Annotation of src/usr.bin/mandoc/roff.c, Revision 1.138
1.138 ! schwarze 1: /* $OpenBSD: roff.c,v 1.137 2015/04/18 17:28:08 schwarze Exp $ */
1.1 schwarze 2: /*
1.137 schwarze 3: * Copyright (c) 2009-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.119 schwarze 4: * Copyright (c) 2010-2015 Ingo Schwarze <schwarze@openbsd.org>
1.1 schwarze 5: *
6: * Permission to use, copy, modify, and distribute this software for any
7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
9: *
1.16 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1 schwarze 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1 schwarze 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.
17: */
1.99 schwarze 18: #include <sys/types.h>
19:
1.1 schwarze 20: #include <assert.h>
1.3 schwarze 21: #include <ctype.h>
1.117 schwarze 22: #include <limits.h>
1.51 schwarze 23: #include <stdio.h>
1.1 schwarze 24: #include <stdlib.h>
25: #include <string.h>
26:
27: #include "mandoc.h"
1.74 schwarze 28: #include "mandoc_aux.h"
1.137 schwarze 29: #include "roff.h"
1.98 daniel 30: #include "libmandoc.h"
1.138 ! schwarze 31: #include "roff_int.h"
1.27 schwarze 32: #include "libroff.h"
1.138 ! schwarze 33: #include "libmdoc.h"
1.1 schwarze 34:
1.37 schwarze 35: /* Maximum number of nested if-else conditionals. */
1.2 schwarze 36: #define RSTACK_MAX 128
37:
1.43 schwarze 38: /* Maximum number of string expansions per line, to break infinite loops. */
39: #define EXPAND_LIMIT 1000
40:
1.138 ! schwarze 41: /* --- data types --------------------------------------------------------- */
! 42:
1.1 schwarze 43: enum rofft {
1.123 schwarze 44: ROFF_ab,
1.20 schwarze 45: ROFF_ad,
1.123 schwarze 46: ROFF_af,
47: ROFF_aln,
48: ROFF_als,
1.2 schwarze 49: ROFF_am,
1.123 schwarze 50: ROFF_am1,
1.2 schwarze 51: ROFF_ami,
1.123 schwarze 52: ROFF_ami1,
1.66 schwarze 53: ROFF_as,
1.123 schwarze 54: ROFF_as1,
55: ROFF_asciify,
56: ROFF_backtrace,
57: ROFF_bd,
58: ROFF_bleedat,
59: ROFF_blm,
60: ROFF_box,
61: ROFF_boxa,
62: ROFF_bp,
63: ROFF_BP,
64: /* MAN_br, MDOC_br */
65: ROFF_break,
66: ROFF_breakchar,
67: ROFF_brnl,
68: ROFF_brp,
69: ROFF_brpnl,
70: ROFF_c2,
1.48 schwarze 71: ROFF_cc,
1.67 schwarze 72: ROFF_ce,
1.123 schwarze 73: ROFF_cf,
74: ROFF_cflags,
75: ROFF_ch,
76: ROFF_char,
77: ROFF_chop,
78: ROFF_class,
79: ROFF_close,
80: ROFF_CL,
81: ROFF_color,
82: ROFF_composite,
83: ROFF_continue,
84: ROFF_cp,
85: ROFF_cropat,
86: ROFF_cs,
87: ROFF_cu,
88: ROFF_da,
89: ROFF_dch,
90: ROFF_Dd,
1.1 schwarze 91: ROFF_de,
1.123 schwarze 92: ROFF_de1,
93: ROFF_defcolor,
1.1 schwarze 94: ROFF_dei,
1.123 schwarze 95: ROFF_dei1,
96: ROFF_device,
97: ROFF_devicem,
98: ROFF_di,
99: ROFF_do,
1.2 schwarze 100: ROFF_ds,
1.123 schwarze 101: ROFF_ds1,
102: ROFF_dwh,
103: ROFF_dt,
104: ROFF_ec,
105: ROFF_ecr,
106: ROFF_ecs,
1.2 schwarze 107: ROFF_el,
1.123 schwarze 108: ROFF_em,
109: ROFF_EN,
110: ROFF_eo,
111: ROFF_EP,
112: ROFF_EQ,
113: ROFF_errprint,
114: ROFF_ev,
115: ROFF_evc,
116: ROFF_ex,
117: ROFF_fallback,
1.58 schwarze 118: ROFF_fam,
1.123 schwarze 119: ROFF_fc,
120: ROFF_fchar,
121: ROFF_fcolor,
122: ROFF_fdeferlig,
123: ROFF_feature,
124: /* MAN_fi; ignored in mdoc(7) */
125: ROFF_fkern,
126: ROFF_fl,
127: ROFF_flig,
128: ROFF_fp,
129: ROFF_fps,
130: ROFF_fschar,
131: ROFF_fspacewidth,
132: ROFF_fspecial,
133: /* MAN_ft; ignored in mdoc(7) */
134: ROFF_ftr,
135: ROFF_fzoom,
136: ROFF_gcolor,
137: ROFF_hc,
138: ROFF_hcode,
139: ROFF_hidechar,
140: ROFF_hla,
141: ROFF_hlm,
