Annotation of mandoc/roff.c, Revision 1.287
1.287 ! schwarze 1: /* $Id: roff.c,v 1.286 2017/01/10 14:09:07 schwarze Exp $ */
1.1 kristaps 2: /*
1.267 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.287 ! schwarze 4: * Copyright (c) 2010-2015, 2017 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"
1.265 schwarze 31: #include "roff.h"
1.226 schwarze 32: #include "libmandoc.h"
1.266 schwarze 33: #include "roff_int.h"
1.109 kristaps 34: #include "libroff.h"
1.82 kristaps 35:
1.170 schwarze 36: /* Maximum number of string expansions per line, to break infinite loops. */
37: #define EXPAND_LIMIT 1000
38:
1.266 schwarze 39: /* --- data types --------------------------------------------------------- */
40:
1.67 kristaps 41: enum rofft {
1.251 schwarze 42: ROFF_ab,
1.103 kristaps 43: ROFF_ad,
1.251 schwarze 44: ROFF_af,
45: ROFF_aln,
46: ROFF_als,
1.80 kristaps 47: ROFF_am,
1.251 schwarze 48: ROFF_am1,
1.80 kristaps 49: ROFF_ami,
1.251 schwarze 50: ROFF_ami1,
1.193 schwarze 51: ROFF_as,
1.251 schwarze 52: ROFF_as1,
53: ROFF_asciify,
54: ROFF_backtrace,
55: ROFF_bd,
56: ROFF_bleedat,
57: ROFF_blm,
58: ROFF_box,
59: ROFF_boxa,
60: ROFF_bp,
61: ROFF_BP,
62: /* MAN_br, MDOC_br */
63: ROFF_break,
64: ROFF_breakchar,
65: ROFF_brnl,
66: ROFF_brp,
67: ROFF_brpnl,
68: ROFF_c2,
1.174 kristaps 69: ROFF_cc,
1.194 schwarze 70: ROFF_ce,
1.251 schwarze 71: ROFF_cf,
72: ROFF_cflags,
73: ROFF_ch,
74: ROFF_char,
75: ROFF_chop,
76: ROFF_class,
77: ROFF_close,
78: ROFF_CL,
79: ROFF_color,
80: ROFF_composite,
81: ROFF_continue,
82: ROFF_cp,
83: ROFF_cropat,
84: ROFF_cs,
85: ROFF_cu,
86: ROFF_da,
87: ROFF_dch,
88: ROFF_Dd,
1.80 kristaps 89: ROFF_de,
1.251 schwarze 90: ROFF_de1,
91: ROFF_defcolor,
1.80 kristaps 92: ROFF_dei,
1.251 schwarze 93: ROFF_dei1,
94: ROFF_device,
95: ROFF_devicem,
96: ROFF_di,
97: ROFF_do,
1.83 schwarze 98: ROFF_ds,
1.251 schwarze 99: ROFF_ds1,
100: ROFF_dwh,
101: ROFF_dt,
102: ROFF_ec,
103: ROFF_ecr,
104: ROFF_ecs,
1.82 kristaps 105: ROFF_el,
1.251 schwarze 106: ROFF_em,
107: ROFF_EN,
108: ROFF_eo,
109: ROFF_EP,
110: ROFF_EQ,
111: ROFF_errprint,
112: ROFF_ev,
113: ROFF_evc,
114: ROFF_ex,
115: ROFF_fallback,
1.185 schwarze 116: ROFF_fam,
1.251 schwarze 117: ROFF_fc,
118: ROFF_fchar,
119: ROFF_fcolor,
120: ROFF_fdeferlig,
121: ROFF_feature,
122: /* MAN_fi; ignored in mdoc(7) */
123: ROFF_fkern,
124: ROFF_fl,
125: ROFF_flig,
126: ROFF_fp,
127: ROFF_fps,
128: ROFF_fschar,
129: ROFF_fspacewidth,
130: ROFF_fspecial,
131: /* MAN_ft; ignored in mdoc(7) */
132: ROFF_ftr,
133: ROFF_fzoom,
134: ROFF_gcolor,
135: ROFF_hc,
136: ROFF_hcode,
137: ROFF_hidechar,
138: ROFF_hla,
139: ROFF_hlm,
140: ROFF_hpf,
141: ROFF_hpfa,
142: ROFF_hpfcode,
1.186 schwarze 143: ROFF_hw,
1.103 kristaps 144: ROFF_hy,
1.251 schwarze 145: ROFF_hylang,
146: ROFF_hylen,
147: ROFF_hym,
148: ROFF_hypp,
149: ROFF_hys,
1.82 kristaps 150: ROFF_ie,
1.75 kristaps 151: ROFF_if,
1.76 kristaps 152: ROFF_ig,
1.251 schwarze 153: /* MAN_in; ignored in mdoc(7) */
154: ROFF_index,
1.123 schwarze 155: ROFF_it,
1.251 schwarze 156: ROFF_itc,
157: ROFF_IX,
158: ROFF_kern,
159: ROFF_kernafter,
160: ROFF_kernbefore,
161: ROFF_kernpair,
162: ROFF_lc,
163: ROFF_lc_ctype,
164: ROFF_lds,
165: ROFF_length,
166: ROFF_letadj,
167: ROFF_lf,
168: ROFF_lg,
169: ROFF_lhang,
170: ROFF_linetabs,
171: /* MAN_ll, MDOC_ll */
172: ROFF_lnr,
173: ROFF_lnrf,
174: ROFF_lpfx,
175: ROFF_ls,
176: ROFF_lsm,
177: ROFF_lt,
178: ROFF_mc,
179: ROFF_mediasize,
180: ROFF_minss,
181: ROFF_mk,
182: ROFF_mso,
1.256 schwarze 183: ROFF_na,
1.103 kristaps 184: ROFF_ne,
1.251 schwarze 185: /* MAN_nf; ignored in mdoc(7) */
1.103 kristaps 186: ROFF_nh,
1.251 schwarze 187: ROFF_nhychar,
188: ROFF_nm,
189: ROFF_nn,
190: ROFF_nop,
1.104 kristaps 191: ROFF_nr,
1.251 schwarze 192: ROFF_nrf,
193: ROFF_nroff,
1.124 schwarze 194: ROFF_ns,
1.251 schwarze 195: ROFF_nx,
196: ROFF_open,
197: ROFF_opena,
198: ROFF_os,
199: ROFF_output,
200: ROFF_padj,
201: ROFF_papersize,
202: ROFF_pc,
203: ROFF_pev,
204: ROFF_pi,
205: ROFF_PI,
1.229 schwarze 206: ROFF_pl,
1.251 schwarze 207: ROFF_pm,
208: ROFF_pn,
209: ROFF_pnr,
210: ROFF_po,
1.124 schwarze 211: ROFF_ps,
1.251 schwarze 212: ROFF_psbb,
213: ROFF_pshape,
214: ROFF_pso,
215: ROFF_ptr,
216: ROFF_pvs,
217: ROFF_rchar,
218: ROFF_rd,
219: ROFF_recursionlimit,
220: ROFF_return,
221: ROFF_rfschar,
222: ROFF_rhang,
223: ROFF_rj,
1.83 schwarze 224: ROFF_rm,
1.251 schwarze 225: ROFF_rn,
226: ROFF_rnn,
1.203 schwarze 227: ROFF_rr,
1.251 schwarze 228: ROFF_rs,
229: ROFF_rt,
230: ROFF_schar,
231: ROFF_sentchar,
232: ROFF_shc,
233: ROFF_shift,
234: ROFF_sizes,
1.105 kristaps 235: ROFF_so,
1.251 schwarze 236: /* MAN_sp, MDOC_sp */
237: ROFF_spacewidth,
238: ROFF_special,
239: ROFF_spreadwarn,
240: ROFF_ss,
241: ROFF_sty,
242: ROFF_substring,
243: ROFF_sv,
244: ROFF_sy,
245: ROFF_T_,
1.124 schwarze 246: ROFF_ta,
1.251 schwarze 247: ROFF_tc,
248: ROFF_TE,
249: ROFF_TH,
250: ROFF_ti,
251: ROFF_tkf,
252: ROFF_tl,
253: ROFF_tm,
254: ROFF_tm1,
255: ROFF_tmc,
1.83 schwarze 256: ROFF_tr,
1.251 schwarze 257: ROFF_track,
258: ROFF_transchar,
259: ROFF_trf,
260: ROFF_trimat,
261: ROFF_trin,
262: ROFF_trnt,
263: ROFF_troff,
1.109 kristaps 264: ROFF_TS,
1.251 schwarze 265: ROFF_uf,
266: ROFF_ul,
267: ROFF_unformat,
268: ROFF_unwatch,
269: ROFF_unwatchn,
270: ROFF_vpt,
271: ROFF_vs,
272: ROFF_warn,
273: ROFF_warnscale,
274: ROFF_watch,
275: ROFF_watchlength,
276: ROFF_watchn,
277: ROFF_wh,
278: ROFF_while,
279: ROFF_write,
280: ROFF_writec,
281: ROFF_writem,
282: ROFF_xflag,
1.76 kristaps 283: ROFF_cblock,
1.106 kristaps 284: ROFF_USERDEF,
1.67 kristaps 285: ROFF_MAX
286: };
287:
1.147 kristaps 288: /*
1.167 kristaps 289: * An incredibly-simple string buffer.
290: */
1.94 kristaps 291: struct roffstr {
1.167 kristaps 292: char *p; /* nil-terminated buffer */
293: size_t sz; /* saved strlen(p) */
1.166 kristaps 294: };
295:
296: /*
1.167 kristaps 297: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 298: */
299: struct roffkv {
300: struct roffstr key;
301: struct roffstr val;
302: struct roffkv *next; /* next in list */
1.94 kristaps 303: };
304:
1.180 schwarze 305: /*
306: * A single number register as part of a singly-linked list.
307: */
308: struct roffreg {
309: struct roffstr key;
1.181 schwarze 310: int val;
1.180 schwarze 311: struct roffreg *next;
312: };
313:
1.67 kristaps 314: struct roff {
1.128 kristaps 315: struct mparse *parse; /* parse point */
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 */
1.273 schwarze 334: int argc; /* number of args of the last macro */
1.223 schwarze 335: char control; /* control character */
1.79 kristaps 336: };
337:
1.67 kristaps 338: struct roffnode {
339: enum rofft tok; /* type of node */
340: struct roffnode *parent; /* up one in stack */
341: int line; /* parse line */
342: int col; /* parse col */
1.106 kristaps 343: char *name; /* node name, e.g. macro name */
1.79 kristaps 344: char *end; /* end-rules: custom token */
345: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 346: int rule; /* current evaluation rule */
1.67 kristaps 347: };
348:
349: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.72 kristaps 350: enum rofft tok, /* tok of macro */ \
1.238 schwarze 351: struct buf *buf, /* input buffer */ \
1.67 kristaps 352: int ln, /* parse line */ \
1.75 kristaps 353: int ppos, /* original pos in buffer */ \
354: int pos, /* current pos in buffer */ \
1.74 kristaps 355: int *offs /* reset offset of buffer data */
1.67 kristaps 356:
357: typedef enum rofferr (*roffproc)(ROFF_ARGS);
358:
359: struct roffmac {
360: const char *name; /* macro name */
1.79 kristaps 361: roffproc proc; /* process new macro */
362: roffproc text; /* process as child text of macro */
363: roffproc sub; /* process as child of macro */
364: int flags;
365: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.85 kristaps 366: struct roffmac *next;
1.67 kristaps 367: };
368:
1.141 kristaps 369: struct predef {
370: const char *name; /* predefined input name */
371: const char *str; /* replacement symbol */
372: };
373:
374: #define PREDEF(__name, __str) \
375: { (__name), (__str) },
376:
1.266 schwarze 377: /* --- function prototypes ------------------------------------------------ */
378:
1.155 kristaps 379: static enum rofft roffhash_find(const char *, size_t);
380: static void roffhash_init(void);
381: static void roffnode_cleanscope(struct roff *);
382: static void roffnode_pop(struct roff *);
383: static void roffnode_push(struct roff *, enum rofft,
384: const char *, int, int);
1.80 kristaps 385: static enum rofferr roff_block(ROFF_ARGS);
386: static enum rofferr roff_block_text(ROFF_ARGS);
387: static enum rofferr roff_block_sub(ROFF_ARGS);
1.251 schwarze 388: static enum rofferr roff_brp(ROFF_ARGS);
1.80 kristaps 389: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 390: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 391: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 392: static enum rofferr roff_cond(ROFF_ARGS);
393: static enum rofferr roff_cond_text(ROFF_ARGS);
394: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 395: static enum rofferr roff_ds(ROFF_ARGS);
1.238 schwarze 396: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 397: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 398: static int roff_evalnum(struct roff *, int,
399: const char *, int *, int *, int);
400: static int roff_evalpar(struct roff *, int,
1.261 schwarze 401: const char *, int *, int *, int);
1.198 schwarze 402: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 403: static void roff_free1(struct roff *);
1.180 schwarze 404: static void roff_freereg(struct roffreg *);
1.167 kristaps 405: static void roff_freestr(struct roffkv *);
1.212 schwarze 406: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 407: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 408: static int roff_getop(const char *, int *, char *);
1.181 schwarze 409: static int roff_getregn(const struct roff *,
410: const char *, size_t);
1.273 schwarze 411: static int roff_getregro(const struct roff *,
412: const char *name);
1.207 schwarze 413: static const char *roff_getstrn(const struct roff *,
1.94 kristaps 414: const char *, size_t);
1.271 schwarze 415: static int roff_hasregn(const struct roff *,
416: const char *, size_t);
1.251 schwarze 417: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 418: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 419: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 schwarze 420: static void roff_man_alloc1(struct roff_man *);
421: static void roff_man_free1(struct roff_man *);
1.89 kristaps 422: static enum rofferr roff_nr(ROFF_ARGS);
1.214 schwarze 423: static enum rofft roff_parse(struct roff *, char *, int *,
424: int, int);
1.238 schwarze 425: static enum rofferr roff_parsetext(struct buf *, int, int *);
426: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 427: static enum rofferr roff_rm(ROFF_ARGS);
1.203 schwarze 428: static enum rofferr roff_rr(ROFF_ARGS);
1.94 kristaps 429: static void roff_setstr(struct roff *,
1.106 kristaps 430: const char *, const char *, int);
1.207 schwarze 431: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 432: size_t, const char *, size_t, int);
1.105 kristaps 433: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 434: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 435: static enum rofferr roff_Dd(ROFF_ARGS);
436: static enum rofferr roff_TH(ROFF_ARGS);
1.109 kristaps 437: static enum rofferr roff_TE(ROFF_ARGS);
438: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 439: static enum rofferr roff_EQ(ROFF_ARGS);
440: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 441: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 442: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 443: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 444:
1.266 schwarze 445: /* --- constant data ------------------------------------------------------ */
446:
1.155 kristaps 447: /* See roffhash_find() */
1.85 kristaps 448:
449: #define ASCII_HI 126
450: #define ASCII_LO 33
451: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
452:
1.261 schwarze 453: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
454: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
455:
1.85 kristaps 456: static struct roffmac *hash[HASHWIDTH];
457:
458: static struct roffmac roffs[ROFF_MAX] = {
1.251 schwarze 459: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.103 kristaps 460: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 461: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
462: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
463: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 464: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 465: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.85 kristaps 466: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 467: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.193 schwarze 468: { "as", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 469: { "as1", roff_ds, NULL, NULL, 0, NULL },
470: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
471: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
472: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
473: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
474: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
475: { "box", roff_unsupp, NULL, NULL, 0, NULL },
476: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
477: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
478: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
479: { "break", roff_unsupp, NULL, NULL, 0, NULL },
480: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
481: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
482: { "brp", roff_brp, NULL, NULL, 0, NULL },
483: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
484: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.174 kristaps 485: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.194 schwarze 486: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 487: { "cf", roff_insec, NULL, NULL, 0, NULL },
488: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
489: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
490: { "char", roff_unsupp, NULL, NULL, 0, NULL },
491: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
492: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
493: { "close", roff_insec, NULL, NULL, 0, NULL },
494: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
495: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
496: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
497: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
498: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
499: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
500: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
501: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
502: { "da", roff_unsupp, NULL, NULL, 0, NULL },
503: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
504: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.85 kristaps 505: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 506: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
507: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 508: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 509: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
510: { "device", roff_unsupp, NULL, NULL, 0, NULL },
511: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
512: { "di", roff_unsupp, NULL, NULL, 0, NULL },
513: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.92 schwarze 514: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 515: { "ds1", roff_ds, NULL, NULL, 0, NULL },
516: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
517: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
518: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
519: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
520: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 521: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.251 schwarze 522: { "em", roff_unsupp, NULL, NULL, 0, NULL },
523: { "EN", roff_EN, NULL, NULL, 0, NULL },
524: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
525: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
526: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
527: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
528: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
529: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
530: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
531: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.185 schwarze 532: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 533: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
534: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
535: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
536: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
538: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
543: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
544: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
545: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
551: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
552: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
553: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
554: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
555: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
556: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.186 schwarze 557: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 558: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 559: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
560: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
561: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
562: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
563: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 564: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
565: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
