwarze 859: }
860:
861: void
862: roff_reset(struct roff *r)
863: {
864:
865: roff_free1(r);
1.99 schwarze 866: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48 schwarze 867: r->control = 0;
1.1 schwarze 868: }
869:
870: void
871: roff_free(struct roff *r)
872: {
873:
874: roff_free1(r);
875: free(r);
876: }
877:
878: struct roff *
1.109 schwarze 879: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.1 schwarze 880: {
881: struct roff *r;
882:
1.11 schwarze 883: r = mandoc_calloc(1, sizeof(struct roff));
1.35 schwarze 884: r->parse = parse;
1.109 schwarze 885: r->mchars = mchars;
1.72 schwarze 886: r->options = options;
1.99 schwarze 887: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2 schwarze 888: r->rstackpos = -1;
1.80 schwarze 889:
1.42 schwarze 890: roffhash_init();
1.37 schwarze 891:
1.1 schwarze 892: return(r);
893: }
894:
1.8 schwarze 895: /*
1.79 schwarze 896: * In the current line, expand escape sequences that tend to get
897: * used in numerical expressions and conditional requests.
898: * Also check the syntax of the remaining escape sequences.
1.8 schwarze 899: */
1.45 schwarze 900: static enum rofferr
1.110 schwarze 901: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8 schwarze 902: {
1.81 schwarze 903: char ubuf[24]; /* buffer to print the number */
1.78 schwarze 904: const char *start; /* start of the string to process */
1.82 schwarze 905: char *stesc; /* start of an escape sequence ('\\') */
1.23 schwarze 906: const char *stnam; /* start of the name, after "[(*" */
907: const char *cp; /* end of the name, e.g. before ']' */
908: const char *res; /* the string to be substituted */
1.110 schwarze 909: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 910: size_t maxl; /* expected length of the escape name */
911: size_t naml; /* actual length of the escape name */
1.109 schwarze 912: enum mandoc_esc esc; /* type of the escape sequence */
913: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 914: int expand_count; /* to avoid infinite loops */
1.79 schwarze 915: int npos; /* position in numeric expression */
1.91 schwarze 916: int arg_complete; /* argument not interrupted by eol */
1.79 schwarze 917: char term; /* character terminating the escape */
1.8 schwarze 918:
1.43 schwarze 919: expand_count = 0;
1.110 schwarze 920: start = buf->buf + pos;
1.78 schwarze 921: stesc = strchr(start, '\0') - 1;
922: while (stesc-- > start) {
1.43 schwarze 923:
1.78 schwarze 924: /* Search backwards for the next backslash. */
925:
1.110 schwarze 926: if (*stesc != '\\')
1.78 schwarze 927: continue;
928:
929: /* If it is escaped, skip it. */
930:
931: for (cp = stesc - 1; cp >= start; cp--)
1.110 schwarze 932: if (*cp != '\\')
1.78 schwarze 933: break;
934:
1.110 schwarze 935: if ((stesc - cp) % 2 == 0) {
1.82 schwarze 936: stesc = (char *)cp;
1.78 schwarze 937: continue;
938: }
1.23 schwarze 939:
1.79 schwarze 940: /* Decide whether to expand or to check only. */
1.23 schwarze 941:
1.79 schwarze 942: term = '\0';
1.78 schwarze 943: cp = stesc + 1;
1.53 schwarze 944: switch (*cp) {
1.80 schwarze 945: case '*':
1.53 schwarze 946: res = NULL;
947: break;
1.80 schwarze 948: case 'B':
1.79 schwarze 949: /* FALLTHROUGH */
1.80 schwarze 950: case 'w':
1.79 schwarze 951: term = cp[1];
952: /* FALLTHROUGH */
1.80 schwarze 953: case 'n':
1.53 schwarze 954: res = ubuf;
955: break;
956: default:
1.109 schwarze 957: esc = mandoc_escape(&cp, &stnam, &inaml);
958: if (esc == ESCAPE_ERROR ||
959: (esc == ESCAPE_SPECIAL &&
960: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.92 schwarze 961: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110 schwarze 962: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 963: "%.*s", (int)(cp - stesc), stesc);
1.78 schwarze 964: continue;
1.42 schwarze 965: }
966:
1.78 schwarze 967: if (EXPAND_LIMIT < ++expand_count) {
968: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110 schwarze 969: ln, (int)(stesc - buf->buf), NULL);
1.78 schwarze 970: return(ROFF_IGN);
971: }
1.23 schwarze 972:
973: /*
974: * The third character decides the length
1.53 schwarze 975: * of the name of the string or register.
1.23 schwarze 976: * Save a pointer to the name.
977: */
978:
1.110 schwarze 979: if (term == '\0') {
1.79 schwarze 980: switch (*++cp) {
1.80 schwarze 981: case '\0':
1.79 schwarze 982: maxl = 0;
983: break;
1.80 schwarze 984: case '(':
1.79 schwarze 985: cp++;
986: maxl = 2;
987: break;
1.80 schwarze 988: case '[':
1.79 schwarze 989: cp++;
990: term = ']';
991: maxl = 0;
992: break;
993: default:
994: maxl = 1;
995: break;
996: }
997: } else {
998: cp += 2;
1.8 schwarze 999: maxl = 0;
1000: }
1.23 schwarze 1001: stnam = cp;
1.8 schwarze 1002:
1.23 schwarze 1003: /* Advance to the end of the name. */
1.8 schwarze 1004:
1.125 ! schwarze 1005: naml = 0;
1.91 schwarze 1006: arg_complete = 1;
1.125 ! schwarze 1007: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1008: if (*cp == '\0') {
1.92 schwarze 1009: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 1010: ln, (int)(stesc - buf->buf), stesc);
1.91 schwarze 1011: arg_complete = 0;
1.79 schwarze 1012: break;
1.42 schwarze 1013: }
1.110 schwarze 1014: if (maxl == 0 && *cp == term) {
1.79 schwarze 1015: cp++;
1.125 ! schwarze 1016: break;
! 1017: }
! 1018: if (*cp++ != '\\' || stesc[1] != 'w') {
! 1019: naml++;
! 1020: continue;
! 1021: }
! 1022: switch (mandoc_escape(&cp, NULL, NULL)) {
! 1023: case ESCAPE_SPECIAL:
! 1024: /* FALLTHROUGH */
! 1025: case ESCAPE_UNICODE:
! 1026: /* FALLTHROUGH */
! 1027: case ESCAPE_NUMBERED:
! 1028: /* FALLTHROUGH */
! 1029: case ESCAPE_OVERSTRIKE:
! 1030: naml++;
! 1031: break;
! 1032: default:
1.8 schwarze 1033: break;
1.79 schwarze 1034: }
1.8 schwarze 1035: }
1036:
1.23 schwarze 1037: /*
1038: * Retrieve the replacement string; if it is
1039: * undefined, resume searching for escapes.
1040: */
1041:
1.79 schwarze 1042: switch (stesc[1]) {
1.80 schwarze 1043: case '*':
1.91 schwarze 1044: if (arg_complete)
1045: res = roff_getstrn(r, stnam, naml);
1.79 schwarze 1046: break;
1.80 schwarze 1047: case 'B':
1.79 schwarze 1048: npos = 0;
1.91 schwarze 1049: ubuf[0] = arg_complete &&
1.106 schwarze 1050: roff_evalnum(r, ln, stnam, &npos, NULL, 0) &&
1.91 schwarze 1051: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1052: ubuf[1] = '\0';
1053: break;
1.80 schwarze 1054: case 'n':
1.91 schwarze 1055: if (arg_complete)
1056: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1057: roff_getregn(r, stnam, naml));
1058: else
1059: ubuf[0] = '\0';
1.79 schwarze 1060: break;
1.80 schwarze 1061: case 'w':
1.91 schwarze 1062: /* use even incomplete args */
1.81 schwarze 1063: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1064: 24 * (int)naml);
1065: break;
1066: }
1.8 schwarze 1067:
1.110 schwarze 1068: if (res == NULL) {
1.92 schwarze 1069: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1070: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1071: "%.*s", (int)naml, stnam);
1.37 schwarze 1072: res = "";
1.118 schwarze 1073: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1074: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1075: ln, (int)(stesc - buf->buf), NULL);
1076: return(ROFF_IGN);
1.8 schwarze 1077: }
1078:
1.23 schwarze 1079: /* Replace the escape sequence by the string. */
1080:
1.82 schwarze 1081: *stesc = '\0';
1.110 schwarze 1082: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1083: buf->buf, res, cp) + 1;
1.8 schwarze 1084:
1.78 schwarze 1085: /* Prepare for the next replacement. */
1.8 schwarze 1086:
1.78 schwarze 1087: start = nbuf + pos;
1.110 schwarze 1088: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1089: free(buf->buf);
1090: buf->buf = nbuf;
1.42 schwarze 1091: }
1.45 schwarze 1092: return(ROFF_CONT);
1.42 schwarze 1093: }
1094:
1095: /*
1.51 schwarze 1096: * Process text streams:
1097: * Convert all breakable hyphens into ASCII_HYPH.
1098: * Decrement and spring input line trap.
