Annotation of mandoc/roff.c, Revision 1.253
1.253 ! schwarze 1: /* $Id: roff.c,v 1.252 2015/01/21 02:16:52 schwarze Exp $ */
1.1 kristaps 2: /*
1.247 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 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.66 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 kristaps 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66 kristaps 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.
1.1 kristaps 17: */
1.66 kristaps 18: #include "config.h"
1.225 schwarze 19:
20: #include <sys/types.h>
1.30 kristaps 21:
1.67 kristaps 22: #include <assert.h>
1.85 kristaps 23: #include <ctype.h>
1.245 schwarze 24: #include <limits.h>
1.178 schwarze 25: #include <stdio.h>
1.1 kristaps 26: #include <stdlib.h>
1.67 kristaps 27: #include <string.h>
1.1 kristaps 28:
1.67 kristaps 29: #include "mandoc.h"
1.201 schwarze 30: #include "mandoc_aux.h"
1.226 schwarze 31: #include "libmandoc.h"
1.109 kristaps 32: #include "libroff.h"
1.33 kristaps 33:
1.141 kristaps 34: /* Maximum number of nested if-else conditionals. */
1.82 kristaps 35: #define RSTACK_MAX 128
36:
1.170 schwarze 37: /* Maximum number of string expansions per line, to break infinite loops. */
38: #define EXPAND_LIMIT 1000
39:
1.67 kristaps 40: enum rofft {
1.251 schwarze 41: ROFF_ab,
1.103 kristaps 42: ROFF_ad,
1.251 schwarze 43: ROFF_af,
44: ROFF_aln,
45: ROFF_als,
1.80 kristaps 46: ROFF_am,
1.251 schwarze 47: ROFF_am1,
1.80 kristaps 48: ROFF_ami,
1.251 schwarze 49: ROFF_ami1,
1.193 schwarze 50: ROFF_as,
1.251 schwarze 51: ROFF_as1,
52: ROFF_asciify,
53: ROFF_backtrace,
54: ROFF_bd,
55: ROFF_bleedat,
56: ROFF_blm,
57: ROFF_box,
58: ROFF_boxa,
59: ROFF_bp,
60: ROFF_BP,
61: /* MAN_br, MDOC_br */
62: ROFF_break,
63: ROFF_breakchar,
64: ROFF_brnl,
65: ROFF_brp,
66: ROFF_brpnl,
67: ROFF_c2,
1.174 kristaps 68: ROFF_cc,
1.194 schwarze 69: ROFF_ce,
1.251 schwarze 70: ROFF_cf,
71: ROFF_cflags,
72: ROFF_ch,
73: ROFF_char,
74: ROFF_chop,
75: ROFF_class,
76: ROFF_close,
77: ROFF_CL,
78: ROFF_color,
79: ROFF_composite,
80: ROFF_continue,
81: ROFF_cp,
82: ROFF_cropat,
83: ROFF_cs,
84: ROFF_cu,
85: ROFF_da,
86: ROFF_dch,
87: ROFF_Dd,
1.80 kristaps 88: ROFF_de,
1.251 schwarze 89: ROFF_de1,
90: ROFF_defcolor,
1.80 kristaps 91: ROFF_dei,
1.251 schwarze 92: ROFF_dei1,
93: ROFF_device,
94: ROFF_devicem,
95: ROFF_di,
96: ROFF_do,
1.83 schwarze 97: ROFF_ds,
1.251 schwarze 98: ROFF_ds1,
99: ROFF_dwh,
100: ROFF_dt,
101: ROFF_ec,
102: ROFF_ecr,
103: ROFF_ecs,
1.82 kristaps 104: ROFF_el,
1.251 schwarze 105: ROFF_em,
106: ROFF_EN,
107: ROFF_eo,
108: ROFF_EP,
109: ROFF_EQ,
110: ROFF_errprint,
111: ROFF_ev,
112: ROFF_evc,
113: ROFF_ex,
114: ROFF_fallback,
1.185 schwarze 115: ROFF_fam,
1.251 schwarze 116: ROFF_fc,
117: ROFF_fchar,
118: ROFF_fcolor,
119: ROFF_fdeferlig,
120: ROFF_feature,
121: /* MAN_fi; ignored in mdoc(7) */
122: ROFF_fkern,
123: ROFF_fl,
124: ROFF_flig,
125: ROFF_fp,
126: ROFF_fps,
127: ROFF_fschar,
128: ROFF_fspacewidth,
129: ROFF_fspecial,
130: /* MAN_ft; ignored in mdoc(7) */
131: ROFF_ftr,
132: ROFF_fzoom,
133: ROFF_gcolor,
134: ROFF_hc,
135: ROFF_hcode,
136: ROFF_hidechar,
137: ROFF_hla,
138: ROFF_hlm,
139: ROFF_hpf,
140: ROFF_hpfa,
141: ROFF_hpfcode,
1.186 schwarze 142: ROFF_hw,
1.103 kristaps 143: ROFF_hy,
1.251 schwarze 144: ROFF_hylang,
145: ROFF_hylen,
146: ROFF_hym,
147: ROFF_hypp,
148: ROFF_hys,
1.82 kristaps 149: ROFF_ie,
1.75 kristaps 150: ROFF_if,
1.76 kristaps 151: ROFF_ig,
1.251 schwarze 152: /* MAN_in; ignored in mdoc(7) */
153: ROFF_index,
1.123 schwarze 154: ROFF_it,
1.251 schwarze 155: ROFF_itc,
156: ROFF_IX,
157: ROFF_kern,
158: ROFF_kernafter,
159: ROFF_kernbefore,
160: ROFF_kernpair,
161: ROFF_lc,
162: ROFF_lc_ctype,
163: ROFF_lds,
164: ROFF_length,
165: ROFF_letadj,
166: ROFF_lf,
167: ROFF_lg,
168: ROFF_lhang,
169: ROFF_linetabs,
170: /* MAN_ll, MDOC_ll */
171: ROFF_lnr,
172: ROFF_lnrf,
173: ROFF_lpfx,
174: ROFF_ls,
175: ROFF_lsm,
176: ROFF_lt,
177: ROFF_mc,
178: ROFF_mediasize,
179: ROFF_minss,
180: ROFF_mk,
181: ROFF_mso,
182: /* MAN_na; ignored in mdoc(7) */
1.103 kristaps 183: ROFF_ne,
1.251 schwarze 184: /* MAN_nf; ignored in mdoc(7) */
1.103 kristaps 185: ROFF_nh,
1.251 schwarze 186: ROFF_nhychar,
187: ROFF_nm,
188: ROFF_nn,
189: ROFF_nop,
1.104 kristaps 190: ROFF_nr,
1.251 schwarze 191: ROFF_nrf,
192: ROFF_nroff,
1.124 schwarze 193: ROFF_ns,
1.251 schwarze 194: ROFF_nx,
195: ROFF_open,
196: ROFF_opena,
197: ROFF_os,
198: ROFF_output,
199: ROFF_padj,
200: ROFF_papersize,
201: ROFF_pc,
202: ROFF_pev,
203: ROFF_pi,
204: ROFF_PI,
1.229 schwarze 205: ROFF_pl,
1.251 schwarze 206: ROFF_pm,
207: ROFF_pn,
208: ROFF_pnr,
209: ROFF_po,
1.124 schwarze 210: ROFF_ps,
1.251 schwarze 211: ROFF_psbb,
212: ROFF_pshape,
213: ROFF_pso,
214: ROFF_ptr,
215: ROFF_pvs,
216: ROFF_rchar,
217: ROFF_rd,
218: ROFF_recursionlimit,
219: ROFF_return,
220: ROFF_rfschar,
221: ROFF_rhang,
222: ROFF_rj,
1.83 schwarze 223: ROFF_rm,
1.251 schwarze 224: ROFF_rn,
225: ROFF_rnn,
1.203 schwarze 226: ROFF_rr,
1.251 schwarze 227: ROFF_rs,
228: ROFF_rt,
229: ROFF_schar,
230: ROFF_sentchar,
231: ROFF_shc,
232: ROFF_shift,
233: ROFF_sizes,
1.105 kristaps 234: ROFF_so,
1.251 schwarze 235: /* MAN_sp, MDOC_sp */
236: ROFF_spacewidth,
237: ROFF_special,
238: ROFF_spreadwarn,
239: ROFF_ss,
240: ROFF_sty,
241: ROFF_substring,
242: ROFF_sv,
243: ROFF_sy,
244: ROFF_T_,
1.124 schwarze 245: ROFF_ta,
1.251 schwarze 246: ROFF_tc,
247: ROFF_TE,
248: ROFF_TH,
249: ROFF_ti,
250: ROFF_tkf,
251: ROFF_tl,
252: ROFF_tm,
253: ROFF_tm1,
254: ROFF_tmc,
1.83 schwarze 255: ROFF_tr,
1.251 schwarze 256: ROFF_track,
257: ROFF_transchar,
258: ROFF_trf,
259: ROFF_trimat,
260: ROFF_trin,
261: ROFF_trnt,
262: ROFF_troff,
1.109 kristaps 263: ROFF_TS,
1.251 schwarze 264: ROFF_uf,
265: ROFF_ul,
266: ROFF_unformat,
267: ROFF_unwatch,
268: ROFF_unwatchn,
269: ROFF_vpt,
270: ROFF_vs,
271: ROFF_warn,
272: ROFF_warnscale,
273: ROFF_watch,
274: ROFF_watchlength,
275: ROFF_watchn,
276: ROFF_wh,
277: ROFF_while,
278: ROFF_write,
279: ROFF_writec,
280: ROFF_writem,
281: ROFF_xflag,
1.76 kristaps 282: ROFF_cblock,
1.106 kristaps 283: ROFF_USERDEF,
1.67 kristaps 284: ROFF_MAX
285: };
286:
1.147 kristaps 287: /*
1.167 kristaps 288: * An incredibly-simple string buffer.
289: */
1.94 kristaps 290: struct roffstr {
1.167 kristaps 291: char *p; /* nil-terminated buffer */
292: size_t sz; /* saved strlen(p) */
1.166 kristaps 293: };
294:
295: /*
1.167 kristaps 296: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 297: */
298: struct roffkv {
299: struct roffstr key;
300: struct roffstr val;
301: struct roffkv *next; /* next in list */
1.94 kristaps 302: };
303:
1.180 schwarze 304: /*
305: * A single number register as part of a singly-linked list.
