Annotation of mandoc/roff.c, Revision 1.285
1.285 ! schwarze 1: /* $Id: roff.c,v 1.284 2016/01/08 17:48:10 schwarze Exp $ */
1.1 kristaps 2: /*
1.267 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.247 schwarze 4: * Copyright (c) 2010-2015 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.66 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 kristaps 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66 kristaps 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.66 kristaps 18: #include "config.h"
1.225 schwarze 19:
20: #include <sys/types.h>
1.30 kristaps 21:
1.67 kristaps 22: #include <assert.h>
1.85 kristaps 23: #include <ctype.h>
1.245 schwarze 24: #include <limits.h>
1.178 schwarze 25: #include <stdio.h>
1.1 kristaps 26: #include <stdlib.h>
1.67 kristaps 27: #include <string.h>
1.1 kristaps 28:
1.67 kristaps 29: #include "mandoc.h"
1.201 schwarze 30: #include "mandoc_aux.h"
1.265 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:
1020: man->last->child = n;
1021: n->parent = man->last;
1.281 schwarze 1022: n->parent->last = n;
1.266 schwarze 1023: break;
1024: default:
1025: abort();
1026: }
1.282 schwarze 1027: man->last = n;
1.266 schwarze 1028:
1029: switch (n->type) {
1030: case ROFFT_HEAD:
1031: n->parent->head = n;
1032: break;
1033: case ROFFT_BODY:
1.282 schwarze 1034: if (n->end != ENDBODY_NOT)
1035: return;
1.266 schwarze 1036: n->parent->body = n;
1037: break;
1038: case ROFFT_TAIL:
1039: n->parent->tail = n;
1040: break;
1041: default:
1.282 schwarze 1042: return;
1.266 schwarze 1043: }
1.282 schwarze 1044:
1045: /*
1046: * Copy over the normalised-data pointer of our parent. Not
1047: * everybody has one, but copying a null pointer is fine.
1048: */
1049:
1050: n->norm = n->parent->norm;
1051: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 1052: }
1053:
1.267 schwarze 1054: void
1055: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
1056: {
1057: struct roff_node *n;
1058:
1059: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
1060: n->string = roff_strdup(man->roff, word);
1061: roff_node_append(man, n);
1062: if (man->macroset == MACROSET_MDOC)
1.285 ! schwarze 1063: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 1064: else
1.285 ! schwarze 1065: n->flags |= NODE_VALID;
1.267 schwarze 1066: man->next = ROFF_NEXT_SIBLING;
1067: }
1068:
1069: void
1070: roff_word_append(struct roff_man *man, const char *word)
1071: {
1072: struct roff_node *n;
1073: char *addstr, *newstr;
1074:
1075: n = man->last;
1076: addstr = roff_strdup(man->roff, word);
1077: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1078: free(addstr);
1079: free(n->string);
1080: n->string = newstr;
1081: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 1082: }
1083:
1084: void
1085: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
1086: {
1087: struct roff_node *n;
1088:
1089: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
1090: roff_node_append(man, n);
1091: man->next = ROFF_NEXT_CHILD;
1092: }
1093:
1094: struct roff_node *
1095: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
1096: {
1097: struct roff_node *n;
1098:
1099: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
1100: roff_node_append(man, n);
1101: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1102: return n;
1.267 schwarze 1103: }
1104:
1.266 schwarze 1105: struct roff_node *
1106: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
1107: {
1108: struct roff_node *n;
1109:
1110: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
1111: roff_node_append(man, n);
1112: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1113: return n;
1.266 schwarze 1114: }
1115:
1116: struct roff_node *
1117: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1118: {
1119: struct roff_node *n;
1120:
1121: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1122: roff_node_append(man, n);
1123: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1124: return n;
1.267 schwarze 1125: }
1126:
1127: void
1128: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
1129: {
1130: struct roff_node *n;
1131:
1132: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
1133: n->eqn = eqn;
1134: if (eqn->ln > man->last->line)
1.285 ! schwarze 1135: n->flags |= NODE_LINE;
1.267 schwarze 1136: roff_node_append(man, n);
1137: man->next = ROFF_NEXT_SIBLING;
1138: }
1139:
1140: void
1141: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
1142: {
1143: struct roff_node *n;
1144:
1145: if (man->macroset == MACROSET_MAN)
1146: man_breakscope(man, TOKEN_NONE);
1147: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
1148: n->span = tbl;
1149: roff_node_append(man, n);
1150: if (man->macroset == MACROSET_MDOC)
1.285 ! schwarze 1151: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 1152: else
1.285 ! schwarze 1153: n->flags |= NODE_VALID;
1.267 schwarze 1154: man->next = ROFF_NEXT_SIBLING;
1.266 schwarze 1155: }
1156:
1157: void
1158: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1159: {
1160:
1161: /* Adjust siblings. */
1162:
1163: if (n->prev)
1164: n->prev->next = n->next;
1165: if (n->next)
1166: n->next->prev = n->prev;
1167:
1168: /* Adjust parent. */
1169:
1170: if (n->parent != NULL) {
1171: if (n->parent->child == n)
1172: n->parent->child = n->next;
1173: if (n->parent->last == n)
1174: n->parent->last = n->prev;
1175: }
1176:
1177: /* Adjust parse point. */
1178:
1179: if (man == NULL)
1180: return;
1181: if (man->last == n) {
1182: if (n->prev == NULL) {
1183: man->last = n->parent;
1184: man->next = ROFF_NEXT_CHILD;
1185: } else {
1186: man->last = n->prev;
1187: man->next = ROFF_NEXT_SIBLING;
1188: }
1189: }
1190: if (man->first == n)
1191: man->first = NULL;
1192: }
1193:
1194: void
1195: roff_node_free(struct roff_node *n)
1196: {
1197:
1198: if (n->args != NULL)
1199: mdoc_argv_free(n->args);
1200: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1201: free(n->norm);
1202: free(n->string);
1203: free(n);
1204: }
1205:
1206: void
1207: roff_node_delete(struct roff_man *man, struct roff_node *n)
1208: {
1209:
1210: while (n->child != NULL)
1211: roff_node_delete(man, n->child);
1212: roff_node_unlink(man, n);
1213: roff_node_free(n);
1.269 schwarze 1214: }
1215:
1216: void
1217: deroff(char **dest, const struct roff_node *n)
1218: {
1219: char *cp;
1220: size_t sz;
1221:
1222: if (n->type != ROFFT_TEXT) {
1223: for (n = n->child; n != NULL; n = n->next)
1224: deroff(dest, n);
1225: return;
1226: }
1227:
1228: /* Skip leading whitespace and escape sequences. */
1229:
1230: cp = n->string;
1231: while (*cp != '\0') {
1232: if ('\\' == *cp) {
1233: cp++;
1234: mandoc_escape((const char **)&cp, NULL, NULL);
1235: } else if (isspace((unsigned char)*cp))
1236: cp++;
1237: else
1238: break;
1239: }
1240:
1241: /* Skip trailing whitespace. */
1242:
1243: for (sz = strlen(cp); sz; sz--)
1244: if ( ! isspace((unsigned char)cp[sz-1]))
1245: break;
1246:
1247: /* Skip empty strings. */
1248:
1249: if (sz == 0)
1250: return;
1251:
1252: if (*dest == NULL) {
1253: *dest = mandoc_strndup(cp, sz);
1254: return;
1255: }
1256:
1257: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1258: free(*dest);
1259: *dest = cp;
1.266 schwarze 1260: }
1261:
1262: /* --- main functions of the roff parser ---------------------------------- */
1263:
1.94 kristaps 1264: /*
1.206 schwarze 1265: * In the current line, expand escape sequences that tend to get
1266: * used in numerical expressions and conditional requests.
1267: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1268: */
1.172 schwarze 1269: static enum rofferr
1.238 schwarze 1270: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1271: {
1.208 schwarze 1272: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1273: const char *start; /* start of the string to process */
1.209 schwarze 1274: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1275: const char *stnam; /* start of the name, after "[(*" */
1276: const char *cp; /* end of the name, e.g. before ']' */
1277: const char *res; /* the string to be substituted */
1.238 schwarze 1278: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1279: size_t maxl; /* expected length of the escape name */
1280: size_t naml; /* actual length of the escape name */
1.237 schwarze 1281: enum mandoc_esc esc; /* type of the escape sequence */
1282: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1283: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1284: int npos; /* position in numeric expression */
1.218 schwarze 1285: int arg_complete; /* argument not interrupted by eol */
1.206 schwarze 1286: char term; /* character terminating the escape */
1.94 kristaps 1287:
1.170 schwarze 1288: expand_count = 0;
1.238 schwarze 1289: start = buf->buf + pos;
1.205 schwarze 1290: stesc = strchr(start, '\0') - 1;
1291: while (stesc-- > start) {
1.170 schwarze 1292:
1.205 schwarze 1293: /* Search backwards for the next backslash. */
1294:
1.238 schwarze 1295: if (*stesc != '\\')
1.205 schwarze 1296: continue;
1297:
1298: /* If it is escaped, skip it. */
1299:
1300: for (cp = stesc - 1; cp >= start; cp--)
1.238 schwarze 1301: if (*cp != '\\')
1.205 schwarze 1302: break;
1303:
1.238 schwarze 1304: if ((stesc - cp) % 2 == 0) {
1.209 schwarze 1305: stesc = (char *)cp;
1.205 schwarze 1306: continue;
1307: }
1.108 schwarze 1308:
1.206 schwarze 1309: /* Decide whether to expand or to check only. */
1.108 schwarze 1310:
1.206 schwarze 1311: term = '\0';
1.205 schwarze 1312: cp = stesc + 1;
1.181 schwarze 1313: switch (*cp) {
1.207 schwarze 1314: case '*':
1.181 schwarze 1315: res = NULL;
1316: break;
1.207 schwarze 1317: case 'B':
1318: case 'w':
1.206 schwarze 1319: term = cp[1];
1320: /* FALLTHROUGH */
1.207 schwarze 1321: case 'n':
1.181 schwarze 1322: res = ubuf;
1323: break;
1324: default:
1.237 schwarze 1325: esc = mandoc_escape(&cp, &stnam, &inaml);
1326: if (esc == ESCAPE_ERROR ||
1327: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1328: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1329: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1330: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1331: "%.*s", (int)(cp - stesc), stesc);
1.205 schwarze 1332: continue;
1.152 kristaps 1333: }
1334:
1.205 schwarze 1335: if (EXPAND_LIMIT < ++expand_count) {
1336: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1337: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1338: return ROFF_IGN;
1.205 schwarze 1339: }
1.108 schwarze 1340:
1341: /*
1342: * The third character decides the length
1.181 schwarze 1343: * of the name of the string or register.