142: ROFF_hpf,
143: ROFF_hpfa,
144: ROFF_hpfcode,
1.59 schwarze 145: ROFF_hw,
1.20 schwarze 146: ROFF_hy,
1.123 schwarze 147: ROFF_hylang,
148: ROFF_hylen,
149: ROFF_hym,
150: ROFF_hypp,
151: ROFF_hys,
1.2 schwarze 152: ROFF_ie,
153: ROFF_if,
1.1 schwarze 154: ROFF_ig,
1.123 schwarze 155: /* MAN_in; ignored in mdoc(7) */
156: ROFF_index,
1.30 schwarze 157: ROFF_it,
1.123 schwarze 158: ROFF_itc,
159: ROFF_IX,
160: ROFF_kern,
161: ROFF_kernafter,
162: ROFF_kernbefore,
163: ROFF_kernpair,
164: ROFF_lc,
165: ROFF_lc_ctype,
166: ROFF_lds,
167: ROFF_length,
168: ROFF_letadj,
169: ROFF_lf,
170: ROFF_lg,
171: ROFF_lhang,
172: ROFF_linetabs,
173: /* MAN_ll, MDOC_ll */
174: ROFF_lnr,
175: ROFF_lnrf,
176: ROFF_lpfx,
177: ROFF_ls,
178: ROFF_lsm,
179: ROFF_lt,
180: ROFF_mc,
181: ROFF_mediasize,
182: ROFF_minss,
183: ROFF_mk,
184: ROFF_mso,
1.128 schwarze 185: ROFF_na,
1.20 schwarze 186: ROFF_ne,
1.123 schwarze 187: /* MAN_nf; ignored in mdoc(7) */
1.20 schwarze 188: ROFF_nh,
1.123 schwarze 189: ROFF_nhychar,
190: ROFF_nm,
191: ROFF_nn,
192: ROFF_nop,
1.14 schwarze 193: ROFF_nr,
1.123 schwarze 194: ROFF_nrf,
195: ROFF_nroff,
1.31 schwarze 196: ROFF_ns,
1.123 schwarze 197: ROFF_nx,
198: ROFF_open,
199: ROFF_opena,
200: ROFF_os,
201: ROFF_output,
202: ROFF_padj,
203: ROFF_papersize,
204: ROFF_pc,
205: ROFF_pev,
206: ROFF_pi,
207: ROFF_PI,
1.101 schwarze 208: ROFF_pl,
1.123 schwarze 209: ROFF_pm,
210: ROFF_pn,
211: ROFF_pnr,
212: ROFF_po,
1.31 schwarze 213: ROFF_ps,
1.123 schwarze 214: ROFF_psbb,
215: ROFF_pshape,
216: ROFF_pso,
217: ROFF_ptr,
218: ROFF_pvs,
219: ROFF_rchar,
220: ROFF_rd,
221: ROFF_recursionlimit,
222: ROFF_return,
223: ROFF_rfschar,
224: ROFF_rhang,
225: ROFF_rj,
1.2 schwarze 226: ROFF_rm,
1.123 schwarze 227: ROFF_rn,
228: ROFF_rnn,
1.76 schwarze 229: ROFF_rr,
1.123 schwarze 230: ROFF_rs,
231: ROFF_rt,
232: ROFF_schar,
233: ROFF_sentchar,
234: ROFF_shc,
235: ROFF_shift,
236: ROFF_sizes,
1.14 schwarze 237: ROFF_so,
1.123 schwarze 238: /* MAN_sp, MDOC_sp */
239: ROFF_spacewidth,
240: ROFF_special,
241: ROFF_spreadwarn,
242: ROFF_ss,
243: ROFF_sty,
244: ROFF_substring,
245: ROFF_sv,
246: ROFF_sy,
247: ROFF_T_,
1.31 schwarze 248: ROFF_ta,
1.123 schwarze 249: ROFF_tc,
250: ROFF_TE,
251: ROFF_TH,
252: ROFF_ti,
253: ROFF_tkf,
254: ROFF_tl,
255: ROFF_tm,
256: ROFF_tm1,
257: ROFF_tmc,
1.2 schwarze 258: ROFF_tr,
1.123 schwarze 259: ROFF_track,
260: ROFF_transchar,
261: ROFF_trf,
262: ROFF_trimat,
263: ROFF_trin,
264: ROFF_trnt,
265: ROFF_troff,
1.27 schwarze 266: ROFF_TS,
1.123 schwarze 267: ROFF_uf,
268: ROFF_ul,
269: ROFF_unformat,
270: ROFF_unwatch,
271: ROFF_unwatchn,
272: ROFF_vpt,
273: ROFF_vs,
274: ROFF_warn,
275: ROFF_warnscale,
276: ROFF_watch,
277: ROFF_watchlength,
278: ROFF_watchn,
279: ROFF_wh,
280: ROFF_while,
281: ROFF_write,
282: ROFF_writec,
283: ROFF_writem,
284: ROFF_xflag,
1.2 schwarze 285: ROFF_cblock,
1.16 schwarze 286: ROFF_USERDEF,
1.1 schwarze 287: ROFF_MAX
288: };
289:
1.41 schwarze 290: /*
1.42 schwarze 291: * An incredibly-simple string buffer.
292: */
1.8 schwarze 293: struct roffstr {
1.42 schwarze 294: char *p; /* nil-terminated buffer */
295: size_t sz; /* saved strlen(p) */
296: };
297:
298: /*
299: * A key-value roffstr pair as part of a singly-linked list.
300: */
301: struct roffkv {
302: struct roffstr key;
303: struct roffstr val;
304: struct roffkv *next; /* next in list */
1.8 schwarze 305: };
306:
1.52 schwarze 307: /*
308: * A single number register as part of a singly-linked list.