1.42 schwarze 1099: */
1100: static enum rofferr
1.110 schwarze 1101: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42 schwarze 1102: {
1103: size_t sz;
1104: const char *start;
1.51 schwarze 1105: char *p;
1106: int isz;
1.42 schwarze 1107: enum mandoc_esc esc;
1108:
1.110 schwarze 1109: start = p = buf->buf + pos;
1.42 schwarze 1110:
1.110 schwarze 1111: while (*p != '\0') {
1.42 schwarze 1112: sz = strcspn(p, "-\\");
1113: p += sz;
1114:
1.110 schwarze 1115: if (*p == '\0')
1.42 schwarze 1116: break;
1117:
1.110 schwarze 1118: if (*p == '\\') {
1.42 schwarze 1119: /* Skip over escapes. */
1120: p++;
1.62 schwarze 1121: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1122: if (esc == ESCAPE_ERROR)
1.42 schwarze 1123: break;
1124: continue;
1125: } else if (p == start) {
1126: p++;
1127: continue;
1128: }
1129:
1.44 schwarze 1130: if (isalpha((unsigned char)p[-1]) &&
1131: isalpha((unsigned char)p[1]))
1.42 schwarze 1132: *p = ASCII_HYPH;
1133: p++;
1.8 schwarze 1134: }
1135:
1.51 schwarze 1136: /* Spring the input line trap. */
1.110 schwarze 1137: if (roffit_lines == 1) {
1138: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1139: free(buf->buf);
1140: buf->buf = p;
1141: buf->sz = isz + 1;
1.51 schwarze 1142: *offs = 0;
1143: free(roffit_macro);
1144: roffit_lines = 0;
1145: return(ROFF_REPARSE);
1.110 schwarze 1146: } else if (roffit_lines > 1)
1.51 schwarze 1147: --roffit_lines;
1.42 schwarze 1148: return(ROFF_CONT);
1.8 schwarze 1149: }
1150:
1.1 schwarze 1151: enum rofferr
1.110 schwarze 1152: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1153: {
1154: enum rofft t;
1.27 schwarze 1155: enum rofferr e;
1.110 schwarze 1156: int pos; /* parse point */
1.115 schwarze 1157: int spos; /* saved parse point for messages */
1.110 schwarze 1158: int ppos; /* original offset in buf->buf */
1159: int ctl; /* macro line (boolean) */
1160:
1161: ppos = pos = *offs;
1.1 schwarze 1162:
1.102 schwarze 1163: /* Handle in-line equation delimiters. */
1164:
1.108 schwarze 1165: if (r->tbl == NULL &&
1166: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1167: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1168: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1169: if (e == ROFF_REPARSE)
1170: return(e);
1171: assert(e == ROFF_CONT);
1172: }
1173:
1174: /* Expand some escape sequences. */
1.8 schwarze 1175:
1.110 schwarze 1176: e = roff_res(r, buf, ln, pos);
1177: if (e == ROFF_IGN)
1.45 schwarze 1178: return(e);
1.110 schwarze 1179: assert(e == ROFF_CONT);
1.8 schwarze 1180:
1.110 schwarze 1181: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1182:
1.8 schwarze 1183: /*
1.2 schwarze 1184: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1185: * text through the macro's filter.
1186: * Equations process all content themselves.
1187: * Tables process almost all content themselves, but we want
1188: * to warn about macros before passing it there.
1.2 schwarze 1189: */
1190:
1.124 schwarze 1191: if (r->last != NULL && ! ctl) {
1.2 schwarze 1192: t = r->last->tok;
1193: assert(roffs[t].text);
1.110 schwarze 1194: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1195: assert(e == ROFF_IGN || e == ROFF_CONT);
1196: if (e != ROFF_CONT)
1.32 schwarze 1197: return(e);
1.54 schwarze 1198: }
1.124 schwarze 1199: if (r->eqn != NULL)
1.110 schwarze 1200: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.124 schwarze 1201: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1202: return(tbl_read(r->tbl, ln, buf->buf, pos));
1203: if ( ! ctl)
1.110 schwarze 1204: return(roff_parsetext(buf, pos, offs));
1.100 schwarze 1205:
1206: /* Skip empty request lines. */
1207:
1.110 schwarze 1208: if (buf->buf[pos] == '"') {
1.100 schwarze 1209: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1210: ln, pos, NULL);
1211: return(ROFF_IGN);
1.110 schwarze 1212: } else if (buf->buf[pos] == '\0')
1.100 schwarze 1213: return(ROFF_IGN);
1.2 schwarze 1214:
1215: /*
1216: * If a scope is open, go to the child handler for that macro,
1217: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1218: * Don't do so if an equation is open.
1.2 schwarze 1219: */
1220:
1221: if (r->last) {
1.1 schwarze 1222: t = r->last->tok;
1223: assert(roffs[t].sub);
1.110 schwarze 1224: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1225: }
1226:
1.115 schwarze 1227: /* No scope is open. This is a new request or macro. */
1228:
1229: spos = pos;
1230: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1231:
1232: /* Tables ignore most macros. */
1233:
1234: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1235: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1236: ln, pos, buf->buf + spos);
1237: return(ROFF_IGN);
1238: }
1239:
1.2 schwarze 1240: /*
1.115 schwarze 1241: * This is neither a roff request nor a user-defined macro.
1242: * Let the standard macro set parsers handle it.
1.2 schwarze 1243: */
1244:
1.115 schwarze 1245: if (t == ROFF_MAX)
1.1 schwarze 1246: return(ROFF_CONT);
1.115 schwarze 1247:
1248: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1249:
1.2 schwarze 1250: assert(roffs[t].proc);
1.110 schwarze 1251: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1252: }
1253:
1.27 schwarze 1254: void
1.2 schwarze 1255: roff_endparse(struct roff *r)
1256: {
1.1 schwarze 1257:
1.27 schwarze 1258: if (r->last)
1.94 schwarze 1259: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1260: r->last->line, r->last->col,
1261: roffs[r->last->tok].name);
1.27 schwarze 1262:
1.32 schwarze 1263: if (r->eqn) {
1.94 schwarze 1264: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1265: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41 schwarze 1266: eqn_end(&r->eqn);
1.32 schwarze 1267: }
1268:
1.27 schwarze 1269: if (r->tbl) {
1.94 schwarze 1270: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1271: r->tbl->line, r->tbl->pos, "TS");
1.41 schwarze 1272: tbl_end(&r->tbl);
1.27 schwarze 1273: }
1.1 schwarze 1274: }
1275:
1276: /*
1277: * Parse a roff node's type from the input buffer. This must be in the
1278: * form of ".foo xxx" in the usual way.
1279: */
1280: static enum rofft
1.87 schwarze 1281: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1282: {
1.87 schwarze 1283: char *cp;
1.16 schwarze 1284: const char *mac;
1285: size_t maclen;
1.1 schwarze 1286: enum rofft t;
1287:
1.87 schwarze 1288: cp = buf + *pos;
1289:
1290: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.1 schwarze 1291: return(ROFF_MAX);
1292:
1.87 schwarze 1293: mac = cp;
1294: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 1295:
1.16 schwarze 1296: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42 schwarze 1297: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1 schwarze 1298:
1.87 schwarze 1299: if (ROFF_MAX != t)
1300: *pos = cp - buf;
1.1 schwarze 1301:
1302: return(t);
1303: }
1304:
1305: static enum rofferr
1.2 schwarze 1306: roff_cblock(ROFF_ARGS)
1.1 schwarze 1307: {
1308:
1.2 schwarze 1309: /*
1310: * A block-close `..' should only be invoked as a child of an
1311: * ignore macro, otherwise raise a warning and just ignore it.