306: */
307: struct roffreg {
308: struct roffstr key;
1.181 schwarze 309: int val;
1.180 schwarze 310: struct roffreg *next;
311: };
312:
1.67 kristaps 313: struct roff {
1.128 kristaps 314: struct mparse *parse; /* parse point */
1.237 schwarze 315: const struct mchars *mchars; /* character table */
1.67 kristaps 316: struct roffnode *last; /* leaf of stack */
1.223 schwarze 317: int *rstack; /* stack of inverted `ie' values */
1.180 schwarze 318: struct roffreg *regtab; /* number registers */
1.166 kristaps 319: struct roffkv *strtab; /* user-defined strings & macros */
1.167 kristaps 320: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
321: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 322: const char *current_string; /* value of last called user macro */
1.118 kristaps 323: struct tbl_node *first_tbl; /* first table parsed */
324: struct tbl_node *last_tbl; /* last table parsed */
325: struct tbl_node *tbl; /* current table being parsed */
1.125 kristaps 326: struct eqn_node *last_eqn; /* last equation parsed */
327: struct eqn_node *first_eqn; /* first equation parsed */
328: struct eqn_node *eqn; /* current equation being parsed */
1.230 schwarze 329: int eqn_inline; /* current equation is inline */
1.223 schwarze 330: int options; /* parse options */
331: int rstacksz; /* current size limit of rstack */
332: int rstackpos; /* position in rstack */
1.227 schwarze 333: int format; /* current file in mdoc or man format */
1.223 schwarze 334: char control; /* control character */
1.79 kristaps 335: };
336:
1.67 kristaps 337: struct roffnode {
338: enum rofft tok; /* type of node */
339: struct roffnode *parent; /* up one in stack */
340: int line; /* parse line */
341: int col; /* parse col */
1.106 kristaps 342: char *name; /* node name, e.g. macro name */
1.79 kristaps 343: char *end; /* end-rules: custom token */
344: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 345: int rule; /* current evaluation rule */
1.67 kristaps 346: };
347:
348: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.72 kristaps 349: enum rofft tok, /* tok of macro */ \
1.238 schwarze 350: struct buf *buf, /* input buffer */ \
1.67 kristaps 351: int ln, /* parse line */ \
1.75 kristaps 352: int ppos, /* original pos in buffer */ \
353: int pos, /* current pos in buffer */ \
1.74 kristaps 354: int *offs /* reset offset of buffer data */
1.67 kristaps 355:
356: typedef enum rofferr (*roffproc)(ROFF_ARGS);
357:
358: struct roffmac {
359: const char *name; /* macro name */
1.79 kristaps 360: roffproc proc; /* process new macro */
361: roffproc text; /* process as child text of macro */
362: roffproc sub; /* process as child of macro */
363: int flags;
364: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.85 kristaps 365: struct roffmac *next;
1.67 kristaps 366: };
367:
1.141 kristaps 368: struct predef {
369: const char *name; /* predefined input name */
370: const char *str; /* replacement symbol */
371: };
372:
373: #define PREDEF(__name, __str) \
374: { (__name), (__str) },
375:
1.155 kristaps 376: static enum rofft roffhash_find(const char *, size_t);
377: static void roffhash_init(void);
378: static void roffnode_cleanscope(struct roff *);
379: static void roffnode_pop(struct roff *);
380: static void roffnode_push(struct roff *, enum rofft,
381: const char *, int, int);
1.80 kristaps 382: static enum rofferr roff_block(ROFF_ARGS);
383: static enum rofferr roff_block_text(ROFF_ARGS);
384: static enum rofferr roff_block_sub(ROFF_ARGS);
1.251 schwarze 385: static enum rofferr roff_brp(ROFF_ARGS);
1.80 kristaps 386: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 387: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 388: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 389: static enum rofferr roff_cond(ROFF_ARGS);
390: static enum rofferr roff_cond_text(ROFF_ARGS);
391: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 392: static enum rofferr roff_ds(ROFF_ARGS);
1.238 schwarze 393: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.234 kristaps 394: static int roff_evalcond(struct roff *r, int,
395: const char *, int *);
396: static int roff_evalnum(struct roff *, int,
397: const char *, int *, int *, int);
398: static int roff_evalpar(struct roff *, int,
399: const char *, int *, int *);
1.198 schwarze 400: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 401: static void roff_free1(struct roff *);
1.180 schwarze 402: static void roff_freereg(struct roffreg *);
1.167 kristaps 403: static void roff_freestr(struct roffkv *);
1.212 schwarze 404: static size_t roff_getname(struct roff *, char **, int, int);
1.184 schwarze 405: static int roff_getnum(const char *, int *, int *);
406: static int roff_getop(const char *, int *, char *);
1.181 schwarze 407: static int roff_getregn(const struct roff *,
408: const char *, size_t);
1.192 schwarze 409: static int roff_getregro(const char *name);
1.207 schwarze 410: static const char *roff_getstrn(const struct roff *,
1.94 kristaps 411: const char *, size_t);
1.251 schwarze 412: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 413: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 414: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.89 kristaps 415: static enum rofferr roff_nr(ROFF_ARGS);
1.214 schwarze 416: static enum rofft roff_parse(struct roff *, char *, int *,
417: int, int);
1.238 schwarze 418: static enum rofferr roff_parsetext(struct buf *, int, int *);
419: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 420: static enum rofferr roff_rm(ROFF_ARGS);
1.203 schwarze 421: static enum rofferr roff_rr(ROFF_ARGS);
1.94 kristaps 422: static void roff_setstr(struct roff *,
1.106 kristaps 423: const char *, const char *, int);
1.207 schwarze 424: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 425: size_t, const char *, size_t, int);
1.105 kristaps 426: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 427: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 428: static enum rofferr roff_Dd(ROFF_ARGS);
429: static enum rofferr roff_TH(ROFF_ARGS);
1.109 kristaps 430: static enum rofferr roff_TE(ROFF_ARGS);
431: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 432: static enum rofferr roff_EQ(ROFF_ARGS);
433: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 434: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 435: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 436: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 437:
1.155 kristaps 438: /* See roffhash_find() */
1.85 kristaps 439:
440: #define ASCII_HI 126
441: #define ASCII_LO 33
442: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
443:
444: static struct roffmac *hash[HASHWIDTH];
445:
446: static struct roffmac roffs[ROFF_MAX] = {
1.251 schwarze 447: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.103 kristaps 448: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 449: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
450: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
451: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 452: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 453: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.85 kristaps 454: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 455: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.193 schwarze 456: { "as", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 457: { "as1", roff_ds, NULL, NULL, 0, NULL },
458: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
459: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
460: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
461: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
462: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
463: { "box", roff_unsupp, NULL, NULL, 0, NULL },
464: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
465: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
466: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
467: { "break", roff_unsupp, NULL, NULL, 0, NULL },
468: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
469: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
470: { "brp", roff_brp, NULL, NULL, 0, NULL },
471: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
472: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.174 kristaps 473: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.194 schwarze 474: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 475: { "cf", roff_insec, NULL, NULL, 0, NULL },
476: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
477: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
478: { "char", roff_unsupp, NULL, NULL, 0, NULL },
479: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
480: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
481: { "close", roff_insec, NULL, NULL, 0, NULL },
482: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
483: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
484: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
485: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
486: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
487: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
488: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
489: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
490: { "da", roff_unsupp, NULL, NULL, 0, NULL },
491: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
492: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.85 kristaps 493: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 494: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
495: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 496: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 497: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
498: { "device", roff_unsupp, NULL, NULL, 0, NULL },
499: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
500: { "di", roff_unsupp, NULL, NULL, 0, NULL },
501: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.92 schwarze 502: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 503: { "ds1", roff_ds, NULL, NULL, 0, NULL },
504: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
505: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
506: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
507: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
508: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 509: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.251 schwarze 510: { "em", roff_unsupp, NULL, NULL, 0, NULL },
511: { "EN", roff_EN, NULL, NULL, 0, NULL },
512: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
513: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
514: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
515: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
516: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
517: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
518: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
519: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.185 schwarze 520: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 521: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
522: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
523: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
524: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
525: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
526: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
527: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
528: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
529: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
530: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
531: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
532: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
533: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
534: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
535: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
536: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
538: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
543: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
544: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.186 schwarze 545: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 546: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 547: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
551: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 552: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
553: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
554: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 555: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.178 schwarze 556: { "it", roff_it, NULL, NULL, 0, NULL },
1.251 schwarze 557: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
558: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
559: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
560: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
561: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
562: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
563: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
564: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
565: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
566: { "length", roff_unsupp, NULL, NULL, 0, NULL },
567: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
568: { "lf", roff_insec, NULL, NULL, 0, NULL },
569: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
570: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
571: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
572: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
573: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
574: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
575: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
576: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
577: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
578: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
579: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
580: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
581: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
582: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.103 kristaps 583: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
584: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 585: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
586: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
587: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
588: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.104 kristaps 589: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.251 schwarze 590: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
591: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 592: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 593: { "nx", roff_insec, NULL, NULL, 0, NULL },
594: { "open", roff_insec, NULL, NULL, 0, NULL },
595: { "opena", roff_insec, NULL, NULL, 0, NULL },
596: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
597: { "output", roff_unsupp, NULL, NULL, 0, NULL },
598: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
599: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
600: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
601: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
602: { "pi", roff_insec, NULL, NULL, 0, NULL },
603: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.229 schwarze 604: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 605: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
606: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
607: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
608: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 609: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 610: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
611: { "pshape", roff_line_ignore, NULL, NULL, 0, NULL },
612: { "pso", roff_insec, NULL, NULL, 0, NULL },
613: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
615: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
616: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
617: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
618: { "return", roff_unsupp, NULL, NULL, 0, NULL },
619: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
620: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
621: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.122 schwarze 622: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.251 schwarze 623: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
624: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.203 schwarze 625: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.251 schwarze 626: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
627: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
628: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
629: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
630: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
631: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
632: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.105 kristaps 633: { "so", roff_so, NULL, NULL, 0, NULL },
1.251 schwarze 634: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
635: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
636: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
637: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
638: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
639: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
640: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
641: { "sy", roff_insec, NULL, NULL, 0, NULL },
642: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.124 schwarze 643: { "ta", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 644: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
645: { "TE", roff_TE, NULL, NULL, 0, NULL },
646: { "TH", roff_TH, NULL, NULL, 0, NULL },
647: { "ti", roff_line_ignore, NULL, NULL, 0, NULL },
648: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
649: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
650: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
651: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
652: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.164 kristaps 653: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.251 schwarze 654: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
655: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
656: { "trf", roff_insec, NULL, NULL, 0, NULL },
657: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
658: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
659: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
660: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.109 kristaps 661: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.251 schwarze 662: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
663: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
664: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
665: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
666: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
667: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
668: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
670: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
671: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
672: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
673: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
674: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
675: { "while", roff_unsupp, NULL, NULL, 0, NULL },
676: { "write", roff_insec, NULL, NULL, 0, NULL },
677: { "writec", roff_insec, NULL, NULL, 0, NULL },
678: { "writem", roff_insec, NULL, NULL, 0, NULL },
679: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 680: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.106 kristaps 681: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.67 kristaps 682: };
683:
1.200 schwarze 684: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175 schwarze 685: const char *const __mdoc_reserved[] = {
686: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
687: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
688: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
689: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200 schwarze 690: "Dt", "Dv", "Dx", "D1",
691: "Ec", "Ed", "Ef", "Ek", "El", "Em",
692: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175 schwarze 693: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200 schwarze 694: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
695: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175 schwarze 696: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200 schwarze 697: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
698: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
699: "Sc", "Sh", "Sm", "So", "Sq",
1.175 schwarze 700: "Ss", "St", "Sx", "Sy",
701: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200 schwarze 702: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175 schwarze 703: "%P", "%Q", "%R", "%T", "%U", "%V",
704: NULL
705: };
706:
1.200 schwarze 707: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175 schwarze 708: const char *const __man_reserved[] = {
1.200 schwarze 709: "AT", "B", "BI", "BR", "DT",
710: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
711: "LP", "OP", "P", "PD", "PP",
712: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
713: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175 schwarze 714: NULL
715: };
716:
1.141 kristaps 717: /* Array of injected predefined strings. */
718: #define PREDEFS_MAX 38
719: static const struct predef predefs[PREDEFS_MAX] = {
720: #include "predefs.in"
721: };
722:
1.155 kristaps 723: /* See roffhash_find() */
1.85 kristaps 724: #define ROFF_HASH(p) (p[0] - ASCII_LO)
725:
1.178 schwarze 726: static int roffit_lines; /* number of lines to delay */
727: static char *roffit_macro; /* nil-terminated macro line */
728:
1.207 schwarze 729:
1.85 kristaps 730: static void
1.155 kristaps 731: roffhash_init(void)
1.85 kristaps 732: {
733: struct roffmac *n;
734: int buc, i;
735:
1.106 kristaps 736: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.85 kristaps 737: assert(roffs[i].name[0] >= ASCII_LO);
738: assert(roffs[i].name[0] <= ASCII_HI);
739:
740: buc = ROFF_HASH(roffs[i].name);
741:
742: if (NULL != (n = hash[buc])) {
743: for ( ; n->next; n = n->next)
744: /* Do nothing. */ ;
745: n->next = &roffs[i];
746: } else
747: hash[buc] = &roffs[i];
748: }
749: }
750:
1.67 kristaps 751: /*
752: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
753: * the nil-terminated string name could be found.
754: */
755: static enum rofft
1.155 kristaps 756: roffhash_find(const char *p, size_t s)
1.67 kristaps 757: {
1.85 kristaps 758: int buc;
759: struct roffmac *n;
1.67 kristaps 760:
1.85 kristaps 761: /*
762: * libroff has an extremely simple hashtable, for the time
763: * being, which simply keys on the first character, which must
764: * be printable, then walks a chain. It works well enough until
765: * optimised.
766: */
767:
768: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
769: return(ROFF_MAX);
770:
771: buc = ROFF_HASH(p);
772:
773: if (NULL == (n = hash[buc]))
774: return(ROFF_MAX);
775: for ( ; n; n = n->next)
1.106 kristaps 776: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.85 kristaps 777: return((enum rofft)(n - roffs));
1.67 kristaps 778:
779: return(ROFF_MAX);
780: }
781:
782: /*
783: * Pop the current node off of the stack of roff instructions currently
784: * pending.
785: */
786: static void
787: roffnode_pop(struct roff *r)
788: {
789: struct roffnode *p;
790:
1.75 kristaps 791: assert(r->last);
1.207 schwarze 792: p = r->last;
1.82 kristaps 793:
1.75 kristaps 794: r->last = r->last->parent;
1.106 kristaps 795: free(p->name);
796: free(p->end);
1.67 kristaps 797: free(p);
798: }
799:
800: /*
801: * Push a roff node onto the instruction stack. This must later be
802: * removed with roffnode_pop().