309: */
310: struct roffreg {
311: struct roffstr key;
1.53 schwarze 312: int val;
1.52 schwarze 313: struct roffreg *next;
314: };
315:
1.1 schwarze 316: struct roff {
1.35 schwarze 317: struct mparse *parse; /* parse point */
1.109 schwarze 318: const struct mchars *mchars; /* character table */
1.1 schwarze 319: struct roffnode *last; /* leaf of stack */
1.96 schwarze 320: int *rstack; /* stack of inverted `ie' values */
1.52 schwarze 321: struct roffreg *regtab; /* number registers */
1.42 schwarze 322: struct roffkv *strtab; /* user-defined strings & macros */
323: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
324: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.16 schwarze 325: const char *current_string; /* value of last called user macro */
1.27 schwarze 326: struct tbl_node *first_tbl; /* first table parsed */
327: struct tbl_node *last_tbl; /* last table parsed */
328: struct tbl_node *tbl; /* current table being parsed */
1.32 schwarze 329: struct eqn_node *last_eqn; /* last equation parsed */
330: struct eqn_node *first_eqn; /* first equation parsed */
331: struct eqn_node *eqn; /* current equation being parsed */
1.102 schwarze 332: int eqn_inline; /* current equation is inline */
1.96 schwarze 333: int options; /* parse options */
334: int rstacksz; /* current size limit of rstack */
335: int rstackpos; /* position in rstack */
1.99 schwarze 336: int format; /* current file in mdoc or man format */
1.96 schwarze 337: char control; /* control character */
1.1 schwarze 338: };
339:
340: struct roffnode {
341: enum rofft tok; /* type of node */
342: struct roffnode *parent; /* up one in stack */
343: int line; /* parse line */
344: int col; /* parse col */
1.16 schwarze 345: char *name; /* node name, e.g. macro name */
1.2 schwarze 346: char *end; /* end-rules: custom token */
347: int endspan; /* end-rules: next-line or infty */
1.71 schwarze 348: int rule; /* current evaluation rule */
1.1 schwarze 349: };
350:
351: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
352: enum rofft tok, /* tok of macro */ \
1.110 schwarze 353: struct buf *buf, /* input buffer */ \
1.1 schwarze 354: int ln, /* parse line */ \
1.2 schwarze 355: int ppos, /* original pos in buffer */ \
356: int pos, /* current pos in buffer */ \
357: int *offs /* reset offset of buffer data */
1.1 schwarze 358:
359: typedef enum rofferr (*roffproc)(ROFF_ARGS);
360:
361: struct roffmac {
362: const char *name; /* macro name */
1.2 schwarze 363: roffproc proc; /* process new macro */
364: roffproc text; /* process as child text of macro */
365: roffproc sub; /* process as child of macro */
366: int flags;
367: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.3 schwarze 368: struct roffmac *next;
1.1 schwarze 369: };
370:
1.37 schwarze 371: struct predef {
372: const char *name; /* predefined input name */
373: const char *str; /* replacement symbol */
374: };
375:
376: #define PREDEF(__name, __str) \
377: { (__name), (__str) },
378:
1.138 ! schwarze 379: /* --- function prototypes ------------------------------------------------ */
! 380:
1.42 schwarze 381: static enum rofft roffhash_find(const char *, size_t);
382: static void roffhash_init(void);
383: static void roffnode_cleanscope(struct roff *);
384: static void roffnode_pop(struct roff *);
385: static void roffnode_push(struct roff *, enum rofft,
386: const char *, int, int);
1.2 schwarze 387: static enum rofferr roff_block(ROFF_ARGS);
388: static enum rofferr roff_block_text(ROFF_ARGS);
389: static enum rofferr roff_block_sub(ROFF_ARGS);
1.123 schwarze 390: static enum rofferr roff_brp(ROFF_ARGS);
1.2 schwarze 391: static enum rofferr roff_cblock(ROFF_ARGS);
1.48 schwarze 392: static enum rofferr roff_cc(ROFF_ARGS);
1.68 schwarze 393: static void roff_ccond(struct roff *, int, int);
1.2 schwarze 394: static enum rofferr roff_cond(ROFF_ARGS);
395: static enum rofferr roff_cond_text(ROFF_ARGS);
396: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.7 schwarze 397: static enum rofferr roff_ds(ROFF_ARGS);
1.110 schwarze 398: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.106 schwarze 399: static int roff_evalcond(struct roff *r, int,
400: const char *, int *);
401: static int roff_evalnum(struct roff *, int,
402: const char *, int *, int *, int);
403: static int roff_evalpar(struct roff *, int,
1.133 schwarze 404: const char *, int *, int *, int);
1.71 schwarze 405: static int roff_evalstrcond(const char *, int *);
1.42 schwarze 406: static void roff_free1(struct roff *);
1.52 schwarze 407: static void roff_freereg(struct roffreg *);
1.42 schwarze 408: static void roff_freestr(struct roffkv *);
1.85 schwarze 409: static size_t roff_getname(struct roff *, char **, int, int);
1.133 schwarze 410: static int roff_getnum(const char *, int *, int *, int);
1.56 schwarze 411: static int roff_getop(const char *, int *, char *);
1.53 schwarze 412: static int roff_getregn(const struct roff *,
413: const char *, size_t);
1.65 schwarze 414: static int roff_getregro(const char *name);
1.80 schwarze 415: static const char *roff_getstrn(const struct roff *,
1.8 schwarze 416: const char *, size_t);
1.123 schwarze 417: static enum rofferr roff_insec(ROFF_ARGS);
1.51 schwarze 418: static enum rofferr roff_it(ROFF_ARGS);
1.21 schwarze 419: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.137 schwarze 420: static void roff_man_alloc1(struct roff_man *);
421: static void roff_man_free1(struct roff_man *);
1.6 schwarze 422: static enum rofferr roff_nr(ROFF_ARGS);
1.87 schwarze 423: static enum rofft roff_parse(struct roff *, char *, int *,
424: int, int);
1.110 schwarze 425: static enum rofferr roff_parsetext(struct buf *, int, int *);
426: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.29 schwarze 427: static enum rofferr roff_rm(ROFF_ARGS);
1.76 schwarze 428: static enum rofferr roff_rr(ROFF_ARGS);
1.8 schwarze 429: static void roff_setstr(struct roff *,
1.16 schwarze 430: const char *, const char *, int);
1.80 schwarze 431: static void roff_setstrn(struct roffkv **, const char *,
1.42 schwarze 432: size_t, const char *, size_t, int);
1.14 schwarze 433: static enum rofferr roff_so(ROFF_ARGS);
1.42 schwarze 434: static enum rofferr roff_tr(ROFF_ARGS);
1.47 schwarze 435: static enum rofferr roff_Dd(ROFF_ARGS);
436: static enum rofferr roff_TH(ROFF_ARGS);
1.27 schwarze 437: static enum rofferr roff_TE(ROFF_ARGS);
438: static enum rofferr roff_TS(ROFF_ARGS);
1.32 schwarze 439: static enum rofferr roff_EQ(ROFF_ARGS);
440: static enum rofferr roff_EN(ROFF_ARGS);
1.27 schwarze 441: static enum rofferr roff_T_(ROFF_ARGS);
1.123 schwarze 442: static enum rofferr roff_unsupp(ROFF_ARGS);
1.16 schwarze 443: static enum rofferr roff_userdef(ROFF_ARGS);
1.1 schwarze 444:
1.138 ! schwarze 445: /* --- constant data ------------------------------------------------------ */
! 446:
1.42 schwarze 447: /* See roffhash_find() */
1.3 schwarze 448:
449: #define ASCII_HI 126
450: #define ASCII_LO 33
451: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
452:
1.133 schwarze 453: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
454: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
455:
1.3 schwarze 456: static struct roffmac *hash[HASHWIDTH];
457:
458: static struct roffmac roffs[ROFF_MAX] = {
1.123 schwarze 459: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.21 schwarze 460: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 461: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
462: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
463: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 464: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 465: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.3 schwarze 466: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 467: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.66 schwarze 468: { "as", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 469: { "as1", roff_ds, NULL, NULL, 0, NULL },
470: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
471: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
472: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
473: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
474: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
