Annotation of mandoc/roff.c, Revision 1.274
1.274 ! schwarze 1: /* $Id: roff.c,v 1.273 2015/08/29 20:26:04 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"
1.33 kristaps 35:
1.141 kristaps 36: /* Maximum number of nested if-else conditionals. */
1.82 kristaps 37: #define RSTACK_MAX 128
38:
1.170 schwarze 39: /* Maximum number of string expansions per line, to break infinite loops. */
40: #define EXPAND_LIMIT 1000
41:
1.266 schwarze 42: /* --- data types --------------------------------------------------------- */
43:
1.67 kristaps 44: enum rofft {
1.251 schwarze 45: ROFF_ab,
1.103 kristaps 46: ROFF_ad,
1.251 schwarze 47: ROFF_af,
48: ROFF_aln,
49: ROFF_als,
1.80 kristaps 50: ROFF_am,
1.251 schwarze 51: ROFF_am1,
1.80 kristaps 52: ROFF_ami,
1.251 schwarze 53: ROFF_ami1,
1.193 schwarze 54: ROFF_as,
1.251 schwarze 55: ROFF_as1,
56: ROFF_asciify,
57: ROFF_backtrace,
58: ROFF_bd,
59: ROFF_bleedat,
60: ROFF_blm,
61: ROFF_box,
62: ROFF_boxa,
63: ROFF_bp,
64: ROFF_BP,
65: /* MAN_br, MDOC_br */
66: ROFF_break,
67: ROFF_breakchar,
68: ROFF_brnl,
69: ROFF_brp,
70: ROFF_brpnl,
71: ROFF_c2,
1.174 kristaps 72: ROFF_cc,
1.194 schwarze 73: ROFF_ce,
1.251 schwarze 74: ROFF_cf,
75: ROFF_cflags,
76: ROFF_ch,
77: ROFF_char,
78: ROFF_chop,
79: ROFF_class,
80: ROFF_close,
81: ROFF_CL,
82: ROFF_color,
83: ROFF_composite,
84: ROFF_continue,
85: ROFF_cp,
86: ROFF_cropat,
87: ROFF_cs,
88: ROFF_cu,
89: ROFF_da,
90: ROFF_dch,
91: ROFF_Dd,
1.80 kristaps 92: ROFF_de,
1.251 schwarze 93: ROFF_de1,
94: ROFF_defcolor,
1.80 kristaps 95: ROFF_dei,
1.251 schwarze 96: ROFF_dei1,
97: ROFF_device,
98: ROFF_devicem,
99: ROFF_di,
100: ROFF_do,
1.83 schwarze 101: ROFF_ds,
1.251 schwarze 102: ROFF_ds1,
103: ROFF_dwh,
104: ROFF_dt,
105: ROFF_ec,
106: ROFF_ecr,
107: ROFF_ecs,
1.82 kristaps 108: ROFF_el,
1.251 schwarze 109: ROFF_em,
110: ROFF_EN,
111: ROFF_eo,
112: ROFF_EP,
113: ROFF_EQ,
114: ROFF_errprint,
115: ROFF_ev,
116: ROFF_evc,
117: ROFF_ex,
118: ROFF_fallback,
1.185 schwarze 119: ROFF_fam,
1.251 schwarze 120: ROFF_fc,
121: ROFF_fchar,
122: ROFF_fcolor,
123: ROFF_fdeferlig,
124: ROFF_feature,
125: /* MAN_fi; ignored in mdoc(7) */
126: ROFF_fkern,
127: ROFF_fl,
128: ROFF_flig,
129: ROFF_fp,
130: ROFF_fps,
131: ROFF_fschar,
132: ROFF_fspacewidth,
133: ROFF_fspecial,
134: /* MAN_ft; ignored in mdoc(7) */
135: ROFF_ftr,
136: ROFF_fzoom,
137: ROFF_gcolor,
138: ROFF_hc,
139: ROFF_hcode,
140: ROFF_hidechar,
141: ROFF_hla,
142: ROFF_hlm,
143: ROFF_hpf,
144: ROFF_hpfa,
145: ROFF_hpfcode,
1.186 schwarze 146: ROFF_hw,
1.103 kristaps 147: ROFF_hy,
1.251 schwarze 148: ROFF_hylang,
149: ROFF_hylen,
150: ROFF_hym,
151: ROFF_hypp,
152: ROFF_hys,
1.82 kristaps 153: ROFF_ie,
1.75 kristaps 154: ROFF_if,
1.76 kristaps 155: ROFF_ig,
1.251 schwarze 156: /* MAN_in; ignored in mdoc(7) */
157: ROFF_index,
1.123 schwarze 158: ROFF_it,
1.251 schwarze 159: ROFF_itc,
160: ROFF_IX,
161: ROFF_kern,
162: ROFF_kernafter,
163: ROFF_kernbefore,
164: ROFF_kernpair,
165: ROFF_lc,
166: ROFF_lc_ctype,
167: ROFF_lds,
168: ROFF_length,
169: ROFF_letadj,
170: ROFF_lf,
171: ROFF_lg,
172: ROFF_lhang,
173: ROFF_linetabs,
174: /* MAN_ll, MDOC_ll */
175: ROFF_lnr,
176: ROFF_lnrf,
177: ROFF_lpfx,
178: ROFF_ls,
179: ROFF_lsm,
180: ROFF_lt,
181: ROFF_mc,
182: ROFF_mediasize,
183: ROFF_minss,
184: ROFF_mk,
185: ROFF_mso,
1.256 schwarze 186: ROFF_na,
1.103 kristaps 187: ROFF_ne,
1.251 schwarze 188: /* MAN_nf; ignored in mdoc(7) */
1.103 kristaps 189: ROFF_nh,
1.251 schwarze 190: ROFF_nhychar,
191: ROFF_nm,
192: ROFF_nn,
193: ROFF_nop,
1.104 kristaps 194: ROFF_nr,
1.251 schwarze 195: ROFF_nrf,
196: ROFF_nroff,
1.124 schwarze 197: ROFF_ns,
1.251 schwarze 198: ROFF_nx,
199: ROFF_open,
200: ROFF_opena,
201: ROFF_os,
202: ROFF_output,
203: ROFF_padj,
204: ROFF_papersize,
205: ROFF_pc,
206: ROFF_pev,
207: ROFF_pi,
208: ROFF_PI,
1.229 schwarze 209: ROFF_pl,
1.251 schwarze 210: ROFF_pm,
211: ROFF_pn,
212: ROFF_pnr,
213: ROFF_po,
1.124 schwarze 214: ROFF_ps,
1.251 schwarze 215: ROFF_psbb,
216: ROFF_pshape,
217: ROFF_pso,
218: ROFF_ptr,
219: ROFF_pvs,
220: ROFF_rchar,
221: ROFF_rd,
222: ROFF_recursionlimit,
223: ROFF_return,
224: ROFF_rfschar,
225: ROFF_rhang,
226: ROFF_rj,
1.83 schwarze 227: ROFF_rm,
1.251 schwarze 228: ROFF_rn,
229: ROFF_rnn,
1.203 schwarze 230: ROFF_rr,
1.251 schwarze 231: ROFF_rs,
232: ROFF_rt,
233: ROFF_schar,
234: ROFF_sentchar,
235: ROFF_shc,
236: ROFF_shift,
237: ROFF_sizes,
1.105 kristaps 238: ROFF_so,
1.251 schwarze 239: /* MAN_sp, MDOC_sp */
240: ROFF_spacewidth,
241: ROFF_special,
242: ROFF_spreadwarn,
243: ROFF_ss,
244: ROFF_sty,
245: ROFF_substring,
246: ROFF_sv,
247: ROFF_sy,
248: ROFF_T_,
1.124 schwarze 249: ROFF_ta,
1.251 schwarze 250: ROFF_tc,
251: ROFF_TE,
252: ROFF_TH,
253: ROFF_ti,
254: ROFF_tkf,
255: ROFF_tl,
256: ROFF_tm,
257: ROFF_tm1,
258: ROFF_tmc,
1.83 schwarze 259: ROFF_tr,
1.251 schwarze 260: ROFF_track,
261: ROFF_transchar,
262: ROFF_trf,
263: ROFF_trimat,
264: ROFF_trin,
265: ROFF_trnt,
266: ROFF_troff,
1.109 kristaps 267: ROFF_TS,
1.251 schwarze 268: ROFF_uf,
269: ROFF_ul,
270: ROFF_unformat,
271: ROFF_unwatch,
272: ROFF_unwatchn,
273: ROFF_vpt,
274: ROFF_vs,
275: ROFF_warn,
276: ROFF_warnscale,
277: ROFF_watch,
278: ROFF_watchlength,
279: ROFF_watchn,
280: ROFF_wh,
281: ROFF_while,
282: ROFF_write,
283: ROFF_writec,
284: ROFF_writem,
285: ROFF_xflag,
1.76 kristaps 286: ROFF_cblock,
1.106 kristaps 287: ROFF_USERDEF,
1.67 kristaps 288: ROFF_MAX
289: };
290:
1.147 kristaps 291: /*
1.167 kristaps 292: * An incredibly-simple string buffer.
293: */
1.94 kristaps 294: struct roffstr {
1.167 kristaps 295: char *p; /* nil-terminated buffer */
296: size_t sz; /* saved strlen(p) */
1.166 kristaps 297: };
298:
299: /*
1.167 kristaps 300: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 301: */
302: struct roffkv {
303: struct roffstr key;
304: struct roffstr val;
305: struct roffkv *next; /* next in list */
1.94 kristaps 306: };
307:
1.180 schwarze 308: /*
309: * A single number register as part of a singly-linked list.
310: */
311: struct roffreg {
312: struct roffstr key;
1.181 schwarze 313: int val;
1.180 schwarze 314: struct roffreg *next;
315: };
316:
1.67 kristaps 317: struct roff {
1.128 kristaps 318: struct mparse *parse; /* parse point */
1.237 schwarze 319: const struct mchars *mchars; /* character table */
1.67 kristaps 320: struct roffnode *last; /* leaf of stack */
1.223 schwarze 321: int *rstack; /* stack of inverted `ie' values */
1.180 schwarze 322: struct roffreg *regtab; /* number registers */
1.166 kristaps 323: struct roffkv *strtab; /* user-defined strings & macros */
1.167 kristaps 324: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
325: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 326: const char *current_string; /* value of last called user macro */
1.118 kristaps 327: struct tbl_node *first_tbl; /* first table parsed */
328: struct tbl_node *last_tbl; /* last table parsed */
329: struct tbl_node *tbl; /* current table being parsed */
1.125 kristaps 330: struct eqn_node *last_eqn; /* last equation parsed */
331: struct eqn_node *first_eqn; /* first equation parsed */
332: struct eqn_node *eqn; /* current equation being parsed */
1.230 schwarze 333: int eqn_inline; /* current equation is inline */
1.223 schwarze 334: int options; /* parse options */
335: int rstacksz; /* current size limit of rstack */
336: int rstackpos; /* position in rstack */
1.227 schwarze 337: int format; /* current file in mdoc or man format */
1.273 schwarze 338: int argc; /* number of args of the last macro */
1.223 schwarze 339: char control; /* control character */
1.79 kristaps 340: };
341:
1.67 kristaps 342: struct roffnode {
343: enum rofft tok; /* type of node */
344: struct roffnode *parent; /* up one in stack */
345: int line; /* parse line */
346: int col; /* parse col */
1.106 kristaps 347: char *name; /* node name, e.g. macro name */
1.79 kristaps 348: char *end; /* end-rules: custom token */
349: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 350: int rule; /* current evaluation rule */
1.67 kristaps 351: };
352:
353: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.72 kristaps 354: enum rofft tok, /* tok of macro */ \
1.238 schwarze 355: struct buf *buf, /* input buffer */ \
1.67 kristaps 356: int ln, /* parse line */ \
1.75 kristaps 357: int ppos, /* original pos in buffer */ \
358: int pos, /* current pos in buffer */ \
1.74 kristaps 359: int *offs /* reset offset of buffer data */
1.67 kristaps 360:
361: typedef enum rofferr (*roffproc)(ROFF_ARGS);
362:
363: struct roffmac {
364: const char *name; /* macro name */
1.79 kristaps 365: roffproc proc; /* process new macro */
366: roffproc text; /* process as child text of macro */
367: roffproc sub; /* process as child of macro */
368: int flags;
369: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.85 kristaps 370: struct roffmac *next;
1.67 kristaps 371: };
372:
1.141 kristaps 373: struct predef {
374: const char *name; /* predefined input name */
375: const char *str; /* replacement symbol */
376: };
377:
378: #define PREDEF(__name, __str) \
379: { (__name), (__str) },
380:
1.266 schwarze 381: /* --- function prototypes ------------------------------------------------ */
382:
1.155 kristaps 383: static enum rofft roffhash_find(const char *, size_t);
384: static void roffhash_init(void);
385: static void roffnode_cleanscope(struct roff *);
386: static void roffnode_pop(struct roff *);
387: static void roffnode_push(struct roff *, enum rofft,
388: const char *, int, int);
1.80 kristaps 389: static enum rofferr roff_block(ROFF_ARGS);
390: static enum rofferr roff_block_text(ROFF_ARGS);
391: static enum rofferr roff_block_sub(ROFF_ARGS);
1.251 schwarze 392: static enum rofferr roff_brp(ROFF_ARGS);
1.80 kristaps 393: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 394: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 395: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 396: static enum rofferr roff_cond(ROFF_ARGS);
397: static enum rofferr roff_cond_text(ROFF_ARGS);
398: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 399: static enum rofferr roff_ds(ROFF_ARGS);
1.238 schwarze 400: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 401: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 402: static int roff_evalnum(struct roff *, int,
403: const char *, int *, int *, int);
404: static int roff_evalpar(struct roff *, int,
1.261 schwarze 405: const char *, int *, int *, int);
1.198 schwarze 406: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 407: static void roff_free1(struct roff *);
1.180 schwarze 408: static void roff_freereg(struct roffreg *);
1.167 kristaps 409: static void roff_freestr(struct roffkv *);
1.212 schwarze 410: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 411: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 412: static int roff_getop(const char *, int *, char *);
1.181 schwarze 413: static int roff_getregn(const struct roff *,
414: const char *, size_t);
1.273 schwarze 415: static int roff_getregro(const struct roff *,
416: const char *name);
1.207 schwarze 417: static const char *roff_getstrn(const struct roff *,
1.94 kristaps 418: const char *, size_t);
1.271 schwarze 419: static int roff_hasregn(const struct roff *,
420: const char *, size_t);
1.251 schwarze 421: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 422: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 423: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 schwarze 424: static void roff_man_alloc1(struct roff_man *);
425: static void roff_man_free1(struct roff_man *);
1.89 kristaps 426: static enum rofferr roff_nr(ROFF_ARGS);
1.214 schwarze 427: static enum rofft roff_parse(struct roff *, char *, int *,
428: int, int);
1.238 schwarze 429: static enum rofferr roff_parsetext(struct buf *, int, int *);
430: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 431: static enum rofferr roff_rm(ROFF_ARGS);
1.203 schwarze 432: static enum rofferr roff_rr(ROFF_ARGS);
1.94 kristaps 433: static void roff_setstr(struct roff *,
1.106 kristaps 434: const char *, const char *, int);
1.207 schwarze 435: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 436: size_t, const char *, size_t, int);
1.105 kristaps 437: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 438: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 439: static enum rofferr roff_Dd(ROFF_ARGS);
440: static enum rofferr roff_TH(ROFF_ARGS);
1.109 kristaps 441: static enum rofferr roff_TE(ROFF_ARGS);
442: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 443: static enum rofferr roff_EQ(ROFF_ARGS);
444: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 445: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 446: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 447: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 448:
1.266 schwarze 449: /* --- constant data ------------------------------------------------------ */
450:
1.155 kristaps 451: /* See roffhash_find() */
1.85 kristaps 452:
453: #define ASCII_HI 126
454: #define ASCII_LO 33
455: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
456:
1.261 schwarze 457: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
458: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
459:
1.85 kristaps 460: static struct roffmac *hash[HASHWIDTH];
461:
462: static struct roffmac roffs[ROFF_MAX] = {
1.251 schwarze 463: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.103 kristaps 464: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 465: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
466: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
467: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 468: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 469: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.85 kristaps 470: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 471: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.193 schwarze 472: { "as", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 473: { "as1", roff_ds, NULL, NULL, 0, NULL },
474: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
475: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
476: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
477: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
478: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
479: { "box", roff_unsupp, NULL, NULL, 0, NULL },
480: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
481: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
482: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
483: { "break", roff_unsupp, NULL, NULL, 0, NULL },
484: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
485: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
486: { "brp", roff_brp, NULL, NULL, 0, NULL },
487: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
488: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.174 kristaps 489: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.194 schwarze 490: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 491: { "cf", roff_insec, NULL, NULL, 0, NULL },
492: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
493: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
494: { "char", roff_unsupp, NULL, NULL, 0, NULL },
495: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
496: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
497: { "close", roff_insec, NULL, NULL, 0, NULL },
498: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
499: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
500: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
501: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
502: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
503: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
504: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
505: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
506: { "da", roff_unsupp, NULL, NULL, 0, NULL },
507: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
508: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.85 kristaps 509: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 510: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
511: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 512: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 513: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
514: { "device", roff_unsupp, NULL, NULL, 0, NULL },
515: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
516: { "di", roff_unsupp, NULL, NULL, 0, NULL },
517: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.92 schwarze 518: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 519: { "ds1", roff_ds, NULL, NULL, 0, NULL },
520: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
521: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
522: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
523: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
524: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 525: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.251 schwarze 526: { "em", roff_unsupp, NULL, NULL, 0, NULL },
527: { "EN", roff_EN, NULL, NULL, 0, NULL },
528: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
529: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
530: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
531: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
532: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
533: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
534: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
535: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.185 schwarze 536: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 537: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
538: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
539: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
543: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
544: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
545: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
548: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
551: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
552: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