803: */
1.98 schwarze 804: static void
1.106 kristaps 805: roffnode_push(struct roff *r, enum rofft tok, const char *name,
806: int line, int col)
1.67 kristaps 807: {
808: struct roffnode *p;
809:
1.98 schwarze 810: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 811: p->tok = tok;
1.106 kristaps 812: if (name)
813: p->name = mandoc_strdup(name);
1.67 kristaps 814: p->parent = r->last;
815: p->line = line;
816: p->col = col;
1.198 schwarze 817: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 818:
819: r->last = p;
820: }
821:
822: static void
823: roff_free1(struct roff *r)
824: {
1.176 schwarze 825: struct tbl_node *tbl;
1.125 kristaps 826: struct eqn_node *e;
1.167 kristaps 827: int i;
1.67 kristaps 828:
1.176 schwarze 829: while (NULL != (tbl = r->first_tbl)) {
830: r->first_tbl = tbl->next;
831: tbl_free(tbl);
1.109 kristaps 832: }
1.113 kristaps 833: r->first_tbl = r->last_tbl = r->tbl = NULL;
834:
1.125 kristaps 835: while (NULL != (e = r->first_eqn)) {
836: r->first_eqn = e->next;
837: eqn_free(e);
838: }
839: r->first_eqn = r->last_eqn = r->eqn = NULL;
840:
1.67 kristaps 841: while (r->last)
842: roffnode_pop(r);
1.109 kristaps 843:
1.223 schwarze 844: free (r->rstack);
845: r->rstack = NULL;
846: r->rstacksz = 0;
847: r->rstackpos = -1;
848:
849: roff_freereg(r->regtab);
850: r->regtab = NULL;
851:
1.167 kristaps 852: roff_freestr(r->strtab);
853: roff_freestr(r->xmbtab);
854: r->strtab = r->xmbtab = NULL;
855:
856: if (r->xtab)
857: for (i = 0; i < 128; i++)
858: free(r->xtab[i].p);
859: free(r->xtab);
860: r->xtab = NULL;
1.67 kristaps 861: }
862:
863: void
864: roff_reset(struct roff *r)
865: {
866:
867: roff_free1(r);
1.227 schwarze 868: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174 kristaps 869: r->control = 0;
1.67 kristaps 870: }
871:
872: void
873: roff_free(struct roff *r)
874: {
875:
876: roff_free1(r);
877: free(r);
878: }
879:
880: struct roff *
1.237 schwarze 881: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.67 kristaps 882: {
883: struct roff *r;
884:
1.98 schwarze 885: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 886: r->parse = parse;
1.237 schwarze 887: r->mchars = mchars;
1.199 schwarze 888: r->options = options;
1.227 schwarze 889: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 890: r->rstackpos = -1;
1.207 schwarze 891:
1.155 kristaps 892: roffhash_init();
1.141 kristaps 893:
1.67 kristaps 894: return(r);
895: }
896:
1.94 kristaps 897: /*
1.206 schwarze 898: * In the current line, expand escape sequences that tend to get
899: * used in numerical expressions and conditional requests.
900: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 901: */
1.172 schwarze 902: static enum rofferr
1.238 schwarze 903: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 904: {
1.208 schwarze 905: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 906: const char *start; /* start of the string to process */
1.209 schwarze 907: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 908: const char *stnam; /* start of the name, after "[(*" */
909: const char *cp; /* end of the name, e.g. before ']' */
910: const char *res; /* the string to be substituted */
1.238 schwarze 911: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 912: size_t maxl; /* expected length of the escape name */
913: size_t naml; /* actual length of the escape name */
1.237 schwarze 914: enum mandoc_esc esc; /* type of the escape sequence */
915: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 916: int expand_count; /* to avoid infinite loops */
1.206 schwarze 917: int npos; /* position in numeric expression */
1.218 schwarze 918: int arg_complete; /* argument not interrupted by eol */
1.206 schwarze 919: char term; /* character terminating the escape */
1.94 kristaps 920:
1.170 schwarze 921: expand_count = 0;
1.238 schwarze 922: start = buf->buf + pos;
1.205 schwarze 923: stesc = strchr(start, '\0') - 1;
924: while (stesc-- > start) {
1.170 schwarze 925:
1.205 schwarze 926: /* Search backwards for the next backslash. */
927:
1.238 schwarze 928: if (*stesc != '\\')
1.205 schwarze 929: continue;
930:
931: /* If it is escaped, skip it. */
932:
933: for (cp = stesc - 1; cp >= start; cp--)
1.238 schwarze 934: if (*cp != '\\')
1.205 schwarze 935: break;
936:
1.238 schwarze 937: if ((stesc - cp) % 2 == 0) {
1.209 schwarze 938: stesc = (char *)cp;
1.205 schwarze 939: continue;
940: }
1.108 schwarze 941:
1.206 schwarze 942: /* Decide whether to expand or to check only. */
1.108 schwarze 943:
1.206 schwarze 944: term = '\0';
1.205 schwarze 945: cp = stesc + 1;
1.181 schwarze 946: switch (*cp) {
1.207 schwarze 947: case '*':
1.181 schwarze 948: res = NULL;
949: break;
1.207 schwarze 950: case 'B':
1.206 schwarze 951: /* FALLTHROUGH */
1.207 schwarze 952: case 'w':
1.206 schwarze 953: term = cp[1];
954: /* FALLTHROUGH */
1.207 schwarze 955: case 'n':
1.181 schwarze 956: res = ubuf;
957: break;
958: default:
1.237 schwarze 959: esc = mandoc_escape(&cp, &stnam, &inaml);
960: if (esc == ESCAPE_ERROR ||
961: (esc == ESCAPE_SPECIAL &&
962: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.219 schwarze 963: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 964: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 965: "%.*s", (int)(cp - stesc), stesc);
1.205 schwarze 966: continue;
1.152 kristaps 967: }
968:
1.205 schwarze 969: if (EXPAND_LIMIT < ++expand_count) {
970: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 971: ln, (int)(stesc - buf->buf), NULL);
1.205 schwarze 972: return(ROFF_IGN);
973: }
1.108 schwarze 974:
975: /*
976: * The third character decides the length
1.181 schwarze 977: * of the name of the string or register.
1.108 schwarze 978: * Save a pointer to the name.
979: */
980:
1.238 schwarze 981: if (term == '\0') {
1.206 schwarze 982: switch (*++cp) {
1.207 schwarze 983: case '\0':
1.206 schwarze 984: maxl = 0;
985: break;
1.207 schwarze 986: case '(':
1.206 schwarze 987: cp++;
988: maxl = 2;
989: break;
1.207 schwarze 990: case '[':
1.206 schwarze 991: cp++;
992: term = ']';
993: maxl = 0;
994: break;
995: default:
996: maxl = 1;
997: break;
998: }
999: } else {
1000: cp += 2;
1.94 kristaps 1001: maxl = 0;
1002: }
1.108 schwarze 1003: stnam = cp;
1.94 kristaps 1004:
1.108 schwarze 1005: /* Advance to the end of the name. */
1.94 kristaps 1006:
1.253 ! schwarze 1007: naml = 0;
1.218 schwarze 1008: arg_complete = 1;
1.253 ! schwarze 1009: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1010: if (*cp == '\0') {
1.219 schwarze 1011: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1012: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1013: arg_complete = 0;
1.206 schwarze 1014: break;
1.153 kristaps 1015: }
1.238 schwarze 1016: if (maxl == 0 && *cp == term) {
1.206 schwarze 1017: cp++;
1.253 ! schwarze 1018: break;
! 1019: }
! 1020: if (*cp++ != '\\' || stesc[1] != 'w') {
! 1021: naml++;
! 1022: continue;
! 1023: }
! 1024: switch (mandoc_escape(&cp, NULL, NULL)) {
! 1025: case ESCAPE_SPECIAL:
! 1026: /* FALLTHROUGH */
! 1027: case ESCAPE_UNICODE:
! 1028: /* FALLTHROUGH */
! 1029: case ESCAPE_NUMBERED:
! 1030: /* FALLTHROUGH */
! 1031: case ESCAPE_OVERSTRIKE:
! 1032: naml++;
! 1033: break;
! 1034: default:
1.94 kristaps 1035: break;
1.206 schwarze 1036: }
1.94 kristaps 1037: }
1038:
1.108 schwarze 1039: /*
1040: * Retrieve the replacement string; if it is
1041: * undefined, resume searching for escapes.
1042: */
1043:
1.206 schwarze 1044: switch (stesc[1]) {
1.207 schwarze 1045: case '*':
1.218 schwarze 1046: if (arg_complete)
1047: res = roff_getstrn(r, stnam, naml);
1.206 schwarze 1048: break;
1.207 schwarze 1049: case 'B':
1.206 schwarze 1050: npos = 0;
1.218 schwarze 1051: ubuf[0] = arg_complete &&
1.234 kristaps 1052: roff_evalnum(r, ln, stnam, &npos, NULL, 0) &&
1.218 schwarze 1053: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1054: ubuf[1] = '\0';
1055: break;
1.207 schwarze 1056: case 'n':
1.218 schwarze 1057: if (arg_complete)
1058: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1059: roff_getregn(r, stnam, naml));
1060: else
1061: ubuf[0] = '\0';
1.206 schwarze 1062: break;
1.207 schwarze 1063: case 'w':
1.218 schwarze 1064: /* use even incomplete args */
1.208 schwarze 1065: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1066: 24 * (int)naml);
1067: break;
1068: }
1.94 kristaps 1069:
1.238 schwarze 1070: if (res == NULL) {
1.219 schwarze 1071: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1072: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1073: "%.*s", (int)naml, stnam);
1.142 kristaps 1074: res = "";
1.246 schwarze 1075: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1076: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1077: ln, (int)(stesc - buf->buf), NULL);
1078: return(ROFF_IGN);
1.94 kristaps 1079: }
1080:
1.108 schwarze 1081: /* Replace the escape sequence by the string. */
1082:
1.209 schwarze 1083: *stesc = '\0';
1.238 schwarze 1084: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1085: buf->buf, res, cp) + 1;
1.94 kristaps 1086:
1.205 schwarze 1087: /* Prepare for the next replacement. */
1.94 kristaps 1088:
1.205 schwarze 1089: start = nbuf + pos;
1.238 schwarze 1090: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1091: free(buf->buf);
1092: buf->buf = nbuf;
1.154 kristaps 1093: }
1.172 schwarze 1094: return(ROFF_CONT);
1.154 kristaps 1095: }
1096:
1097: /*
1.178 schwarze 1098: * Process text streams:
1099: * Convert all breakable hyphens into ASCII_HYPH.
1100: * Decrement and spring input line trap.
1.154 kristaps 1101: */
1102: static enum rofferr
1.238 schwarze 1103: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154 kristaps 1104: {
1105: size_t sz;
1106: const char *start;
1.178 schwarze 1107: char *p;
1108: int isz;
1.154 kristaps 1109: enum mandoc_esc esc;
1110:
1.238 schwarze 1111: start = p = buf->buf + pos;
1.154 kristaps 1112:
1.238 schwarze 1113: while (*p != '\0') {
1.154 kristaps 1114: sz = strcspn(p, "-\\");
1115: p += sz;
1116:
1.238 schwarze 1117: if (*p == '\0')
1.159 kristaps 1118: break;
1119:
1.238 schwarze 1120: if (*p == '\\') {
1.154 kristaps 1121: /* Skip over escapes. */
1122: p++;
1.189 schwarze 1123: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1124: if (esc == ESCAPE_ERROR)
1.154 kristaps 1125: break;
1.155 kristaps 1126: continue;
1.159 kristaps 1127: } else if (p == start) {
1.158 kristaps 1128: p++;
1.155 kristaps 1129: continue;
1.158 kristaps 1130: }
1.155 kristaps 1131:
1.171 schwarze 1132: if (isalpha((unsigned char)p[-1]) &&
1133: isalpha((unsigned char)p[1]))
1.155 kristaps 1134: *p = ASCII_HYPH;
1135: p++;
1.94 kristaps 1136: }
1137:
1.178 schwarze 1138: /* Spring the input line trap. */
1.238 schwarze 1139: if (roffit_lines == 1) {
1140: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1141: free(buf->buf);
1142: buf->buf = p;
1143: buf->sz = isz + 1;
1.178 schwarze 1144: *offs = 0;
1145: free(roffit_macro);
1146: roffit_lines = 0;
1147: return(ROFF_REPARSE);
1.238 schwarze 1148: } else if (roffit_lines > 1)
1.178 schwarze 1149: --roffit_lines;
1.154 kristaps 1150: return(ROFF_CONT);
1.94 kristaps 1151: }
1152:
1.67 kristaps 1153: enum rofferr
1.238 schwarze 1154: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1155: {
1156: enum rofft t;
1.109 kristaps 1157: enum rofferr e;
1.238 schwarze 1158: int pos; /* parse point */
1.243 schwarze 1159: int spos; /* saved parse point for messages */
1.238 schwarze 1160: int ppos; /* original offset in buf->buf */
1161: int ctl; /* macro line (boolean) */
1162:
1163: ppos = pos = *offs;
1.79 kristaps 1164:
1.230 schwarze 1165: /* Handle in-line equation delimiters. */
1166:
1.236 schwarze 1167: if (r->tbl == NULL &&
1168: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1169: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1170: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1171: if (e == ROFF_REPARSE)
1172: return(e);
1173: assert(e == ROFF_CONT);
1174: }
1175:
1176: /* Expand some escape sequences. */
1.94 kristaps 1177:
1.238 schwarze 1178: e = roff_res(r, buf, ln, pos);
1179: if (e == ROFF_IGN)
1.172 schwarze 1180: return(e);
1.238 schwarze 1181: assert(e == ROFF_CONT);
1.94 kristaps 1182:
1.238 schwarze 1183: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1184:
1.94 kristaps 1185: /*
1.79 kristaps 1186: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1187: * text through the macro's filter.