1.108 schwarze 1344: * Save a pointer to the name.
1345: */
1346:
1.238 schwarze 1347: if (term == '\0') {
1.206 schwarze 1348: switch (*++cp) {
1.207 schwarze 1349: case '\0':
1.206 schwarze 1350: maxl = 0;
1351: break;
1.207 schwarze 1352: case '(':
1.206 schwarze 1353: cp++;
1354: maxl = 2;
1355: break;
1.207 schwarze 1356: case '[':
1.206 schwarze 1357: cp++;
1358: term = ']';
1359: maxl = 0;
1360: break;
1361: default:
1362: maxl = 1;
1363: break;
1364: }
1365: } else {
1366: cp += 2;
1.94 kristaps 1367: maxl = 0;
1368: }
1.108 schwarze 1369: stnam = cp;
1.94 kristaps 1370:
1.108 schwarze 1371: /* Advance to the end of the name. */
1.94 kristaps 1372:
1.253 schwarze 1373: naml = 0;
1.218 schwarze 1374: arg_complete = 1;
1.253 schwarze 1375: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1376: if (*cp == '\0') {
1.219 schwarze 1377: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1378: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1379: arg_complete = 0;
1.206 schwarze 1380: break;
1.153 kristaps 1381: }
1.238 schwarze 1382: if (maxl == 0 && *cp == term) {
1.206 schwarze 1383: cp++;
1.253 schwarze 1384: break;
1385: }
1386: if (*cp++ != '\\' || stesc[1] != 'w') {
1387: naml++;
1388: continue;
1389: }
1390: switch (mandoc_escape(&cp, NULL, NULL)) {
1391: case ESCAPE_SPECIAL:
1392: case ESCAPE_UNICODE:
1393: case ESCAPE_NUMBERED:
1394: case ESCAPE_OVERSTRIKE:
1395: naml++;
1396: break;
1397: default:
1.94 kristaps 1398: break;
1.206 schwarze 1399: }
1.94 kristaps 1400: }
1401:
1.108 schwarze 1402: /*
1403: * Retrieve the replacement string; if it is
1404: * undefined, resume searching for escapes.
1405: */
1406:
1.206 schwarze 1407: switch (stesc[1]) {
1.207 schwarze 1408: case '*':
1.218 schwarze 1409: if (arg_complete)
1410: res = roff_getstrn(r, stnam, naml);
1.206 schwarze 1411: break;
1.207 schwarze 1412: case 'B':
1.206 schwarze 1413: npos = 0;
1.218 schwarze 1414: ubuf[0] = arg_complete &&
1.261 schwarze 1415: roff_evalnum(r, ln, stnam, &npos,
1416: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1417: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1418: ubuf[1] = '\0';
1419: break;
1.207 schwarze 1420: case 'n':
1.218 schwarze 1421: if (arg_complete)
1422: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1423: roff_getregn(r, stnam, naml));
1424: else
1425: ubuf[0] = '\0';
1.206 schwarze 1426: break;
1.207 schwarze 1427: case 'w':
1.218 schwarze 1428: /* use even incomplete args */
1.208 schwarze 1429: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1430: 24 * (int)naml);
1431: break;
1432: }
1.94 kristaps 1433:
1.238 schwarze 1434: if (res == NULL) {
1.219 schwarze 1435: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1436: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1437: "%.*s", (int)naml, stnam);
1.142 kristaps 1438: res = "";
1.246 schwarze 1439: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1440: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1441: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1442: return ROFF_IGN;
1.94 kristaps 1443: }
1444:
1.108 schwarze 1445: /* Replace the escape sequence by the string. */
1446:
1.209 schwarze 1447: *stesc = '\0';
1.238 schwarze 1448: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1449: buf->buf, res, cp) + 1;
1.94 kristaps 1450:
1.205 schwarze 1451: /* Prepare for the next replacement. */
1.94 kristaps 1452:
1.205 schwarze 1453: start = nbuf + pos;
1.238 schwarze 1454: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1455: free(buf->buf);
1456: buf->buf = nbuf;
1.154 kristaps 1457: }
1.277 schwarze 1458: return ROFF_CONT;
1.154 kristaps 1459: }
1460:
1461: /*
1.275 schwarze 1462: * Process text streams.
1.154 kristaps 1463: */
1464: static enum rofferr
1.238 schwarze 1465: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154 kristaps 1466: {
1467: size_t sz;
1468: const char *start;
1.178 schwarze 1469: char *p;
1470: int isz;
1.154 kristaps 1471: enum mandoc_esc esc;
1472:
1.275 schwarze 1473: /* Spring the input line trap. */
1474:
1475: if (roffit_lines == 1) {
1476: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1477: free(buf->buf);
1478: buf->buf = p;
1479: buf->sz = isz + 1;
1480: *offs = 0;
1481: free(roffit_macro);
1482: roffit_lines = 0;
1.277 schwarze 1483: return ROFF_REPARSE;
1.275 schwarze 1484: } else if (roffit_lines > 1)
1485: --roffit_lines;
1486:
1487: /* Convert all breakable hyphens into ASCII_HYPH. */
1488:
1.238 schwarze 1489: start = p = buf->buf + pos;
1.154 kristaps 1490:
1.238 schwarze 1491: while (*p != '\0') {
1.154 kristaps 1492: sz = strcspn(p, "-\\");
1493: p += sz;
1494:
1.238 schwarze 1495: if (*p == '\0')
1.159 kristaps 1496: break;
1497:
1.238 schwarze 1498: if (*p == '\\') {
1.154 kristaps 1499: /* Skip over escapes. */
1500: p++;
1.189 schwarze 1501: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1502: if (esc == ESCAPE_ERROR)
1.154 kristaps 1503: break;
1.264 schwarze 1504: while (*p == '-')
1505: p++;
1.155 kristaps 1506: continue;
1.159 kristaps 1507: } else if (p == start) {
1.158 kristaps 1508: p++;
1.155 kristaps 1509: continue;
1.158 kristaps 1510: }
1.155 kristaps 1511:
1.171 schwarze 1512: if (isalpha((unsigned char)p[-1]) &&
1513: isalpha((unsigned char)p[1]))
1.155 kristaps 1514: *p = ASCII_HYPH;
1515: p++;
1.94 kristaps 1516: }
1.277 schwarze 1517: return ROFF_CONT;
1.94 kristaps 1518: }
1519:
1.67 kristaps 1520: enum rofferr
1.238 schwarze 1521: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1522: {
1523: enum rofft t;
1.109 kristaps 1524: enum rofferr e;
1.238 schwarze 1525: int pos; /* parse point */
1.243 schwarze 1526: int spos; /* saved parse point for messages */
1.238 schwarze 1527: int ppos; /* original offset in buf->buf */
1528: int ctl; /* macro line (boolean) */
1529:
1530: ppos = pos = *offs;
1.79 kristaps 1531:
1.230 schwarze 1532: /* Handle in-line equation delimiters. */
1533:
1.236 schwarze 1534: if (r->tbl == NULL &&
1535: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1536: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1537: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1538: if (e == ROFF_REPARSE)
1.277 schwarze 1539: return e;
1.230 schwarze 1540: assert(e == ROFF_CONT);
1541: }
1542:
1543: /* Expand some escape sequences. */
1.94 kristaps 1544:
1.238 schwarze 1545: e = roff_res(r, buf, ln, pos);
1546: if (e == ROFF_IGN)
1.277 schwarze 1547: return e;
1.238 schwarze 1548: assert(e == ROFF_CONT);
1.94 kristaps 1549:
1.238 schwarze 1550: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1551:
1.94 kristaps 1552: /*
1.79 kristaps 1553: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1554: * text through the macro's filter.
1555: * Equations process all content themselves.
1556: * Tables process almost all content themselves, but we want
1557: * to warn about macros before passing it there.
1.79 kristaps 1558: */
1.74 kristaps 1559:
1.252 schwarze 1560: if (r->last != NULL && ! ctl) {
1.78 kristaps 1561: t = r->last->tok;
1562: assert(roffs[t].text);
1.238 schwarze 1563: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1564: assert(e == ROFF_IGN || e == ROFF_CONT);
1565: if (e != ROFF_CONT)
1.277 schwarze 1566: return e;
1.182 schwarze 1567: }
1.252 schwarze 1568: if (r->eqn != NULL)
1.277 schwarze 1569: return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.252 schwarze 1570: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.277 schwarze 1571: return tbl_read(r->tbl, ln, buf->buf, ppos);
1.252 schwarze 1572: if ( ! ctl)
1.277 schwarze 1573: return roff_parsetext(buf, pos, offs);
1.228 schwarze 1574:
1575: /* Skip empty request lines. */
1576:
1.238 schwarze 1577: if (buf->buf[pos] == '"') {
1.228 schwarze 1578: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1579: ln, pos, NULL);
1.277 schwarze 1580: return ROFF_IGN;
1.238 schwarze 1581: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1582: return ROFF_IGN;
1.67 kristaps 1583:
1.79 kristaps 1584: /*
1585: * If a scope is open, go to the child handler for that macro,
1586: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1587: * Don't do so if an equation is open.
1.79 kristaps 1588: */
1.78 kristaps 1589:
1.79 kristaps 1590: if (r->last) {
1591: t = r->last->tok;
1592: assert(roffs[t].sub);
1.277 schwarze 1593: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1594: }
1.78 kristaps 1595:
1.243 schwarze 1596: /* No scope is open. This is a new request or macro. */
1597:
1598: spos = pos;
1599: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1600:
1601: /* Tables ignore most macros. */
1602:
1603: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1604: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1605: ln, pos, buf->buf + spos);
1.257 schwarze 1606: if (t == ROFF_TS)
1.277 schwarze 1607: return ROFF_IGN;
1.257 schwarze 1608: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1609: pos++;
1610: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1611: pos++;
1.277 schwarze 1612: return tbl_read(r->tbl, ln, buf->buf, pos);
1.243 schwarze 1613: }
1614:
1.79 kristaps 1615: /*
1.243 schwarze 1616: * This is neither a roff request nor a user-defined macro.
1617: * Let the standard macro set parsers handle it.