1312: */
1313:
1.110 schwarze 1314: if (r->last == NULL) {
1.94 schwarze 1315: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1316: ln, ppos, "..");
1.2 schwarze 1317: return(ROFF_IGN);
1318: }
1.1 schwarze 1319:
1.2 schwarze 1320: switch (r->last->tok) {
1.80 schwarze 1321: case ROFF_am:
1.93 schwarze 1322: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.2 schwarze 1323: /* FALLTHROUGH */
1.80 schwarze 1324: case ROFF_ami:
1.2 schwarze 1325: /* FALLTHROUGH */
1.80 schwarze 1326: case ROFF_de:
1.23 schwarze 1327: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2 schwarze 1328: /* FALLTHROUGH */
1.80 schwarze 1329: case ROFF_dei:
1.2 schwarze 1330: /* FALLTHROUGH */
1.80 schwarze 1331: case ROFF_ig:
1.2 schwarze 1332: break;
1333: default:
1.94 schwarze 1334: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1335: ln, ppos, "..");
1.1 schwarze 1336: return(ROFF_IGN);
1.2 schwarze 1337: }
1338:
1.110 schwarze 1339: if (buf->buf[pos] != '\0')
1.90 schwarze 1340: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 1341: ".. %s", buf->buf + pos);
1.2 schwarze 1342:
1343: roffnode_pop(r);
1344: roffnode_cleanscope(r);
1345: return(ROFF_IGN);
1346:
1347: }
1.1 schwarze 1348:
1.2 schwarze 1349: static void
1350: roffnode_cleanscope(struct roff *r)
1351: {
1.1 schwarze 1352:
1.2 schwarze 1353: while (r->last) {
1.46 schwarze 1354: if (--r->last->endspan != 0)
1.2 schwarze 1355: break;
1356: roffnode_pop(r);
1357: }
1358: }
1.1 schwarze 1359:
1.68 schwarze 1360: static void
1361: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 1362: {
1.1 schwarze 1363:
1.2 schwarze 1364: if (NULL == r->last) {
1.94 schwarze 1365: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1366: ln, ppos, "\\}");
1.68 schwarze 1367: return;
1.2 schwarze 1368: }
1.1 schwarze 1369:
1.2 schwarze 1370: switch (r->last->tok) {
1.80 schwarze 1371: case ROFF_el:
1.2 schwarze 1372: /* FALLTHROUGH */
1.80 schwarze 1373: case ROFF_ie:
1.2 schwarze 1374: /* FALLTHROUGH */
1.80 schwarze 1375: case ROFF_if:
1.2 schwarze 1376: break;
1377: default:
1.94 schwarze 1378: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1379: ln, ppos, "\\}");
1.68 schwarze 1380: return;
1.2 schwarze 1381: }
1.1 schwarze 1382:
1.2 schwarze 1383: if (r->last->endspan > -1) {
1.94 schwarze 1384: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1385: ln, ppos, "\\}");
1.68 schwarze 1386: return;
1.2 schwarze 1387: }
1388:
1389: roffnode_pop(r);
1390: roffnode_cleanscope(r);
1.68 schwarze 1391: return;
1.1 schwarze 1392: }
1393:
1394: static enum rofferr
1.2 schwarze 1395: roff_block(ROFF_ARGS)
1.1 schwarze 1396: {
1.93 schwarze 1397: const char *name;
1398: char *iname, *cp;
1.86 schwarze 1399: size_t namesz;
1.16 schwarze 1400:
1.93 schwarze 1401: /* Ignore groff compatibility mode for now. */
1.2 schwarze 1402:
1.110 schwarze 1403: if (tok == ROFF_de1)
1.93 schwarze 1404: tok = ROFF_de;
1.123 schwarze 1405: else if (tok == ROFF_dei1)
1406: tok = ROFF_dei;
1.110 schwarze 1407: else if (tok == ROFF_am1)
1.93 schwarze 1408: tok = ROFF_am;
1.123 schwarze 1409: else if (tok == ROFF_ami1)
1410: tok = ROFF_ami;
1.93 schwarze 1411:
1412: /* Parse the macro name argument. */
1413:
1.110 schwarze 1414: cp = buf->buf + pos;
1415: if (tok == ROFF_ig) {
1.93 schwarze 1416: iname = NULL;
1417: namesz = 0;
1418: } else {
1419: iname = cp;
1420: namesz = roff_getname(r, &cp, ln, ppos);
1421: iname[namesz] = '\0';
1422: }
1.22 schwarze 1423:
1.93 schwarze 1424: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 1425:
1.110 schwarze 1426: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1427: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1428: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1429: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1430: "%.*s", (int)namesz, iname);
1431: namesz = 0;
1432: } else
1433: namesz = strlen(name);
1434: } else
1435: name = iname;
1.22 schwarze 1436:
1.110 schwarze 1437: if (namesz == 0 && tok != ROFF_ig) {
1.93 schwarze 1438: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1439: ln, ppos, roffs[tok].name);
1440: return(ROFF_IGN);
1441: }
1.2 schwarze 1442:
1.16 schwarze 1443: roffnode_push(r, tok, name, ln, ppos);
1444:
1445: /*
1446: * At the beginning of a `de' macro, clear the existing string
1447: * with the same name, if there is one. New content will be
1.66 schwarze 1448: * appended from roff_block_text() in multiline mode.
1.16 schwarze 1449: */
1.22 schwarze 1450:
1.110 schwarze 1451: if (tok == ROFF_de || tok == ROFF_dei)
1.86 schwarze 1452: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2 schwarze 1453:
1.110 schwarze 1454: if (*cp == '\0')
1.2 schwarze 1455: return(ROFF_IGN);
1.1 schwarze 1456:
1.93 schwarze 1457: /* Get the custom end marker. */
1.22 schwarze 1458:
1.93 schwarze 1459: iname = cp;
1.86 schwarze 1460: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 1461:
1462: /* Resolve the end marker if it is indirect. */
1463:
1.110 schwarze 1464: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1465: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1466: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1467: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1468: "%.*s", (int)namesz, iname);
1469: namesz = 0;
1470: } else
1471: namesz = strlen(name);
1472: } else
1473: name = iname;
1474:
1.86 schwarze 1475: if (namesz)
1476: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 1477:
1.110 schwarze 1478: if (*cp != '\0')
1.90 schwarze 1479: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1480: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.1 schwarze 1481:
1482: return(ROFF_IGN);
1483: }
1484:
1485: static enum rofferr
1.2 schwarze 1486: roff_block_sub(ROFF_ARGS)
1.1 schwarze 1487: {
1.2 schwarze 1488: enum rofft t;
1489: int i, j;
1490:
1491: /*
1492: * First check whether a custom macro exists at this level. If
1493: * it does, then check against it. This is some of groff's
1494: * stranger behaviours. If we encountered a custom end-scope
1495: * tag and that tag also happens to be a "real" macro, then we
1496: * need to try interpreting it again as a real macro. If it's
1497: * not, then return ignore. Else continue.
1498: */
1499:
1500: if (r->last->end) {
1.35 schwarze 1501: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 1502: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 1503: break;
1.1 schwarze 1504:
1.110 schwarze 1505: if (r->last->end[j] == '\0' &&
1506: (buf->buf[i] == '\0' ||
1507: buf->buf[i] == ' ' ||
1508: buf->buf[i] == '\t')) {
1.2 schwarze 1509: roffnode_pop(r);
1510: roffnode_cleanscope(r);
1.1 schwarze 1511:
1.110 schwarze 1512: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 1513: i++;
1514:
1515: pos = i;
1.110 schwarze 1516: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1517: ROFF_MAX)
1.2 schwarze 1518: return(ROFF_RERUN);
1519: return(ROFF_IGN);
1520: }
1.1 schwarze 1521: }
1522:
1.2 schwarze 1523: /*
1524: * If we have no custom end-query or lookup failed, then try
1525: * pulling it out of the hashtable.
1526: */
1.1 schwarze 1527:
1.110 schwarze 1528: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 1529:
1.110 schwarze 1530: if (t != ROFF_cblock) {
1531: if (tok != ROFF_ig)
1532: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.1 schwarze 1533: return(ROFF_IGN);
1.16 schwarze 1534: }
1.1 schwarze 1535:
1.2 schwarze 1536: assert(roffs[t].proc);
1.110 schwarze 1537: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1538: }
1539:
1540: static enum rofferr
1541: roff_block_text(ROFF_ARGS)
1542: {
1543:
1.110 schwarze 1544: if (tok != ROFF_ig)
1545: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 1546:
1.2 schwarze 1547: return(ROFF_IGN);
1548: }
1549:
1550: static enum rofferr
1551: roff_cond_sub(ROFF_ARGS)
1552: {
1553: enum rofft t;
1.37 schwarze 1554: char *ep;
1.71 schwarze 1555: int rr;
1.2 schwarze 1556:
1557: rr = r->last->rule;
1.37 schwarze 1558: roffnode_cleanscope(r);
1.110 schwarze 1559: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2 schwarze 1560:
1.37 schwarze 1561: /*
1.50 schwarze 1562: * Fully handle known macros when they are structurally
1563: * required or when the conditional evaluated to true.
1.5 schwarze 1564: */
1565:
1.110 schwarze 1566: if ((t != ROFF_MAX) &&
1567: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.50 schwarze 1568: assert(roffs[t].proc);
1.110 schwarze 1569: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.50 schwarze 1570: }
1.39 schwarze 1571:
1.69 schwarze 1572: /*
1573: * If `\}' occurs on a macro line without a preceding macro,
1574: * drop the line completely.
1575: */
1576:
1.110 schwarze 1577: ep = buf->buf + pos;
1578: if (ep[0] == '\\' && ep[1] == '}')
1.71 schwarze 1579: rr = 0;
1.69 schwarze 1580:
1.50 schwarze 1581: /* Always check for the closing delimiter `\}'. */
1.39 schwarze 1582:
1.110 schwarze 1583: while ((ep = strchr(ep, '\\')) != NULL) {
1584: if (*(++ep) == '}') {
1.70 schwarze 1585: *ep = '&';
1.110 schwarze 1586: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1587: }
1.119 schwarze 1588: if (*ep != '\0')
1589: ++ep;
1.50 schwarze 1590: }
1.71 schwarze 1591: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1592: }
1593:
1594: static enum rofferr
1595: roff_cond_text(ROFF_ARGS)
1596: {
1.37 schwarze 1597: char *ep;
1.71 schwarze 1598: int rr;
1.2 schwarze 1599:
1600: rr = r->last->rule;
1.37 schwarze 1601: roffnode_cleanscope(r);
1.1 schwarze 1602:
1.110 schwarze 1603: ep = buf->buf + pos;
1604: while ((ep = strchr(ep, '\\')) != NULL) {
1605: if (*(++ep) == '}') {
1.70 schwarze 1606: *ep = '&';
1.110 schwarze 1607: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1608: }
1.119 schwarze 1609: if (*ep != '\0')
1610: ++ep;
1.2 schwarze 1611: }
1.71 schwarze 1612: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1613: }
1614:
1.77 schwarze 1615: /*
1616: * Parse a single signed integer number. Stop at the first non-digit.
1617: * If there is at least one digit, return success and advance the
1618: * parse point, else return failure and let the parse point unchanged.
1619: * Ignore overflows, treat them just like the C language.