553: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
554: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
555: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
556: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
557: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
558: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
559: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
560: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.186 schwarze 561: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 562: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 563: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
564: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
565: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
566: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
567: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 568: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
569: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
570: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 571: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.178 schwarze 572: { "it", roff_it, NULL, NULL, 0, NULL },
1.251 schwarze 573: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
574: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
575: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
576: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
577: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
578: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
579: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
580: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
581: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
582: { "length", roff_unsupp, NULL, NULL, 0, NULL },
583: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
584: { "lf", roff_insec, NULL, NULL, 0, NULL },
585: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
586: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
587: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
588: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
589: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
590: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
591: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
592: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
593: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
594: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
595: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
596: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
597: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
598: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.256 schwarze 599: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 600: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
601: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 602: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
603: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
604: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
605: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.104 kristaps 606: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.251 schwarze 607: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
608: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 609: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 610: { "nx", roff_insec, NULL, NULL, 0, NULL },
611: { "open", roff_insec, NULL, NULL, 0, NULL },
612: { "opena", roff_insec, NULL, NULL, 0, NULL },
613: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "output", roff_unsupp, NULL, NULL, 0, NULL },
615: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
616: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
617: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
618: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
619: { "pi", roff_insec, NULL, NULL, 0, NULL },
620: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.229 schwarze 621: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 622: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
623: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
624: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
625: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 626: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 627: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.255 schwarze 628: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 629: { "pso", roff_insec, NULL, NULL, 0, NULL },
630: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
631: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
632: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
633: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
634: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
635: { "return", roff_unsupp, NULL, NULL, 0, NULL },
636: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
637: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
638: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.122 schwarze 639: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.251 schwarze 640: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
641: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.203 schwarze 642: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.251 schwarze 643: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
644: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
645: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
646: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
647: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
648: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
649: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.105 kristaps 650: { "so", roff_so, NULL, NULL, 0, NULL },
1.251 schwarze 651: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
652: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
653: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
654: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
655: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
656: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
657: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
658: { "sy", roff_insec, NULL, NULL, 0, NULL },
659: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.255 schwarze 660: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 661: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
662: { "TE", roff_TE, NULL, NULL, 0, NULL },
663: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.255 schwarze 664: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 665: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
666: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
667: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
668: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.164 kristaps 670: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.251 schwarze 671: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
672: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
673: { "trf", roff_insec, NULL, NULL, 0, NULL },
674: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
675: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
676: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
677: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.109 kristaps 678: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.251 schwarze 679: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
680: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
681: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
682: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
683: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
684: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
685: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
686: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
687: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
688: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
689: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
690: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
691: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
692: { "while", roff_unsupp, NULL, NULL, 0, NULL },
693: { "write", roff_insec, NULL, NULL, 0, NULL },
694: { "writec", roff_insec, NULL, NULL, 0, NULL },
695: { "writem", roff_insec, NULL, NULL, 0, NULL },
696: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 697: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.106 kristaps 698: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.67 kristaps 699: };
700:
1.200 schwarze 701: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175 schwarze 702: const char *const __mdoc_reserved[] = {
703: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
704: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
705: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
706: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200 schwarze 707: "Dt", "Dv", "Dx", "D1",
708: "Ec", "Ed", "Ef", "Ek", "El", "Em",
709: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175 schwarze 710: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200 schwarze 711: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
712: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175 schwarze 713: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200 schwarze 714: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
715: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
716: "Sc", "Sh", "Sm", "So", "Sq",
1.175 schwarze 717: "Ss", "St", "Sx", "Sy",
718: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200 schwarze 719: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175 schwarze 720: "%P", "%Q", "%R", "%T", "%U", "%V",
721: NULL
722: };
723:
1.200 schwarze 724: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175 schwarze 725: const char *const __man_reserved[] = {
1.200 schwarze 726: "AT", "B", "BI", "BR", "DT",
727: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
728: "LP", "OP", "P", "PD", "PP",
729: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
730: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175 schwarze 731: NULL
732: };
733:
1.141 kristaps 734: /* Array of injected predefined strings. */
735: #define PREDEFS_MAX 38
736: static const struct predef predefs[PREDEFS_MAX] = {
737: #include "predefs.in"
738: };
739:
1.155 kristaps 740: /* See roffhash_find() */
1.85 kristaps 741: #define ROFF_HASH(p) (p[0] - ASCII_LO)
742:
1.178 schwarze 743: static int roffit_lines; /* number of lines to delay */
744: static char *roffit_macro; /* nil-terminated macro line */
745:
1.207 schwarze 746:
1.266 schwarze 747: /* --- request table ------------------------------------------------------ */
748:
1.85 kristaps 749: static void
1.155 kristaps 750: roffhash_init(void)
1.85 kristaps 751: {
752: struct roffmac *n;
753: int buc, i;
754:
1.106 kristaps 755: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.85 kristaps 756: assert(roffs[i].name[0] >= ASCII_LO);
757: assert(roffs[i].name[0] <= ASCII_HI);
758:
759: buc = ROFF_HASH(roffs[i].name);
760:
761: if (NULL != (n = hash[buc])) {
762: for ( ; n->next; n = n->next)
763: /* Do nothing. */ ;
764: n->next = &roffs[i];
765: } else
766: hash[buc] = &roffs[i];
767: }
768: }
769:
1.67 kristaps 770: /*
771: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
772: * the nil-terminated string name could be found.
773: */
774: static enum rofft
1.155 kristaps 775: roffhash_find(const char *p, size_t s)
1.67 kristaps 776: {
1.85 kristaps 777: int buc;
778: struct roffmac *n;
1.67 kristaps 779:
1.85 kristaps 780: /*
781: * libroff has an extremely simple hashtable, for the time
782: * being, which simply keys on the first character, which must
783: * be printable, then walks a chain. It works well enough until
784: * optimised.
785: */
786:
787: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
788: return(ROFF_MAX);
789:
790: buc = ROFF_HASH(p);
791:
792: if (NULL == (n = hash[buc]))
793: return(ROFF_MAX);
794: for ( ; n; n = n->next)
1.106 kristaps 795: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.85 kristaps 796: return((enum rofft)(n - roffs));
1.67 kristaps 797:
798: return(ROFF_MAX);
799: }
800:
1.266 schwarze 801: /* --- stack of request blocks -------------------------------------------- */
802:
1.67 kristaps 803: /*
804: * Pop the current node off of the stack of roff instructions currently
805: * pending.
806: */
807: static void
808: roffnode_pop(struct roff *r)
809: {
810: struct roffnode *p;
811:
1.75 kristaps 812: assert(r->last);
1.207 schwarze 813: p = r->last;
1.82 kristaps 814:
1.75 kristaps 815: r->last = r->last->parent;
1.106 kristaps 816: free(p->name);
817: free(p->end);
1.67 kristaps 818: free(p);
819: }
820:
821: /*
822: * Push a roff node onto the instruction stack. This must later be
823: * removed with roffnode_pop().
824: */
1.98 schwarze 825: static void
1.106 kristaps 826: roffnode_push(struct roff *r, enum rofft tok, const char *name,
827: int line, int col)
1.67 kristaps 828: {
829: struct roffnode *p;
830:
1.98 schwarze 831: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 832: p->tok = tok;
1.106 kristaps 833: if (name)
834: p->name = mandoc_strdup(name);
1.67 kristaps 835: p->parent = r->last;
836: p->line = line;
837: p->col = col;
1.198 schwarze 838: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 839:
840: r->last = p;
841: }
842:
1.266 schwarze 843: /* --- roff parser state data management ---------------------------------- */
844:
1.67 kristaps 845: static void
846: roff_free1(struct roff *r)
847: {
1.176 schwarze 848: struct tbl_node *tbl;
1.125 kristaps 849: struct eqn_node *e;
1.167 kristaps 850: int i;
1.67 kristaps 851:
1.176 schwarze 852: while (NULL != (tbl = r->first_tbl)) {
853: r->first_tbl = tbl->next;
854: tbl_free(tbl);
1.109 kristaps 855: }
1.113 kristaps 856: r->first_tbl = r->last_tbl = r->tbl = NULL;
857:
1.125 kristaps 858: while (NULL != (e = r->first_eqn)) {
859: r->first_eqn = e->next;
860: eqn_free(e);
861: }
862: r->first_eqn = r->last_eqn = r->eqn = NULL;
863:
1.67 kristaps 864: while (r->last)
865: roffnode_pop(r);
1.109 kristaps 866:
1.223 schwarze 867: free (r->rstack);
868: r->rstack = NULL;
869: r->rstacksz = 0;
870: r->rstackpos = -1;
871:
872: roff_freereg(r->regtab);
873: r->regtab = NULL;
874:
1.167 kristaps 875: roff_freestr(r->strtab);
876: roff_freestr(r->xmbtab);
877: r->strtab = r->xmbtab = NULL;
878:
879: if (r->xtab)
880: for (i = 0; i < 128; i++)
881: free(r->xtab[i].p);
882: free(r->xtab);
883: r->xtab = NULL;
1.67 kristaps 884: }
885:
886: void
887: roff_reset(struct roff *r)
888: {
889:
890: roff_free1(r);
1.227 schwarze 891: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174 kristaps 892: r->control = 0;
1.67 kristaps 893: }
894:
895: void
896: roff_free(struct roff *r)
897: {
898:
899: roff_free1(r);
900: free(r);
901: }
902:
903: struct roff *
1.237 schwarze 904: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.67 kristaps 905: {
906: struct roff *r;
907:
1.98 schwarze 908: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 909: r->parse = parse;
1.237 schwarze 910: r->mchars = mchars;
1.199 schwarze 911: r->options = options;
1.227 schwarze 912: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 913: r->rstackpos = -1;
1.207 schwarze 914:
1.155 kristaps 915: roffhash_init();
1.141 kristaps 916:
1.67 kristaps 917: return(r);
1.265 schwarze 918: }
919:
1.266 schwarze 920: /* --- syntax tree state data management ---------------------------------- */
921:
1.265 schwarze 922: static void
923: roff_man_free1(struct roff_man *man)
924: {
925:
1.266 schwarze 926: if (man->first != NULL)
927: roff_node_delete(man, man->first);
1.265 schwarze 928: free(man->meta.msec);
929: free(man->meta.vol);
930: free(man->meta.os);
931: free(man->meta.arch);
932: free(man->meta.title);
933: free(man->meta.name);
934: free(man->meta.date);
935: }
936:
937: static void
938: roff_man_alloc1(struct roff_man *man)
939: {
940:
941: memset(&man->meta, 0, sizeof(man->meta));
942: man->first = mandoc_calloc(1, sizeof(*man->first));
943: man->first->type = ROFFT_ROOT;
944: man->last = man->first;
945: man->last_es = NULL;
946: man->flags = 0;
947: man->macroset = MACROSET_NONE;
948: man->lastsec = man->lastnamed = SEC_NONE;
949: man->next = ROFF_NEXT_CHILD;
950: }
951:
952: void
953: roff_man_reset(struct roff_man *man)
954: {
955:
956: roff_man_free1(man);
957: roff_man_alloc1(man);
958: }
959:
960: void
961: roff_man_free(struct roff_man *man)
962: {
963:
964: roff_man_free1(man);
965: free(man);
966: }
967:
968: struct roff_man *
969: roff_man_alloc(struct roff *roff, struct mparse *parse,
970: const char *defos, int quick)
971: {
972: struct roff_man *man;
973:
974: man = mandoc_calloc(1, sizeof(*man));
975: man->parse = parse;
976: man->roff = roff;
977: man->defos = defos;
978: man->quick = quick;
979: roff_man_alloc1(man);
980: return(man);
1.67 kristaps 981: }
982:
1.266 schwarze 983: /* --- syntax tree handling ----------------------------------------------- */
984:
985: struct roff_node *
986: roff_node_alloc(struct roff_man *man, int line, int pos,
987: enum roff_type type, int tok)
988: {
989: struct roff_node *n;
990:
991: n = mandoc_calloc(1, sizeof(*n));
992: n->line = line;
993: n->pos = pos;
994: n->tok = tok;
995: n->type = type;
996: n->sec = man->lastsec;
997:
998: if (man->flags & MDOC_SYNOPSIS)
999: n->flags |= MDOC_SYNPRETTY;
1000: else
1001: n->flags &= ~MDOC_SYNPRETTY;
1002: if (man->flags & MDOC_NEWLINE)
1003: n->flags |= MDOC_LINE;
1004: man->flags &= ~MDOC_NEWLINE;
1005:
1006: return(n);
1007: }
1008:
1009: void
1010: roff_node_append(struct roff_man *man, struct roff_node *n)
1011: {
1012:
1013: switch (man->next) {
1014: case ROFF_NEXT_SIBLING:
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;
1022: break;
1023: default:
1024: abort();
1025: /* NOTREACHED */
1026: }
1027: n->parent->nchild++;
1.270 schwarze 1028: n->parent->last = n;
1.266 schwarze 1029:
1030: /*
1031: * Copy over the normalised-data pointer of our parent. Not
1032: * everybody has one, but copying a null pointer is fine.