566: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 567: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.178 schwarze 568: { "it", roff_it, NULL, NULL, 0, NULL },
1.251 schwarze 569: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
570: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
571: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
572: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
573: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
574: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
575: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
576: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
577: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
578: { "length", roff_unsupp, NULL, NULL, 0, NULL },
579: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
580: { "lf", roff_insec, NULL, NULL, 0, NULL },
581: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
582: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
583: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
584: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
585: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
586: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
587: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
588: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
589: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
590: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
591: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
592: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
593: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
594: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.256 schwarze 595: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 596: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
597: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 598: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
599: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
600: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
601: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.104 kristaps 602: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.251 schwarze 603: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
604: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 605: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 606: { "nx", roff_insec, NULL, NULL, 0, NULL },
607: { "open", roff_insec, NULL, NULL, 0, NULL },
608: { "opena", roff_insec, NULL, NULL, 0, NULL },
609: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
610: { "output", roff_unsupp, NULL, NULL, 0, NULL },
611: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
612: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
615: { "pi", roff_insec, NULL, NULL, 0, NULL },
616: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.229 schwarze 617: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 618: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
619: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
620: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
621: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 622: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 623: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.255 schwarze 624: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 625: { "pso", roff_insec, NULL, NULL, 0, NULL },
626: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
627: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
628: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
629: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
630: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
631: { "return", roff_unsupp, NULL, NULL, 0, NULL },
632: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
633: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
634: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.122 schwarze 635: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.251 schwarze 636: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
637: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.203 schwarze 638: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.251 schwarze 639: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
640: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
641: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
642: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
643: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
644: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
645: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.105 kristaps 646: { "so", roff_so, NULL, NULL, 0, NULL },
1.251 schwarze 647: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
648: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
649: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
650: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
651: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
652: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
653: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
654: { "sy", roff_insec, NULL, NULL, 0, NULL },
655: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.255 schwarze 656: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 657: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
658: { "TE", roff_TE, NULL, NULL, 0, NULL },
659: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.255 schwarze 660: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 661: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
662: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
663: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
664: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
665: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.164 kristaps 666: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.251 schwarze 667: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
668: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "trf", roff_insec, NULL, NULL, 0, NULL },
670: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
671: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
672: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
673: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.109 kristaps 674: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.251 schwarze 675: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
676: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
677: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
678: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
679: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
680: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
681: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
682: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
683: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
684: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
685: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
686: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
687: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
688: { "while", roff_unsupp, NULL, NULL, 0, NULL },
689: { "write", roff_insec, NULL, NULL, 0, NULL },
690: { "writec", roff_insec, NULL, NULL, 0, NULL },
691: { "writem", roff_insec, NULL, NULL, 0, NULL },
692: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 693: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.106 kristaps 694: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.67 kristaps 695: };
696:
1.200 schwarze 697: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175 schwarze 698: const char *const __mdoc_reserved[] = {
699: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
700: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
701: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
702: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200 schwarze 703: "Dt", "Dv", "Dx", "D1",
704: "Ec", "Ed", "Ef", "Ek", "El", "Em",
705: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175 schwarze 706: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200 schwarze 707: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
708: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175 schwarze 709: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200 schwarze 710: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
711: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
712: "Sc", "Sh", "Sm", "So", "Sq",
1.175 schwarze 713: "Ss", "St", "Sx", "Sy",
714: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200 schwarze 715: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175 schwarze 716: "%P", "%Q", "%R", "%T", "%U", "%V",
717: NULL
718: };
719:
1.200 schwarze 720: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175 schwarze 721: const char *const __man_reserved[] = {
1.200 schwarze 722: "AT", "B", "BI", "BR", "DT",
723: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
724: "LP", "OP", "P", "PD", "PP",
725: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
726: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175 schwarze 727: NULL
728: };
729:
1.141 kristaps 730: /* Array of injected predefined strings. */
731: #define PREDEFS_MAX 38
732: static const struct predef predefs[PREDEFS_MAX] = {
733: #include "predefs.in"
734: };
735:
1.155 kristaps 736: /* See roffhash_find() */
1.85 kristaps 737: #define ROFF_HASH(p) (p[0] - ASCII_LO)
738:
1.178 schwarze 739: static int roffit_lines; /* number of lines to delay */
740: static char *roffit_macro; /* nil-terminated macro line */
741:
1.207 schwarze 742:
1.266 schwarze 743: /* --- request table ------------------------------------------------------ */
744:
1.85 kristaps 745: static void
1.155 kristaps 746: roffhash_init(void)
1.85 kristaps 747: {
748: struct roffmac *n;
749: int buc, i;
750:
1.106 kristaps 751: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.85 kristaps 752: assert(roffs[i].name[0] >= ASCII_LO);
753: assert(roffs[i].name[0] <= ASCII_HI);
754:
755: buc = ROFF_HASH(roffs[i].name);
756:
757: if (NULL != (n = hash[buc])) {
758: for ( ; n->next; n = n->next)
759: /* Do nothing. */ ;
760: n->next = &roffs[i];
761: } else
762: hash[buc] = &roffs[i];
763: }
764: }
765:
1.67 kristaps 766: /*
767: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
768: * the nil-terminated string name could be found.
769: */
770: static enum rofft
1.155 kristaps 771: roffhash_find(const char *p, size_t s)
1.67 kristaps 772: {
1.85 kristaps 773: int buc;
774: struct roffmac *n;
1.67 kristaps 775:
1.85 kristaps 776: /*
777: * libroff has an extremely simple hashtable, for the time
778: * being, which simply keys on the first character, which must
779: * be printable, then walks a chain. It works well enough until
780: * optimised.
781: */
782:
783: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
1.277 schwarze 784: return ROFF_MAX;
1.85 kristaps 785:
786: buc = ROFF_HASH(p);
787:
788: if (NULL == (n = hash[buc]))
1.277 schwarze 789: return ROFF_MAX;
1.85 kristaps 790: for ( ; n; n = n->next)
1.106 kristaps 791: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.277 schwarze 792: return (enum rofft)(n - roffs);
1.67 kristaps 793:
1.277 schwarze 794: return ROFF_MAX;
1.67 kristaps 795: }
796:
1.266 schwarze 797: /* --- stack of request blocks -------------------------------------------- */
798:
1.67 kristaps 799: /*
800: * Pop the current node off of the stack of roff instructions currently
801: * pending.
802: */
803: static void
804: roffnode_pop(struct roff *r)
805: {
806: struct roffnode *p;
807:
1.75 kristaps 808: assert(r->last);
1.207 schwarze 809: p = r->last;
1.82 kristaps 810:
1.75 kristaps 811: r->last = r->last->parent;
1.106 kristaps 812: free(p->name);
813: free(p->end);
1.67 kristaps 814: free(p);
815: }
816:
817: /*
818: * Push a roff node onto the instruction stack. This must later be
819: * removed with roffnode_pop().
820: */
1.98 schwarze 821: static void
1.106 kristaps 822: roffnode_push(struct roff *r, enum rofft tok, const char *name,
823: int line, int col)
1.67 kristaps 824: {
825: struct roffnode *p;
826:
1.98 schwarze 827: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 828: p->tok = tok;
1.106 kristaps 829: if (name)
830: p->name = mandoc_strdup(name);
1.67 kristaps 831: p->parent = r->last;
832: p->line = line;
833: p->col = col;
1.198 schwarze 834: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 835:
836: r->last = p;
837: }
838:
1.266 schwarze 839: /* --- roff parser state data management ---------------------------------- */
840:
1.67 kristaps 841: static void
842: roff_free1(struct roff *r)
843: {
1.176 schwarze 844: struct tbl_node *tbl;
1.125 kristaps 845: struct eqn_node *e;
1.167 kristaps 846: int i;
1.67 kristaps 847:
1.176 schwarze 848: while (NULL != (tbl = r->first_tbl)) {
849: r->first_tbl = tbl->next;
850: tbl_free(tbl);
1.109 kristaps 851: }
1.113 kristaps 852: r->first_tbl = r->last_tbl = r->tbl = NULL;
853:
1.125 kristaps 854: while (NULL != (e = r->first_eqn)) {
855: r->first_eqn = e->next;
856: eqn_free(e);
857: }
858: r->first_eqn = r->last_eqn = r->eqn = NULL;
859:
1.67 kristaps 860: while (r->last)
861: roffnode_pop(r);
1.109 kristaps 862:
1.223 schwarze 863: free (r->rstack);
864: r->rstack = NULL;
865: r->rstacksz = 0;
866: r->rstackpos = -1;
867:
868: roff_freereg(r->regtab);
869: r->regtab = NULL;
870:
1.167 kristaps 871: roff_freestr(r->strtab);
872: roff_freestr(r->xmbtab);
873: r->strtab = r->xmbtab = NULL;
874:
875: if (r->xtab)
876: for (i = 0; i < 128; i++)
877: free(r->xtab[i].p);
878: free(r->xtab);
879: r->xtab = NULL;
1.67 kristaps 880: }
881:
882: void
883: roff_reset(struct roff *r)
884: {
885:
886: roff_free1(r);
1.227 schwarze 887: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174 kristaps 888: r->control = 0;
1.67 kristaps 889: }
890:
891: void
892: roff_free(struct roff *r)
893: {
894:
895: roff_free1(r);
896: free(r);
897: }
898:
899: struct roff *
1.279 schwarze 900: roff_alloc(struct mparse *parse, int options)
1.67 kristaps 901: {
902: struct roff *r;
903:
1.98 schwarze 904: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 905: r->parse = parse;
1.199 schwarze 906: r->options = options;
1.227 schwarze 907: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 908: r->rstackpos = -1;
1.207 schwarze 909:
1.155 kristaps 910: roffhash_init();
1.141 kristaps 911:
1.277 schwarze 912: return r;
1.265 schwarze 913: }
914:
1.266 schwarze 915: /* --- syntax tree state data management ---------------------------------- */
916:
1.265 schwarze 917: static void
918: roff_man_free1(struct roff_man *man)
919: {
920:
1.266 schwarze 921: if (man->first != NULL)
922: roff_node_delete(man, man->first);
1.265 schwarze 923: free(man->meta.msec);
924: free(man->meta.vol);
925: free(man->meta.os);
926: free(man->meta.arch);
927: free(man->meta.title);
928: free(man->meta.name);
929: free(man->meta.date);
930: }
931:
932: static void
933: roff_man_alloc1(struct roff_man *man)
934: {
935:
936: memset(&man->meta, 0, sizeof(man->meta));
937: man->first = mandoc_calloc(1, sizeof(*man->first));
938: man->first->type = ROFFT_ROOT;
939: man->last = man->first;
940: man->last_es = NULL;
941: man->flags = 0;
942: man->macroset = MACROSET_NONE;
943: man->lastsec = man->lastnamed = SEC_NONE;
944: man->next = ROFF_NEXT_CHILD;
945: }
946:
947: void
948: roff_man_reset(struct roff_man *man)
949: {
950:
951: roff_man_free1(man);
952: roff_man_alloc1(man);
953: }
954:
955: void
956: roff_man_free(struct roff_man *man)
957: {
958:
959: roff_man_free1(man);
960: free(man);
961: }
962:
963: struct roff_man *
964: roff_man_alloc(struct roff *roff, struct mparse *parse,
965: const char *defos, int quick)
966: {
967: struct roff_man *man;
968:
969: man = mandoc_calloc(1, sizeof(*man));
970: man->parse = parse;
971: man->roff = roff;
972: man->defos = defos;
973: man->quick = quick;
974: roff_man_alloc1(man);
1.277 schwarze 975: return man;
1.67 kristaps 976: }
977:
1.266 schwarze 978: /* --- syntax tree handling ----------------------------------------------- */
979:
980: struct roff_node *
981: roff_node_alloc(struct roff_man *man, int line, int pos,
982: enum roff_type type, int tok)
983: {
984: struct roff_node *n;
985:
986: n = mandoc_calloc(1, sizeof(*n));
987: n->line = line;
988: n->pos = pos;
989: n->tok = tok;
990: n->type = type;
991: n->sec = man->lastsec;
992:
993: if (man->flags & MDOC_SYNOPSIS)
1.285 schwarze 994: n->flags |= NODE_SYNPRETTY;
1.266 schwarze 995: else
1.285 schwarze 996: n->flags &= ~NODE_SYNPRETTY;
1.266 schwarze 997: if (man->flags & MDOC_NEWLINE)
1.285 schwarze 998: n->flags |= NODE_LINE;
1.266 schwarze 999: man->flags &= ~MDOC_NEWLINE;
1000:
1.277 schwarze 1001: return n;
1.266 schwarze 1002: }
1003:
1004: void
1005: roff_node_append(struct roff_man *man, struct roff_node *n)
1006: {
1007:
1008: switch (man->next) {
1009: case ROFF_NEXT_SIBLING:
1.281 schwarze 1010: if (man->last->next != NULL) {
1011: n->next = man->last->next;
1012: man->last->next->prev = n;
1013: } else
1014: man->last->parent->last = n;
1.266 schwarze 1015: man->last->next = n;
1016: n->prev = man->last;
1017: n->parent = man->last->parent;
1018: break;
1019: case ROFF_NEXT_CHILD:
1.287 ! schwarze 1020: if (man->last->child != NULL) {
! 1021: n->next = man->last->child;
! 1022: man->last->child->prev = n;
! 1023: } else
! 1024: man->last->last = n;
1.266 schwarze 1025: man->last->child = n;
1026: n->parent = man->last;
1027: break;
1028: default:
1029: abort();
1030: }
1.282 schwarze 1031: man->last = n;
1.266 schwarze 1032:
1033: switch (n->type) {
1034: case ROFFT_HEAD:
1035: n->parent->head = n;
1036: break;
1037: case ROFFT_BODY:
1.282 schwarze 1038: if (n->end != ENDBODY_NOT)
1039: return;
1.266 schwarze 1040: n->parent->body = n;
1041: break;
1042: case ROFFT_TAIL:
1043: n->parent->tail = n;
1044: break;
1045: default:
1.282 schwarze 1046: return;
1.266 schwarze 1047: }
1.282 schwarze 1048:
1049: /*
1050: * Copy over the normalised-data pointer of our parent. Not
1051: * everybody has one, but copying a null pointer is fine.