1620: */
1.56 schwarze 1621: static int
1622: roff_getnum(const char *v, int *pos, int *res)
1623: {
1.79 schwarze 1624: int myres, n, p;
1625:
1626: if (NULL == res)
1627: res = &myres;
1.56 schwarze 1628:
1629: p = *pos;
1630: n = v[p] == '-';
1631: if (n)
1632: p++;
1633:
1634: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 1635: *res = 10 * *res + v[p] - '0';
1.56 schwarze 1636: if (p == *pos + n)
1637: return 0;
1638:
1639: if (n)
1640: *res = -*res;
1641:
1642: *pos = p;
1643: return 1;
1644: }
1645:
1.71 schwarze 1646: /*
1647: * Evaluate a string comparison condition.
1648: * The first character is the delimiter.
1649: * Succeed if the string up to its second occurrence
1650: * matches the string up to its third occurence.
1651: * Advance the cursor after the third occurrence
1652: * or lacking that, to the end of the line.
1653: */
1654: static int
1655: roff_evalstrcond(const char *v, int *pos)
1656: {
1657: const char *s1, *s2, *s3;
1658: int match;
1659:
1660: match = 0;
1661: s1 = v + *pos; /* initial delimiter */
1662: s2 = s1 + 1; /* for scanning the first string */
1663: s3 = strchr(s2, *s1); /* for scanning the second string */
1664:
1665: if (NULL == s3) /* found no middle delimiter */
1666: goto out;
1667:
1668: while ('\0' != *++s3) {
1669: if (*s2 != *s3) { /* mismatch */
1670: s3 = strchr(s3, *s1);
1671: break;
1672: }
1673: if (*s3 == *s1) { /* found the final delimiter */
1674: match = 1;
1675: break;
1676: }
1677: s2++;
1678: }
1679:
1680: out:
1681: if (NULL == s3)
1682: s3 = strchr(s2, '\0');
1.114 schwarze 1683: else if (*s3 != '\0')
1.71 schwarze 1684: s3++;
1685: *pos = s3 - v;
1686: return(match);
1687: }
1688:
1.77 schwarze 1689: /*
1690: * Evaluate an optionally negated single character, numerical,
1691: * or string condition.
1692: */
1.71 schwarze 1693: static int
1.106 schwarze 1694: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.5 schwarze 1695: {
1.113 schwarze 1696: int number, savepos, wanttrue;
1.5 schwarze 1697:
1.71 schwarze 1698: if ('!' == v[*pos]) {
1699: wanttrue = 0;
1700: (*pos)++;
1701: } else
1702: wanttrue = 1;
1703:
1.5 schwarze 1704: switch (v[*pos]) {
1.112 schwarze 1705: case '\0':
1706: return(0);
1.80 schwarze 1707: case 'n':
1.71 schwarze 1708: /* FALLTHROUGH */
1.80 schwarze 1709: case 'o':
1.5 schwarze 1710: (*pos)++;
1.71 schwarze 1711: return(wanttrue);
1.80 schwarze 1712: case 'c':
1.71 schwarze 1713: /* FALLTHROUGH */
1.80 schwarze 1714: case 'd':
1.71 schwarze 1715: /* FALLTHROUGH */
1.80 schwarze 1716: case 'e':
1.5 schwarze 1717: /* FALLTHROUGH */
1.80 schwarze 1718: case 'r':
1.5 schwarze 1719: /* FALLTHROUGH */
1.80 schwarze 1720: case 't':
1.111 schwarze 1721: /* FALLTHROUGH */
1722: case 'v':
1.5 schwarze 1723: (*pos)++;
1.71 schwarze 1724: return(!wanttrue);
1.5 schwarze 1725: default:
1726: break;
1727: }
1728:
1.113 schwarze 1729: savepos = *pos;
1.106 schwarze 1730: if (roff_evalnum(r, ln, v, pos, &number, 0))
1.77 schwarze 1731: return((number > 0) == wanttrue);
1.113 schwarze 1732: else if (*pos == savepos)
1733: return(roff_evalstrcond(v, pos) == wanttrue);
1.77 schwarze 1734: else
1.113 schwarze 1735: return (0);
1.5 schwarze 1736: }
1737:
1.2 schwarze 1738: static enum rofferr
1.21 schwarze 1739: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 1740: {
1.30 schwarze 1741:
1.21 schwarze 1742: return(ROFF_IGN);
1743: }
1744:
1745: static enum rofferr
1.123 schwarze 1746: roff_insec(ROFF_ARGS)
1747: {
1748:
1749: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1750: ln, ppos, roffs[tok].name);
1751: return(ROFF_IGN);
1752: }
1753:
1754: static enum rofferr
1755: roff_unsupp(ROFF_ARGS)
1756: {
1757:
1758: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1759: ln, ppos, roffs[tok].name);
1760: return(ROFF_IGN);
1761: }
1762:
1763: static enum rofferr
1.2 schwarze 1764: roff_cond(ROFF_ARGS)
1765: {
1.46 schwarze 1766:
1767: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 1768:
1.80 schwarze 1769: /*
1.35 schwarze 1770: * An `.el' has no conditional body: it will consume the value
1771: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 1772: * defaults to DENY.
1.35 schwarze 1773: *
1774: * If we're not an `el', however, then evaluate the conditional.
1775: */
1.1 schwarze 1776:
1.110 schwarze 1777: r->last->rule = tok == ROFF_el ?
1.80 schwarze 1778: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 1779: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 1780:
1.35 schwarze 1781: /*
1782: * An if-else will put the NEGATION of the current evaluated
1783: * conditional into the stack of rules.
1784: */
1785:
1.110 schwarze 1786: if (tok == ROFF_ie) {
1.96 schwarze 1787: if (r->rstackpos + 1 == r->rstacksz) {
1788: r->rstacksz += 16;
1789: r->rstack = mandoc_reallocarray(r->rstack,
1790: r->rstacksz, sizeof(int));
1.35 schwarze 1791: }
1.71 schwarze 1792: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 1793: }
1.5 schwarze 1794:
1795: /* If the parent has false as its rule, then so do we. */
1796:
1.71 schwarze 1797: if (r->last->parent && !r->last->parent->rule)
1798: r->last->rule = 0;
1.5 schwarze 1799:
1800: /*
1.46 schwarze 1801: * Determine scope.
1802: * If there is nothing on the line after the conditional,
1803: * not even whitespace, use next-line scope.
1.5 schwarze 1804: */
1.2 schwarze 1805:
1.110 schwarze 1806: if (buf->buf[pos] == '\0') {
1.46 schwarze 1807: r->last->endspan = 2;
1808: goto out;
1809: }
1810:
1.110 schwarze 1811: while (buf->buf[pos] == ' ')
1.46 schwarze 1812: pos++;
1813:
1814: /* An opening brace requests multiline scope. */
1.2 schwarze 1815:
1.110 schwarze 1816: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 1817: r->last->endspan = -1;
1818: pos += 2;
1.46 schwarze 1819: goto out;
1.80 schwarze 1820: }
1.2 schwarze 1821:
1822: /*
1.46 schwarze 1823: * Anything else following the conditional causes
1824: * single-line scope. Warn if the scope contains
1825: * nothing but trailing whitespace.
1.2 schwarze 1826: */
1827:
1.110 schwarze 1828: if (buf->buf[pos] == '\0')
1.89 schwarze 1829: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1830: ln, ppos, roffs[tok].name);
1.2 schwarze 1831:
1.46 schwarze 1832: r->last->endspan = 1;
1.1 schwarze 1833:
1.46 schwarze 1834: out:
1.2 schwarze 1835: *offs = pos;
1836: return(ROFF_RERUN);
1.1 schwarze 1837: }
1838:
1.2 schwarze 1839: static enum rofferr
1.7 schwarze 1840: roff_ds(ROFF_ARGS)
1841: {
1.85 schwarze 1842: char *string;
1843: const char *name;
1844: size_t namesz;
1.10 schwarze 1845:
1.123 schwarze 1846: /* Ignore groff compatibility mode for now. */
1847:
1848: if (tok == ROFF_ds1)
1849: tok = ROFF_ds;
1850: else if (tok == ROFF_as1)
1851: tok = ROFF_as;
1852:
1.10 schwarze 1853: /*
1.85 schwarze 1854: * The first word is the name of the string.
1855: * If it is empty or terminated by an escape sequence,
1856: * abort the `ds' request without defining anything.
1.10 schwarze 1857: */
1.7 schwarze 1858:
1.110 schwarze 1859: name = string = buf->buf + pos;
1860: if (*name == '\0')
1.7 schwarze 1861: return(ROFF_IGN);
1862:
1.85 schwarze 1863: namesz = roff_getname(r, &string, ln, pos);
1.110 schwarze 1864: if (name[namesz] == '\\')
1.85 schwarze 1865: return(ROFF_IGN);
1866:
1867: /* Read past the initial double-quote, if any. */
1.110 schwarze 1868: if (*string == '"')
1.7 schwarze 1869: string++;
1870:
1.10 schwarze 1871: /* The rest is the value. */
1.85 schwarze 1872: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
1873: ROFF_as == tok);
1.7 schwarze 1874: return(ROFF_IGN);
1875: }
1876:
1.77 schwarze 1877: /*
1878: * Parse a single operator, one or two characters long.
1879: * If the operator is recognized, return success and advance the
1880: * parse point, else return failure and let the parse point unchanged.
