Annotation of src/usr.bin/mandoc/roff.c, Revision 1.128
1.128 ! schwarze 1: /* $OpenBSD: roff.c,v 1.127 2015/01/23 20:17:25 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,
1.128 ! schwarze 180: ROFF_na,
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 },
1.128 ! schwarze 581: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 582: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
583: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 584: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
585: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
586: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
587: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.14 schwarze 588: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.123 schwarze 589: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
590: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 591: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 592: { "nx", roff_insec, NULL, NULL, 0, NULL },
593: { "open", roff_insec, NULL, NULL, 0, NULL },
594: { "opena", roff_insec, NULL, NULL, 0, NULL },
595: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
596: { "output", roff_unsupp, NULL, NULL, 0, NULL },
597: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
598: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
599: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
600: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
601: { "pi", roff_insec, NULL, NULL, 0, NULL },
602: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.101 schwarze 603: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 604: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
605: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
606: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
607: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 608: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 609: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.127 schwarze 610: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 611: { "pso", roff_insec, NULL, NULL, 0, NULL },
612: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
615: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
616: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
617: { "return", roff_unsupp, NULL, NULL, 0, NULL },
618: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
619: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
620: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.29 schwarze 621: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.123 schwarze 622: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
623: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.76 schwarze 624: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.123 schwarze 625: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
626: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
627: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
628: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
629: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
630: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
631: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.14 schwarze 632: { "so", roff_so, NULL, NULL, 0, NULL },
1.123 schwarze 633: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
634: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
635: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
636: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
637: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
638: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
639: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
640: { "sy", roff_insec, NULL, NULL, 0, NULL },
641: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.127 schwarze 642: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 643: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
644: { "TE", roff_TE, NULL, NULL, 0, NULL },
645: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.127 schwarze 646: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 647: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
648: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
649: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
650: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
651: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.42 schwarze 652: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.123 schwarze 653: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
654: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
655: { "trf", roff_insec, NULL, NULL, 0, NULL },
656: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
657: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
658: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
659: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.27 schwarze 660: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.123 schwarze 661: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
662: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
663: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
664: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
665: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
666: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
667: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
668: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
670: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
671: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
672: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
673: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
674: { "while", roff_unsupp, NULL, NULL, 0, NULL },
675: { "write", roff_insec, NULL, NULL, 0, NULL },
676: { "writec", roff_insec, NULL, NULL, 0, NULL },
677: { "writem", roff_insec, NULL, NULL, 0, NULL },
678: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 679: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.16 schwarze 680: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1 schwarze 681: };
682:
1.73 schwarze 683: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47 schwarze 684: const char *const __mdoc_reserved[] = {
685: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
686: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
687: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
688: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73 schwarze 689: "Dt", "Dv", "Dx", "D1",
690: "Ec", "Ed", "Ef", "Ek", "El", "Em",
691: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47 schwarze 692: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73 schwarze 693: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
694: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47 schwarze 695: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73 schwarze 696: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
697: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
698: "Sc", "Sh", "Sm", "So", "Sq",
1.47 schwarze 699: "Ss", "St", "Sx", "Sy",
700: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73 schwarze 701: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47 schwarze 702: "%P", "%Q", "%R", "%T", "%U", "%V",
703: NULL
704: };
705:
1.73 schwarze 706: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47 schwarze 707: const char *const __man_reserved[] = {
1.73 schwarze 708: "AT", "B", "BI", "BR", "DT",
709: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
710: "LP", "OP", "P", "PD", "PP",
711: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
712: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47 schwarze 713: NULL
714: };
715:
1.37 schwarze 716: /* Array of injected predefined strings. */
717: #define PREDEFS_MAX 38
718: static const struct predef predefs[PREDEFS_MAX] = {
719: #include "predefs.in"
720: };
721:
1.42 schwarze 722: /* See roffhash_find() */
1.3 schwarze 723: #define ROFF_HASH(p) (p[0] - ASCII_LO)
724:
1.51 schwarze 725: static int roffit_lines; /* number of lines to delay */
726: static char *roffit_macro; /* nil-terminated macro line */
727:
1.80 schwarze 728:
1.3 schwarze 729: static void
1.42 schwarze 730: roffhash_init(void)
1.3 schwarze 731: {
732: struct roffmac *n;
733: int buc, i;
734:
1.16 schwarze 735: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3 schwarze 736: assert(roffs[i].name[0] >= ASCII_LO);
737: assert(roffs[i].name[0] <= ASCII_HI);
738:
739: buc = ROFF_HASH(roffs[i].name);
740:
741: if (NULL != (n = hash[buc])) {
742: for ( ; n->next; n = n->next)
743: /* Do nothing. */ ;
744: n->next = &roffs[i];
745: } else
746: hash[buc] = &roffs[i];
747: }
748: }
749:
1.1 schwarze 750: /*
751: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
752: * the nil-terminated string name could be found.
753: */
754: static enum rofft
1.42 schwarze 755: roffhash_find(const char *p, size_t s)
1.1 schwarze 756: {
1.3 schwarze 757: int buc;
758: struct roffmac *n;
1.1 schwarze 759:
1.3 schwarze 760: /*
761: * libroff has an extremely simple hashtable, for the time
762: * being, which simply keys on the first character, which must
763: * be printable, then walks a chain. It works well enough until
764: * optimised.
765: */
766:
767: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
768: return(ROFF_MAX);
769:
770: buc = ROFF_HASH(p);
771:
772: if (NULL == (n = hash[buc]))
773: return(ROFF_MAX);
774: for ( ; n; n = n->next)
1.16 schwarze 775: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3 schwarze 776: return((enum rofft)(n - roffs));
1.1 schwarze 777:
778: return(ROFF_MAX);
779: }
780:
781: /*
782: * Pop the current node off of the stack of roff instructions currently
783: * pending.
784: */
785: static void
786: roffnode_pop(struct roff *r)
787: {
788: struct roffnode *p;
789:
1.2 schwarze 790: assert(r->last);
1.80 schwarze 791: p = r->last;
1.2 schwarze 792:
793: r->last = r->last->parent;
1.16 schwarze 794: free(p->name);
795: free(p->end);
1.1 schwarze 796: free(p);
797: }
798:
799: /*
800: * Push a roff node onto the instruction stack. This must later be
801: * removed with roffnode_pop().
802: */
1.11 schwarze 803: static void
1.16 schwarze 804: roffnode_push(struct roff *r, enum rofft tok, const char *name,
805: int line, int col)
1.1 schwarze 806: {
807: struct roffnode *p;
808:
1.11 schwarze 809: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 810: p->tok = tok;
1.16 schwarze 811: if (name)
812: p->name = mandoc_strdup(name);
1.1 schwarze 813: p->parent = r->last;
814: p->line = line;
815: p->col = col;
1.71 schwarze 816: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 817:
818: r->last = p;
819: }
820:
821: static void
822: roff_free1(struct roff *r)
823: {
1.49 schwarze 824: struct tbl_node *tbl;
1.32 schwarze 825: struct eqn_node *e;
1.42 schwarze 826: int i;
1.27 schwarze 827:
1.49 schwarze 828: while (NULL != (tbl = r->first_tbl)) {
829: r->first_tbl = tbl->next;
830: tbl_free(tbl);
1.27 schwarze 831: }
832: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 833:
1.32 schwarze 834: while (NULL != (e = r->first_eqn)) {
835: r->first_eqn = e->next;
836: eqn_free(e);
837: }
838: r->first_eqn = r->last_eqn = r->eqn = NULL;
839:
1.1 schwarze 840: while (r->last)
841: roffnode_pop(r);
1.27 schwarze 842:
1.96 schwarze 843: free (r->rstack);
844: r->rstack = NULL;
845: r->rstacksz = 0;
846: r->rstackpos = -1;
847:
848: roff_freereg(r->regtab);
849: r->regtab = NULL;
850:
1.42 schwarze 851: roff_freestr(r->strtab);
852: roff_freestr(r->xmbtab);
853: r->strtab = r->xmbtab = NULL;
854:
855: if (r->xtab)
856: for (i = 0; i < 128; i++)
857: free(r->xtab[i].p);
858: free(r->xtab);
859: r->xtab = NULL;
1.1 schwarze 860: }
861:
862: void
863: roff_reset(struct roff *r)
864: {
865:
866: roff_free1(r);
1.99 schwarze 867: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48 schwarze 868: r->control = 0;
1.1 schwarze 869: }
870:
871: void
872: roff_free(struct roff *r)
873: {
874:
875: roff_free1(r);
876: free(r);
877: }
878:
879: struct roff *
1.109 schwarze 880: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.1 schwarze 881: {
882: struct roff *r;
883:
1.11 schwarze 884: r = mandoc_calloc(1, sizeof(struct roff));
1.35 schwarze 885: r->parse = parse;
1.109 schwarze 886: r->mchars = mchars;
1.72 schwarze 887: r->options = options;
1.99 schwarze 888: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2 schwarze 889: r->rstackpos = -1;
1.80 schwarze 890:
1.42 schwarze 891: roffhash_init();
1.37 schwarze 892:
1.1 schwarze 893: return(r);
894: }
895:
1.8 schwarze 896: /*
1.79 schwarze 897: * In the current line, expand escape sequences that tend to get
898: * used in numerical expressions and conditional requests.
899: * Also check the syntax of the remaining escape sequences.