1188: * Equations process all content themselves.
1189: * Tables process almost all content themselves, but we want
1190: * to warn about macros before passing it there.
1.79 kristaps 1191: */
1.74 kristaps 1192:
1.252 schwarze 1193: if (r->last != NULL && ! ctl) {
1.78 kristaps 1194: t = r->last->tok;
1195: assert(roffs[t].text);
1.238 schwarze 1196: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1197: assert(e == ROFF_IGN || e == ROFF_CONT);
1198: if (e != ROFF_CONT)
1.125 kristaps 1199: return(e);
1.182 schwarze 1200: }
1.252 schwarze 1201: if (r->eqn != NULL)
1.238 schwarze 1202: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.252 schwarze 1203: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1204: return(tbl_read(r->tbl, ln, buf->buf, pos));
1205: if ( ! ctl)
1.238 schwarze 1206: return(roff_parsetext(buf, pos, offs));
1.228 schwarze 1207:
1208: /* Skip empty request lines. */
1209:
1.238 schwarze 1210: if (buf->buf[pos] == '"') {
1.228 schwarze 1211: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1212: ln, pos, NULL);
1213: return(ROFF_IGN);
1.238 schwarze 1214: } else if (buf->buf[pos] == '\0')
1.228 schwarze 1215: return(ROFF_IGN);
1.67 kristaps 1216:
1.79 kristaps 1217: /*
1218: * If a scope is open, go to the child handler for that macro,
1219: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1220: * Don't do so if an equation is open.
1.79 kristaps 1221: */
1.78 kristaps 1222:
1.79 kristaps 1223: if (r->last) {
1224: t = r->last->tok;
1225: assert(roffs[t].sub);
1.238 schwarze 1226: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1227: }
1.78 kristaps 1228:
1.243 schwarze 1229: /* No scope is open. This is a new request or macro. */
1230:
1231: spos = pos;
1232: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1233:
1234: /* Tables ignore most macros. */
1235:
1236: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1237: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1238: ln, pos, buf->buf + spos);
1239: return(ROFF_IGN);
1240: }
1241:
1.79 kristaps 1242: /*
1.243 schwarze 1243: * This is neither a roff request nor a user-defined macro.
1244: * Let the standard macro set parsers handle it.
1.79 kristaps 1245: */
1.67 kristaps 1246:
1.243 schwarze 1247: if (t == ROFF_MAX)
1.79 kristaps 1248: return(ROFF_CONT);
1.243 schwarze 1249:
1250: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1251:
1.75 kristaps 1252: assert(roffs[t].proc);
1.238 schwarze 1253: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.74 kristaps 1254: }
1255:
1.117 kristaps 1256: void
1.74 kristaps 1257: roff_endparse(struct roff *r)
1258: {
1259:
1.110 kristaps 1260: if (r->last)
1.221 schwarze 1261: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1262: r->last->line, r->last->col,
1263: roffs[r->last->tok].name);
1.117 kristaps 1264:
1.125 kristaps 1265: if (r->eqn) {
1.221 schwarze 1266: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1267: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151 kristaps 1268: eqn_end(&r->eqn);
1.125 kristaps 1269: }
1270:
1.117 kristaps 1271: if (r->tbl) {
1.221 schwarze 1272: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1273: r->tbl->line, r->tbl->pos, "TS");
1.151 kristaps 1274: tbl_end(&r->tbl);
1.117 kristaps 1275: }
1.67 kristaps 1276: }
1277:
1278: /*
1279: * Parse a roff node's type from the input buffer. This must be in the
1280: * form of ".foo xxx" in the usual way.
1281: */
1282: static enum rofft
1.214 schwarze 1283: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1284: {
1.214 schwarze 1285: char *cp;
1.106 kristaps 1286: const char *mac;
1287: size_t maclen;
1.67 kristaps 1288: enum rofft t;
1289:
1.214 schwarze 1290: cp = buf + *pos;
1291:
1292: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.67 kristaps 1293: return(ROFF_MAX);
1294:
1.214 schwarze 1295: mac = cp;
1296: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1297:
1.106 kristaps 1298: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.155 kristaps 1299: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.67 kristaps 1300:
1.214 schwarze 1301: if (ROFF_MAX != t)
1302: *pos = cp - buf;
1.67 kristaps 1303:
1304: return(t);
1305: }
1306:
1307: static enum rofferr
1.76 kristaps 1308: roff_cblock(ROFF_ARGS)
1.67 kristaps 1309: {
1310:
1.79 kristaps 1311: /*
1312: * A block-close `..' should only be invoked as a child of an
1313: * ignore macro, otherwise raise a warning and just ignore it.
1314: */
1315:
1.238 schwarze 1316: if (r->last == NULL) {
1.221 schwarze 1317: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1318: ln, ppos, "..");
1.76 kristaps 1319: return(ROFF_IGN);
1320: }
1.67 kristaps 1321:
1.81 kristaps 1322: switch (r->last->tok) {
1.207 schwarze 1323: case ROFF_am:
1.220 schwarze 1324: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.81 kristaps 1325: /* FALLTHROUGH */
1.207 schwarze 1326: case ROFF_ami:
1.81 kristaps 1327: /* FALLTHROUGH */
1.207 schwarze 1328: case ROFF_de:
1.108 schwarze 1329: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.81 kristaps 1330: /* FALLTHROUGH */
1.207 schwarze 1331: case ROFF_dei:
1.81 kristaps 1332: /* FALLTHROUGH */
1.207 schwarze 1333: case ROFF_ig:
1.81 kristaps 1334: break;
1335: default:
1.221 schwarze 1336: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1337: ln, ppos, "..");
1.67 kristaps 1338: return(ROFF_IGN);
1.76 kristaps 1339: }
1.67 kristaps 1340:
1.238 schwarze 1341: if (buf->buf[pos] != '\0')
1.217 schwarze 1342: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1343: ".. %s", buf->buf + pos);
1.71 kristaps 1344:
1345: roffnode_pop(r);
1.76 kristaps 1346: roffnode_cleanscope(r);
1347: return(ROFF_IGN);
1.71 kristaps 1348:
1.67 kristaps 1349: }
1350:
1.76 kristaps 1351: static void
1352: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1353: {
1354:
1.76 kristaps 1355: while (r->last) {
1.173 schwarze 1356: if (--r->last->endspan != 0)
1.76 kristaps 1357: break;
1358: roffnode_pop(r);
1359: }
1.67 kristaps 1360: }
1361:
1.195 schwarze 1362: static void
1363: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1364: {
1365:
1.76 kristaps 1366: if (NULL == r->last) {
1.221 schwarze 1367: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1368: ln, ppos, "\\}");
1.195 schwarze 1369: return;
1.76 kristaps 1370: }
1371:
1.82 kristaps 1372: switch (r->last->tok) {
1.207 schwarze 1373: case ROFF_el:
1.82 kristaps 1374: /* FALLTHROUGH */
1.207 schwarze 1375: case ROFF_ie:
1.82 kristaps 1376: /* FALLTHROUGH */
1.207 schwarze 1377: case ROFF_if:
1.82 kristaps 1378: break;
1379: default:
1.221 schwarze 1380: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1381: ln, ppos, "\\}");
1.195 schwarze 1382: return;
1.75 kristaps 1383: }
1384:
1.76 kristaps 1385: if (r->last->endspan > -1) {
1.221 schwarze 1386: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1387: ln, ppos, "\\}");
1.195 schwarze 1388: return;
1.76 kristaps 1389: }
1390:
1.75 kristaps 1391: roffnode_pop(r);
1.76 kristaps 1392: roffnode_cleanscope(r);
1.195 schwarze 1393: return;
1.76 kristaps 1394: }
1395:
1396: static enum rofferr
1.80 kristaps 1397: roff_block(ROFF_ARGS)
1.76 kristaps 1398: {
1.220 schwarze 1399: const char *name;
1400: char *iname, *cp;
1.213 schwarze 1401: size_t namesz;
1.106 kristaps 1402:
1.220 schwarze 1403: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1404:
1.238 schwarze 1405: if (tok == ROFF_de1)
1.220 schwarze 1406: tok = ROFF_de;
1.251 schwarze 1407: else if (tok == ROFF_dei1)
1408: tok = ROFF_dei;
1.238 schwarze 1409: else if (tok == ROFF_am1)
1.220 schwarze 1410: tok = ROFF_am;
1.251 schwarze 1411: else if (tok == ROFF_ami1)
1412: tok = ROFF_ami;
1.220 schwarze 1413:
1414: /* Parse the macro name argument. */
1415:
1.238 schwarze 1416: cp = buf->buf + pos;
1417: if (tok == ROFF_ig) {
1.220 schwarze 1418: iname = NULL;
1419: namesz = 0;
1420: } else {
1421: iname = cp;
1422: namesz = roff_getname(r, &cp, ln, ppos);
1423: iname[namesz] = '\0';
1424: }
1.107 kristaps 1425:
1.220 schwarze 1426: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1427:
1.238 schwarze 1428: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1429: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1430: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1431: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1432: "%.*s", (int)namesz, iname);
1433: namesz = 0;
1434: } else
1435: namesz = strlen(name);
1436: } else
1437: name = iname;
1.107 kristaps 1438:
1.238 schwarze 1439: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1440: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1441: ln, ppos, roffs[tok].name);
1442: return(ROFF_IGN);
1443: }
1.80 kristaps 1444:
1.106 kristaps 1445: roffnode_push(r, tok, name, ln, ppos);
1446:
1447: /*
1448: * At the beginning of a `de' macro, clear the existing string
1449: * with the same name, if there is one. New content will be
1.193 schwarze 1450: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1451: */
1.107 kristaps 1452:
1.238 schwarze 1453: if (tok == ROFF_de || tok == ROFF_dei)
1.213 schwarze 1454: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76 kristaps 1455:
1.238 schwarze 1456: if (*cp == '\0')
1.78 kristaps 1457: return(ROFF_IGN);
1458:
1.220 schwarze 1459: /* Get the custom end marker. */
1.107 kristaps 1460:
1.220 schwarze 1461: iname = cp;
1.213 schwarze 1462: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1463:
1464: /* Resolve the end marker if it is indirect. */
1465:
1.238 schwarze 1466: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1467: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1468: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1469: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1470: "%.*s", (int)namesz, iname);
1471: namesz = 0;
1472: } else
1473: namesz = strlen(name);
1474: } else
1475: name = iname;
1476:
1.213 schwarze 1477: if (namesz)
1478: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1479:
1.238 schwarze 1480: if (*cp != '\0')
1.217 schwarze 1481: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1482: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.74 kristaps 1483:
1.78 kristaps 1484: return(ROFF_IGN);
1485: }
1486:
1487: static enum rofferr
1.80 kristaps 1488: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1489: {
1490: enum rofft t;
1491: int i, j;
1492:
1493: /*
1494: * First check whether a custom macro exists at this level. If
1495: * it does, then check against it. This is some of groff's
1496: * stranger behaviours. If we encountered a custom end-scope
1497: * tag and that tag also happens to be a "real" macro, then we
1498: * need to try interpreting it again as a real macro. If it's
1499: * not, then return ignore. Else continue.
1500: */
1501:
1502: if (r->last->end) {
1.130 kristaps 1503: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1504: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1505: break;
1506:
1.238 schwarze 1507: if (r->last->end[j] == '\0' &&
1508: (buf->buf[i] == '\0' ||
1509: buf->buf[i] == ' ' ||
1510: buf->buf[i] == '\t')) {
1.79 kristaps 1511: roffnode_pop(r);
1512: roffnode_cleanscope(r);
1513:
1.238 schwarze 1514: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1515: i++;
1516:
1517: pos = i;
1.238 schwarze 1518: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1519: ROFF_MAX)
1.79 kristaps 1520: return(ROFF_RERUN);
1521: return(ROFF_IGN);
1522: }
1523: }
1524:
1525: /*
1526: * If we have no custom end-query or lookup failed, then try
1527: * pulling it out of the hashtable.
