Annotation of mandoc/roff.c, Revision 1.267
1.267 ! schwarze 1: /* $Id: roff.c,v 1.266 2015/04/19 13:50:26 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 */
1.223 schwarze 338: char control; /* control character */
1.79 kristaps 339: };
340:
1.67 kristaps 341: struct roffnode {
342: enum rofft tok; /* type of node */
343: struct roffnode *parent; /* up one in stack */
344: int line; /* parse line */
345: int col; /* parse col */
1.106 kristaps 346: char *name; /* node name, e.g. macro name */
1.79 kristaps 347: char *end; /* end-rules: custom token */
348: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 349: int rule; /* current evaluation rule */
1.67 kristaps 350: };
351:
352: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.72 kristaps 353: enum rofft tok, /* tok of macro */ \
1.238 schwarze 354: struct buf *buf, /* input buffer */ \
1.67 kristaps 355: int ln, /* parse line */ \
1.75 kristaps 356: int ppos, /* original pos in buffer */ \
357: int pos, /* current pos in buffer */ \
1.74 kristaps 358: int *offs /* reset offset of buffer data */
1.67 kristaps 359:
360: typedef enum rofferr (*roffproc)(ROFF_ARGS);
361:
362: struct roffmac {
363: const char *name; /* macro name */
1.79 kristaps 364: roffproc proc; /* process new macro */
365: roffproc text; /* process as child text of macro */
366: roffproc sub; /* process as child of macro */
367: int flags;
368: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.85 kristaps 369: struct roffmac *next;
1.67 kristaps 370: };
371:
1.141 kristaps 372: struct predef {
373: const char *name; /* predefined input name */
374: const char *str; /* replacement symbol */
375: };
376:
377: #define PREDEF(__name, __str) \
378: { (__name), (__str) },
379:
1.266 schwarze 380: /* --- function prototypes ------------------------------------------------ */
381:
1.155 kristaps 382: static enum rofft roffhash_find(const char *, size_t);
383: static void roffhash_init(void);
384: static void roffnode_cleanscope(struct roff *);
385: static void roffnode_pop(struct roff *);
386: static void roffnode_push(struct roff *, enum rofft,
387: const char *, int, int);
1.80 kristaps 388: static enum rofferr roff_block(ROFF_ARGS);
389: static enum rofferr roff_block_text(ROFF_ARGS);
390: static enum rofferr roff_block_sub(ROFF_ARGS);
1.251 schwarze 391: static enum rofferr roff_brp(ROFF_ARGS);
1.80 kristaps 392: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 393: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 394: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 395: static enum rofferr roff_cond(ROFF_ARGS);
396: static enum rofferr roff_cond_text(ROFF_ARGS);
397: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 398: static enum rofferr roff_ds(ROFF_ARGS);
1.238 schwarze 399: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.234 kristaps 400: static int roff_evalcond(struct roff *r, int,
401: const char *, int *);
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);
1.192 schwarze 415: static int roff_getregro(const char *name);
1.207 schwarze 416: static const char *roff_getstrn(const struct roff *,
1.94 kristaps 417: const char *, size_t);
1.251 schwarze 418: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 419: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 420: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 schwarze 421: static void roff_man_alloc1(struct roff_man *);
422: static void roff_man_free1(struct roff_man *);
1.89 kristaps 423: static enum rofferr roff_nr(ROFF_ARGS);
1.214 schwarze 424: static enum rofft roff_parse(struct roff *, char *, int *,
425: int, int);
1.238 schwarze 426: static enum rofferr roff_parsetext(struct buf *, int, int *);
427: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 428: static enum rofferr roff_rm(ROFF_ARGS);
1.203 schwarze 429: static enum rofferr roff_rr(ROFF_ARGS);
1.94 kristaps 430: static void roff_setstr(struct roff *,
1.106 kristaps 431: const char *, const char *, int);
1.207 schwarze 432: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 433: size_t, const char *, size_t, int);
1.105 kristaps 434: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 435: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 436: static enum rofferr roff_Dd(ROFF_ARGS);
437: static enum rofferr roff_TH(ROFF_ARGS);
1.109 kristaps 438: static enum rofferr roff_TE(ROFF_ARGS);
439: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 440: static enum rofferr roff_EQ(ROFF_ARGS);
441: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 442: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 443: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 444: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 445:
1.266 schwarze 446: /* --- constant data ------------------------------------------------------ */
447:
1.155 kristaps 448: /* See roffhash_find() */
1.85 kristaps 449:
450: #define ASCII_HI 126
451: #define ASCII_LO 33
452: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
453:
1.261 schwarze 454: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
455: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
456:
1.85 kristaps 457: static struct roffmac *hash[HASHWIDTH];
458:
459: static struct roffmac roffs[ROFF_MAX] = {
1.251 schwarze 460: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.103 kristaps 461: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 462: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
463: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
464: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 465: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 466: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.85 kristaps 467: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 468: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.193 schwarze 469: { "as", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 470: { "as1", roff_ds, NULL, NULL, 0, NULL },
471: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
472: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
473: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
474: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
475: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
476: { "box", roff_unsupp, NULL, NULL, 0, NULL },
477: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
478: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
479: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
480: { "break", roff_unsupp, NULL, NULL, 0, NULL },
481: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
482: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
483: { "brp", roff_brp, NULL, NULL, 0, NULL },
484: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
485: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.174 kristaps 486: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.194 schwarze 487: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 488: { "cf", roff_insec, NULL, NULL, 0, NULL },
489: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
490: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
491: { "char", roff_unsupp, NULL, NULL, 0, NULL },
492: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
493: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
494: { "close", roff_insec, NULL, NULL, 0, NULL },
495: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
496: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
497: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
498: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
499: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
500: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
501: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
502: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
503: { "da", roff_unsupp, NULL, NULL, 0, NULL },
504: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
505: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.85 kristaps 506: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 507: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
508: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 509: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 510: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
511: { "device", roff_unsupp, NULL, NULL, 0, NULL },
512: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
513: { "di", roff_unsupp, NULL, NULL, 0, NULL },
514: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.92 schwarze 515: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 516: { "ds1", roff_ds, NULL, NULL, 0, NULL },
517: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
518: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
519: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
520: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
521: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 522: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.251 schwarze 523: { "em", roff_unsupp, NULL, NULL, 0, NULL },
524: { "EN", roff_EN, NULL, NULL, 0, NULL },
525: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
526: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
527: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
528: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
529: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
530: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
531: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
532: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.185 schwarze 533: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 534: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
535: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
536: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
538: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
543: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
544: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
545: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
551: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
552: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
553: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
554: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
555: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
556: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
557: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.186 schwarze 558: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 559: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 560: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
561: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
562: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
563: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
564: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 565: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
566: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
567: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 568: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.178 schwarze 569: { "it", roff_it, NULL, NULL, 0, NULL },
1.251 schwarze 570: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
571: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
572: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
573: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
574: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
575: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
576: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
577: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
578: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
579: { "length", roff_unsupp, NULL, NULL, 0, NULL },
580: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
581: { "lf", roff_insec, NULL, NULL, 0, NULL },
582: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
583: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
584: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
585: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
586: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
587: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
588: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
589: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
590: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
591: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
592: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
593: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
594: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
595: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.256 schwarze 596: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 597: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
598: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 599: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
600: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
601: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
602: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.104 kristaps 603: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.251 schwarze 604: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
605: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 606: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 607: { "nx", roff_insec, NULL, NULL, 0, NULL },
608: { "open", roff_insec, NULL, NULL, 0, NULL },
609: { "opena", roff_insec, NULL, NULL, 0, NULL },
610: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
611: { "output", roff_unsupp, NULL, NULL, 0, NULL },
612: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
615: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
616: { "pi", roff_insec, NULL, NULL, 0, NULL },
617: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.229 schwarze 618: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 619: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
620: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
621: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
622: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 623: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 624: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.255 schwarze 625: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 626: { "pso", roff_insec, NULL, NULL, 0, NULL },
627: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
628: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
629: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
630: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
631: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
632: { "return", roff_unsupp, NULL, NULL, 0, NULL },
633: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
634: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
635: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.122 schwarze 636: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.251 schwarze 637: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
638: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.203 schwarze 639: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.251 schwarze 640: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
641: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
642: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
643: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
644: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
645: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
646: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.105 kristaps 647: { "so", roff_so, NULL, NULL, 0, NULL },
1.251 schwarze 648: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
649: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
650: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
651: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
652: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
653: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
654: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
655: { "sy", roff_insec, NULL, NULL, 0, NULL },
656: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.255 schwarze 657: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 658: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
659: { "TE", roff_TE, NULL, NULL, 0, NULL },
660: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.255 schwarze 661: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 662: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
663: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
664: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
665: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
666: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.164 kristaps 667: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.251 schwarze 668: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
670: { "trf", roff_insec, NULL, NULL, 0, NULL },
671: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
672: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
673: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
674: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.109 kristaps 675: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.251 schwarze 676: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
677: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
678: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
679: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
680: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
681: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
682: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
683: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
684: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
685: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
686: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
687: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
688: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
689: { "while", roff_unsupp, NULL, NULL, 0, NULL },
690: { "write", roff_insec, NULL, NULL, 0, NULL },
691: { "writec", roff_insec, NULL, NULL, 0, NULL },
692: { "writem", roff_insec, NULL, NULL, 0, NULL },
693: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 694: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.106 kristaps 695: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.67 kristaps 696: };
697:
1.200 schwarze 698: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175 schwarze 699: const char *const __mdoc_reserved[] = {
700: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
701: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
702: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
703: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200 schwarze 704: "Dt", "Dv", "Dx", "D1",
705: "Ec", "Ed", "Ef", "Ek", "El", "Em",
706: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175 schwarze 707: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200 schwarze 708: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
709: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175 schwarze 710: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200 schwarze 711: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
712: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
713: "Sc", "Sh", "Sm", "So", "Sq",
1.175 schwarze 714: "Ss", "St", "Sx", "Sy",
715: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200 schwarze 716: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175 schwarze 717: "%P", "%Q", "%R", "%T", "%U", "%V",
718: NULL
719: };
720:
1.200 schwarze 721: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175 schwarze 722: const char *const __man_reserved[] = {
1.200 schwarze 723: "AT", "B", "BI", "BR", "DT",
724: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
725: "LP", "OP", "P", "PD", "PP",
726: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
727: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175 schwarze 728: NULL
729: };
730:
1.141 kristaps 731: /* Array of injected predefined strings. */
732: #define PREDEFS_MAX 38
733: static const struct predef predefs[PREDEFS_MAX] = {
734: #include "predefs.in"
735: };
736:
1.155 kristaps 737: /* See roffhash_find() */
1.85 kristaps 738: #define ROFF_HASH(p) (p[0] - ASCII_LO)
739:
1.178 schwarze 740: static int roffit_lines; /* number of lines to delay */
741: static char *roffit_macro; /* nil-terminated macro line */
742:
1.207 schwarze 743:
1.266 schwarze 744: /* --- request table ------------------------------------------------------ */
745:
1.85 kristaps 746: static void
1.155 kristaps 747: roffhash_init(void)
1.85 kristaps 748: {
749: struct roffmac *n;
750: int buc, i;
751:
1.106 kristaps 752: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.85 kristaps 753: assert(roffs[i].name[0] >= ASCII_LO);
754: assert(roffs[i].name[0] <= ASCII_HI);
755:
756: buc = ROFF_HASH(roffs[i].name);
757:
758: if (NULL != (n = hash[buc])) {
759: for ( ; n->next; n = n->next)
760: /* Do nothing. */ ;
761: n->next = &roffs[i];
762: } else
763: hash[buc] = &roffs[i];
764: }
765: }
766:
1.67 kristaps 767: /*
768: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
769: * the nil-terminated string name could be found.
770: */
771: static enum rofft
1.155 kristaps 772: roffhash_find(const char *p, size_t s)
1.67 kristaps 773: {
1.85 kristaps 774: int buc;
775: struct roffmac *n;
1.67 kristaps 776:
1.85 kristaps 777: /*
778: * libroff has an extremely simple hashtable, for the time
779: * being, which simply keys on the first character, which must
780: * be printable, then walks a chain. It works well enough until
781: * optimised.