1.8 schwarze 900: */
1.45 schwarze 901: static enum rofferr
1.110 schwarze 902: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8 schwarze 903: {
1.81 schwarze 904: char ubuf[24]; /* buffer to print the number */
1.78 schwarze 905: const char *start; /* start of the string to process */
1.82 schwarze 906: char *stesc; /* start of an escape sequence ('\\') */
1.23 schwarze 907: const char *stnam; /* start of the name, after "[(*" */
908: const char *cp; /* end of the name, e.g. before ']' */
909: const char *res; /* the string to be substituted */
1.110 schwarze 910: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 911: size_t maxl; /* expected length of the escape name */
912: size_t naml; /* actual length of the escape name */
1.109 schwarze 913: enum mandoc_esc esc; /* type of the escape sequence */
914: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 915: int expand_count; /* to avoid infinite loops */
1.79 schwarze 916: int npos; /* position in numeric expression */
1.91 schwarze 917: int arg_complete; /* argument not interrupted by eol */
1.79 schwarze 918: char term; /* character terminating the escape */
1.8 schwarze 919:
1.43 schwarze 920: expand_count = 0;
1.110 schwarze 921: start = buf->buf + pos;
1.78 schwarze 922: stesc = strchr(start, '\0') - 1;
923: while (stesc-- > start) {
1.43 schwarze 924:
1.78 schwarze 925: /* Search backwards for the next backslash. */
926:
1.110 schwarze 927: if (*stesc != '\\')
1.78 schwarze 928: continue;
929:
930: /* If it is escaped, skip it. */
931:
932: for (cp = stesc - 1; cp >= start; cp--)
1.110 schwarze 933: if (*cp != '\\')
1.78 schwarze 934: break;
935:
1.110 schwarze 936: if ((stesc - cp) % 2 == 0) {
1.82 schwarze 937: stesc = (char *)cp;
1.78 schwarze 938: continue;
939: }
1.23 schwarze 940:
1.79 schwarze 941: /* Decide whether to expand or to check only. */
1.23 schwarze 942:
1.79 schwarze 943: term = '\0';
1.78 schwarze 944: cp = stesc + 1;
1.53 schwarze 945: switch (*cp) {
1.80 schwarze 946: case '*':
1.53 schwarze 947: res = NULL;
948: break;
1.80 schwarze 949: case 'B':
1.79 schwarze 950: /* FALLTHROUGH */
1.80 schwarze 951: case 'w':
1.79 schwarze 952: term = cp[1];
953: /* FALLTHROUGH */
1.80 schwarze 954: case 'n':
1.53 schwarze 955: res = ubuf;
956: break;
957: default:
1.109 schwarze 958: esc = mandoc_escape(&cp, &stnam, &inaml);
959: if (esc == ESCAPE_ERROR ||
960: (esc == ESCAPE_SPECIAL &&
961: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.92 schwarze 962: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110 schwarze 963: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 964: "%.*s", (int)(cp - stesc), stesc);
1.78 schwarze 965: continue;
1.42 schwarze 966: }
967:
1.78 schwarze 968: if (EXPAND_LIMIT < ++expand_count) {
969: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110 schwarze 970: ln, (int)(stesc - buf->buf), NULL);
1.78 schwarze 971: return(ROFF_IGN);
972: }
1.23 schwarze 973:
974: /*
975: * The third character decides the length
1.53 schwarze 976: * of the name of the string or register.
1.23 schwarze 977: * Save a pointer to the name.
978: */
979:
1.110 schwarze 980: if (term == '\0') {
1.79 schwarze 981: switch (*++cp) {
1.80 schwarze 982: case '\0':
1.79 schwarze 983: maxl = 0;
984: break;
1.80 schwarze 985: case '(':
1.79 schwarze 986: cp++;
987: maxl = 2;
988: break;
1.80 schwarze 989: case '[':
1.79 schwarze 990: cp++;
991: term = ']';
992: maxl = 0;
993: break;
994: default:
995: maxl = 1;
996: break;
997: }
998: } else {
999: cp += 2;
1.8 schwarze 1000: maxl = 0;
1001: }
1.23 schwarze 1002: stnam = cp;
1.8 schwarze 1003:
1.23 schwarze 1004: /* Advance to the end of the name. */
1.8 schwarze 1005:
1.125 schwarze 1006: naml = 0;
1.91 schwarze 1007: arg_complete = 1;
1.125 schwarze 1008: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1009: if (*cp == '\0') {
1.92 schwarze 1010: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 1011: ln, (int)(stesc - buf->buf), stesc);
1.91 schwarze 1012: arg_complete = 0;
1.79 schwarze 1013: break;
1.42 schwarze 1014: }
1.110 schwarze 1015: if (maxl == 0 && *cp == term) {
1.79 schwarze 1016: cp++;
1.125 schwarze 1017: break;
1018: }
1019: if (*cp++ != '\\' || stesc[1] != 'w') {
1020: naml++;
1021: continue;
1022: }
1023: switch (mandoc_escape(&cp, NULL, NULL)) {
1024: case ESCAPE_SPECIAL:
1025: /* FALLTHROUGH */
1026: case ESCAPE_UNICODE:
1027: /* FALLTHROUGH */
1028: case ESCAPE_NUMBERED:
1029: /* FALLTHROUGH */
1030: case ESCAPE_OVERSTRIKE:
1031: naml++;
1032: break;
1033: default:
1.8 schwarze 1034: break;
1.79 schwarze 1035: }
1.8 schwarze 1036: }
1037:
1.23 schwarze 1038: /*
1039: * Retrieve the replacement string; if it is
1040: * undefined, resume searching for escapes.
1041: */
1042:
1.79 schwarze 1043: switch (stesc[1]) {
1.80 schwarze 1044: case '*':
1.91 schwarze 1045: if (arg_complete)
1046: res = roff_getstrn(r, stnam, naml);
1.79 schwarze 1047: break;
1.80 schwarze 1048: case 'B':
1.79 schwarze 1049: npos = 0;
1.91 schwarze 1050: ubuf[0] = arg_complete &&
1.106 schwarze 1051: roff_evalnum(r, ln, stnam, &npos, NULL, 0) &&
1.91 schwarze 1052: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1053: ubuf[1] = '\0';
1054: break;
1.80 schwarze 1055: case 'n':
1.91 schwarze 1056: if (arg_complete)
1057: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1058: roff_getregn(r, stnam, naml));
1059: else
1060: ubuf[0] = '\0';
1.79 schwarze 1061: break;
1.80 schwarze 1062: case 'w':
1.91 schwarze 1063: /* use even incomplete args */
1.81 schwarze 1064: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1065: 24 * (int)naml);
1066: break;
1067: }
1.8 schwarze 1068:
1.110 schwarze 1069: if (res == NULL) {
1.92 schwarze 1070: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1071: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1072: "%.*s", (int)naml, stnam);
1.37 schwarze 1073: res = "";
1.118 schwarze 1074: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1075: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1076: ln, (int)(stesc - buf->buf), NULL);
1077: return(ROFF_IGN);
1.8 schwarze 1078: }
1079:
1.23 schwarze 1080: /* Replace the escape sequence by the string. */
1081:
1.82 schwarze 1082: *stesc = '\0';
1.110 schwarze 1083: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1084: buf->buf, res, cp) + 1;
1.8 schwarze 1085:
1.78 schwarze 1086: /* Prepare for the next replacement. */
1.8 schwarze 1087:
1.78 schwarze 1088: start = nbuf + pos;
1.110 schwarze 1089: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1090: free(buf->buf);
1091: buf->buf = nbuf;
1.42 schwarze 1092: }
1.45 schwarze 1093: return(ROFF_CONT);
1.42 schwarze 1094: }
1095:
1096: /*
1.51 schwarze 1097: * Process text streams:
1098: * Convert all breakable hyphens into ASCII_HYPH.
1099: * Decrement and spring input line trap.
1.42 schwarze 1100: */
1101: static enum rofferr
1.110 schwarze 1102: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42 schwarze 1103: {
1104: size_t sz;
1105: const char *start;
1.51 schwarze 1106: char *p;
1107: int isz;
1.42 schwarze 1108: enum mandoc_esc esc;
1109:
1.110 schwarze 1110: start = p = buf->buf + pos;
1.42 schwarze 1111:
1.110 schwarze 1112: while (*p != '\0') {
1.42 schwarze 1113: sz = strcspn(p, "-\\");
1114: p += sz;
1115:
1.110 schwarze 1116: if (*p == '\0')
1.42 schwarze 1117: break;
1118:
1.110 schwarze 1119: if (*p == '\\') {
1.42 schwarze 1120: /* Skip over escapes. */
1121: p++;
1.62 schwarze 1122: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1123: if (esc == ESCAPE_ERROR)
1.42 schwarze 1124: break;
1125: continue;
1126: } else if (p == start) {
1127: p++;
1128: continue;
1129: }
1130:
1.44 schwarze 1131: if (isalpha((unsigned char)p[-1]) &&
1132: isalpha((unsigned char)p[1]))
1.42 schwarze 1133: *p = ASCII_HYPH;
1134: p++;
1.8 schwarze 1135: }
1136:
1.51 schwarze 1137: /* Spring the input line trap. */
1.110 schwarze 1138: if (roffit_lines == 1) {
1139: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1140: free(buf->buf);
1141: buf->buf = p;
1142: buf->sz = isz + 1;
1.51 schwarze 1143: *offs = 0;
1144: free(roffit_macro);
1145: roffit_lines = 0;
1146: return(ROFF_REPARSE);
1.110 schwarze 1147: } else if (roffit_lines > 1)
1.51 schwarze 1148: --roffit_lines;
1.42 schwarze 1149: return(ROFF_CONT);
1.8 schwarze 1150: }
1151:
1.1 schwarze 1152: enum rofferr
1.110 schwarze 1153: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1154: {
1155: enum rofft t;
1.27 schwarze 1156: enum rofferr e;
1.110 schwarze 1157: int pos; /* parse point */
1.115 schwarze 1158: int spos; /* saved parse point for messages */
1.110 schwarze 1159: int ppos; /* original offset in buf->buf */
1160: int ctl; /* macro line (boolean) */
1161:
1162: ppos = pos = *offs;
1.1 schwarze 1163:
1.102 schwarze 1164: /* Handle in-line equation delimiters. */
1165:
1.108 schwarze 1166: if (r->tbl == NULL &&
1167: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1168: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1169: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1170: if (e == ROFF_REPARSE)
1171: return(e);
1172: assert(e == ROFF_CONT);
1173: }
1174:
1175: /* Expand some escape sequences. */
1.8 schwarze 1176:
1.110 schwarze 1177: e = roff_res(r, buf, ln, pos);
1178: if (e == ROFF_IGN)
1.45 schwarze 1179: return(e);
1.110 schwarze 1180: assert(e == ROFF_CONT);
1.8 schwarze 1181:
1.110 schwarze 1182: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1183:
1.8 schwarze 1184: /*
1.2 schwarze 1185: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1186: * text through the macro's filter.
1187: * Equations process all content themselves.
1188: * Tables process almost all content themselves, but we want
1189: * to warn about macros before passing it there.