1052: */
1053:
1054: n->norm = n->parent->norm;
1055: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 1056: }
1057:
1.267 schwarze 1058: void
1059: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
1060: {
1061: struct roff_node *n;
1062:
1063: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
1064: n->string = roff_strdup(man->roff, word);
1065: roff_node_append(man, n);
1.286 schwarze 1066: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 1067: man->next = ROFF_NEXT_SIBLING;
1068: }
1069:
1070: void
1071: roff_word_append(struct roff_man *man, const char *word)
1072: {
1073: struct roff_node *n;
1074: char *addstr, *newstr;
1075:
1076: n = man->last;
1077: addstr = roff_strdup(man->roff, word);
1078: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1079: free(addstr);
1080: free(n->string);
1081: n->string = newstr;
1082: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 1083: }
1084:
1085: void
1086: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
1087: {
1088: struct roff_node *n;
1089:
1090: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
1091: roff_node_append(man, n);
1092: man->next = ROFF_NEXT_CHILD;
1093: }
1094:
1095: struct roff_node *
1096: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
1097: {
1098: struct roff_node *n;
1099:
1100: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
1101: roff_node_append(man, n);
1102: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1103: return n;
1.267 schwarze 1104: }
1105:
1.266 schwarze 1106: struct roff_node *
1107: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
1108: {
1109: struct roff_node *n;
1110:
1111: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
1112: roff_node_append(man, n);
1113: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1114: return n;
1.266 schwarze 1115: }
1116:
1117: struct roff_node *
1118: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1119: {
1120: struct roff_node *n;
1121:
1122: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1123: roff_node_append(man, n);
1124: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1125: return n;
1.267 schwarze 1126: }
1127:
1128: void
1129: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
1130: {
1131: struct roff_node *n;
1132:
1133: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
1134: n->eqn = eqn;
1135: if (eqn->ln > man->last->line)
1.285 schwarze 1136: n->flags |= NODE_LINE;
1.267 schwarze 1137: roff_node_append(man, n);
1138: man->next = ROFF_NEXT_SIBLING;
1139: }
1140:
1141: void
1142: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
1143: {
1144: struct roff_node *n;
1145:
1146: if (man->macroset == MACROSET_MAN)
1147: man_breakscope(man, TOKEN_NONE);
1148: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
1149: n->span = tbl;
1150: roff_node_append(man, n);
1.286 schwarze 1151: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 1152: man->next = ROFF_NEXT_SIBLING;
1.266 schwarze 1153: }
1154:
1155: void
1156: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1157: {
1158:
1159: /* Adjust siblings. */
1160:
1161: if (n->prev)
1162: n->prev->next = n->next;
1163: if (n->next)
1164: n->next->prev = n->prev;
1165:
1166: /* Adjust parent. */
1167:
1168: if (n->parent != NULL) {
1169: if (n->parent->child == n)
1170: n->parent->child = n->next;
1171: if (n->parent->last == n)
1172: n->parent->last = n->prev;
1173: }
1174:
1175: /* Adjust parse point. */
1176:
1177: if (man == NULL)
1178: return;
1179: if (man->last == n) {
1180: if (n->prev == NULL) {
1181: man->last = n->parent;
1182: man->next = ROFF_NEXT_CHILD;
1183: } else {
1184: man->last = n->prev;
1185: man->next = ROFF_NEXT_SIBLING;
1186: }
1187: }
1188: if (man->first == n)
1189: man->first = NULL;
1190: }
1191:
1192: void
1193: roff_node_free(struct roff_node *n)
1194: {
1195:
1196: if (n->args != NULL)
1197: mdoc_argv_free(n->args);
1198: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1199: free(n->norm);
1200: free(n->string);
1201: free(n);
1202: }
1203:
1204: void
1205: roff_node_delete(struct roff_man *man, struct roff_node *n)
1206: {
1207:
1208: while (n->child != NULL)
1209: roff_node_delete(man, n->child);
1210: roff_node_unlink(man, n);
1211: roff_node_free(n);
1.269 schwarze 1212: }
1213:
1214: void
1215: deroff(char **dest, const struct roff_node *n)
1216: {
1217: char *cp;
1218: size_t sz;
1219:
1220: if (n->type != ROFFT_TEXT) {
1221: for (n = n->child; n != NULL; n = n->next)
1222: deroff(dest, n);
1223: return;
1224: }
1225:
1226: /* Skip leading whitespace and escape sequences. */
1227:
1228: cp = n->string;
1229: while (*cp != '\0') {
1230: if ('\\' == *cp) {
1231: cp++;
1232: mandoc_escape((const char **)&cp, NULL, NULL);
1233: } else if (isspace((unsigned char)*cp))
1234: cp++;
1235: else
1236: break;
1237: }
1238:
1239: /* Skip trailing whitespace. */
1240:
1241: for (sz = strlen(cp); sz; sz--)
1242: if ( ! isspace((unsigned char)cp[sz-1]))
1243: break;
1244:
1245: /* Skip empty strings. */
1246:
1247: if (sz == 0)
1248: return;
1249:
1250: if (*dest == NULL) {
1251: *dest = mandoc_strndup(cp, sz);
1252: return;
1253: }
1254:
1255: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1256: free(*dest);
1257: *dest = cp;
1.266 schwarze 1258: }
1259:
1260: /* --- main functions of the roff parser ---------------------------------- */
1261:
1.94 kristaps 1262: /*
1.206 schwarze 1263: * In the current line, expand escape sequences that tend to get
1264: * used in numerical expressions and conditional requests.
1265: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1266: */
1.172 schwarze 1267: static enum rofferr
1.238 schwarze 1268: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1269: {
1.208 schwarze 1270: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1271: const char *start; /* start of the string to process */
1.209 schwarze 1272: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1273: const char *stnam; /* start of the name, after "[(*" */
1274: const char *cp; /* end of the name, e.g. before ']' */
1275: const char *res; /* the string to be substituted */
1.238 schwarze 1276: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1277: size_t maxl; /* expected length of the escape name */
1278: size_t naml; /* actual length of the escape name */
1.237 schwarze 1279: enum mandoc_esc esc; /* type of the escape sequence */
1280: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1281: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1282: int npos; /* position in numeric expression */
1.218 schwarze 1283: int arg_complete; /* argument not interrupted by eol */
1.206 schwarze 1284: char term; /* character terminating the escape */
1.94 kristaps 1285:
1.170 schwarze 1286: expand_count = 0;
1.238 schwarze 1287: start = buf->buf + pos;
1.205 schwarze 1288: stesc = strchr(start, '\0') - 1;
1289: while (stesc-- > start) {
1.170 schwarze 1290:
1.205 schwarze 1291: /* Search backwards for the next backslash. */
1292:
1.238 schwarze 1293: if (*stesc != '\\')
1.205 schwarze 1294: continue;
1295:
1296: /* If it is escaped, skip it. */
1297:
1298: for (cp = stesc - 1; cp >= start; cp--)
1.238 schwarze 1299: if (*cp != '\\')
1.205 schwarze 1300: break;
1301:
1.238 schwarze 1302: if ((stesc - cp) % 2 == 0) {
1.209 schwarze 1303: stesc = (char *)cp;
1.205 schwarze 1304: continue;
1305: }
1.108 schwarze 1306:
1.206 schwarze 1307: /* Decide whether to expand or to check only. */
1.108 schwarze 1308:
1.206 schwarze 1309: term = '\0';
1.205 schwarze 1310: cp = stesc + 1;
1.181 schwarze 1311: switch (*cp) {
1.207 schwarze 1312: case '*':
1.181 schwarze 1313: res = NULL;
1314: break;
1.207 schwarze 1315: case 'B':
1316: case 'w':
1.206 schwarze 1317: term = cp[1];
1318: /* FALLTHROUGH */
1.207 schwarze 1319: case 'n':
1.181 schwarze 1320: res = ubuf;
1321: break;
1322: default:
1.237 schwarze 1323: esc = mandoc_escape(&cp, &stnam, &inaml);
1324: if (esc == ESCAPE_ERROR ||
1325: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1326: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1327: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1328: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1329: "%.*s", (int)(cp - stesc), stesc);
1.205 schwarze 1330: continue;
1.152 kristaps 1331: }
1332:
1.205 schwarze 1333: if (EXPAND_LIMIT < ++expand_count) {
1334: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1335: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1336: return ROFF_IGN;
1.205 schwarze 1337: }
1.108 schwarze 1338:
1339: /*
1340: * The third character decides the length
1.181 schwarze 1341: * of the name of the string or register.
1.108 schwarze 1342: * Save a pointer to the name.
1343: */
1344:
1.238 schwarze 1345: if (term == '\0') {
1.206 schwarze 1346: switch (*++cp) {
1.207 schwarze 1347: case '\0':
1.206 schwarze 1348: maxl = 0;
1349: break;
1.207 schwarze 1350: case '(':
1.206 schwarze 1351: cp++;
1352: maxl = 2;
1353: break;
1.207 schwarze 1354: case '[':
1.206 schwarze 1355: cp++;
1356: term = ']';
1357: maxl = 0;
1358: break;
1359: default:
1360: maxl = 1;
1361: break;
1362: }
1363: } else {
1364: cp += 2;
1.94 kristaps 1365: maxl = 0;
1366: }
1.108 schwarze 1367: stnam = cp;
1.94 kristaps 1368:
1.108 schwarze 1369: /* Advance to the end of the name. */
1.94 kristaps 1370:
1.253 schwarze 1371: naml = 0;
1.218 schwarze 1372: arg_complete = 1;
1.253 schwarze 1373: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1374: if (*cp == '\0') {
1.219 schwarze 1375: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1376: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1377: arg_complete = 0;
1.206 schwarze 1378: break;
1.153 kristaps 1379: }
1.238 schwarze 1380: if (maxl == 0 && *cp == term) {
1.206 schwarze 1381: cp++;
1.253 schwarze 1382: break;
1383: }
1384: if (*cp++ != '\\' || stesc[1] != 'w') {
1385: naml++;
1386: continue;
1387: }
1388: switch (mandoc_escape(&cp, NULL, NULL)) {
1389: case ESCAPE_SPECIAL:
1390: case ESCAPE_UNICODE:
1391: case ESCAPE_NUMBERED:
1392: case ESCAPE_OVERSTRIKE:
1393: naml++;
1394: break;
1395: default:
1.94 kristaps 1396: break;
1.206 schwarze 1397: }
1.94 kristaps 1398: }
1399:
1.108 schwarze 1400: /*
1401: * Retrieve the replacement string; if it is
1402: * undefined, resume searching for escapes.