1881: */
1882: static int
1883: roff_getop(const char *v, int *pos, char *res)
1884: {
1885:
1886: *res = v[*pos];
1887:
1888: switch (*res) {
1.80 schwarze 1889: case '+':
1.77 schwarze 1890: /* FALLTHROUGH */
1.80 schwarze 1891: case '-':
1.77 schwarze 1892: /* FALLTHROUGH */
1.80 schwarze 1893: case '*':
1.77 schwarze 1894: /* FALLTHROUGH */
1.80 schwarze 1895: case '/':
1.77 schwarze 1896: /* FALLTHROUGH */
1.80 schwarze 1897: case '%':
1.77 schwarze 1898: /* FALLTHROUGH */
1.80 schwarze 1899: case '&':
1.77 schwarze 1900: /* FALLTHROUGH */
1.80 schwarze 1901: case ':':
1.77 schwarze 1902: break;
1903: case '<':
1904: switch (v[*pos + 1]) {
1.80 schwarze 1905: case '=':
1.77 schwarze 1906: *res = 'l';
1907: (*pos)++;
1908: break;
1.80 schwarze 1909: case '>':
1.77 schwarze 1910: *res = '!';
1911: (*pos)++;
1912: break;
1.80 schwarze 1913: case '?':
1.77 schwarze 1914: *res = 'i';
1915: (*pos)++;
1916: break;
1917: default:
1918: break;
1919: }
1920: break;
1921: case '>':
1922: switch (v[*pos + 1]) {
1.80 schwarze 1923: case '=':
1.77 schwarze 1924: *res = 'g';
1925: (*pos)++;
1926: break;
1.80 schwarze 1927: case '?':
1.77 schwarze 1928: *res = 'a';
1929: (*pos)++;
1930: break;
1931: default:
1932: break;
1933: }
1934: break;
1935: case '=':
1936: if ('=' == v[*pos + 1])
1937: (*pos)++;
1938: break;
1939: default:
1940: return(0);
1941: }
1942: (*pos)++;
1943:
1944: return(*res);
1945: }
1946:
1947: /*
1948: * Evaluate either a parenthesized numeric expression
1949: * or a single signed integer number.
1950: */
1951: static int
1.106 schwarze 1952: roff_evalpar(struct roff *r, int ln,
1953: const char *v, int *pos, int *res)
1.77 schwarze 1954: {
1955:
1956: if ('(' != v[*pos])
1957: return(roff_getnum(v, pos, res));
1958:
1959: (*pos)++;
1.106 schwarze 1960: if ( ! roff_evalnum(r, ln, v, pos, res, 1))
1.77 schwarze 1961: return(0);
1962:
1.79 schwarze 1963: /*
1964: * Omission of the closing parenthesis
1965: * is an error in validation mode,
1966: * but ignored in evaluation mode.
1967: */
1968:
1.77 schwarze 1969: if (')' == v[*pos])
1970: (*pos)++;
1.79 schwarze 1971: else if (NULL == res)
1972: return(0);
1.77 schwarze 1973:
1974: return(1);
1975: }
1976:
1977: /*
1978: * Evaluate a complete numeric expression.
1979: * Proceed left to right, there is no concept of precedence.
1980: */
1981: static int
1.106 schwarze 1982: roff_evalnum(struct roff *r, int ln, const char *v,
1983: int *pos, int *res, int skipwhite)
1.77 schwarze 1984: {
1985: int mypos, operand2;
1986: char operator;
1987:
1988: if (NULL == pos) {
1989: mypos = 0;
1990: pos = &mypos;
1991: }
1992:
1993: if (skipwhite)
1994: while (isspace((unsigned char)v[*pos]))
1995: (*pos)++;
1996:
1.106 schwarze 1997: if ( ! roff_evalpar(r, ln, v, pos, res))
1.77 schwarze 1998: return(0);
1999:
2000: while (1) {
2001: if (skipwhite)
2002: while (isspace((unsigned char)v[*pos]))
2003: (*pos)++;
2004:
2005: if ( ! roff_getop(v, pos, &operator))
2006: break;
2007:
2008: if (skipwhite)
2009: while (isspace((unsigned char)v[*pos]))
2010: (*pos)++;
2011:
1.106 schwarze 2012: if ( ! roff_evalpar(r, ln, v, pos, &operand2))
1.77 schwarze 2013: return(0);
2014:
2015: if (skipwhite)
2016: while (isspace((unsigned char)v[*pos]))
2017: (*pos)++;
1.79 schwarze 2018:
2019: if (NULL == res)
2020: continue;
1.77 schwarze 2021:
2022: switch (operator) {
1.80 schwarze 2023: case '+':
1.77 schwarze 2024: *res += operand2;
2025: break;
1.80 schwarze 2026: case '-':
1.77 schwarze 2027: *res -= operand2;
2028: break;
1.80 schwarze 2029: case '*':
1.77 schwarze 2030: *res *= operand2;
2031: break;
1.80 schwarze 2032: case '/':
1.116 schwarze 2033: if (operand2 == 0) {
1.106 schwarze 2034: mandoc_msg(MANDOCERR_DIVZERO,
2035: r->parse, ln, *pos, v);
2036: *res = 0;
2037: break;
2038: }
1.77 schwarze 2039: *res /= operand2;
2040: break;
1.80 schwarze 2041: case '%':
1.116 schwarze 2042: if (operand2 == 0) {
2043: mandoc_msg(MANDOCERR_DIVZERO,
2044: r->parse, ln, *pos, v);
2045: *res = 0;
2046: break;
2047: }
1.77 schwarze 2048: *res %= operand2;
2049: break;
1.80 schwarze 2050: case '<':
1.77 schwarze 2051: *res = *res < operand2;
2052: break;
1.80 schwarze 2053: case '>':
1.77 schwarze 2054: *res = *res > operand2;
2055: break;
1.80 schwarze 2056: case 'l':
1.77 schwarze 2057: *res = *res <= operand2;
2058: break;
1.80 schwarze 2059: case 'g':
1.77 schwarze 2060: *res = *res >= operand2;
2061: break;
1.80 schwarze 2062: case '=':
1.77 schwarze 2063: *res = *res == operand2;
2064: break;
1.80 schwarze 2065: case '!':
1.77 schwarze 2066: *res = *res != operand2;
2067: break;
1.80 schwarze 2068: case '&':
1.77 schwarze 2069: *res = *res && operand2;
2070: break;
1.80 schwarze 2071: case ':':
1.77 schwarze 2072: *res = *res || operand2;
2073: break;
1.80 schwarze 2074: case 'i':
1.77 schwarze 2075: if (operand2 < *res)
2076: *res = operand2;
2077: break;
1.80 schwarze 2078: case 'a':
1.77 schwarze 2079: if (operand2 > *res)
2080: *res = operand2;
2081: break;
2082: default:
2083: abort();
2084: }
2085: }
2086: return(1);
2087: }
2088:
1.52 schwarze 2089: void
1.60 schwarze 2090: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 2091: {
1.52 schwarze 2092: struct roffreg *reg;
2093:
2094: /* Search for an existing register with the same name. */
2095: reg = r->regtab;
2096:
2097: while (reg && strcmp(name, reg->key.p))
2098: reg = reg->next;
1.41 schwarze 2099:
1.52 schwarze 2100: if (NULL == reg) {
2101: /* Create a new register. */
2102: reg = mandoc_malloc(sizeof(struct roffreg));
2103: reg->key.p = mandoc_strdup(name);
2104: reg->key.sz = strlen(name);
1.60 schwarze 2105: reg->val = 0;
1.52 schwarze 2106: reg->next = r->regtab;
2107: r->regtab = reg;
2108: }
2109:
1.60 schwarze 2110: if ('+' == sign)
2111: reg->val += val;
2112: else if ('-' == sign)
2113: reg->val -= val;
2114: else
2115: reg->val = val;
1.41 schwarze 2116: }
2117:
1.65 schwarze 2118: /*
2119: * Handle some predefined read-only number registers.
2120: * For now, return -1 if the requested register is not predefined;
2121: * in case a predefined read-only register having the value -1
2122: * were to turn up, another special value would have to be chosen.