1.2 schwarze 1190: */
1191:
1.124 schwarze 1192: if (r->last != NULL && ! ctl) {
1.2 schwarze 1193: t = r->last->tok;
1194: assert(roffs[t].text);
1.110 schwarze 1195: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1196: assert(e == ROFF_IGN || e == ROFF_CONT);
1197: if (e != ROFF_CONT)
1.32 schwarze 1198: return(e);
1.54 schwarze 1199: }
1.124 schwarze 1200: if (r->eqn != NULL)
1.110 schwarze 1201: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.124 schwarze 1202: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1203: return(tbl_read(r->tbl, ln, buf->buf, pos));
1204: if ( ! ctl)
1.110 schwarze 1205: return(roff_parsetext(buf, pos, offs));
1.100 schwarze 1206:
1207: /* Skip empty request lines. */
1208:
1.110 schwarze 1209: if (buf->buf[pos] == '"') {
1.100 schwarze 1210: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1211: ln, pos, NULL);
1212: return(ROFF_IGN);
1.110 schwarze 1213: } else if (buf->buf[pos] == '\0')
1.100 schwarze 1214: return(ROFF_IGN);
1.2 schwarze 1215:
1216: /*
1217: * If a scope is open, go to the child handler for that macro,
1218: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1219: * Don't do so if an equation is open.
1.2 schwarze 1220: */
1221:
1222: if (r->last) {
1.1 schwarze 1223: t = r->last->tok;
1224: assert(roffs[t].sub);
1.110 schwarze 1225: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1226: }
1227:
1.115 schwarze 1228: /* No scope is open. This is a new request or macro. */
1229:
1230: spos = pos;
1231: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1232:
1233: /* Tables ignore most macros. */
1234:
1235: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1236: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1237: ln, pos, buf->buf + spos);
1238: return(ROFF_IGN);
1239: }
1240:
1.2 schwarze 1241: /*
1.115 schwarze 1242: * This is neither a roff request nor a user-defined macro.
1243: * Let the standard macro set parsers handle it.
1.2 schwarze 1244: */
1245:
1.115 schwarze 1246: if (t == ROFF_MAX)
1.1 schwarze 1247: return(ROFF_CONT);
1.115 schwarze 1248:
1249: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1250:
1.2 schwarze 1251: assert(roffs[t].proc);
1.110 schwarze 1252: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1253: }
1254:
1.27 schwarze 1255: void
1.2 schwarze 1256: roff_endparse(struct roff *r)
1257: {
1.1 schwarze 1258:
1.27 schwarze 1259: if (r->last)
1.94 schwarze 1260: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1261: r->last->line, r->last->col,
1262: roffs[r->last->tok].name);
1.27 schwarze 1263:
1.32 schwarze 1264: if (r->eqn) {
1.94 schwarze 1265: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1266: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41 schwarze 1267: eqn_end(&r->eqn);
1.32 schwarze 1268: }
1269:
1.27 schwarze 1270: if (r->tbl) {
1.94 schwarze 1271: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1272: r->tbl->line, r->tbl->pos, "TS");
1.41 schwarze 1273: tbl_end(&r->tbl);
1.27 schwarze 1274: }
1.1 schwarze 1275: }
1276:
1277: /*
1278: * Parse a roff node's type from the input buffer. This must be in the
1279: * form of ".foo xxx" in the usual way.
1280: */
1281: static enum rofft
1.87 schwarze 1282: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1283: {
1.87 schwarze 1284: char *cp;
1.16 schwarze 1285: const char *mac;
1286: size_t maclen;
1.1 schwarze 1287: enum rofft t;
1288:
1.87 schwarze 1289: cp = buf + *pos;
1290:
1291: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.1 schwarze 1292: return(ROFF_MAX);
1293:
1.87 schwarze 1294: mac = cp;
1295: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 1296:
1.16 schwarze 1297: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42 schwarze 1298: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1 schwarze 1299:
1.87 schwarze 1300: if (ROFF_MAX != t)
1301: *pos = cp - buf;
1.1 schwarze 1302:
1303: return(t);
1304: }
1305:
1306: static enum rofferr
1.2 schwarze 1307: roff_cblock(ROFF_ARGS)
1.1 schwarze 1308: {
1309:
1.2 schwarze 1310: /*
1311: * A block-close `..' should only be invoked as a child of an
1312: * ignore macro, otherwise raise a warning and just ignore it.
1313: */
1314:
1.110 schwarze 1315: if (r->last == NULL) {
1.94 schwarze 1316: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1317: ln, ppos, "..");
1.2 schwarze 1318: return(ROFF_IGN);
1319: }
1.1 schwarze 1320:
1.2 schwarze 1321: switch (r->last->tok) {
1.80 schwarze 1322: case ROFF_am:
1.93 schwarze 1323: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.2 schwarze 1324: /* FALLTHROUGH */
1.80 schwarze 1325: case ROFF_ami:
1.2 schwarze 1326: /* FALLTHROUGH */
1.80 schwarze 1327: case ROFF_de:
1.23 schwarze 1328: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2 schwarze 1329: /* FALLTHROUGH */
1.80 schwarze 1330: case ROFF_dei:
1.2 schwarze 1331: /* FALLTHROUGH */
1.80 schwarze 1332: case ROFF_ig:
1.2 schwarze 1333: break;
1334: default:
1.94 schwarze 1335: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1336: ln, ppos, "..");
1.1 schwarze 1337: return(ROFF_IGN);
1.2 schwarze 1338: }
1339:
1.110 schwarze 1340: if (buf->buf[pos] != '\0')
1.90 schwarze 1341: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 1342: ".. %s", buf->buf + pos);
1.2 schwarze 1343:
1344: roffnode_pop(r);
1345: roffnode_cleanscope(r);
1346: return(ROFF_IGN);
1347:
1348: }
1.1 schwarze 1349:
1.2 schwarze 1350: static void
1351: roffnode_cleanscope(struct roff *r)
1352: {
1.1 schwarze 1353:
1.2 schwarze 1354: while (r->last) {
1.46 schwarze 1355: if (--r->last->endspan != 0)
1.2 schwarze 1356: break;
1357: roffnode_pop(r);
1358: }
1359: }
1.1 schwarze 1360:
1.68 schwarze 1361: static void
1362: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 1363: {
1.1 schwarze 1364:
1.2 schwarze 1365: if (NULL == r->last) {
1.94 schwarze 1366: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1367: ln, ppos, "\\}");
1.68 schwarze 1368: return;
1.2 schwarze 1369: }
1.1 schwarze 1370:
1.2 schwarze 1371: switch (r->last->tok) {
1.80 schwarze 1372: case ROFF_el:
1.2 schwarze 1373: /* FALLTHROUGH */
1.80 schwarze 1374: case ROFF_ie:
1.2 schwarze 1375: /* FALLTHROUGH */
1.80 schwarze 1376: case ROFF_if:
1.2 schwarze 1377: break;
1378: default:
1.94 schwarze 1379: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1380: ln, ppos, "\\}");
1.68 schwarze 1381: return;
1.2 schwarze 1382: }
1.1 schwarze 1383:
1.2 schwarze 1384: if (r->last->endspan > -1) {
1.94 schwarze 1385: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1386: ln, ppos, "\\}");
1.68 schwarze 1387: return;
1.2 schwarze 1388: }
1389:
1390: roffnode_pop(r);
1391: roffnode_cleanscope(r);
1.68 schwarze 1392: return;
1.1 schwarze 1393: }
1394:
1395: static enum rofferr
1.2 schwarze 1396: roff_block(ROFF_ARGS)
1.1 schwarze 1397: {
1.93 schwarze 1398: const char *name;
1399: char *iname, *cp;
1.86 schwarze 1400: size_t namesz;
1.16 schwarze 1401:
1.93 schwarze 1402: /* Ignore groff compatibility mode for now. */
1.2 schwarze 1403:
1.110 schwarze 1404: if (tok == ROFF_de1)
1.93 schwarze 1405: tok = ROFF_de;
1.123 schwarze 1406: else if (tok == ROFF_dei1)
1407: tok = ROFF_dei;
1.110 schwarze 1408: else if (tok == ROFF_am1)
1.93 schwarze 1409: tok = ROFF_am;
1.123 schwarze 1410: else if (tok == ROFF_ami1)
1411: tok = ROFF_ami;
1.93 schwarze 1412:
1413: /* Parse the macro name argument. */
1414:
1.110 schwarze 1415: cp = buf->buf + pos;
1416: if (tok == ROFF_ig) {
1.93 schwarze 1417: iname = NULL;
1418: namesz = 0;
1419: } else {
1420: iname = cp;
1421: namesz = roff_getname(r, &cp, ln, ppos);
1422: iname[namesz] = '\0';
1423: }
1.22 schwarze 1424:
1.93 schwarze 1425: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 1426:
1.110 schwarze 1427: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1428: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1429: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1430: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1431: "%.*s", (int)namesz, iname);
1432: namesz = 0;
1433: } else
1434: namesz = strlen(name);
1435: } else
1436: name = iname;
1.22 schwarze 1437:
1.110 schwarze 1438: if (namesz == 0 && tok != ROFF_ig) {
1.93 schwarze 1439: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1440: ln, ppos, roffs[tok].name);
1441: return(ROFF_IGN);
1442: }
1.2 schwarze 1443:
1.16 schwarze 1444: roffnode_push(r, tok, name, ln, ppos);
1445:
1446: /*
1447: * At the beginning of a `de' macro, clear the existing string
1448: * with the same name, if there is one. New content will be
1.66 schwarze 1449: * appended from roff_block_text() in multiline mode.
1.16 schwarze 1450: */
1.22 schwarze 1451:
1.110 schwarze 1452: if (tok == ROFF_de || tok == ROFF_dei)
1.86 schwarze 1453: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2 schwarze 1454:
1.110 schwarze 1455: if (*cp == '\0')
1.2 schwarze 1456: return(ROFF_IGN);
1.1 schwarze 1457:
1.93 schwarze 1458: /* Get the custom end marker. */
1.22 schwarze 1459:
1.93 schwarze 1460: iname = cp;
1.86 schwarze 1461: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 1462:
1463: /* Resolve the end marker if it is indirect. */
1464:
1.110 schwarze 1465: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1466: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1467: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1468: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1469: "%.*s", (int)namesz, iname);
1470: namesz = 0;
1471: } else
1472: namesz = strlen(name);
1473: } else
1474: name = iname;
1475:
1.86 schwarze 1476: if (namesz)
1477: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 1478:
1.110 schwarze 1479: if (*cp != '\0')
1.90 schwarze 1480: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1481: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.1 schwarze 1482:
1483: return(ROFF_IGN);
1484: }
1485:
1486: static enum rofferr
1.2 schwarze 1487: roff_block_sub(ROFF_ARGS)
1.1 schwarze 1488: {
1.2 schwarze 1489: enum rofft t;
1490: int i, j;
1491:
1492: /*
1493: * First check whether a custom macro exists at this level. If
1494: * it does, then check against it. This is some of groff's
1495: * stranger behaviours. If we encountered a custom end-scope
1496: * tag and that tag also happens to be a "real" macro, then we
1497: * need to try interpreting it again as a real macro. If it's
1498: * not, then return ignore. Else continue.