1403: */
1404:
1.206 schwarze 1405: switch (stesc[1]) {
1.207 schwarze 1406: case '*':
1.218 schwarze 1407: if (arg_complete)
1408: res = roff_getstrn(r, stnam, naml);
1.206 schwarze 1409: break;
1.207 schwarze 1410: case 'B':
1.206 schwarze 1411: npos = 0;
1.218 schwarze 1412: ubuf[0] = arg_complete &&
1.261 schwarze 1413: roff_evalnum(r, ln, stnam, &npos,
1414: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1415: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1416: ubuf[1] = '\0';
1417: break;
1.207 schwarze 1418: case 'n':
1.218 schwarze 1419: if (arg_complete)
1420: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1421: roff_getregn(r, stnam, naml));
1422: else
1423: ubuf[0] = '\0';
1.206 schwarze 1424: break;
1.207 schwarze 1425: case 'w':
1.218 schwarze 1426: /* use even incomplete args */
1.208 schwarze 1427: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1428: 24 * (int)naml);
1429: break;
1430: }
1.94 kristaps 1431:
1.238 schwarze 1432: if (res == NULL) {
1.219 schwarze 1433: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1434: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1435: "%.*s", (int)naml, stnam);
1.142 kristaps 1436: res = "";
1.246 schwarze 1437: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1438: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1439: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1440: return ROFF_IGN;
1.94 kristaps 1441: }
1442:
1.108 schwarze 1443: /* Replace the escape sequence by the string. */
1444:
1.209 schwarze 1445: *stesc = '\0';
1.238 schwarze 1446: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1447: buf->buf, res, cp) + 1;
1.94 kristaps 1448:
1.205 schwarze 1449: /* Prepare for the next replacement. */
1.94 kristaps 1450:
1.205 schwarze 1451: start = nbuf + pos;
1.238 schwarze 1452: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1453: free(buf->buf);
1454: buf->buf = nbuf;
1.154 kristaps 1455: }
1.277 schwarze 1456: return ROFF_CONT;
1.154 kristaps 1457: }
1458:
1459: /*
1.275 schwarze 1460: * Process text streams.
1.154 kristaps 1461: */
1462: static enum rofferr
1.238 schwarze 1463: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154 kristaps 1464: {
1465: size_t sz;
1466: const char *start;
1.178 schwarze 1467: char *p;
1468: int isz;
1.154 kristaps 1469: enum mandoc_esc esc;
1470:
1.275 schwarze 1471: /* Spring the input line trap. */
1472:
1473: if (roffit_lines == 1) {
1474: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1475: free(buf->buf);
1476: buf->buf = p;
1477: buf->sz = isz + 1;
1478: *offs = 0;
1479: free(roffit_macro);
1480: roffit_lines = 0;
1.277 schwarze 1481: return ROFF_REPARSE;
1.275 schwarze 1482: } else if (roffit_lines > 1)
1483: --roffit_lines;
1484:
1485: /* Convert all breakable hyphens into ASCII_HYPH. */
1486:
1.238 schwarze 1487: start = p = buf->buf + pos;
1.154 kristaps 1488:
1.238 schwarze 1489: while (*p != '\0') {
1.154 kristaps 1490: sz = strcspn(p, "-\\");
1491: p += sz;
1492:
1.238 schwarze 1493: if (*p == '\0')
1.159 kristaps 1494: break;
1495:
1.238 schwarze 1496: if (*p == '\\') {
1.154 kristaps 1497: /* Skip over escapes. */
1498: p++;
1.189 schwarze 1499: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1500: if (esc == ESCAPE_ERROR)
1.154 kristaps 1501: break;
1.264 schwarze 1502: while (*p == '-')
1503: p++;
1.155 kristaps 1504: continue;
1.159 kristaps 1505: } else if (p == start) {
1.158 kristaps 1506: p++;
1.155 kristaps 1507: continue;
1.158 kristaps 1508: }
1.155 kristaps 1509:
1.171 schwarze 1510: if (isalpha((unsigned char)p[-1]) &&
1511: isalpha((unsigned char)p[1]))
1.155 kristaps 1512: *p = ASCII_HYPH;
1513: p++;
1.94 kristaps 1514: }
1.277 schwarze 1515: return ROFF_CONT;
1.94 kristaps 1516: }
1517:
1.67 kristaps 1518: enum rofferr
1.238 schwarze 1519: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1520: {
1521: enum rofft t;
1.109 kristaps 1522: enum rofferr e;
1.238 schwarze 1523: int pos; /* parse point */
1.243 schwarze 1524: int spos; /* saved parse point for messages */
1.238 schwarze 1525: int ppos; /* original offset in buf->buf */
1526: int ctl; /* macro line (boolean) */
1527:
1528: ppos = pos = *offs;
1.79 kristaps 1529:
1.230 schwarze 1530: /* Handle in-line equation delimiters. */
1531:
1.236 schwarze 1532: if (r->tbl == NULL &&
1533: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1534: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1535: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1536: if (e == ROFF_REPARSE)
1.277 schwarze 1537: return e;
1.230 schwarze 1538: assert(e == ROFF_CONT);
1539: }
1540:
1541: /* Expand some escape sequences. */
1.94 kristaps 1542:
1.238 schwarze 1543: e = roff_res(r, buf, ln, pos);
1544: if (e == ROFF_IGN)
1.277 schwarze 1545: return e;
1.238 schwarze 1546: assert(e == ROFF_CONT);
1.94 kristaps 1547:
1.238 schwarze 1548: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1549:
1.94 kristaps 1550: /*
1.79 kristaps 1551: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1552: * text through the macro's filter.
1553: * Equations process all content themselves.
1554: * Tables process almost all content themselves, but we want
1555: * to warn about macros before passing it there.
1.79 kristaps 1556: */
1.74 kristaps 1557:
1.252 schwarze 1558: if (r->last != NULL && ! ctl) {
1.78 kristaps 1559: t = r->last->tok;
1560: assert(roffs[t].text);
1.238 schwarze 1561: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1562: assert(e == ROFF_IGN || e == ROFF_CONT);
1563: if (e != ROFF_CONT)
1.277 schwarze 1564: return e;
1.182 schwarze 1565: }
1.252 schwarze 1566: if (r->eqn != NULL)
1.277 schwarze 1567: return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.252 schwarze 1568: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.277 schwarze 1569: return tbl_read(r->tbl, ln, buf->buf, ppos);
1.252 schwarze 1570: if ( ! ctl)
1.277 schwarze 1571: return roff_parsetext(buf, pos, offs);
1.228 schwarze 1572:
1573: /* Skip empty request lines. */
1574:
1.238 schwarze 1575: if (buf->buf[pos] == '"') {
1.228 schwarze 1576: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1577: ln, pos, NULL);
1.277 schwarze 1578: return ROFF_IGN;
1.238 schwarze 1579: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1580: return ROFF_IGN;
1.67 kristaps 1581:
1.79 kristaps 1582: /*
1583: * If a scope is open, go to the child handler for that macro,
1584: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1585: * Don't do so if an equation is open.
1.79 kristaps 1586: */
1.78 kristaps 1587:
1.79 kristaps 1588: if (r->last) {
1589: t = r->last->tok;
1590: assert(roffs[t].sub);
1.277 schwarze 1591: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1592: }
1.78 kristaps 1593:
1.243 schwarze 1594: /* No scope is open. This is a new request or macro. */
1595:
1596: spos = pos;
1597: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1598:
1599: /* Tables ignore most macros. */
1600:
1601: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1602: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1603: ln, pos, buf->buf + spos);
1.257 schwarze 1604: if (t == ROFF_TS)
1.277 schwarze 1605: return ROFF_IGN;
1.257 schwarze 1606: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1607: pos++;
1608: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1609: pos++;
1.277 schwarze 1610: return tbl_read(r->tbl, ln, buf->buf, pos);
1.243 schwarze 1611: }
1612:
1.79 kristaps 1613: /*
1.243 schwarze 1614: * This is neither a roff request nor a user-defined macro.
1615: * Let the standard macro set parsers handle it.
1.79 kristaps 1616: */
1.67 kristaps 1617:
1.243 schwarze 1618: if (t == ROFF_MAX)
1.277 schwarze 1619: return ROFF_CONT;
1.243 schwarze 1620:
1621: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1622:
1.75 kristaps 1623: assert(roffs[t].proc);
1.277 schwarze 1624: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.74 kristaps 1625: }
1626:
1.117 kristaps 1627: void
1.74 kristaps 1628: roff_endparse(struct roff *r)
1629: {
1630:
1.110 kristaps 1631: if (r->last)
1.221 schwarze 1632: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1633: r->last->line, r->last->col,
1634: roffs[r->last->tok].name);
1.117 kristaps 1635:
1.125 kristaps 1636: if (r->eqn) {
1.221 schwarze 1637: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1638: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151 kristaps 1639: eqn_end(&r->eqn);
1.125 kristaps 1640: }
1641:
1.117 kristaps 1642: if (r->tbl) {
1.221 schwarze 1643: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1644: r->tbl->line, r->tbl->pos, "TS");
1.151 kristaps 1645: tbl_end(&r->tbl);
1.117 kristaps 1646: }
1.67 kristaps 1647: }
1648:
1649: /*
1650: * Parse a roff node's type from the input buffer. This must be in the
1651: * form of ".foo xxx" in the usual way.
1652: */
1653: static enum rofft
1.214 schwarze 1654: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1655: {
1.214 schwarze 1656: char *cp;
1.106 kristaps 1657: const char *mac;
1658: size_t maclen;
1.67 kristaps 1659: enum rofft t;
1660:
1.214 schwarze 1661: cp = buf + *pos;
1662:
1663: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.277 schwarze 1664: return ROFF_MAX;
1.67 kristaps 1665:
1.214 schwarze 1666: mac = cp;
1667: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1668:
1.106 kristaps 1669: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.155 kristaps 1670: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.67 kristaps 1671:
1.214 schwarze 1672: if (ROFF_MAX != t)
1673: *pos = cp - buf;
1.67 kristaps 1674:
1.277 schwarze 1675: return t;
1.67 kristaps 1676: }
1677:
1.266 schwarze 1678: /* --- handling of request blocks ----------------------------------------- */
1679:
1.67 kristaps 1680: static enum rofferr
1.76 kristaps 1681: roff_cblock(ROFF_ARGS)
1.67 kristaps 1682: {
1683:
1.79 kristaps 1684: /*
1685: * A block-close `..' should only be invoked as a child of an
1686: * ignore macro, otherwise raise a warning and just ignore it.
1687: */
1688:
1.238 schwarze 1689: if (r->last == NULL) {
1.221 schwarze 1690: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1691: ln, ppos, "..");
1.277 schwarze 1692: return ROFF_IGN;
1.76 kristaps 1693: }
1.67 kristaps 1694:
1.81 kristaps 1695: switch (r->last->tok) {
1.207 schwarze 1696: case ROFF_am:
1.220 schwarze 1697: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1698: case ROFF_ami:
1699: case ROFF_de:
1.108 schwarze 1700: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1701: case ROFF_dei:
1702: case ROFF_ig:
1.81 kristaps 1703: break;
1704: default:
1.221 schwarze 1705: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1706: ln, ppos, "..");
1.277 schwarze 1707: return ROFF_IGN;
1.76 kristaps 1708: }
1.67 kristaps 1709:
1.238 schwarze 1710: if (buf->buf[pos] != '\0')
1.217 schwarze 1711: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1712: ".. %s", buf->buf + pos);
1.71 kristaps 1713:
1714: roffnode_pop(r);
1.76 kristaps 1715: roffnode_cleanscope(r);
1.277 schwarze 1716: return ROFF_IGN;
1.71 kristaps 1717:
1.67 kristaps 1718: }
1719:
1.76 kristaps 1720: static void
1721: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1722: {
1723:
1.76 kristaps 1724: while (r->last) {
1.173 schwarze 1725: if (--r->last->endspan != 0)
1.76 kristaps 1726: break;
1727: roffnode_pop(r);
1728: }
1.67 kristaps 1729: }
1730:
1.195 schwarze 1731: static void
1732: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1733: {
1734:
1.76 kristaps 1735: if (NULL == r->last) {
1.221 schwarze 1736: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1737: ln, ppos, "\\}");
1.195 schwarze 1738: return;
1.76 kristaps 1739: }
1740:
1.82 kristaps 1741: switch (r->last->tok) {
1.207 schwarze 1742: case ROFF_el:
1743: case ROFF_ie:
1744: case ROFF_if:
1.82 kristaps 1745: break;
1746: default:
1.221 schwarze 1747: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1748: ln, ppos, "\\}");
1.195 schwarze 1749: return;
1.75 kristaps 1750: }
1751:
1.76 kristaps 1752: if (r->last->endspan > -1) {
1.221 schwarze 1753: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1754: ln, ppos, "\\}");
1.195 schwarze 1755: return;
1.76 kristaps 1756: }
1757:
1.75 kristaps 1758: roffnode_pop(r);
1.76 kristaps 1759: roffnode_cleanscope(r);
1.195 schwarze 1760: return;
1.76 kristaps 1761: }
1762:
1763: static enum rofferr
1.80 kristaps 1764: roff_block(ROFF_ARGS)
1.76 kristaps 1765: {
1.220 schwarze 1766: const char *name;
1767: char *iname, *cp;
1.213 schwarze 1768: size_t namesz;
1.106 kristaps 1769:
1.220 schwarze 1770: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1771:
1.238 schwarze 1772: if (tok == ROFF_de1)
1.220 schwarze 1773: tok = ROFF_de;
1.251 schwarze 1774: else if (tok == ROFF_dei1)
1775: tok = ROFF_dei;
1.238 schwarze 1776: else if (tok == ROFF_am1)
1.220 schwarze 1777: tok = ROFF_am;
1.251 schwarze 1778: else if (tok == ROFF_ami1)
1779: tok = ROFF_ami;
1.220 schwarze 1780:
1781: /* Parse the macro name argument. */
1782:
1.238 schwarze 1783: cp = buf->buf + pos;
1784: if (tok == ROFF_ig) {
1.220 schwarze 1785: iname = NULL;
1786: namesz = 0;
1787: } else {
1788: iname = cp;
1789: namesz = roff_getname(r, &cp, ln, ppos);
1790: iname[namesz] = '\0';
1791: }
1.107 kristaps 1792:
1.220 schwarze 1793: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1794:
1.238 schwarze 1795: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1796: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1797: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1798: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1799: "%.*s", (int)namesz, iname);
1800: namesz = 0;
1801: } else
1802: namesz = strlen(name);
1803: } else
1804: name = iname;
1.107 kristaps 1805:
1.238 schwarze 1806: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1807: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1808: ln, ppos, roffs[tok].name);
1.277 schwarze 1809: return ROFF_IGN;
1.220 schwarze 1810: }
1.80 kristaps 1811:
1.106 kristaps 1812: roffnode_push(r, tok, name, ln, ppos);
1813:
1814: /*
1815: * At the beginning of a `de' macro, clear the existing string
1816: * with the same name, if there is one. New content will be
1.193 schwarze 1817: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1818: */
1.107 kristaps 1819:
1.238 schwarze 1820: if (tok == ROFF_de || tok == ROFF_dei)
1.213 schwarze 1821: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76 kristaps 1822:
1.238 schwarze 1823: if (*cp == '\0')
1.277 schwarze 1824: return ROFF_IGN;
1.78 kristaps 1825:
1.220 schwarze 1826: /* Get the custom end marker. */
1.107 kristaps 1827:
1.220 schwarze 1828: iname = cp;
1.213 schwarze 1829: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1830:
1831: /* Resolve the end marker if it is indirect. */
1832:
1.238 schwarze 1833: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1834: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1835: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1836: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1837: "%.*s", (int)namesz, iname);
1838: namesz = 0;
1839: } else
1840: namesz = strlen(name);
1841: } else
1842: name = iname;
1843:
1.213 schwarze 1844: if (namesz)
1845: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1846:
1.238 schwarze 1847: if (*cp != '\0')
1.217 schwarze 1848: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1849: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.74 kristaps 1850:
1.277 schwarze 1851: return ROFF_IGN;
1.78 kristaps 1852: }
1853:
1854: static enum rofferr
1.80 kristaps 1855: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1856: {
1857: enum rofft t;
1858: int i, j;
1859:
1860: /*
1861: * First check whether a custom macro exists at this level. If
1862: * it does, then check against it. This is some of groff's
1863: * stranger behaviours. If we encountered a custom end-scope
1864: * tag and that tag also happens to be a "real" macro, then we
1865: * need to try interpreting it again as a real macro. If it's
1866: * not, then return ignore. Else continue.