475: { "box", roff_unsupp, NULL, NULL, 0, NULL },
476: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
477: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
478: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
479: { "break", roff_unsupp, NULL, NULL, 0, NULL },
480: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
481: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
482: { "brp", roff_brp, NULL, NULL, 0, NULL },
483: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
484: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.48 schwarze 485: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.67 schwarze 486: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 487: { "cf", roff_insec, NULL, NULL, 0, NULL },
488: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
489: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
490: { "char", roff_unsupp, NULL, NULL, 0, NULL },
491: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
492: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
493: { "close", roff_insec, NULL, NULL, 0, NULL },
494: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
495: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
496: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
497: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
498: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
499: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
500: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
501: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
502: { "da", roff_unsupp, NULL, NULL, 0, NULL },
503: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
504: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.3 schwarze 505: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 506: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
507: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 508: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 509: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
510: { "device", roff_unsupp, NULL, NULL, 0, NULL },
511: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
512: { "di", roff_unsupp, NULL, NULL, 0, NULL },
513: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.7 schwarze 514: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.123 schwarze 515: { "ds1", roff_ds, NULL, NULL, 0, NULL },
516: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
517: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
518: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
519: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
520: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.3 schwarze 521: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.123 schwarze 522: { "em", roff_unsupp, NULL, NULL, 0, NULL },
523: { "EN", roff_EN, NULL, NULL, 0, NULL },
524: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
525: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
526: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
527: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
528: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
529: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
530: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
531: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.58 schwarze 532: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 533: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
534: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
535: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
536: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
538: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
543: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
544: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
545: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
551: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
552: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
553: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
554: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
555: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
556: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.59 schwarze 557: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 558: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 559: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
560: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
561: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
562: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
563: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 564: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
565: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
566: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123 schwarze 567: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.51 schwarze 568: { "it", roff_it, NULL, NULL, 0, NULL },
1.123 schwarze 569: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
570: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
571: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
572: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
573: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
574: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
575: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
576: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
577: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
578: { "length", roff_unsupp, NULL, NULL, 0, NULL },
579: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
580: { "lf", roff_insec, NULL, NULL, 0, NULL },
581: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
582: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
583: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
584: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
585: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
586: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
587: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
588: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
589: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
590: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
591: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
592: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
593: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
594: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.128 schwarze 595: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.21 schwarze 596: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
597: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 598: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
599: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
600: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
601: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.14 schwarze 602: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.123 schwarze 603: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
604: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 605: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 606: { "nx", roff_insec, NULL, NULL, 0, NULL },
607: { "open", roff_insec, NULL, NULL, 0, NULL },
608: { "opena", roff_insec, NULL, NULL, 0, NULL },
609: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
610: { "output", roff_unsupp, NULL, NULL, 0, NULL },
611: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
612: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
615: { "pi", roff_insec, NULL, NULL, 0, NULL },
616: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.101 schwarze 617: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 618: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
619: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
620: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
621: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.31 schwarze 622: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.123 schwarze 623: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.127 schwarze 624: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 625: { "pso", roff_insec, NULL, NULL, 0, NULL },
626: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
627: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
628: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