1499: */
1500:
1501: if (r->last->end) {
1.35 schwarze 1502: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 1503: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 1504: break;
1.1 schwarze 1505:
1.110 schwarze 1506: if (r->last->end[j] == '\0' &&
1507: (buf->buf[i] == '\0' ||
1508: buf->buf[i] == ' ' ||
1509: buf->buf[i] == '\t')) {
1.2 schwarze 1510: roffnode_pop(r);
1511: roffnode_cleanscope(r);
1.1 schwarze 1512:
1.110 schwarze 1513: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 1514: i++;
1515:
1516: pos = i;
1.110 schwarze 1517: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1518: ROFF_MAX)
1.2 schwarze 1519: return(ROFF_RERUN);
1520: return(ROFF_IGN);
1521: }
1.1 schwarze 1522: }
1523:
1.2 schwarze 1524: /*
1525: * If we have no custom end-query or lookup failed, then try
1526: * pulling it out of the hashtable.
1527: */
1.1 schwarze 1528:
1.110 schwarze 1529: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 1530:
1.110 schwarze 1531: if (t != ROFF_cblock) {
1532: if (tok != ROFF_ig)
1533: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.1 schwarze 1534: return(ROFF_IGN);
1.16 schwarze 1535: }
1.1 schwarze 1536:
1.2 schwarze 1537: assert(roffs[t].proc);
1.110 schwarze 1538: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1539: }
1540:
1541: static enum rofferr
1542: roff_block_text(ROFF_ARGS)
1543: {
1544:
1.110 schwarze 1545: if (tok != ROFF_ig)
1546: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 1547:
1.2 schwarze 1548: return(ROFF_IGN);
1549: }
1550:
1551: static enum rofferr
1552: roff_cond_sub(ROFF_ARGS)
1553: {
1554: enum rofft t;
1.37 schwarze 1555: char *ep;
1.71 schwarze 1556: int rr;
1.2 schwarze 1557:
1558: rr = r->last->rule;
1.37 schwarze 1559: roffnode_cleanscope(r);
1.110 schwarze 1560: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2 schwarze 1561:
1.37 schwarze 1562: /*
1.50 schwarze 1563: * Fully handle known macros when they are structurally
1564: * required or when the conditional evaluated to true.
1.5 schwarze 1565: */
1566:
1.110 schwarze 1567: if ((t != ROFF_MAX) &&
1568: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.50 schwarze 1569: assert(roffs[t].proc);
1.110 schwarze 1570: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.50 schwarze 1571: }
1.39 schwarze 1572:
1.69 schwarze 1573: /*
1574: * If `\}' occurs on a macro line without a preceding macro,
1575: * drop the line completely.
1576: */
1577:
1.110 schwarze 1578: ep = buf->buf + pos;
1579: if (ep[0] == '\\' && ep[1] == '}')
1.71 schwarze 1580: rr = 0;
1.69 schwarze 1581:
1.50 schwarze 1582: /* Always check for the closing delimiter `\}'. */
1.39 schwarze 1583:
1.110 schwarze 1584: while ((ep = strchr(ep, '\\')) != NULL) {
1585: if (*(++ep) == '}') {
1.70 schwarze 1586: *ep = '&';
1.110 schwarze 1587: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1588: }
1.119 schwarze 1589: if (*ep != '\0')
1590: ++ep;
1.50 schwarze 1591: }
1.71 schwarze 1592: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1593: }
1594:
1595: static enum rofferr
1596: roff_cond_text(ROFF_ARGS)
1597: {
1.37 schwarze 1598: char *ep;
1.71 schwarze 1599: int rr;
1.2 schwarze 1600:
1601: rr = r->last->rule;
1.37 schwarze 1602: roffnode_cleanscope(r);
1.1 schwarze 1603:
1.110 schwarze 1604: ep = buf->buf + pos;
1605: while ((ep = strchr(ep, '\\')) != NULL) {
1606: if (*(++ep) == '}') {
1.70 schwarze 1607: *ep = '&';
1.110 schwarze 1608: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1609: }
1.119 schwarze 1610: if (*ep != '\0')
1611: ++ep;
1.2 schwarze 1612: }
1.71 schwarze 1613: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1614: }
1615:
1.77 schwarze 1616: /*
1617: * Parse a single signed integer number. Stop at the first non-digit.
1618: * If there is at least one digit, return success and advance the
1619: * parse point, else return failure and let the parse point unchanged.
1620: * Ignore overflows, treat them just like the C language.
1621: */
1.56 schwarze 1622: static int
1623: roff_getnum(const char *v, int *pos, int *res)
1624: {
1.79 schwarze 1625: int myres, n, p;
1626:
1627: if (NULL == res)
1628: res = &myres;
1.56 schwarze 1629:
1630: p = *pos;
1631: n = v[p] == '-';
1632: if (n)
1633: p++;
1634:
1635: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 1636: *res = 10 * *res + v[p] - '0';
1.56 schwarze 1637: if (p == *pos + n)
1638: return 0;
1639:
1640: if (n)
1641: *res = -*res;
1642:
1.126 schwarze 1643: /* Each number may be followed by one optional scaling unit. */
1644:
1645: switch (v[p]) {
1646: case 'f':
1647: *res *= 65536;
1648: break;
1649: case 'i':
1650: *res *= 240;
1651: break;
1652: case 'c':
1653: *res *= 240;
1654: *res /= 2.54;
1655: break;
1656: case 'v':
1657: /* FALLTROUGH */
1658: case 'P':
1659: *res *= 40;
1660: break;
1661: case 'm':
1662: /* FALLTROUGH */
1663: case 'n':
1664: *res *= 24;
1665: break;
1666: case 'p':
1667: *res *= 10;
1668: *res /= 3;
1669: break;
1670: case 'u':
1671: break;
1672: case 'M':
1673: *res *= 6;
1674: *res /= 25;
1675: break;
1676: default:
1677: p--;
1678: break;
1679: }
1680:
1681: *pos = p + 1;
1682: return(1);
1.56 schwarze 1683: }
1684:
1.71 schwarze 1685: /*
1686: * Evaluate a string comparison condition.
1687: * The first character is the delimiter.
1688: * Succeed if the string up to its second occurrence
1689: * matches the string up to its third occurence.
1690: * Advance the cursor after the third occurrence
1691: * or lacking that, to the end of the line.
1692: */
1693: static int
1694: roff_evalstrcond(const char *v, int *pos)
1695: {
1696: const char *s1, *s2, *s3;
1697: int match;
1698:
1699: match = 0;
1700: s1 = v + *pos; /* initial delimiter */
1701: s2 = s1 + 1; /* for scanning the first string */
1702: s3 = strchr(s2, *s1); /* for scanning the second string */
1703:
1704: if (NULL == s3) /* found no middle delimiter */
1705: goto out;
1706:
1707: while ('\0' != *++s3) {
1708: if (*s2 != *s3) { /* mismatch */
1709: s3 = strchr(s3, *s1);
1710: break;
1711: }
1712: if (*s3 == *s1) { /* found the final delimiter */
1713: match = 1;
1714: break;
1715: }
1716: s2++;
1717: }
1718:
1719: out:
1720: if (NULL == s3)
1721: s3 = strchr(s2, '\0');
1.114 schwarze 1722: else if (*s3 != '\0')
1.71 schwarze 1723: s3++;
1724: *pos = s3 - v;
1725: return(match);
1726: }
1727:
1.77 schwarze 1728: /*
1729: * Evaluate an optionally negated single character, numerical,
1730: * or string condition.
1731: */
1.71 schwarze 1732: static int
1.106 schwarze 1733: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.5 schwarze 1734: {
1.113 schwarze 1735: int number, savepos, wanttrue;
1.5 schwarze 1736:
1.71 schwarze 1737: if ('!' == v[*pos]) {
1738: wanttrue = 0;
1739: (*pos)++;
1740: } else
1741: wanttrue = 1;
1742:
1.5 schwarze 1743: switch (v[*pos]) {
1.112 schwarze 1744: case '\0':
1745: return(0);
1.80 schwarze 1746: case 'n':
1.71 schwarze 1747: /* FALLTHROUGH */
1.80 schwarze 1748: case 'o':
1.5 schwarze 1749: (*pos)++;
1.71 schwarze 1750: return(wanttrue);
1.80 schwarze 1751: case 'c':
1.71 schwarze 1752: /* FALLTHROUGH */
1.80 schwarze 1753: case 'd':
1.71 schwarze 1754: /* FALLTHROUGH */
1.80 schwarze 1755: case 'e':
1.5 schwarze 1756: /* FALLTHROUGH */
1.80 schwarze 1757: case 'r':
1.5 schwarze 1758: /* FALLTHROUGH */
1.80 schwarze 1759: case 't':
1.111 schwarze 1760: /* FALLTHROUGH */
1761: case 'v':
1.5 schwarze 1762: (*pos)++;
1.71 schwarze 1763: return(!wanttrue);
1.5 schwarze 1764: default:
1765: break;
1766: }
1767:
1.113 schwarze 1768: savepos = *pos;
1.106 schwarze 1769: if (roff_evalnum(r, ln, v, pos, &number, 0))
1.77 schwarze 1770: return((number > 0) == wanttrue);
1.113 schwarze 1771: else if (*pos == savepos)
1772: return(roff_evalstrcond(v, pos) == wanttrue);
1.77 schwarze 1773: else
1.113 schwarze 1774: return (0);
1.5 schwarze 1775: }
1776:
1.2 schwarze 1777: static enum rofferr
1.21 schwarze 1778: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 1779: {
1.30 schwarze 1780:
1.21 schwarze 1781: return(ROFF_IGN);
1782: }
1783:
1784: static enum rofferr
1.123 schwarze 1785: roff_insec(ROFF_ARGS)
1786: {
1787:
1788: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1789: ln, ppos, roffs[tok].name);
1790: return(ROFF_IGN);
1791: }
1792:
1793: static enum rofferr
1794: roff_unsupp(ROFF_ARGS)
1795: {
1796:
1797: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1798: ln, ppos, roffs[tok].name);
1799: return(ROFF_IGN);
1800: }
1801:
1802: static enum rofferr
1.2 schwarze 1803: roff_cond(ROFF_ARGS)
1804: {
1.46 schwarze 1805:
1806: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 1807:
1.80 schwarze 1808: /*
1.35 schwarze 1809: * An `.el' has no conditional body: it will consume the value
1810: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 1811: * defaults to DENY.