1.79 kristaps 1618: */
1.67 kristaps 1619:
1.243 schwarze 1620: if (t == ROFF_MAX)
1.277 schwarze 1621: return ROFF_CONT;
1.243 schwarze 1622:
1623: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1624:
1.75 kristaps 1625: assert(roffs[t].proc);
1.277 schwarze 1626: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.74 kristaps 1627: }
1628:
1.117 kristaps 1629: void
1.74 kristaps 1630: roff_endparse(struct roff *r)
1631: {
1632:
1.110 kristaps 1633: if (r->last)
1.221 schwarze 1634: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1635: r->last->line, r->last->col,
1636: roffs[r->last->tok].name);
1.117 kristaps 1637:
1.125 kristaps 1638: if (r->eqn) {
1.221 schwarze 1639: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1640: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151 kristaps 1641: eqn_end(&r->eqn);
1.125 kristaps 1642: }
1643:
1.117 kristaps 1644: if (r->tbl) {
1.221 schwarze 1645: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1646: r->tbl->line, r->tbl->pos, "TS");
1.151 kristaps 1647: tbl_end(&r->tbl);
1.117 kristaps 1648: }
1.67 kristaps 1649: }
1650:
1651: /*
1652: * Parse a roff node's type from the input buffer. This must be in the
1653: * form of ".foo xxx" in the usual way.
1654: */
1655: static enum rofft
1.214 schwarze 1656: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1657: {
1.214 schwarze 1658: char *cp;
1.106 kristaps 1659: const char *mac;
1660: size_t maclen;
1.67 kristaps 1661: enum rofft t;
1662:
1.214 schwarze 1663: cp = buf + *pos;
1664:
1665: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.277 schwarze 1666: return ROFF_MAX;
1.67 kristaps 1667:
1.214 schwarze 1668: mac = cp;
1669: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1670:
1.106 kristaps 1671: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.155 kristaps 1672: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.67 kristaps 1673:
1.214 schwarze 1674: if (ROFF_MAX != t)
1675: *pos = cp - buf;
1.67 kristaps 1676:
1.277 schwarze 1677: return t;
1.67 kristaps 1678: }
1679:
1.266 schwarze 1680: /* --- handling of request blocks ----------------------------------------- */
1681:
1.67 kristaps 1682: static enum rofferr
1.76 kristaps 1683: roff_cblock(ROFF_ARGS)
1.67 kristaps 1684: {
1685:
1.79 kristaps 1686: /*
1687: * A block-close `..' should only be invoked as a child of an
1688: * ignore macro, otherwise raise a warning and just ignore it.
1689: */
1690:
1.238 schwarze 1691: if (r->last == NULL) {
1.221 schwarze 1692: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1693: ln, ppos, "..");
1.277 schwarze 1694: return ROFF_IGN;
1.76 kristaps 1695: }
1.67 kristaps 1696:
1.81 kristaps 1697: switch (r->last->tok) {
1.207 schwarze 1698: case ROFF_am:
1.220 schwarze 1699: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1700: case ROFF_ami:
1701: case ROFF_de:
1.108 schwarze 1702: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1703: case ROFF_dei:
1704: case ROFF_ig:
1.81 kristaps 1705: break;
1706: default:
1.221 schwarze 1707: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1708: ln, ppos, "..");
1.277 schwarze 1709: return ROFF_IGN;
1.76 kristaps 1710: }
1.67 kristaps 1711:
1.238 schwarze 1712: if (buf->buf[pos] != '\0')
1.217 schwarze 1713: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1714: ".. %s", buf->buf + pos);
1.71 kristaps 1715:
1716: roffnode_pop(r);
1.76 kristaps 1717: roffnode_cleanscope(r);
1.277 schwarze 1718: return ROFF_IGN;
1.71 kristaps 1719:
1.67 kristaps 1720: }
1721:
1.76 kristaps 1722: static void
1723: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1724: {
1725:
1.76 kristaps 1726: while (r->last) {
1.173 schwarze 1727: if (--r->last->endspan != 0)
1.76 kristaps 1728: break;
1729: roffnode_pop(r);
1730: }
1.67 kristaps 1731: }
1732:
1.195 schwarze 1733: static void
1734: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1735: {
1736:
1.76 kristaps 1737: if (NULL == r->last) {
1.221 schwarze 1738: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1739: ln, ppos, "\\}");
1.195 schwarze 1740: return;
1.76 kristaps 1741: }
1742:
1.82 kristaps 1743: switch (r->last->tok) {
1.207 schwarze 1744: case ROFF_el:
1745: case ROFF_ie:
1746: case ROFF_if:
1.82 kristaps 1747: break;
1748: default:
1.221 schwarze 1749: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1750: ln, ppos, "\\}");
1.195 schwarze 1751: return;
1.75 kristaps 1752: }
1753:
1.76 kristaps 1754: if (r->last->endspan > -1) {
1.221 schwarze 1755: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1756: ln, ppos, "\\}");
1.195 schwarze 1757: return;
1.76 kristaps 1758: }
1759:
1.75 kristaps 1760: roffnode_pop(r);
1.76 kristaps 1761: roffnode_cleanscope(r);
1.195 schwarze 1762: return;
1.76 kristaps 1763: }
1764:
1765: static enum rofferr
1.80 kristaps 1766: roff_block(ROFF_ARGS)
1.76 kristaps 1767: {
1.220 schwarze 1768: const char *name;
1769: char *iname, *cp;
1.213 schwarze 1770: size_t namesz;
1.106 kristaps 1771:
1.220 schwarze 1772: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1773:
1.238 schwarze 1774: if (tok == ROFF_de1)
1.220 schwarze 1775: tok = ROFF_de;
1.251 schwarze 1776: else if (tok == ROFF_dei1)
1777: tok = ROFF_dei;
1.238 schwarze 1778: else if (tok == ROFF_am1)
1.220 schwarze 1779: tok = ROFF_am;
1.251 schwarze 1780: else if (tok == ROFF_ami1)
1781: tok = ROFF_ami;
1.220 schwarze 1782:
1783: /* Parse the macro name argument. */
1784:
1.238 schwarze 1785: cp = buf->buf + pos;
1786: if (tok == ROFF_ig) {
1.220 schwarze 1787: iname = NULL;
1788: namesz = 0;
1789: } else {
1790: iname = cp;
1791: namesz = roff_getname(r, &cp, ln, ppos);
1792: iname[namesz] = '\0';
1793: }
1.107 kristaps 1794:
1.220 schwarze 1795: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1796:
1.238 schwarze 1797: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1798: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1799: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1800: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1801: "%.*s", (int)namesz, iname);
1802: namesz = 0;
1803: } else
1804: namesz = strlen(name);
1805: } else
1806: name = iname;
1.107 kristaps 1807:
1.238 schwarze 1808: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1809: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1810: ln, ppos, roffs[tok].name);
1.277 schwarze 1811: return ROFF_IGN;
1.220 schwarze 1812: }
1.80 kristaps 1813:
1.106 kristaps 1814: roffnode_push(r, tok, name, ln, ppos);
1815:
1816: /*
1817: * At the beginning of a `de' macro, clear the existing string
1818: * with the same name, if there is one. New content will be
1.193 schwarze 1819: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1820: */
1.107 kristaps 1821:
1.238 schwarze 1822: if (tok == ROFF_de || tok == ROFF_dei)
1.213 schwarze 1823: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76 kristaps 1824:
1.238 schwarze 1825: if (*cp == '\0')
1.277 schwarze 1826: return ROFF_IGN;
1.78 kristaps 1827:
1.220 schwarze 1828: /* Get the custom end marker. */
1.107 kristaps 1829:
1.220 schwarze 1830: iname = cp;
1.213 schwarze 1831: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1832:
1833: /* Resolve the end marker if it is indirect. */
1834:
1.238 schwarze 1835: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1836: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1837: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1838: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1839: "%.*s", (int)namesz, iname);
1840: namesz = 0;
1841: } else
1842: namesz = strlen(name);
1843: } else
1844: name = iname;
1845:
1.213 schwarze 1846: if (namesz)
1847: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1848:
1.238 schwarze 1849: if (*cp != '\0')
1.217 schwarze 1850: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1851: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.74 kristaps 1852:
1.277 schwarze 1853: return ROFF_IGN;
1.78 kristaps 1854: }
1855:
1856: static enum rofferr
1.80 kristaps 1857: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1858: {
1859: enum rofft t;
1860: int i, j;
1861:
1862: /*
1863: * First check whether a custom macro exists at this level. If
1864: * it does, then check against it. This is some of groff's
1865: * stranger behaviours. If we encountered a custom end-scope
1866: * tag and that tag also happens to be a "real" macro, then we
1867: * need to try interpreting it again as a real macro. If it's
1868: * not, then return ignore. Else continue.
1869: */
1870:
1871: if (r->last->end) {
1.130 kristaps 1872: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1873: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1874: break;
1875:
1.238 schwarze 1876: if (r->last->end[j] == '\0' &&
1877: (buf->buf[i] == '\0' ||
1878: buf->buf[i] == ' ' ||
1879: buf->buf[i] == '\t')) {
1.79 kristaps 1880: roffnode_pop(r);
1881: roffnode_cleanscope(r);
1882:
1.238 schwarze 1883: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1884: i++;
1885:
1886: pos = i;
1.238 schwarze 1887: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1888: ROFF_MAX)
1.277 schwarze 1889: return ROFF_RERUN;
1890: return ROFF_IGN;
1.79 kristaps 1891: }
1892: }
1893:
1894: /*
1895: * If we have no custom end-query or lookup failed, then try
1896: * pulling it out of the hashtable.
1897: */
1898:
1.238 schwarze 1899: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1900:
1.238 schwarze 1901: if (t != ROFF_cblock) {
1902: if (tok != ROFF_ig)
1903: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 1904: return ROFF_IGN;
1.106 kristaps 1905: }
1.79 kristaps 1906:
1907: assert(roffs[t].proc);
1.277 schwarze 1908: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1909: }
1910:
1911: static enum rofferr
1.80 kristaps 1912: roff_block_text(ROFF_ARGS)
1.78 kristaps 1913: {
1914:
1.238 schwarze 1915: if (tok != ROFF_ig)
1916: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1917:
1.277 schwarze 1918: return ROFF_IGN;
1.78 kristaps 1919: }
1920:
1921: static enum rofferr
1.82 kristaps 1922: roff_cond_sub(ROFF_ARGS)
1923: {
1924: enum rofft t;
1.139 kristaps 1925: char *ep;
1.198 schwarze 1926: int rr;
1.82 kristaps 1927:
1928: rr = r->last->rule;
1.139 kristaps 1929: roffnode_cleanscope(r);
1.238 schwarze 1930: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82 kristaps 1931:
1.139 kristaps 1932: /*
1.177 schwarze 1933: * Fully handle known macros when they are structurally
1934: * required or when the conditional evaluated to true.
1.87 kristaps 1935: */
1936:
1.238 schwarze 1937: if ((t != ROFF_MAX) &&
1938: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.177 schwarze 1939: assert(roffs[t].proc);
1.277 schwarze 1940: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.177 schwarze 1941: }
1.144 kristaps 1942:
1.196 schwarze 1943: /*
1944: * If `\}' occurs on a macro line without a preceding macro,
1945: * drop the line completely.