2123: */
2124: static int
2125: roff_getregro(const char *name)
2126: {
2127:
2128: switch (*name) {
1.80 schwarze 2129: case 'A': /* ASCII approximation mode is always off. */
1.65 schwarze 2130: return(0);
1.80 schwarze 2131: case 'g': /* Groff compatibility mode is always on. */
1.65 schwarze 2132: return(1);
1.80 schwarze 2133: case 'H': /* Fixed horizontal resolution. */
1.65 schwarze 2134: return (24);
1.80 schwarze 2135: case 'j': /* Always adjust left margin only. */
1.65 schwarze 2136: return(0);
1.80 schwarze 2137: case 'T': /* Some output device is always defined. */
1.65 schwarze 2138: return(1);
1.80 schwarze 2139: case 'V': /* Fixed vertical resolution. */
1.65 schwarze 2140: return (40);
2141: default:
2142: return (-1);
2143: }
2144: }
2145:
1.53 schwarze 2146: int
1.52 schwarze 2147: roff_getreg(const struct roff *r, const char *name)
1.41 schwarze 2148: {
1.52 schwarze 2149: struct roffreg *reg;
1.65 schwarze 2150: int val;
2151:
2152: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
2153: val = roff_getregro(name + 1);
2154: if (-1 != val)
2155: return (val);
2156: }
1.52 schwarze 2157:
2158: for (reg = r->regtab; reg; reg = reg->next)
2159: if (0 == strcmp(name, reg->key.p))
1.53 schwarze 2160: return(reg->val);
2161:
2162: return(0);
2163: }
2164:
2165: static int
2166: roff_getregn(const struct roff *r, const char *name, size_t len)
2167: {
2168: struct roffreg *reg;
1.65 schwarze 2169: int val;
2170:
2171: if ('.' == name[0] && 2 == len) {
2172: val = roff_getregro(name + 1);
2173: if (-1 != val)
2174: return (val);
2175: }
1.53 schwarze 2176:
2177: for (reg = r->regtab; reg; reg = reg->next)
2178: if (len == reg->key.sz &&
2179: 0 == strncmp(name, reg->key.p, len))
2180: return(reg->val);
1.41 schwarze 2181:
1.52 schwarze 2182: return(0);
1.41 schwarze 2183: }
2184:
1.52 schwarze 2185: static void
2186: roff_freereg(struct roffreg *reg)
1.41 schwarze 2187: {
1.52 schwarze 2188: struct roffreg *old_reg;
1.41 schwarze 2189:
1.52 schwarze 2190: while (NULL != reg) {
2191: free(reg->key.p);
2192: old_reg = reg;
2193: reg = reg->next;
2194: free(old_reg);
2195: }
1.41 schwarze 2196: }
1.7 schwarze 2197:
2198: static enum rofferr
1.6 schwarze 2199: roff_nr(ROFF_ARGS)
1.1 schwarze 2200: {
1.85 schwarze 2201: char *key, *val;
2202: size_t keysz;
1.37 schwarze 2203: int iv;
1.60 schwarze 2204: char sign;
1.6 schwarze 2205:
1.110 schwarze 2206: key = val = buf->buf + pos;
2207: if (*key == '\0')
1.85 schwarze 2208: return(ROFF_IGN);
2209:
2210: keysz = roff_getname(r, &val, ln, pos);
1.110 schwarze 2211: if (key[keysz] == '\\')
1.85 schwarze 2212: return(ROFF_IGN);
2213: key[keysz] = '\0';
1.6 schwarze 2214:
1.60 schwarze 2215: sign = *val;
1.110 schwarze 2216: if (sign == '+' || sign == '-')
1.60 schwarze 2217: val++;
2218:
1.106 schwarze 2219: if (roff_evalnum(r, ln, val, NULL, &iv, 0))
1.77 schwarze 2220: roff_setreg(r, key, iv, sign);
1.1 schwarze 2221:
1.76 schwarze 2222: return(ROFF_IGN);
2223: }
2224:
2225: static enum rofferr
2226: roff_rr(ROFF_ARGS)
2227: {
2228: struct roffreg *reg, **prev;
1.85 schwarze 2229: char *name, *cp;
2230: size_t namesz;
1.76 schwarze 2231:
1.110 schwarze 2232: name = cp = buf->buf + pos;
2233: if (*name == '\0')
1.85 schwarze 2234: return(ROFF_IGN);
2235: namesz = roff_getname(r, &cp, ln, pos);
2236: name[namesz] = '\0';
1.76 schwarze 2237:
2238: prev = &r->regtab;
2239: while (1) {
2240: reg = *prev;
1.110 schwarze 2241: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 2242: break;
2243: prev = ®->next;
2244: }
1.110 schwarze 2245: if (reg != NULL) {
1.76 schwarze 2246: *prev = reg->next;
2247: free(reg->key.p);
2248: free(reg);
2249: }
1.29 schwarze 2250: return(ROFF_IGN);
2251: }
2252:
2253: static enum rofferr
2254: roff_rm(ROFF_ARGS)
2255: {
2256: const char *name;
2257: char *cp;
1.85 schwarze 2258: size_t namesz;
1.29 schwarze 2259:
1.110 schwarze 2260: cp = buf->buf + pos;
2261: while (*cp != '\0') {
1.85 schwarze 2262: name = cp;
1.110 schwarze 2263: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 2264: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110 schwarze 2265: if (name[namesz] == '\\')
1.85 schwarze 2266: break;
1.29 schwarze 2267: }
1.51 schwarze 2268: return(ROFF_IGN);
2269: }
2270:
2271: static enum rofferr
2272: roff_it(ROFF_ARGS)
2273: {
2274: char *cp;
2275: size_t len;
2276: int iv;
2277:
2278: /* Parse the number of lines. */
1.110 schwarze 2279: cp = buf->buf + pos;
1.51 schwarze 2280: len = strcspn(cp, " \t");
2281: cp[len] = '\0';
2282: if ((iv = mandoc_strntoi(cp, len, 10)) <= 0) {
1.95 schwarze 2283: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110 schwarze 2284: ln, ppos, buf->buf + 1);
1.51 schwarze 2285: return(ROFF_IGN);
2286: }
2287: cp += len + 1;
2288:
2289: /* Arm the input line trap. */
2290: roffit_lines = iv;
2291: roffit_macro = mandoc_strdup(cp);
1.2 schwarze 2292: return(ROFF_IGN);
1.47 schwarze 2293: }
2294:
2295: static enum rofferr
2296: roff_Dd(ROFF_ARGS)
2297: {
2298: const char *const *cp;
2299:
1.99 schwarze 2300: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47 schwarze 2301: for (cp = __mdoc_reserved; *cp; cp++)
2302: roff_setstr(r, *cp, NULL, 0);
2303:
1.99 schwarze 2304: if (r->format == 0)
2305: r->format = MPARSE_MDOC;
2306:
1.47 schwarze 2307: return(ROFF_CONT);
2308: }
2309:
2310: static enum rofferr
2311: roff_TH(ROFF_ARGS)
2312: {
2313: const char *const *cp;
2314:
1.99 schwarze 2315: if ((r->options & MPARSE_QUICK) == 0)
1.47 schwarze 2316: for (cp = __man_reserved; *cp; cp++)
2317: roff_setstr(r, *cp, NULL, 0);
2318:
1.99 schwarze 2319: if (r->format == 0)
2320: r->format = MPARSE_MAN;
2321:
1.47 schwarze 2322: return(ROFF_CONT);
1.14 schwarze 2323: }
2324:
2325: static enum rofferr
1.27 schwarze 2326: roff_TE(ROFF_ARGS)
2327: {
2328:
2329: if (NULL == r->tbl)
1.94 schwarze 2330: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2331: ln, ppos, "TE");
1.27 schwarze 2332: else
1.41 schwarze 2333: tbl_end(&r->tbl);
1.27 schwarze 2334:
2335: return(ROFF_IGN);
2336: }
2337:
2338: static enum rofferr
2339: roff_T_(ROFF_ARGS)
2340: {
2341:
2342: if (NULL == r->tbl)
1.94 schwarze 2343: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2344: ln, ppos, "T&");
1.27 schwarze 2345: else
2346: tbl_restart(ppos, ln, r->tbl);
2347:
2348: return(ROFF_IGN);
2349: }
2350:
1.102 schwarze 2351: /*
2352: * Handle in-line equation delimiters.
2353: */
2354: static enum rofferr
1.110 schwarze 2355: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 2356: {
1.105 schwarze 2357: char *cp1, *cp2;
2358: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 2359:
1.102 schwarze 2360: /*
2361: * Outside equations, look for an opening delimiter.
2362: * If we are inside an equation, we already know it is
2363: * in-line, or this function wouldn't have been called;
2364: * so look for a closing delimiter.
2365: */
2366:
1.110 schwarze 2367: cp1 = buf->buf + pos;
1.102 schwarze 2368: cp2 = strchr(cp1, r->eqn == NULL ?
2369: r->last_eqn->odelim : r->last_eqn->cdelim);
2370: if (cp2 == NULL)
2371: return(ROFF_CONT);
2372:
1.105 schwarze 2373: *cp2++ = '\0';
2374: bef_pr = bef_nl = aft_nl = aft_pr = "";
2375:
2376: /* Handle preceding text, protecting whitespace. */
2377:
1.110 schwarze 2378: if (*buf->buf != '\0') {
1.105 schwarze 2379: if (r->eqn == NULL)
2380: bef_pr = "\\&";
2381: bef_nl = "\n";
2382: }
2383:
2384: /*
2385: * Prepare replacing the delimiter with an equation macro
2386: * and drop leading white space from the equation.