629: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
630: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
631: { "return", roff_unsupp, NULL, NULL, 0, NULL },
632: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
633: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
634: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.29 schwarze 635: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.123 schwarze 636: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
637: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.76 schwarze 638: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.123 schwarze 639: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
640: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
641: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
642: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
643: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
644: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
645: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.14 schwarze 646: { "so", roff_so, NULL, NULL, 0, NULL },
1.123 schwarze 647: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
648: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
649: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
650: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
651: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
652: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
653: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
654: { "sy", roff_insec, NULL, NULL, 0, NULL },
655: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.127 schwarze 656: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 657: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
658: { "TE", roff_TE, NULL, NULL, 0, NULL },
659: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.127 schwarze 660: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.123 schwarze 661: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
662: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
663: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
664: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
665: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.42 schwarze 666: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.123 schwarze 667: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
668: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
669: { "trf", roff_insec, NULL, NULL, 0, NULL },
670: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
671: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
672: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
673: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.27 schwarze 674: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.123 schwarze 675: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
676: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
677: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
678: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
679: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
680: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
681: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
682: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
683: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
684: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
685: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
686: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
687: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
688: { "while", roff_unsupp, NULL, NULL, 0, NULL },
689: { "write", roff_insec, NULL, NULL, 0, NULL },
690: { "writec", roff_insec, NULL, NULL, 0, NULL },
691: { "writem", roff_insec, NULL, NULL, 0, NULL },
692: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.3 schwarze 693: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.16 schwarze 694: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1 schwarze 695: };
696:
1.73 schwarze 697: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47 schwarze 698: const char *const __mdoc_reserved[] = {
699: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
700: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
701: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
702: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73 schwarze 703: "Dt", "Dv", "Dx", "D1",
704: "Ec", "Ed", "Ef", "Ek", "El", "Em",
705: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47 schwarze 706: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73 schwarze 707: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
708: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47 schwarze 709: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73 schwarze 710: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
711: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
712: "Sc", "Sh", "Sm", "So", "Sq",
1.47 schwarze 713: "Ss", "St", "Sx", "Sy",
714: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73 schwarze 715: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47 schwarze 716: "%P", "%Q", "%R", "%T", "%U", "%V",
717: NULL
718: };
719:
1.73 schwarze 720: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47 schwarze 721: const char *const __man_reserved[] = {
1.73 schwarze 722: "AT", "B", "BI", "BR", "DT",
723: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
724: "LP", "OP", "P", "PD", "PP",
725: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
726: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47 schwarze 727: NULL
728: };
729:
1.37 schwarze 730: /* Array of injected predefined strings. */
731: #define PREDEFS_MAX 38
732: static const struct predef predefs[PREDEFS_MAX] = {
733: #include "predefs.in"
734: };
735:
1.42 schwarze 736: /* See roffhash_find() */
1.3 schwarze 737: #define ROFF_HASH(p) (p[0] - ASCII_LO)
738:
1.51 schwarze 739: static int roffit_lines; /* number of lines to delay */
740: static char *roffit_macro; /* nil-terminated macro line */
741:
1.80 schwarze 742:
1.138 ! schwarze 743: /* --- request table ------------------------------------------------------ */
! 744:
1.3 schwarze 745: static void
1.42 schwarze 746: roffhash_init(void)
1.3 schwarze 747: {
748: struct roffmac *n;
749: int buc, i;
750:
1.16 schwarze 751: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3 schwarze 752: assert(roffs[i].name[0] >= ASCII_LO);
753: assert(roffs[i].name[0] <= ASCII_HI);
754:
755: buc = ROFF_HASH(roffs[i].name);
756:
757: if (NULL != (n = hash[buc])) {
758: for ( ; n->next; n = n->next)
759: /* Do nothing. */ ;
760: n->next = &roffs[i];
761: } else
762: hash[buc] = &roffs[i];
763: }
764: }
765:
1.1 schwarze 766: /*
767: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
768: * the nil-terminated string name could be found.
769: */
770: static enum rofft
1.42 schwarze 771: roffhash_find(const char *p, size_t s)
1.1 schwarze 772: {
1.3 schwarze 773: int buc;
774: struct roffmac *n;
1.1 schwarze 775:
1.3 schwarze 776: /*
777: * libroff has an extremely simple hashtable, for the time
778: * being, which simply keys on the first character, which must
779: * be printable, then walks a chain. It works well enough until
780: * optimised.
781: */
782:
783: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
784: return(ROFF_MAX);
785:
786: buc = ROFF_HASH(p);
787:
788: if (NULL == (n = hash[buc]))
789: return(ROFF_MAX);
790: for ( ; n; n = n->next)
1.16 schwarze 791: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3 schwarze 792: return((enum rofft)(n - roffs));
1.1 schwarze 793:
794: return(ROFF_MAX);
795: }
796:
1.138 ! schwarze 797: /* --- stack of request blocks -------------------------------------------- */
! 798:
1.1 schwarze 799: /*
800: * Pop the current node off of the stack of roff instructions currently
801: * pending.
802: */
803: static void
804: roffnode_pop(struct roff *r)
805: {
806: struct roffnode *p;
807:
1.2 schwarze 808: assert(r->last);
1.80 schwarze 809: p = r->last;
1.2 schwarze 810:
811: r->last = r->last->parent;
1.16 schwarze 812: free(p->name);
813: free(p->end);
1.1 schwarze 814: free(p);
815: }
816:
817: /*
818: * Push a roff node onto the instruction stack. This must later be
819: * removed with roffnode_pop().