782: */
783:
784: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
785: return(ROFF_MAX);
786:
787: buc = ROFF_HASH(p);
788:
789: if (NULL == (n = hash[buc]))
790: return(ROFF_MAX);
791: for ( ; n; n = n->next)
1.106 kristaps 792: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.85 kristaps 793: return((enum rofft)(n - roffs));
1.67 kristaps 794:
795: return(ROFF_MAX);
796: }
797:
1.266 schwarze 798: /* --- stack of request blocks -------------------------------------------- */
799:
1.67 kristaps 800: /*
801: * Pop the current node off of the stack of roff instructions currently
802: * pending.
803: */
804: static void
805: roffnode_pop(struct roff *r)
806: {
807: struct roffnode *p;
808:
1.75 kristaps 809: assert(r->last);
1.207 schwarze 810: p = r->last;
1.82 kristaps 811:
1.75 kristaps 812: r->last = r->last->parent;
1.106 kristaps 813: free(p->name);
814: free(p->end);
1.67 kristaps 815: free(p);
816: }
817:
818: /*
819: * Push a roff node onto the instruction stack. This must later be
820: * removed with roffnode_pop().
821: */
1.98 schwarze 822: static void
1.106 kristaps 823: roffnode_push(struct roff *r, enum rofft tok, const char *name,
824: int line, int col)
1.67 kristaps 825: {
826: struct roffnode *p;
827:
1.98 schwarze 828: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 829: p->tok = tok;
1.106 kristaps 830: if (name)
831: p->name = mandoc_strdup(name);
1.67 kristaps 832: p->parent = r->last;
833: p->line = line;
834: p->col = col;
1.198 schwarze 835: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 836:
837: r->last = p;
838: }
839:
1.266 schwarze 840: /* --- roff parser state data management ---------------------------------- */
841:
1.67 kristaps 842: static void
843: roff_free1(struct roff *r)
844: {
1.176 schwarze 845: struct tbl_node *tbl;
1.125 kristaps 846: struct eqn_node *e;
1.167 kristaps 847: int i;
1.67 kristaps 848:
1.176 schwarze 849: while (NULL != (tbl = r->first_tbl)) {
850: r->first_tbl = tbl->next;
851: tbl_free(tbl);
1.109 kristaps 852: }
1.113 kristaps 853: r->first_tbl = r->last_tbl = r->tbl = NULL;
854:
1.125 kristaps 855: while (NULL != (e = r->first_eqn)) {
856: r->first_eqn = e->next;
857: eqn_free(e);
858: }
859: r->first_eqn = r->last_eqn = r->eqn = NULL;
860:
1.67 kristaps 861: while (r->last)
862: roffnode_pop(r);
1.109 kristaps 863:
1.223 schwarze 864: free (r->rstack);
865: r->rstack = NULL;
866: r->rstacksz = 0;
867: r->rstackpos = -1;
868:
869: roff_freereg(r->regtab);
870: r->regtab = NULL;
871:
1.167 kristaps 872: roff_freestr(r->strtab);
873: roff_freestr(r->xmbtab);
874: r->strtab = r->xmbtab = NULL;
875:
876: if (r->xtab)
877: for (i = 0; i < 128; i++)
878: free(r->xtab[i].p);
879: free(r->xtab);
880: r->xtab = NULL;
1.67 kristaps 881: }
882:
883: void
884: roff_reset(struct roff *r)
885: {
886:
887: roff_free1(r);
1.227 schwarze 888: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174 kristaps 889: r->control = 0;
1.67 kristaps 890: }
891:
892: void
893: roff_free(struct roff *r)
894: {
895:
896: roff_free1(r);
897: free(r);
898: }
899:
900: struct roff *
1.237 schwarze 901: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.67 kristaps 902: {
903: struct roff *r;
904:
1.98 schwarze 905: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 906: r->parse = parse;
1.237 schwarze 907: r->mchars = mchars;
1.199 schwarze 908: r->options = options;
1.227 schwarze 909: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 910: r->rstackpos = -1;
1.207 schwarze 911:
1.155 kristaps 912: roffhash_init();
1.141 kristaps 913:
1.67 kristaps 914: return(r);
1.265 schwarze 915: }
916:
1.266 schwarze 917: /* --- syntax tree state data management ---------------------------------- */
918:
1.265 schwarze 919: static void
920: roff_man_free1(struct roff_man *man)
921: {
922:
1.266 schwarze 923: if (man->first != NULL)
924: roff_node_delete(man, man->first);
1.265 schwarze 925: free(man->meta.msec);
926: free(man->meta.vol);
927: free(man->meta.os);
928: free(man->meta.arch);
929: free(man->meta.title);
930: free(man->meta.name);
931: free(man->meta.date);
932: }
933:
934: static void
935: roff_man_alloc1(struct roff_man *man)
936: {
937:
938: memset(&man->meta, 0, sizeof(man->meta));
939: man->first = mandoc_calloc(1, sizeof(*man->first));
940: man->first->type = ROFFT_ROOT;
941: man->last = man->first;
942: man->last_es = NULL;
943: man->flags = 0;
944: man->macroset = MACROSET_NONE;
945: man->lastsec = man->lastnamed = SEC_NONE;
946: man->next = ROFF_NEXT_CHILD;
947: }
948:
949: void
950: roff_man_reset(struct roff_man *man)
951: {
952:
953: roff_man_free1(man);
954: roff_man_alloc1(man);
955: }
956:
957: void
958: roff_man_free(struct roff_man *man)
959: {
960:
961: roff_man_free1(man);
962: free(man);
963: }
964:
965: struct roff_man *
966: roff_man_alloc(struct roff *roff, struct mparse *parse,
967: const char *defos, int quick)
968: {
969: struct roff_man *man;
970:
971: man = mandoc_calloc(1, sizeof(*man));
972: man->parse = parse;
973: man->roff = roff;
974: man->defos = defos;
975: man->quick = quick;
976: roff_man_alloc1(man);
977: return(man);
1.67 kristaps 978: }
979:
1.266 schwarze 980: /* --- syntax tree handling ----------------------------------------------- */
981:
982: struct roff_node *
983: roff_node_alloc(struct roff_man *man, int line, int pos,
984: enum roff_type type, int tok)
985: {
986: struct roff_node *n;
987:
988: n = mandoc_calloc(1, sizeof(*n));
989: n->line = line;
990: n->pos = pos;
991: n->tok = tok;
992: n->type = type;
993: n->sec = man->lastsec;
994:
995: if (man->flags & MDOC_SYNOPSIS)
996: n->flags |= MDOC_SYNPRETTY;
997: else
998: n->flags &= ~MDOC_SYNPRETTY;
999: if (man->flags & MDOC_NEWLINE)
1000: n->flags |= MDOC_LINE;
1001: man->flags &= ~MDOC_NEWLINE;
1002:
1003: return(n);
1004: }
1005:
1006: void
1007: roff_node_append(struct roff_man *man, struct roff_node *n)
1008: {
1009:
1010: switch (man->next) {
1011: case ROFF_NEXT_SIBLING:
1012: man->last->next = n;
1013: n->prev = man->last;
1014: n->parent = man->last->parent;
1015: break;
1016: case ROFF_NEXT_CHILD:
1017: man->last->child = n;
1018: n->parent = man->last;
1019: break;
1020: default:
1021: abort();
1022: /* NOTREACHED */
1023: }
1024: n->parent->nchild++;
1025:
1026: /*
1027: * Copy over the normalised-data pointer of our parent. Not
1028: * everybody has one, but copying a null pointer is fine.
1029: */
1030:
1031: switch (n->type) {
1032: case ROFFT_BODY:
1033: if (n->end != ENDBODY_NOT)
1034: break;
1035: /* FALLTHROUGH */
1036: case ROFFT_TAIL:
1037: /* FALLTHROUGH */
1038: case ROFFT_HEAD:
1039: n->norm = n->parent->norm;
1040: break;
1041: default:
1042: break;
1043: }
1044:
1045: if (man->macroset == MACROSET_MDOC)
1046: mdoc_valid_pre(man, n);
1047:
1048: switch (n->type) {
1049: case ROFFT_HEAD:
1050: assert(n->parent->type == ROFFT_BLOCK);
1051: n->parent->head = n;
1052: break;
1053: case ROFFT_BODY:
1054: if (n->end)
1055: break;
1056: assert(n->parent->type == ROFFT_BLOCK);
1057: n->parent->body = n;
1058: break;
1059: case ROFFT_TAIL:
1060: assert(n->parent->type == ROFFT_BLOCK);
1061: n->parent->tail = n;
1062: break;
1063: default:
1064: break;
1065: }
1066: man->last = n;
1067: }
1068:
1.267 ! schwarze 1069: void
! 1070: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
! 1071: {
! 1072: struct roff_node *n;
! 1073:
! 1074: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
! 1075: n->string = roff_strdup(man->roff, word);
! 1076: roff_node_append(man, n);
! 1077: if (man->macroset == MACROSET_MDOC)
! 1078: mdoc_valid_post(man);
! 1079: else
! 1080: man_valid_post(man);
! 1081: man->next = ROFF_NEXT_SIBLING;
! 1082: }
! 1083:
! 1084: void
! 1085: roff_word_append(struct roff_man *man, const char *word)
! 1086: {
! 1087: struct roff_node *n;
! 1088: char *addstr, *newstr;
! 1089:
! 1090: n = man->last;
! 1091: addstr = roff_strdup(man->roff, word);
! 1092: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
! 1093: free(addstr);
! 1094: free(n->string);
! 1095: n->string = newstr;
! 1096: man->next = ROFF_NEXT_SIBLING;
! 1097: }
! 1098:
1.266 schwarze 1099: struct roff_node *
1100: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
1101: {
1102: struct roff_node *n;
1103:
1104: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
1105: roff_node_append(man, n);
1106: man->next = ROFF_NEXT_CHILD;
1107: return(n);
1108: }
1109:
1110: struct roff_node *
1111: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1112: {
1113: struct roff_node *n;
1114:
1115: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1116: roff_node_append(man, n);
1117: man->next = ROFF_NEXT_CHILD;
1118: return(n);
1.267 ! schwarze 1119: }
! 1120:
! 1121: void
! 1122: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
! 1123: {
! 1124: struct roff_node *n;
! 1125:
! 1126: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
! 1127: n->eqn = eqn;
! 1128: if (eqn->ln > man->last->line)
! 1129: n->flags |= MDOC_LINE;
! 1130: roff_node_append(man, n);
! 1131: man->next = ROFF_NEXT_SIBLING;
! 1132: }
! 1133:
! 1134: void
! 1135: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
! 1136: {
! 1137: struct roff_node *n;
! 1138:
! 1139: if (man->macroset == MACROSET_MAN)
! 1140: man_breakscope(man, TOKEN_NONE);
! 1141: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
! 1142: n->span = tbl;
! 1143: roff_node_append(man, n);
! 1144: if (man->macroset == MACROSET_MDOC)
! 1145: mdoc_valid_post(man);
! 1146: else
! 1147: man_valid_post(man);
! 1148: man->next = ROFF_NEXT_SIBLING;
1.266 schwarze 1149: }
1150:
1151: void
1152: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1153: {
1154:
1155: /* Adjust siblings. */
1156:
1157: if (n->prev)
1158: n->prev->next = n->next;
1159: if (n->next)
1160: n->next->prev = n->prev;
1161:
1162: /* Adjust parent. */
1163:
1164: if (n->parent != NULL) {
1165: n->parent->nchild--;
1166: if (n->parent->child == n)
1167: n->parent->child = n->next;
1168: if (n->parent->last == n)
1169: n->parent->last = n->prev;
1170: }
1171:
1172: /* Adjust parse point. */
1173:
1174: if (man == NULL)
1175: return;
1176: if (man->last == n) {
1177: if (n->prev == NULL) {
1178: man->last = n->parent;
1179: man->next = ROFF_NEXT_CHILD;
1180: } else {
1181: man->last = n->prev;
1182: man->next = ROFF_NEXT_SIBLING;
1183: }
1184: }
1185: if (man->first == n)
1186: man->first = NULL;
1187: }
1188:
1189: void
1190: roff_node_free(struct roff_node *n)
1191: {
1192:
1193: if (n->args != NULL)
1194: mdoc_argv_free(n->args);
1195: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1196: free(n->norm);
1197: free(n->string);
1198: free(n);
1199: }
1200:
1201: void
1202: roff_node_delete(struct roff_man *man, struct roff_node *n)
1203: {
1204:
1205: while (n->child != NULL)
1206: roff_node_delete(man, n->child);
1207: assert(n->nchild == 0);
1208: roff_node_unlink(man, n);
1209: roff_node_free(n);
1210: }
1211:
1212: /* --- main functions of the roff parser ---------------------------------- */
1213:
1.94 kristaps 1214: /*
1.206 schwarze 1215: * In the current line, expand escape sequences that tend to get
1216: * used in numerical expressions and conditional requests.