2387: */
1.102 schwarze 2388:
1.105 schwarze 2389: if (r->eqn == NULL) {
2390: while (*cp2 == ' ')
2391: cp2++;
2392: mac = ".EQ";
2393: } else
2394: mac = ".EN";
2395:
2396: /* Handle following text, protecting whitespace. */
2397:
2398: if (*cp2 != '\0') {
2399: aft_nl = "\n";
2400: if (r->eqn != NULL)
2401: aft_pr = "\\&";
2402: }
2403:
2404: /* Do the actual replacement. */
2405:
1.110 schwarze 2406: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 2407: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 2408: free(buf->buf);
2409: buf->buf = cp1;
1.102 schwarze 2410:
2411: /* Toggle the in-line state of the eqn subsystem. */
2412:
2413: r->eqn_inline = r->eqn == NULL;
2414: return(ROFF_REPARSE);
1.41 schwarze 2415: }
2416:
1.107 schwarze 2417: static enum rofferr
2418: roff_EQ(ROFF_ARGS)
1.32 schwarze 2419: {
1.41 schwarze 2420: struct eqn_node *e;
1.32 schwarze 2421:
1.110 schwarze 2422: assert(r->eqn == NULL);
1.107 schwarze 2423: e = eqn_alloc(ppos, ln, r->parse);
1.32 schwarze 2424:
1.102 schwarze 2425: if (r->last_eqn) {
1.32 schwarze 2426: r->last_eqn->next = e;
1.102 schwarze 2427: e->delim = r->last_eqn->delim;
2428: e->odelim = r->last_eqn->odelim;
2429: e->cdelim = r->last_eqn->cdelim;
2430: } else
1.32 schwarze 2431: r->first_eqn = r->last_eqn = e;
2432:
2433: r->eqn = r->last_eqn = e;
1.41 schwarze 2434:
1.110 schwarze 2435: if (buf->buf[pos] != '\0')
1.107 schwarze 2436: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 2437: ".EQ %s", buf->buf + pos);
1.41 schwarze 2438:
1.32 schwarze 2439: return(ROFF_IGN);
2440: }
2441:
2442: static enum rofferr
2443: roff_EN(ROFF_ARGS)
2444: {
2445:
1.94 schwarze 2446: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.32 schwarze 2447: return(ROFF_IGN);
2448: }
2449:
2450: static enum rofferr
1.27 schwarze 2451: roff_TS(ROFF_ARGS)
2452: {
1.49 schwarze 2453: struct tbl_node *tbl;
1.27 schwarze 2454:
2455: if (r->tbl) {
1.94 schwarze 2456: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2457: ln, ppos, "TS breaks TS");
1.41 schwarze 2458: tbl_end(&r->tbl);
1.27 schwarze 2459: }
2460:
1.49 schwarze 2461: tbl = tbl_alloc(ppos, ln, r->parse);
1.27 schwarze 2462:
2463: if (r->last_tbl)
1.49 schwarze 2464: r->last_tbl->next = tbl;
1.27 schwarze 2465: else
1.49 schwarze 2466: r->first_tbl = r->last_tbl = tbl;
1.27 schwarze 2467:
1.49 schwarze 2468: r->tbl = r->last_tbl = tbl;
1.27 schwarze 2469: return(ROFF_IGN);
1.123 schwarze 2470: }
2471:
2472: static enum rofferr
2473: roff_brp(ROFF_ARGS)
2474: {
2475:
2476: buf->buf[pos - 1] = '\0';
2477: return(ROFF_CONT);
1.27 schwarze 2478: }
2479:
2480: static enum rofferr
1.48 schwarze 2481: roff_cc(ROFF_ARGS)
2482: {
2483: const char *p;
2484:
1.110 schwarze 2485: p = buf->buf + pos;
1.48 schwarze 2486:
1.110 schwarze 2487: if (*p == '\0' || (r->control = *p++) == '.')
1.48 schwarze 2488: r->control = 0;
2489:
1.110 schwarze 2490: if (*p != '\0')
1.48 schwarze 2491: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
2492:
2493: return(ROFF_IGN);
2494: }
2495:
2496: static enum rofferr
1.42 schwarze 2497: roff_tr(ROFF_ARGS)
2498: {
2499: const char *p, *first, *second;
2500: size_t fsz, ssz;
2501: enum mandoc_esc esc;
2502:
1.110 schwarze 2503: p = buf->buf + pos;
1.42 schwarze 2504:
1.110 schwarze 2505: if (*p == '\0') {
1.42 schwarze 2506: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
2507: return(ROFF_IGN);
2508: }
2509:
1.110 schwarze 2510: while (*p != '\0') {
1.42 schwarze 2511: fsz = ssz = 1;
2512:
2513: first = p++;
1.110 schwarze 2514: if (*first == '\\') {
1.42 schwarze 2515: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2516: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2517: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2518: ln, (int)(p - buf->buf), first);
1.42 schwarze 2519: return(ROFF_IGN);
2520: }
2521: fsz = (size_t)(p - first);
2522: }
2523:
2524: second = p++;
1.110 schwarze 2525: if (*second == '\\') {
1.42 schwarze 2526: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2527: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2528: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2529: ln, (int)(p - buf->buf), second);
1.42 schwarze 2530: return(ROFF_IGN);
2531: }
2532: ssz = (size_t)(p - second);
1.110 schwarze 2533: } else if (*second == '\0') {
1.80 schwarze 2534: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse,
1.110 schwarze 2535: ln, (int)(p - buf->buf), NULL);
1.42 schwarze 2536: second = " ";
2537: p--;
2538: }
2539:
2540: if (fsz > 1) {
1.80 schwarze 2541: roff_setstrn(&r->xmbtab, first, fsz,
2542: second, ssz, 0);
1.42 schwarze 2543: continue;
2544: }
2545:
1.110 schwarze 2546: if (r->xtab == NULL)
1.80 schwarze 2547: r->xtab = mandoc_calloc(128,
2548: sizeof(struct roffstr));
1.42 schwarze 2549:
2550: free(r->xtab[(int)*first].p);
2551: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2552: r->xtab[(int)*first].sz = ssz;
2553: }
2554:
2555: return(ROFF_IGN);
2556: }
2557:
2558: static enum rofferr
1.14 schwarze 2559: roff_so(ROFF_ARGS)
2560: {
1.121 schwarze 2561: char *name, *cp;
1.15 schwarze 2562:
1.110 schwarze 2563: name = buf->buf + pos;
1.97 schwarze 2564: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14 schwarze 2565:
1.22 schwarze 2566: /*
2567: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2568: * opening anything that's not in our cwd or anything beneath
2569: * it. Thus, explicitly disallow traversing up the file-system
2570: * or using absolute paths.
2571: */
2572:
1.110 schwarze 2573: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83 schwarze 2574: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2575: ".so %s", name);
1.121 schwarze 2576: buf->sz = mandoc_asprintf(&cp,
2577: ".sp\nSee the file %s.\n.sp", name) + 1;
2578: free(buf->buf);
2579: buf->buf = cp;
2580: *offs = 0;
2581: return(ROFF_REPARSE);
1.14 schwarze 2582: }
2583:
2584: *offs = pos;
2585: return(ROFF_SO);
1.7 schwarze 2586: }
2587:
1.16 schwarze 2588: static enum rofferr
2589: roff_userdef(ROFF_ARGS)
1.12 schwarze 2590: {
1.16 schwarze 2591: const char *arg[9];
2592: char *cp, *n1, *n2;
1.25 schwarze 2593: int i;
1.12 schwarze 2594:
1.16 schwarze 2595: /*
2596: * Collect pointers to macro argument strings
1.61 schwarze 2597: * and NUL-terminate them.
1.16 schwarze 2598: */
1.110 schwarze 2599: cp = buf->buf + pos;
1.25 schwarze 2600: for (i = 0; i < 9; i++)
1.110 schwarze 2601: arg[i] = *cp == '\0' ? "" :
1.35 schwarze 2602: mandoc_getarg(r->parse, &cp, ln, &pos);
1.16 schwarze 2603:
2604: /*
2605: * Expand macro arguments.
1.12 schwarze 2606: */
1.110 schwarze 2607: buf->sz = 0;
1.16 schwarze 2608: n1 = cp = mandoc_strdup(r->current_string);
1.110 schwarze 2609: while ((cp = strstr(cp, "\\$")) != NULL) {
1.16 schwarze 2610: i = cp[2] - '1';
2611: if (0 > i || 8 < i) {
2612: /* Not an argument invocation. */
2613: cp += 2;
2614: continue;
2615: }
1.82 schwarze 2616: *cp = '\0';
1.110 schwarze 2617: buf->sz = mandoc_asprintf(&n2, "%s%s%s",
1.82 schwarze 2618: n1, arg[i], cp + 3) + 1;
1.16 schwarze 2619: cp = n2 + (cp - n1);
2620: free(n1);
2621: n1 = n2;
1.12 schwarze 2622: }
2623:
1.16 schwarze 2624: /*
2625: * Replace the macro invocation
2626: * by the expanded macro.
2627: */
1.110 schwarze 2628: free(buf->buf);
2629: buf->buf = n1;
2630: if (buf->sz == 0)
2631: buf->sz = strlen(buf->buf) + 1;
1.120 schwarze 2632: *offs = 0;
1.16 schwarze 2633:
1.110 schwarze 2634: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.16 schwarze 2635: ROFF_REPARSE : ROFF_APPEND);
1.12 schwarze 2636: }
1.28 schwarze 2637:
1.85 schwarze 2638: static size_t
1.28 schwarze 2639: roff_getname(struct roff *r, char **cpp, int ln, int pos)
2640: {
2641: char *name, *cp;
1.85 schwarze 2642: size_t namesz;
1.28 schwarze 2643:
2644: name = *cpp;
2645: if ('\0' == *name)
1.85 schwarze 2646: return(0);
1.28 schwarze 2647:
1.85 schwarze 2648: /* Read until end of name and terminate it with NUL. */
2649: for (cp = name; 1; cp++) {
2650: if ('\0' == *cp || ' ' == *cp) {
2651: namesz = cp - name;
2652: break;
2653: }
1.28 schwarze 2654: if ('\\' != *cp)
2655: continue;
1.88 schwarze 2656: namesz = cp - name;
2657: if ('{' == cp[1] || '}' == cp[1])
2658: break;
1.28 schwarze 2659: cp++;
2660: if ('\\' == *cp)
2661: continue;
1.97 schwarze 2662: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
2663: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 2664: mandoc_escape((const char **)&cp, NULL, NULL);
2665: break;
1.28 schwarze 2666: }
2667:
2668: /* Read past spaces. */
2669: while (' ' == *cp)
2670: cp++;
2671:
2672: *cpp = cp;
1.85 schwarze 2673: return(namesz);
1.28 schwarze 2674: }
2675:
1.16 schwarze 2676: /*
2677: * Store *string into the user-defined string called *name.