1867: */
1868:
1869: if (r->last->end) {
1.130 kristaps 1870: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1871: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1872: break;
1873:
1.238 schwarze 1874: if (r->last->end[j] == '\0' &&
1875: (buf->buf[i] == '\0' ||
1876: buf->buf[i] == ' ' ||
1877: buf->buf[i] == '\t')) {
1.79 kristaps 1878: roffnode_pop(r);
1879: roffnode_cleanscope(r);
1880:
1.238 schwarze 1881: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1882: i++;
1883:
1884: pos = i;
1.238 schwarze 1885: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1886: ROFF_MAX)
1.277 schwarze 1887: return ROFF_RERUN;
1888: return ROFF_IGN;
1.79 kristaps 1889: }
1890: }
1891:
1892: /*
1893: * If we have no custom end-query or lookup failed, then try
1894: * pulling it out of the hashtable.
1895: */
1896:
1.238 schwarze 1897: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1898:
1.238 schwarze 1899: if (t != ROFF_cblock) {
1900: if (tok != ROFF_ig)
1901: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 1902: return ROFF_IGN;
1.106 kristaps 1903: }
1.79 kristaps 1904:
1905: assert(roffs[t].proc);
1.277 schwarze 1906: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1907: }
1908:
1909: static enum rofferr
1.80 kristaps 1910: roff_block_text(ROFF_ARGS)
1.78 kristaps 1911: {
1912:
1.238 schwarze 1913: if (tok != ROFF_ig)
1914: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1915:
1.277 schwarze 1916: return ROFF_IGN;
1.78 kristaps 1917: }
1918:
1919: static enum rofferr
1.82 kristaps 1920: roff_cond_sub(ROFF_ARGS)
1921: {
1922: enum rofft t;
1.139 kristaps 1923: char *ep;
1.198 schwarze 1924: int rr;
1.82 kristaps 1925:
1926: rr = r->last->rule;
1.139 kristaps 1927: roffnode_cleanscope(r);
1.238 schwarze 1928: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82 kristaps 1929:
1.139 kristaps 1930: /*
1.177 schwarze 1931: * Fully handle known macros when they are structurally
1932: * required or when the conditional evaluated to true.
1.87 kristaps 1933: */
1934:
1.238 schwarze 1935: if ((t != ROFF_MAX) &&
1936: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.177 schwarze 1937: assert(roffs[t].proc);
1.277 schwarze 1938: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.177 schwarze 1939: }
1.144 kristaps 1940:
1.196 schwarze 1941: /*
1942: * If `\}' occurs on a macro line without a preceding macro,
1943: * drop the line completely.
1944: */
1945:
1.238 schwarze 1946: ep = buf->buf + pos;
1947: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1948: rr = 0;
1.196 schwarze 1949:
1.177 schwarze 1950: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1951:
1.238 schwarze 1952: while ((ep = strchr(ep, '\\')) != NULL) {
1953: if (*(++ep) == '}') {
1.197 schwarze 1954: *ep = '&';
1.238 schwarze 1955: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1956: }
1.247 schwarze 1957: if (*ep != '\0')
1958: ++ep;
1.177 schwarze 1959: }
1.277 schwarze 1960: return rr ? ROFF_CONT : ROFF_IGN;
1.82 kristaps 1961: }
1962:
1963: static enum rofferr
1964: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1965: {
1.140 kristaps 1966: char *ep;
1.198 schwarze 1967: int rr;
1.82 kristaps 1968:
1969: rr = r->last->rule;
1.140 kristaps 1970: roffnode_cleanscope(r);
1.82 kristaps 1971:
1.238 schwarze 1972: ep = buf->buf + pos;
1973: while ((ep = strchr(ep, '\\')) != NULL) {
1974: if (*(++ep) == '}') {
1.197 schwarze 1975: *ep = '&';
1.238 schwarze 1976: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1977: }
1.247 schwarze 1978: if (*ep != '\0')
1979: ++ep;
1.78 kristaps 1980: }
1.277 schwarze 1981: return rr ? ROFF_CONT : ROFF_IGN;
1.74 kristaps 1982: }
1983:
1.266 schwarze 1984: /* --- handling of numeric and conditional expressions -------------------- */
1985:
1.204 schwarze 1986: /*
1987: * Parse a single signed integer number. Stop at the first non-digit.
1988: * If there is at least one digit, return success and advance the
1989: * parse point, else return failure and let the parse point unchanged.
1990: * Ignore overflows, treat them just like the C language.
1991: */
1.184 schwarze 1992: static int
1.261 schwarze 1993: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 1994: {
1.261 schwarze 1995: int myres, scaled, n, p;
1.206 schwarze 1996:
1997: if (NULL == res)
1998: res = &myres;
1.184 schwarze 1999:
2000: p = *pos;
2001: n = v[p] == '-';
1.261 schwarze 2002: if (n || v[p] == '+')
1.184 schwarze 2003: p++;
2004:
1.261 schwarze 2005: if (flags & ROFFNUM_WHITE)
2006: while (isspace((unsigned char)v[p]))
2007: p++;
2008:
1.184 schwarze 2009: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2010: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2011: if (p == *pos + n)
2012: return 0;
2013:
2014: if (n)
2015: *res = -*res;
2016:
1.254 schwarze 2017: /* Each number may be followed by one optional scaling unit. */
2018:
2019: switch (v[p]) {
2020: case 'f':
1.261 schwarze 2021: scaled = *res * 65536;
1.254 schwarze 2022: break;
2023: case 'i':
1.261 schwarze 2024: scaled = *res * 240;
1.254 schwarze 2025: break;
2026: case 'c':
1.261 schwarze 2027: scaled = *res * 240 / 2.54;
1.254 schwarze 2028: break;
2029: case 'v':
2030: case 'P':
1.261 schwarze 2031: scaled = *res * 40;
1.254 schwarze 2032: break;
2033: case 'm':
2034: case 'n':
1.261 schwarze 2035: scaled = *res * 24;
1.254 schwarze 2036: break;
2037: case 'p':
1.261 schwarze 2038: scaled = *res * 10 / 3;
1.254 schwarze 2039: break;
2040: case 'u':
1.261 schwarze 2041: scaled = *res;
1.254 schwarze 2042: break;
2043: case 'M':
1.261 schwarze 2044: scaled = *res * 6 / 25;
1.254 schwarze 2045: break;
2046: default:
1.261 schwarze 2047: scaled = *res;
1.254 schwarze 2048: p--;
2049: break;
2050: }
1.261 schwarze 2051: if (flags & ROFFNUM_SCALE)
2052: *res = scaled;
1.254 schwarze 2053:
2054: *pos = p + 1;
1.277 schwarze 2055: return 1;
1.184 schwarze 2056: }
2057:
1.198 schwarze 2058: /*
2059: * Evaluate a string comparison condition.
2060: * The first character is the delimiter.
2061: * Succeed if the string up to its second occurrence
2062: * matches the string up to its third occurence.
2063: * Advance the cursor after the third occurrence
2064: * or lacking that, to the end of the line.
2065: */
2066: static int
2067: roff_evalstrcond(const char *v, int *pos)
2068: {
2069: const char *s1, *s2, *s3;
2070: int match;
2071:
2072: match = 0;
2073: s1 = v + *pos; /* initial delimiter */
2074: s2 = s1 + 1; /* for scanning the first string */
2075: s3 = strchr(s2, *s1); /* for scanning the second string */
2076:
2077: if (NULL == s3) /* found no middle delimiter */
2078: goto out;
2079:
2080: while ('\0' != *++s3) {
2081: if (*s2 != *s3) { /* mismatch */
2082: s3 = strchr(s3, *s1);
2083: break;
2084: }
2085: if (*s3 == *s1) { /* found the final delimiter */
2086: match = 1;
2087: break;
2088: }
2089: s2++;
2090: }
2091:
2092: out:
2093: if (NULL == s3)
2094: s3 = strchr(s2, '\0');
1.242 schwarze 2095: else if (*s3 != '\0')
1.198 schwarze 2096: s3++;
2097: *pos = s3 - v;
1.277 schwarze 2098: return match;
1.198 schwarze 2099: }
2100:
1.204 schwarze 2101: /*
2102: * Evaluate an optionally negated single character, numerical,
2103: * or string condition.
2104: */
1.198 schwarze 2105: static int
1.271 schwarze 2106: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2107: {
1.271 schwarze 2108: char *cp, *name;
2109: size_t sz;
1.241 schwarze 2110: int number, savepos, wanttrue;
1.88 kristaps 2111:
1.198 schwarze 2112: if ('!' == v[*pos]) {
2113: wanttrue = 0;
2114: (*pos)++;
2115: } else
2116: wanttrue = 1;
2117:
1.88 kristaps 2118: switch (v[*pos]) {
1.240 schwarze 2119: case '\0':
1.277 schwarze 2120: return 0;
1.207 schwarze 2121: case 'n':
2122: case 'o':
1.88 kristaps 2123: (*pos)++;
1.277 schwarze 2124: return wanttrue;
1.207 schwarze 2125: case 'c':
2126: case 'd':
2127: case 'e':
2128: case 't':
1.239 schwarze 2129: case 'v':
1.88 kristaps 2130: (*pos)++;
1.277 schwarze 2131: return !wanttrue;
1.271 schwarze 2132: case 'r':
2133: cp = name = v + ++*pos;
2134: sz = roff_getname(r, &cp, ln, *pos);
2135: *pos = cp - v;
1.277 schwarze 2136: return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.88 kristaps 2137: default:
2138: break;
2139: }
2140:
1.241 schwarze 2141: savepos = *pos;
1.261 schwarze 2142: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2143: return (number > 0) == wanttrue;
1.241 schwarze 2144: else if (*pos == savepos)
1.277 schwarze 2145: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2146: else
1.277 schwarze 2147: return 0;
1.88 kristaps 2148: }
2149:
1.74 kristaps 2150: static enum rofferr
1.103 kristaps 2151: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2152: {
1.123 schwarze 2153:
1.277 schwarze 2154: return ROFF_IGN;
1.89 kristaps 2155: }
2156:
1.104 kristaps 2157: static enum rofferr
1.251 schwarze 2158: roff_insec(ROFF_ARGS)
2159: {
2160:
2161: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
2162: ln, ppos, roffs[tok].name);
1.277 schwarze 2163: return ROFF_IGN;
1.251 schwarze 2164: }
2165:
2166: static enum rofferr
2167: roff_unsupp(ROFF_ARGS)
2168: {
2169:
2170: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
2171: ln, ppos, roffs[tok].name);
1.277 schwarze 2172: return ROFF_IGN;
1.251 schwarze 2173: }
2174:
2175: static enum rofferr
1.82 kristaps 2176: roff_cond(ROFF_ARGS)
1.74 kristaps 2177: {
1.173 schwarze 2178:
2179: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2180:
1.207 schwarze 2181: /*
1.134 kristaps 2182: * An `.el' has no conditional body: it will consume the value
2183: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2184: * defaults to DENY.
1.134 kristaps 2185: *
2186: * If we're not an `el', however, then evaluate the conditional.
2187: */
1.133 kristaps 2188:
1.238 schwarze 2189: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2190: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2191: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2192:
1.134 kristaps 2193: /*
2194: * An if-else will put the NEGATION of the current evaluated
2195: * conditional into the stack of rules.
2196: */
2197:
1.238 schwarze 2198: if (tok == ROFF_ie) {
1.223 schwarze 2199: if (r->rstackpos + 1 == r->rstacksz) {
2200: r->rstacksz += 16;
2201: r->rstack = mandoc_reallocarray(r->rstack,
2202: r->rstacksz, sizeof(int));
1.134 kristaps 2203: }
1.198 schwarze 2204: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2205: }
1.88 kristaps 2206:
2207: /* If the parent has false as its rule, then so do we. */
2208:
1.198 schwarze 2209: if (r->last->parent && !r->last->parent->rule)
2210: r->last->rule = 0;
1.88 kristaps 2211:
2212: /*
1.173 schwarze 2213: * Determine scope.
2214: * If there is nothing on the line after the conditional,
2215: * not even whitespace, use next-line scope.
1.88 kristaps 2216: */
1.74 kristaps 2217:
1.238 schwarze 2218: if (buf->buf[pos] == '\0') {
1.173 schwarze 2219: r->last->endspan = 2;
2220: goto out;
2221: }
2222:
1.238 schwarze 2223: while (buf->buf[pos] == ' ')
1.173 schwarze 2224: pos++;
2225:
2226: /* An opening brace requests multiline scope. */
1.75 kristaps 2227:
1.238 schwarze 2228: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2229: r->last->endspan = -1;
2230: pos += 2;
1.272 schwarze 2231: while (buf->buf[pos] == ' ')
2232: pos++;
1.173 schwarze 2233: goto out;
1.207 schwarze 2234: }
1.74 kristaps 2235:
1.77 kristaps 2236: /*
1.173 schwarze 2237: * Anything else following the conditional causes
2238: * single-line scope. Warn if the scope contains
2239: * nothing but trailing whitespace.