1946: */
1947:
1.238 schwarze 1948: ep = buf->buf + pos;
1949: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1950: rr = 0;
1.196 schwarze 1951:
1.177 schwarze 1952: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1953:
1.238 schwarze 1954: while ((ep = strchr(ep, '\\')) != NULL) {
1955: if (*(++ep) == '}') {
1.197 schwarze 1956: *ep = '&';
1.238 schwarze 1957: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1958: }
1.247 schwarze 1959: if (*ep != '\0')
1960: ++ep;
1.177 schwarze 1961: }
1.277 schwarze 1962: return rr ? ROFF_CONT : ROFF_IGN;
1.82 kristaps 1963: }
1964:
1965: static enum rofferr
1966: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1967: {
1.140 kristaps 1968: char *ep;
1.198 schwarze 1969: int rr;
1.82 kristaps 1970:
1971: rr = r->last->rule;
1.140 kristaps 1972: roffnode_cleanscope(r);
1.82 kristaps 1973:
1.238 schwarze 1974: ep = buf->buf + pos;
1975: while ((ep = strchr(ep, '\\')) != NULL) {
1976: if (*(++ep) == '}') {
1.197 schwarze 1977: *ep = '&';
1.238 schwarze 1978: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1979: }
1.247 schwarze 1980: if (*ep != '\0')
1981: ++ep;
1.78 kristaps 1982: }
1.277 schwarze 1983: return rr ? ROFF_CONT : ROFF_IGN;
1.74 kristaps 1984: }
1985:
1.266 schwarze 1986: /* --- handling of numeric and conditional expressions -------------------- */
1987:
1.204 schwarze 1988: /*
1989: * Parse a single signed integer number. Stop at the first non-digit.
1990: * If there is at least one digit, return success and advance the
1991: * parse point, else return failure and let the parse point unchanged.
1992: * Ignore overflows, treat them just like the C language.
1993: */
1.184 schwarze 1994: static int
1.261 schwarze 1995: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 1996: {
1.261 schwarze 1997: int myres, scaled, n, p;
1.206 schwarze 1998:
1999: if (NULL == res)
2000: res = &myres;
1.184 schwarze 2001:
2002: p = *pos;
2003: n = v[p] == '-';
1.261 schwarze 2004: if (n || v[p] == '+')
1.184 schwarze 2005: p++;
2006:
1.261 schwarze 2007: if (flags & ROFFNUM_WHITE)
2008: while (isspace((unsigned char)v[p]))
2009: p++;
2010:
1.184 schwarze 2011: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2012: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2013: if (p == *pos + n)
2014: return 0;
2015:
2016: if (n)
2017: *res = -*res;
2018:
1.254 schwarze 2019: /* Each number may be followed by one optional scaling unit. */
2020:
2021: switch (v[p]) {
2022: case 'f':
1.261 schwarze 2023: scaled = *res * 65536;
1.254 schwarze 2024: break;
2025: case 'i':
1.261 schwarze 2026: scaled = *res * 240;
1.254 schwarze 2027: break;
2028: case 'c':
1.261 schwarze 2029: scaled = *res * 240 / 2.54;
1.254 schwarze 2030: break;
2031: case 'v':
2032: case 'P':
1.261 schwarze 2033: scaled = *res * 40;
1.254 schwarze 2034: break;
2035: case 'm':
2036: case 'n':
1.261 schwarze 2037: scaled = *res * 24;
1.254 schwarze 2038: break;
2039: case 'p':
1.261 schwarze 2040: scaled = *res * 10 / 3;
1.254 schwarze 2041: break;
2042: case 'u':
1.261 schwarze 2043: scaled = *res;
1.254 schwarze 2044: break;
2045: case 'M':
1.261 schwarze 2046: scaled = *res * 6 / 25;
1.254 schwarze 2047: break;
2048: default:
1.261 schwarze 2049: scaled = *res;
1.254 schwarze 2050: p--;
2051: break;
2052: }
1.261 schwarze 2053: if (flags & ROFFNUM_SCALE)
2054: *res = scaled;
1.254 schwarze 2055:
2056: *pos = p + 1;
1.277 schwarze 2057: return 1;
1.184 schwarze 2058: }
2059:
1.198 schwarze 2060: /*
2061: * Evaluate a string comparison condition.
2062: * The first character is the delimiter.
2063: * Succeed if the string up to its second occurrence
2064: * matches the string up to its third occurence.
2065: * Advance the cursor after the third occurrence
2066: * or lacking that, to the end of the line.
2067: */
2068: static int
2069: roff_evalstrcond(const char *v, int *pos)
2070: {
2071: const char *s1, *s2, *s3;
2072: int match;
2073:
2074: match = 0;
2075: s1 = v + *pos; /* initial delimiter */
2076: s2 = s1 + 1; /* for scanning the first string */
2077: s3 = strchr(s2, *s1); /* for scanning the second string */
2078:
2079: if (NULL == s3) /* found no middle delimiter */
2080: goto out;
2081:
2082: while ('\0' != *++s3) {
2083: if (*s2 != *s3) { /* mismatch */
2084: s3 = strchr(s3, *s1);
2085: break;
2086: }
2087: if (*s3 == *s1) { /* found the final delimiter */
2088: match = 1;
2089: break;
2090: }
2091: s2++;
2092: }
2093:
2094: out:
2095: if (NULL == s3)
2096: s3 = strchr(s2, '\0');
1.242 schwarze 2097: else if (*s3 != '\0')
1.198 schwarze 2098: s3++;
2099: *pos = s3 - v;
1.277 schwarze 2100: return match;
1.198 schwarze 2101: }
2102:
1.204 schwarze 2103: /*
2104: * Evaluate an optionally negated single character, numerical,
2105: * or string condition.
2106: */
1.198 schwarze 2107: static int
1.271 schwarze 2108: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2109: {
1.271 schwarze 2110: char *cp, *name;
2111: size_t sz;
1.241 schwarze 2112: int number, savepos, wanttrue;
1.88 kristaps 2113:
1.198 schwarze 2114: if ('!' == v[*pos]) {
2115: wanttrue = 0;
2116: (*pos)++;
2117: } else
2118: wanttrue = 1;
2119:
1.88 kristaps 2120: switch (v[*pos]) {
1.240 schwarze 2121: case '\0':
1.277 schwarze 2122: return 0;
1.207 schwarze 2123: case 'n':
2124: case 'o':
1.88 kristaps 2125: (*pos)++;
1.277 schwarze 2126: return wanttrue;
1.207 schwarze 2127: case 'c':
2128: case 'd':
2129: case 'e':
2130: case 't':
1.239 schwarze 2131: case 'v':
1.88 kristaps 2132: (*pos)++;
1.277 schwarze 2133: return !wanttrue;
1.271 schwarze 2134: case 'r':
2135: cp = name = v + ++*pos;
2136: sz = roff_getname(r, &cp, ln, *pos);
2137: *pos = cp - v;
1.277 schwarze 2138: return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.88 kristaps 2139: default:
2140: break;
2141: }
2142:
1.241 schwarze 2143: savepos = *pos;
1.261 schwarze 2144: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2145: return (number > 0) == wanttrue;
1.241 schwarze 2146: else if (*pos == savepos)
1.277 schwarze 2147: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2148: else
1.277 schwarze 2149: return 0;
1.88 kristaps 2150: }
2151:
1.74 kristaps 2152: static enum rofferr
1.103 kristaps 2153: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2154: {
1.123 schwarze 2155:
1.277 schwarze 2156: return ROFF_IGN;
1.89 kristaps 2157: }
2158:
1.104 kristaps 2159: static enum rofferr
1.251 schwarze 2160: roff_insec(ROFF_ARGS)
2161: {
2162:
2163: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
2164: ln, ppos, roffs[tok].name);
1.277 schwarze 2165: return ROFF_IGN;
1.251 schwarze 2166: }
2167:
2168: static enum rofferr
2169: roff_unsupp(ROFF_ARGS)
2170: {
2171:
2172: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
2173: ln, ppos, roffs[tok].name);
1.277 schwarze 2174: return ROFF_IGN;
1.251 schwarze 2175: }
2176:
2177: static enum rofferr
1.82 kristaps 2178: roff_cond(ROFF_ARGS)
1.74 kristaps 2179: {
1.173 schwarze 2180:
2181: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2182:
1.207 schwarze 2183: /*
1.134 kristaps 2184: * An `.el' has no conditional body: it will consume the value
2185: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2186: * defaults to DENY.
1.134 kristaps 2187: *
2188: * If we're not an `el', however, then evaluate the conditional.
2189: */
1.133 kristaps 2190:
1.238 schwarze 2191: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2192: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2193: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2194:
1.134 kristaps 2195: /*
2196: * An if-else will put the NEGATION of the current evaluated
2197: * conditional into the stack of rules.
2198: */
2199:
1.238 schwarze 2200: if (tok == ROFF_ie) {
1.223 schwarze 2201: if (r->rstackpos + 1 == r->rstacksz) {
2202: r->rstacksz += 16;
2203: r->rstack = mandoc_reallocarray(r->rstack,
2204: r->rstacksz, sizeof(int));
1.134 kristaps 2205: }
1.198 schwarze 2206: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2207: }
1.88 kristaps 2208:
2209: /* If the parent has false as its rule, then so do we. */
2210:
1.198 schwarze 2211: if (r->last->parent && !r->last->parent->rule)
2212: r->last->rule = 0;
1.88 kristaps 2213:
2214: /*
1.173 schwarze 2215: * Determine scope.
2216: * If there is nothing on the line after the conditional,
2217: * not even whitespace, use next-line scope.
1.88 kristaps 2218: */
1.74 kristaps 2219:
1.238 schwarze 2220: if (buf->buf[pos] == '\0') {
1.173 schwarze 2221: r->last->endspan = 2;
2222: goto out;
2223: }
2224:
1.238 schwarze 2225: while (buf->buf[pos] == ' ')
1.173 schwarze 2226: pos++;
2227:
2228: /* An opening brace requests multiline scope. */
1.75 kristaps 2229:
1.238 schwarze 2230: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2231: r->last->endspan = -1;
2232: pos += 2;
1.272 schwarze 2233: while (buf->buf[pos] == ' ')
2234: pos++;
1.173 schwarze 2235: goto out;
1.207 schwarze 2236: }
1.74 kristaps 2237:
1.77 kristaps 2238: /*
1.173 schwarze 2239: * Anything else following the conditional causes
2240: * single-line scope. Warn if the scope contains
2241: * nothing but trailing whitespace.