1033: */
1034:
1035: switch (n->type) {
1036: case ROFFT_BODY:
1037: if (n->end != ENDBODY_NOT)
1038: break;
1039: /* FALLTHROUGH */
1040: case ROFFT_TAIL:
1041: /* FALLTHROUGH */
1042: case ROFFT_HEAD:
1043: n->norm = n->parent->norm;
1044: break;
1045: default:
1046: break;
1047: }
1048:
1049: if (man->macroset == MACROSET_MDOC)
1050: mdoc_valid_pre(man, n);
1051:
1052: switch (n->type) {
1053: case ROFFT_HEAD:
1054: assert(n->parent->type == ROFFT_BLOCK);
1055: n->parent->head = n;
1056: break;
1057: case ROFFT_BODY:
1058: if (n->end)
1059: break;
1060: assert(n->parent->type == ROFFT_BLOCK);
1061: n->parent->body = n;
1062: break;
1063: case ROFFT_TAIL:
1064: assert(n->parent->type == ROFFT_BLOCK);
1065: n->parent->tail = n;
1066: break;
1067: default:
1068: break;
1069: }
1070: man->last = n;
1071: }
1072:
1.267 schwarze 1073: void
1074: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
1075: {
1076: struct roff_node *n;
1077:
1078: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
1079: n->string = roff_strdup(man->roff, word);
1080: roff_node_append(man, n);
1081: if (man->macroset == MACROSET_MDOC)
1082: mdoc_valid_post(man);
1083: else
1084: man_valid_post(man);
1085: man->next = ROFF_NEXT_SIBLING;
1086: }
1087:
1088: void
1089: roff_word_append(struct roff_man *man, const char *word)
1090: {
1091: struct roff_node *n;
1092: char *addstr, *newstr;
1093:
1094: n = man->last;
1095: addstr = roff_strdup(man->roff, word);
1096: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
1097: free(addstr);
1098: free(n->string);
1099: n->string = newstr;
1100: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 1101: }
1102:
1103: void
1104: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
1105: {
1106: struct roff_node *n;
1107:
1108: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
1109: roff_node_append(man, n);
1110: man->next = ROFF_NEXT_CHILD;
1111: }
1112:
1113: struct roff_node *
1114: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
1115: {
1116: struct roff_node *n;
1117:
1118: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
1119: roff_node_append(man, n);
1120: man->next = ROFF_NEXT_CHILD;
1121: return(n);
1.267 schwarze 1122: }
1123:
1.266 schwarze 1124: struct roff_node *
1125: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
1126: {
1127: struct roff_node *n;
1128:
1129: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
1130: roff_node_append(man, n);
1131: man->next = ROFF_NEXT_CHILD;
1132: return(n);
1133: }
1134:
1135: struct roff_node *
1136: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1137: {
1138: struct roff_node *n;
1139:
1140: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1141: roff_node_append(man, n);
1142: man->next = ROFF_NEXT_CHILD;
1143: return(n);
1.267 schwarze 1144: }
1145:
1146: void
1147: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
1148: {
1149: struct roff_node *n;
1150:
1151: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
1152: n->eqn = eqn;
1153: if (eqn->ln > man->last->line)
1154: n->flags |= MDOC_LINE;
1155: roff_node_append(man, n);
1156: man->next = ROFF_NEXT_SIBLING;
1157: }
1158:
1159: void
1160: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
1161: {
1162: struct roff_node *n;
1163:
1164: if (man->macroset == MACROSET_MAN)
1165: man_breakscope(man, TOKEN_NONE);
1166: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
1167: n->span = tbl;
1168: roff_node_append(man, n);
1169: if (man->macroset == MACROSET_MDOC)
1170: mdoc_valid_post(man);
1171: else
1172: man_valid_post(man);
1173: man->next = ROFF_NEXT_SIBLING;
1.266 schwarze 1174: }
1175:
1176: void
1177: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1178: {
1179:
1180: /* Adjust siblings. */
1181:
1182: if (n->prev)
1183: n->prev->next = n->next;
1184: if (n->next)
1185: n->next->prev = n->prev;
1186:
1187: /* Adjust parent. */
1188:
1189: if (n->parent != NULL) {
1190: n->parent->nchild--;
1191: if (n->parent->child == n)
1192: n->parent->child = n->next;
1193: if (n->parent->last == n)
1194: n->parent->last = n->prev;
1195: }
1196:
1197: /* Adjust parse point. */
1198:
1199: if (man == NULL)
1200: return;
1201: if (man->last == n) {
1202: if (n->prev == NULL) {
1203: man->last = n->parent;
1204: man->next = ROFF_NEXT_CHILD;
1205: } else {
1206: man->last = n->prev;
1207: man->next = ROFF_NEXT_SIBLING;
1208: }
1209: }
1210: if (man->first == n)
1211: man->first = NULL;
1212: }
1213:
1214: void
1215: roff_node_free(struct roff_node *n)
1216: {
1217:
1218: if (n->args != NULL)
1219: mdoc_argv_free(n->args);
1220: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1221: free(n->norm);
1222: free(n->string);
1223: free(n);
1224: }
1225:
1226: void
1227: roff_node_delete(struct roff_man *man, struct roff_node *n)
1228: {
1229:
1230: while (n->child != NULL)
1231: roff_node_delete(man, n->child);
1232: assert(n->nchild == 0);
1233: roff_node_unlink(man, n);
1234: roff_node_free(n);
1.269 schwarze 1235: }
1236:
1237: void
1238: deroff(char **dest, const struct roff_node *n)
1239: {
1240: char *cp;
1241: size_t sz;
1242:
1243: if (n->type != ROFFT_TEXT) {
1244: for (n = n->child; n != NULL; n = n->next)
1245: deroff(dest, n);
1246: return;
1247: }
1248:
1249: /* Skip leading whitespace and escape sequences. */
1250:
1251: cp = n->string;
1252: while (*cp != '\0') {
1253: if ('\\' == *cp) {
1254: cp++;
1255: mandoc_escape((const char **)&cp, NULL, NULL);
1256: } else if (isspace((unsigned char)*cp))
1257: cp++;
1258: else
1259: break;
1260: }
1261:
1262: /* Skip trailing whitespace. */
1263:
1264: for (sz = strlen(cp); sz; sz--)
1265: if ( ! isspace((unsigned char)cp[sz-1]))
1266: break;
1267:
1268: /* Skip empty strings. */
1269:
1270: if (sz == 0)
1271: return;
1272:
1273: if (*dest == NULL) {
1274: *dest = mandoc_strndup(cp, sz);
1275: return;
1276: }
1277:
1278: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1279: free(*dest);
1280: *dest = cp;
1.266 schwarze 1281: }
1282:
1283: /* --- main functions of the roff parser ---------------------------------- */
1284:
1.94 kristaps 1285: /*
1.206 schwarze 1286: * In the current line, expand escape sequences that tend to get
1287: * used in numerical expressions and conditional requests.
1288: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1289: */
1.172 schwarze 1290: static enum rofferr
1.238 schwarze 1291: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1292: {
1.208 schwarze 1293: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1294: const char *start; /* start of the string to process */
1.209 schwarze 1295: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1296: const char *stnam; /* start of the name, after "[(*" */
1297: const char *cp; /* end of the name, e.g. before ']' */
1298: const char *res; /* the string to be substituted */
1.238 schwarze 1299: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1300: size_t maxl; /* expected length of the escape name */
1301: size_t naml; /* actual length of the escape name */
1.237 schwarze 1302: enum mandoc_esc esc; /* type of the escape sequence */
1303: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1304: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1305: int npos; /* position in numeric expression */
1.218 schwarze 1306: int arg_complete; /* argument not interrupted by eol */
1.206 schwarze 1307: char term; /* character terminating the escape */
1.94 kristaps 1308:
1.170 schwarze 1309: expand_count = 0;
1.238 schwarze 1310: start = buf->buf + pos;
1.205 schwarze 1311: stesc = strchr(start, '\0') - 1;
1312: while (stesc-- > start) {
1.170 schwarze 1313:
1.205 schwarze 1314: /* Search backwards for the next backslash. */
1315:
1.238 schwarze 1316: if (*stesc != '\\')
1.205 schwarze 1317: continue;
1318:
1319: /* If it is escaped, skip it. */
1320:
1321: for (cp = stesc - 1; cp >= start; cp--)
1.238 schwarze 1322: if (*cp != '\\')
1.205 schwarze 1323: break;
1324:
1.238 schwarze 1325: if ((stesc - cp) % 2 == 0) {
1.209 schwarze 1326: stesc = (char *)cp;
1.205 schwarze 1327: continue;
1328: }
1.108 schwarze 1329:
1.206 schwarze 1330: /* Decide whether to expand or to check only. */
1.108 schwarze 1331:
1.206 schwarze 1332: term = '\0';
1.205 schwarze 1333: cp = stesc + 1;
1.181 schwarze 1334: switch (*cp) {
1.207 schwarze 1335: case '*':
1.181 schwarze 1336: res = NULL;
1337: break;
1.207 schwarze 1338: case 'B':
1.206 schwarze 1339: /* FALLTHROUGH */
1.207 schwarze 1340: case 'w':
1.206 schwarze 1341: term = cp[1];
1342: /* FALLTHROUGH */
1.207 schwarze 1343: case 'n':
1.181 schwarze 1344: res = ubuf;
1345: break;
1346: default:
1.237 schwarze 1347: esc = mandoc_escape(&cp, &stnam, &inaml);
1348: if (esc == ESCAPE_ERROR ||
1349: (esc == ESCAPE_SPECIAL &&
1350: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.219 schwarze 1351: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1352: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1353: "%.*s", (int)(cp - stesc), stesc);
1.205 schwarze 1354: continue;
1.152 kristaps 1355: }
1356:
1.205 schwarze 1357: if (EXPAND_LIMIT < ++expand_count) {
1358: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1359: ln, (int)(stesc - buf->buf), NULL);
1.205 schwarze 1360: return(ROFF_IGN);
1361: }
1.108 schwarze 1362:
1363: /*
1364: * The third character decides the length
1.181 schwarze 1365: * of the name of the string or register.
1.108 schwarze 1366: * Save a pointer to the name.
1367: */
1368:
1.238 schwarze 1369: if (term == '\0') {
1.206 schwarze 1370: switch (*++cp) {
1.207 schwarze 1371: case '\0':
1.206 schwarze 1372: maxl = 0;
1373: break;
1.207 schwarze 1374: case '(':
1.206 schwarze 1375: cp++;
1376: maxl = 2;
1377: break;
1.207 schwarze 1378: case '[':
1.206 schwarze 1379: cp++;
1380: term = ']';
1381: maxl = 0;
1382: break;
1383: default:
1384: maxl = 1;
1385: break;
1386: }
1387: } else {
1388: cp += 2;
1.94 kristaps 1389: maxl = 0;
1390: }
1.108 schwarze 1391: stnam = cp;
1.94 kristaps 1392:
1.108 schwarze 1393: /* Advance to the end of the name. */
1.94 kristaps 1394:
1.253 schwarze 1395: naml = 0;
1.218 schwarze 1396: arg_complete = 1;
1.253 schwarze 1397: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1398: if (*cp == '\0') {
1.219 schwarze 1399: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1400: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1401: arg_complete = 0;
1.206 schwarze 1402: break;
1.153 kristaps 1403: }
1.238 schwarze 1404: if (maxl == 0 && *cp == term) {
1.206 schwarze 1405: cp++;
1.253 schwarze 1406: break;
1407: }
1408: if (*cp++ != '\\' || stesc[1] != 'w') {
1409: naml++;
1410: continue;
1411: }
1412: switch (mandoc_escape(&cp, NULL, NULL)) {
1413: case ESCAPE_SPECIAL:
1414: /* FALLTHROUGH */
1415: case ESCAPE_UNICODE:
1416: /* FALLTHROUGH */
1417: case ESCAPE_NUMBERED:
1418: /* FALLTHROUGH */
1419: case ESCAPE_OVERSTRIKE:
1420: naml++;
1421: break;
1422: default:
1.94 kristaps 1423: break;
1.206 schwarze 1424: }
1.94 kristaps 1425: }
1426:
1.108 schwarze 1427: /*
1428: * Retrieve the replacement string; if it is
1429: * undefined, resume searching for escapes.