820: */
1.11 schwarze 821: static void
1.16 schwarze 822: roffnode_push(struct roff *r, enum rofft tok, const char *name,
823: int line, int col)
1.1 schwarze 824: {
825: struct roffnode *p;
826:
1.11 schwarze 827: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 828: p->tok = tok;
1.16 schwarze 829: if (name)
830: p->name = mandoc_strdup(name);
1.1 schwarze 831: p->parent = r->last;
832: p->line = line;
833: p->col = col;
1.71 schwarze 834: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 835:
836: r->last = p;
837: }
838:
1.138 ! schwarze 839: /* --- roff parser state data management ---------------------------------- */
! 840:
1.1 schwarze 841: static void
842: roff_free1(struct roff *r)
843: {
1.49 schwarze 844: struct tbl_node *tbl;
1.32 schwarze 845: struct eqn_node *e;
1.42 schwarze 846: int i;
1.27 schwarze 847:
1.49 schwarze 848: while (NULL != (tbl = r->first_tbl)) {
849: r->first_tbl = tbl->next;
850: tbl_free(tbl);
1.27 schwarze 851: }
852: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 853:
1.32 schwarze 854: while (NULL != (e = r->first_eqn)) {
855: r->first_eqn = e->next;
856: eqn_free(e);
857: }
858: r->first_eqn = r->last_eqn = r->eqn = NULL;
859:
1.1 schwarze 860: while (r->last)
861: roffnode_pop(r);
1.27 schwarze 862:
1.96 schwarze 863: free (r->rstack);
864: r->rstack = NULL;
865: r->rstacksz = 0;
866: r->rstackpos = -1;
867:
868: roff_freereg(r->regtab);
869: r->regtab = NULL;
870:
1.42 schwarze 871: roff_freestr(r->strtab);
872: roff_freestr(r->xmbtab);
873: r->strtab = r->xmbtab = NULL;
874:
875: if (r->xtab)
876: for (i = 0; i < 128; i++)
877: free(r->xtab[i].p);
878: free(r->xtab);
879: r->xtab = NULL;
1.1 schwarze 880: }
881:
882: void
883: roff_reset(struct roff *r)
884: {
885:
886: roff_free1(r);
1.99 schwarze 887: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48 schwarze 888: r->control = 0;
1.1 schwarze 889: }
890:
891: void
892: roff_free(struct roff *r)
893: {
894:
895: roff_free1(r);
896: free(r);
897: }
898:
899: struct roff *
1.109 schwarze 900: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.1 schwarze 901: {
902: struct roff *r;
903:
1.11 schwarze 904: r = mandoc_calloc(1, sizeof(struct roff));
1.35 schwarze 905: r->parse = parse;
1.109 schwarze 906: r->mchars = mchars;
1.72 schwarze 907: r->options = options;
1.99 schwarze 908: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2 schwarze 909: r->rstackpos = -1;
1.80 schwarze 910:
1.42 schwarze 911: roffhash_init();
1.37 schwarze 912:
1.1 schwarze 913: return(r);
1.137 schwarze 914: }
915:
1.138 ! schwarze 916: /* --- syntax tree state data management ---------------------------------- */
! 917:
1.137 schwarze 918: static void
919: roff_man_free1(struct roff_man *man)
920: {
921:
1.138 ! schwarze 922: if (man->first != NULL)
! 923: roff_node_delete(man, man->first);
1.137 schwarze 924: free(man->meta.msec);
925: free(man->meta.vol);
926: free(man->meta.os);
927: free(man->meta.arch);
928: free(man->meta.title);
929: free(man->meta.name);
930: free(man->meta.date);
931: }
932:
933: static void
934: roff_man_alloc1(struct roff_man *man)
935: {
936:
937: memset(&man->meta, 0, sizeof(man->meta));
938: man->first = mandoc_calloc(1, sizeof(*man->first));
939: man->first->type = ROFFT_ROOT;
940: man->last = man->first;
941: man->last_es = NULL;
942: man->flags = 0;
943: man->macroset = MACROSET_NONE;
944: man->lastsec = man->lastnamed = SEC_NONE;
945: man->next = ROFF_NEXT_CHILD;
946: }
947:
948: void
949: roff_man_reset(struct roff_man *man)
950: {
951:
952: roff_man_free1(man);
953: roff_man_alloc1(man);
954: }
955:
956: void
957: roff_man_free(struct roff_man *man)
958: {
959:
960: roff_man_free1(man);
961: free(man);
962: }
963:
964: struct roff_man *
965: roff_man_alloc(struct roff *roff, struct mparse *parse,
966: const char *defos, int quick)
967: {
968: struct roff_man *man;
969:
970: man = mandoc_calloc(1, sizeof(*man));
971: man->parse = parse;
972: man->roff = roff;
973: man->defos = defos;
974: man->quick = quick;
975: roff_man_alloc1(man);
976: return(man);
1.1 schwarze 977: }
978:
1.138 ! schwarze 979: /* --- syntax tree handling ----------------------------------------------- */
! 980:
! 981: struct roff_node *
! 982: roff_node_alloc(struct roff_man *man, int line, int pos,
! 983: enum roff_type type, int tok)
! 984: {
! 985: struct roff_node *n;
! 986:
! 987: n = mandoc_calloc(1, sizeof(*n));
! 988: n->line = line;
! 989: n->pos = pos;
! 990: n->tok = tok;
! 991: n->type = type;
! 992: n->sec = man->lastsec;
! 993:
! 994: if (man->flags & MDOC_SYNOPSIS)
! 995: n->flags |= MDOC_SYNPRETTY;
! 996: else
! 997: n->flags &= ~MDOC_SYNPRETTY;
! 998: if (man->flags & MDOC_NEWLINE)
! 999: n->flags |= MDOC_LINE;
! 1000: man->flags &= ~MDOC_NEWLINE;
! 1001:
! 1002: return(n);
! 1003: }
! 1004:
! 1005: void
! 1006: roff_node_append(struct roff_man *man, struct roff_node *n)
! 1007: {
! 1008:
! 1009: switch (man->next) {
! 1010: case ROFF_NEXT_SIBLING:
! 1011: man->last->next = n;
! 1012: n->prev = man->last;
! 1013: n->parent = man->last->parent;
! 1014: break;
! 1015: case ROFF_NEXT_CHILD:
! 1016: man->last->child = n;
! 1017: n->parent = man->last;
! 1018: break;
! 1019: default:
! 1020: abort();
! 1021: /* NOTREACHED */
! 1022: }
! 1023: n->parent->nchild++;
! 1024:
! 1025: /*
! 1026: * Copy over the normalised-data pointer of our parent. Not
! 1027: * everybody has one, but copying a null pointer is fine.