1217: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1218: */
1.172 schwarze 1219: static enum rofferr
1.238 schwarze 1220: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1221: {
1.208 schwarze 1222: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1223: const char *start; /* start of the string to process */
1.209 schwarze 1224: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1225: const char *stnam; /* start of the name, after "[(*" */
1226: const char *cp; /* end of the name, e.g. before ']' */
1227: const char *res; /* the string to be substituted */
1.238 schwarze 1228: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1229: size_t maxl; /* expected length of the escape name */
1230: size_t naml; /* actual length of the escape name */
1.237 schwarze 1231: enum mandoc_esc esc; /* type of the escape sequence */
1232: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1233: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1234: int npos; /* position in numeric expression */
1.218 schwarze 1235: int arg_complete; /* argument not interrupted by eol */
1.206 schwarze 1236: char term; /* character terminating the escape */
1.94 kristaps 1237:
1.170 schwarze 1238: expand_count = 0;
1.238 schwarze 1239: start = buf->buf + pos;
1.205 schwarze 1240: stesc = strchr(start, '\0') - 1;
1241: while (stesc-- > start) {
1.170 schwarze 1242:
1.205 schwarze 1243: /* Search backwards for the next backslash. */
1244:
1.238 schwarze 1245: if (*stesc != '\\')
1.205 schwarze 1246: continue;
1247:
1248: /* If it is escaped, skip it. */
1249:
1250: for (cp = stesc - 1; cp >= start; cp--)
1.238 schwarze 1251: if (*cp != '\\')
1.205 schwarze 1252: break;
1253:
1.238 schwarze 1254: if ((stesc - cp) % 2 == 0) {
1.209 schwarze 1255: stesc = (char *)cp;
1.205 schwarze 1256: continue;
1257: }
1.108 schwarze 1258:
1.206 schwarze 1259: /* Decide whether to expand or to check only. */
1.108 schwarze 1260:
1.206 schwarze 1261: term = '\0';
1.205 schwarze 1262: cp = stesc + 1;
1.181 schwarze 1263: switch (*cp) {
1.207 schwarze 1264: case '*':
1.181 schwarze 1265: res = NULL;
1266: break;
1.207 schwarze 1267: case 'B':
1.206 schwarze 1268: /* FALLTHROUGH */
1.207 schwarze 1269: case 'w':
1.206 schwarze 1270: term = cp[1];
1271: /* FALLTHROUGH */
1.207 schwarze 1272: case 'n':
1.181 schwarze 1273: res = ubuf;
1274: break;
1275: default:
1.237 schwarze 1276: esc = mandoc_escape(&cp, &stnam, &inaml);
1277: if (esc == ESCAPE_ERROR ||
1278: (esc == ESCAPE_SPECIAL &&
1279: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.219 schwarze 1280: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1281: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1282: "%.*s", (int)(cp - stesc), stesc);
1.205 schwarze 1283: continue;
1.152 kristaps 1284: }
1285:
1.205 schwarze 1286: if (EXPAND_LIMIT < ++expand_count) {
1287: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1288: ln, (int)(stesc - buf->buf), NULL);
1.205 schwarze 1289: return(ROFF_IGN);
1290: }
1.108 schwarze 1291:
1292: /*
1293: * The third character decides the length
1.181 schwarze 1294: * of the name of the string or register.
1.108 schwarze 1295: * Save a pointer to the name.
1296: */
1297:
1.238 schwarze 1298: if (term == '\0') {
1.206 schwarze 1299: switch (*++cp) {
1.207 schwarze 1300: case '\0':
1.206 schwarze 1301: maxl = 0;
1302: break;
1.207 schwarze 1303: case '(':
1.206 schwarze 1304: cp++;
1305: maxl = 2;
1306: break;
1.207 schwarze 1307: case '[':
1.206 schwarze 1308: cp++;
1309: term = ']';
1310: maxl = 0;
1311: break;
1312: default:
1313: maxl = 1;
1314: break;
1315: }
1316: } else {
1317: cp += 2;
1.94 kristaps 1318: maxl = 0;
1319: }
1.108 schwarze 1320: stnam = cp;
1.94 kristaps 1321:
1.108 schwarze 1322: /* Advance to the end of the name. */
1.94 kristaps 1323:
1.253 schwarze 1324: naml = 0;
1.218 schwarze 1325: arg_complete = 1;
1.253 schwarze 1326: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1327: if (*cp == '\0') {
1.219 schwarze 1328: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1329: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1330: arg_complete = 0;
1.206 schwarze 1331: break;
1.153 kristaps 1332: }
1.238 schwarze 1333: if (maxl == 0 && *cp == term) {
1.206 schwarze 1334: cp++;
1.253 schwarze 1335: break;
1336: }
1337: if (*cp++ != '\\' || stesc[1] != 'w') {
1338: naml++;
1339: continue;
1340: }
1341: switch (mandoc_escape(&cp, NULL, NULL)) {
1342: case ESCAPE_SPECIAL:
1343: /* FALLTHROUGH */
1344: case ESCAPE_UNICODE:
1345: /* FALLTHROUGH */
1346: case ESCAPE_NUMBERED:
1347: /* FALLTHROUGH */
1348: case ESCAPE_OVERSTRIKE:
1349: naml++;
1350: break;
1351: default:
1.94 kristaps 1352: break;
1.206 schwarze 1353: }
1.94 kristaps 1354: }
1355:
1.108 schwarze 1356: /*
1357: * Retrieve the replacement string; if it is
1358: * undefined, resume searching for escapes.
1359: */
1360:
1.206 schwarze 1361: switch (stesc[1]) {
1.207 schwarze 1362: case '*':
1.218 schwarze 1363: if (arg_complete)
1364: res = roff_getstrn(r, stnam, naml);
1.206 schwarze 1365: break;
1.207 schwarze 1366: case 'B':
1.206 schwarze 1367: npos = 0;
1.218 schwarze 1368: ubuf[0] = arg_complete &&
1.261 schwarze 1369: roff_evalnum(r, ln, stnam, &npos,
1370: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1371: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1372: ubuf[1] = '\0';
1373: break;
1.207 schwarze 1374: case 'n':
1.218 schwarze 1375: if (arg_complete)
1376: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1377: roff_getregn(r, stnam, naml));
1378: else
1379: ubuf[0] = '\0';
1.206 schwarze 1380: break;
1.207 schwarze 1381: case 'w':
1.218 schwarze 1382: /* use even incomplete args */
1.208 schwarze 1383: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1384: 24 * (int)naml);
1385: break;
1386: }
1.94 kristaps 1387:
1.238 schwarze 1388: if (res == NULL) {
1.219 schwarze 1389: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1390: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1391: "%.*s", (int)naml, stnam);
1.142 kristaps 1392: res = "";
1.246 schwarze 1393: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1394: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1395: ln, (int)(stesc - buf->buf), NULL);
1396: return(ROFF_IGN);
1.94 kristaps 1397: }
1398:
1.108 schwarze 1399: /* Replace the escape sequence by the string. */
1400:
1.209 schwarze 1401: *stesc = '\0';
1.238 schwarze 1402: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1403: buf->buf, res, cp) + 1;
1.94 kristaps 1404:
1.205 schwarze 1405: /* Prepare for the next replacement. */
1.94 kristaps 1406:
1.205 schwarze 1407: start = nbuf + pos;
1.238 schwarze 1408: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1409: free(buf->buf);
1410: buf->buf = nbuf;
1.154 kristaps 1411: }
1.172 schwarze 1412: return(ROFF_CONT);
1.154 kristaps 1413: }
1414:
1415: /*
1.178 schwarze 1416: * Process text streams:
1417: * Convert all breakable hyphens into ASCII_HYPH.
1418: * Decrement and spring input line trap.
1.154 kristaps 1419: */
1420: static enum rofferr
1.238 schwarze 1421: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154 kristaps 1422: {
1423: size_t sz;
1424: const char *start;
1.178 schwarze 1425: char *p;
1426: int isz;
1.154 kristaps 1427: enum mandoc_esc esc;
1428:
1.238 schwarze 1429: start = p = buf->buf + pos;
1.154 kristaps 1430:
1.238 schwarze 1431: while (*p != '\0') {
1.154 kristaps 1432: sz = strcspn(p, "-\\");
1433: p += sz;
1434:
1.238 schwarze 1435: if (*p == '\0')
1.159 kristaps 1436: break;
1437:
1.238 schwarze 1438: if (*p == '\\') {
1.154 kristaps 1439: /* Skip over escapes. */
1440: p++;
1.189 schwarze 1441: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1442: if (esc == ESCAPE_ERROR)
1.154 kristaps 1443: break;
1.264 schwarze 1444: while (*p == '-')
1445: p++;
1.155 kristaps 1446: continue;
1.159 kristaps 1447: } else if (p == start) {
1.158 kristaps 1448: p++;
1.155 kristaps 1449: continue;
1.158 kristaps 1450: }
1.155 kristaps 1451:
1.171 schwarze 1452: if (isalpha((unsigned char)p[-1]) &&
1453: isalpha((unsigned char)p[1]))
1.155 kristaps 1454: *p = ASCII_HYPH;
1455: p++;
1.94 kristaps 1456: }
1457:
1.178 schwarze 1458: /* Spring the input line trap. */
1.238 schwarze 1459: if (roffit_lines == 1) {
1460: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1461: free(buf->buf);
1462: buf->buf = p;
1463: buf->sz = isz + 1;
1.178 schwarze 1464: *offs = 0;
1465: free(roffit_macro);
1466: roffit_lines = 0;
1467: return(ROFF_REPARSE);
1.238 schwarze 1468: } else if (roffit_lines > 1)
1.178 schwarze 1469: --roffit_lines;
1.154 kristaps 1470: return(ROFF_CONT);
1.94 kristaps 1471: }
1472:
1.67 kristaps 1473: enum rofferr
1.238 schwarze 1474: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1475: {
1476: enum rofft t;
1.109 kristaps 1477: enum rofferr e;
1.238 schwarze 1478: int pos; /* parse point */
1.243 schwarze 1479: int spos; /* saved parse point for messages */
1.238 schwarze 1480: int ppos; /* original offset in buf->buf */
1481: int ctl; /* macro line (boolean) */
1482:
1483: ppos = pos = *offs;
1.79 kristaps 1484:
1.230 schwarze 1485: /* Handle in-line equation delimiters. */
1486:
1.236 schwarze 1487: if (r->tbl == NULL &&
1488: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1489: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1490: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1491: if (e == ROFF_REPARSE)
1492: return(e);
1493: assert(e == ROFF_CONT);
1494: }
1495:
1496: /* Expand some escape sequences. */
1.94 kristaps 1497:
1.238 schwarze 1498: e = roff_res(r, buf, ln, pos);
1499: if (e == ROFF_IGN)
1.172 schwarze 1500: return(e);
1.238 schwarze 1501: assert(e == ROFF_CONT);
1.94 kristaps 1502:
1.238 schwarze 1503: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1504:
1.94 kristaps 1505: /*
1.79 kristaps 1506: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1507: * text through the macro's filter.
1508: * Equations process all content themselves.
1509: * Tables process almost all content themselves, but we want
1510: * to warn about macros before passing it there.
1.79 kristaps 1511: */
1.74 kristaps 1512:
1.252 schwarze 1513: if (r->last != NULL && ! ctl) {
1.78 kristaps 1514: t = r->last->tok;
1515: assert(roffs[t].text);
1.238 schwarze 1516: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1517: assert(e == ROFF_IGN || e == ROFF_CONT);
1518: if (e != ROFF_CONT)
1.125 kristaps 1519: return(e);
1.182 schwarze 1520: }
1.252 schwarze 1521: if (r->eqn != NULL)
1.238 schwarze 1522: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.252 schwarze 1523: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.259 schwarze 1524: return(tbl_read(r->tbl, ln, buf->buf, ppos));
1.252 schwarze 1525: if ( ! ctl)
1.238 schwarze 1526: return(roff_parsetext(buf, pos, offs));
1.228 schwarze 1527:
1528: /* Skip empty request lines. */
1529:
1.238 schwarze 1530: if (buf->buf[pos] == '"') {
1.228 schwarze 1531: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1532: ln, pos, NULL);
1533: return(ROFF_IGN);
1.238 schwarze 1534: } else if (buf->buf[pos] == '\0')
1.228 schwarze 1535: return(ROFF_IGN);
1.67 kristaps 1536:
1.79 kristaps 1537: /*
1538: * If a scope is open, go to the child handler for that macro,
1539: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1540: * Don't do so if an equation is open.