1430: */
1431:
1.206 schwarze 1432: switch (stesc[1]) {
1.207 schwarze 1433: case '*':
1.218 schwarze 1434: if (arg_complete)
1435: res = roff_getstrn(r, stnam, naml);
1.206 schwarze 1436: break;
1.207 schwarze 1437: case 'B':
1.206 schwarze 1438: npos = 0;
1.218 schwarze 1439: ubuf[0] = arg_complete &&
1.261 schwarze 1440: roff_evalnum(r, ln, stnam, &npos,
1441: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1442: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1443: ubuf[1] = '\0';
1444: break;
1.207 schwarze 1445: case 'n':
1.218 schwarze 1446: if (arg_complete)
1447: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1448: roff_getregn(r, stnam, naml));
1449: else
1450: ubuf[0] = '\0';
1.206 schwarze 1451: break;
1.207 schwarze 1452: case 'w':
1.218 schwarze 1453: /* use even incomplete args */
1.208 schwarze 1454: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1455: 24 * (int)naml);
1456: break;
1457: }
1.94 kristaps 1458:
1.238 schwarze 1459: if (res == NULL) {
1.219 schwarze 1460: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1461: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1462: "%.*s", (int)naml, stnam);
1.142 kristaps 1463: res = "";
1.246 schwarze 1464: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1465: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1466: ln, (int)(stesc - buf->buf), NULL);
1467: return(ROFF_IGN);
1.94 kristaps 1468: }
1469:
1.108 schwarze 1470: /* Replace the escape sequence by the string. */
1471:
1.209 schwarze 1472: *stesc = '\0';
1.238 schwarze 1473: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1474: buf->buf, res, cp) + 1;
1.94 kristaps 1475:
1.205 schwarze 1476: /* Prepare for the next replacement. */
1.94 kristaps 1477:
1.205 schwarze 1478: start = nbuf + pos;
1.238 schwarze 1479: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1480: free(buf->buf);
1481: buf->buf = nbuf;
1.154 kristaps 1482: }
1.172 schwarze 1483: return(ROFF_CONT);
1.154 kristaps 1484: }
1485:
1486: /*
1.178 schwarze 1487: * Process text streams:
1488: * Convert all breakable hyphens into ASCII_HYPH.
1489: * Decrement and spring input line trap.
1.154 kristaps 1490: */
1491: static enum rofferr
1.238 schwarze 1492: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154 kristaps 1493: {
1494: size_t sz;
1495: const char *start;
1.178 schwarze 1496: char *p;
1497: int isz;
1.154 kristaps 1498: enum mandoc_esc esc;
1499:
1.238 schwarze 1500: start = p = buf->buf + pos;
1.154 kristaps 1501:
1.238 schwarze 1502: while (*p != '\0') {
1.154 kristaps 1503: sz = strcspn(p, "-\\");
1504: p += sz;
1505:
1.238 schwarze 1506: if (*p == '\0')
1.159 kristaps 1507: break;
1508:
1.238 schwarze 1509: if (*p == '\\') {
1.154 kristaps 1510: /* Skip over escapes. */
1511: p++;
1.189 schwarze 1512: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1513: if (esc == ESCAPE_ERROR)
1.154 kristaps 1514: break;
1.264 schwarze 1515: while (*p == '-')
1516: p++;
1.155 kristaps 1517: continue;
1.159 kristaps 1518: } else if (p == start) {
1.158 kristaps 1519: p++;
1.155 kristaps 1520: continue;
1.158 kristaps 1521: }
1.155 kristaps 1522:
1.171 schwarze 1523: if (isalpha((unsigned char)p[-1]) &&
1524: isalpha((unsigned char)p[1]))
1.155 kristaps 1525: *p = ASCII_HYPH;
1526: p++;
1.94 kristaps 1527: }
1528:
1.178 schwarze 1529: /* Spring the input line trap. */
1.238 schwarze 1530: if (roffit_lines == 1) {
1531: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1532: free(buf->buf);
1533: buf->buf = p;
1534: buf->sz = isz + 1;
1.178 schwarze 1535: *offs = 0;
1536: free(roffit_macro);
1537: roffit_lines = 0;
1538: return(ROFF_REPARSE);
1.238 schwarze 1539: } else if (roffit_lines > 1)
1.178 schwarze 1540: --roffit_lines;
1.154 kristaps 1541: return(ROFF_CONT);
1.94 kristaps 1542: }
1543:
1.67 kristaps 1544: enum rofferr
1.238 schwarze 1545: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1546: {
1547: enum rofft t;
1.109 kristaps 1548: enum rofferr e;
1.238 schwarze 1549: int pos; /* parse point */
1.243 schwarze 1550: int spos; /* saved parse point for messages */
1.238 schwarze 1551: int ppos; /* original offset in buf->buf */
1552: int ctl; /* macro line (boolean) */
1553:
1554: ppos = pos = *offs;
1.79 kristaps 1555:
1.230 schwarze 1556: /* Handle in-line equation delimiters. */
1557:
1.236 schwarze 1558: if (r->tbl == NULL &&
1559: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1560: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1561: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1562: if (e == ROFF_REPARSE)
1563: return(e);
1564: assert(e == ROFF_CONT);
1565: }
1566:
1567: /* Expand some escape sequences. */
1.94 kristaps 1568:
1.238 schwarze 1569: e = roff_res(r, buf, ln, pos);
1570: if (e == ROFF_IGN)
1.172 schwarze 1571: return(e);
1.238 schwarze 1572: assert(e == ROFF_CONT);
1.94 kristaps 1573:
1.238 schwarze 1574: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1575:
1.94 kristaps 1576: /*
1.79 kristaps 1577: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1578: * text through the macro's filter.
1579: * Equations process all content themselves.
1580: * Tables process almost all content themselves, but we want
1581: * to warn about macros before passing it there.
1.79 kristaps 1582: */
1.74 kristaps 1583:
1.252 schwarze 1584: if (r->last != NULL && ! ctl) {
1.78 kristaps 1585: t = r->last->tok;
1586: assert(roffs[t].text);
1.238 schwarze 1587: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1588: assert(e == ROFF_IGN || e == ROFF_CONT);
1589: if (e != ROFF_CONT)
1.125 kristaps 1590: return(e);
1.182 schwarze 1591: }
1.252 schwarze 1592: if (r->eqn != NULL)
1.238 schwarze 1593: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.252 schwarze 1594: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.259 schwarze 1595: return(tbl_read(r->tbl, ln, buf->buf, ppos));
1.252 schwarze 1596: if ( ! ctl)
1.238 schwarze 1597: return(roff_parsetext(buf, pos, offs));
1.228 schwarze 1598:
1599: /* Skip empty request lines. */
1600:
1.238 schwarze 1601: if (buf->buf[pos] == '"') {
1.228 schwarze 1602: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1603: ln, pos, NULL);
1604: return(ROFF_IGN);
1.238 schwarze 1605: } else if (buf->buf[pos] == '\0')
1.228 schwarze 1606: return(ROFF_IGN);
1.67 kristaps 1607:
1.79 kristaps 1608: /*
1609: * If a scope is open, go to the child handler for that macro,
1610: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1611: * Don't do so if an equation is open.
1.79 kristaps 1612: */
1.78 kristaps 1613:
1.79 kristaps 1614: if (r->last) {
1615: t = r->last->tok;
1616: assert(roffs[t].sub);
1.238 schwarze 1617: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1618: }
1.78 kristaps 1619:
1.243 schwarze 1620: /* No scope is open. This is a new request or macro. */
1621:
1622: spos = pos;
1623: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1624:
1625: /* Tables ignore most macros. */
1626:
1627: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1628: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1629: ln, pos, buf->buf + spos);
1.257 schwarze 1630: if (t == ROFF_TS)
1631: return(ROFF_IGN);
1632: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1633: pos++;
1634: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1635: pos++;
1636: return(tbl_read(r->tbl, ln, buf->buf, pos));
1.243 schwarze 1637: }
1638:
1.79 kristaps 1639: /*
1.243 schwarze 1640: * This is neither a roff request nor a user-defined macro.
1641: * Let the standard macro set parsers handle it.
1.79 kristaps 1642: */
1.67 kristaps 1643:
1.243 schwarze 1644: if (t == ROFF_MAX)
1.79 kristaps 1645: return(ROFF_CONT);
1.243 schwarze 1646:
1647: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1648:
1.75 kristaps 1649: assert(roffs[t].proc);
1.238 schwarze 1650: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.74 kristaps 1651: }
1652:
1.117 kristaps 1653: void
1.74 kristaps 1654: roff_endparse(struct roff *r)
1655: {
1656:
1.110 kristaps 1657: if (r->last)
1.221 schwarze 1658: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1659: r->last->line, r->last->col,
1660: roffs[r->last->tok].name);
1.117 kristaps 1661:
1.125 kristaps 1662: if (r->eqn) {
1.221 schwarze 1663: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1664: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151 kristaps 1665: eqn_end(&r->eqn);
1.125 kristaps 1666: }
1667:
1.117 kristaps 1668: if (r->tbl) {
1.221 schwarze 1669: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1670: r->tbl->line, r->tbl->pos, "TS");
1.151 kristaps 1671: tbl_end(&r->tbl);
1.117 kristaps 1672: }
1.67 kristaps 1673: }
1674:
1675: /*
1676: * Parse a roff node's type from the input buffer. This must be in the
1677: * form of ".foo xxx" in the usual way.
1678: */
1679: static enum rofft
1.214 schwarze 1680: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1681: {
1.214 schwarze 1682: char *cp;
1.106 kristaps 1683: const char *mac;
1684: size_t maclen;
1.67 kristaps 1685: enum rofft t;
1686:
1.214 schwarze 1687: cp = buf + *pos;
1688:
1689: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.67 kristaps 1690: return(ROFF_MAX);
1691:
1.214 schwarze 1692: mac = cp;
1693: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1694:
1.106 kristaps 1695: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.155 kristaps 1696: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.67 kristaps 1697:
1.214 schwarze 1698: if (ROFF_MAX != t)
1699: *pos = cp - buf;
1.67 kristaps 1700:
1701: return(t);
1702: }
1703:
1.266 schwarze 1704: /* --- handling of request blocks ----------------------------------------- */
1705:
1.67 kristaps 1706: static enum rofferr
1.76 kristaps 1707: roff_cblock(ROFF_ARGS)
1.67 kristaps 1708: {
1709:
1.79 kristaps 1710: /*
1711: * A block-close `..' should only be invoked as a child of an
1712: * ignore macro, otherwise raise a warning and just ignore it.
1713: */
1714:
1.238 schwarze 1715: if (r->last == NULL) {
1.221 schwarze 1716: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1717: ln, ppos, "..");
1.76 kristaps 1718: return(ROFF_IGN);
1719: }
1.67 kristaps 1720:
1.81 kristaps 1721: switch (r->last->tok) {
1.207 schwarze 1722: case ROFF_am:
1.220 schwarze 1723: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.81 kristaps 1724: /* FALLTHROUGH */
1.207 schwarze 1725: case ROFF_ami:
1.81 kristaps 1726: /* FALLTHROUGH */
1.207 schwarze 1727: case ROFF_de:
1.108 schwarze 1728: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.81 kristaps 1729: /* FALLTHROUGH */
1.207 schwarze 1730: case ROFF_dei:
1.81 kristaps 1731: /* FALLTHROUGH */
1.207 schwarze 1732: case ROFF_ig:
1.81 kristaps 1733: break;
1734: default:
1.221 schwarze 1735: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1736: ln, ppos, "..");
1.67 kristaps 1737: return(ROFF_IGN);
1.76 kristaps 1738: }
1.67 kristaps 1739:
1.238 schwarze 1740: if (buf->buf[pos] != '\0')
1.217 schwarze 1741: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1742: ".. %s", buf->buf + pos);
1.71 kristaps 1743:
1744: roffnode_pop(r);
1.76 kristaps 1745: roffnode_cleanscope(r);
1746: return(ROFF_IGN);
1.71 kristaps 1747:
1.67 kristaps 1748: }
1749:
1.76 kristaps 1750: static void
1751: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1752: {
1753:
1.76 kristaps 1754: while (r->last) {
1.173 schwarze 1755: if (--r->last->endspan != 0)
1.76 kristaps 1756: break;
1757: roffnode_pop(r);
1758: }
1.67 kristaps 1759: }
1760:
1.195 schwarze 1761: static void
1762: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1763: {
1764:
1.76 kristaps 1765: if (NULL == r->last) {
1.221 schwarze 1766: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1767: ln, ppos, "\\}");
1.195 schwarze 1768: return;
1.76 kristaps 1769: }
1770:
1.82 kristaps 1771: switch (r->last->tok) {
1.207 schwarze 1772: case ROFF_el:
1.82 kristaps 1773: /* FALLTHROUGH */
1.207 schwarze 1774: case ROFF_ie:
1.82 kristaps 1775: /* FALLTHROUGH */
1.207 schwarze 1776: case ROFF_if:
1.82 kristaps 1777: break;
1778: default:
1.221 schwarze 1779: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1780: ln, ppos, "\\}");
1.195 schwarze 1781: return;
1.75 kristaps 1782: }
1783:
1.76 kristaps 1784: if (r->last->endspan > -1) {
1.221 schwarze 1785: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1786: ln, ppos, "\\}");
1.195 schwarze 1787: return;
1.76 kristaps 1788: }
1789:
1.75 kristaps 1790: roffnode_pop(r);
1.76 kristaps 1791: roffnode_cleanscope(r);
1.195 schwarze 1792: return;
1.76 kristaps 1793: }
1794:
1795: static enum rofferr
1.80 kristaps 1796: roff_block(ROFF_ARGS)
1.76 kristaps 1797: {
1.220 schwarze 1798: const char *name;
1799: char *iname, *cp;
1.213 schwarze 1800: size_t namesz;
1.106 kristaps 1801:
1.220 schwarze 1802: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1803:
1.238 schwarze 1804: if (tok == ROFF_de1)
1.220 schwarze 1805: tok = ROFF_de;
1.251 schwarze 1806: else if (tok == ROFF_dei1)
1807: tok = ROFF_dei;
1.238 schwarze 1808: else if (tok == ROFF_am1)
1.220 schwarze 1809: tok = ROFF_am;
1.251 schwarze 1810: else if (tok == ROFF_ami1)
1811: tok = ROFF_ami;
1.220 schwarze 1812:
1813: /* Parse the macro name argument. */
1814:
1.238 schwarze 1815: cp = buf->buf + pos;
1816: if (tok == ROFF_ig) {
1.220 schwarze 1817: iname = NULL;
1818: namesz = 0;
1819: } else {
1820: iname = cp;
1821: namesz = roff_getname(r, &cp, ln, ppos);
1822: iname[namesz] = '\0';
1823: }
1.107 kristaps 1824:
1.220 schwarze 1825: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1826:
1.238 schwarze 1827: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1828: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1829: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1830: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1831: "%.*s", (int)namesz, iname);
1832: namesz = 0;
1833: } else
1834: namesz = strlen(name);
1835: } else
1836: name = iname;
1.107 kristaps 1837:
1.238 schwarze 1838: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1839: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1840: ln, ppos, roffs[tok].name);
1841: return(ROFF_IGN);
1842: }
1.80 kristaps 1843:
1.106 kristaps 1844: roffnode_push(r, tok, name, ln, ppos);
1845:
1846: /*
1847: * At the beginning of a `de' macro, clear the existing string
1848: * with the same name, if there is one. New content will be
1.193 schwarze 1849: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1850: */
1.107 kristaps 1851:
1.238 schwarze 1852: if (tok == ROFF_de || tok == ROFF_dei)
1.213 schwarze 1853: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76 kristaps 1854:
1.238 schwarze 1855: if (*cp == '\0')
1.78 kristaps 1856: return(ROFF_IGN);
1857:
1.220 schwarze 1858: /* Get the custom end marker. */
1.107 kristaps 1859:
1.220 schwarze 1860: iname = cp;
1.213 schwarze 1861: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1862:
1863: /* Resolve the end marker if it is indirect. */
1864:
1.238 schwarze 1865: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1866: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1867: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1868: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1869: "%.*s", (int)namesz, iname);
1870: namesz = 0;
1871: } else
1872: namesz = strlen(name);
1873: } else
1874: name = iname;
1875:
1.213 schwarze 1876: if (namesz)
1877: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1878:
1.238 schwarze 1879: if (*cp != '\0')
1.217 schwarze 1880: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1881: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.74 kristaps 1882:
1.78 kristaps 1883: return(ROFF_IGN);
1884: }
1885:
1886: static enum rofferr
1.80 kristaps 1887: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1888: {
1889: enum rofft t;
1890: int i, j;
1891:
1892: /*
1893: * First check whether a custom macro exists at this level. If
1894: * it does, then check against it. This is some of groff's
1895: * stranger behaviours. If we encountered a custom end-scope
1896: * tag and that tag also happens to be a "real" macro, then we
1897: * need to try interpreting it again as a real macro. If it's
1898: * not, then return ignore. Else continue.