1.35 schwarze 1812: *
1813: * If we're not an `el', however, then evaluate the conditional.
1814: */
1.1 schwarze 1815:
1.110 schwarze 1816: r->last->rule = tok == ROFF_el ?
1.80 schwarze 1817: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 1818: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 1819:
1.35 schwarze 1820: /*
1821: * An if-else will put the NEGATION of the current evaluated
1822: * conditional into the stack of rules.
1823: */
1824:
1.110 schwarze 1825: if (tok == ROFF_ie) {
1.96 schwarze 1826: if (r->rstackpos + 1 == r->rstacksz) {
1827: r->rstacksz += 16;
1828: r->rstack = mandoc_reallocarray(r->rstack,
1829: r->rstacksz, sizeof(int));
1.35 schwarze 1830: }
1.71 schwarze 1831: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 1832: }
1.5 schwarze 1833:
1834: /* If the parent has false as its rule, then so do we. */
1835:
1.71 schwarze 1836: if (r->last->parent && !r->last->parent->rule)
1837: r->last->rule = 0;
1.5 schwarze 1838:
1839: /*
1.46 schwarze 1840: * Determine scope.
1841: * If there is nothing on the line after the conditional,
1842: * not even whitespace, use next-line scope.
1.5 schwarze 1843: */
1.2 schwarze 1844:
1.110 schwarze 1845: if (buf->buf[pos] == '\0') {
1.46 schwarze 1846: r->last->endspan = 2;
1847: goto out;
1848: }
1849:
1.110 schwarze 1850: while (buf->buf[pos] == ' ')
1.46 schwarze 1851: pos++;
1852:
1853: /* An opening brace requests multiline scope. */
1.2 schwarze 1854:
1.110 schwarze 1855: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 1856: r->last->endspan = -1;
1857: pos += 2;
1.46 schwarze 1858: goto out;
1.80 schwarze 1859: }
1.2 schwarze 1860:
1861: /*
1.46 schwarze 1862: * Anything else following the conditional causes
1863: * single-line scope. Warn if the scope contains
1864: * nothing but trailing whitespace.
1.2 schwarze 1865: */
1866:
1.110 schwarze 1867: if (buf->buf[pos] == '\0')
1.89 schwarze 1868: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1869: ln, ppos, roffs[tok].name);
1.2 schwarze 1870:
1.46 schwarze 1871: r->last->endspan = 1;
1.1 schwarze 1872:
1.46 schwarze 1873: out:
1.2 schwarze 1874: *offs = pos;
1875: return(ROFF_RERUN);
1.1 schwarze 1876: }
1877:
1.2 schwarze 1878: static enum rofferr
1.7 schwarze 1879: roff_ds(ROFF_ARGS)
1880: {
1.85 schwarze 1881: char *string;
1882: const char *name;
1883: size_t namesz;
1.10 schwarze 1884:
1.123 schwarze 1885: /* Ignore groff compatibility mode for now. */
1886:
1887: if (tok == ROFF_ds1)
1888: tok = ROFF_ds;
1889: else if (tok == ROFF_as1)
1890: tok = ROFF_as;
1891:
1.10 schwarze 1892: /*
1.85 schwarze 1893: * The first word is the name of the string.
1894: * If it is empty or terminated by an escape sequence,
1895: * abort the `ds' request without defining anything.
1.10 schwarze 1896: */
1.7 schwarze 1897:
1.110 schwarze 1898: name = string = buf->buf + pos;
1899: if (*name == '\0')
1.7 schwarze 1900: return(ROFF_IGN);
1901:
1.85 schwarze 1902: namesz = roff_getname(r, &string, ln, pos);
1.110 schwarze 1903: if (name[namesz] == '\\')
1.85 schwarze 1904: return(ROFF_IGN);
1905:
1906: /* Read past the initial double-quote, if any. */
1.110 schwarze 1907: if (*string == '"')
1.7 schwarze 1908: string++;
1909:
1.10 schwarze 1910: /* The rest is the value. */
1.85 schwarze 1911: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
1912: ROFF_as == tok);
1.7 schwarze 1913: return(ROFF_IGN);
1914: }
1915:
1.77 schwarze 1916: /*
1917: * Parse a single operator, one or two characters long.
1918: * If the operator is recognized, return success and advance the
1919: * parse point, else return failure and let the parse point unchanged.
1920: */
1921: static int
1922: roff_getop(const char *v, int *pos, char *res)
1923: {
1924:
1925: *res = v[*pos];
1926:
1927: switch (*res) {
1.80 schwarze 1928: case '+':
1.77 schwarze 1929: /* FALLTHROUGH */
1.80 schwarze 1930: case '-':
1.77 schwarze 1931: /* FALLTHROUGH */
1.80 schwarze 1932: case '*':
1.77 schwarze 1933: /* FALLTHROUGH */
1.80 schwarze 1934: case '/':
1.77 schwarze 1935: /* FALLTHROUGH */
1.80 schwarze 1936: case '%':
1.77 schwarze 1937: /* FALLTHROUGH */
1.80 schwarze 1938: case '&':
1.77 schwarze 1939: /* FALLTHROUGH */
1.80 schwarze 1940: case ':':
1.77 schwarze 1941: break;
1942: case '<':
1943: switch (v[*pos + 1]) {
1.80 schwarze 1944: case '=':
1.77 schwarze 1945: *res = 'l';
1946: (*pos)++;
1947: break;
1.80 schwarze 1948: case '>':
1.77 schwarze 1949: *res = '!';
1950: (*pos)++;
1951: break;
1.80 schwarze 1952: case '?':
1.77 schwarze 1953: *res = 'i';
1954: (*pos)++;
1955: break;
1956: default:
1957: break;
1958: }
1959: break;
1960: case '>':
1961: switch (v[*pos + 1]) {
1.80 schwarze 1962: case '=':
1.77 schwarze 1963: *res = 'g';
1964: (*pos)++;
1965: break;
1.80 schwarze 1966: case '?':
1.77 schwarze 1967: *res = 'a';
1968: (*pos)++;
1969: break;
1970: default:
1971: break;
1972: }
1973: break;
1974: case '=':
1975: if ('=' == v[*pos + 1])
1976: (*pos)++;
1977: break;
1978: default:
1979: return(0);
1980: }
1981: (*pos)++;
1982:
1983: return(*res);
1984: }
1985:
1986: /*
1987: * Evaluate either a parenthesized numeric expression
1988: * or a single signed integer number.
1989: */
1990: static int
1.106 schwarze 1991: roff_evalpar(struct roff *r, int ln,
1992: const char *v, int *pos, int *res)
1.77 schwarze 1993: {
1994:
1995: if ('(' != v[*pos])
1996: return(roff_getnum(v, pos, res));
1997:
1998: (*pos)++;
1.106 schwarze 1999: if ( ! roff_evalnum(r, ln, v, pos, res, 1))
1.77 schwarze 2000: return(0);
2001:
1.79 schwarze 2002: /*
2003: * Omission of the closing parenthesis
2004: * is an error in validation mode,
2005: * but ignored in evaluation mode.
2006: */
2007:
1.77 schwarze 2008: if (')' == v[*pos])
2009: (*pos)++;
1.79 schwarze 2010: else if (NULL == res)
2011: return(0);
1.77 schwarze 2012:
2013: return(1);
2014: }
2015:
2016: /*
2017: * Evaluate a complete numeric expression.
2018: * Proceed left to right, there is no concept of precedence.
2019: */
2020: static int
1.106 schwarze 2021: roff_evalnum(struct roff *r, int ln, const char *v,
2022: int *pos, int *res, int skipwhite)
1.77 schwarze 2023: {
2024: int mypos, operand2;
2025: char operator;
2026:
2027: if (NULL == pos) {
2028: mypos = 0;
2029: pos = &mypos;
2030: }
2031:
2032: if (skipwhite)
2033: while (isspace((unsigned char)v[*pos]))
2034: (*pos)++;
2035:
1.106 schwarze 2036: if ( ! roff_evalpar(r, ln, v, pos, res))
1.77 schwarze 2037: return(0);
2038:
2039: while (1) {
2040: if (skipwhite)
2041: while (isspace((unsigned char)v[*pos]))
2042: (*pos)++;
2043:
2044: if ( ! roff_getop(v, pos, &operator))
2045: break;
2046:
2047: if (skipwhite)
2048: while (isspace((unsigned char)v[*pos]))
2049: (*pos)++;
2050:
1.106 schwarze 2051: if ( ! roff_evalpar(r, ln, v, pos, &operand2))
1.77 schwarze 2052: return(0);
2053:
2054: if (skipwhite)
2055: while (isspace((unsigned char)v[*pos]))
2056: (*pos)++;
1.79 schwarze 2057:
2058: if (NULL == res)
2059: continue;
1.77 schwarze 2060:
2061: switch (operator) {
1.80 schwarze 2062: case '+':
1.77 schwarze 2063: *res += operand2;
2064: break;
1.80 schwarze 2065: case '-':
1.77 schwarze 2066: *res -= operand2;
2067: break;
1.80 schwarze 2068: case '*':
1.77 schwarze 2069: *res *= operand2;
2070: break;
1.80 schwarze 2071: case '/':
1.116 schwarze 2072: if (operand2 == 0) {
1.106 schwarze 2073: mandoc_msg(MANDOCERR_DIVZERO,
2074: r->parse, ln, *pos, v);
2075: *res = 0;
2076: break;
2077: }
1.77 schwarze 2078: *res /= operand2;
2079: break;
1.80 schwarze 2080: case '%':
1.116 schwarze 2081: if (operand2 == 0) {
2082: mandoc_msg(MANDOCERR_DIVZERO,
2083: r->parse, ln, *pos, v);
2084: *res = 0;
2085: break;
2086: }
1.77 schwarze 2087: *res %= operand2;
2088: break;
1.80 schwarze 2089: case '<':
1.77 schwarze 2090: *res = *res < operand2;
2091: break;
1.80 schwarze 2092: case '>':
1.77 schwarze 2093: *res = *res > operand2;
2094: break;
1.80 schwarze 2095: case 'l':
1.77 schwarze 2096: *res = *res <= operand2;
2097: break;
1.80 schwarze 2098: case 'g':
1.77 schwarze 2099: *res = *res >= operand2;
2100: break;
1.80 schwarze 2101: case '=':
1.77 schwarze 2102: *res = *res == operand2;
2103: break;
1.80 schwarze 2104: case '!':
1.77 schwarze 2105: *res = *res != operand2;
2106: break;
1.80 schwarze 2107: case '&':
1.77 schwarze 2108: *res = *res && operand2;
2109: break;
1.80 schwarze 2110: case ':':
1.77 schwarze 2111: *res = *res || operand2;
2112: break;
1.80 schwarze 2113: case 'i':
1.77 schwarze 2114: if (operand2 < *res)
2115: *res = operand2;
2116: break;
1.80 schwarze 2117: case 'a':
1.77 schwarze 2118: if (operand2 > *res)
2119: *res = operand2;
2120: break;
2121: default:
2122: abort();
2123: }
2124: }
2125: return(1);
2126: }
2127:
1.52 schwarze 2128: void
1.60 schwarze 2129: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 2130: {
1.52 schwarze 2131: struct roffreg *reg;
2132:
2133: /* Search for an existing register with the same name. */
2134: reg = r->regtab;
2135:
2136: while (reg && strcmp(name, reg->key.p))
2137: reg = reg->next;
1.41 schwarze 2138:
1.52 schwarze 2139: if (NULL == reg) {
2140: /* Create a new register. */
2141: reg = mandoc_malloc(sizeof(struct roffreg));
2142: reg->key.p = mandoc_strdup(name);
2143: reg->key.sz = strlen(name);
1.60 schwarze 2144: reg->val = 0;
1.52 schwarze 2145: reg->next = r->regtab;
2146: r->regtab = reg;
2147: }
2148:
1.60 schwarze 2149: if ('+' == sign)
2150: reg->val += val;
2151: else if ('-' == sign)
2152: reg->val -= val;
2153: else
2154: reg->val = val;
1.41 schwarze 2155: }
2156:
1.65 schwarze 2157: /*
2158: * Handle some predefined read-only number registers.