2678: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 2679: * append == 0: replace mode
2680: * append == 1: single-line append mode
2681: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 2682: */
1.8 schwarze 2683: static void
1.16 schwarze 2684: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 2685: int append)
1.7 schwarze 2686: {
1.42 schwarze 2687:
2688: roff_setstrn(&r->strtab, name, strlen(name), string,
1.80 schwarze 2689: string ? strlen(string) : 0, append);
1.42 schwarze 2690: }
2691:
2692: static void
2693: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 2694: const char *string, size_t stringsz, int append)
1.42 schwarze 2695: {
2696: struct roffkv *n;
2697: char *c;
2698: int i;
2699: size_t oldch, newch;
1.7 schwarze 2700:
1.16 schwarze 2701: /* Search for an existing string with the same name. */
1.42 schwarze 2702: n = *r;
2703:
1.84 schwarze 2704: while (n && (namesz != n->key.sz ||
2705: strncmp(n->key.p, name, namesz)))
1.7 schwarze 2706: n = n->next;
1.8 schwarze 2707:
2708: if (NULL == n) {
1.16 schwarze 2709: /* Create a new string table entry. */
1.42 schwarze 2710: n = mandoc_malloc(sizeof(struct roffkv));
2711: n->key.p = mandoc_strndup(name, namesz);
2712: n->key.sz = namesz;
2713: n->val.p = NULL;
2714: n->val.sz = 0;
2715: n->next = *r;
2716: *r = n;
1.66 schwarze 2717: } else if (0 == append) {
1.42 schwarze 2718: free(n->val.p);
2719: n->val.p = NULL;
2720: n->val.sz = 0;
1.16 schwarze 2721: }
2722:
2723: if (NULL == string)
2724: return;
2725:
2726: /*
2727: * One additional byte for the '\n' in multiline mode,
2728: * and one for the terminating '\0'.
2729: */
1.66 schwarze 2730: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 2731:
2732: if (NULL == n->val.p) {
2733: n->val.p = mandoc_malloc(newch);
2734: *n->val.p = '\0';
1.16 schwarze 2735: oldch = 0;
2736: } else {
1.42 schwarze 2737: oldch = n->val.sz;
2738: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 2739: }
2740:
2741: /* Skip existing content in the destination buffer. */
1.42 schwarze 2742: c = n->val.p + (int)oldch;
1.16 schwarze 2743:
2744: /* Append new content to the destination buffer. */
1.42 schwarze 2745: i = 0;
2746: while (i < (int)stringsz) {
1.16 schwarze 2747: /*
2748: * Rudimentary roff copy mode:
2749: * Handle escaped backslashes.
2750: */
1.42 schwarze 2751: if ('\\' == string[i] && '\\' == string[i + 1])
2752: i++;
2753: *c++ = string[i++];
1.16 schwarze 2754: }
1.8 schwarze 2755:
1.16 schwarze 2756: /* Append terminating bytes. */
1.66 schwarze 2757: if (1 < append)
1.16 schwarze 2758: *c++ = '\n';
1.42 schwarze 2759:
1.16 schwarze 2760: *c = '\0';
1.42 schwarze 2761: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 2762: }
2763:
1.8 schwarze 2764: static const char *
2765: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7 schwarze 2766: {
1.42 schwarze 2767: const struct roffkv *n;
1.64 schwarze 2768: int i;
1.7 schwarze 2769:
1.42 schwarze 2770: for (n = r->strtab; n; n = n->next)
1.80 schwarze 2771: if (0 == strncmp(name, n->key.p, len) &&
2772: '\0' == n->key.p[(int)len])
1.42 schwarze 2773: return(n->val.p);
1.64 schwarze 2774:
2775: for (i = 0; i < PREDEFS_MAX; i++)
2776: if (0 == strncmp(name, predefs[i].name, len) &&
2777: '\0' == predefs[i].name[(int)len])
2778: return(predefs[i].str);
1.8 schwarze 2779:
1.42 schwarze 2780: return(NULL);
1.7 schwarze 2781: }
2782:
1.8 schwarze 2783: static void
1.42 schwarze 2784: roff_freestr(struct roffkv *r)
1.7 schwarze 2785: {
1.42 schwarze 2786: struct roffkv *n, *nn;
1.7 schwarze 2787:
1.42 schwarze 2788: for (n = r; n; n = nn) {
2789: free(n->key.p);
2790: free(n->val.p);
1.7 schwarze 2791: nn = n->next;
2792: free(n);
2793: }
1.27 schwarze 2794: }
2795:
2796: const struct tbl_span *
2797: roff_span(const struct roff *r)
2798: {
1.80 schwarze 2799:
1.27 schwarze 2800: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32 schwarze 2801: }
2802:
2803: const struct eqn *
2804: roff_eqn(const struct roff *r)
2805: {
1.80 schwarze 2806:
1.32 schwarze 2807: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.42 schwarze 2808: }
2809:
2810: /*
2811: * Duplicate an input string, making the appropriate character
2812: * conversations (as stipulated by `tr') along the way.
2813: * Returns a heap-allocated string with all the replacements made.
2814: */
2815: char *
2816: roff_strdup(const struct roff *r, const char *p)
2817: {
2818: const struct roffkv *cp;
2819: char *res;
2820: const char *pp;
2821: size_t ssz, sz;
2822: enum mandoc_esc esc;
2823:
2824: if (NULL == r->xmbtab && NULL == r->xtab)
2825: return(mandoc_strdup(p));
2826: else if ('\0' == *p)
2827: return(mandoc_strdup(""));
2828:
2829: /*
2830: * Step through each character looking for term matches
2831: * (remember that a `tr' can be invoked with an escape, which is
2832: * a glyph but the escape is multi-character).
2833: * We only do this if the character hash has been initialised
2834: * and the string is >0 length.
2835: */
2836:
2837: res = NULL;
2838: ssz = 0;
2839:
2840: while ('\0' != *p) {
2841: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
2842: sz = r->xtab[(int)*p].sz;
2843: res = mandoc_realloc(res, ssz + sz + 1);
2844: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
2845: ssz += sz;
2846: p++;
2847: continue;
2848: } else if ('\\' != *p) {
2849: res = mandoc_realloc(res, ssz + 2);
2850: res[ssz++] = *p++;
2851: continue;
2852: }
2853:
2854: /* Search for term matches. */
2855: for (cp = r->xmbtab; cp; cp = cp->next)
2856: if (0 == strncmp(p, cp->key.p, cp->key.sz))
2857: break;
2858:
2859: if (NULL != cp) {
2860: /*
2861: * A match has been found.
2862: * Append the match to the array and move
2863: * forward by its keysize.
2864: */
1.80 schwarze 2865: res = mandoc_realloc(res,
2866: ssz + cp->val.sz + 1);
1.42 schwarze 2867: memcpy(res + ssz, cp->val.p, cp->val.sz);
2868: ssz += cp->val.sz;
2869: p += (int)cp->key.sz;
2870: continue;
2871: }
2872:
2873: /*
2874: * Handle escapes carefully: we need to copy
2875: * over just the escape itself, or else we might
2876: * do replacements within the escape itself.
2877: * Make sure to pass along the bogus string.
2878: */
2879: pp = p++;
2880: esc = mandoc_escape(&p, NULL, NULL);
2881: if (ESCAPE_ERROR == esc) {
2882: sz = strlen(pp);
2883: res = mandoc_realloc(res, ssz + sz + 1);
2884: memcpy(res + ssz, pp, sz);
2885: break;
2886: }
1.80 schwarze 2887: /*
2888: * We bail out on bad escapes.
1.42 schwarze 2889: * No need to warn: we already did so when
2890: * roff_res() was called.
2891: */
2892: sz = (int)(p - pp);
2893: res = mandoc_realloc(res, ssz + sz + 1);
2894: memcpy(res + ssz, pp, sz);
2895: ssz += sz;
2896: }
2897:
2898: res[(int)ssz] = '\0';
2899: return(res);
1.99 schwarze 2900: }
2901:
2902: int
2903: roff_getformat(const struct roff *r)
2904: {
2905:
2906: return(r->format);
1.48 schwarze 2907: }
2908:
2909: /*
1.80 schwarze 2910: * Find out whether a line is a macro line or not.
1.48 schwarze 2911: * If it is, adjust the current position and return one; if it isn't,
2912: * return zero and don't change the current position.
2913: * If the control character has been set with `.cc', then let that grain
2914: * precedence.
2915: * This is slighly contrary to groff, where using the non-breaking
2916: * control character when `cc' has been invoked will cause the
2917: * non-breaking macro contents to be printed verbatim.
2918: */
2919: int
2920: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
2921: {
2922: int pos;
2923:
2924: pos = *ppos;
2925:
2926: if (0 != r->control && cp[pos] == r->control)
2927: pos++;
2928: else if (0 != r->control)
2929: return(0);
2930: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
2931: pos += 2;
2932: else if ('.' == cp[pos] || '\'' == cp[pos])
2933: pos++;
2934: else
2935: return(0);
2936:
2937: while (' ' == cp[pos] || '\t' == cp[pos])
2938: pos++;
2939:
2940: *ppos = pos;
2941: return(1);
1.1 schwarze 2942: }
![[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