1899: */
1900:
1901: if (r->last->end) {
1.130 kristaps 1902: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1903: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1904: break;
1905:
1.238 schwarze 1906: if (r->last->end[j] == '\0' &&
1907: (buf->buf[i] == '\0' ||
1908: buf->buf[i] == ' ' ||
1909: buf->buf[i] == '\t')) {
1.79 kristaps 1910: roffnode_pop(r);
1911: roffnode_cleanscope(r);
1912:
1.238 schwarze 1913: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1914: i++;
1915:
1916: pos = i;
1.238 schwarze 1917: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1918: ROFF_MAX)
1.79 kristaps 1919: return(ROFF_RERUN);
1920: return(ROFF_IGN);
1921: }
1922: }
1923:
1924: /*
1925: * If we have no custom end-query or lookup failed, then try
1926: * pulling it out of the hashtable.
1927: */
1928:
1.238 schwarze 1929: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1930:
1.238 schwarze 1931: if (t != ROFF_cblock) {
1932: if (tok != ROFF_ig)
1933: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.79 kristaps 1934: return(ROFF_IGN);
1.106 kristaps 1935: }
1.79 kristaps 1936:
1937: assert(roffs[t].proc);
1.238 schwarze 1938: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1939: }
1940:
1941: static enum rofferr
1.80 kristaps 1942: roff_block_text(ROFF_ARGS)
1.78 kristaps 1943: {
1944:
1.238 schwarze 1945: if (tok != ROFF_ig)
1946: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1947:
1.78 kristaps 1948: return(ROFF_IGN);
1949: }
1950:
1951: static enum rofferr
1.82 kristaps 1952: roff_cond_sub(ROFF_ARGS)
1953: {
1954: enum rofft t;
1.139 kristaps 1955: char *ep;
1.198 schwarze 1956: int rr;
1.82 kristaps 1957:
1958: rr = r->last->rule;
1.139 kristaps 1959: roffnode_cleanscope(r);
1.238 schwarze 1960: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82 kristaps 1961:
1.139 kristaps 1962: /*
1.177 schwarze 1963: * Fully handle known macros when they are structurally
1964: * required or when the conditional evaluated to true.
1.87 kristaps 1965: */
1966:
1.238 schwarze 1967: if ((t != ROFF_MAX) &&
1968: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.177 schwarze 1969: assert(roffs[t].proc);
1.238 schwarze 1970: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.177 schwarze 1971: }
1.144 kristaps 1972:
1.196 schwarze 1973: /*
1974: * If `\}' occurs on a macro line without a preceding macro,
1975: * drop the line completely.
1976: */
1977:
1.238 schwarze 1978: ep = buf->buf + pos;
1979: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1980: rr = 0;
1.196 schwarze 1981:
1.177 schwarze 1982: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1983:
1.238 schwarze 1984: while ((ep = strchr(ep, '\\')) != NULL) {
1985: if (*(++ep) == '}') {
1.197 schwarze 1986: *ep = '&';
1.238 schwarze 1987: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1988: }
1.247 schwarze 1989: if (*ep != '\0')
1990: ++ep;
1.177 schwarze 1991: }
1.198 schwarze 1992: return(rr ? ROFF_CONT : ROFF_IGN);
1.82 kristaps 1993: }
1994:
1995: static enum rofferr
1996: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1997: {
1.140 kristaps 1998: char *ep;
1.198 schwarze 1999: int rr;
1.82 kristaps 2000:
2001: rr = r->last->rule;
1.140 kristaps 2002: roffnode_cleanscope(r);
1.82 kristaps 2003:
1.238 schwarze 2004: ep = buf->buf + pos;
2005: while ((ep = strchr(ep, '\\')) != NULL) {
2006: if (*(++ep) == '}') {
1.197 schwarze 2007: *ep = '&';
1.238 schwarze 2008: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 2009: }
1.247 schwarze 2010: if (*ep != '\0')
2011: ++ep;
1.78 kristaps 2012: }
1.198 schwarze 2013: return(rr ? ROFF_CONT : ROFF_IGN);
1.74 kristaps 2014: }
2015:
1.266 schwarze 2016: /* --- handling of numeric and conditional expressions -------------------- */
2017:
1.204 schwarze 2018: /*
2019: * Parse a single signed integer number. Stop at the first non-digit.
2020: * If there is at least one digit, return success and advance the
2021: * parse point, else return failure and let the parse point unchanged.
2022: * Ignore overflows, treat them just like the C language.
2023: */
1.184 schwarze 2024: static int
1.261 schwarze 2025: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 2026: {
1.261 schwarze 2027: int myres, scaled, n, p;
1.206 schwarze 2028:
2029: if (NULL == res)
2030: res = &myres;
1.184 schwarze 2031:
2032: p = *pos;
2033: n = v[p] == '-';
1.261 schwarze 2034: if (n || v[p] == '+')
1.184 schwarze 2035: p++;
2036:
1.261 schwarze 2037: if (flags & ROFFNUM_WHITE)
2038: while (isspace((unsigned char)v[p]))
2039: p++;
2040:
1.184 schwarze 2041: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2042: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2043: if (p == *pos + n)
2044: return 0;
2045:
2046: if (n)
2047: *res = -*res;
2048:
1.254 schwarze 2049: /* Each number may be followed by one optional scaling unit. */
2050:
2051: switch (v[p]) {
2052: case 'f':
1.261 schwarze 2053: scaled = *res * 65536;
1.254 schwarze 2054: break;
2055: case 'i':
1.261 schwarze 2056: scaled = *res * 240;
1.254 schwarze 2057: break;
2058: case 'c':
1.261 schwarze 2059: scaled = *res * 240 / 2.54;
1.254 schwarze 2060: break;
2061: case 'v':
2062: /* FALLTROUGH */
2063: case 'P':
1.261 schwarze 2064: scaled = *res * 40;
1.254 schwarze 2065: break;
2066: case 'm':
2067: /* FALLTROUGH */
2068: case 'n':
1.261 schwarze 2069: scaled = *res * 24;
1.254 schwarze 2070: break;
2071: case 'p':
1.261 schwarze 2072: scaled = *res * 10 / 3;
1.254 schwarze 2073: break;
2074: case 'u':
1.261 schwarze 2075: scaled = *res;
1.254 schwarze 2076: break;
2077: case 'M':
1.261 schwarze 2078: scaled = *res * 6 / 25;
1.254 schwarze 2079: break;
2080: default:
1.261 schwarze 2081: scaled = *res;
1.254 schwarze 2082: p--;
2083: break;
2084: }
1.261 schwarze 2085: if (flags & ROFFNUM_SCALE)
2086: *res = scaled;
1.254 schwarze 2087:
2088: *pos = p + 1;
2089: return(1);
1.184 schwarze 2090: }
2091:
1.198 schwarze 2092: /*
2093: * Evaluate a string comparison condition.
2094: * The first character is the delimiter.
2095: * Succeed if the string up to its second occurrence
2096: * matches the string up to its third occurence.
2097: * Advance the cursor after the third occurrence
2098: * or lacking that, to the end of the line.
2099: */
2100: static int
2101: roff_evalstrcond(const char *v, int *pos)
2102: {
2103: const char *s1, *s2, *s3;
2104: int match;
2105:
2106: match = 0;
2107: s1 = v + *pos; /* initial delimiter */
2108: s2 = s1 + 1; /* for scanning the first string */
2109: s3 = strchr(s2, *s1); /* for scanning the second string */
2110:
2111: if (NULL == s3) /* found no middle delimiter */
2112: goto out;
2113:
2114: while ('\0' != *++s3) {
2115: if (*s2 != *s3) { /* mismatch */
2116: s3 = strchr(s3, *s1);
2117: break;
2118: }
2119: if (*s3 == *s1) { /* found the final delimiter */
2120: match = 1;
2121: break;
2122: }
2123: s2++;
2124: }
2125:
2126: out:
2127: if (NULL == s3)
2128: s3 = strchr(s2, '\0');
1.242 schwarze 2129: else if (*s3 != '\0')
1.198 schwarze 2130: s3++;
2131: *pos = s3 - v;
2132: return(match);
2133: }
2134:
1.204 schwarze 2135: /*
2136: * Evaluate an optionally negated single character, numerical,
2137: * or string condition.
2138: */
1.198 schwarze 2139: static int
1.271 schwarze 2140: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2141: {
1.271 schwarze 2142: char *cp, *name;
2143: size_t sz;
1.241 schwarze 2144: int number, savepos, wanttrue;
1.88 kristaps 2145:
1.198 schwarze 2146: if ('!' == v[*pos]) {
2147: wanttrue = 0;
2148: (*pos)++;
2149: } else
2150: wanttrue = 1;
2151:
1.88 kristaps 2152: switch (v[*pos]) {
1.240 schwarze 2153: case '\0':
2154: return(0);
1.207 schwarze 2155: case 'n':
1.198 schwarze 2156: /* FALLTHROUGH */
1.207 schwarze 2157: case 'o':
1.88 kristaps 2158: (*pos)++;
1.198 schwarze 2159: return(wanttrue);
1.207 schwarze 2160: case 'c':
1.198 schwarze 2161: /* FALLTHROUGH */
1.207 schwarze 2162: case 'd':
1.198 schwarze 2163: /* FALLTHROUGH */
1.207 schwarze 2164: case 'e':
1.88 kristaps 2165: /* FALLTHROUGH */
1.207 schwarze 2166: case 't':
1.239 schwarze 2167: /* FALLTHROUGH */
2168: case 'v':
1.88 kristaps 2169: (*pos)++;
1.198 schwarze 2170: return(!wanttrue);
1.271 schwarze 2171: case 'r':
2172: cp = name = v + ++*pos;
2173: sz = roff_getname(r, &cp, ln, *pos);
2174: *pos = cp - v;
2175: return((sz && roff_hasregn(r, name, sz)) == wanttrue);
1.88 kristaps 2176: default:
2177: break;
2178: }
2179:
1.241 schwarze 2180: savepos = *pos;
1.261 schwarze 2181: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.204 schwarze 2182: return((number > 0) == wanttrue);
1.241 schwarze 2183: else if (*pos == savepos)
2184: return(roff_evalstrcond(v, pos) == wanttrue);
1.204 schwarze 2185: else
1.241 schwarze 2186: return (0);
1.88 kristaps 2187: }
2188:
1.74 kristaps 2189: static enum rofferr
1.103 kristaps 2190: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2191: {
1.123 schwarze 2192:
1.89 kristaps 2193: return(ROFF_IGN);
2194: }
2195:
1.104 kristaps 2196: static enum rofferr
1.251 schwarze 2197: roff_insec(ROFF_ARGS)
2198: {
2199:
2200: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
2201: ln, ppos, roffs[tok].name);
2202: return(ROFF_IGN);
2203: }
2204:
2205: static enum rofferr
2206: roff_unsupp(ROFF_ARGS)
2207: {
2208:
2209: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
2210: ln, ppos, roffs[tok].name);
2211: return(ROFF_IGN);
2212: }
2213:
2214: static enum rofferr
1.82 kristaps 2215: roff_cond(ROFF_ARGS)
1.74 kristaps 2216: {
1.173 schwarze 2217:
2218: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2219:
1.207 schwarze 2220: /*
1.134 kristaps 2221: * An `.el' has no conditional body: it will consume the value
2222: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2223: * defaults to DENY.
1.134 kristaps 2224: *
2225: * If we're not an `el', however, then evaluate the conditional.
2226: */
1.133 kristaps 2227:
1.238 schwarze 2228: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2229: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2230: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2231:
1.134 kristaps 2232: /*
2233: * An if-else will put the NEGATION of the current evaluated
2234: * conditional into the stack of rules.
2235: */
2236:
1.238 schwarze 2237: if (tok == ROFF_ie) {
1.223 schwarze 2238: if (r->rstackpos + 1 == r->rstacksz) {
2239: r->rstacksz += 16;
2240: r->rstack = mandoc_reallocarray(r->rstack,
2241: r->rstacksz, sizeof(int));
1.134 kristaps 2242: }
1.198 schwarze 2243: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2244: }
1.88 kristaps 2245:
2246: /* If the parent has false as its rule, then so do we. */
2247:
1.198 schwarze 2248: if (r->last->parent && !r->last->parent->rule)
2249: r->last->rule = 0;
1.88 kristaps 2250:
2251: /*
1.173 schwarze 2252: * Determine scope.
2253: * If there is nothing on the line after the conditional,
2254: * not even whitespace, use next-line scope.