1.79 kristaps 1541: */
1.78 kristaps 1542:
1.79 kristaps 1543: if (r->last) {
1544: t = r->last->tok;
1545: assert(roffs[t].sub);
1.238 schwarze 1546: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1547: }
1.78 kristaps 1548:
1.243 schwarze 1549: /* No scope is open. This is a new request or macro. */
1550:
1551: spos = pos;
1552: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1553:
1554: /* Tables ignore most macros. */
1555:
1556: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1557: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1558: ln, pos, buf->buf + spos);
1.257 schwarze 1559: if (t == ROFF_TS)
1560: return(ROFF_IGN);
1561: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1562: pos++;
1563: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1564: pos++;
1565: return(tbl_read(r->tbl, ln, buf->buf, pos));
1.243 schwarze 1566: }
1567:
1.79 kristaps 1568: /*
1.243 schwarze 1569: * This is neither a roff request nor a user-defined macro.
1570: * Let the standard macro set parsers handle it.
1.79 kristaps 1571: */
1.67 kristaps 1572:
1.243 schwarze 1573: if (t == ROFF_MAX)
1.79 kristaps 1574: return(ROFF_CONT);
1.243 schwarze 1575:
1576: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1577:
1.75 kristaps 1578: assert(roffs[t].proc);
1.238 schwarze 1579: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.74 kristaps 1580: }
1581:
1.117 kristaps 1582: void
1.74 kristaps 1583: roff_endparse(struct roff *r)
1584: {
1585:
1.110 kristaps 1586: if (r->last)
1.221 schwarze 1587: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1588: r->last->line, r->last->col,
1589: roffs[r->last->tok].name);
1.117 kristaps 1590:
1.125 kristaps 1591: if (r->eqn) {
1.221 schwarze 1592: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1593: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151 kristaps 1594: eqn_end(&r->eqn);
1.125 kristaps 1595: }
1596:
1.117 kristaps 1597: if (r->tbl) {
1.221 schwarze 1598: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1599: r->tbl->line, r->tbl->pos, "TS");
1.151 kristaps 1600: tbl_end(&r->tbl);
1.117 kristaps 1601: }
1.67 kristaps 1602: }
1603:
1604: /*
1605: * Parse a roff node's type from the input buffer. This must be in the
1606: * form of ".foo xxx" in the usual way.
1607: */
1608: static enum rofft
1.214 schwarze 1609: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1610: {
1.214 schwarze 1611: char *cp;
1.106 kristaps 1612: const char *mac;
1613: size_t maclen;
1.67 kristaps 1614: enum rofft t;
1615:
1.214 schwarze 1616: cp = buf + *pos;
1617:
1618: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.67 kristaps 1619: return(ROFF_MAX);
1620:
1.214 schwarze 1621: mac = cp;
1622: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1623:
1.106 kristaps 1624: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.155 kristaps 1625: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.67 kristaps 1626:
1.214 schwarze 1627: if (ROFF_MAX != t)
1628: *pos = cp - buf;
1.67 kristaps 1629:
1630: return(t);
1631: }
1632:
1.266 schwarze 1633: /* --- handling of request blocks ----------------------------------------- */
1634:
1.67 kristaps 1635: static enum rofferr
1.76 kristaps 1636: roff_cblock(ROFF_ARGS)
1.67 kristaps 1637: {
1638:
1.79 kristaps 1639: /*
1640: * A block-close `..' should only be invoked as a child of an
1641: * ignore macro, otherwise raise a warning and just ignore it.
1642: */
1643:
1.238 schwarze 1644: if (r->last == NULL) {
1.221 schwarze 1645: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1646: ln, ppos, "..");
1.76 kristaps 1647: return(ROFF_IGN);
1648: }
1.67 kristaps 1649:
1.81 kristaps 1650: switch (r->last->tok) {
1.207 schwarze 1651: case ROFF_am:
1.220 schwarze 1652: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.81 kristaps 1653: /* FALLTHROUGH */
1.207 schwarze 1654: case ROFF_ami:
1.81 kristaps 1655: /* FALLTHROUGH */
1.207 schwarze 1656: case ROFF_de:
1.108 schwarze 1657: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.81 kristaps 1658: /* FALLTHROUGH */
1.207 schwarze 1659: case ROFF_dei:
1.81 kristaps 1660: /* FALLTHROUGH */
1.207 schwarze 1661: case ROFF_ig:
1.81 kristaps 1662: break;
1663: default:
1.221 schwarze 1664: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1665: ln, ppos, "..");
1.67 kristaps 1666: return(ROFF_IGN);
1.76 kristaps 1667: }
1.67 kristaps 1668:
1.238 schwarze 1669: if (buf->buf[pos] != '\0')
1.217 schwarze 1670: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1671: ".. %s", buf->buf + pos);
1.71 kristaps 1672:
1673: roffnode_pop(r);
1.76 kristaps 1674: roffnode_cleanscope(r);
1675: return(ROFF_IGN);
1.71 kristaps 1676:
1.67 kristaps 1677: }
1678:
1.76 kristaps 1679: static void
1680: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1681: {
1682:
1.76 kristaps 1683: while (r->last) {
1.173 schwarze 1684: if (--r->last->endspan != 0)
1.76 kristaps 1685: break;
1686: roffnode_pop(r);
1687: }
1.67 kristaps 1688: }
1689:
1.195 schwarze 1690: static void
1691: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1692: {
1693:
1.76 kristaps 1694: if (NULL == r->last) {
1.221 schwarze 1695: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1696: ln, ppos, "\\}");
1.195 schwarze 1697: return;
1.76 kristaps 1698: }
1699:
1.82 kristaps 1700: switch (r->last->tok) {
1.207 schwarze 1701: case ROFF_el:
1.82 kristaps 1702: /* FALLTHROUGH */
1.207 schwarze 1703: case ROFF_ie:
1.82 kristaps 1704: /* FALLTHROUGH */
1.207 schwarze 1705: case ROFF_if:
1.82 kristaps 1706: break;
1707: default:
1.221 schwarze 1708: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1709: ln, ppos, "\\}");
1.195 schwarze 1710: return;
1.75 kristaps 1711: }
1712:
1.76 kristaps 1713: if (r->last->endspan > -1) {
1.221 schwarze 1714: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1715: ln, ppos, "\\}");
1.195 schwarze 1716: return;
1.76 kristaps 1717: }
1718:
1.75 kristaps 1719: roffnode_pop(r);
1.76 kristaps 1720: roffnode_cleanscope(r);
1.195 schwarze 1721: return;
1.76 kristaps 1722: }
1723:
1724: static enum rofferr
1.80 kristaps 1725: roff_block(ROFF_ARGS)
1.76 kristaps 1726: {
1.220 schwarze 1727: const char *name;
1728: char *iname, *cp;
1.213 schwarze 1729: size_t namesz;
1.106 kristaps 1730:
1.220 schwarze 1731: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1732:
1.238 schwarze 1733: if (tok == ROFF_de1)
1.220 schwarze 1734: tok = ROFF_de;
1.251 schwarze 1735: else if (tok == ROFF_dei1)
1736: tok = ROFF_dei;
1.238 schwarze 1737: else if (tok == ROFF_am1)
1.220 schwarze 1738: tok = ROFF_am;
1.251 schwarze 1739: else if (tok == ROFF_ami1)
1740: tok = ROFF_ami;
1.220 schwarze 1741:
1742: /* Parse the macro name argument. */
1743:
1.238 schwarze 1744: cp = buf->buf + pos;
1745: if (tok == ROFF_ig) {
1.220 schwarze 1746: iname = NULL;
1747: namesz = 0;
1748: } else {
1749: iname = cp;
1750: namesz = roff_getname(r, &cp, ln, ppos);
1751: iname[namesz] = '\0';
1752: }
1.107 kristaps 1753:
1.220 schwarze 1754: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1755:
1.238 schwarze 1756: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1757: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1758: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1759: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1760: "%.*s", (int)namesz, iname);
1761: namesz = 0;
1762: } else
1763: namesz = strlen(name);
1764: } else
1765: name = iname;
1.107 kristaps 1766:
1.238 schwarze 1767: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1768: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1769: ln, ppos, roffs[tok].name);
1770: return(ROFF_IGN);
1771: }
1.80 kristaps 1772:
1.106 kristaps 1773: roffnode_push(r, tok, name, ln, ppos);
1774:
1775: /*
1776: * At the beginning of a `de' macro, clear the existing string
1777: * with the same name, if there is one. New content will be
1.193 schwarze 1778: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1779: */
1.107 kristaps 1780:
1.238 schwarze 1781: if (tok == ROFF_de || tok == ROFF_dei)
1.213 schwarze 1782: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76 kristaps 1783:
1.238 schwarze 1784: if (*cp == '\0')
1.78 kristaps 1785: return(ROFF_IGN);
1786:
1.220 schwarze 1787: /* Get the custom end marker. */
1.107 kristaps 1788:
1.220 schwarze 1789: iname = cp;
1.213 schwarze 1790: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1791:
1792: /* Resolve the end marker if it is indirect. */
1793:
1.238 schwarze 1794: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1795: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1796: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1797: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1798: "%.*s", (int)namesz, iname);
1799: namesz = 0;
1800: } else
1801: namesz = strlen(name);
1802: } else
1803: name = iname;
1804:
1.213 schwarze 1805: if (namesz)
1806: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1807:
1.238 schwarze 1808: if (*cp != '\0')
1.217 schwarze 1809: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1810: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.74 kristaps 1811:
1.78 kristaps 1812: return(ROFF_IGN);
1813: }
1814:
1815: static enum rofferr
1.80 kristaps 1816: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1817: {
1818: enum rofft t;
1819: int i, j;
1820:
1821: /*
1822: * First check whether a custom macro exists at this level. If
1823: * it does, then check against it. This is some of groff's
1824: * stranger behaviours. If we encountered a custom end-scope
1825: * tag and that tag also happens to be a "real" macro, then we
1826: * need to try interpreting it again as a real macro. If it's
1827: * not, then return ignore. Else continue.
1828: */
1829:
1830: if (r->last->end) {
1.130 kristaps 1831: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1832: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1833: break;
1834:
1.238 schwarze 1835: if (r->last->end[j] == '\0' &&
1836: (buf->buf[i] == '\0' ||
1837: buf->buf[i] == ' ' ||
1838: buf->buf[i] == '\t')) {
1.79 kristaps 1839: roffnode_pop(r);
1840: roffnode_cleanscope(r);
1841:
1.238 schwarze 1842: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1843: i++;
1844:
1845: pos = i;
1.238 schwarze 1846: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1847: ROFF_MAX)
1.79 kristaps 1848: return(ROFF_RERUN);
1849: return(ROFF_IGN);
1850: }
1851: }
1852:
1853: /*
1854: * If we have no custom end-query or lookup failed, then try
1855: * pulling it out of the hashtable.
1856: */
1857:
1.238 schwarze 1858: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1859:
1.238 schwarze 1860: if (t != ROFF_cblock) {
1861: if (tok != ROFF_ig)
1862: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.79 kristaps 1863: return(ROFF_IGN);
1.106 kristaps 1864: }
1.79 kristaps 1865:
1866: assert(roffs[t].proc);
1.238 schwarze 1867: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1868: }
1869:
1870: static enum rofferr
1.80 kristaps 1871: roff_block_text(ROFF_ARGS)
1.78 kristaps 1872: {
1873:
1.238 schwarze 1874: if (tok != ROFF_ig)
1875: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1876:
1.78 kristaps 1877: return(ROFF_IGN);
1878: }
1879:
1880: static enum rofferr
1.82 kristaps 1881: roff_cond_sub(ROFF_ARGS)
1882: {
1883: enum rofft t;
1.139 kristaps 1884: char *ep;
1.198 schwarze 1885: int rr;
1.82 kristaps 1886:
1887: rr = r->last->rule;
1.139 kristaps 1888: roffnode_cleanscope(r);
1.238 schwarze 1889: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82 kristaps 1890:
1.139 kristaps 1891: /*
1.177 schwarze 1892: * Fully handle known macros when they are structurally
1893: * required or when the conditional evaluated to true.