1.77 kristaps 2240: */
2241:
1.238 schwarze 2242: if (buf->buf[pos] == '\0')
1.216 schwarze 2243: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
2244: ln, ppos, roffs[tok].name);
1.77 kristaps 2245:
1.173 schwarze 2246: r->last->endspan = 1;
1.74 kristaps 2247:
1.173 schwarze 2248: out:
1.75 kristaps 2249: *offs = pos;
1.277 schwarze 2250: return ROFF_RERUN;
1.83 schwarze 2251: }
2252:
2253: static enum rofferr
1.92 schwarze 2254: roff_ds(ROFF_ARGS)
2255: {
1.212 schwarze 2256: char *string;
2257: const char *name;
2258: size_t namesz;
1.96 kristaps 2259:
1.251 schwarze 2260: /* Ignore groff compatibility mode for now. */
2261:
2262: if (tok == ROFF_ds1)
2263: tok = ROFF_ds;
2264: else if (tok == ROFF_as1)
2265: tok = ROFF_as;
2266:
1.96 kristaps 2267: /*
1.212 schwarze 2268: * The first word is the name of the string.
2269: * If it is empty or terminated by an escape sequence,
2270: * abort the `ds' request without defining anything.
1.96 kristaps 2271: */
1.92 schwarze 2272:
1.238 schwarze 2273: name = string = buf->buf + pos;
2274: if (*name == '\0')
1.277 schwarze 2275: return ROFF_IGN;
1.92 schwarze 2276:
1.212 schwarze 2277: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2278: if (name[namesz] == '\\')
1.277 schwarze 2279: return ROFF_IGN;
1.212 schwarze 2280:
2281: /* Read past the initial double-quote, if any. */
1.238 schwarze 2282: if (*string == '"')
1.92 schwarze 2283: string++;
2284:
1.96 kristaps 2285: /* The rest is the value. */
1.212 schwarze 2286: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2287: ROFF_as == tok);
1.277 schwarze 2288: return ROFF_IGN;
1.92 schwarze 2289: }
2290:
1.204 schwarze 2291: /*
2292: * Parse a single operator, one or two characters long.
2293: * If the operator is recognized, return success and advance the
2294: * parse point, else return failure and let the parse point unchanged.
2295: */
2296: static int
2297: roff_getop(const char *v, int *pos, char *res)
2298: {
2299:
2300: *res = v[*pos];
2301:
2302: switch (*res) {
1.207 schwarze 2303: case '+':
2304: case '-':
2305: case '*':
2306: case '/':
2307: case '%':
2308: case '&':
2309: case ':':
1.204 schwarze 2310: break;
2311: case '<':
2312: switch (v[*pos + 1]) {
1.207 schwarze 2313: case '=':
1.204 schwarze 2314: *res = 'l';
2315: (*pos)++;
2316: break;
1.207 schwarze 2317: case '>':
1.204 schwarze 2318: *res = '!';
2319: (*pos)++;
2320: break;
1.207 schwarze 2321: case '?':
1.204 schwarze 2322: *res = 'i';
2323: (*pos)++;
2324: break;
2325: default:
2326: break;
2327: }
2328: break;
2329: case '>':
2330: switch (v[*pos + 1]) {
1.207 schwarze 2331: case '=':
1.204 schwarze 2332: *res = 'g';
2333: (*pos)++;
2334: break;
1.207 schwarze 2335: case '?':
1.204 schwarze 2336: *res = 'a';
2337: (*pos)++;
2338: break;
2339: default:
2340: break;
2341: }
2342: break;
2343: case '=':
2344: if ('=' == v[*pos + 1])
2345: (*pos)++;
2346: break;
2347: default:
1.277 schwarze 2348: return 0;
1.204 schwarze 2349: }
2350: (*pos)++;
2351:
1.277 schwarze 2352: return *res;
1.204 schwarze 2353: }
2354:
2355: /*
2356: * Evaluate either a parenthesized numeric expression
2357: * or a single signed integer number.
2358: */
2359: static int
1.235 schwarze 2360: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2361: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2362: {
2363:
2364: if ('(' != v[*pos])
1.277 schwarze 2365: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2366:
2367: (*pos)++;
1.261 schwarze 2368: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2369: return 0;
1.204 schwarze 2370:
1.206 schwarze 2371: /*
2372: * Omission of the closing parenthesis
2373: * is an error in validation mode,
2374: * but ignored in evaluation mode.
2375: */
2376:
1.204 schwarze 2377: if (')' == v[*pos])
2378: (*pos)++;
1.206 schwarze 2379: else if (NULL == res)
1.277 schwarze 2380: return 0;
1.204 schwarze 2381:
1.277 schwarze 2382: return 1;
1.204 schwarze 2383: }
2384:
2385: /*
2386: * Evaluate a complete numeric expression.
2387: * Proceed left to right, there is no concept of precedence.
2388: */
2389: static int
1.235 schwarze 2390: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2391: int *pos, int *res, int flags)
1.204 schwarze 2392: {
2393: int mypos, operand2;
2394: char operator;
2395:
2396: if (NULL == pos) {
2397: mypos = 0;
2398: pos = &mypos;
2399: }
2400:
1.261 schwarze 2401: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2402: while (isspace((unsigned char)v[*pos]))
2403: (*pos)++;
2404:
1.261 schwarze 2405: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2406: return 0;
1.204 schwarze 2407:
2408: while (1) {
1.261 schwarze 2409: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2410: while (isspace((unsigned char)v[*pos]))
2411: (*pos)++;
2412:
2413: if ( ! roff_getop(v, pos, &operator))
2414: break;
2415:
1.261 schwarze 2416: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2417: while (isspace((unsigned char)v[*pos]))
2418: (*pos)++;
2419:
1.261 schwarze 2420: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2421: return 0;
1.204 schwarze 2422:
1.261 schwarze 2423: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2424: while (isspace((unsigned char)v[*pos]))
2425: (*pos)++;
1.206 schwarze 2426:
2427: if (NULL == res)
2428: continue;
1.204 schwarze 2429:
2430: switch (operator) {
1.207 schwarze 2431: case '+':
1.204 schwarze 2432: *res += operand2;
2433: break;
1.207 schwarze 2434: case '-':
1.204 schwarze 2435: *res -= operand2;
2436: break;
1.207 schwarze 2437: case '*':
1.204 schwarze 2438: *res *= operand2;
2439: break;
1.207 schwarze 2440: case '/':
1.244 schwarze 2441: if (operand2 == 0) {
1.235 schwarze 2442: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2443: r->parse, ln, *pos, v);
2444: *res = 0;
2445: break;
2446: }
1.204 schwarze 2447: *res /= operand2;
2448: break;
1.207 schwarze 2449: case '%':
1.244 schwarze 2450: if (operand2 == 0) {
2451: mandoc_msg(MANDOCERR_DIVZERO,
2452: r->parse, ln, *pos, v);
2453: *res = 0;
2454: break;
2455: }
1.204 schwarze 2456: *res %= operand2;
2457: break;
1.207 schwarze 2458: case '<':
1.204 schwarze 2459: *res = *res < operand2;
2460: break;
1.207 schwarze 2461: case '>':
1.204 schwarze 2462: *res = *res > operand2;
2463: break;
1.207 schwarze 2464: case 'l':
1.204 schwarze 2465: *res = *res <= operand2;
2466: break;
1.207 schwarze 2467: case 'g':
1.204 schwarze 2468: *res = *res >= operand2;
2469: break;
1.207 schwarze 2470: case '=':
1.204 schwarze 2471: *res = *res == operand2;
2472: break;
1.207 schwarze 2473: case '!':
1.204 schwarze 2474: *res = *res != operand2;
2475: break;
1.207 schwarze 2476: case '&':
1.204 schwarze 2477: *res = *res && operand2;
2478: break;
1.207 schwarze 2479: case ':':
1.204 schwarze 2480: *res = *res || operand2;
2481: break;
1.207 schwarze 2482: case 'i':
1.204 schwarze 2483: if (operand2 < *res)
2484: *res = operand2;
2485: break;
1.207 schwarze 2486: case 'a':
1.204 schwarze 2487: if (operand2 > *res)
2488: *res = operand2;
2489: break;
2490: default:
2491: abort();
2492: }
2493: }
1.277 schwarze 2494: return 1;
1.204 schwarze 2495: }
2496:
1.266 schwarze 2497: /* --- register management ------------------------------------------------ */
2498:
1.180 schwarze 2499: void
1.187 schwarze 2500: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2501: {
1.180 schwarze 2502: struct roffreg *reg;
2503:
2504: /* Search for an existing register with the same name. */
2505: reg = r->regtab;
2506:
2507: while (reg && strcmp(name, reg->key.p))
2508: reg = reg->next;
1.147 kristaps 2509:
1.180 schwarze 2510: if (NULL == reg) {
2511: /* Create a new register. */
2512: reg = mandoc_malloc(sizeof(struct roffreg));
2513: reg->key.p = mandoc_strdup(name);
2514: reg->key.sz = strlen(name);
1.187 schwarze 2515: reg->val = 0;
1.180 schwarze 2516: reg->next = r->regtab;
2517: r->regtab = reg;
2518: }
2519:
1.187 schwarze 2520: if ('+' == sign)
2521: reg->val += val;
2522: else if ('-' == sign)
2523: reg->val -= val;
2524: else
2525: reg->val = val;
1.147 kristaps 2526: }
2527:
1.192 schwarze 2528: /*
2529: * Handle some predefined read-only number registers.
2530: * For now, return -1 if the requested register is not predefined;
2531: * in case a predefined read-only register having the value -1
2532: * were to turn up, another special value would have to be chosen.
2533: */
2534: static int
1.273 schwarze 2535: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2536: {
2537:
2538: switch (*name) {
1.273 schwarze 2539: case '$': /* Number of arguments of the last macro evaluated. */
1.277 schwarze 2540: return r->argc;
1.207 schwarze 2541: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2542: return 0;
1.207 schwarze 2543: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2544: return 1;
1.207 schwarze 2545: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2546: return 24;
1.207 schwarze 2547: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2548: return 0;
1.207 schwarze 2549: case 'T': /* Some output device is always defined. */
1.277 schwarze 2550: return 1;
1.207 schwarze 2551: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2552: return 40;
1.192 schwarze 2553: default:
1.277 schwarze 2554: return -1;
1.192 schwarze 2555: }
2556: }
2557:
1.181 schwarze 2558: int
1.180 schwarze 2559: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2560: {
1.180 schwarze 2561: struct roffreg *reg;
1.192 schwarze 2562: int val;
2563:
2564: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.273 schwarze 2565: val = roff_getregro(r, name + 1);
1.192 schwarze 2566: if (-1 != val)
1.277 schwarze 2567: return val;
1.192 schwarze 2568: }
1.180 schwarze 2569:
2570: for (reg = r->regtab; reg; reg = reg->next)
2571: if (0 == strcmp(name, reg->key.p))
1.277 schwarze 2572: return reg->val;
1.181 schwarze 2573:
1.277 schwarze 2574: return 0;
1.181 schwarze 2575: }
2576:
2577: static int
2578: roff_getregn(const struct roff *r, const char *name, size_t len)
2579: {
2580: struct roffreg *reg;
1.192 schwarze 2581: int val;
2582:
2583: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2584: val = roff_getregro(r, name + 1);
1.192 schwarze 2585: if (-1 != val)
1.277 schwarze 2586: return val;
1.192 schwarze 2587: }
1.181 schwarze 2588:
2589: for (reg = r->regtab; reg; reg = reg->next)
2590: if (len == reg->key.sz &&
2591: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2592: return reg->val;
1.271 schwarze 2593:
1.277 schwarze 2594: return 0;
1.271 schwarze 2595: }
2596:
2597: static int
2598: roff_hasregn(const struct roff *r, const char *name, size_t len)
2599: {
2600: struct roffreg *reg;
2601: int val;
2602:
2603: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2604: val = roff_getregro(r, name + 1);
1.271 schwarze 2605: if (-1 != val)
1.277 schwarze 2606: return 1;
1.271 schwarze 2607: }
2608:
2609: for (reg = r->regtab; reg; reg = reg->next)
2610: if (len == reg->key.sz &&
2611: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2612: return 1;
1.147 kristaps 2613:
1.277 schwarze 2614: return 0;
1.147 kristaps 2615: }
2616:
1.180 schwarze 2617: static void
2618: roff_freereg(struct roffreg *reg)
1.147 kristaps 2619: {
1.180 schwarze 2620: struct roffreg *old_reg;
1.147 kristaps 2621:
1.180 schwarze 2622: while (NULL != reg) {
2623: free(reg->key.p);
2624: old_reg = reg;
2625: reg = reg->next;
2626: free(old_reg);
2627: }
1.147 kristaps 2628: }
1.92 schwarze 2629:
2630: static enum rofferr
1.89 kristaps 2631: roff_nr(ROFF_ARGS)
1.83 schwarze 2632: {
1.212 schwarze 2633: char *key, *val;
2634: size_t keysz;
1.138 kristaps 2635: int iv;
1.187 schwarze 2636: char sign;
1.89 kristaps 2637:
1.238 schwarze 2638: key = val = buf->buf + pos;
2639: if (*key == '\0')
1.277 schwarze 2640: return ROFF_IGN;
1.212 schwarze 2641:
2642: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2643: if (key[keysz] == '\\')
1.277 schwarze 2644: return ROFF_IGN;
1.212 schwarze 2645: key[keysz] = '\0';
1.89 kristaps 2646:
1.187 schwarze 2647: sign = *val;
1.238 schwarze 2648: if (sign == '+' || sign == '-')
1.187 schwarze 2649: val++;
2650:
1.261 schwarze 2651: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204 schwarze 2652: roff_setreg(r, key, iv, sign);
1.109 kristaps 2653:
1.277 schwarze 2654: return ROFF_IGN;
1.203 schwarze 2655: }
2656:
2657: static enum rofferr
2658: roff_rr(ROFF_ARGS)
2659: {
2660: struct roffreg *reg, **prev;
1.212 schwarze 2661: char *name, *cp;
2662: size_t namesz;
1.203 schwarze 2663:
1.238 schwarze 2664: name = cp = buf->buf + pos;
2665: if (*name == '\0')
1.277 schwarze 2666: return ROFF_IGN;
1.212 schwarze 2667: namesz = roff_getname(r, &cp, ln, pos);
2668: name[namesz] = '\0';
1.203 schwarze 2669:
2670: prev = &r->regtab;
2671: while (1) {
2672: reg = *prev;
1.238 schwarze 2673: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2674: break;
2675: prev = ®->next;
2676: }
1.238 schwarze 2677: if (reg != NULL) {
1.203 schwarze 2678: *prev = reg->next;
2679: free(reg->key.p);
2680: free(reg);
2681: }
1.277 schwarze 2682: return ROFF_IGN;
1.122 schwarze 2683: }
2684:
1.266 schwarze 2685: /* --- handler functions for roff requests -------------------------------- */
2686:
1.122 schwarze 2687: static enum rofferr
2688: roff_rm(ROFF_ARGS)
2689: {
2690: const char *name;
2691: char *cp;
1.212 schwarze 2692: size_t namesz;
1.122 schwarze 2693:
1.238 schwarze 2694: cp = buf->buf + pos;
2695: while (*cp != '\0') {
1.212 schwarze 2696: name = cp;
1.238 schwarze 2697: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2698: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238 schwarze 2699: if (name[namesz] == '\\')
1.212 schwarze 2700: break;
1.122 schwarze 2701: }
1.277 schwarze 2702: return ROFF_IGN;
1.178 schwarze 2703: }
2704:
2705: static enum rofferr
2706: roff_it(ROFF_ARGS)
2707: {
2708: int iv;
2709:
2710: /* Parse the number of lines. */
1.261 schwarze 2711:
2712: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2713: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2714: ln, ppos, buf->buf + 1);
1.277 schwarze 2715: return ROFF_IGN;
1.178 schwarze 2716: }
2717:
1.262 schwarze 2718: while (isspace((unsigned char)buf->buf[pos]))
2719: pos++;
2720:
2721: /*
2722: * Arm the input line trap.