1.88 kristaps 2255: */
1.74 kristaps 2256:
1.238 schwarze 2257: if (buf->buf[pos] == '\0') {
1.173 schwarze 2258: r->last->endspan = 2;
2259: goto out;
2260: }
2261:
1.238 schwarze 2262: while (buf->buf[pos] == ' ')
1.173 schwarze 2263: pos++;
2264:
2265: /* An opening brace requests multiline scope. */
1.75 kristaps 2266:
1.238 schwarze 2267: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2268: r->last->endspan = -1;
2269: pos += 2;
1.272 schwarze 2270: while (buf->buf[pos] == ' ')
2271: pos++;
1.173 schwarze 2272: goto out;
1.207 schwarze 2273: }
1.74 kristaps 2274:
1.77 kristaps 2275: /*
1.173 schwarze 2276: * Anything else following the conditional causes
2277: * single-line scope. Warn if the scope contains
2278: * nothing but trailing whitespace.
1.77 kristaps 2279: */
2280:
1.238 schwarze 2281: if (buf->buf[pos] == '\0')
1.216 schwarze 2282: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
2283: ln, ppos, roffs[tok].name);
1.77 kristaps 2284:
1.173 schwarze 2285: r->last->endspan = 1;
1.74 kristaps 2286:
1.173 schwarze 2287: out:
1.75 kristaps 2288: *offs = pos;
2289: return(ROFF_RERUN);
1.83 schwarze 2290: }
2291:
2292: static enum rofferr
1.92 schwarze 2293: roff_ds(ROFF_ARGS)
2294: {
1.212 schwarze 2295: char *string;
2296: const char *name;
2297: size_t namesz;
1.96 kristaps 2298:
1.251 schwarze 2299: /* Ignore groff compatibility mode for now. */
2300:
2301: if (tok == ROFF_ds1)
2302: tok = ROFF_ds;
2303: else if (tok == ROFF_as1)
2304: tok = ROFF_as;
2305:
1.96 kristaps 2306: /*
1.212 schwarze 2307: * The first word is the name of the string.
2308: * If it is empty or terminated by an escape sequence,
2309: * abort the `ds' request without defining anything.
1.96 kristaps 2310: */
1.92 schwarze 2311:
1.238 schwarze 2312: name = string = buf->buf + pos;
2313: if (*name == '\0')
1.92 schwarze 2314: return(ROFF_IGN);
2315:
1.212 schwarze 2316: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2317: if (name[namesz] == '\\')
1.212 schwarze 2318: return(ROFF_IGN);
2319:
2320: /* Read past the initial double-quote, if any. */
1.238 schwarze 2321: if (*string == '"')
1.92 schwarze 2322: string++;
2323:
1.96 kristaps 2324: /* The rest is the value. */
1.212 schwarze 2325: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2326: ROFF_as == tok);
1.92 schwarze 2327: return(ROFF_IGN);
2328: }
2329:
1.204 schwarze 2330: /*
2331: * Parse a single operator, one or two characters long.
2332: * If the operator is recognized, return success and advance the
2333: * parse point, else return failure and let the parse point unchanged.
2334: */
2335: static int
2336: roff_getop(const char *v, int *pos, char *res)
2337: {
2338:
2339: *res = v[*pos];
2340:
2341: switch (*res) {
1.207 schwarze 2342: case '+':
1.204 schwarze 2343: /* FALLTHROUGH */
1.207 schwarze 2344: case '-':
1.204 schwarze 2345: /* FALLTHROUGH */
1.207 schwarze 2346: case '*':
1.204 schwarze 2347: /* FALLTHROUGH */
1.207 schwarze 2348: case '/':
1.204 schwarze 2349: /* FALLTHROUGH */
1.207 schwarze 2350: case '%':
1.204 schwarze 2351: /* FALLTHROUGH */
1.207 schwarze 2352: case '&':
1.204 schwarze 2353: /* FALLTHROUGH */
1.207 schwarze 2354: case ':':
1.204 schwarze 2355: break;
2356: case '<':
2357: switch (v[*pos + 1]) {
1.207 schwarze 2358: case '=':
1.204 schwarze 2359: *res = 'l';
2360: (*pos)++;
2361: break;
1.207 schwarze 2362: case '>':
1.204 schwarze 2363: *res = '!';
2364: (*pos)++;
2365: break;
1.207 schwarze 2366: case '?':
1.204 schwarze 2367: *res = 'i';
2368: (*pos)++;
2369: break;
2370: default:
2371: break;
2372: }
2373: break;
2374: case '>':
2375: switch (v[*pos + 1]) {
1.207 schwarze 2376: case '=':
1.204 schwarze 2377: *res = 'g';
2378: (*pos)++;
2379: break;
1.207 schwarze 2380: case '?':
1.204 schwarze 2381: *res = 'a';
2382: (*pos)++;
2383: break;
2384: default:
2385: break;
2386: }
2387: break;
2388: case '=':
2389: if ('=' == v[*pos + 1])
2390: (*pos)++;
2391: break;
2392: default:
2393: return(0);
2394: }
2395: (*pos)++;
2396:
2397: return(*res);
2398: }
2399:
2400: /*
2401: * Evaluate either a parenthesized numeric expression
2402: * or a single signed integer number.
2403: */
2404: static int
1.235 schwarze 2405: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2406: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2407: {
2408:
2409: if ('(' != v[*pos])
1.261 schwarze 2410: return(roff_getnum(v, pos, res, flags));
1.204 schwarze 2411:
2412: (*pos)++;
1.261 schwarze 2413: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.204 schwarze 2414: return(0);
2415:
1.206 schwarze 2416: /*
2417: * Omission of the closing parenthesis
2418: * is an error in validation mode,
2419: * but ignored in evaluation mode.
2420: */
2421:
1.204 schwarze 2422: if (')' == v[*pos])
2423: (*pos)++;
1.206 schwarze 2424: else if (NULL == res)
2425: return(0);
1.204 schwarze 2426:
2427: return(1);
2428: }
2429:
2430: /*
2431: * Evaluate a complete numeric expression.
2432: * Proceed left to right, there is no concept of precedence.
2433: */
2434: static int
1.235 schwarze 2435: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2436: int *pos, int *res, int flags)
1.204 schwarze 2437: {
2438: int mypos, operand2;
2439: char operator;
2440:
2441: if (NULL == pos) {
2442: mypos = 0;
2443: pos = &mypos;
2444: }
2445:
1.261 schwarze 2446: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2447: while (isspace((unsigned char)v[*pos]))
2448: (*pos)++;
2449:
1.261 schwarze 2450: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.204 schwarze 2451: return(0);
2452:
2453: while (1) {
1.261 schwarze 2454: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2455: while (isspace((unsigned char)v[*pos]))
2456: (*pos)++;
2457:
2458: if ( ! roff_getop(v, pos, &operator))
2459: break;
2460:
1.261 schwarze 2461: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2462: while (isspace((unsigned char)v[*pos]))
2463: (*pos)++;
2464:
1.261 schwarze 2465: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.204 schwarze 2466: return(0);
2467:
1.261 schwarze 2468: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2469: while (isspace((unsigned char)v[*pos]))
2470: (*pos)++;
1.206 schwarze 2471:
2472: if (NULL == res)
2473: continue;
1.204 schwarze 2474:
2475: switch (operator) {
1.207 schwarze 2476: case '+':
1.204 schwarze 2477: *res += operand2;
2478: break;
1.207 schwarze 2479: case '-':
1.204 schwarze 2480: *res -= operand2;
2481: break;
1.207 schwarze 2482: case '*':
1.204 schwarze 2483: *res *= operand2;
2484: break;
1.207 schwarze 2485: case '/':
1.244 schwarze 2486: if (operand2 == 0) {
1.235 schwarze 2487: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2488: r->parse, ln, *pos, v);
2489: *res = 0;
2490: break;
2491: }
1.204 schwarze 2492: *res /= operand2;
2493: break;
1.207 schwarze 2494: case '%':
1.244 schwarze 2495: if (operand2 == 0) {
2496: mandoc_msg(MANDOCERR_DIVZERO,
2497: r->parse, ln, *pos, v);
2498: *res = 0;
2499: break;
2500: }
1.204 schwarze 2501: *res %= operand2;
2502: break;
1.207 schwarze 2503: case '<':
1.204 schwarze 2504: *res = *res < operand2;
2505: break;
1.207 schwarze 2506: case '>':
1.204 schwarze 2507: *res = *res > operand2;
2508: break;
1.207 schwarze 2509: case 'l':
1.204 schwarze 2510: *res = *res <= operand2;
2511: break;
1.207 schwarze 2512: case 'g':
1.204 schwarze 2513: *res = *res >= operand2;
2514: break;
1.207 schwarze 2515: case '=':
1.204 schwarze 2516: *res = *res == operand2;
2517: break;
1.207 schwarze 2518: case '!':
1.204 schwarze 2519: *res = *res != operand2;
2520: break;
1.207 schwarze 2521: case '&':
1.204 schwarze 2522: *res = *res && operand2;
2523: break;
1.207 schwarze 2524: case ':':
1.204 schwarze 2525: *res = *res || operand2;
2526: break;
1.207 schwarze 2527: case 'i':
1.204 schwarze 2528: if (operand2 < *res)
2529: *res = operand2;
2530: break;
1.207 schwarze 2531: case 'a':
1.204 schwarze 2532: if (operand2 > *res)
2533: *res = operand2;
2534: break;
2535: default:
2536: abort();
2537: }
2538: }
2539: return(1);
2540: }
2541:
1.266 schwarze 2542: /* --- register management ------------------------------------------------ */
2543:
1.180 schwarze 2544: void
1.187 schwarze 2545: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2546: {
1.180 schwarze 2547: struct roffreg *reg;
2548:
2549: /* Search for an existing register with the same name. */
2550: reg = r->regtab;
2551:
2552: while (reg && strcmp(name, reg->key.p))
2553: reg = reg->next;
1.147 kristaps 2554:
1.180 schwarze 2555: if (NULL == reg) {
2556: /* Create a new register. */
2557: reg = mandoc_malloc(sizeof(struct roffreg));
2558: reg->key.p = mandoc_strdup(name);
2559: reg->key.sz = strlen(name);
1.187 schwarze 2560: reg->val = 0;
1.180 schwarze 2561: reg->next = r->regtab;
2562: r->regtab = reg;
2563: }
2564:
1.187 schwarze 2565: if ('+' == sign)
2566: reg->val += val;
2567: else if ('-' == sign)
2568: reg->val -= val;
2569: else
2570: reg->val = val;
1.147 kristaps 2571: }
2572:
1.192 schwarze 2573: /*
2574: * Handle some predefined read-only number registers.
2575: * For now, return -1 if the requested register is not predefined;
2576: * in case a predefined read-only register having the value -1
2577: * were to turn up, another special value would have to be chosen.
2578: */
2579: static int
1.273 schwarze 2580: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2581: {
2582:
2583: switch (*name) {
1.273 schwarze 2584: case '$': /* Number of arguments of the last macro evaluated. */
2585: return(r->argc);
1.207 schwarze 2586: case 'A': /* ASCII approximation mode is always off. */
1.192 schwarze 2587: return(0);
1.207 schwarze 2588: case 'g': /* Groff compatibility mode is always on. */
1.192 schwarze 2589: return(1);
1.207 schwarze 2590: case 'H': /* Fixed horizontal resolution. */
1.192 schwarze 2591: return (24);
1.207 schwarze 2592: case 'j': /* Always adjust left margin only. */
1.192 schwarze 2593: return(0);
1.207 schwarze 2594: case 'T': /* Some output device is always defined. */
1.192 schwarze 2595: return(1);
1.207 schwarze 2596: case 'V': /* Fixed vertical resolution. */
1.192 schwarze 2597: return (40);
2598: default:
2599: return (-1);
2600: }
2601: }
2602:
1.181 schwarze 2603: int
1.180 schwarze 2604: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2605: {
1.180 schwarze 2606: struct roffreg *reg;
1.192 schwarze 2607: int val;
2608:
2609: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.273 schwarze 2610: val = roff_getregro(r, name + 1);
1.192 schwarze 2611: if (-1 != val)
2612: return (val);
2613: }
1.180 schwarze 2614:
2615: for (reg = r->regtab; reg; reg = reg->next)
2616: if (0 == strcmp(name, reg->key.p))
1.181 schwarze 2617: return(reg->val);
2618:
2619: return(0);
2620: }
2621:
2622: static int
2623: roff_getregn(const struct roff *r, const char *name, size_t len)
2624: {
2625: struct roffreg *reg;
1.192 schwarze 2626: int val;
2627:
2628: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2629: val = roff_getregro(r, name + 1);
1.192 schwarze 2630: if (-1 != val)
2631: return (val);
2632: }
1.181 schwarze 2633:
2634: for (reg = r->regtab; reg; reg = reg->next)
2635: if (len == reg->key.sz &&
2636: 0 == strncmp(name, reg->key.p, len))
2637: return(reg->val);
1.271 schwarze 2638:
2639: return(0);
2640: }
2641:
2642: static int
2643: roff_hasregn(const struct roff *r, const char *name, size_t len)
2644: {
2645: struct roffreg *reg;
2646: int val;
2647:
2648: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2649: val = roff_getregro(r, name + 1);
1.271 schwarze 2650: if (-1 != val)
2651: return(1);
2652: }
2653:
2654: for (reg = r->regtab; reg; reg = reg->next)
2655: if (len == reg->key.sz &&
2656: 0 == strncmp(name, reg->key.p, len))
2657: return(1);
1.147 kristaps 2658:
1.180 schwarze 2659: return(0);
1.147 kristaps 2660: }
2661:
1.180 schwarze 2662: static void
2663: roff_freereg(struct roffreg *reg)
1.147 kristaps 2664: {
1.180 schwarze 2665: struct roffreg *old_reg;
1.147 kristaps 2666:
1.180 schwarze 2667: while (NULL != reg) {
2668: free(reg->key.p);
2669: old_reg = reg;
2670: reg = reg->next;
2671: free(old_reg);
2672: }
1.147 kristaps 2673: }
1.92 schwarze 2674:
2675: static enum rofferr
1.89 kristaps 2676: roff_nr(ROFF_ARGS)
1.83 schwarze 2677: {
1.212 schwarze 2678: char *key, *val;
2679: size_t keysz;
1.138 kristaps 2680: int iv;
1.187 schwarze 2681: char sign;
1.89 kristaps 2682:
1.238 schwarze 2683: key = val = buf->buf + pos;
2684: if (*key == '\0')
1.212 schwarze 2685: return(ROFF_IGN);
2686:
2687: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2688: if (key[keysz] == '\\')
1.212 schwarze 2689: return(ROFF_IGN);
2690: key[keysz] = '\0';
1.89 kristaps 2691:
1.187 schwarze 2692: sign = *val;
1.238 schwarze 2693: if (sign == '+' || sign == '-')
1.187 schwarze 2694: val++;
2695:
1.261 schwarze 2696: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204 schwarze 2697: roff_setreg(r, key, iv, sign);
1.109 kristaps 2698:
1.203 schwarze 2699: return(ROFF_IGN);
2700: }
2701:
2702: static enum rofferr
2703: roff_rr(ROFF_ARGS)
2704: {
2705: struct roffreg *reg, **prev;
1.212 schwarze 2706: char *name, *cp;
2707: size_t namesz;
1.203 schwarze 2708:
1.238 schwarze 2709: name = cp = buf->buf + pos;
2710: if (*name == '\0')
1.212 schwarze 2711: return(ROFF_IGN);
2712: namesz = roff_getname(r, &cp, ln, pos);
2713: name[namesz] = '\0';
1.203 schwarze 2714:
2715: prev = &r->regtab;
2716: while (1) {
2717: reg = *prev;
1.238 schwarze 2718: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2719: break;
2720: prev = ®->next;
2721: }
1.238 schwarze 2722: if (reg != NULL) {
1.203 schwarze 2723: *prev = reg->next;
2724: free(reg->key.p);
2725: free(reg);
2726: }
1.122 schwarze 2727: return(ROFF_IGN);
2728: }
2729:
1.266 schwarze 2730: /* --- handler functions for roff requests -------------------------------- */
2731:
1.122 schwarze 2732: static enum rofferr
2733: roff_rm(ROFF_ARGS)
2734: {
2735: const char *name;
2736: char *cp;
1.212 schwarze 2737: size_t namesz;
1.122 schwarze 2738:
1.238 schwarze 2739: cp = buf->buf + pos;
2740: while (*cp != '\0') {
1.212 schwarze 2741: name = cp;
1.238 schwarze 2742: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2743: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238 schwarze 2744: if (name[namesz] == '\\')
1.212 schwarze 2745: break;
1.122 schwarze 2746: }
1.178 schwarze 2747: return(ROFF_IGN);
2748: }
2749:
2750: static enum rofferr
2751: roff_it(ROFF_ARGS)
2752: {
2753: int iv;
2754:
2755: /* Parse the number of lines. */
1.261 schwarze 2756:
2757: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2758: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2759: ln, ppos, buf->buf + 1);
1.178 schwarze 2760: return(ROFF_IGN);
2761: }
2762:
1.262 schwarze 2763: while (isspace((unsigned char)buf->buf[pos]))
2764: pos++;
2765:
2766: /*
2767: * Arm the input line trap.