2159: * For now, return -1 if the requested register is not predefined;
2160: * in case a predefined read-only register having the value -1
2161: * were to turn up, another special value would have to be chosen.
2162: */
2163: static int
2164: roff_getregro(const char *name)
2165: {
2166:
2167: switch (*name) {
1.80 schwarze 2168: case 'A': /* ASCII approximation mode is always off. */
1.65 schwarze 2169: return(0);
1.80 schwarze 2170: case 'g': /* Groff compatibility mode is always on. */
1.65 schwarze 2171: return(1);
1.80 schwarze 2172: case 'H': /* Fixed horizontal resolution. */
1.65 schwarze 2173: return (24);
1.80 schwarze 2174: case 'j': /* Always adjust left margin only. */
1.65 schwarze 2175: return(0);
1.80 schwarze 2176: case 'T': /* Some output device is always defined. */
1.65 schwarze 2177: return(1);
1.80 schwarze 2178: case 'V': /* Fixed vertical resolution. */
1.65 schwarze 2179: return (40);
2180: default:
2181: return (-1);
2182: }
2183: }
2184:
1.53 schwarze 2185: int
1.52 schwarze 2186: roff_getreg(const struct roff *r, const char *name)
1.41 schwarze 2187: {
1.52 schwarze 2188: struct roffreg *reg;
1.65 schwarze 2189: int val;
2190:
2191: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
2192: val = roff_getregro(name + 1);
2193: if (-1 != val)
2194: return (val);
2195: }
1.52 schwarze 2196:
2197: for (reg = r->regtab; reg; reg = reg->next)
2198: if (0 == strcmp(name, reg->key.p))
1.53 schwarze 2199: return(reg->val);
2200:
2201: return(0);
2202: }
2203:
2204: static int
2205: roff_getregn(const struct roff *r, const char *name, size_t len)
2206: {
2207: struct roffreg *reg;
1.65 schwarze 2208: int val;
2209:
2210: if ('.' == name[0] && 2 == len) {
2211: val = roff_getregro(name + 1);
2212: if (-1 != val)
2213: return (val);
2214: }
1.53 schwarze 2215:
2216: for (reg = r->regtab; reg; reg = reg->next)
2217: if (len == reg->key.sz &&
2218: 0 == strncmp(name, reg->key.p, len))
2219: return(reg->val);
1.41 schwarze 2220:
1.52 schwarze 2221: return(0);
1.41 schwarze 2222: }
2223:
1.52 schwarze 2224: static void
2225: roff_freereg(struct roffreg *reg)
1.41 schwarze 2226: {
1.52 schwarze 2227: struct roffreg *old_reg;
1.41 schwarze 2228:
1.52 schwarze 2229: while (NULL != reg) {
2230: free(reg->key.p);
2231: old_reg = reg;
2232: reg = reg->next;
2233: free(old_reg);
2234: }
1.41 schwarze 2235: }
1.7 schwarze 2236:
2237: static enum rofferr
1.6 schwarze 2238: roff_nr(ROFF_ARGS)
1.1 schwarze 2239: {
1.85 schwarze 2240: char *key, *val;
2241: size_t keysz;
1.37 schwarze 2242: int iv;
1.60 schwarze 2243: char sign;
1.6 schwarze 2244:
1.110 schwarze 2245: key = val = buf->buf + pos;
2246: if (*key == '\0')
1.85 schwarze 2247: return(ROFF_IGN);
2248:
2249: keysz = roff_getname(r, &val, ln, pos);
1.110 schwarze 2250: if (key[keysz] == '\\')
1.85 schwarze 2251: return(ROFF_IGN);
2252: key[keysz] = '\0';
1.6 schwarze 2253:
1.60 schwarze 2254: sign = *val;
1.110 schwarze 2255: if (sign == '+' || sign == '-')
1.60 schwarze 2256: val++;
2257:
1.106 schwarze 2258: if (roff_evalnum(r, ln, val, NULL, &iv, 0))
1.77 schwarze 2259: roff_setreg(r, key, iv, sign);
1.1 schwarze 2260:
1.76 schwarze 2261: return(ROFF_IGN);
2262: }
2263:
2264: static enum rofferr
2265: roff_rr(ROFF_ARGS)
2266: {
2267: struct roffreg *reg, **prev;
1.85 schwarze 2268: char *name, *cp;
2269: size_t namesz;
1.76 schwarze 2270:
1.110 schwarze 2271: name = cp = buf->buf + pos;
2272: if (*name == '\0')
1.85 schwarze 2273: return(ROFF_IGN);
2274: namesz = roff_getname(r, &cp, ln, pos);
2275: name[namesz] = '\0';
1.76 schwarze 2276:
2277: prev = &r->regtab;
2278: while (1) {
2279: reg = *prev;
1.110 schwarze 2280: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 2281: break;
2282: prev = ®->next;
2283: }
1.110 schwarze 2284: if (reg != NULL) {
1.76 schwarze 2285: *prev = reg->next;
2286: free(reg->key.p);
2287: free(reg);
2288: }
1.29 schwarze 2289: return(ROFF_IGN);
2290: }
2291:
2292: static enum rofferr
2293: roff_rm(ROFF_ARGS)
2294: {
2295: const char *name;
2296: char *cp;
1.85 schwarze 2297: size_t namesz;
1.29 schwarze 2298:
1.110 schwarze 2299: cp = buf->buf + pos;
2300: while (*cp != '\0') {
1.85 schwarze 2301: name = cp;
1.110 schwarze 2302: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 2303: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110 schwarze 2304: if (name[namesz] == '\\')
1.85 schwarze 2305: break;
1.29 schwarze 2306: }
1.51 schwarze 2307: return(ROFF_IGN);
2308: }
2309:
2310: static enum rofferr
2311: roff_it(ROFF_ARGS)
2312: {
2313: char *cp;
2314: size_t len;
2315: int iv;
2316:
2317: /* Parse the number of lines. */
1.110 schwarze 2318: cp = buf->buf + pos;
1.51 schwarze 2319: len = strcspn(cp, " \t");
2320: cp[len] = '\0';
2321: if ((iv = mandoc_strntoi(cp, len, 10)) <= 0) {
1.95 schwarze 2322: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110 schwarze 2323: ln, ppos, buf->buf + 1);
1.51 schwarze 2324: return(ROFF_IGN);
2325: }
2326: cp += len + 1;
2327:
2328: /* Arm the input line trap. */
2329: roffit_lines = iv;
2330: roffit_macro = mandoc_strdup(cp);
1.2 schwarze 2331: return(ROFF_IGN);
1.47 schwarze 2332: }
2333:
2334: static enum rofferr
2335: roff_Dd(ROFF_ARGS)
2336: {
2337: const char *const *cp;
2338:
1.99 schwarze 2339: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47 schwarze 2340: for (cp = __mdoc_reserved; *cp; cp++)
2341: roff_setstr(r, *cp, NULL, 0);
2342:
1.99 schwarze 2343: if (r->format == 0)
2344: r->format = MPARSE_MDOC;
2345:
1.47 schwarze 2346: return(ROFF_CONT);
2347: }
2348:
2349: static enum rofferr
2350: roff_TH(ROFF_ARGS)
2351: {
2352: const char *const *cp;
2353:
1.99 schwarze 2354: if ((r->options & MPARSE_QUICK) == 0)
1.47 schwarze 2355: for (cp = __man_reserved; *cp; cp++)
2356: roff_setstr(r, *cp, NULL, 0);
2357:
1.99 schwarze 2358: if (r->format == 0)
2359: r->format = MPARSE_MAN;
2360:
1.47 schwarze 2361: return(ROFF_CONT);
1.14 schwarze 2362: }
2363:
2364: static enum rofferr
1.27 schwarze 2365: roff_TE(ROFF_ARGS)
2366: {
2367:
2368: if (NULL == r->tbl)
1.94 schwarze 2369: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2370: ln, ppos, "TE");
1.27 schwarze 2371: else
1.41 schwarze 2372: tbl_end(&r->tbl);
1.27 schwarze 2373:
2374: return(ROFF_IGN);
2375: }
2376:
2377: static enum rofferr
2378: roff_T_(ROFF_ARGS)
2379: {
2380:
2381: if (NULL == r->tbl)
1.94 schwarze 2382: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2383: ln, ppos, "T&");
1.27 schwarze 2384: else
2385: tbl_restart(ppos, ln, r->tbl);
2386:
2387: return(ROFF_IGN);
2388: }
2389:
1.102 schwarze 2390: /*
2391: * Handle in-line equation delimiters.
2392: */
2393: static enum rofferr
1.110 schwarze 2394: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 2395: {
1.105 schwarze 2396: char *cp1, *cp2;
2397: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 2398:
1.102 schwarze 2399: /*
2400: * Outside equations, look for an opening delimiter.