2723: * Special-casing "an-trap" is an ugly workaround to cope
2724: * with DocBook stupidly fiddling with man(7) internals.
2725: */
1.261 schwarze 2726:
1.178 schwarze 2727: roffit_lines = iv;
1.262 schwarze 2728: roffit_macro = mandoc_strdup(iv != 1 ||
2729: strcmp(buf->buf + pos, "an-trap") ?
2730: buf->buf + pos : "br");
1.277 schwarze 2731: return ROFF_IGN;
1.175 schwarze 2732: }
2733:
2734: static enum rofferr
2735: roff_Dd(ROFF_ARGS)
2736: {
2737: const char *const *cp;
2738:
1.227 schwarze 2739: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175 schwarze 2740: for (cp = __mdoc_reserved; *cp; cp++)
2741: roff_setstr(r, *cp, NULL, 0);
2742:
1.227 schwarze 2743: if (r->format == 0)
2744: r->format = MPARSE_MDOC;
2745:
1.277 schwarze 2746: return ROFF_CONT;
1.175 schwarze 2747: }
2748:
2749: static enum rofferr
2750: roff_TH(ROFF_ARGS)
2751: {
2752: const char *const *cp;
2753:
1.227 schwarze 2754: if ((r->options & MPARSE_QUICK) == 0)
1.175 schwarze 2755: for (cp = __man_reserved; *cp; cp++)
2756: roff_setstr(r, *cp, NULL, 0);
2757:
1.227 schwarze 2758: if (r->format == 0)
2759: r->format = MPARSE_MAN;
2760:
1.277 schwarze 2761: return ROFF_CONT;
1.109 kristaps 2762: }
2763:
2764: static enum rofferr
2765: roff_TE(ROFF_ARGS)
2766: {
2767:
2768: if (NULL == r->tbl)
1.221 schwarze 2769: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2770: ln, ppos, "TE");
1.258 schwarze 2771: else if ( ! tbl_end(&r->tbl)) {
2772: free(buf->buf);
2773: buf->buf = mandoc_strdup(".sp");
2774: buf->sz = 4;
1.277 schwarze 2775: return ROFF_REPARSE;
1.258 schwarze 2776: }
1.277 schwarze 2777: return ROFF_IGN;
1.112 kristaps 2778: }
2779:
2780: static enum rofferr
2781: roff_T_(ROFF_ARGS)
2782: {
2783:
2784: if (NULL == r->tbl)
1.221 schwarze 2785: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2786: ln, ppos, "T&");
1.112 kristaps 2787: else
1.116 kristaps 2788: tbl_restart(ppos, ln, r->tbl);
1.112 kristaps 2789:
1.277 schwarze 2790: return ROFF_IGN;
1.109 kristaps 2791: }
2792:
1.230 schwarze 2793: /*
2794: * Handle in-line equation delimiters.
2795: */
2796: static enum rofferr
1.238 schwarze 2797: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2798: {
1.233 schwarze 2799: char *cp1, *cp2;
2800: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2801:
1.230 schwarze 2802: /*
2803: * Outside equations, look for an opening delimiter.
2804: * If we are inside an equation, we already know it is
2805: * in-line, or this function wouldn't have been called;
2806: * so look for a closing delimiter.
2807: */
2808:
1.238 schwarze 2809: cp1 = buf->buf + pos;
1.230 schwarze 2810: cp2 = strchr(cp1, r->eqn == NULL ?
2811: r->last_eqn->odelim : r->last_eqn->cdelim);
2812: if (cp2 == NULL)
1.277 schwarze 2813: return ROFF_CONT;
1.230 schwarze 2814:
1.233 schwarze 2815: *cp2++ = '\0';
2816: bef_pr = bef_nl = aft_nl = aft_pr = "";
2817:
2818: /* Handle preceding text, protecting whitespace. */
2819:
1.238 schwarze 2820: if (*buf->buf != '\0') {
1.233 schwarze 2821: if (r->eqn == NULL)
2822: bef_pr = "\\&";
2823: bef_nl = "\n";
2824: }
2825:
2826: /*
2827: * Prepare replacing the delimiter with an equation macro
2828: * and drop leading white space from the equation.
2829: */
1.230 schwarze 2830:
1.233 schwarze 2831: if (r->eqn == NULL) {
2832: while (*cp2 == ' ')
2833: cp2++;
2834: mac = ".EQ";
2835: } else
2836: mac = ".EN";
2837:
2838: /* Handle following text, protecting whitespace. */
2839:
2840: if (*cp2 != '\0') {
2841: aft_nl = "\n";
2842: if (r->eqn != NULL)
2843: aft_pr = "\\&";
2844: }
2845:
2846: /* Do the actual replacement. */
2847:
1.238 schwarze 2848: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2849: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2850: free(buf->buf);
2851: buf->buf = cp1;
1.230 schwarze 2852:
2853: /* Toggle the in-line state of the eqn subsystem. */
2854:
2855: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 2856: return ROFF_REPARSE;
1.151 kristaps 2857: }
2858:
1.235 schwarze 2859: static enum rofferr
2860: roff_EQ(ROFF_ARGS)
1.125 kristaps 2861: {
1.151 kristaps 2862: struct eqn_node *e;
1.125 kristaps 2863:
1.238 schwarze 2864: assert(r->eqn == NULL);
1.235 schwarze 2865: e = eqn_alloc(ppos, ln, r->parse);
1.125 kristaps 2866:
1.230 schwarze 2867: if (r->last_eqn) {
1.125 kristaps 2868: r->last_eqn->next = e;
1.230 schwarze 2869: e->delim = r->last_eqn->delim;
2870: e->odelim = r->last_eqn->odelim;
2871: e->cdelim = r->last_eqn->cdelim;
2872: } else
1.125 kristaps 2873: r->first_eqn = r->last_eqn = e;
2874:
2875: r->eqn = r->last_eqn = e;
1.151 kristaps 2876:
1.238 schwarze 2877: if (buf->buf[pos] != '\0')
1.235 schwarze 2878: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2879: ".EQ %s", buf->buf + pos);
1.151 kristaps 2880:
1.277 schwarze 2881: return ROFF_IGN;
1.125 kristaps 2882: }
2883:
2884: static enum rofferr
2885: roff_EN(ROFF_ARGS)
2886: {
2887:
1.221 schwarze 2888: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.277 schwarze 2889: return ROFF_IGN;
1.125 kristaps 2890: }
2891:
2892: static enum rofferr
1.109 kristaps 2893: roff_TS(ROFF_ARGS)
2894: {
1.176 schwarze 2895: struct tbl_node *tbl;
1.89 kristaps 2896:
1.115 kristaps 2897: if (r->tbl) {
1.221 schwarze 2898: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2899: ln, ppos, "TS breaks TS");
1.151 kristaps 2900: tbl_end(&r->tbl);
1.115 kristaps 2901: }
1.83 schwarze 2902:
1.176 schwarze 2903: tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2904:
2905: if (r->last_tbl)
1.176 schwarze 2906: r->last_tbl->next = tbl;
1.113 kristaps 2907: else
1.176 schwarze 2908: r->first_tbl = r->last_tbl = tbl;
1.113 kristaps 2909:
1.176 schwarze 2910: r->tbl = r->last_tbl = tbl;
1.277 schwarze 2911: return ROFF_IGN;
1.251 schwarze 2912: }
2913:
2914: static enum rofferr
2915: roff_brp(ROFF_ARGS)
2916: {
2917:
2918: buf->buf[pos - 1] = '\0';
1.277 schwarze 2919: return ROFF_CONT;
1.92 schwarze 2920: }
2921:
1.105 kristaps 2922: static enum rofferr
1.174 kristaps 2923: roff_cc(ROFF_ARGS)
2924: {
2925: const char *p;
2926:
1.238 schwarze 2927: p = buf->buf + pos;
1.174 kristaps 2928:
1.238 schwarze 2929: if (*p == '\0' || (r->control = *p++) == '.')
1.174 kristaps 2930: r->control = 0;
2931:
1.238 schwarze 2932: if (*p != '\0')
1.260 schwarze 2933: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2934: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 2935:
1.277 schwarze 2936: return ROFF_IGN;
1.174 kristaps 2937: }
2938:
2939: static enum rofferr
1.164 kristaps 2940: roff_tr(ROFF_ARGS)
2941: {
2942: const char *p, *first, *second;
2943: size_t fsz, ssz;
2944: enum mandoc_esc esc;
2945:
1.238 schwarze 2946: p = buf->buf + pos;
1.164 kristaps 2947:
1.238 schwarze 2948: if (*p == '\0') {
1.260 schwarze 2949: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 2950: return ROFF_IGN;
1.164 kristaps 2951: }
2952:
1.238 schwarze 2953: while (*p != '\0') {
1.164 kristaps 2954: fsz = ssz = 1;
2955:
2956: first = p++;
1.238 schwarze 2957: if (*first == '\\') {
1.164 kristaps 2958: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2959: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2960: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2961: ln, (int)(p - buf->buf), first);
1.277 schwarze 2962: return ROFF_IGN;
1.164 kristaps 2963: }
2964: fsz = (size_t)(p - first);
2965: }
2966:
2967: second = p++;
1.238 schwarze 2968: if (*second == '\\') {
1.164 kristaps 2969: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2970: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2971: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2972: ln, (int)(p - buf->buf), second);
1.277 schwarze 2973: return ROFF_IGN;
1.164 kristaps 2974: }
2975: ssz = (size_t)(p - second);
1.238 schwarze 2976: } else if (*second == '\0') {
1.260 schwarze 2977: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2978: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 2979: second = " ";
1.165 kristaps 2980: p--;
1.164 kristaps 2981: }
2982:
1.167 kristaps 2983: if (fsz > 1) {
1.207 schwarze 2984: roff_setstrn(&r->xmbtab, first, fsz,
2985: second, ssz, 0);
1.167 kristaps 2986: continue;
2987: }
2988:
1.238 schwarze 2989: if (r->xtab == NULL)
1.207 schwarze 2990: r->xtab = mandoc_calloc(128,
2991: sizeof(struct roffstr));
1.167 kristaps 2992:
2993: free(r->xtab[(int)*first].p);
2994: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2995: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 2996: }
2997:
1.277 schwarze 2998: return ROFF_IGN;
1.164 kristaps 2999: }
3000:
3001: static enum rofferr
1.105 kristaps 3002: roff_so(ROFF_ARGS)
3003: {
1.249 schwarze 3004: char *name, *cp;
1.105 kristaps 3005:
1.238 schwarze 3006: name = buf->buf + pos;
1.224 schwarze 3007: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3008:
3009: /*
3010: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3011: * opening anything that's not in our cwd or anything beneath
3012: * it. Thus, explicitly disallow traversing up the file-system
3013: * or using absolute paths.
3014: */
3015:
1.238 schwarze 3016: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3017: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3018: ".so %s", name);
1.249 schwarze 3019: buf->sz = mandoc_asprintf(&cp,
3020: ".sp\nSee the file %s.\n.sp", name) + 1;
3021: free(buf->buf);
3022: buf->buf = cp;
3023: *offs = 0;
1.277 schwarze 3024: return ROFF_REPARSE;
1.105 kristaps 3025: }
3026:
3027: *offs = pos;
1.277 schwarze 3028: return ROFF_SO;
1.105 kristaps 3029: }
1.92 schwarze 3030:
1.266 schwarze 3031: /* --- user defined strings and macros ------------------------------------ */
3032:
1.106 kristaps 3033: static enum rofferr
3034: roff_userdef(ROFF_ARGS)
1.99 kristaps 3035: {
1.263 schwarze 3036: const char *arg[9], *ap;
1.106 kristaps 3037: char *cp, *n1, *n2;
1.274 schwarze 3038: int i, ib, ie;
1.263 schwarze 3039: size_t asz, rsz;
1.106 kristaps 3040:
3041: /*
3042: * Collect pointers to macro argument strings
1.188 schwarze 3043: * and NUL-terminate them.
1.106 kristaps 3044: */
1.263 schwarze 3045:
1.273 schwarze 3046: r->argc = 0;
1.238 schwarze 3047: cp = buf->buf + pos;
1.273 schwarze 3048: for (i = 0; i < 9; i++) {
3049: if (*cp == '\0')
3050: arg[i] = "";
3051: else {
3052: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3053: r->argc = i + 1;
3054: }
3055: }
1.99 kristaps 3056:
1.106 kristaps 3057: /*
3058: * Expand macro arguments.