2768: * Special-casing "an-trap" is an ugly workaround to cope
2769: * with DocBook stupidly fiddling with man(7) internals.
2770: */
1.261 schwarze 2771:
1.178 schwarze 2772: roffit_lines = iv;
1.262 schwarze 2773: roffit_macro = mandoc_strdup(iv != 1 ||
2774: strcmp(buf->buf + pos, "an-trap") ?
2775: buf->buf + pos : "br");
1.109 kristaps 2776: return(ROFF_IGN);
1.175 schwarze 2777: }
2778:
2779: static enum rofferr
2780: roff_Dd(ROFF_ARGS)
2781: {
2782: const char *const *cp;
2783:
1.227 schwarze 2784: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175 schwarze 2785: for (cp = __mdoc_reserved; *cp; cp++)
2786: roff_setstr(r, *cp, NULL, 0);
2787:
1.227 schwarze 2788: if (r->format == 0)
2789: r->format = MPARSE_MDOC;
2790:
1.175 schwarze 2791: return(ROFF_CONT);
2792: }
2793:
2794: static enum rofferr
2795: roff_TH(ROFF_ARGS)
2796: {
2797: const char *const *cp;
2798:
1.227 schwarze 2799: if ((r->options & MPARSE_QUICK) == 0)
1.175 schwarze 2800: for (cp = __man_reserved; *cp; cp++)
2801: roff_setstr(r, *cp, NULL, 0);
2802:
1.227 schwarze 2803: if (r->format == 0)
2804: r->format = MPARSE_MAN;
2805:
1.175 schwarze 2806: return(ROFF_CONT);
1.109 kristaps 2807: }
2808:
2809: static enum rofferr
2810: roff_TE(ROFF_ARGS)
2811: {
2812:
2813: if (NULL == r->tbl)
1.221 schwarze 2814: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2815: ln, ppos, "TE");
1.258 schwarze 2816: else if ( ! tbl_end(&r->tbl)) {
2817: free(buf->buf);
2818: buf->buf = mandoc_strdup(".sp");
2819: buf->sz = 4;
2820: return(ROFF_REPARSE);
2821: }
1.112 kristaps 2822: return(ROFF_IGN);
2823: }
2824:
2825: static enum rofferr
2826: roff_T_(ROFF_ARGS)
2827: {
2828:
2829: if (NULL == r->tbl)
1.221 schwarze 2830: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2831: ln, ppos, "T&");
1.112 kristaps 2832: else
1.116 kristaps 2833: tbl_restart(ppos, ln, r->tbl);
1.112 kristaps 2834:
1.109 kristaps 2835: return(ROFF_IGN);
2836: }
2837:
1.230 schwarze 2838: /*
2839: * Handle in-line equation delimiters.
2840: */
2841: static enum rofferr
1.238 schwarze 2842: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2843: {
1.233 schwarze 2844: char *cp1, *cp2;
2845: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2846:
1.230 schwarze 2847: /*
2848: * Outside equations, look for an opening delimiter.
2849: * If we are inside an equation, we already know it is
2850: * in-line, or this function wouldn't have been called;
2851: * so look for a closing delimiter.
2852: */
2853:
1.238 schwarze 2854: cp1 = buf->buf + pos;
1.230 schwarze 2855: cp2 = strchr(cp1, r->eqn == NULL ?
2856: r->last_eqn->odelim : r->last_eqn->cdelim);
2857: if (cp2 == NULL)
2858: return(ROFF_CONT);
2859:
1.233 schwarze 2860: *cp2++ = '\0';
2861: bef_pr = bef_nl = aft_nl = aft_pr = "";
2862:
2863: /* Handle preceding text, protecting whitespace. */
2864:
1.238 schwarze 2865: if (*buf->buf != '\0') {
1.233 schwarze 2866: if (r->eqn == NULL)
2867: bef_pr = "\\&";
2868: bef_nl = "\n";
2869: }
2870:
2871: /*
2872: * Prepare replacing the delimiter with an equation macro
2873: * and drop leading white space from the equation.
2874: */
1.230 schwarze 2875:
1.233 schwarze 2876: if (r->eqn == NULL) {
2877: while (*cp2 == ' ')
2878: cp2++;
2879: mac = ".EQ";
2880: } else
2881: mac = ".EN";
2882:
2883: /* Handle following text, protecting whitespace. */
2884:
2885: if (*cp2 != '\0') {
2886: aft_nl = "\n";
2887: if (r->eqn != NULL)
2888: aft_pr = "\\&";
2889: }
2890:
2891: /* Do the actual replacement. */
2892:
1.238 schwarze 2893: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2894: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2895: free(buf->buf);
2896: buf->buf = cp1;
1.230 schwarze 2897:
2898: /* Toggle the in-line state of the eqn subsystem. */
2899:
2900: r->eqn_inline = r->eqn == NULL;
2901: return(ROFF_REPARSE);
1.151 kristaps 2902: }
2903:
1.235 schwarze 2904: static enum rofferr
2905: roff_EQ(ROFF_ARGS)
1.125 kristaps 2906: {
1.151 kristaps 2907: struct eqn_node *e;
1.125 kristaps 2908:
1.238 schwarze 2909: assert(r->eqn == NULL);
1.235 schwarze 2910: e = eqn_alloc(ppos, ln, r->parse);
1.125 kristaps 2911:
1.230 schwarze 2912: if (r->last_eqn) {
1.125 kristaps 2913: r->last_eqn->next = e;
1.230 schwarze 2914: e->delim = r->last_eqn->delim;
2915: e->odelim = r->last_eqn->odelim;
2916: e->cdelim = r->last_eqn->cdelim;
2917: } else
1.125 kristaps 2918: r->first_eqn = r->last_eqn = e;
2919:
2920: r->eqn = r->last_eqn = e;
1.151 kristaps 2921:
1.238 schwarze 2922: if (buf->buf[pos] != '\0')
1.235 schwarze 2923: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2924: ".EQ %s", buf->buf + pos);
1.151 kristaps 2925:
1.125 kristaps 2926: return(ROFF_IGN);
2927: }
2928:
2929: static enum rofferr
2930: roff_EN(ROFF_ARGS)
2931: {
2932:
1.221 schwarze 2933: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.125 kristaps 2934: return(ROFF_IGN);
2935: }
2936:
2937: static enum rofferr
1.109 kristaps 2938: roff_TS(ROFF_ARGS)
2939: {
1.176 schwarze 2940: struct tbl_node *tbl;
1.89 kristaps 2941:
1.115 kristaps 2942: if (r->tbl) {
1.221 schwarze 2943: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2944: ln, ppos, "TS breaks TS");
1.151 kristaps 2945: tbl_end(&r->tbl);
1.115 kristaps 2946: }
1.83 schwarze 2947:
1.176 schwarze 2948: tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2949:
2950: if (r->last_tbl)
1.176 schwarze 2951: r->last_tbl->next = tbl;
1.113 kristaps 2952: else
1.176 schwarze 2953: r->first_tbl = r->last_tbl = tbl;
1.113 kristaps 2954:
1.176 schwarze 2955: r->tbl = r->last_tbl = tbl;
1.83 schwarze 2956: return(ROFF_IGN);
1.251 schwarze 2957: }
2958:
2959: static enum rofferr
2960: roff_brp(ROFF_ARGS)
2961: {
2962:
2963: buf->buf[pos - 1] = '\0';
2964: return(ROFF_CONT);
1.92 schwarze 2965: }
2966:
1.105 kristaps 2967: static enum rofferr
1.174 kristaps 2968: roff_cc(ROFF_ARGS)
2969: {
2970: const char *p;
2971:
1.238 schwarze 2972: p = buf->buf + pos;
1.174 kristaps 2973:
1.238 schwarze 2974: if (*p == '\0' || (r->control = *p++) == '.')
1.174 kristaps 2975: r->control = 0;
2976:
1.238 schwarze 2977: if (*p != '\0')
1.260 schwarze 2978: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2979: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 2980:
2981: return(ROFF_IGN);
2982: }
2983:
2984: static enum rofferr
1.164 kristaps 2985: roff_tr(ROFF_ARGS)
2986: {
2987: const char *p, *first, *second;
2988: size_t fsz, ssz;
2989: enum mandoc_esc esc;
2990:
1.238 schwarze 2991: p = buf->buf + pos;
1.164 kristaps 2992:
1.238 schwarze 2993: if (*p == '\0') {
1.260 schwarze 2994: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.164 kristaps 2995: return(ROFF_IGN);
2996: }
2997:
1.238 schwarze 2998: while (*p != '\0') {
1.164 kristaps 2999: fsz = ssz = 1;
3000:
3001: first = p++;
1.238 schwarze 3002: if (*first == '\\') {
1.164 kristaps 3003: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3004: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3005: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3006: ln, (int)(p - buf->buf), first);
1.164 kristaps 3007: return(ROFF_IGN);
3008: }
3009: fsz = (size_t)(p - first);
3010: }
3011:
3012: second = p++;
1.238 schwarze 3013: if (*second == '\\') {
1.164 kristaps 3014: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3015: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3016: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3017: ln, (int)(p - buf->buf), second);
1.164 kristaps 3018: return(ROFF_IGN);
3019: }
3020: ssz = (size_t)(p - second);
1.238 schwarze 3021: } else if (*second == '\0') {
1.260 schwarze 3022: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
3023: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 3024: second = " ";
1.165 kristaps 3025: p--;
1.164 kristaps 3026: }
3027:
1.167 kristaps 3028: if (fsz > 1) {
1.207 schwarze 3029: roff_setstrn(&r->xmbtab, first, fsz,
3030: second, ssz, 0);
1.167 kristaps 3031: continue;
3032: }
3033:
1.238 schwarze 3034: if (r->xtab == NULL)
1.207 schwarze 3035: r->xtab = mandoc_calloc(128,
3036: sizeof(struct roffstr));
1.167 kristaps 3037:
3038: free(r->xtab[(int)*first].p);
3039: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3040: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 3041: }
3042:
3043: return(ROFF_IGN);
3044: }
3045:
3046: static enum rofferr
1.105 kristaps 3047: roff_so(ROFF_ARGS)
3048: {
1.249 schwarze 3049: char *name, *cp;
1.105 kristaps 3050:
1.238 schwarze 3051: name = buf->buf + pos;
1.224 schwarze 3052: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3053:
3054: /*
3055: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3056: * opening anything that's not in our cwd or anything beneath
3057: * it. Thus, explicitly disallow traversing up the file-system
3058: * or using absolute paths.
3059: */
3060:
1.238 schwarze 3061: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3062: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3063: ".so %s", name);
1.249 schwarze 3064: buf->sz = mandoc_asprintf(&cp,
3065: ".sp\nSee the file %s.\n.sp", name) + 1;
3066: free(buf->buf);
3067: buf->buf = cp;
3068: *offs = 0;
3069: return(ROFF_REPARSE);
1.105 kristaps 3070: }
3071:
3072: *offs = pos;
3073: return(ROFF_SO);
3074: }
1.92 schwarze 3075:
1.266 schwarze 3076: /* --- user defined strings and macros ------------------------------------ */
3077:
1.106 kristaps 3078: static enum rofferr
3079: roff_userdef(ROFF_ARGS)
1.99 kristaps 3080: {
1.263 schwarze 3081: const char *arg[9], *ap;
1.106 kristaps 3082: char *cp, *n1, *n2;
1.274 ! schwarze 3083: int i, ib, ie;
1.263 schwarze 3084: size_t asz, rsz;
1.106 kristaps 3085:
3086: /*
3087: * Collect pointers to macro argument strings
1.188 schwarze 3088: * and NUL-terminate them.