1.77 kristaps 2242: */
2243:
1.238 schwarze 2244: if (buf->buf[pos] == '\0')
1.216 schwarze 2245: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
2246: ln, ppos, roffs[tok].name);
1.77 kristaps 2247:
1.173 schwarze 2248: r->last->endspan = 1;
1.74 kristaps 2249:
1.173 schwarze 2250: out:
1.75 kristaps 2251: *offs = pos;
1.277 schwarze 2252: return ROFF_RERUN;
1.83 schwarze 2253: }
2254:
2255: static enum rofferr
1.92 schwarze 2256: roff_ds(ROFF_ARGS)
2257: {
1.212 schwarze 2258: char *string;
2259: const char *name;
2260: size_t namesz;
1.96 kristaps 2261:
1.251 schwarze 2262: /* Ignore groff compatibility mode for now. */
2263:
2264: if (tok == ROFF_ds1)
2265: tok = ROFF_ds;
2266: else if (tok == ROFF_as1)
2267: tok = ROFF_as;
2268:
1.96 kristaps 2269: /*
1.212 schwarze 2270: * The first word is the name of the string.
2271: * If it is empty or terminated by an escape sequence,
2272: * abort the `ds' request without defining anything.
1.96 kristaps 2273: */
1.92 schwarze 2274:
1.238 schwarze 2275: name = string = buf->buf + pos;
2276: if (*name == '\0')
1.277 schwarze 2277: return ROFF_IGN;
1.92 schwarze 2278:
1.212 schwarze 2279: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2280: if (name[namesz] == '\\')
1.277 schwarze 2281: return ROFF_IGN;
1.212 schwarze 2282:
2283: /* Read past the initial double-quote, if any. */
1.238 schwarze 2284: if (*string == '"')
1.92 schwarze 2285: string++;
2286:
1.96 kristaps 2287: /* The rest is the value. */
1.212 schwarze 2288: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2289: ROFF_as == tok);
1.277 schwarze 2290: return ROFF_IGN;
1.92 schwarze 2291: }
2292:
1.204 schwarze 2293: /*
2294: * Parse a single operator, one or two characters long.
2295: * If the operator is recognized, return success and advance the
2296: * parse point, else return failure and let the parse point unchanged.
2297: */
2298: static int
2299: roff_getop(const char *v, int *pos, char *res)
2300: {
2301:
2302: *res = v[*pos];
2303:
2304: switch (*res) {
1.207 schwarze 2305: case '+':
2306: case '-':
2307: case '*':
2308: case '/':
2309: case '%':
2310: case '&':
2311: case ':':
1.204 schwarze 2312: break;
2313: case '<':
2314: switch (v[*pos + 1]) {
1.207 schwarze 2315: case '=':
1.204 schwarze 2316: *res = 'l';
2317: (*pos)++;
2318: break;
1.207 schwarze 2319: case '>':
1.204 schwarze 2320: *res = '!';
2321: (*pos)++;
2322: break;
1.207 schwarze 2323: case '?':
1.204 schwarze 2324: *res = 'i';
2325: (*pos)++;
2326: break;
2327: default:
2328: break;
2329: }
2330: break;
2331: case '>':
2332: switch (v[*pos + 1]) {
1.207 schwarze 2333: case '=':
1.204 schwarze 2334: *res = 'g';
2335: (*pos)++;
2336: break;
1.207 schwarze 2337: case '?':
1.204 schwarze 2338: *res = 'a';
2339: (*pos)++;
2340: break;
2341: default:
2342: break;
2343: }
2344: break;
2345: case '=':
2346: if ('=' == v[*pos + 1])
2347: (*pos)++;
2348: break;
2349: default:
1.277 schwarze 2350: return 0;
1.204 schwarze 2351: }
2352: (*pos)++;
2353:
1.277 schwarze 2354: return *res;
1.204 schwarze 2355: }
2356:
2357: /*
2358: * Evaluate either a parenthesized numeric expression
2359: * or a single signed integer number.
2360: */
2361: static int
1.235 schwarze 2362: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2363: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2364: {
2365:
2366: if ('(' != v[*pos])
1.277 schwarze 2367: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2368:
2369: (*pos)++;
1.261 schwarze 2370: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2371: return 0;
1.204 schwarze 2372:
1.206 schwarze 2373: /*
2374: * Omission of the closing parenthesis
2375: * is an error in validation mode,
2376: * but ignored in evaluation mode.
2377: */
2378:
1.204 schwarze 2379: if (')' == v[*pos])
2380: (*pos)++;
1.206 schwarze 2381: else if (NULL == res)
1.277 schwarze 2382: return 0;
1.204 schwarze 2383:
1.277 schwarze 2384: return 1;
1.204 schwarze 2385: }
2386:
2387: /*
2388: * Evaluate a complete numeric expression.
2389: * Proceed left to right, there is no concept of precedence.
2390: */
2391: static int
1.235 schwarze 2392: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2393: int *pos, int *res, int flags)
1.204 schwarze 2394: {
2395: int mypos, operand2;
2396: char operator;
2397:
2398: if (NULL == pos) {
2399: mypos = 0;
2400: pos = &mypos;
2401: }
2402:
1.261 schwarze 2403: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2404: while (isspace((unsigned char)v[*pos]))
2405: (*pos)++;
2406:
1.261 schwarze 2407: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2408: return 0;
1.204 schwarze 2409:
2410: while (1) {
1.261 schwarze 2411: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2412: while (isspace((unsigned char)v[*pos]))
2413: (*pos)++;
2414:
2415: if ( ! roff_getop(v, pos, &operator))
2416: break;
2417:
1.261 schwarze 2418: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2419: while (isspace((unsigned char)v[*pos]))
2420: (*pos)++;
2421:
1.261 schwarze 2422: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2423: return 0;
1.204 schwarze 2424:
1.261 schwarze 2425: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2426: while (isspace((unsigned char)v[*pos]))
2427: (*pos)++;
1.206 schwarze 2428:
2429: if (NULL == res)
2430: continue;
1.204 schwarze 2431:
2432: switch (operator) {
1.207 schwarze 2433: case '+':
1.204 schwarze 2434: *res += operand2;
2435: break;
1.207 schwarze 2436: case '-':
1.204 schwarze 2437: *res -= operand2;
2438: break;
1.207 schwarze 2439: case '*':
1.204 schwarze 2440: *res *= operand2;
2441: break;
1.207 schwarze 2442: case '/':
1.244 schwarze 2443: if (operand2 == 0) {
1.235 schwarze 2444: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2445: r->parse, ln, *pos, v);
2446: *res = 0;
2447: break;
2448: }
1.204 schwarze 2449: *res /= operand2;
2450: break;
1.207 schwarze 2451: case '%':
1.244 schwarze 2452: if (operand2 == 0) {
2453: mandoc_msg(MANDOCERR_DIVZERO,
2454: r->parse, ln, *pos, v);
2455: *res = 0;
2456: break;
2457: }
1.204 schwarze 2458: *res %= operand2;
2459: break;
1.207 schwarze 2460: case '<':
1.204 schwarze 2461: *res = *res < operand2;
2462: break;
1.207 schwarze 2463: case '>':
1.204 schwarze 2464: *res = *res > operand2;
2465: break;
1.207 schwarze 2466: case 'l':
1.204 schwarze 2467: *res = *res <= operand2;
2468: break;
1.207 schwarze 2469: case 'g':
1.204 schwarze 2470: *res = *res >= operand2;
2471: break;
1.207 schwarze 2472: case '=':
1.204 schwarze 2473: *res = *res == operand2;
2474: break;
1.207 schwarze 2475: case '!':
1.204 schwarze 2476: *res = *res != operand2;
2477: break;
1.207 schwarze 2478: case '&':
1.204 schwarze 2479: *res = *res && operand2;
2480: break;
1.207 schwarze 2481: case ':':
1.204 schwarze 2482: *res = *res || operand2;
2483: break;
1.207 schwarze 2484: case 'i':
1.204 schwarze 2485: if (operand2 < *res)
2486: *res = operand2;
2487: break;
1.207 schwarze 2488: case 'a':
1.204 schwarze 2489: if (operand2 > *res)
2490: *res = operand2;
2491: break;
2492: default:
2493: abort();
2494: }
2495: }
1.277 schwarze 2496: return 1;
1.204 schwarze 2497: }
2498:
1.266 schwarze 2499: /* --- register management ------------------------------------------------ */
2500:
1.180 schwarze 2501: void
1.187 schwarze 2502: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2503: {
1.180 schwarze 2504: struct roffreg *reg;
2505:
2506: /* Search for an existing register with the same name. */
2507: reg = r->regtab;
2508:
2509: while (reg && strcmp(name, reg->key.p))
2510: reg = reg->next;
1.147 kristaps 2511:
1.180 schwarze 2512: if (NULL == reg) {
2513: /* Create a new register. */
2514: reg = mandoc_malloc(sizeof(struct roffreg));
2515: reg->key.p = mandoc_strdup(name);
2516: reg->key.sz = strlen(name);
1.187 schwarze 2517: reg->val = 0;
1.180 schwarze 2518: reg->next = r->regtab;
2519: r->regtab = reg;
2520: }
2521:
1.187 schwarze 2522: if ('+' == sign)
2523: reg->val += val;
2524: else if ('-' == sign)
2525: reg->val -= val;
2526: else
2527: reg->val = val;
1.147 kristaps 2528: }
2529:
1.192 schwarze 2530: /*
2531: * Handle some predefined read-only number registers.
2532: * For now, return -1 if the requested register is not predefined;
2533: * in case a predefined read-only register having the value -1
2534: * were to turn up, another special value would have to be chosen.