1528: */
1529:
1.238 schwarze 1530: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1531:
1.238 schwarze 1532: if (t != ROFF_cblock) {
1533: if (tok != ROFF_ig)
1534: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.79 kristaps 1535: return(ROFF_IGN);
1.106 kristaps 1536: }
1.79 kristaps 1537:
1538: assert(roffs[t].proc);
1.238 schwarze 1539: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1540: }
1541:
1542: static enum rofferr
1.80 kristaps 1543: roff_block_text(ROFF_ARGS)
1.78 kristaps 1544: {
1545:
1.238 schwarze 1546: if (tok != ROFF_ig)
1547: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1548:
1.78 kristaps 1549: return(ROFF_IGN);
1550: }
1551:
1552: static enum rofferr
1.82 kristaps 1553: roff_cond_sub(ROFF_ARGS)
1554: {
1555: enum rofft t;
1.139 kristaps 1556: char *ep;
1.198 schwarze 1557: int rr;
1.82 kristaps 1558:
1559: rr = r->last->rule;
1.139 kristaps 1560: roffnode_cleanscope(r);
1.238 schwarze 1561: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82 kristaps 1562:
1.139 kristaps 1563: /*
1.177 schwarze 1564: * Fully handle known macros when they are structurally
1565: * required or when the conditional evaluated to true.
1.87 kristaps 1566: */
1567:
1.238 schwarze 1568: if ((t != ROFF_MAX) &&
1569: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.177 schwarze 1570: assert(roffs[t].proc);
1.238 schwarze 1571: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.177 schwarze 1572: }
1.144 kristaps 1573:
1.196 schwarze 1574: /*
1575: * If `\}' occurs on a macro line without a preceding macro,
1576: * drop the line completely.
1577: */
1578:
1.238 schwarze 1579: ep = buf->buf + pos;
1580: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1581: rr = 0;
1.196 schwarze 1582:
1.177 schwarze 1583: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1584:
1.238 schwarze 1585: while ((ep = strchr(ep, '\\')) != NULL) {
1586: if (*(++ep) == '}') {
1.197 schwarze 1587: *ep = '&';
1.238 schwarze 1588: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1589: }
1.247 schwarze 1590: if (*ep != '\0')
1591: ++ep;
1.177 schwarze 1592: }
1.198 schwarze 1593: return(rr ? ROFF_CONT : ROFF_IGN);
1.82 kristaps 1594: }
1595:
1596: static enum rofferr
1597: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1598: {
1.140 kristaps 1599: char *ep;
1.198 schwarze 1600: int rr;
1.82 kristaps 1601:
1602: rr = r->last->rule;
1.140 kristaps 1603: roffnode_cleanscope(r);
1.82 kristaps 1604:
1.238 schwarze 1605: ep = buf->buf + pos;
1606: while ((ep = strchr(ep, '\\')) != NULL) {
1607: if (*(++ep) == '}') {
1.197 schwarze 1608: *ep = '&';
1.238 schwarze 1609: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1610: }
1.247 schwarze 1611: if (*ep != '\0')
1612: ++ep;
1.78 kristaps 1613: }
1.198 schwarze 1614: return(rr ? ROFF_CONT : ROFF_IGN);
1.74 kristaps 1615: }
1616:
1.204 schwarze 1617: /*
1618: * Parse a single signed integer number. Stop at the first non-digit.
1619: * If there is at least one digit, return success and advance the
1620: * parse point, else return failure and let the parse point unchanged.
1621: * Ignore overflows, treat them just like the C language.
1622: */
1.184 schwarze 1623: static int
1624: roff_getnum(const char *v, int *pos, int *res)
1625: {
1.206 schwarze 1626: int myres, n, p;
1627:
1628: if (NULL == res)
1629: res = &myres;
1.184 schwarze 1630:
1631: p = *pos;
1632: n = v[p] == '-';
1633: if (n)
1634: p++;
1635:
1636: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 1637: *res = 10 * *res + v[p] - '0';
1.184 schwarze 1638: if (p == *pos + n)
1639: return 0;
1640:
1641: if (n)
1642: *res = -*res;
1643:
1644: *pos = p;
1645: return 1;
1646: }
1647:
1.198 schwarze 1648: /*
1649: * Evaluate a string comparison condition.
1650: * The first character is the delimiter.
1651: * Succeed if the string up to its second occurrence
1652: * matches the string up to its third occurence.
1653: * Advance the cursor after the third occurrence
1654: * or lacking that, to the end of the line.
1655: */
1656: static int
1657: roff_evalstrcond(const char *v, int *pos)
1658: {
1659: const char *s1, *s2, *s3;
1660: int match;
1661:
1662: match = 0;
1663: s1 = v + *pos; /* initial delimiter */
1664: s2 = s1 + 1; /* for scanning the first string */
1665: s3 = strchr(s2, *s1); /* for scanning the second string */
1666:
1667: if (NULL == s3) /* found no middle delimiter */
1668: goto out;
1669:
1670: while ('\0' != *++s3) {
1671: if (*s2 != *s3) { /* mismatch */
1672: s3 = strchr(s3, *s1);
1673: break;
1674: }
1675: if (*s3 == *s1) { /* found the final delimiter */
1676: match = 1;
1677: break;
1678: }
1679: s2++;
1680: }
1681:
1682: out:
1683: if (NULL == s3)
1684: s3 = strchr(s2, '\0');
1.242 schwarze 1685: else if (*s3 != '\0')
1.198 schwarze 1686: s3++;
1687: *pos = s3 - v;
1688: return(match);
1689: }
1690:
1.204 schwarze 1691: /*
1692: * Evaluate an optionally negated single character, numerical,
1693: * or string condition.
1694: */
1.198 schwarze 1695: static int
1.234 kristaps 1696: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.88 kristaps 1697: {
1.241 schwarze 1698: int number, savepos, wanttrue;
1.88 kristaps 1699:
1.198 schwarze 1700: if ('!' == v[*pos]) {
1701: wanttrue = 0;
1702: (*pos)++;
1703: } else
1704: wanttrue = 1;
1705:
1.88 kristaps 1706: switch (v[*pos]) {
1.240 schwarze 1707: case '\0':
1708: return(0);
1.207 schwarze 1709: case 'n':
1.198 schwarze 1710: /* FALLTHROUGH */
1.207 schwarze 1711: case 'o':
1.88 kristaps 1712: (*pos)++;
1.198 schwarze 1713: return(wanttrue);
1.207 schwarze 1714: case 'c':
1.198 schwarze 1715: /* FALLTHROUGH */
1.207 schwarze 1716: case 'd':
1.198 schwarze 1717: /* FALLTHROUGH */
1.207 schwarze 1718: case 'e':
1.88 kristaps 1719: /* FALLTHROUGH */
1.207 schwarze 1720: case 'r':
1.88 kristaps 1721: /* FALLTHROUGH */
1.207 schwarze 1722: case 't':
1.239 schwarze 1723: /* FALLTHROUGH */
1724: case 'v':
1.88 kristaps 1725: (*pos)++;
1.198 schwarze 1726: return(!wanttrue);
1.88 kristaps 1727: default:
1728: break;
1729: }
1730:
1.241 schwarze 1731: savepos = *pos;
1.234 kristaps 1732: if (roff_evalnum(r, ln, v, pos, &number, 0))
1.204 schwarze 1733: return((number > 0) == wanttrue);
1.241 schwarze 1734: else if (*pos == savepos)
1735: return(roff_evalstrcond(v, pos) == wanttrue);
1.204 schwarze 1736: else
1.241 schwarze 1737: return (0);
1.88 kristaps 1738: }
1739:
1.74 kristaps 1740: static enum rofferr
1.103 kristaps 1741: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 1742: {
1.123 schwarze 1743:
1.89 kristaps 1744: return(ROFF_IGN);
1745: }
1746:
1.104 kristaps 1747: static enum rofferr
1.251 schwarze 1748: roff_insec(ROFF_ARGS)
1749: {
1750:
1751: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1752: ln, ppos, roffs[tok].name);
1753: return(ROFF_IGN);
1754: }
1755:
1756: static enum rofferr
1757: roff_unsupp(ROFF_ARGS)
1758: {
1759:
1760: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1761: ln, ppos, roffs[tok].name);
1762: return(ROFF_IGN);
1763: }
1764:
1765: static enum rofferr
1.82 kristaps 1766: roff_cond(ROFF_ARGS)
1.74 kristaps 1767: {
1.173 schwarze 1768:
1769: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 1770:
1.207 schwarze 1771: /*
1.134 kristaps 1772: * An `.el' has no conditional body: it will consume the value
1773: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 1774: * defaults to DENY.
1.134 kristaps 1775: *
1776: * If we're not an `el', however, then evaluate the conditional.
1777: */
1.133 kristaps 1778:
1.238 schwarze 1779: r->last->rule = tok == ROFF_el ?
1.207 schwarze 1780: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 1781: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 1782:
1.134 kristaps 1783: /*
1784: * An if-else will put the NEGATION of the current evaluated
1785: * conditional into the stack of rules.
1786: */
1787:
1.238 schwarze 1788: if (tok == ROFF_ie) {
1.223 schwarze 1789: if (r->rstackpos + 1 == r->rstacksz) {
1790: r->rstacksz += 16;
1791: r->rstack = mandoc_reallocarray(r->rstack,
1792: r->rstacksz, sizeof(int));
1.134 kristaps 1793: }
1.198 schwarze 1794: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 1795: }
1.88 kristaps 1796:
1797: /* If the parent has false as its rule, then so do we. */
1798:
1.198 schwarze 1799: if (r->last->parent && !r->last->parent->rule)
1800: r->last->rule = 0;
1.88 kristaps 1801:
1802: /*
1.173 schwarze 1803: * Determine scope.
1804: * If there is nothing on the line after the conditional,
1805: * not even whitespace, use next-line scope.
1.88 kristaps 1806: */
1.74 kristaps 1807:
1.238 schwarze 1808: if (buf->buf[pos] == '\0') {
1.173 schwarze 1809: r->last->endspan = 2;
1810: goto out;
1811: }
1812:
1.238 schwarze 1813: while (buf->buf[pos] == ' ')
1.173 schwarze 1814: pos++;
1815:
1816: /* An opening brace requests multiline scope. */
1.75 kristaps 1817:
1.238 schwarze 1818: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 1819: r->last->endspan = -1;
1820: pos += 2;
1.173 schwarze 1821: goto out;
1.207 schwarze 1822: }
1.74 kristaps 1823:
1.77 kristaps 1824: /*
1.173 schwarze 1825: * Anything else following the conditional causes
1826: * single-line scope. Warn if the scope contains
1827: * nothing but trailing whitespace.
1.77 kristaps 1828: */
1829:
1.238 schwarze 1830: if (buf->buf[pos] == '\0')
1.216 schwarze 1831: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1832: ln, ppos, roffs[tok].name);
1.77 kristaps 1833:
1.173 schwarze 1834: r->last->endspan = 1;
1.74 kristaps 1835:
1.173 schwarze 1836: out:
1.75 kristaps 1837: *offs = pos;
1838: return(ROFF_RERUN);
1.83 schwarze 1839: }
1840:
1841: static enum rofferr
1.92 schwarze 1842: roff_ds(ROFF_ARGS)
1843: {
1.212 schwarze 1844: char *string;
1845: const char *name;
1846: size_t namesz;
1.96 kristaps 1847:
1.251 schwarze 1848: /* Ignore groff compatibility mode for now. */
1849:
1850: if (tok == ROFF_ds1)
1851: tok = ROFF_ds;
1852: else if (tok == ROFF_as1)
1853: tok = ROFF_as;
1854:
1.96 kristaps 1855: /*
1.212 schwarze 1856: * The first word is the name of the string.
1857: * If it is empty or terminated by an escape sequence,
1858: * abort the `ds' request without defining anything.
1.96 kristaps 1859: */
1.92 schwarze 1860:
1.238 schwarze 1861: name = string = buf->buf + pos;
1862: if (*name == '\0')
1.92 schwarze 1863: return(ROFF_IGN);
1864:
1.212 schwarze 1865: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 1866: if (name[namesz] == '\\')
1.212 schwarze 1867: return(ROFF_IGN);
1868:
1869: /* Read past the initial double-quote, if any. */
1.238 schwarze 1870: if (*string == '"')
1.92 schwarze 1871: string++;
1872:
1.96 kristaps 1873: /* The rest is the value. */
1.212 schwarze 1874: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
1875: ROFF_as == tok);
1.92 schwarze 1876: return(ROFF_IGN);
1877: }
1878:
1.204 schwarze 1879: /*
1880: * Parse a single operator, one or two characters long.