2401: * If we are inside an equation, we already know it is
2402: * in-line, or this function wouldn't have been called;
2403: * so look for a closing delimiter.
2404: */
2405:
1.110 schwarze 2406: cp1 = buf->buf + pos;
1.102 schwarze 2407: cp2 = strchr(cp1, r->eqn == NULL ?
2408: r->last_eqn->odelim : r->last_eqn->cdelim);
2409: if (cp2 == NULL)
2410: return(ROFF_CONT);
2411:
1.105 schwarze 2412: *cp2++ = '\0';
2413: bef_pr = bef_nl = aft_nl = aft_pr = "";
2414:
2415: /* Handle preceding text, protecting whitespace. */
2416:
1.110 schwarze 2417: if (*buf->buf != '\0') {
1.105 schwarze 2418: if (r->eqn == NULL)
2419: bef_pr = "\\&";
2420: bef_nl = "\n";
2421: }
2422:
2423: /*
2424: * Prepare replacing the delimiter with an equation macro
2425: * and drop leading white space from the equation.
2426: */
1.102 schwarze 2427:
1.105 schwarze 2428: if (r->eqn == NULL) {
2429: while (*cp2 == ' ')
2430: cp2++;
2431: mac = ".EQ";
2432: } else
2433: mac = ".EN";
2434:
2435: /* Handle following text, protecting whitespace. */
2436:
2437: if (*cp2 != '\0') {
2438: aft_nl = "\n";
2439: if (r->eqn != NULL)
2440: aft_pr = "\\&";
2441: }
2442:
2443: /* Do the actual replacement. */
2444:
1.110 schwarze 2445: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 2446: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 2447: free(buf->buf);
2448: buf->buf = cp1;
1.102 schwarze 2449:
2450: /* Toggle the in-line state of the eqn subsystem. */
2451:
2452: r->eqn_inline = r->eqn == NULL;
2453: return(ROFF_REPARSE);
1.41 schwarze 2454: }
2455:
1.107 schwarze 2456: static enum rofferr
2457: roff_EQ(ROFF_ARGS)
1.32 schwarze 2458: {
1.41 schwarze 2459: struct eqn_node *e;
1.32 schwarze 2460:
1.110 schwarze 2461: assert(r->eqn == NULL);
1.107 schwarze 2462: e = eqn_alloc(ppos, ln, r->parse);
1.32 schwarze 2463:
1.102 schwarze 2464: if (r->last_eqn) {
1.32 schwarze 2465: r->last_eqn->next = e;
1.102 schwarze 2466: e->delim = r->last_eqn->delim;
2467: e->odelim = r->last_eqn->odelim;
2468: e->cdelim = r->last_eqn->cdelim;
2469: } else
1.32 schwarze 2470: r->first_eqn = r->last_eqn = e;
2471:
2472: r->eqn = r->last_eqn = e;
1.41 schwarze 2473:
1.110 schwarze 2474: if (buf->buf[pos] != '\0')
1.107 schwarze 2475: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 2476: ".EQ %s", buf->buf + pos);
1.41 schwarze 2477:
1.32 schwarze 2478: return(ROFF_IGN);
2479: }
2480:
2481: static enum rofferr
2482: roff_EN(ROFF_ARGS)
2483: {
2484:
1.94 schwarze 2485: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.32 schwarze 2486: return(ROFF_IGN);
2487: }
2488:
2489: static enum rofferr
1.27 schwarze 2490: roff_TS(ROFF_ARGS)
2491: {
1.49 schwarze 2492: struct tbl_node *tbl;
1.27 schwarze 2493:
2494: if (r->tbl) {
1.94 schwarze 2495: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2496: ln, ppos, "TS breaks TS");
1.41 schwarze 2497: tbl_end(&r->tbl);
1.27 schwarze 2498: }
2499:
1.49 schwarze 2500: tbl = tbl_alloc(ppos, ln, r->parse);
1.27 schwarze 2501:
2502: if (r->last_tbl)
1.49 schwarze 2503: r->last_tbl->next = tbl;
1.27 schwarze 2504: else
1.49 schwarze 2505: r->first_tbl = r->last_tbl = tbl;
1.27 schwarze 2506:
1.49 schwarze 2507: r->tbl = r->last_tbl = tbl;
1.27 schwarze 2508: return(ROFF_IGN);
1.123 schwarze 2509: }
2510:
2511: static enum rofferr
2512: roff_brp(ROFF_ARGS)
2513: {
2514:
2515: buf->buf[pos - 1] = '\0';
2516: return(ROFF_CONT);
1.27 schwarze 2517: }
2518:
2519: static enum rofferr
1.48 schwarze 2520: roff_cc(ROFF_ARGS)
2521: {
2522: const char *p;
2523:
1.110 schwarze 2524: p = buf->buf + pos;
1.48 schwarze 2525:
1.110 schwarze 2526: if (*p == '\0' || (r->control = *p++) == '.')
1.48 schwarze 2527: r->control = 0;
2528:
1.110 schwarze 2529: if (*p != '\0')
1.48 schwarze 2530: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
2531:
2532: return(ROFF_IGN);
2533: }
2534:
2535: static enum rofferr
1.42 schwarze 2536: roff_tr(ROFF_ARGS)
2537: {
2538: const char *p, *first, *second;
2539: size_t fsz, ssz;
2540: enum mandoc_esc esc;
2541:
1.110 schwarze 2542: p = buf->buf + pos;
1.42 schwarze 2543:
1.110 schwarze 2544: if (*p == '\0') {
1.42 schwarze 2545: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
2546: return(ROFF_IGN);
2547: }
2548:
1.110 schwarze 2549: while (*p != '\0') {
1.42 schwarze 2550: fsz = ssz = 1;
2551:
2552: first = p++;
1.110 schwarze 2553: if (*first == '\\') {
1.42 schwarze 2554: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2555: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2556: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2557: ln, (int)(p - buf->buf), first);
1.42 schwarze 2558: return(ROFF_IGN);
2559: }
2560: fsz = (size_t)(p - first);
2561: }
2562:
2563: second = p++;
1.110 schwarze 2564: if (*second == '\\') {
1.42 schwarze 2565: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2566: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2567: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2568: ln, (int)(p - buf->buf), second);
1.42 schwarze 2569: return(ROFF_IGN);
2570: }
2571: ssz = (size_t)(p - second);
1.110 schwarze 2572: } else if (*second == '\0') {
1.80 schwarze 2573: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse,
1.110 schwarze 2574: ln, (int)(p - buf->buf), NULL);
1.42 schwarze 2575: second = " ";
2576: p--;
2577: }
2578:
2579: if (fsz > 1) {
1.80 schwarze 2580: roff_setstrn(&r->xmbtab, first, fsz,
2581: second, ssz, 0);
1.42 schwarze 2582: continue;
2583: }
2584:
1.110 schwarze 2585: if (r->xtab == NULL)
1.80 schwarze 2586: r->xtab = mandoc_calloc(128,
2587: sizeof(struct roffstr));
1.42 schwarze 2588:
2589: free(r->xtab[(int)*first].p);
2590: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2591: r->xtab[(int)*first].sz = ssz;
2592: }
2593:
2594: return(ROFF_IGN);
2595: }
2596:
2597: static enum rofferr
1.14 schwarze 2598: roff_so(ROFF_ARGS)
2599: {
1.121 schwarze 2600: char *name, *cp;
1.15 schwarze 2601:
1.110 schwarze 2602: name = buf->buf + pos;
1.97 schwarze 2603: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14 schwarze 2604:
1.22 schwarze 2605: /*
2606: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2607: * opening anything that's not in our cwd or anything beneath
2608: * it. Thus, explicitly disallow traversing up the file-system
2609: * or using absolute paths.
2610: */
2611:
1.110 schwarze 2612: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83 schwarze 2613: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2614: ".so %s", name);
1.121 schwarze 2615: buf->sz = mandoc_asprintf(&cp,
2616: ".sp\nSee the file %s.\n.sp", name) + 1;
2617: free(buf->buf);
2618: buf->buf = cp;
2619: *offs = 0;
2620: return(ROFF_REPARSE);
1.14 schwarze 2621: }
2622:
2623: *offs = pos;
2624: return(ROFF_SO);
1.7 schwarze 2625: }
2626:
1.16 schwarze 2627: static enum rofferr
2628: roff_userdef(ROFF_ARGS)
1.12 schwarze 2629: {
1.16 schwarze 2630: const char *arg[9];
2631: char *cp, *n1, *n2;
1.25 schwarze 2632: int i;
1.12 schwarze 2633:
1.16 schwarze 2634: /*
2635: * Collect pointers to macro argument strings
1.61 schwarze 2636: * and NUL-terminate them.
1.16 schwarze 2637: */
1.110 schwarze 2638: cp = buf->buf + pos;
1.25 schwarze 2639: for (i = 0; i < 9; i++)
1.110 schwarze 2640: arg[i] = *cp == '\0' ? "" :
1.35 schwarze 2641: mandoc_getarg(r->parse, &cp, ln, &pos);
1.16 schwarze 2642:
2643: /*
2644: * Expand macro arguments.
1.12 schwarze 2645: */
1.110 schwarze 2646: buf->sz = 0;
1.16 schwarze 2647: n1 = cp = mandoc_strdup(r->current_string);
1.110 schwarze 2648: while ((cp = strstr(cp, "\\$")) != NULL) {
1.16 schwarze 2649: i = cp[2] - '1';
2650: if (0 > i || 8 < i) {
2651: /* Not an argument invocation. */
2652: cp += 2;
2653: continue;
2654: }
1.82 schwarze 2655: *cp = '\0';
1.110 schwarze 2656: buf->sz = mandoc_asprintf(&n2, "%s%s%s",
1.82 schwarze 2657: n1, arg[i], cp + 3) + 1;
1.16 schwarze 2658: cp = n2 + (cp - n1);
2659: free(n1);
2660: n1 = n2;
1.12 schwarze 2661: }
2662:
1.16 schwarze 2663: /*
2664: * Replace the macro invocation
2665: * by the expanded macro.