2535: */
2536: static int
1.273 schwarze 2537: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2538: {
2539:
2540: switch (*name) {
1.273 schwarze 2541: case '$': /* Number of arguments of the last macro evaluated. */
1.277 schwarze 2542: return r->argc;
1.207 schwarze 2543: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2544: return 0;
1.207 schwarze 2545: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2546: return 1;
1.207 schwarze 2547: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2548: return 24;
1.207 schwarze 2549: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2550: return 0;
1.207 schwarze 2551: case 'T': /* Some output device is always defined. */
1.277 schwarze 2552: return 1;
1.207 schwarze 2553: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2554: return 40;
1.192 schwarze 2555: default:
1.277 schwarze 2556: return -1;
1.192 schwarze 2557: }
2558: }
2559:
1.181 schwarze 2560: int
1.180 schwarze 2561: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2562: {
1.180 schwarze 2563: struct roffreg *reg;
1.192 schwarze 2564: int val;
2565:
2566: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.273 schwarze 2567: val = roff_getregro(r, name + 1);
1.192 schwarze 2568: if (-1 != val)
1.277 schwarze 2569: return val;
1.192 schwarze 2570: }
1.180 schwarze 2571:
2572: for (reg = r->regtab; reg; reg = reg->next)
2573: if (0 == strcmp(name, reg->key.p))
1.277 schwarze 2574: return reg->val;
1.181 schwarze 2575:
1.277 schwarze 2576: return 0;
1.181 schwarze 2577: }
2578:
2579: static int
2580: roff_getregn(const struct roff *r, const char *name, size_t len)
2581: {
2582: struct roffreg *reg;
1.192 schwarze 2583: int val;
2584:
2585: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2586: val = roff_getregro(r, name + 1);
1.192 schwarze 2587: if (-1 != val)
1.277 schwarze 2588: return val;
1.192 schwarze 2589: }
1.181 schwarze 2590:
2591: for (reg = r->regtab; reg; reg = reg->next)
2592: if (len == reg->key.sz &&
2593: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2594: return reg->val;
1.271 schwarze 2595:
1.277 schwarze 2596: return 0;
1.271 schwarze 2597: }
2598:
2599: static int
2600: roff_hasregn(const struct roff *r, const char *name, size_t len)
2601: {
2602: struct roffreg *reg;
2603: int val;
2604:
2605: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2606: val = roff_getregro(r, name + 1);
1.271 schwarze 2607: if (-1 != val)
1.277 schwarze 2608: return 1;
1.271 schwarze 2609: }
2610:
2611: for (reg = r->regtab; reg; reg = reg->next)
2612: if (len == reg->key.sz &&
2613: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2614: return 1;
1.147 kristaps 2615:
1.277 schwarze 2616: return 0;
1.147 kristaps 2617: }
2618:
1.180 schwarze 2619: static void
2620: roff_freereg(struct roffreg *reg)
1.147 kristaps 2621: {
1.180 schwarze 2622: struct roffreg *old_reg;
1.147 kristaps 2623:
1.180 schwarze 2624: while (NULL != reg) {
2625: free(reg->key.p);
2626: old_reg = reg;
2627: reg = reg->next;
2628: free(old_reg);
2629: }
1.147 kristaps 2630: }
1.92 schwarze 2631:
2632: static enum rofferr
1.89 kristaps 2633: roff_nr(ROFF_ARGS)
1.83 schwarze 2634: {
1.212 schwarze 2635: char *key, *val;
2636: size_t keysz;
1.138 kristaps 2637: int iv;
1.187 schwarze 2638: char sign;
1.89 kristaps 2639:
1.238 schwarze 2640: key = val = buf->buf + pos;
2641: if (*key == '\0')
1.277 schwarze 2642: return ROFF_IGN;
1.212 schwarze 2643:
2644: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2645: if (key[keysz] == '\\')
1.277 schwarze 2646: return ROFF_IGN;
1.212 schwarze 2647: key[keysz] = '\0';
1.89 kristaps 2648:
1.187 schwarze 2649: sign = *val;
1.238 schwarze 2650: if (sign == '+' || sign == '-')
1.187 schwarze 2651: val++;
2652:
1.261 schwarze 2653: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204 schwarze 2654: roff_setreg(r, key, iv, sign);
1.109 kristaps 2655:
1.277 schwarze 2656: return ROFF_IGN;
1.203 schwarze 2657: }
2658:
2659: static enum rofferr
2660: roff_rr(ROFF_ARGS)
2661: {
2662: struct roffreg *reg, **prev;
1.212 schwarze 2663: char *name, *cp;
2664: size_t namesz;
1.203 schwarze 2665:
1.238 schwarze 2666: name = cp = buf->buf + pos;
2667: if (*name == '\0')
1.277 schwarze 2668: return ROFF_IGN;
1.212 schwarze 2669: namesz = roff_getname(r, &cp, ln, pos);
2670: name[namesz] = '\0';
1.203 schwarze 2671:
2672: prev = &r->regtab;
2673: while (1) {
2674: reg = *prev;
1.238 schwarze 2675: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2676: break;
2677: prev = ®->next;
2678: }
1.238 schwarze 2679: if (reg != NULL) {
1.203 schwarze 2680: *prev = reg->next;
2681: free(reg->key.p);
2682: free(reg);
2683: }
1.277 schwarze 2684: return ROFF_IGN;
1.122 schwarze 2685: }
2686:
1.266 schwarze 2687: /* --- handler functions for roff requests -------------------------------- */
2688:
1.122 schwarze 2689: static enum rofferr
2690: roff_rm(ROFF_ARGS)
2691: {
2692: const char *name;
2693: char *cp;
1.212 schwarze 2694: size_t namesz;
1.122 schwarze 2695:
1.238 schwarze 2696: cp = buf->buf + pos;
2697: while (*cp != '\0') {
1.212 schwarze 2698: name = cp;
1.238 schwarze 2699: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2700: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238 schwarze 2701: if (name[namesz] == '\\')
1.212 schwarze 2702: break;
1.122 schwarze 2703: }
1.277 schwarze 2704: return ROFF_IGN;
1.178 schwarze 2705: }
2706:
2707: static enum rofferr
2708: roff_it(ROFF_ARGS)
2709: {
2710: int iv;
2711:
2712: /* Parse the number of lines. */
1.261 schwarze 2713:
2714: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2715: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2716: ln, ppos, buf->buf + 1);
1.277 schwarze 2717: return ROFF_IGN;
1.178 schwarze 2718: }
2719:
1.262 schwarze 2720: while (isspace((unsigned char)buf->buf[pos]))
2721: pos++;
2722:
2723: /*
2724: * Arm the input line trap.
2725: * Special-casing "an-trap" is an ugly workaround to cope
2726: * with DocBook stupidly fiddling with man(7) internals.
2727: */
1.261 schwarze 2728:
1.178 schwarze 2729: roffit_lines = iv;
1.262 schwarze 2730: roffit_macro = mandoc_strdup(iv != 1 ||
2731: strcmp(buf->buf + pos, "an-trap") ?
2732: buf->buf + pos : "br");
1.277 schwarze 2733: return ROFF_IGN;
1.175 schwarze 2734: }
2735:
2736: static enum rofferr
2737: roff_Dd(ROFF_ARGS)
2738: {
2739: const char *const *cp;
2740:
1.227 schwarze 2741: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175 schwarze 2742: for (cp = __mdoc_reserved; *cp; cp++)
2743: roff_setstr(r, *cp, NULL, 0);
2744:
1.227 schwarze 2745: if (r->format == 0)
2746: r->format = MPARSE_MDOC;
2747:
1.277 schwarze 2748: return ROFF_CONT;
1.175 schwarze 2749: }
2750:
2751: static enum rofferr
2752: roff_TH(ROFF_ARGS)
2753: {
2754: const char *const *cp;
2755:
1.227 schwarze 2756: if ((r->options & MPARSE_QUICK) == 0)
1.175 schwarze 2757: for (cp = __man_reserved; *cp; cp++)
2758: roff_setstr(r, *cp, NULL, 0);
2759:
1.227 schwarze 2760: if (r->format == 0)
2761: r->format = MPARSE_MAN;
2762:
1.277 schwarze 2763: return ROFF_CONT;
1.109 kristaps 2764: }
2765:
2766: static enum rofferr
2767: roff_TE(ROFF_ARGS)
2768: {
2769:
2770: if (NULL == r->tbl)
1.221 schwarze 2771: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2772: ln, ppos, "TE");
1.258 schwarze 2773: else if ( ! tbl_end(&r->tbl)) {
2774: free(buf->buf);
2775: buf->buf = mandoc_strdup(".sp");
2776: buf->sz = 4;
1.277 schwarze 2777: return ROFF_REPARSE;
1.258 schwarze 2778: }
1.277 schwarze 2779: return ROFF_IGN;
1.112 kristaps 2780: }
2781:
2782: static enum rofferr
2783: roff_T_(ROFF_ARGS)
2784: {
2785:
2786: if (NULL == r->tbl)
1.221 schwarze 2787: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2788: ln, ppos, "T&");
1.112 kristaps 2789: else
1.116 kristaps 2790: tbl_restart(ppos, ln, r->tbl);
1.112 kristaps 2791:
1.277 schwarze 2792: return ROFF_IGN;
1.109 kristaps 2793: }
2794:
1.230 schwarze 2795: /*
2796: * Handle in-line equation delimiters.
2797: */
2798: static enum rofferr
1.238 schwarze 2799: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2800: {
1.233 schwarze 2801: char *cp1, *cp2;
2802: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2803:
1.230 schwarze 2804: /*
2805: * Outside equations, look for an opening delimiter.
2806: * If we are inside an equation, we already know it is
2807: * in-line, or this function wouldn't have been called;
2808: * so look for a closing delimiter.
2809: */
2810:
1.238 schwarze 2811: cp1 = buf->buf + pos;
1.230 schwarze 2812: cp2 = strchr(cp1, r->eqn == NULL ?
2813: r->last_eqn->odelim : r->last_eqn->cdelim);
2814: if (cp2 == NULL)
1.277 schwarze 2815: return ROFF_CONT;
1.230 schwarze 2816:
1.233 schwarze 2817: *cp2++ = '\0';
2818: bef_pr = bef_nl = aft_nl = aft_pr = "";
2819:
2820: /* Handle preceding text, protecting whitespace. */
2821:
1.238 schwarze 2822: if (*buf->buf != '\0') {
1.233 schwarze 2823: if (r->eqn == NULL)
2824: bef_pr = "\\&";
2825: bef_nl = "\n";
2826: }
2827:
2828: /*
2829: * Prepare replacing the delimiter with an equation macro
2830: * and drop leading white space from the equation.