1881: * If the operator is recognized, return success and advance the
1882: * parse point, else return failure and let the parse point unchanged.
1883: */
1884: static int
1885: roff_getop(const char *v, int *pos, char *res)
1886: {
1887:
1888: *res = v[*pos];
1889:
1890: switch (*res) {
1.207 schwarze 1891: case '+':
1.204 schwarze 1892: /* FALLTHROUGH */
1.207 schwarze 1893: case '-':
1.204 schwarze 1894: /* FALLTHROUGH */
1.207 schwarze 1895: case '*':
1.204 schwarze 1896: /* FALLTHROUGH */
1.207 schwarze 1897: case '/':
1.204 schwarze 1898: /* FALLTHROUGH */
1.207 schwarze 1899: case '%':
1.204 schwarze 1900: /* FALLTHROUGH */
1.207 schwarze 1901: case '&':
1.204 schwarze 1902: /* FALLTHROUGH */
1.207 schwarze 1903: case ':':
1.204 schwarze 1904: break;
1905: case '<':
1906: switch (v[*pos + 1]) {
1.207 schwarze 1907: case '=':
1.204 schwarze 1908: *res = 'l';
1909: (*pos)++;
1910: break;
1.207 schwarze 1911: case '>':
1.204 schwarze 1912: *res = '!';
1913: (*pos)++;
1914: break;
1.207 schwarze 1915: case '?':
1.204 schwarze 1916: *res = 'i';
1917: (*pos)++;
1918: break;
1919: default:
1920: break;
1921: }
1922: break;
1923: case '>':
1924: switch (v[*pos + 1]) {
1.207 schwarze 1925: case '=':
1.204 schwarze 1926: *res = 'g';
1927: (*pos)++;
1928: break;
1.207 schwarze 1929: case '?':
1.204 schwarze 1930: *res = 'a';
1931: (*pos)++;
1932: break;
1933: default:
1934: break;
1935: }
1936: break;
1937: case '=':
1938: if ('=' == v[*pos + 1])
1939: (*pos)++;
1940: break;
1941: default:
1942: return(0);
1943: }
1944: (*pos)++;
1945:
1946: return(*res);
1947: }
1948:
1949: /*
1950: * Evaluate either a parenthesized numeric expression
1951: * or a single signed integer number.
1952: */
1953: static int
1.235 schwarze 1954: roff_evalpar(struct roff *r, int ln,
1.234 kristaps 1955: const char *v, int *pos, int *res)
1.204 schwarze 1956: {
1957:
1958: if ('(' != v[*pos])
1959: return(roff_getnum(v, pos, res));
1960:
1961: (*pos)++;
1.234 kristaps 1962: if ( ! roff_evalnum(r, ln, v, pos, res, 1))
1.204 schwarze 1963: return(0);
1964:
1.206 schwarze 1965: /*
1966: * Omission of the closing parenthesis
1967: * is an error in validation mode,
1968: * but ignored in evaluation mode.
1969: */
1970:
1.204 schwarze 1971: if (')' == v[*pos])
1972: (*pos)++;
1.206 schwarze 1973: else if (NULL == res)
1974: return(0);
1.204 schwarze 1975:
1976: return(1);
1977: }
1978:
1979: /*
1980: * Evaluate a complete numeric expression.
1981: * Proceed left to right, there is no concept of precedence.
1982: */
1983: static int
1.235 schwarze 1984: roff_evalnum(struct roff *r, int ln, const char *v,
1.234 kristaps 1985: int *pos, int *res, int skipwhite)
1.204 schwarze 1986: {
1987: int mypos, operand2;
1988: char operator;
1989:
1990: if (NULL == pos) {
1991: mypos = 0;
1992: pos = &mypos;
1993: }
1994:
1995: if (skipwhite)
1996: while (isspace((unsigned char)v[*pos]))
1997: (*pos)++;
1998:
1.234 kristaps 1999: if ( ! roff_evalpar(r, ln, v, pos, res))
1.204 schwarze 2000: return(0);
2001:
2002: while (1) {
2003: if (skipwhite)
2004: while (isspace((unsigned char)v[*pos]))
2005: (*pos)++;
2006:
2007: if ( ! roff_getop(v, pos, &operator))
2008: break;
2009:
2010: if (skipwhite)
2011: while (isspace((unsigned char)v[*pos]))
2012: (*pos)++;
2013:
1.234 kristaps 2014: if ( ! roff_evalpar(r, ln, v, pos, &operand2))
1.204 schwarze 2015: return(0);
2016:
2017: if (skipwhite)
2018: while (isspace((unsigned char)v[*pos]))
2019: (*pos)++;
1.206 schwarze 2020:
2021: if (NULL == res)
2022: continue;
1.204 schwarze 2023:
2024: switch (operator) {
1.207 schwarze 2025: case '+':
1.204 schwarze 2026: *res += operand2;
2027: break;
1.207 schwarze 2028: case '-':
1.204 schwarze 2029: *res -= operand2;
2030: break;
1.207 schwarze 2031: case '*':
1.204 schwarze 2032: *res *= operand2;
2033: break;
1.207 schwarze 2034: case '/':
1.244 schwarze 2035: if (operand2 == 0) {
1.235 schwarze 2036: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2037: r->parse, ln, *pos, v);
2038: *res = 0;
2039: break;
2040: }
1.204 schwarze 2041: *res /= operand2;
2042: break;
1.207 schwarze 2043: case '%':
1.244 schwarze 2044: if (operand2 == 0) {
2045: mandoc_msg(MANDOCERR_DIVZERO,
2046: r->parse, ln, *pos, v);
2047: *res = 0;
2048: break;
2049: }
1.204 schwarze 2050: *res %= operand2;
2051: break;
1.207 schwarze 2052: case '<':
1.204 schwarze 2053: *res = *res < operand2;
2054: break;
1.207 schwarze 2055: case '>':
1.204 schwarze 2056: *res = *res > operand2;
2057: break;
1.207 schwarze 2058: case 'l':
1.204 schwarze 2059: *res = *res <= operand2;
2060: break;
1.207 schwarze 2061: case 'g':
1.204 schwarze 2062: *res = *res >= operand2;
2063: break;
1.207 schwarze 2064: case '=':
1.204 schwarze 2065: *res = *res == operand2;
2066: break;
1.207 schwarze 2067: case '!':
1.204 schwarze 2068: *res = *res != operand2;
2069: break;
1.207 schwarze 2070: case '&':
1.204 schwarze 2071: *res = *res && operand2;
2072: break;
1.207 schwarze 2073: case ':':
1.204 schwarze 2074: *res = *res || operand2;
2075: break;
1.207 schwarze 2076: case 'i':
1.204 schwarze 2077: if (operand2 < *res)
2078: *res = operand2;
2079: break;
1.207 schwarze 2080: case 'a':
1.204 schwarze 2081: if (operand2 > *res)
2082: *res = operand2;
2083: break;
2084: default:
2085: abort();
2086: }
2087: }
2088: return(1);
2089: }
2090:
1.180 schwarze 2091: void
1.187 schwarze 2092: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2093: {
1.180 schwarze 2094: struct roffreg *reg;
2095:
2096: /* Search for an existing register with the same name. */
2097: reg = r->regtab;
2098:
2099: while (reg && strcmp(name, reg->key.p))
2100: reg = reg->next;
1.147 kristaps 2101:
1.180 schwarze 2102: if (NULL == reg) {
2103: /* Create a new register. */
2104: reg = mandoc_malloc(sizeof(struct roffreg));
2105: reg->key.p = mandoc_strdup(name);
2106: reg->key.sz = strlen(name);
1.187 schwarze 2107: reg->val = 0;
1.180 schwarze 2108: reg->next = r->regtab;
2109: r->regtab = reg;
2110: }
2111:
1.187 schwarze 2112: if ('+' == sign)
2113: reg->val += val;
2114: else if ('-' == sign)
2115: reg->val -= val;
2116: else
2117: reg->val = val;
1.147 kristaps 2118: }
2119:
1.192 schwarze 2120: /*
2121: * Handle some predefined read-only number registers.
2122: * For now, return -1 if the requested register is not predefined;
2123: * in case a predefined read-only register having the value -1
2124: * were to turn up, another special value would have to be chosen.
2125: */
2126: static int
2127: roff_getregro(const char *name)
2128: {
2129:
2130: switch (*name) {
1.207 schwarze 2131: case 'A': /* ASCII approximation mode is always off. */
1.192 schwarze 2132: return(0);
1.207 schwarze 2133: case 'g': /* Groff compatibility mode is always on. */
1.192 schwarze 2134: return(1);
1.207 schwarze 2135: case 'H': /* Fixed horizontal resolution. */
1.192 schwarze 2136: return (24);
1.207 schwarze 2137: case 'j': /* Always adjust left margin only. */
1.192 schwarze 2138: return(0);
1.207 schwarze 2139: case 'T': /* Some output device is always defined. */
1.192 schwarze 2140: return(1);
1.207 schwarze 2141: case 'V': /* Fixed vertical resolution. */
1.192 schwarze 2142: return (40);
2143: default:
2144: return (-1);
2145: }
2146: }
2147:
1.181 schwarze 2148: int
1.180 schwarze 2149: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2150: {
1.180 schwarze 2151: struct roffreg *reg;
1.192 schwarze 2152: int val;
2153:
2154: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
2155: val = roff_getregro(name + 1);
2156: if (-1 != val)
2157: return (val);
2158: }
1.180 schwarze 2159:
2160: for (reg = r->regtab; reg; reg = reg->next)
2161: if (0 == strcmp(name, reg->key.p))
1.181 schwarze 2162: return(reg->val);
2163:
2164: return(0);
2165: }
2166:
2167: static int
2168: roff_getregn(const struct roff *r, const char *name, size_t len)
2169: {
2170: struct roffreg *reg;
1.192 schwarze 2171: int val;
2172:
2173: if ('.' == name[0] && 2 == len) {
2174: val = roff_getregro(name + 1);
2175: if (-1 != val)
2176: return (val);
2177: }
1.181 schwarze 2178:
2179: for (reg = r->regtab; reg; reg = reg->next)
2180: if (len == reg->key.sz &&
2181: 0 == strncmp(name, reg->key.p, len))
2182: return(reg->val);
1.147 kristaps 2183:
1.180 schwarze 2184: return(0);
1.147 kristaps 2185: }
2186:
1.180 schwarze 2187: static void
2188: roff_freereg(struct roffreg *reg)
1.147 kristaps 2189: {
1.180 schwarze 2190: struct roffreg *old_reg;
1.147 kristaps 2191:
1.180 schwarze 2192: while (NULL != reg) {
2193: free(reg->key.p);
2194: old_reg = reg;
2195: reg = reg->next;
2196: free(old_reg);
2197: }
1.147 kristaps 2198: }
1.92 schwarze 2199:
2200: static enum rofferr
1.89 kristaps 2201: roff_nr(ROFF_ARGS)
1.83 schwarze 2202: {
1.212 schwarze 2203: char *key, *val;
2204: size_t keysz;
1.138 kristaps 2205: int iv;
1.187 schwarze 2206: char sign;
1.89 kristaps 2207:
1.238 schwarze 2208: key = val = buf->buf + pos;
2209: if (*key == '\0')
1.212 schwarze 2210: return(ROFF_IGN);
2211:
2212: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2213: if (key[keysz] == '\\')
1.212 schwarze 2214: return(ROFF_IGN);
2215: key[keysz] = '\0';
1.89 kristaps 2216:
1.187 schwarze 2217: sign = *val;
1.238 schwarze 2218: if (sign == '+' || sign == '-')
1.187 schwarze 2219: val++;
2220:
1.234 kristaps 2221: if (roff_evalnum(r, ln, val, NULL, &iv, 0))
1.204 schwarze 2222: roff_setreg(r, key, iv, sign);
1.109 kristaps 2223:
1.203 schwarze 2224: return(ROFF_IGN);
2225: }
2226:
2227: static enum rofferr
2228: roff_rr(ROFF_ARGS)
2229: {
2230: struct roffreg *reg, **prev;
1.212 schwarze 2231: char *name, *cp;
2232: size_t namesz;
1.203 schwarze 2233:
1.238 schwarze 2234: name = cp = buf->buf + pos;
2235: if (*name == '\0')
1.212 schwarze 2236: return(ROFF_IGN);
2237: namesz = roff_getname(r, &cp, ln, pos);
2238: name[namesz] = '\0';
1.203 schwarze 2239:
2240: prev = &r->regtab;
2241: while (1) {
2242: reg = *prev;
1.238 schwarze 2243: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2244: break;
2245: prev = ®->next;
2246: }
1.238 schwarze 2247: if (reg != NULL) {
1.203 schwarze 2248: *prev = reg->next;
2249: free(reg->key.p);
2250: free(reg);
2251: }
1.122 schwarze 2252: return(ROFF_IGN);
2253: }