2666: */
1.110 schwarze 2667: free(buf->buf);
2668: buf->buf = n1;
2669: if (buf->sz == 0)
2670: buf->sz = strlen(buf->buf) + 1;
1.120 schwarze 2671: *offs = 0;
1.16 schwarze 2672:
1.110 schwarze 2673: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.16 schwarze 2674: ROFF_REPARSE : ROFF_APPEND);
1.12 schwarze 2675: }
1.28 schwarze 2676:
1.85 schwarze 2677: static size_t
1.28 schwarze 2678: roff_getname(struct roff *r, char **cpp, int ln, int pos)
2679: {
2680: char *name, *cp;
1.85 schwarze 2681: size_t namesz;
1.28 schwarze 2682:
2683: name = *cpp;
2684: if ('\0' == *name)
1.85 schwarze 2685: return(0);
1.28 schwarze 2686:
1.85 schwarze 2687: /* Read until end of name and terminate it with NUL. */
2688: for (cp = name; 1; cp++) {
2689: if ('\0' == *cp || ' ' == *cp) {
2690: namesz = cp - name;
2691: break;
2692: }
1.28 schwarze 2693: if ('\\' != *cp)
2694: continue;
1.88 schwarze 2695: namesz = cp - name;
2696: if ('{' == cp[1] || '}' == cp[1])
2697: break;
1.28 schwarze 2698: cp++;
2699: if ('\\' == *cp)
2700: continue;
1.97 schwarze 2701: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
2702: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 2703: mandoc_escape((const char **)&cp, NULL, NULL);
2704: break;
1.28 schwarze 2705: }
2706:
2707: /* Read past spaces. */
2708: while (' ' == *cp)
2709: cp++;
2710:
2711: *cpp = cp;
1.85 schwarze 2712: return(namesz);
1.28 schwarze 2713: }
2714:
1.16 schwarze 2715: /*
2716: * Store *string into the user-defined string called *name.
2717: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 2718: * append == 0: replace mode
2719: * append == 1: single-line append mode
2720: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 2721: */
1.8 schwarze 2722: static void
1.16 schwarze 2723: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 2724: int append)
1.7 schwarze 2725: {
1.42 schwarze 2726:
2727: roff_setstrn(&r->strtab, name, strlen(name), string,
1.80 schwarze 2728: string ? strlen(string) : 0, append);
1.42 schwarze 2729: }
2730:
2731: static void
2732: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 2733: const char *string, size_t stringsz, int append)
1.42 schwarze 2734: {
2735: struct roffkv *n;
2736: char *c;
2737: int i;
2738: size_t oldch, newch;
1.7 schwarze 2739:
1.16 schwarze 2740: /* Search for an existing string with the same name. */
1.42 schwarze 2741: n = *r;
2742:
1.84 schwarze 2743: while (n && (namesz != n->key.sz ||
2744: strncmp(n->key.p, name, namesz)))
1.7 schwarze 2745: n = n->next;
1.8 schwarze 2746:
2747: if (NULL == n) {
1.16 schwarze 2748: /* Create a new string table entry. */
1.42 schwarze 2749: n = mandoc_malloc(sizeof(struct roffkv));
2750: n->key.p = mandoc_strndup(name, namesz);
2751: n->key.sz = namesz;
2752: n->val.p = NULL;
2753: n->val.sz = 0;
2754: n->next = *r;
2755: *r = n;
1.66 schwarze 2756: } else if (0 == append) {
1.42 schwarze 2757: free(n->val.p);
2758: n->val.p = NULL;
2759: n->val.sz = 0;
1.16 schwarze 2760: }
2761:
2762: if (NULL == string)
2763: return;
2764:
2765: /*
2766: * One additional byte for the '\n' in multiline mode,
2767: * and one for the terminating '\0'.
2768: */
1.66 schwarze 2769: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 2770:
2771: if (NULL == n->val.p) {
2772: n->val.p = mandoc_malloc(newch);
2773: *n->val.p = '\0';
1.16 schwarze 2774: oldch = 0;
2775: } else {
1.42 schwarze 2776: oldch = n->val.sz;
2777: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 2778: }
2779:
2780: /* Skip existing content in the destination buffer. */
1.42 schwarze 2781: c = n->val.p + (int)oldch;
1.16 schwarze 2782:
2783: /* Append new content to the destination buffer. */
1.42 schwarze 2784: i = 0;
2785: while (i < (int)stringsz) {
1.16 schwarze 2786: /*
2787: * Rudimentary roff copy mode:
2788: * Handle escaped backslashes.
2789: */
1.42 schwarze 2790: if ('\\' == string[i] && '\\' == string[i + 1])
2791: i++;
2792: *c++ = string[i++];
1.16 schwarze 2793: }
1.8 schwarze 2794:
1.16 schwarze 2795: /* Append terminating bytes. */
1.66 schwarze 2796: if (1 < append)
1.16 schwarze 2797: *c++ = '\n';
1.42 schwarze 2798:
1.16 schwarze 2799: *c = '\0';
1.42 schwarze 2800: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 2801: }
2802:
1.8 schwarze 2803: static const char *
2804: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7 schwarze 2805: {
1.42 schwarze 2806: const struct roffkv *n;
1.64 schwarze 2807: int i;
1.7 schwarze 2808:
1.42 schwarze 2809: for (n = r->strtab; n; n = n->next)
1.80 schwarze 2810: if (0 == strncmp(name, n->key.p, len) &&
2811: '\0' == n->key.p[(int)len])
1.42 schwarze 2812: return(n->val.p);
1.64 schwarze 2813:
2814: for (i = 0; i < PREDEFS_MAX; i++)
2815: if (0 == strncmp(name, predefs[i].name, len) &&
2816: '\0' == predefs[i].name[(int)len])
2817: return(predefs[i].str);
1.8 schwarze 2818:
1.42 schwarze 2819: return(NULL);
1.7 schwarze 2820: }
2821:
1.8 schwarze 2822: static void
1.42 schwarze 2823: roff_freestr(struct roffkv *r)
1.7 schwarze 2824: {
1.42 schwarze 2825: struct roffkv *n, *nn;
1.7 schwarze 2826:
1.42 schwarze 2827: for (n = r; n; n = nn) {
2828: free(n->key.p);
2829: free(n->val.p);
1.7 schwarze 2830: nn = n->next;
2831: free(n);
2832: }
1.27 schwarze 2833: }
2834:
2835: const struct tbl_span *
2836: roff_span(const struct roff *r)
2837: {
1.80 schwarze 2838:
1.27 schwarze 2839: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32 schwarze 2840: }
2841:
2842: const struct eqn *
2843: roff_eqn(const struct roff *r)
2844: {
1.80 schwarze 2845:
1.32 schwarze 2846: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.42 schwarze 2847: }
2848:
2849: /*
2850: * Duplicate an input string, making the appropriate character
2851: * conversations (as stipulated by `tr') along the way.
2852: * Returns a heap-allocated string with all the replacements made.
2853: */
2854: char *
2855: roff_strdup(const struct roff *r, const char *p)
2856: {
2857: const struct roffkv *cp;
2858: char *res;
2859: const char *pp;
2860: size_t ssz, sz;
2861: enum mandoc_esc esc;
2862:
2863: if (NULL == r->xmbtab && NULL == r->xtab)
2864: return(mandoc_strdup(p));
2865: else if ('\0' == *p)
2866: return(mandoc_strdup(""));
2867:
2868: /*
2869: * Step through each character looking for term matches
2870: * (remember that a `tr' can be invoked with an escape, which is
2871: * a glyph but the escape is multi-character).
2872: * We only do this if the character hash has been initialised
2873: * and the string is >0 length.
2874: */
2875:
2876: res = NULL;
2877: ssz = 0;
2878:
2879: while ('\0' != *p) {
2880: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
2881: sz = r->xtab[(int)*p].sz;
2882: res = mandoc_realloc(res, ssz + sz + 1);
2883: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
2884: ssz += sz;
2885: p++;
2886: continue;
2887: } else if ('\\' != *p) {
2888: res = mandoc_realloc(res, ssz + 2);
2889: res[ssz++] = *p++;
2890: continue;
2891: }
2892:
2893: /* Search for term matches. */
2894: for (cp = r->xmbtab; cp; cp = cp->next)
2895: if (0 == strncmp(p, cp->key.p, cp->key.sz))
2896: break;
2897:
2898: if (NULL != cp) {
2899: /*
2900: * A match has been found.
2901: * Append the match to the array and move
2902: * forward by its keysize.
2903: */
1.80 schwarze 2904: res = mandoc_realloc(res,
2905: ssz + cp->val.sz + 1);
1.42 schwarze 2906: memcpy(res + ssz, cp->val.p, cp->val.sz);
2907: ssz += cp->val.sz;
2908: p += (int)cp->key.sz;
2909: continue;
2910: }
2911:
2912: /*
2913: * Handle escapes carefully: we need to copy
2914: * over just the escape itself, or else we might
2915: * do replacements within the escape itself.
2916: * Make sure to pass along the bogus string.
2917: */
2918: pp = p++;
2919: esc = mandoc_escape(&p, NULL, NULL);
2920: if (ESCAPE_ERROR == esc) {
2921: sz = strlen(pp);
2922: res = mandoc_realloc(res, ssz + sz + 1);
2923: memcpy(res + ssz, pp, sz);
2924: break;
2925: }
1.80 schwarze 2926: /*
2927: * We bail out on bad escapes.
1.42 schwarze 2928: * No need to warn: we already did so when
2929: * roff_res() was called.
2930: */
2931: sz = (int)(p - pp);
2932: res = mandoc_realloc(res, ssz + sz + 1);
2933: memcpy(res + ssz, pp, sz);
2934: ssz += sz;
2935: }
2936:
2937: res[(int)ssz] = '\0';
2938: return(res);
1.99 schwarze 2939: }
2940:
2941: int
2942: roff_getformat(const struct roff *r)
2943: {
2944:
2945: return(r->format);
1.48 schwarze 2946: }
2947:
2948: /*
1.80 schwarze 2949: * Find out whether a line is a macro line or not.
1.48 schwarze 2950: * If it is, adjust the current position and return one; if it isn't,
2951: * return zero and don't change the current position.
2952: * If the control character has been set with `.cc', then let that grain
2953: * precedence.
2954: * This is slighly contrary to groff, where using the non-breaking
2955: * control character when `cc' has been invoked will cause the
2956: * non-breaking macro contents to be printed verbatim.
2957: */
2958: int
2959: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
2960: {
2961: int pos;
2962:
2963: pos = *ppos;
2964:
2965: if (0 != r->control && cp[pos] == r->control)
2966: pos++;
2967: else if (0 != r->control)
2968: return(0);
2969: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
2970: pos += 2;
2971: else if ('.' == cp[pos] || '\'' == cp[pos])
2972: pos++;
2973: else
2974: return(0);
2975:
2976: while (' ' == cp[pos] || '\t' == cp[pos])
2977: pos++;
2978:
2979: *ppos = pos;
2980: return(1);
1.1 schwarze 2981: }
![[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