2831: */
1.230 schwarze 2832:
1.233 schwarze 2833: if (r->eqn == NULL) {
2834: while (*cp2 == ' ')
2835: cp2++;
2836: mac = ".EQ";
2837: } else
2838: mac = ".EN";
2839:
2840: /* Handle following text, protecting whitespace. */
2841:
2842: if (*cp2 != '\0') {
2843: aft_nl = "\n";
2844: if (r->eqn != NULL)
2845: aft_pr = "\\&";
2846: }
2847:
2848: /* Do the actual replacement. */
2849:
1.238 schwarze 2850: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2851: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2852: free(buf->buf);
2853: buf->buf = cp1;
1.230 schwarze 2854:
2855: /* Toggle the in-line state of the eqn subsystem. */
2856:
2857: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 2858: return ROFF_REPARSE;
1.151 kristaps 2859: }
2860:
1.235 schwarze 2861: static enum rofferr
2862: roff_EQ(ROFF_ARGS)
1.125 kristaps 2863: {
1.151 kristaps 2864: struct eqn_node *e;
1.125 kristaps 2865:
1.238 schwarze 2866: assert(r->eqn == NULL);
1.235 schwarze 2867: e = eqn_alloc(ppos, ln, r->parse);
1.125 kristaps 2868:
1.230 schwarze 2869: if (r->last_eqn) {
1.125 kristaps 2870: r->last_eqn->next = e;
1.230 schwarze 2871: e->delim = r->last_eqn->delim;
2872: e->odelim = r->last_eqn->odelim;
2873: e->cdelim = r->last_eqn->cdelim;
2874: } else
1.125 kristaps 2875: r->first_eqn = r->last_eqn = e;
2876:
2877: r->eqn = r->last_eqn = e;
1.151 kristaps 2878:
1.238 schwarze 2879: if (buf->buf[pos] != '\0')
1.235 schwarze 2880: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2881: ".EQ %s", buf->buf + pos);
1.151 kristaps 2882:
1.277 schwarze 2883: return ROFF_IGN;
1.125 kristaps 2884: }
2885:
2886: static enum rofferr
2887: roff_EN(ROFF_ARGS)
2888: {
2889:
1.221 schwarze 2890: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.277 schwarze 2891: return ROFF_IGN;
1.125 kristaps 2892: }
2893:
2894: static enum rofferr
1.109 kristaps 2895: roff_TS(ROFF_ARGS)
2896: {
1.176 schwarze 2897: struct tbl_node *tbl;
1.89 kristaps 2898:
1.115 kristaps 2899: if (r->tbl) {
1.221 schwarze 2900: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2901: ln, ppos, "TS breaks TS");
1.151 kristaps 2902: tbl_end(&r->tbl);
1.115 kristaps 2903: }
1.83 schwarze 2904:
1.176 schwarze 2905: tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2906:
2907: if (r->last_tbl)
1.176 schwarze 2908: r->last_tbl->next = tbl;
1.113 kristaps 2909: else
1.176 schwarze 2910: r->first_tbl = r->last_tbl = tbl;
1.113 kristaps 2911:
1.176 schwarze 2912: r->tbl = r->last_tbl = tbl;
1.277 schwarze 2913: return ROFF_IGN;
1.251 schwarze 2914: }
2915:
2916: static enum rofferr
2917: roff_brp(ROFF_ARGS)
2918: {
2919:
2920: buf->buf[pos - 1] = '\0';
1.277 schwarze 2921: return ROFF_CONT;
1.92 schwarze 2922: }
2923:
1.105 kristaps 2924: static enum rofferr
1.174 kristaps 2925: roff_cc(ROFF_ARGS)
2926: {
2927: const char *p;
2928:
1.238 schwarze 2929: p = buf->buf + pos;
1.174 kristaps 2930:
1.238 schwarze 2931: if (*p == '\0' || (r->control = *p++) == '.')
1.174 kristaps 2932: r->control = 0;
2933:
1.238 schwarze 2934: if (*p != '\0')
1.260 schwarze 2935: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2936: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 2937:
1.277 schwarze 2938: return ROFF_IGN;
1.174 kristaps 2939: }
2940:
2941: static enum rofferr
1.164 kristaps 2942: roff_tr(ROFF_ARGS)
2943: {
2944: const char *p, *first, *second;
2945: size_t fsz, ssz;
2946: enum mandoc_esc esc;
2947:
1.238 schwarze 2948: p = buf->buf + pos;
1.164 kristaps 2949:
1.238 schwarze 2950: if (*p == '\0') {
1.260 schwarze 2951: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 2952: return ROFF_IGN;
1.164 kristaps 2953: }
2954:
1.238 schwarze 2955: while (*p != '\0') {
1.164 kristaps 2956: fsz = ssz = 1;
2957:
2958: first = p++;
1.238 schwarze 2959: if (*first == '\\') {
1.164 kristaps 2960: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2961: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2962: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2963: ln, (int)(p - buf->buf), first);
1.277 schwarze 2964: return ROFF_IGN;
1.164 kristaps 2965: }
2966: fsz = (size_t)(p - first);
2967: }
2968:
2969: second = p++;
1.238 schwarze 2970: if (*second == '\\') {
1.164 kristaps 2971: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2972: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2973: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2974: ln, (int)(p - buf->buf), second);
1.277 schwarze 2975: return ROFF_IGN;
1.164 kristaps 2976: }
2977: ssz = (size_t)(p - second);
1.238 schwarze 2978: } else if (*second == '\0') {
1.260 schwarze 2979: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2980: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 2981: second = " ";
1.165 kristaps 2982: p--;
1.164 kristaps 2983: }
2984:
1.167 kristaps 2985: if (fsz > 1) {
1.207 schwarze 2986: roff_setstrn(&r->xmbtab, first, fsz,
2987: second, ssz, 0);
1.167 kristaps 2988: continue;
2989: }
2990:
1.238 schwarze 2991: if (r->xtab == NULL)
1.207 schwarze 2992: r->xtab = mandoc_calloc(128,
2993: sizeof(struct roffstr));
1.167 kristaps 2994:
2995: free(r->xtab[(int)*first].p);
2996: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2997: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 2998: }
2999:
1.277 schwarze 3000: return ROFF_IGN;
1.164 kristaps 3001: }
3002:
3003: static enum rofferr
1.105 kristaps 3004: roff_so(ROFF_ARGS)
3005: {
1.249 schwarze 3006: char *name, *cp;
1.105 kristaps 3007:
1.238 schwarze 3008: name = buf->buf + pos;
1.224 schwarze 3009: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3010:
3011: /*
3012: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3013: * opening anything that's not in our cwd or anything beneath
3014: * it. Thus, explicitly disallow traversing up the file-system
3015: * or using absolute paths.
3016: */
3017:
1.238 schwarze 3018: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3019: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3020: ".so %s", name);
1.249 schwarze 3021: buf->sz = mandoc_asprintf(&cp,
3022: ".sp\nSee the file %s.\n.sp", name) + 1;
3023: free(buf->buf);
3024: buf->buf = cp;
3025: *offs = 0;
1.277 schwarze 3026: return ROFF_REPARSE;
1.105 kristaps 3027: }
3028:
3029: *offs = pos;
1.277 schwarze 3030: return ROFF_SO;
1.105 kristaps 3031: }
1.92 schwarze 3032:
1.266 schwarze 3033: /* --- user defined strings and macros ------------------------------------ */
3034:
1.106 kristaps 3035: static enum rofferr
3036: roff_userdef(ROFF_ARGS)
1.99 kristaps 3037: {
1.263 schwarze 3038: const char *arg[9], *ap;
1.106 kristaps 3039: char *cp, *n1, *n2;
1.274 schwarze 3040: int i, ib, ie;
1.263 schwarze 3041: size_t asz, rsz;
1.106 kristaps 3042:
3043: /*
3044: * Collect pointers to macro argument strings
1.188 schwarze 3045: * and NUL-terminate them.
1.106 kristaps 3046: */
1.263 schwarze 3047:
1.273 schwarze 3048: r->argc = 0;
1.238 schwarze 3049: cp = buf->buf + pos;
1.273 schwarze 3050: for (i = 0; i < 9; i++) {
3051: if (*cp == '\0')
3052: arg[i] = "";
3053: else {
3054: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3055: r->argc = i + 1;
3056: }
3057: }
1.99 kristaps 3058:
1.106 kristaps 3059: /*
3060: * Expand macro arguments.
1.99 kristaps 3061: */
1.263 schwarze 3062:
3063: buf->sz = strlen(r->current_string) + 1;
3064: n1 = cp = mandoc_malloc(buf->sz);
3065: memcpy(n1, r->current_string, buf->sz);
3066: while (*cp != '\0') {
3067:
3068: /* Scan ahead for the next argument invocation. */
3069:
3070: if (*cp++ != '\\')
3071: continue;
3072: if (*cp++ != '$')
3073: continue;
1.274 schwarze 3074: if (*cp == '*') { /* \\$* inserts all arguments */
3075: ib = 0;
3076: ie = r->argc - 1;
3077: } else { /* \\$1 .. \\$9 insert one argument */
3078: ib = ie = *cp - '1';
3079: if (ib < 0 || ib > 8)
3080: continue;
3081: }
1.263 schwarze 3082: cp -= 2;
3083:
3084: /*
3085: * Determine the size of the expanded argument,
3086: * taking escaping of quotes into account.
3087: */
3088:
1.274 schwarze 3089: asz = ie > ib ? ie - ib : 0; /* for blanks */
3090: for (i = ib; i <= ie; i++) {
3091: for (ap = arg[i]; *ap != '\0'; ap++) {
3092: asz++;
3093: if (*ap == '"')
3094: asz += 3;
3095: }
1.263 schwarze 3096: }
3097: if (asz != 3) {
3098:
3099: /*
3100: * Determine the size of the rest of the
3101: * unexpanded macro, including the NUL.
3102: */
3103:
3104: rsz = buf->sz - (cp - n1) - 3;
3105:
3106: /*
3107: * When shrinking, move before
3108: * releasing the storage.
3109: */
3110:
3111: if (asz < 3)
3112: memmove(cp + asz, cp + 3, rsz);
3113:
3114: /*
3115: * Resize the storage for the macro
3116: * and readjust the parse pointer.
3117: */
3118:
3119: buf->sz += asz - 3;
3120: n2 = mandoc_realloc(n1, buf->sz);
3121: cp = n2 + (cp - n1);
3122: n1 = n2;
3123:
3124: /*
3125: * When growing, make room
3126: * for the expanded argument.
3127: */
3128:
3129: if (asz > 3)
3130: memmove(cp + asz, cp + 3, rsz);
3131: }
3132:
3133: /* Copy the expanded argument, escaping quotes. */
3134:
3135: n2 = cp;
1.274 schwarze 3136: for (i = ib; i <= ie; i++) {
3137: for (ap = arg[i]; *ap != '\0'; ap++) {
3138: if (*ap == '"') {
3139: memcpy(n2, "\\(dq", 4);
3140: n2 += 4;
3141: } else
3142: *n2++ = *ap;
3143: }
3144: if (i < ie)
3145: *n2++ = ' ';
1.106 kristaps 3146: }
1.99 kristaps 3147: }
3148:
1.106 kristaps 3149: /*
3150: * Replace the macro invocation
3151: * by the expanded macro.