! 1028: */
! 1029:
! 1030: switch (n->type) {
! 1031: case ROFFT_BODY:
! 1032: if (n->end != ENDBODY_NOT)
! 1033: break;
! 1034: /* FALLTHROUGH */
! 1035: case ROFFT_TAIL:
! 1036: /* FALLTHROUGH */
! 1037: case ROFFT_HEAD:
! 1038: n->norm = n->parent->norm;
! 1039: break;
! 1040: default:
! 1041: break;
! 1042: }
! 1043:
! 1044: if (man->macroset == MACROSET_MDOC)
! 1045: mdoc_valid_pre(man, n);
! 1046:
! 1047: switch (n->type) {
! 1048: case ROFFT_HEAD:
! 1049: assert(n->parent->type == ROFFT_BLOCK);
! 1050: n->parent->head = n;
! 1051: break;
! 1052: case ROFFT_BODY:
! 1053: if (n->end)
! 1054: break;
! 1055: assert(n->parent->type == ROFFT_BLOCK);
! 1056: n->parent->body = n;
! 1057: break;
! 1058: case ROFFT_TAIL:
! 1059: assert(n->parent->type == ROFFT_BLOCK);
! 1060: n->parent->tail = n;
! 1061: break;
! 1062: default:
! 1063: break;
! 1064: }
! 1065: man->last = n;
! 1066: }
! 1067:
! 1068: struct roff_node *
! 1069: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
! 1070: {
! 1071: struct roff_node *n;
! 1072:
! 1073: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
! 1074: roff_node_append(man, n);
! 1075: man->next = ROFF_NEXT_CHILD;
! 1076: return(n);
! 1077: }
! 1078:
! 1079: struct roff_node *
! 1080: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
! 1081: {
! 1082: struct roff_node *n;
! 1083:
! 1084: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
! 1085: roff_node_append(man, n);
! 1086: man->next = ROFF_NEXT_CHILD;
! 1087: return(n);
! 1088: }
! 1089:
! 1090: void
! 1091: roff_node_unlink(struct roff_man *man, struct roff_node *n)
! 1092: {
! 1093:
! 1094: /* Adjust siblings. */
! 1095:
! 1096: if (n->prev)
! 1097: n->prev->next = n->next;
! 1098: if (n->next)
! 1099: n->next->prev = n->prev;
! 1100:
! 1101: /* Adjust parent. */
! 1102:
! 1103: if (n->parent != NULL) {
! 1104: n->parent->nchild--;
! 1105: if (n->parent->child == n)
! 1106: n->parent->child = n->next;
! 1107: if (n->parent->last == n)
! 1108: n->parent->last = n->prev;
! 1109: }
! 1110:
! 1111: /* Adjust parse point. */
! 1112:
! 1113: if (man == NULL)
! 1114: return;
! 1115: if (man->last == n) {
! 1116: if (n->prev == NULL) {
! 1117: man->last = n->parent;
! 1118: man->next = ROFF_NEXT_CHILD;
! 1119: } else {
! 1120: man->last = n->prev;
! 1121: man->next = ROFF_NEXT_SIBLING;
! 1122: }
! 1123: }
! 1124: if (man->first == n)
! 1125: man->first = NULL;
! 1126: }
! 1127:
! 1128: void
! 1129: roff_node_free(struct roff_node *n)
! 1130: {
! 1131:
! 1132: if (n->args != NULL)
! 1133: mdoc_argv_free(n->args);
! 1134: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
! 1135: free(n->norm);
! 1136: free(n->string);
! 1137: free(n);
! 1138: }
! 1139:
! 1140: void
! 1141: roff_node_delete(struct roff_man *man, struct roff_node *n)
! 1142: {
! 1143:
! 1144: while (n->child != NULL)
! 1145: roff_node_delete(man, n->child);
! 1146: assert(n->nchild == 0);
! 1147: roff_node_unlink(man, n);
! 1148: roff_node_free(n);
! 1149: }
! 1150:
! 1151: /* --- main functions of the roff parser ---------------------------------- */
! 1152:
1.8 schwarze 1153: /*
1.79 schwarze 1154: * In the current line, expand escape sequences that tend to get
1155: * used in numerical expressions and conditional requests.
1156: * Also check the syntax of the remaining escape sequences.
1.8 schwarze 1157: */
1.45 schwarze 1158: static enum rofferr
1.110 schwarze 1159: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8 schwarze 1160: {
1.81 schwarze 1161: char ubuf[24]; /* buffer to print the number */
1.78 schwarze 1162: const char *start; /* start of the string to process */
1.82 schwarze 1163: char *stesc; /* start of an escape sequence ('\\') */
1.23 schwarze 1164: const char *stnam; /* start of the name, after "[(*" */
1165: const char *cp; /* end of the name, e.g. before ']' */
1166: const char *res; /* the string to be substituted */
1.110 schwarze 1167: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 1168: size_t maxl; /* expected length of the escape name */
1169: size_t naml; /* actual length of the escape name */
1.109 schwarze 1170: enum mandoc_esc esc; /* type of the escape sequence */
1171: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 1172: int expand_count; /* to avoid infinite loops */
1.79 schwarze 1173: int npos; /* position in numeric expression */
1.91 schwarze 1174: int arg_complete; /* argument not interrupted by eol */
1.79 schwarze 1175: char term; /* character terminating the escape */
1.8 schwarze 1176:
1.43 schwarze 1177: expand_count = 0;
1.110 schwarze 1178: start = buf->buf + pos;
1.78 schwarze 1179: stesc = strchr(start, '\0') - 1;
1180: while (stesc-- > start) {
1.43 schwarze 1181:
1.78 schwarze 1182: /* Search backwards for the next backslash. */
1183:
1.110 schwarze 1184: if (*stesc != '\\')
1.78 schwarze 1185: continue;
1186:
1187: /* If it is escaped, skip it. */
1188:
1189: for (cp = stesc - 1; cp >= start; cp--)
1.110 schwarze 1190: if (*cp != '\\')
1.78 schwarze 1191: break;
1192:
1.110 schwarze 1193: if ((stesc - cp) % 2 == 0) {
1.82 schwarze 1194: stesc = (char *)cp;
1.78 schwarze 1195: continue;
1196: }
1.23 schwarze 1197:
1.79 schwarze 1198: /* Decide whether to expand or to check only. */
1.23 schwarze 1199:
1.79 schwarze 1200: term = '\0';
1.78 schwarze 1201: cp = stesc + 1;
1.53 schwarze 1202: switch (*cp) {
1.80 schwarze 1203: case '*':
1.53 schwarze 1204: res = NULL;
1205: break;
1.80 schwarze 1206: case 'B':
1.79 schwarze 1207: /* FALLTHROUGH */
1.80 schwarze 1208: case 'w':
1.79 schwarze 1209: term = cp[1];
1210: /* FALLTHROUGH */
1.80 schwarze 1211: case 'n':
1.53 schwarze 1212: res = ubuf;
1213: break;
1214: default:
1.109 schwarze 1215: esc = mandoc_escape(&cp, &stnam, &inaml);
1216: if (esc == ESCAPE_ERROR ||
1217: (esc == ESCAPE_SPECIAL &&
1218: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.92 schwarze 1219: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110 schwarze 1220: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1221: "%.*s", (int)(cp - stesc), stesc);
1.78 schwarze 1222: continue;
1.42 schwarze 1223: }
1224:
1.78 schwarze 1225: if (EXPAND_LIMIT < ++expand_count) {
1226: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110 schwarze 1227: ln, (int)(stesc - buf->buf), NULL);
1.78 schwarze 1228: return(ROFF_IGN);
1229: }
1.23 schwarze 1230:
1231: /*
1232: * The third character decides the length
1.53 schwarze 1233: * of the name of the string or register.