1.106 kristaps 3089: */
1.263 schwarze 3090:
1.273 schwarze 3091: r->argc = 0;
1.238 schwarze 3092: cp = buf->buf + pos;
1.273 schwarze 3093: for (i = 0; i < 9; i++) {
3094: if (*cp == '\0')
3095: arg[i] = "";
3096: else {
3097: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3098: r->argc = i + 1;
3099: }
3100: }
1.99 kristaps 3101:
1.106 kristaps 3102: /*
3103: * Expand macro arguments.
1.99 kristaps 3104: */
1.263 schwarze 3105:
3106: buf->sz = strlen(r->current_string) + 1;
3107: n1 = cp = mandoc_malloc(buf->sz);
3108: memcpy(n1, r->current_string, buf->sz);
3109: while (*cp != '\0') {
3110:
3111: /* Scan ahead for the next argument invocation. */
3112:
3113: if (*cp++ != '\\')
3114: continue;
3115: if (*cp++ != '$')
3116: continue;
1.274 ! schwarze 3117: if (*cp == '*') { /* \\$* inserts all arguments */
! 3118: ib = 0;
! 3119: ie = r->argc - 1;
! 3120: } else { /* \\$1 .. \\$9 insert one argument */
! 3121: ib = ie = *cp - '1';
! 3122: if (ib < 0 || ib > 8)
! 3123: continue;
! 3124: }
1.263 schwarze 3125: cp -= 2;
3126:
3127: /*
3128: * Determine the size of the expanded argument,
3129: * taking escaping of quotes into account.
3130: */
3131:
1.274 ! schwarze 3132: asz = ie > ib ? ie - ib : 0; /* for blanks */
! 3133: for (i = ib; i <= ie; i++) {
! 3134: for (ap = arg[i]; *ap != '\0'; ap++) {
! 3135: asz++;
! 3136: if (*ap == '"')
! 3137: asz += 3;
! 3138: }
1.263 schwarze 3139: }
3140: if (asz != 3) {
3141:
3142: /*
3143: * Determine the size of the rest of the
3144: * unexpanded macro, including the NUL.
3145: */
3146:
3147: rsz = buf->sz - (cp - n1) - 3;
3148:
3149: /*
3150: * When shrinking, move before
3151: * releasing the storage.
3152: */
3153:
3154: if (asz < 3)
3155: memmove(cp + asz, cp + 3, rsz);
3156:
3157: /*
3158: * Resize the storage for the macro
3159: * and readjust the parse pointer.
3160: */
3161:
3162: buf->sz += asz - 3;
3163: n2 = mandoc_realloc(n1, buf->sz);
3164: cp = n2 + (cp - n1);
3165: n1 = n2;
3166:
3167: /*
3168: * When growing, make room
3169: * for the expanded argument.
3170: */
3171:
3172: if (asz > 3)
3173: memmove(cp + asz, cp + 3, rsz);
3174: }
3175:
3176: /* Copy the expanded argument, escaping quotes. */
3177:
3178: n2 = cp;
1.274 ! schwarze 3179: for (i = ib; i <= ie; i++) {
! 3180: for (ap = arg[i]; *ap != '\0'; ap++) {
! 3181: if (*ap == '"') {
! 3182: memcpy(n2, "\\(dq", 4);
! 3183: n2 += 4;
! 3184: } else
! 3185: *n2++ = *ap;
! 3186: }
! 3187: if (i < ie)
! 3188: *n2++ = ' ';
1.106 kristaps 3189: }
1.99 kristaps 3190: }
3191:
1.106 kristaps 3192: /*
3193: * Replace the macro invocation
3194: * by the expanded macro.
3195: */
1.263 schwarze 3196:
1.238 schwarze 3197: free(buf->buf);
3198: buf->buf = n1;
1.248 schwarze 3199: *offs = 0;
1.106 kristaps 3200:
1.238 schwarze 3201: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.106 kristaps 3202: ROFF_REPARSE : ROFF_APPEND);
1.99 kristaps 3203: }
1.121 schwarze 3204:
1.212 schwarze 3205: static size_t
1.121 schwarze 3206: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3207: {
3208: char *name, *cp;
1.212 schwarze 3209: size_t namesz;
1.121 schwarze 3210:
3211: name = *cpp;
3212: if ('\0' == *name)
1.212 schwarze 3213: return(0);
1.121 schwarze 3214:
1.212 schwarze 3215: /* Read until end of name and terminate it with NUL. */
3216: for (cp = name; 1; cp++) {
3217: if ('\0' == *cp || ' ' == *cp) {
3218: namesz = cp - name;
3219: break;
3220: }
1.121 schwarze 3221: if ('\\' != *cp)
3222: continue;
1.215 schwarze 3223: namesz = cp - name;
3224: if ('{' == cp[1] || '}' == cp[1])
3225: break;
1.121 schwarze 3226: cp++;
3227: if ('\\' == *cp)
3228: continue;
1.224 schwarze 3229: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3230: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3231: mandoc_escape((const char **)&cp, NULL, NULL);
3232: break;
1.121 schwarze 3233: }
3234:
3235: /* Read past spaces. */
3236: while (' ' == *cp)
3237: cp++;
3238:
3239: *cpp = cp;
1.212 schwarze 3240: return(namesz);
1.121 schwarze 3241: }
3242:
1.106 kristaps 3243: /*
3244: * Store *string into the user-defined string called *name.
3245: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3246: * append == 0: replace mode
3247: * append == 1: single-line append mode
3248: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3249: */
1.94 kristaps 3250: static void
1.106 kristaps 3251: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3252: int append)
1.92 schwarze 3253: {
1.164 kristaps 3254:
3255: roff_setstrn(&r->strtab, name, strlen(name), string,
1.207 schwarze 3256: string ? strlen(string) : 0, append);
1.164 kristaps 3257: }
3258:
3259: static void
1.166 kristaps 3260: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3261: const char *string, size_t stringsz, int append)
1.164 kristaps 3262: {
1.166 kristaps 3263: struct roffkv *n;
1.164 kristaps 3264: char *c;
3265: int i;
3266: size_t oldch, newch;
1.92 schwarze 3267:
1.106 kristaps 3268: /* Search for an existing string with the same name. */
1.164 kristaps 3269: n = *r;
3270:
1.211 schwarze 3271: while (n && (namesz != n->key.sz ||
3272: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3273: n = n->next;
1.94 kristaps 3274:
3275: if (NULL == n) {
1.106 kristaps 3276: /* Create a new string table entry. */
1.166 kristaps 3277: n = mandoc_malloc(sizeof(struct roffkv));
3278: n->key.p = mandoc_strndup(name, namesz);
3279: n->key.sz = namesz;
3280: n->val.p = NULL;
3281: n->val.sz = 0;
1.164 kristaps 3282: n->next = *r;
3283: *r = n;
1.193 schwarze 3284: } else if (0 == append) {
1.166 kristaps 3285: free(n->val.p);
3286: n->val.p = NULL;
3287: n->val.sz = 0;
1.106 kristaps 3288: }
3289:
3290: if (NULL == string)
3291: return;
3292:
3293: /*
3294: * One additional byte for the '\n' in multiline mode,
3295: * and one for the terminating '\0'.
3296: */
1.193 schwarze 3297: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3298:
1.166 kristaps 3299: if (NULL == n->val.p) {
3300: n->val.p = mandoc_malloc(newch);
3301: *n->val.p = '\0';
1.106 kristaps 3302: oldch = 0;
3303: } else {
1.166 kristaps 3304: oldch = n->val.sz;
3305: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3306: }
3307:
3308: /* Skip existing content in the destination buffer. */
1.166 kristaps 3309: c = n->val.p + (int)oldch;
1.106 kristaps 3310:
3311: /* Append new content to the destination buffer. */
1.164 kristaps 3312: i = 0;
3313: while (i < (int)stringsz) {
1.106 kristaps 3314: /*
3315: * Rudimentary roff copy mode:
3316: * Handle escaped backslashes.
3317: */
1.164 kristaps 3318: if ('\\' == string[i] && '\\' == string[i + 1])
3319: i++;
3320: *c++ = string[i++];
1.106 kristaps 3321: }
1.94 kristaps 3322:
1.106 kristaps 3323: /* Append terminating bytes. */
1.193 schwarze 3324: if (1 < append)
1.106 kristaps 3325: *c++ = '\n';
1.163 kristaps 3326:
1.106 kristaps 3327: *c = '\0';
1.166 kristaps 3328: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3329: }
3330:
1.94 kristaps 3331: static const char *
3332: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92 schwarze 3333: {
1.166 kristaps 3334: const struct roffkv *n;
1.191 schwarze 3335: int i;
1.92 schwarze 3336:
1.164 kristaps 3337: for (n = r->strtab; n; n = n->next)
1.207 schwarze 3338: if (0 == strncmp(name, n->key.p, len) &&
3339: '\0' == n->key.p[(int)len])
1.166 kristaps 3340: return(n->val.p);
1.191 schwarze 3341:
3342: for (i = 0; i < PREDEFS_MAX; i++)
3343: if (0 == strncmp(name, predefs[i].name, len) &&
3344: '\0' == predefs[i].name[(int)len])
3345: return(predefs[i].str);
1.94 kristaps 3346:
1.157 kristaps 3347: return(NULL);
1.92 schwarze 3348: }
3349:
1.94 kristaps 3350: static void
1.167 kristaps 3351: roff_freestr(struct roffkv *r)
1.92 schwarze 3352: {
1.166 kristaps 3353: struct roffkv *n, *nn;
1.92 schwarze 3354:
1.167 kristaps 3355: for (n = r; n; n = nn) {
1.166 kristaps 3356: free(n->key.p);
3357: free(n->val.p);
1.92 schwarze 3358: nn = n->next;
3359: free(n);
3360: }
1.114 kristaps 3361: }
1.266 schwarze 3362:
3363: /* --- accessors and utility functions ------------------------------------ */
1.114 kristaps 3364:
3365: const struct tbl_span *
3366: roff_span(const struct roff *r)
3367: {
1.207 schwarze 3368:
1.114 kristaps 3369: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.125 kristaps 3370: }
3371:
3372: const struct eqn *
3373: roff_eqn(const struct roff *r)
3374: {
1.207 schwarze 3375:
1.125 kristaps 3376: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.164 kristaps 3377: }
3378:
3379: /*
3380: * Duplicate an input string, making the appropriate character
3381: * conversations (as stipulated by `tr') along the way.
3382: * Returns a heap-allocated string with all the replacements made.
3383: */
3384: char *
3385: roff_strdup(const struct roff *r, const char *p)
3386: {
1.166 kristaps 3387: const struct roffkv *cp;
1.164 kristaps 3388: char *res;
3389: const char *pp;
3390: size_t ssz, sz;
3391: enum mandoc_esc esc;
3392:
1.167 kristaps 3393: if (NULL == r->xmbtab && NULL == r->xtab)
1.164 kristaps 3394: return(mandoc_strdup(p));
3395: else if ('\0' == *p)
3396: return(mandoc_strdup(""));
3397:
3398: /*
3399: * Step through each character looking for term matches
3400: * (remember that a `tr' can be invoked with an escape, which is
3401: * a glyph but the escape is multi-character).
3402: * We only do this if the character hash has been initialised
3403: * and the string is >0 length.
3404: */
3405:
3406: res = NULL;
3407: ssz = 0;
3408:
3409: while ('\0' != *p) {
1.167 kristaps 3410: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3411: sz = r->xtab[(int)*p].sz;
3412: res = mandoc_realloc(res, ssz + sz + 1);
3413: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3414: ssz += sz;
3415: p++;
3416: continue;
3417: } else if ('\\' != *p) {
3418: res = mandoc_realloc(res, ssz + 2);
3419: res[ssz++] = *p++;
3420: continue;
3421: }
3422:
1.164 kristaps 3423: /* Search for term matches. */
1.167 kristaps 3424: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3425: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3426: break;
3427:
3428: if (NULL != cp) {
3429: /*
3430: * A match has been found.
3431: * Append the match to the array and move
3432: * forward by its keysize.
3433: */
1.207 schwarze 3434: res = mandoc_realloc(res,
3435: ssz + cp->val.sz + 1);
1.166 kristaps 3436: memcpy(res + ssz, cp->val.p, cp->val.sz);
3437: ssz += cp->val.sz;
3438: p += (int)cp->key.sz;
1.164 kristaps 3439: continue;
3440: }
3441:
1.167 kristaps 3442: /*
3443: * Handle escapes carefully: we need to copy
3444: * over just the escape itself, or else we might
3445: * do replacements within the escape itself.
3446: * Make sure to pass along the bogus string.
3447: */
3448: pp = p++;
3449: esc = mandoc_escape(&p, NULL, NULL);
3450: if (ESCAPE_ERROR == esc) {
3451: sz = strlen(pp);
1.164 kristaps 3452: res = mandoc_realloc(res, ssz + sz + 1);
3453: memcpy(res + ssz, pp, sz);
1.167 kristaps 3454: break;
1.164 kristaps 3455: }
1.207 schwarze 3456: /*
3457: * We bail out on bad escapes.
1.167 kristaps 3458: * No need to warn: we already did so when
3459: * roff_res() was called.
3460: */
3461: sz = (int)(p - pp);
3462: res = mandoc_realloc(res, ssz + sz + 1);
3463: memcpy(res + ssz, pp, sz);
3464: ssz += sz;
1.164 kristaps 3465: }
3466:
3467: res[(int)ssz] = '\0';
3468: return(res);
1.227 schwarze 3469: }
3470:
3471: int
3472: roff_getformat(const struct roff *r)
3473: {
3474:
3475: return(r->format);
1.174 kristaps 3476: }
3477:
3478: /*
1.207 schwarze 3479: * Find out whether a line is a macro line or not.
1.174 kristaps 3480: * If it is, adjust the current position and return one; if it isn't,
3481: * return zero and don't change the current position.
3482: * If the control character has been set with `.cc', then let that grain
3483: * precedence.
3484: * This is slighly contrary to groff, where using the non-breaking
3485: * control character when `cc' has been invoked will cause the
3486: * non-breaking macro contents to be printed verbatim.
3487: */
3488: int
3489: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3490: {
3491: int pos;
3492:
3493: pos = *ppos;
3494:
3495: if (0 != r->control && cp[pos] == r->control)
3496: pos++;
3497: else if (0 != r->control)
3498: return(0);
3499: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3500: pos += 2;
3501: else if ('.' == cp[pos] || '\'' == cp[pos])
3502: pos++;
3503: else
3504: return(0);
3505:
3506: while (' ' == cp[pos] || '\t' == cp[pos])
3507: pos++;
3508:
3509: *ppos = pos;
3510: return(1);
1.74 kristaps 3511: }
CVSweb