1.87 kristaps 1894: */
1895:
1.238 schwarze 1896: if ((t != ROFF_MAX) &&
1897: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.177 schwarze 1898: assert(roffs[t].proc);
1.238 schwarze 1899: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.177 schwarze 1900: }
1.144 kristaps 1901:
1.196 schwarze 1902: /*
1903: * If `\}' occurs on a macro line without a preceding macro,
1904: * drop the line completely.
1905: */
1906:
1.238 schwarze 1907: ep = buf->buf + pos;
1908: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1909: rr = 0;
1.196 schwarze 1910:
1.177 schwarze 1911: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1912:
1.238 schwarze 1913: while ((ep = strchr(ep, '\\')) != NULL) {
1914: if (*(++ep) == '}') {
1.197 schwarze 1915: *ep = '&';
1.238 schwarze 1916: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1917: }
1.247 schwarze 1918: if (*ep != '\0')
1919: ++ep;
1.177 schwarze 1920: }
1.198 schwarze 1921: return(rr ? ROFF_CONT : ROFF_IGN);
1.82 kristaps 1922: }
1923:
1924: static enum rofferr
1925: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1926: {
1.140 kristaps 1927: char *ep;
1.198 schwarze 1928: int rr;
1.82 kristaps 1929:
1930: rr = r->last->rule;
1.140 kristaps 1931: roffnode_cleanscope(r);
1.82 kristaps 1932:
1.238 schwarze 1933: ep = buf->buf + pos;
1934: while ((ep = strchr(ep, '\\')) != NULL) {
1935: if (*(++ep) == '}') {
1.197 schwarze 1936: *ep = '&';
1.238 schwarze 1937: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1938: }
1.247 schwarze 1939: if (*ep != '\0')
1940: ++ep;
1.78 kristaps 1941: }
1.198 schwarze 1942: return(rr ? ROFF_CONT : ROFF_IGN);
1.74 kristaps 1943: }
1944:
1.266 schwarze 1945: /* --- handling of numeric and conditional expressions -------------------- */
1946:
1.204 schwarze 1947: /*
1948: * Parse a single signed integer number. Stop at the first non-digit.
1949: * If there is at least one digit, return success and advance the
1950: * parse point, else return failure and let the parse point unchanged.
1951: * Ignore overflows, treat them just like the C language.
1952: */
1.184 schwarze 1953: static int
1.261 schwarze 1954: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 1955: {
1.261 schwarze 1956: int myres, scaled, n, p;
1.206 schwarze 1957:
1958: if (NULL == res)
1959: res = &myres;
1.184 schwarze 1960:
1961: p = *pos;
1962: n = v[p] == '-';
1.261 schwarze 1963: if (n || v[p] == '+')
1.184 schwarze 1964: p++;
1965:
1.261 schwarze 1966: if (flags & ROFFNUM_WHITE)
1967: while (isspace((unsigned char)v[p]))
1968: p++;
1969:
1.184 schwarze 1970: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 1971: *res = 10 * *res + v[p] - '0';
1.184 schwarze 1972: if (p == *pos + n)
1973: return 0;
1974:
1975: if (n)
1976: *res = -*res;
1977:
1.254 schwarze 1978: /* Each number may be followed by one optional scaling unit. */
1979:
1980: switch (v[p]) {
1981: case 'f':
1.261 schwarze 1982: scaled = *res * 65536;
1.254 schwarze 1983: break;
1984: case 'i':
1.261 schwarze 1985: scaled = *res * 240;
1.254 schwarze 1986: break;
1987: case 'c':
1.261 schwarze 1988: scaled = *res * 240 / 2.54;
1.254 schwarze 1989: break;
1990: case 'v':
1991: /* FALLTROUGH */
1992: case 'P':
1.261 schwarze 1993: scaled = *res * 40;
1.254 schwarze 1994: break;
1995: case 'm':
1996: /* FALLTROUGH */
1997: case 'n':
1.261 schwarze 1998: scaled = *res * 24;
1.254 schwarze 1999: break;
2000: case 'p':
1.261 schwarze 2001: scaled = *res * 10 / 3;
1.254 schwarze 2002: break;
2003: case 'u':
1.261 schwarze 2004: scaled = *res;
1.254 schwarze 2005: break;
2006: case 'M':
1.261 schwarze 2007: scaled = *res * 6 / 25;
1.254 schwarze 2008: break;
2009: default:
1.261 schwarze 2010: scaled = *res;
1.254 schwarze 2011: p--;
2012: break;
2013: }
1.261 schwarze 2014: if (flags & ROFFNUM_SCALE)
2015: *res = scaled;
1.254 schwarze 2016:
2017: *pos = p + 1;
2018: return(1);
1.184 schwarze 2019: }
2020:
1.198 schwarze 2021: /*
2022: * Evaluate a string comparison condition.
2023: * The first character is the delimiter.
2024: * Succeed if the string up to its second occurrence
2025: * matches the string up to its third occurence.
2026: * Advance the cursor after the third occurrence
2027: * or lacking that, to the end of the line.
2028: */
2029: static int
2030: roff_evalstrcond(const char *v, int *pos)
2031: {
2032: const char *s1, *s2, *s3;
2033: int match;
2034:
2035: match = 0;
2036: s1 = v + *pos; /* initial delimiter */
2037: s2 = s1 + 1; /* for scanning the first string */
2038: s3 = strchr(s2, *s1); /* for scanning the second string */
2039:
2040: if (NULL == s3) /* found no middle delimiter */
2041: goto out;
2042:
2043: while ('\0' != *++s3) {
2044: if (*s2 != *s3) { /* mismatch */
2045: s3 = strchr(s3, *s1);
2046: break;
2047: }
2048: if (*s3 == *s1) { /* found the final delimiter */
2049: match = 1;
2050: break;
2051: }
2052: s2++;
2053: }
2054:
2055: out:
2056: if (NULL == s3)
2057: s3 = strchr(s2, '\0');
1.242 schwarze 2058: else if (*s3 != '\0')
1.198 schwarze 2059: s3++;
2060: *pos = s3 - v;
2061: return(match);
2062: }
2063:
1.204 schwarze 2064: /*
2065: * Evaluate an optionally negated single character, numerical,
2066: * or string condition.
2067: */
1.198 schwarze 2068: static int
1.234 kristaps 2069: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.88 kristaps 2070: {
1.241 schwarze 2071: int number, savepos, wanttrue;
1.88 kristaps 2072:
1.198 schwarze 2073: if ('!' == v[*pos]) {
2074: wanttrue = 0;
2075: (*pos)++;
2076: } else
2077: wanttrue = 1;
2078:
1.88 kristaps 2079: switch (v[*pos]) {
1.240 schwarze 2080: case '\0':
2081: return(0);
1.207 schwarze 2082: case 'n':
1.198 schwarze 2083: /* FALLTHROUGH */
1.207 schwarze 2084: case 'o':
1.88 kristaps 2085: (*pos)++;
1.198 schwarze 2086: return(wanttrue);
1.207 schwarze 2087: case 'c':
1.198 schwarze 2088: /* FALLTHROUGH */
1.207 schwarze 2089: case 'd':
1.198 schwarze 2090: /* FALLTHROUGH */
1.207 schwarze 2091: case 'e':
1.88 kristaps 2092: /* FALLTHROUGH */
1.207 schwarze 2093: case 'r':
1.88 kristaps 2094: /* FALLTHROUGH */
1.207 schwarze 2095: case 't':
1.239 schwarze 2096: /* FALLTHROUGH */
2097: case 'v':
1.88 kristaps 2098: (*pos)++;
1.198 schwarze 2099: return(!wanttrue);
1.88 kristaps 2100: default:
2101: break;
2102: }
2103:
1.241 schwarze 2104: savepos = *pos;
1.261 schwarze 2105: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.204 schwarze 2106: return((number > 0) == wanttrue);
1.241 schwarze 2107: else if (*pos == savepos)
2108: return(roff_evalstrcond(v, pos) == wanttrue);
1.204 schwarze 2109: else
1.241 schwarze 2110: return (0);
1.88 kristaps 2111: }
2112:
1.74 kristaps 2113: static enum rofferr
1.103 kristaps 2114: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2115: {
1.123 schwarze 2116:
1.89 kristaps 2117: return(ROFF_IGN);
2118: }
2119:
1.104 kristaps 2120: static enum rofferr
1.251 schwarze 2121: roff_insec(ROFF_ARGS)
2122: {
2123:
2124: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
2125: ln, ppos, roffs[tok].name);
2126: return(ROFF_IGN);
2127: }
2128:
2129: static enum rofferr
2130: roff_unsupp(ROFF_ARGS)
2131: {
2132:
2133: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
2134: ln, ppos, roffs[tok].name);
2135: return(ROFF_IGN);
2136: }
2137:
2138: static enum rofferr
1.82 kristaps 2139: roff_cond(ROFF_ARGS)
1.74 kristaps 2140: {
1.173 schwarze 2141:
2142: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2143:
1.207 schwarze 2144: /*
1.134 kristaps 2145: * An `.el' has no conditional body: it will consume the value
2146: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2147: * defaults to DENY.
1.134 kristaps 2148: *
2149: * If we're not an `el', however, then evaluate the conditional.
2150: */
1.133 kristaps 2151:
1.238 schwarze 2152: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2153: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2154: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2155:
1.134 kristaps 2156: /*
2157: * An if-else will put the NEGATION of the current evaluated
2158: * conditional into the stack of rules.
2159: */
2160:
1.238 schwarze 2161: if (tok == ROFF_ie) {
1.223 schwarze 2162: if (r->rstackpos + 1 == r->rstacksz) {
2163: r->rstacksz += 16;
2164: r->rstack = mandoc_reallocarray(r->rstack,
2165: r->rstacksz, sizeof(int));
1.134 kristaps 2166: }
1.198 schwarze 2167: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2168: }
1.88 kristaps 2169:
2170: /* If the parent has false as its rule, then so do we. */
2171:
1.198 schwarze 2172: if (r->last->parent && !r->last->parent->rule)
2173: r->last->rule = 0;
1.88 kristaps 2174:
2175: /*
1.173 schwarze 2176: * Determine scope.
2177: * If there is nothing on the line after the conditional,
2178: * not even whitespace, use next-line scope.
1.88 kristaps 2179: */
1.74 kristaps 2180:
1.238 schwarze 2181: if (buf->buf[pos] == '\0') {
1.173 schwarze 2182: r->last->endspan = 2;
2183: goto out;
2184: }
2185:
1.238 schwarze 2186: while (buf->buf[pos] == ' ')
1.173 schwarze 2187: pos++;
2188:
2189: /* An opening brace requests multiline scope. */
1.75 kristaps 2190:
1.238 schwarze 2191: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2192: r->last->endspan = -1;
2193: pos += 2;
1.173 schwarze 2194: goto out;
1.207 schwarze 2195: }
1.74 kristaps 2196:
1.77 kristaps 2197: /*
1.173 schwarze 2198: * Anything else following the conditional causes
2199: * single-line scope. Warn if the scope contains
2200: * nothing but trailing whitespace.
1.77 kristaps 2201: */
2202:
1.238 schwarze 2203: if (buf->buf[pos] == '\0')
1.216 schwarze 2204: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
2205: ln, ppos, roffs[tok].name);
1.77 kristaps 2206:
1.173 schwarze 2207: r->last->endspan = 1;
1.74 kristaps 2208:
1.173 schwarze 2209: out:
1.75 kristaps 2210: *offs = pos;
2211: return(ROFF_RERUN);
1.83 schwarze 2212: }
2213:
2214: static enum rofferr
1.92 schwarze 2215: roff_ds(ROFF_ARGS)
2216: {
1.212 schwarze 2217: char *string;
2218: const char *name;
2219: size_t namesz;
1.96 kristaps 2220:
1.251 schwarze 2221: /* Ignore groff compatibility mode for now. */
2222:
2223: if (tok == ROFF_ds1)
2224: tok = ROFF_ds;
2225: else if (tok == ROFF_as1)
2226: tok = ROFF_as;
2227:
1.96 kristaps 2228: /*
1.212 schwarze 2229: * The first word is the name of the string.