2254:
2255: static enum rofferr
2256: roff_rm(ROFF_ARGS)
2257: {
2258: const char *name;
2259: char *cp;
1.212 schwarze 2260: size_t namesz;
1.122 schwarze 2261:
1.238 schwarze 2262: cp = buf->buf + pos;
2263: while (*cp != '\0') {
1.212 schwarze 2264: name = cp;
1.238 schwarze 2265: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2266: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238 schwarze 2267: if (name[namesz] == '\\')
1.212 schwarze 2268: break;
1.122 schwarze 2269: }
1.178 schwarze 2270: return(ROFF_IGN);
2271: }
2272:
2273: static enum rofferr
2274: roff_it(ROFF_ARGS)
2275: {
2276: char *cp;
2277: size_t len;
2278: int iv;
2279:
2280: /* Parse the number of lines. */
1.238 schwarze 2281: cp = buf->buf + pos;
1.178 schwarze 2282: len = strcspn(cp, " \t");
2283: cp[len] = '\0';
2284: if ((iv = mandoc_strntoi(cp, len, 10)) <= 0) {
1.222 schwarze 2285: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2286: ln, ppos, buf->buf + 1);
1.178 schwarze 2287: return(ROFF_IGN);
2288: }
2289: cp += len + 1;
2290:
2291: /* Arm the input line trap. */
2292: roffit_lines = iv;
2293: roffit_macro = mandoc_strdup(cp);
1.109 kristaps 2294: return(ROFF_IGN);
1.175 schwarze 2295: }
2296:
2297: static enum rofferr
2298: roff_Dd(ROFF_ARGS)
2299: {
2300: const char *const *cp;
2301:
1.227 schwarze 2302: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175 schwarze 2303: for (cp = __mdoc_reserved; *cp; cp++)
2304: roff_setstr(r, *cp, NULL, 0);
2305:
1.227 schwarze 2306: if (r->format == 0)
2307: r->format = MPARSE_MDOC;
2308:
1.175 schwarze 2309: return(ROFF_CONT);
2310: }
2311:
2312: static enum rofferr
2313: roff_TH(ROFF_ARGS)
2314: {
2315: const char *const *cp;
2316:
1.227 schwarze 2317: if ((r->options & MPARSE_QUICK) == 0)
1.175 schwarze 2318: for (cp = __man_reserved; *cp; cp++)
2319: roff_setstr(r, *cp, NULL, 0);
2320:
1.227 schwarze 2321: if (r->format == 0)
2322: r->format = MPARSE_MAN;
2323:
1.175 schwarze 2324: return(ROFF_CONT);
1.109 kristaps 2325: }
2326:
2327: static enum rofferr
2328: roff_TE(ROFF_ARGS)
2329: {
2330:
2331: if (NULL == r->tbl)
1.221 schwarze 2332: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2333: ln, ppos, "TE");
1.115 kristaps 2334: else
1.151 kristaps 2335: tbl_end(&r->tbl);
1.109 kristaps 2336:
1.112 kristaps 2337: return(ROFF_IGN);
2338: }
2339:
2340: static enum rofferr
2341: roff_T_(ROFF_ARGS)
2342: {
2343:
2344: if (NULL == r->tbl)
1.221 schwarze 2345: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2346: ln, ppos, "T&");
1.112 kristaps 2347: else
1.116 kristaps 2348: tbl_restart(ppos, ln, r->tbl);
1.112 kristaps 2349:
1.109 kristaps 2350: return(ROFF_IGN);
2351: }
2352:
1.230 schwarze 2353: /*
2354: * Handle in-line equation delimiters.
2355: */
2356: static enum rofferr
1.238 schwarze 2357: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2358: {
1.233 schwarze 2359: char *cp1, *cp2;
2360: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2361:
1.230 schwarze 2362: /*
2363: * Outside equations, look for an opening delimiter.
2364: * If we are inside an equation, we already know it is
2365: * in-line, or this function wouldn't have been called;
2366: * so look for a closing delimiter.
2367: */
2368:
1.238 schwarze 2369: cp1 = buf->buf + pos;
1.230 schwarze 2370: cp2 = strchr(cp1, r->eqn == NULL ?
2371: r->last_eqn->odelim : r->last_eqn->cdelim);
2372: if (cp2 == NULL)
2373: return(ROFF_CONT);
2374:
1.233 schwarze 2375: *cp2++ = '\0';
2376: bef_pr = bef_nl = aft_nl = aft_pr = "";
2377:
2378: /* Handle preceding text, protecting whitespace. */
2379:
1.238 schwarze 2380: if (*buf->buf != '\0') {
1.233 schwarze 2381: if (r->eqn == NULL)
2382: bef_pr = "\\&";
2383: bef_nl = "\n";
2384: }
2385:
2386: /*
2387: * Prepare replacing the delimiter with an equation macro
2388: * and drop leading white space from the equation.
2389: */
1.230 schwarze 2390:
1.233 schwarze 2391: if (r->eqn == NULL) {
2392: while (*cp2 == ' ')
2393: cp2++;
2394: mac = ".EQ";
2395: } else
2396: mac = ".EN";
2397:
2398: /* Handle following text, protecting whitespace. */
2399:
2400: if (*cp2 != '\0') {
2401: aft_nl = "\n";
2402: if (r->eqn != NULL)
2403: aft_pr = "\\&";
2404: }
2405:
2406: /* Do the actual replacement. */
2407:
1.238 schwarze 2408: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2409: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2410: free(buf->buf);
2411: buf->buf = cp1;
1.230 schwarze 2412:
2413: /* Toggle the in-line state of the eqn subsystem. */
2414:
2415: r->eqn_inline = r->eqn == NULL;
2416: return(ROFF_REPARSE);
1.151 kristaps 2417: }
2418:
1.235 schwarze 2419: static enum rofferr
2420: roff_EQ(ROFF_ARGS)
1.125 kristaps 2421: {
1.151 kristaps 2422: struct eqn_node *e;
1.125 kristaps 2423:
1.238 schwarze 2424: assert(r->eqn == NULL);
1.235 schwarze 2425: e = eqn_alloc(ppos, ln, r->parse);
1.125 kristaps 2426:
1.230 schwarze 2427: if (r->last_eqn) {
1.125 kristaps 2428: r->last_eqn->next = e;
1.230 schwarze 2429: e->delim = r->last_eqn->delim;
2430: e->odelim = r->last_eqn->odelim;
2431: e->cdelim = r->last_eqn->cdelim;
2432: } else
1.125 kristaps 2433: r->first_eqn = r->last_eqn = e;
2434:
2435: r->eqn = r->last_eqn = e;
1.151 kristaps 2436:
1.238 schwarze 2437: if (buf->buf[pos] != '\0')
1.235 schwarze 2438: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2439: ".EQ %s", buf->buf + pos);
1.151 kristaps 2440:
1.125 kristaps 2441: return(ROFF_IGN);
2442: }
2443:
2444: static enum rofferr
2445: roff_EN(ROFF_ARGS)
2446: {
2447:
1.221 schwarze 2448: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.125 kristaps 2449: return(ROFF_IGN);
2450: }
2451:
2452: static enum rofferr
1.109 kristaps 2453: roff_TS(ROFF_ARGS)
2454: {
1.176 schwarze 2455: struct tbl_node *tbl;
1.89 kristaps 2456:
1.115 kristaps 2457: if (r->tbl) {
1.221 schwarze 2458: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2459: ln, ppos, "TS breaks TS");
1.151 kristaps 2460: tbl_end(&r->tbl);
1.115 kristaps 2461: }
1.83 schwarze 2462:
1.176 schwarze 2463: tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2464:
2465: if (r->last_tbl)
1.176 schwarze 2466: r->last_tbl->next = tbl;
1.113 kristaps 2467: else
1.176 schwarze 2468: r->first_tbl = r->last_tbl = tbl;
1.113 kristaps 2469:
1.176 schwarze 2470: r->tbl = r->last_tbl = tbl;
1.83 schwarze 2471: return(ROFF_IGN);
1.251 schwarze 2472: }
2473:
2474: static enum rofferr
2475: roff_brp(ROFF_ARGS)
2476: {
2477:
2478: buf->buf[pos - 1] = '\0';
2479: return(ROFF_CONT);
1.92 schwarze 2480: }
2481:
1.105 kristaps 2482: static enum rofferr
1.174 kristaps 2483: roff_cc(ROFF_ARGS)
2484: {
2485: const char *p;
2486:
1.238 schwarze 2487: p = buf->buf + pos;
1.174 kristaps 2488:
1.238 schwarze 2489: if (*p == '\0' || (r->control = *p++) == '.')
1.174 kristaps 2490: r->control = 0;
2491:
1.238 schwarze 2492: if (*p != '\0')
1.174 kristaps 2493: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
2494:
2495: return(ROFF_IGN);
2496: }
2497:
2498: static enum rofferr
1.164 kristaps 2499: roff_tr(ROFF_ARGS)
2500: {
2501: const char *p, *first, *second;
2502: size_t fsz, ssz;
2503: enum mandoc_esc esc;
2504:
1.238 schwarze 2505: p = buf->buf + pos;
1.164 kristaps 2506:
1.238 schwarze 2507: if (*p == '\0') {
1.164 kristaps 2508: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
2509: return(ROFF_IGN);
2510: }
2511:
1.238 schwarze 2512: while (*p != '\0') {
1.164 kristaps 2513: fsz = ssz = 1;
2514:
2515: first = p++;
1.238 schwarze 2516: if (*first == '\\') {
1.164 kristaps 2517: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2518: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2519: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2520: ln, (int)(p - buf->buf), first);
1.164 kristaps 2521: return(ROFF_IGN);
2522: }
2523: fsz = (size_t)(p - first);
2524: }
2525:
2526: second = p++;
1.238 schwarze 2527: if (*second == '\\') {
1.164 kristaps 2528: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2529: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2530: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2531: ln, (int)(p - buf->buf), second);
1.164 kristaps 2532: return(ROFF_IGN);
2533: }
2534: ssz = (size_t)(p - second);
1.238 schwarze 2535: } else if (*second == '\0') {
1.207 schwarze 2536: mandoc_msg(MANDOCERR_ARGCOUNT, r->parse,
1.238 schwarze 2537: ln, (int)(p - buf->buf), NULL);
1.164 kristaps 2538: second = " ";
1.165 kristaps 2539: p--;
1.164 kristaps 2540: }
2541:
1.167 kristaps 2542: if (fsz > 1) {
1.207 schwarze 2543: roff_setstrn(&r->xmbtab, first, fsz,
2544: second, ssz, 0);
1.167 kristaps 2545: continue;
2546: }
2547:
1.238 schwarze 2548: if (r->xtab == NULL)
1.207 schwarze 2549: r->xtab = mandoc_calloc(128,
2550: sizeof(struct roffstr));
1.167 kristaps 2551:
2552: free(r->xtab[(int)*first].p);
2553: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2554: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 2555: }
2556:
2557: return(ROFF_IGN);
2558: }
2559:
2560: static enum rofferr
1.105 kristaps 2561: roff_so(ROFF_ARGS)
2562: {
1.249 schwarze 2563: char *name, *cp;
1.105 kristaps 2564:
1.238 schwarze 2565: name = buf->buf + pos;
1.224 schwarze 2566: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 2567:
2568: /*
2569: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2570: * opening anything that's not in our cwd or anything beneath
2571: * it. Thus, explicitly disallow traversing up the file-system
2572: * or using absolute paths.
2573: */
2574:
1.238 schwarze 2575: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 2576: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2577: ".so %s", name);
1.249 schwarze 2578: buf->sz = mandoc_asprintf(&cp,
2579: ".sp\nSee the file %s.\n.sp", name) + 1;
2580: free(buf->buf);
2581: buf->buf = cp;
2582: *offs = 0;
2583: return(ROFF_REPARSE);
1.105 kristaps 2584: }
2585:
2586: *offs = pos;
2587: return(ROFF_SO);
2588: }
1.92 schwarze 2589:
1.106 kristaps 2590: static enum rofferr
2591: roff_userdef(ROFF_ARGS)
1.99 kristaps 2592: {
1.106 kristaps 2593: const char *arg[9];
2594: char *cp, *n1, *n2;
1.119 schwarze 2595: int i;
1.106 kristaps 2596:
2597: /*
2598: * Collect pointers to macro argument strings
1.188 schwarze 2599: * and NUL-terminate them.