3152: */
1.263 schwarze 3153:
1.238 schwarze 3154: free(buf->buf);
3155: buf->buf = n1;
1.248 schwarze 3156: *offs = 0;
1.106 kristaps 3157:
1.277 schwarze 3158: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3159: ROFF_REPARSE : ROFF_APPEND;
1.99 kristaps 3160: }
1.121 schwarze 3161:
1.212 schwarze 3162: static size_t
1.121 schwarze 3163: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3164: {
3165: char *name, *cp;
1.212 schwarze 3166: size_t namesz;
1.121 schwarze 3167:
3168: name = *cpp;
3169: if ('\0' == *name)
1.277 schwarze 3170: return 0;
1.121 schwarze 3171:
1.212 schwarze 3172: /* Read until end of name and terminate it with NUL. */
3173: for (cp = name; 1; cp++) {
3174: if ('\0' == *cp || ' ' == *cp) {
3175: namesz = cp - name;
3176: break;
3177: }
1.121 schwarze 3178: if ('\\' != *cp)
3179: continue;
1.215 schwarze 3180: namesz = cp - name;
3181: if ('{' == cp[1] || '}' == cp[1])
3182: break;
1.121 schwarze 3183: cp++;
3184: if ('\\' == *cp)
3185: continue;
1.224 schwarze 3186: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3187: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3188: mandoc_escape((const char **)&cp, NULL, NULL);
3189: break;
1.121 schwarze 3190: }
3191:
3192: /* Read past spaces. */
3193: while (' ' == *cp)
3194: cp++;
3195:
3196: *cpp = cp;
1.277 schwarze 3197: return namesz;
1.121 schwarze 3198: }
3199:
1.106 kristaps 3200: /*
3201: * Store *string into the user-defined string called *name.
3202: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3203: * append == 0: replace mode
3204: * append == 1: single-line append mode
3205: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3206: */
1.94 kristaps 3207: static void
1.106 kristaps 3208: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3209: int append)
1.92 schwarze 3210: {
1.164 kristaps 3211:
3212: roff_setstrn(&r->strtab, name, strlen(name), string,
1.207 schwarze 3213: string ? strlen(string) : 0, append);
1.164 kristaps 3214: }
3215:
3216: static void
1.166 kristaps 3217: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3218: const char *string, size_t stringsz, int append)
1.164 kristaps 3219: {
1.166 kristaps 3220: struct roffkv *n;
1.164 kristaps 3221: char *c;
3222: int i;
3223: size_t oldch, newch;
1.92 schwarze 3224:
1.106 kristaps 3225: /* Search for an existing string with the same name. */
1.164 kristaps 3226: n = *r;
3227:
1.211 schwarze 3228: while (n && (namesz != n->key.sz ||
3229: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3230: n = n->next;
1.94 kristaps 3231:
3232: if (NULL == n) {
1.106 kristaps 3233: /* Create a new string table entry. */
1.166 kristaps 3234: n = mandoc_malloc(sizeof(struct roffkv));
3235: n->key.p = mandoc_strndup(name, namesz);
3236: n->key.sz = namesz;
3237: n->val.p = NULL;
3238: n->val.sz = 0;
1.164 kristaps 3239: n->next = *r;
3240: *r = n;
1.193 schwarze 3241: } else if (0 == append) {
1.166 kristaps 3242: free(n->val.p);
3243: n->val.p = NULL;
3244: n->val.sz = 0;
1.106 kristaps 3245: }
3246:
3247: if (NULL == string)
3248: return;
3249:
3250: /*
3251: * One additional byte for the '\n' in multiline mode,
3252: * and one for the terminating '\0'.
3253: */
1.193 schwarze 3254: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3255:
1.166 kristaps 3256: if (NULL == n->val.p) {
3257: n->val.p = mandoc_malloc(newch);
3258: *n->val.p = '\0';
1.106 kristaps 3259: oldch = 0;
3260: } else {
1.166 kristaps 3261: oldch = n->val.sz;
3262: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3263: }
3264:
3265: /* Skip existing content in the destination buffer. */
1.166 kristaps 3266: c = n->val.p + (int)oldch;
1.106 kristaps 3267:
3268: /* Append new content to the destination buffer. */
1.164 kristaps 3269: i = 0;
3270: while (i < (int)stringsz) {
1.106 kristaps 3271: /*
3272: * Rudimentary roff copy mode:
3273: * Handle escaped backslashes.
3274: */
1.164 kristaps 3275: if ('\\' == string[i] && '\\' == string[i + 1])
3276: i++;
3277: *c++ = string[i++];
1.106 kristaps 3278: }
1.94 kristaps 3279:
1.106 kristaps 3280: /* Append terminating bytes. */
1.193 schwarze 3281: if (1 < append)
1.106 kristaps 3282: *c++ = '\n';
1.163 kristaps 3283:
1.106 kristaps 3284: *c = '\0';
1.166 kristaps 3285: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3286: }
3287:
1.94 kristaps 3288: static const char *
3289: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92 schwarze 3290: {
1.166 kristaps 3291: const struct roffkv *n;
1.191 schwarze 3292: int i;
1.92 schwarze 3293:
1.164 kristaps 3294: for (n = r->strtab; n; n = n->next)
1.207 schwarze 3295: if (0 == strncmp(name, n->key.p, len) &&
3296: '\0' == n->key.p[(int)len])
1.277 schwarze 3297: return n->val.p;
1.191 schwarze 3298:
3299: for (i = 0; i < PREDEFS_MAX; i++)
3300: if (0 == strncmp(name, predefs[i].name, len) &&
3301: '\0' == predefs[i].name[(int)len])
1.277 schwarze 3302: return predefs[i].str;
1.94 kristaps 3303:
1.277 schwarze 3304: return NULL;
1.92 schwarze 3305: }
3306:
1.94 kristaps 3307: static void
1.167 kristaps 3308: roff_freestr(struct roffkv *r)
1.92 schwarze 3309: {
1.166 kristaps 3310: struct roffkv *n, *nn;
1.92 schwarze 3311:
1.167 kristaps 3312: for (n = r; n; n = nn) {
1.166 kristaps 3313: free(n->key.p);
3314: free(n->val.p);
1.92 schwarze 3315: nn = n->next;
3316: free(n);
3317: }
1.114 kristaps 3318: }
1.266 schwarze 3319:
3320: /* --- accessors and utility functions ------------------------------------ */
1.114 kristaps 3321:
3322: const struct tbl_span *
3323: roff_span(const struct roff *r)
3324: {
1.207 schwarze 3325:
1.277 schwarze 3326: return r->tbl ? tbl_span(r->tbl) : NULL;
1.125 kristaps 3327: }
3328:
3329: const struct eqn *
3330: roff_eqn(const struct roff *r)
3331: {
1.207 schwarze 3332:
1.277 schwarze 3333: return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.164 kristaps 3334: }
3335:
3336: /*
3337: * Duplicate an input string, making the appropriate character
3338: * conversations (as stipulated by `tr') along the way.
3339: * Returns a heap-allocated string with all the replacements made.
3340: */
3341: char *
3342: roff_strdup(const struct roff *r, const char *p)
3343: {
1.166 kristaps 3344: const struct roffkv *cp;
1.164 kristaps 3345: char *res;
3346: const char *pp;
3347: size_t ssz, sz;
3348: enum mandoc_esc esc;
3349:
1.167 kristaps 3350: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3351: return mandoc_strdup(p);
1.164 kristaps 3352: else if ('\0' == *p)
1.277 schwarze 3353: return mandoc_strdup("");
1.164 kristaps 3354:
3355: /*
3356: * Step through each character looking for term matches
3357: * (remember that a `tr' can be invoked with an escape, which is
3358: * a glyph but the escape is multi-character).
3359: * We only do this if the character hash has been initialised
3360: * and the string is >0 length.
3361: */
3362:
3363: res = NULL;
3364: ssz = 0;
3365:
3366: while ('\0' != *p) {
1.167 kristaps 3367: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3368: sz = r->xtab[(int)*p].sz;
3369: res = mandoc_realloc(res, ssz + sz + 1);
3370: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3371: ssz += sz;
3372: p++;
3373: continue;
3374: } else if ('\\' != *p) {
3375: res = mandoc_realloc(res, ssz + 2);
3376: res[ssz++] = *p++;
3377: continue;
3378: }
3379:
1.164 kristaps 3380: /* Search for term matches. */
1.167 kristaps 3381: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3382: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3383: break;
3384:
3385: if (NULL != cp) {
3386: /*
3387: * A match has been found.
3388: * Append the match to the array and move
3389: * forward by its keysize.
3390: */
1.207 schwarze 3391: res = mandoc_realloc(res,
3392: ssz + cp->val.sz + 1);
1.166 kristaps 3393: memcpy(res + ssz, cp->val.p, cp->val.sz);
3394: ssz += cp->val.sz;
3395: p += (int)cp->key.sz;
1.164 kristaps 3396: continue;
3397: }
3398:
1.167 kristaps 3399: /*
3400: * Handle escapes carefully: we need to copy
3401: * over just the escape itself, or else we might
3402: * do replacements within the escape itself.
3403: * Make sure to pass along the bogus string.
3404: */
3405: pp = p++;
3406: esc = mandoc_escape(&p, NULL, NULL);
3407: if (ESCAPE_ERROR == esc) {
3408: sz = strlen(pp);
1.164 kristaps 3409: res = mandoc_realloc(res, ssz + sz + 1);
3410: memcpy(res + ssz, pp, sz);
1.167 kristaps 3411: break;
1.164 kristaps 3412: }
1.207 schwarze 3413: /*
3414: * We bail out on bad escapes.
1.167 kristaps 3415: * No need to warn: we already did so when
3416: * roff_res() was called.
3417: */
3418: sz = (int)(p - pp);
3419: res = mandoc_realloc(res, ssz + sz + 1);
3420: memcpy(res + ssz, pp, sz);
3421: ssz += sz;
1.164 kristaps 3422: }
3423:
3424: res[(int)ssz] = '\0';
1.277 schwarze 3425: return res;
1.227 schwarze 3426: }
3427:
3428: int
3429: roff_getformat(const struct roff *r)
3430: {
3431:
1.277 schwarze 3432: return r->format;
1.174 kristaps 3433: }
3434:
3435: /*
1.207 schwarze 3436: * Find out whether a line is a macro line or not.
1.174 kristaps 3437: * If it is, adjust the current position and return one; if it isn't,
3438: * return zero and don't change the current position.
3439: * If the control character has been set with `.cc', then let that grain
3440: * precedence.
3441: * This is slighly contrary to groff, where using the non-breaking
3442: * control character when `cc' has been invoked will cause the
3443: * non-breaking macro contents to be printed verbatim.
3444: */
3445: int
3446: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3447: {
3448: int pos;
3449:
3450: pos = *ppos;
3451:
3452: if (0 != r->control && cp[pos] == r->control)
3453: pos++;
3454: else if (0 != r->control)
1.277 schwarze 3455: return 0;
1.174 kristaps 3456: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3457: pos += 2;
3458: else if ('.' == cp[pos] || '\'' == cp[pos])
3459: pos++;
3460: else
1.277 schwarze 3461: return 0;
1.174 kristaps 3462:
3463: while (' ' == cp[pos] || '\t' == cp[pos])
3464: pos++;
3465:
3466: *ppos = pos;
1.277 schwarze 3467: return 1;
1.74 kristaps 3468: }
CVSweb