2230: * If it is empty or terminated by an escape sequence,
2231: * abort the `ds' request without defining anything.
1.96 kristaps 2232: */
1.92 schwarze 2233:
1.238 schwarze 2234: name = string = buf->buf + pos;
2235: if (*name == '\0')
1.92 schwarze 2236: return(ROFF_IGN);
2237:
1.212 schwarze 2238: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2239: if (name[namesz] == '\\')
1.212 schwarze 2240: return(ROFF_IGN);
2241:
2242: /* Read past the initial double-quote, if any. */
1.238 schwarze 2243: if (*string == '"')
1.92 schwarze 2244: string++;
2245:
1.96 kristaps 2246: /* The rest is the value. */
1.212 schwarze 2247: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2248: ROFF_as == tok);
1.92 schwarze 2249: return(ROFF_IGN);
2250: }
2251:
1.204 schwarze 2252: /*
2253: * Parse a single operator, one or two characters long.
2254: * If the operator is recognized, return success and advance the
2255: * parse point, else return failure and let the parse point unchanged.
2256: */
2257: static int
2258: roff_getop(const char *v, int *pos, char *res)
2259: {
2260:
2261: *res = v[*pos];
2262:
2263: switch (*res) {
1.207 schwarze 2264: case '+':
1.204 schwarze 2265: /* FALLTHROUGH */
1.207 schwarze 2266: case '-':
1.204 schwarze 2267: /* FALLTHROUGH */
1.207 schwarze 2268: case '*':
1.204 schwarze 2269: /* FALLTHROUGH */
1.207 schwarze 2270: case '/':
1.204 schwarze 2271: /* FALLTHROUGH */
1.207 schwarze 2272: case '%':
1.204 schwarze 2273: /* FALLTHROUGH */
1.207 schwarze 2274: case '&':
1.204 schwarze 2275: /* FALLTHROUGH */
1.207 schwarze 2276: case ':':
1.204 schwarze 2277: break;
2278: case '<':
2279: switch (v[*pos + 1]) {
1.207 schwarze 2280: case '=':
1.204 schwarze 2281: *res = 'l';
2282: (*pos)++;
2283: break;
1.207 schwarze 2284: case '>':
1.204 schwarze 2285: *res = '!';
2286: (*pos)++;
2287: break;
1.207 schwarze 2288: case '?':
1.204 schwarze 2289: *res = 'i';
2290: (*pos)++;
2291: break;
2292: default:
2293: break;
2294: }
2295: break;
2296: case '>':
2297: switch (v[*pos + 1]) {
1.207 schwarze 2298: case '=':
1.204 schwarze 2299: *res = 'g';
2300: (*pos)++;
2301: break;
1.207 schwarze 2302: case '?':
1.204 schwarze 2303: *res = 'a';
2304: (*pos)++;
2305: break;
2306: default:
2307: break;
2308: }
2309: break;
2310: case '=':
2311: if ('=' == v[*pos + 1])
2312: (*pos)++;
2313: break;
2314: default:
2315: return(0);
2316: }
2317: (*pos)++;
2318:
2319: return(*res);
2320: }
2321:
2322: /*
2323: * Evaluate either a parenthesized numeric expression
2324: * or a single signed integer number.
2325: */
2326: static int
1.235 schwarze 2327: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2328: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2329: {
2330:
2331: if ('(' != v[*pos])
1.261 schwarze 2332: return(roff_getnum(v, pos, res, flags));
1.204 schwarze 2333:
2334: (*pos)++;
1.261 schwarze 2335: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.204 schwarze 2336: return(0);
2337:
1.206 schwarze 2338: /*
2339: * Omission of the closing parenthesis
2340: * is an error in validation mode,
2341: * but ignored in evaluation mode.
2342: */
2343:
1.204 schwarze 2344: if (')' == v[*pos])
2345: (*pos)++;
1.206 schwarze 2346: else if (NULL == res)
2347: return(0);
1.204 schwarze 2348:
2349: return(1);
2350: }
2351:
2352: /*
2353: * Evaluate a complete numeric expression.
2354: * Proceed left to right, there is no concept of precedence.
2355: */
2356: static int
1.235 schwarze 2357: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2358: int *pos, int *res, int flags)
1.204 schwarze 2359: {
2360: int mypos, operand2;
2361: char operator;
2362:
2363: if (NULL == pos) {
2364: mypos = 0;
2365: pos = &mypos;
2366: }
2367:
1.261 schwarze 2368: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2369: while (isspace((unsigned char)v[*pos]))
2370: (*pos)++;
2371:
1.261 schwarze 2372: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.204 schwarze 2373: return(0);
2374:
2375: while (1) {
1.261 schwarze 2376: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2377: while (isspace((unsigned char)v[*pos]))
2378: (*pos)++;
2379:
2380: if ( ! roff_getop(v, pos, &operator))
2381: break;
2382:
1.261 schwarze 2383: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2384: while (isspace((unsigned char)v[*pos]))
2385: (*pos)++;
2386:
1.261 schwarze 2387: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.204 schwarze 2388: return(0);
2389:
1.261 schwarze 2390: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2391: while (isspace((unsigned char)v[*pos]))
2392: (*pos)++;
1.206 schwarze 2393:
2394: if (NULL == res)
2395: continue;
1.204 schwarze 2396:
2397: switch (operator) {
1.207 schwarze 2398: case '+':
1.204 schwarze 2399: *res += operand2;
2400: break;
1.207 schwarze 2401: case '-':
1.204 schwarze 2402: *res -= operand2;
2403: break;
1.207 schwarze 2404: case '*':
1.204 schwarze 2405: *res *= operand2;
2406: break;
1.207 schwarze 2407: case '/':
1.244 schwarze 2408: if (operand2 == 0) {
1.235 schwarze 2409: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2410: r->parse, ln, *pos, v);
2411: *res = 0;
2412: break;
2413: }
1.204 schwarze 2414: *res /= operand2;
2415: break;
1.207 schwarze 2416: case '%':
1.244 schwarze 2417: if (operand2 == 0) {
2418: mandoc_msg(MANDOCERR_DIVZERO,
2419: r->parse, ln, *pos, v);
2420: *res = 0;
2421: break;
2422: }
1.204 schwarze 2423: *res %= operand2;
2424: break;
1.207 schwarze 2425: case '<':
1.204 schwarze 2426: *res = *res < operand2;
2427: break;
1.207 schwarze 2428: case '>':
1.204 schwarze 2429: *res = *res > operand2;
2430: break;
1.207 schwarze 2431: case 'l':
1.204 schwarze 2432: *res = *res <= operand2;
2433: break;
1.207 schwarze 2434: case 'g':
1.204 schwarze 2435: *res = *res >= operand2;
2436: break;
1.207 schwarze 2437: case '=':
1.204 schwarze 2438: *res = *res == operand2;
2439: break;
1.207 schwarze 2440: case '!':
1.204 schwarze 2441: *res = *res != operand2;
2442: break;
1.207 schwarze 2443: case '&':
1.204 schwarze 2444: *res = *res && operand2;
2445: break;
1.207 schwarze 2446: case ':':
1.204 schwarze 2447: *res = *res || operand2;
2448: break;
1.207 schwarze 2449: case 'i':
1.204 schwarze 2450: if (operand2 < *res)
2451: *res = operand2;
2452: break;
1.207 schwarze 2453: case 'a':
1.204 schwarze 2454: if (operand2 > *res)
2455: *res = operand2;
2456: break;
2457: default:
2458: abort();
2459: }
2460: }
2461: return(1);
2462: }
2463:
1.266 schwarze 2464: /* --- register management ------------------------------------------------ */
2465:
1.180 schwarze 2466: void
1.187 schwarze 2467: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2468: {
1.180 schwarze 2469: struct roffreg *reg;
2470:
2471: /* Search for an existing register with the same name. */
2472: reg = r->regtab;
2473:
2474: while (reg && strcmp(name, reg->key.p))
2475: reg = reg->next;
1.147 kristaps 2476:
1.180 schwarze 2477: if (NULL == reg) {
2478: /* Create a new register. */
2479: reg = mandoc_malloc(sizeof(struct roffreg));
2480: reg->key.p = mandoc_strdup(name);
2481: reg->key.sz = strlen(name);
1.187 schwarze 2482: reg->val = 0;
1.180 schwarze 2483: reg->next = r->regtab;
2484: r->regtab = reg;
2485: }
2486:
1.187 schwarze 2487: if ('+' == sign)
2488: reg->val += val;
2489: else if ('-' == sign)
2490: reg->val -= val;
2491: else
2492: reg->val = val;
1.147 kristaps 2493: }
2494:
1.192 schwarze 2495: /*
2496: * Handle some predefined read-only number registers.
2497: * For now, return -1 if the requested register is not predefined;
2498: * in case a predefined read-only register having the value -1
2499: * were to turn up, another special value would have to be chosen.
2500: */
2501: static int
2502: roff_getregro(const char *name)
2503: {
2504:
2505: switch (*name) {
1.207 schwarze 2506: case 'A': /* ASCII approximation mode is always off. */
1.192 schwarze 2507: return(0);
1.207 schwarze 2508: case 'g': /* Groff compatibility mode is always on. */
1.192 schwarze 2509: return(1);
1.207 schwarze 2510: case 'H': /* Fixed horizontal resolution. */
1.192 schwarze 2511: return (24);
1.207 schwarze 2512: case 'j': /* Always adjust left margin only. */
1.192 schwarze 2513: return(0);
1.207 schwarze 2514: case 'T': /* Some output device is always defined. */
1.192 schwarze 2515: return(1);
1.207 schwarze 2516: case 'V': /* Fixed vertical resolution. */
1.192 schwarze 2517: return (40);
2518: default:
2519: return (-1);
2520: }
2521: }
2522:
1.181 schwarze 2523: int
1.180 schwarze 2524: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2525: {
1.180 schwarze 2526: struct roffreg *reg;
1.192 schwarze 2527: int val;
2528:
2529: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
2530: val = roff_getregro(name + 1);
2531: if (-1 != val)
2532: return (val);
2533: }
1.180 schwarze 2534:
2535: for (reg = r->regtab; reg; reg = reg->next)
2536: if (0 == strcmp(name, reg->key.p))
1.181 schwarze 2537: return(reg->val);
2538:
2539: return(0);
2540: }
2541:
2542: static int
2543: roff_getregn(const struct roff *r, const char *name, size_t len)
2544: {
2545: struct roffreg *reg;
1.192 schwarze 2546: int val;
2547:
2548: if ('.' == name[0] && 2 == len) {
2549: val = roff_getregro(name + 1);
2550: if (-1 != val)
2551: return (val);
2552: }
1.181 schwarze 2553:
2554: for (reg = r->regtab; reg; reg = reg->next)
2555: if (len == reg->key.sz &&
2556: 0 == strncmp(name, reg->key.p, len))
2557: return(reg->val);
1.147 kristaps 2558:
1.180 schwarze 2559: return(0);
1.147 kristaps 2560: }
2561:
1.180 schwarze 2562: static void
2563: roff_freereg(struct roffreg *reg)
1.147 kristaps 2564: {
1.180 schwarze 2565: struct roffreg *old_reg;
1.147 kristaps 2566:
1.180 schwarze 2567: while (NULL != reg) {
2568: free(reg->key.p);
2569: old_reg = reg;
2570: reg = reg->next;
2571: free(old_reg);
2572: }
1.147 kristaps 2573: }
1.92 schwarze 2574:
2575: static enum rofferr
1.89 kristaps 2576: roff_nr(ROFF_ARGS)
1.83 schwarze 2577: {
1.212 schwarze 2578: char *key, *val;
2579: size_t keysz;
1.138 kristaps 2580: int iv;
1.187 schwarze 2581: char sign;
1.89 kristaps 2582:
1.238 schwarze 2583: key = val = buf->buf + pos;
2584: if (*key == '\0')
1.212 schwarze 2585: return(ROFF_IGN);
2586:
2587: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2588: if (key[keysz] == '\\')
1.212 schwarze 2589: return(ROFF_IGN);
2590: key[keysz] = '\0';
1.89 kristaps 2591:
1.187 schwarze 2592: sign = *val;
1.238 schwarze 2593: if (sign == '+' || sign == '-')
1.187 schwarze 2594: val++;
2595:
1.261 schwarze 2596: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204 schwarze 2597: roff_setreg(r, key, iv, sign);
1.109 kristaps 2598:
1.203 schwarze 2599: return(ROFF_IGN);
2600: }
2601:
2602: static enum rofferr
2603: roff_rr(ROFF_ARGS)
2604: {
2605: struct roffreg *reg, **prev;
1.212 schwarze 2606: char *name, *cp;
2607: size_t namesz;
1.203 schwarze 2608:
1.238 schwarze 2609: name = cp = buf->buf + pos;
2610: if (*name == '\0')
1.212 schwarze 2611: return(ROFF_IGN);
2612: namesz = roff_getname(r, &cp, ln, pos);
2613: name[namesz] = '\0';
1.203 schwarze 2614:
2615: prev = &r->regtab;
2616: while (1) {
2617: reg = *prev;
1.238 schwarze 2618: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2619: break;
2620: prev = ®->next;
2621: }
1.238 schwarze 2622: if (reg != NULL) {
1.203 schwarze 2623: *prev = reg->next;
2624: free(reg->key.p);
2625: free(reg);
2626: }
1.122 schwarze 2627: return(ROFF_IGN);
2628: }
2629:
1.266 schwarze 2630: /* --- handler functions for roff requests -------------------------------- */
2631:
1.122 schwarze 2632: static enum rofferr
2633: roff_rm(ROFF_ARGS)
2634: {
2635: const char *name;
2636: char *cp;
1.212 schwarze 2637: size_t namesz;
1.122 schwarze 2638:
1.238 schwarze 2639: cp = buf->buf + pos;
2640: while (*cp != '\0') {
1.212 schwarze 2641: name = cp;
1.238 schwarze 2642: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2643: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238 schwarze 2644: if (name[namesz] == '\\')
1.212 schwarze 2645: break;
1.122 schwarze 2646: }
1.178 schwarze 2647: return(ROFF_IGN);
2648: }
2649:
2650: static enum rofferr
2651: roff_it(ROFF_ARGS)
2652: {
2653: int iv;
2654:
2655: /* Parse the number of lines. */
1.261 schwarze 2656:
2657: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2658: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2659: ln, ppos, buf->buf + 1);
1.178 schwarze 2660: return(ROFF_IGN);
2661: }
2662:
1.262 schwarze 2663: while (isspace((unsigned char)buf->buf[pos]))
2664: pos++;
2665:
2666: /*
2667: * Arm the input line trap.