1.99 kristaps 3059: */
1.263 schwarze 3060:
3061: buf->sz = strlen(r->current_string) + 1;
3062: n1 = cp = mandoc_malloc(buf->sz);
3063: memcpy(n1, r->current_string, buf->sz);
3064: while (*cp != '\0') {
3065:
3066: /* Scan ahead for the next argument invocation. */
3067:
3068: if (*cp++ != '\\')
3069: continue;
3070: if (*cp++ != '$')
3071: continue;
1.274 schwarze 3072: if (*cp == '*') { /* \\$* inserts all arguments */
3073: ib = 0;
3074: ie = r->argc - 1;
3075: } else { /* \\$1 .. \\$9 insert one argument */
3076: ib = ie = *cp - '1';
3077: if (ib < 0 || ib > 8)
3078: continue;
3079: }
1.263 schwarze 3080: cp -= 2;
3081:
3082: /*
3083: * Determine the size of the expanded argument,
3084: * taking escaping of quotes into account.
3085: */
3086:
1.274 schwarze 3087: asz = ie > ib ? ie - ib : 0; /* for blanks */
3088: for (i = ib; i <= ie; i++) {
3089: for (ap = arg[i]; *ap != '\0'; ap++) {
3090: asz++;
3091: if (*ap == '"')
3092: asz += 3;
3093: }
1.263 schwarze 3094: }
3095: if (asz != 3) {
3096:
3097: /*
3098: * Determine the size of the rest of the
3099: * unexpanded macro, including the NUL.
3100: */
3101:
3102: rsz = buf->sz - (cp - n1) - 3;
3103:
3104: /*
3105: * When shrinking, move before
3106: * releasing the storage.
3107: */
3108:
3109: if (asz < 3)
3110: memmove(cp + asz, cp + 3, rsz);
3111:
3112: /*
3113: * Resize the storage for the macro
3114: * and readjust the parse pointer.
3115: */
3116:
3117: buf->sz += asz - 3;
3118: n2 = mandoc_realloc(n1, buf->sz);
3119: cp = n2 + (cp - n1);
3120: n1 = n2;
3121:
3122: /*
3123: * When growing, make room
3124: * for the expanded argument.
3125: */
3126:
3127: if (asz > 3)
3128: memmove(cp + asz, cp + 3, rsz);
3129: }
3130:
3131: /* Copy the expanded argument, escaping quotes. */
3132:
3133: n2 = cp;
1.274 schwarze 3134: for (i = ib; i <= ie; i++) {
3135: for (ap = arg[i]; *ap != '\0'; ap++) {
3136: if (*ap == '"') {
3137: memcpy(n2, "\\(dq", 4);
3138: n2 += 4;
3139: } else
3140: *n2++ = *ap;
3141: }
3142: if (i < ie)
3143: *n2++ = ' ';
1.106 kristaps 3144: }
1.99 kristaps 3145: }
3146:
1.106 kristaps 3147: /*
3148: * Replace the macro invocation
3149: * by the expanded macro.
3150: */
1.263 schwarze 3151:
1.238 schwarze 3152: free(buf->buf);
3153: buf->buf = n1;
1.248 schwarze 3154: *offs = 0;
1.106 kristaps 3155:
1.277 schwarze 3156: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3157: ROFF_REPARSE : ROFF_APPEND;
1.99 kristaps 3158: }
1.121 schwarze 3159:
1.212 schwarze 3160: static size_t
1.121 schwarze 3161: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3162: {
3163: char *name, *cp;
1.212 schwarze 3164: size_t namesz;
1.121 schwarze 3165:
3166: name = *cpp;
3167: if ('\0' == *name)
1.277 schwarze 3168: return 0;
1.121 schwarze 3169:
1.212 schwarze 3170: /* Read until end of name and terminate it with NUL. */
3171: for (cp = name; 1; cp++) {
3172: if ('\0' == *cp || ' ' == *cp) {
3173: namesz = cp - name;
3174: break;
3175: }
1.121 schwarze 3176: if ('\\' != *cp)
3177: continue;
1.215 schwarze 3178: namesz = cp - name;
3179: if ('{' == cp[1] || '}' == cp[1])
3180: break;
1.121 schwarze 3181: cp++;
3182: if ('\\' == *cp)
3183: continue;
1.224 schwarze 3184: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3185: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3186: mandoc_escape((const char **)&cp, NULL, NULL);
3187: break;
1.121 schwarze 3188: }
3189:
3190: /* Read past spaces. */
3191: while (' ' == *cp)
3192: cp++;
3193:
3194: *cpp = cp;
1.277 schwarze 3195: return namesz;
1.121 schwarze 3196: }
3197:
1.106 kristaps 3198: /*
3199: * Store *string into the user-defined string called *name.
3200: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3201: * append == 0: replace mode
3202: * append == 1: single-line append mode
3203: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3204: */
1.94 kristaps 3205: static void
1.106 kristaps 3206: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3207: int append)
1.92 schwarze 3208: {
1.164 kristaps 3209:
3210: roff_setstrn(&r->strtab, name, strlen(name), string,
1.207 schwarze 3211: string ? strlen(string) : 0, append);
1.164 kristaps 3212: }
3213:
3214: static void
1.166 kristaps 3215: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3216: const char *string, size_t stringsz, int append)
1.164 kristaps 3217: {
1.166 kristaps 3218: struct roffkv *n;
1.164 kristaps 3219: char *c;
3220: int i;
3221: size_t oldch, newch;
1.92 schwarze 3222:
1.106 kristaps 3223: /* Search for an existing string with the same name. */
1.164 kristaps 3224: n = *r;
3225:
1.211 schwarze 3226: while (n && (namesz != n->key.sz ||
3227: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3228: n = n->next;
1.94 kristaps 3229:
3230: if (NULL == n) {
1.106 kristaps 3231: /* Create a new string table entry. */
1.166 kristaps 3232: n = mandoc_malloc(sizeof(struct roffkv));
3233: n->key.p = mandoc_strndup(name, namesz);
3234: n->key.sz = namesz;
3235: n->val.p = NULL;
3236: n->val.sz = 0;
1.164 kristaps 3237: n->next = *r;
3238: *r = n;
1.193 schwarze 3239: } else if (0 == append) {
1.166 kristaps 3240: free(n->val.p);
3241: n->val.p = NULL;
3242: n->val.sz = 0;
1.106 kristaps 3243: }
3244:
3245: if (NULL == string)
3246: return;
3247:
3248: /*
3249: * One additional byte for the '\n' in multiline mode,
3250: * and one for the terminating '\0'.
3251: */
1.193 schwarze 3252: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3253:
1.166 kristaps 3254: if (NULL == n->val.p) {
3255: n->val.p = mandoc_malloc(newch);
3256: *n->val.p = '\0';
1.106 kristaps 3257: oldch = 0;
3258: } else {
1.166 kristaps 3259: oldch = n->val.sz;
3260: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3261: }
3262:
3263: /* Skip existing content in the destination buffer. */
1.166 kristaps 3264: c = n->val.p + (int)oldch;
1.106 kristaps 3265:
3266: /* Append new content to the destination buffer. */
1.164 kristaps 3267: i = 0;
3268: while (i < (int)stringsz) {
1.106 kristaps 3269: /*
3270: * Rudimentary roff copy mode:
3271: * Handle escaped backslashes.
3272: */
1.164 kristaps 3273: if ('\\' == string[i] && '\\' == string[i + 1])
3274: i++;
3275: *c++ = string[i++];
1.106 kristaps 3276: }
1.94 kristaps 3277:
1.106 kristaps 3278: /* Append terminating bytes. */
1.193 schwarze 3279: if (1 < append)
1.106 kristaps 3280: *c++ = '\n';
1.163 kristaps 3281:
1.106 kristaps 3282: *c = '\0';
1.166 kristaps 3283: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3284: }
3285:
1.94 kristaps 3286: static const char *
3287: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92 schwarze 3288: {
1.166 kristaps 3289: const struct roffkv *n;
1.191 schwarze 3290: int i;
1.92 schwarze 3291:
1.164 kristaps 3292: for (n = r->strtab; n; n = n->next)
1.207 schwarze 3293: if (0 == strncmp(name, n->key.p, len) &&
3294: '\0' == n->key.p[(int)len])
1.277 schwarze 3295: return n->val.p;
1.191 schwarze 3296:
3297: for (i = 0; i < PREDEFS_MAX; i++)
3298: if (0 == strncmp(name, predefs[i].name, len) &&
3299: '\0' == predefs[i].name[(int)len])
1.277 schwarze 3300: return predefs[i].str;
1.94 kristaps 3301:
1.277 schwarze 3302: return NULL;
1.92 schwarze 3303: }
3304:
1.94 kristaps 3305: static void
1.167 kristaps 3306: roff_freestr(struct roffkv *r)
1.92 schwarze 3307: {
1.166 kristaps 3308: struct roffkv *n, *nn;
1.92 schwarze 3309:
1.167 kristaps 3310: for (n = r; n; n = nn) {
1.166 kristaps 3311: free(n->key.p);
3312: free(n->val.p);
1.92 schwarze 3313: nn = n->next;
3314: free(n);
3315: }
1.114 kristaps 3316: }
1.266 schwarze 3317:
3318: /* --- accessors and utility functions ------------------------------------ */
1.114 kristaps 3319:
3320: const struct tbl_span *
3321: roff_span(const struct roff *r)
3322: {
1.207 schwarze 3323:
1.277 schwarze 3324: return r->tbl ? tbl_span(r->tbl) : NULL;
1.125 kristaps 3325: }
3326:
3327: const struct eqn *
3328: roff_eqn(const struct roff *r)
3329: {
1.207 schwarze 3330:
1.277 schwarze 3331: return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.164 kristaps 3332: }
3333:
3334: /*
3335: * Duplicate an input string, making the appropriate character
3336: * conversations (as stipulated by `tr') along the way.
3337: * Returns a heap-allocated string with all the replacements made.
3338: */
3339: char *
3340: roff_strdup(const struct roff *r, const char *p)
3341: {
1.166 kristaps 3342: const struct roffkv *cp;
1.164 kristaps 3343: char *res;
3344: const char *pp;
3345: size_t ssz, sz;
3346: enum mandoc_esc esc;
3347:
1.167 kristaps 3348: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3349: return mandoc_strdup(p);
1.164 kristaps 3350: else if ('\0' == *p)
1.277 schwarze 3351: return mandoc_strdup("");
1.164 kristaps 3352:
3353: /*
3354: * Step through each character looking for term matches
3355: * (remember that a `tr' can be invoked with an escape, which is
3356: * a glyph but the escape is multi-character).
3357: * We only do this if the character hash has been initialised
3358: * and the string is >0 length.
3359: */
3360:
3361: res = NULL;
3362: ssz = 0;
3363:
3364: while ('\0' != *p) {
1.167 kristaps 3365: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3366: sz = r->xtab[(int)*p].sz;
3367: res = mandoc_realloc(res, ssz + sz + 1);
3368: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3369: ssz += sz;
3370: p++;
3371: continue;
3372: } else if ('\\' != *p) {
3373: res = mandoc_realloc(res, ssz + 2);
3374: res[ssz++] = *p++;
3375: continue;
3376: }
3377:
1.164 kristaps 3378: /* Search for term matches. */
1.167 kristaps 3379: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3380: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3381: break;
3382:
3383: if (NULL != cp) {
3384: /*
3385: * A match has been found.
3386: * Append the match to the array and move
3387: * forward by its keysize.
3388: */
1.207 schwarze 3389: res = mandoc_realloc(res,
3390: ssz + cp->val.sz + 1);
1.166 kristaps 3391: memcpy(res + ssz, cp->val.p, cp->val.sz);
3392: ssz += cp->val.sz;
3393: p += (int)cp->key.sz;
1.164 kristaps 3394: continue;
3395: }
3396:
1.167 kristaps 3397: /*
3398: * Handle escapes carefully: we need to copy
3399: * over just the escape itself, or else we might
3400: * do replacements within the escape itself.
3401: * Make sure to pass along the bogus string.
3402: */
3403: pp = p++;
3404: esc = mandoc_escape(&p, NULL, NULL);
3405: if (ESCAPE_ERROR == esc) {
3406: sz = strlen(pp);
1.164 kristaps 3407: res = mandoc_realloc(res, ssz + sz + 1);
3408: memcpy(res + ssz, pp, sz);
1.167 kristaps 3409: break;
1.164 kristaps 3410: }
1.207 schwarze 3411: /*
3412: * We bail out on bad escapes.
1.167 kristaps 3413: * No need to warn: we already did so when
3414: * roff_res() was called.
3415: */
3416: sz = (int)(p - pp);
3417: res = mandoc_realloc(res, ssz + sz + 1);
3418: memcpy(res + ssz, pp, sz);
3419: ssz += sz;
1.164 kristaps 3420: }
3421:
3422: res[(int)ssz] = '\0';
1.277 schwarze 3423: return res;
1.227 schwarze 3424: }
3425:
3426: int
3427: roff_getformat(const struct roff *r)
3428: {
3429:
1.277 schwarze 3430: return r->format;
1.174 kristaps 3431: }
3432:
3433: /*
1.207 schwarze 3434: * Find out whether a line is a macro line or not.
1.174 kristaps 3435: * If it is, adjust the current position and return one; if it isn't,
3436: * return zero and don't change the current position.
3437: * If the control character has been set with `.cc', then let that grain
3438: * precedence.
3439: * This is slighly contrary to groff, where using the non-breaking
3440: * control character when `cc' has been invoked will cause the
3441: * non-breaking macro contents to be printed verbatim.
3442: */
3443: int
3444: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3445: {
3446: int pos;
3447:
3448: pos = *ppos;
3449:
3450: if (0 != r->control && cp[pos] == r->control)
3451: pos++;
3452: else if (0 != r->control)
1.277 schwarze 3453: return 0;
1.174 kristaps 3454: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3455: pos += 2;
3456: else if ('.' == cp[pos] || '\'' == cp[pos])
3457: pos++;
3458: else
1.277 schwarze 3459: return 0;
1.174 kristaps 3460:
3461: while (' ' == cp[pos] || '\t' == cp[pos])
3462: pos++;
3463:
3464: *ppos = pos;
1.277 schwarze 3465: return 1;
1.74 kristaps 3466: }
CVSweb