1.106 kristaps 2600: */
1.238 schwarze 2601: cp = buf->buf + pos;
1.119 schwarze 2602: for (i = 0; i < 9; i++)
1.238 schwarze 2603: arg[i] = *cp == '\0' ? "" :
1.136 kristaps 2604: mandoc_getarg(r->parse, &cp, ln, &pos);
1.99 kristaps 2605:
1.106 kristaps 2606: /*
2607: * Expand macro arguments.
1.99 kristaps 2608: */
1.238 schwarze 2609: buf->sz = 0;
1.106 kristaps 2610: n1 = cp = mandoc_strdup(r->current_string);
1.238 schwarze 2611: while ((cp = strstr(cp, "\\$")) != NULL) {
1.106 kristaps 2612: i = cp[2] - '1';
2613: if (0 > i || 8 < i) {
2614: /* Not an argument invocation. */
2615: cp += 2;
2616: continue;
2617: }
1.209 schwarze 2618: *cp = '\0';
1.238 schwarze 2619: buf->sz = mandoc_asprintf(&n2, "%s%s%s",
1.209 schwarze 2620: n1, arg[i], cp + 3) + 1;
1.106 kristaps 2621: cp = n2 + (cp - n1);
2622: free(n1);
2623: n1 = n2;
1.99 kristaps 2624: }
2625:
1.106 kristaps 2626: /*
2627: * Replace the macro invocation
2628: * by the expanded macro.
2629: */
1.238 schwarze 2630: free(buf->buf);
2631: buf->buf = n1;
2632: if (buf->sz == 0)
2633: buf->sz = strlen(buf->buf) + 1;
1.248 schwarze 2634: *offs = 0;
1.106 kristaps 2635:
1.238 schwarze 2636: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.106 kristaps 2637: ROFF_REPARSE : ROFF_APPEND);
1.99 kristaps 2638: }
1.121 schwarze 2639:
1.212 schwarze 2640: static size_t
1.121 schwarze 2641: roff_getname(struct roff *r, char **cpp, int ln, int pos)
2642: {
2643: char *name, *cp;
1.212 schwarze 2644: size_t namesz;
1.121 schwarze 2645:
2646: name = *cpp;
2647: if ('\0' == *name)
1.212 schwarze 2648: return(0);
1.121 schwarze 2649:
1.212 schwarze 2650: /* Read until end of name and terminate it with NUL. */
2651: for (cp = name; 1; cp++) {
2652: if ('\0' == *cp || ' ' == *cp) {
2653: namesz = cp - name;
2654: break;
2655: }
1.121 schwarze 2656: if ('\\' != *cp)
2657: continue;
1.215 schwarze 2658: namesz = cp - name;
2659: if ('{' == cp[1] || '}' == cp[1])
2660: break;
1.121 schwarze 2661: cp++;
2662: if ('\\' == *cp)
2663: continue;
1.224 schwarze 2664: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
2665: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 2666: mandoc_escape((const char **)&cp, NULL, NULL);
2667: break;
1.121 schwarze 2668: }
2669:
2670: /* Read past spaces. */
2671: while (' ' == *cp)
2672: cp++;
2673:
2674: *cpp = cp;
1.212 schwarze 2675: return(namesz);
1.121 schwarze 2676: }
2677:
1.106 kristaps 2678: /*
2679: * Store *string into the user-defined string called *name.
2680: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 2681: * append == 0: replace mode
2682: * append == 1: single-line append mode
2683: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 2684: */
1.94 kristaps 2685: static void
1.106 kristaps 2686: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 2687: int append)
1.92 schwarze 2688: {
1.164 kristaps 2689:
2690: roff_setstrn(&r->strtab, name, strlen(name), string,
1.207 schwarze 2691: string ? strlen(string) : 0, append);
1.164 kristaps 2692: }
2693:
2694: static void
1.166 kristaps 2695: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 2696: const char *string, size_t stringsz, int append)
1.164 kristaps 2697: {
1.166 kristaps 2698: struct roffkv *n;
1.164 kristaps 2699: char *c;
2700: int i;
2701: size_t oldch, newch;
1.92 schwarze 2702:
1.106 kristaps 2703: /* Search for an existing string with the same name. */
1.164 kristaps 2704: n = *r;
2705:
1.211 schwarze 2706: while (n && (namesz != n->key.sz ||
2707: strncmp(n->key.p, name, namesz)))
1.92 schwarze 2708: n = n->next;
1.94 kristaps 2709:
2710: if (NULL == n) {
1.106 kristaps 2711: /* Create a new string table entry. */
1.166 kristaps 2712: n = mandoc_malloc(sizeof(struct roffkv));
2713: n->key.p = mandoc_strndup(name, namesz);
2714: n->key.sz = namesz;
2715: n->val.p = NULL;
2716: n->val.sz = 0;
1.164 kristaps 2717: n->next = *r;
2718: *r = n;
1.193 schwarze 2719: } else if (0 == append) {
1.166 kristaps 2720: free(n->val.p);
2721: n->val.p = NULL;
2722: n->val.sz = 0;
1.106 kristaps 2723: }
2724:
2725: if (NULL == string)
2726: return;
2727:
2728: /*
2729: * One additional byte for the '\n' in multiline mode,
2730: * and one for the terminating '\0'.
2731: */
1.193 schwarze 2732: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 2733:
1.166 kristaps 2734: if (NULL == n->val.p) {
2735: n->val.p = mandoc_malloc(newch);
2736: *n->val.p = '\0';
1.106 kristaps 2737: oldch = 0;
2738: } else {
1.166 kristaps 2739: oldch = n->val.sz;
2740: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 2741: }
2742:
2743: /* Skip existing content in the destination buffer. */
1.166 kristaps 2744: c = n->val.p + (int)oldch;
1.106 kristaps 2745:
2746: /* Append new content to the destination buffer. */
1.164 kristaps 2747: i = 0;
2748: while (i < (int)stringsz) {
1.106 kristaps 2749: /*
2750: * Rudimentary roff copy mode:
2751: * Handle escaped backslashes.
2752: */
1.164 kristaps 2753: if ('\\' == string[i] && '\\' == string[i + 1])
2754: i++;
2755: *c++ = string[i++];
1.106 kristaps 2756: }
1.94 kristaps 2757:
1.106 kristaps 2758: /* Append terminating bytes. */
1.193 schwarze 2759: if (1 < append)
1.106 kristaps 2760: *c++ = '\n';
1.163 kristaps 2761:
1.106 kristaps 2762: *c = '\0';
1.166 kristaps 2763: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 2764: }
2765:
1.94 kristaps 2766: static const char *
2767: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92 schwarze 2768: {
1.166 kristaps 2769: const struct roffkv *n;
1.191 schwarze 2770: int i;
1.92 schwarze 2771:
1.164 kristaps 2772: for (n = r->strtab; n; n = n->next)
1.207 schwarze 2773: if (0 == strncmp(name, n->key.p, len) &&
2774: '\0' == n->key.p[(int)len])
1.166 kristaps 2775: return(n->val.p);
1.191 schwarze 2776:
2777: for (i = 0; i < PREDEFS_MAX; i++)
2778: if (0 == strncmp(name, predefs[i].name, len) &&
2779: '\0' == predefs[i].name[(int)len])
2780: return(predefs[i].str);
1.94 kristaps 2781:
1.157 kristaps 2782: return(NULL);
1.92 schwarze 2783: }
2784:
1.94 kristaps 2785: static void
1.167 kristaps 2786: roff_freestr(struct roffkv *r)
1.92 schwarze 2787: {
1.166 kristaps 2788: struct roffkv *n, *nn;
1.92 schwarze 2789:
1.167 kristaps 2790: for (n = r; n; n = nn) {
1.166 kristaps 2791: free(n->key.p);
2792: free(n->val.p);
1.92 schwarze 2793: nn = n->next;
2794: free(n);
2795: }
1.114 kristaps 2796: }
2797:
2798: const struct tbl_span *
2799: roff_span(const struct roff *r)
2800: {
1.207 schwarze 2801:
1.114 kristaps 2802: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.125 kristaps 2803: }
2804:
2805: const struct eqn *
2806: roff_eqn(const struct roff *r)
2807: {
1.207 schwarze 2808:
1.125 kristaps 2809: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.164 kristaps 2810: }
2811:
2812: /*
2813: * Duplicate an input string, making the appropriate character
2814: * conversations (as stipulated by `tr') along the way.
2815: * Returns a heap-allocated string with all the replacements made.
2816: */
2817: char *
2818: roff_strdup(const struct roff *r, const char *p)
2819: {
1.166 kristaps 2820: const struct roffkv *cp;
1.164 kristaps 2821: char *res;
2822: const char *pp;
2823: size_t ssz, sz;
2824: enum mandoc_esc esc;
2825:
1.167 kristaps 2826: if (NULL == r->xmbtab && NULL == r->xtab)
1.164 kristaps 2827: return(mandoc_strdup(p));
2828: else if ('\0' == *p)
2829: return(mandoc_strdup(""));
2830:
2831: /*
2832: * Step through each character looking for term matches
2833: * (remember that a `tr' can be invoked with an escape, which is
2834: * a glyph but the escape is multi-character).
2835: * We only do this if the character hash has been initialised
2836: * and the string is >0 length.
2837: */
2838:
2839: res = NULL;
2840: ssz = 0;
2841:
2842: while ('\0' != *p) {
1.167 kristaps 2843: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
2844: sz = r->xtab[(int)*p].sz;
2845: res = mandoc_realloc(res, ssz + sz + 1);
2846: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
2847: ssz += sz;
2848: p++;
2849: continue;
2850: } else if ('\\' != *p) {
2851: res = mandoc_realloc(res, ssz + 2);
2852: res[ssz++] = *p++;
2853: continue;
2854: }
2855:
1.164 kristaps 2856: /* Search for term matches. */
1.167 kristaps 2857: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 2858: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 2859: break;
2860:
2861: if (NULL != cp) {
2862: /*
2863: * A match has been found.
2864: * Append the match to the array and move
2865: * forward by its keysize.
2866: */
1.207 schwarze 2867: res = mandoc_realloc(res,
2868: ssz + cp->val.sz + 1);
1.166 kristaps 2869: memcpy(res + ssz, cp->val.p, cp->val.sz);
2870: ssz += cp->val.sz;
2871: p += (int)cp->key.sz;
1.164 kristaps 2872: continue;
2873: }
2874:
1.167 kristaps 2875: /*
2876: * Handle escapes carefully: we need to copy
2877: * over just the escape itself, or else we might
2878: * do replacements within the escape itself.
2879: * Make sure to pass along the bogus string.
2880: */
2881: pp = p++;
2882: esc = mandoc_escape(&p, NULL, NULL);
2883: if (ESCAPE_ERROR == esc) {
2884: sz = strlen(pp);
1.164 kristaps 2885: res = mandoc_realloc(res, ssz + sz + 1);
2886: memcpy(res + ssz, pp, sz);
1.167 kristaps 2887: break;
1.164 kristaps 2888: }
1.207 schwarze 2889: /*
2890: * We bail out on bad escapes.
1.167 kristaps 2891: * No need to warn: we already did so when
2892: * roff_res() was called.
2893: */
2894: sz = (int)(p - pp);
2895: res = mandoc_realloc(res, ssz + sz + 1);
2896: memcpy(res + ssz, pp, sz);
2897: ssz += sz;
1.164 kristaps 2898: }
2899:
2900: res[(int)ssz] = '\0';
2901: return(res);
1.227 schwarze 2902: }
2903:
2904: int
2905: roff_getformat(const struct roff *r)
2906: {
2907:
2908: return(r->format);
1.174 kristaps 2909: }
2910:
2911: /*
1.207 schwarze 2912: * Find out whether a line is a macro line or not.
1.174 kristaps 2913: * If it is, adjust the current position and return one; if it isn't,
2914: * return zero and don't change the current position.
2915: * If the control character has been set with `.cc', then let that grain
2916: * precedence.
2917: * This is slighly contrary to groff, where using the non-breaking
2918: * control character when `cc' has been invoked will cause the
2919: * non-breaking macro contents to be printed verbatim.
2920: */
2921: int
2922: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
2923: {
2924: int pos;
2925:
2926: pos = *ppos;
2927:
2928: if (0 != r->control && cp[pos] == r->control)
2929: pos++;
2930: else if (0 != r->control)
2931: return(0);
2932: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
2933: pos += 2;
2934: else if ('.' == cp[pos] || '\'' == cp[pos])
2935: pos++;
2936: else
2937: return(0);
2938:
2939: while (' ' == cp[pos] || '\t' == cp[pos])
2940: pos++;
2941:
2942: *ppos = pos;
2943: return(1);
1.74 kristaps 2944: }
CVSweb