2668: * Special-casing "an-trap" is an ugly workaround to cope
2669: * with DocBook stupidly fiddling with man(7) internals.
2670: */
1.261 schwarze 2671:
1.178 schwarze 2672: roffit_lines = iv;
1.262 schwarze 2673: roffit_macro = mandoc_strdup(iv != 1 ||
2674: strcmp(buf->buf + pos, "an-trap") ?
2675: buf->buf + pos : "br");
1.109 kristaps 2676: return(ROFF_IGN);
1.175 schwarze 2677: }
2678:
2679: static enum rofferr
2680: roff_Dd(ROFF_ARGS)
2681: {
2682: const char *const *cp;
2683:
1.227 schwarze 2684: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175 schwarze 2685: for (cp = __mdoc_reserved; *cp; cp++)
2686: roff_setstr(r, *cp, NULL, 0);
2687:
1.227 schwarze 2688: if (r->format == 0)
2689: r->format = MPARSE_MDOC;
2690:
1.175 schwarze 2691: return(ROFF_CONT);
2692: }
2693:
2694: static enum rofferr
2695: roff_TH(ROFF_ARGS)
2696: {
2697: const char *const *cp;
2698:
1.227 schwarze 2699: if ((r->options & MPARSE_QUICK) == 0)
1.175 schwarze 2700: for (cp = __man_reserved; *cp; cp++)
2701: roff_setstr(r, *cp, NULL, 0);
2702:
1.227 schwarze 2703: if (r->format == 0)
2704: r->format = MPARSE_MAN;
2705:
1.175 schwarze 2706: return(ROFF_CONT);
1.109 kristaps 2707: }
2708:
2709: static enum rofferr
2710: roff_TE(ROFF_ARGS)
2711: {
2712:
2713: if (NULL == r->tbl)
1.221 schwarze 2714: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2715: ln, ppos, "TE");
1.258 schwarze 2716: else if ( ! tbl_end(&r->tbl)) {
2717: free(buf->buf);
2718: buf->buf = mandoc_strdup(".sp");
2719: buf->sz = 4;
2720: return(ROFF_REPARSE);
2721: }
1.112 kristaps 2722: return(ROFF_IGN);
2723: }
2724:
2725: static enum rofferr
2726: roff_T_(ROFF_ARGS)
2727: {
2728:
2729: if (NULL == r->tbl)
1.221 schwarze 2730: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2731: ln, ppos, "T&");
1.112 kristaps 2732: else
1.116 kristaps 2733: tbl_restart(ppos, ln, r->tbl);
1.112 kristaps 2734:
1.109 kristaps 2735: return(ROFF_IGN);
2736: }
2737:
1.230 schwarze 2738: /*
2739: * Handle in-line equation delimiters.
2740: */
2741: static enum rofferr
1.238 schwarze 2742: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2743: {
1.233 schwarze 2744: char *cp1, *cp2;
2745: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2746:
1.230 schwarze 2747: /*
2748: * Outside equations, look for an opening delimiter.
2749: * If we are inside an equation, we already know it is
2750: * in-line, or this function wouldn't have been called;
2751: * so look for a closing delimiter.
2752: */
2753:
1.238 schwarze 2754: cp1 = buf->buf + pos;
1.230 schwarze 2755: cp2 = strchr(cp1, r->eqn == NULL ?
2756: r->last_eqn->odelim : r->last_eqn->cdelim);
2757: if (cp2 == NULL)
2758: return(ROFF_CONT);
2759:
1.233 schwarze 2760: *cp2++ = '\0';
2761: bef_pr = bef_nl = aft_nl = aft_pr = "";
2762:
2763: /* Handle preceding text, protecting whitespace. */
2764:
1.238 schwarze 2765: if (*buf->buf != '\0') {
1.233 schwarze 2766: if (r->eqn == NULL)
2767: bef_pr = "\\&";
2768: bef_nl = "\n";
2769: }
2770:
2771: /*
2772: * Prepare replacing the delimiter with an equation macro
2773: * and drop leading white space from the equation.
2774: */
1.230 schwarze 2775:
1.233 schwarze 2776: if (r->eqn == NULL) {
2777: while (*cp2 == ' ')
2778: cp2++;
2779: mac = ".EQ";
2780: } else
2781: mac = ".EN";
2782:
2783: /* Handle following text, protecting whitespace. */
2784:
2785: if (*cp2 != '\0') {
2786: aft_nl = "\n";
2787: if (r->eqn != NULL)
2788: aft_pr = "\\&";
2789: }
2790:
2791: /* Do the actual replacement. */
2792:
1.238 schwarze 2793: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2794: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2795: free(buf->buf);
2796: buf->buf = cp1;
1.230 schwarze 2797:
2798: /* Toggle the in-line state of the eqn subsystem. */
2799:
2800: r->eqn_inline = r->eqn == NULL;
2801: return(ROFF_REPARSE);
1.151 kristaps 2802: }
2803:
1.235 schwarze 2804: static enum rofferr
2805: roff_EQ(ROFF_ARGS)
1.125 kristaps 2806: {
1.151 kristaps 2807: struct eqn_node *e;
1.125 kristaps 2808:
1.238 schwarze 2809: assert(r->eqn == NULL);
1.235 schwarze 2810: e = eqn_alloc(ppos, ln, r->parse);
1.125 kristaps 2811:
1.230 schwarze 2812: if (r->last_eqn) {
1.125 kristaps 2813: r->last_eqn->next = e;
1.230 schwarze 2814: e->delim = r->last_eqn->delim;
2815: e->odelim = r->last_eqn->odelim;
2816: e->cdelim = r->last_eqn->cdelim;
2817: } else
1.125 kristaps 2818: r->first_eqn = r->last_eqn = e;
2819:
2820: r->eqn = r->last_eqn = e;
1.151 kristaps 2821:
1.238 schwarze 2822: if (buf->buf[pos] != '\0')
1.235 schwarze 2823: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2824: ".EQ %s", buf->buf + pos);
1.151 kristaps 2825:
1.125 kristaps 2826: return(ROFF_IGN);
2827: }
2828:
2829: static enum rofferr
2830: roff_EN(ROFF_ARGS)
2831: {
2832:
1.221 schwarze 2833: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.125 kristaps 2834: return(ROFF_IGN);
2835: }
2836:
2837: static enum rofferr
1.109 kristaps 2838: roff_TS(ROFF_ARGS)
2839: {
1.176 schwarze 2840: struct tbl_node *tbl;
1.89 kristaps 2841:
1.115 kristaps 2842: if (r->tbl) {
1.221 schwarze 2843: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2844: ln, ppos, "TS breaks TS");
1.151 kristaps 2845: tbl_end(&r->tbl);
1.115 kristaps 2846: }
1.83 schwarze 2847:
1.176 schwarze 2848: tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2849:
2850: if (r->last_tbl)
1.176 schwarze 2851: r->last_tbl->next = tbl;
1.113 kristaps 2852: else
1.176 schwarze 2853: r->first_tbl = r->last_tbl = tbl;
1.113 kristaps 2854:
1.176 schwarze 2855: r->tbl = r->last_tbl = tbl;
1.83 schwarze 2856: return(ROFF_IGN);
1.251 schwarze 2857: }
2858:
2859: static enum rofferr
2860: roff_brp(ROFF_ARGS)
2861: {
2862:
2863: buf->buf[pos - 1] = '\0';
2864: return(ROFF_CONT);
1.92 schwarze 2865: }
2866:
1.105 kristaps 2867: static enum rofferr
1.174 kristaps 2868: roff_cc(ROFF_ARGS)
2869: {
2870: const char *p;
2871:
1.238 schwarze 2872: p = buf->buf + pos;
1.174 kristaps 2873:
1.238 schwarze 2874: if (*p == '\0' || (r->control = *p++) == '.')
1.174 kristaps 2875: r->control = 0;
2876:
1.238 schwarze 2877: if (*p != '\0')
1.260 schwarze 2878: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2879: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 2880:
2881: return(ROFF_IGN);
2882: }
2883:
2884: static enum rofferr
1.164 kristaps 2885: roff_tr(ROFF_ARGS)
2886: {
2887: const char *p, *first, *second;
2888: size_t fsz, ssz;
2889: enum mandoc_esc esc;
2890:
1.238 schwarze 2891: p = buf->buf + pos;
1.164 kristaps 2892:
1.238 schwarze 2893: if (*p == '\0') {
1.260 schwarze 2894: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.164 kristaps 2895: return(ROFF_IGN);
2896: }
2897:
1.238 schwarze 2898: while (*p != '\0') {
1.164 kristaps 2899: fsz = ssz = 1;
2900:
2901: first = p++;
1.238 schwarze 2902: if (*first == '\\') {
1.164 kristaps 2903: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2904: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2905: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2906: ln, (int)(p - buf->buf), first);
1.164 kristaps 2907: return(ROFF_IGN);
2908: }
2909: fsz = (size_t)(p - first);
2910: }
2911:
2912: second = p++;
1.238 schwarze 2913: if (*second == '\\') {
1.164 kristaps 2914: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2915: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2916: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2917: ln, (int)(p - buf->buf), second);
1.164 kristaps 2918: return(ROFF_IGN);
2919: }
2920: ssz = (size_t)(p - second);
1.238 schwarze 2921: } else if (*second == '\0') {
1.260 schwarze 2922: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2923: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 2924: second = " ";
1.165 kristaps 2925: p--;
1.164 kristaps 2926: }
2927:
1.167 kristaps 2928: if (fsz > 1) {
1.207 schwarze 2929: roff_setstrn(&r->xmbtab, first, fsz,
2930: second, ssz, 0);
1.167 kristaps 2931: continue;
2932: }
2933:
1.238 schwarze 2934: if (r->xtab == NULL)
1.207 schwarze 2935: r->xtab = mandoc_calloc(128,
2936: sizeof(struct roffstr));
1.167 kristaps 2937:
2938: free(r->xtab[(int)*first].p);
2939: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2940: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 2941: }
2942:
2943: return(ROFF_IGN);
2944: }
2945:
2946: static enum rofferr
1.105 kristaps 2947: roff_so(ROFF_ARGS)
2948: {
1.249 schwarze 2949: char *name, *cp;
1.105 kristaps 2950:
1.238 schwarze 2951: name = buf->buf + pos;
1.224 schwarze 2952: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 2953:
2954: /*
2955: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2956: * opening anything that's not in our cwd or anything beneath
2957: * it. Thus, explicitly disallow traversing up the file-system
2958: * or using absolute paths.
2959: */
2960:
1.238 schwarze 2961: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 2962: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2963: ".so %s", name);
1.249 schwarze 2964: buf->sz = mandoc_asprintf(&cp,
2965: ".sp\nSee the file %s.\n.sp", name) + 1;
2966: free(buf->buf);
2967: buf->buf = cp;
2968: *offs = 0;
2969: return(ROFF_REPARSE);
1.105 kristaps 2970: }
2971:
2972: *offs = pos;
2973: return(ROFF_SO);
2974: }
1.92 schwarze 2975:
1.266 schwarze 2976: /* --- user defined strings and macros ------------------------------------ */
2977:
1.106 kristaps 2978: static enum rofferr
2979: roff_userdef(ROFF_ARGS)
1.99 kristaps 2980: {
1.263 schwarze 2981: const char *arg[9], *ap;
1.106 kristaps 2982: char *cp, *n1, *n2;
1.119 schwarze 2983: int i;
1.263 schwarze 2984: size_t asz, rsz;
1.106 kristaps 2985:
2986: /*
2987: * Collect pointers to macro argument strings
1.188 schwarze 2988: * and NUL-terminate them.
1.106 kristaps 2989: */
1.263 schwarze 2990:
1.238 schwarze 2991: cp = buf->buf + pos;
1.119 schwarze 2992: for (i = 0; i < 9; i++)
1.238 schwarze 2993: arg[i] = *cp == '\0' ? "" :
1.136 kristaps 2994: mandoc_getarg(r->parse, &cp, ln, &pos);
1.99 kristaps 2995:
1.106 kristaps 2996: /*
2997: * Expand macro arguments.
1.99 kristaps 2998: */
1.263 schwarze 2999:
3000: buf->sz = strlen(r->current_string) + 1;
3001: n1 = cp = mandoc_malloc(buf->sz);
3002: memcpy(n1, r->current_string, buf->sz);
3003: while (*cp != '\0') {
3004:
3005: /* Scan ahead for the next argument invocation. */
3006:
3007: if (*cp++ != '\\')
3008: continue;
3009: if (*cp++ != '$')
3010: continue;
3011: i = *cp - '1';
3012: if (0 > i || 8 < i)
1.106 kristaps 3013: continue;
1.263 schwarze 3014: cp -= 2;
3015:
3016: /*
3017: * Determine the size of the expanded argument,
3018: * taking escaping of quotes into account.
3019: */
3020:
3021: asz = 0;
3022: for (ap = arg[i]; *ap != '\0'; ap++) {
3023: asz++;
3024: if (*ap == '"')
3025: asz += 3;
3026: }
3027: if (asz != 3) {
3028:
3029: /*
3030: * Determine the size of the rest of the
3031: * unexpanded macro, including the NUL.
3032: */
3033:
3034: rsz = buf->sz - (cp - n1) - 3;
3035:
3036: /*
3037: * When shrinking, move before
3038: * releasing the storage.
3039: */
3040:
3041: if (asz < 3)
3042: memmove(cp + asz, cp + 3, rsz);
3043:
3044: /*
3045: * Resize the storage for the macro
3046: * and readjust the parse pointer.
3047: */
3048:
3049: buf->sz += asz - 3;
3050: n2 = mandoc_realloc(n1, buf->sz);
3051: cp = n2 + (cp - n1);
3052: n1 = n2;
3053:
3054: /*
3055: * When growing, make room
3056: * for the expanded argument.
3057: */
3058:
3059: if (asz > 3)
3060: memmove(cp + asz, cp + 3, rsz);
3061: }
3062:
3063: /* Copy the expanded argument, escaping quotes. */
3064:
3065: n2 = cp;
3066: for (ap = arg[i]; *ap != '\0'; ap++) {
3067: if (*ap == '"') {
3068: memcpy(n2, "\\(dq", 4);
3069: n2 += 4;
3070: } else
3071: *n2++ = *ap;
1.106 kristaps 3072: }
1.99 kristaps 3073: }
3074:
1.106 kristaps 3075: /*
3076: * Replace the macro invocation
3077: * by the expanded macro.
3078: */
1.263 schwarze 3079:
1.238 schwarze 3080: free(buf->buf);
3081: buf->buf = n1;
1.248 schwarze 3082: *offs = 0;
1.106 kristaps 3083:
1.238 schwarze 3084: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.106 kristaps 3085: ROFF_REPARSE : ROFF_APPEND);
1.99 kristaps 3086: }
1.121 schwarze 3087:
1.212 schwarze 3088: static size_t
1.121 schwarze 3089: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3090: {
3091: char *name, *cp;
1.212 schwarze 3092: size_t namesz;
1.121 schwarze 3093:
3094: name = *cpp;
3095: if ('\0' == *name)
1.212 schwarze 3096: return(0);
1.121 schwarze 3097:
1.212 schwarze 3098: /* Read until end of name and terminate it with NUL. */
3099: for (cp = name; 1; cp++) {
3100: if ('\0' == *cp || ' ' == *cp) {
3101: namesz = cp - name;
3102: break;
3103: }
1.121 schwarze 3104: if ('\\' != *cp)
3105: continue;
1.215 schwarze 3106: namesz = cp - name;
3107: if ('{' == cp[1] || '}' == cp[1])
3108: break;
1.121 schwarze 3109: cp++;
3110: if ('\\' == *cp)
3111: continue;
1.224 schwarze 3112: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3113: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3114: mandoc_escape((const char **)&cp, NULL, NULL);
3115: break;
1.121 schwarze 3116: }
3117:
3118: /* Read past spaces. */
3119: while (' ' == *cp)
3120: cp++;
3121:
3122: *cpp = cp;
1.212 schwarze 3123: return(namesz);
1.121 schwarze 3124: }
3125:
1.106 kristaps 3126: /*
3127: * Store *string into the user-defined string called *name.
3128: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3129: * append == 0: replace mode
3130: * append == 1: single-line append mode
3131: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3132: */
1.94 kristaps 3133: static void
1.106 kristaps 3134: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3135: int append)
1.92 schwarze 3136: {
1.164 kristaps 3137:
3138: roff_setstrn(&r->strtab, name, strlen(name), string,
1.207 schwarze 3139: string ? strlen(string) : 0, append);
1.164 kristaps 3140: }
3141:
3142: static void
1.166 kristaps 3143: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3144: const char *string, size_t stringsz, int append)
1.164 kristaps 3145: {
1.166 kristaps 3146: struct roffkv *n;
1.164 kristaps 3147: char *c;
3148: int i;
3149: size_t oldch, newch;
1.92 schwarze 3150:
1.106 kristaps 3151: /* Search for an existing string with the same name. */
1.164 kristaps 3152: n = *r;
3153:
1.211 schwarze 3154: while (n && (namesz != n->key.sz ||
3155: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3156: n = n->next;
1.94 kristaps 3157:
3158: if (NULL == n) {
1.106 kristaps 3159: /* Create a new string table entry. */
1.166 kristaps 3160: n = mandoc_malloc(sizeof(struct roffkv));
3161: n->key.p = mandoc_strndup(name, namesz);
3162: n->key.sz = namesz;
3163: n->val.p = NULL;
3164: n->val.sz = 0;
1.164 kristaps 3165: n->next = *r;
3166: *r = n;
1.193 schwarze 3167: } else if (0 == append) {
1.166 kristaps 3168: free(n->val.p);
3169: n->val.p = NULL;
3170: n->val.sz = 0;
1.106 kristaps 3171: }
3172:
3173: if (NULL == string)
3174: return;
3175:
3176: /*
3177: * One additional byte for the '\n' in multiline mode,
3178: * and one for the terminating '\0'.
3179: */
1.193 schwarze 3180: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3181:
1.166 kristaps 3182: if (NULL == n->val.p) {
3183: n->val.p = mandoc_malloc(newch);
3184: *n->val.p = '\0';
1.106 kristaps 3185: oldch = 0;
3186: } else {
1.166 kristaps 3187: oldch = n->val.sz;
3188: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3189: }
3190:
3191: /* Skip existing content in the destination buffer. */
1.166 kristaps 3192: c = n->val.p + (int)oldch;
1.106 kristaps 3193:
3194: /* Append new content to the destination buffer. */
1.164 kristaps 3195: i = 0;
3196: while (i < (int)stringsz) {
1.106 kristaps 3197: /*
3198: * Rudimentary roff copy mode:
3199: * Handle escaped backslashes.
3200: */
1.164 kristaps 3201: if ('\\' == string[i] && '\\' == string[i + 1])
3202: i++;
3203: *c++ = string[i++];
1.106 kristaps 3204: }
1.94 kristaps 3205:
1.106 kristaps 3206: /* Append terminating bytes. */
1.193 schwarze 3207: if (1 < append)
1.106 kristaps 3208: *c++ = '\n';
1.163 kristaps 3209:
1.106 kristaps 3210: *c = '\0';
1.166 kristaps 3211: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3212: }
3213:
1.94 kristaps 3214: static const char *
3215: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92 schwarze 3216: {
1.166 kristaps 3217: const struct roffkv *n;
1.191 schwarze 3218: int i;
1.92 schwarze 3219:
1.164 kristaps 3220: for (n = r->strtab; n; n = n->next)
1.207 schwarze 3221: if (0 == strncmp(name, n->key.p, len) &&
3222: '\0' == n->key.p[(int)len])
1.166 kristaps 3223: return(n->val.p);
1.191 schwarze 3224:
3225: for (i = 0; i < PREDEFS_MAX; i++)
3226: if (0 == strncmp(name, predefs[i].name, len) &&
3227: '\0' == predefs[i].name[(int)len])
3228: return(predefs[i].str);
1.94 kristaps 3229:
1.157 kristaps 3230: return(NULL);
1.92 schwarze 3231: }
3232:
1.94 kristaps 3233: static void
1.167 kristaps 3234: roff_freestr(struct roffkv *r)
1.92 schwarze 3235: {
1.166 kristaps 3236: struct roffkv *n, *nn;
1.92 schwarze 3237:
1.167 kristaps 3238: for (n = r; n; n = nn) {
1.166 kristaps 3239: free(n->key.p);
3240: free(n->val.p);
1.92 schwarze 3241: nn = n->next;
3242: free(n);
3243: }
1.114 kristaps 3244: }
1.266 schwarze 3245:
3246: /* --- accessors and utility functions ------------------------------------ */
1.114 kristaps 3247:
3248: const struct tbl_span *
3249: roff_span(const struct roff *r)
3250: {
1.207 schwarze 3251:
1.114 kristaps 3252: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.125 kristaps 3253: }
3254:
3255: const struct eqn *
3256: roff_eqn(const struct roff *r)
3257: {
1.207 schwarze 3258:
1.125 kristaps 3259: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.164 kristaps 3260: }
3261:
3262: /*
3263: * Duplicate an input string, making the appropriate character
3264: * conversations (as stipulated by `tr') along the way.
3265: * Returns a heap-allocated string with all the replacements made.
3266: */
3267: char *
3268: roff_strdup(const struct roff *r, const char *p)
3269: {
1.166 kristaps 3270: const struct roffkv *cp;
1.164 kristaps 3271: char *res;
3272: const char *pp;
3273: size_t ssz, sz;
3274: enum mandoc_esc esc;
3275:
1.167 kristaps 3276: if (NULL == r->xmbtab && NULL == r->xtab)
1.164 kristaps 3277: return(mandoc_strdup(p));
3278: else if ('\0' == *p)
3279: return(mandoc_strdup(""));
3280:
3281: /*
3282: * Step through each character looking for term matches
3283: * (remember that a `tr' can be invoked with an escape, which is
3284: * a glyph but the escape is multi-character).
3285: * We only do this if the character hash has been initialised
3286: * and the string is >0 length.
3287: */
3288:
3289: res = NULL;
3290: ssz = 0;
3291:
3292: while ('\0' != *p) {
1.167 kristaps 3293: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3294: sz = r->xtab[(int)*p].sz;
3295: res = mandoc_realloc(res, ssz + sz + 1);
3296: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3297: ssz += sz;
CVSweb