1.23 schwarze 1234: * Save a pointer to the name.
1235: */
1236:
1.110 schwarze 1237: if (term == '\0') {
1.79 schwarze 1238: switch (*++cp) {
1.80 schwarze 1239: case '\0':
1.79 schwarze 1240: maxl = 0;
1241: break;
1.80 schwarze 1242: case '(':
1.79 schwarze 1243: cp++;
1244: maxl = 2;
1245: break;
1.80 schwarze 1246: case '[':
1.79 schwarze 1247: cp++;
1248: term = ']';
1249: maxl = 0;
1250: break;
1251: default:
1252: maxl = 1;
1253: break;
1254: }
1255: } else {
1256: cp += 2;
1.8 schwarze 1257: maxl = 0;
1258: }
1.23 schwarze 1259: stnam = cp;
1.8 schwarze 1260:
1.23 schwarze 1261: /* Advance to the end of the name. */
1.8 schwarze 1262:
1.125 schwarze 1263: naml = 0;
1.91 schwarze 1264: arg_complete = 1;
1.125 schwarze 1265: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1266: if (*cp == '\0') {
1.92 schwarze 1267: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 1268: ln, (int)(stesc - buf->buf), stesc);
1.91 schwarze 1269: arg_complete = 0;
1.79 schwarze 1270: break;
1.42 schwarze 1271: }
1.110 schwarze 1272: if (maxl == 0 && *cp == term) {
1.79 schwarze 1273: cp++;
1.125 schwarze 1274: break;
1275: }
1276: if (*cp++ != '\\' || stesc[1] != 'w') {
1277: naml++;
1278: continue;
1279: }
1280: switch (mandoc_escape(&cp, NULL, NULL)) {
1281: case ESCAPE_SPECIAL:
1282: /* FALLTHROUGH */
1283: case ESCAPE_UNICODE:
1284: /* FALLTHROUGH */
1285: case ESCAPE_NUMBERED:
1286: /* FALLTHROUGH */
1287: case ESCAPE_OVERSTRIKE:
1288: naml++;
1289: break;
1290: default:
1.8 schwarze 1291: break;
1.79 schwarze 1292: }
1.8 schwarze 1293: }
1294:
1.23 schwarze 1295: /*
1296: * Retrieve the replacement string; if it is
1297: * undefined, resume searching for escapes.
1298: */
1299:
1.79 schwarze 1300: switch (stesc[1]) {
1.80 schwarze 1301: case '*':
1.91 schwarze 1302: if (arg_complete)
1303: res = roff_getstrn(r, stnam, naml);
1.79 schwarze 1304: break;
1.80 schwarze 1305: case 'B':
1.79 schwarze 1306: npos = 0;
1.91 schwarze 1307: ubuf[0] = arg_complete &&
1.133 schwarze 1308: roff_evalnum(r, ln, stnam, &npos,
1309: NULL, ROFFNUM_SCALE) &&
1.91 schwarze 1310: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1311: ubuf[1] = '\0';
1312: break;
1.80 schwarze 1313: case 'n':
1.91 schwarze 1314: if (arg_complete)
1315: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1316: roff_getregn(r, stnam, naml));
1317: else
1318: ubuf[0] = '\0';
1.79 schwarze 1319: break;
1.80 schwarze 1320: case 'w':
1.91 schwarze 1321: /* use even incomplete args */
1.81 schwarze 1322: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1323: 24 * (int)naml);
1324: break;
1325: }
1.8 schwarze 1326:
1.110 schwarze 1327: if (res == NULL) {
1.92 schwarze 1328: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1329: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1330: "%.*s", (int)naml, stnam);
1.37 schwarze 1331: res = "";
1.118 schwarze 1332: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1333: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1334: ln, (int)(stesc - buf->buf), NULL);
1335: return(ROFF_IGN);
1.8 schwarze 1336: }
1337:
1.23 schwarze 1338: /* Replace the escape sequence by the string. */
1339:
1.82 schwarze 1340: *stesc = '\0';
1.110 schwarze 1341: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1342: buf->buf, res, cp) + 1;
1.8 schwarze 1343:
1.78 schwarze 1344: /* Prepare for the next replacement. */
1.8 schwarze 1345:
1.78 schwarze 1346: start = nbuf + pos;
1.110 schwarze 1347: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1348: free(buf->buf);
1349: buf->buf = nbuf;
1.42 schwarze 1350: }
1.45 schwarze 1351: return(ROFF_CONT);
1.42 schwarze 1352: }
1353:
1354: /*
1.51 schwarze 1355: * Process text streams:
1356: * Convert all breakable hyphens into ASCII_HYPH.
1357: * Decrement and spring input line trap.
1.42 schwarze 1358: */
1359: static enum rofferr
1.110 schwarze 1360: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42 schwarze 1361: {
1362: size_t sz;
1363: const char *start;
1.51 schwarze 1364: char *p;
1365: int isz;
1.42 schwarze 1366: enum mandoc_esc esc;
1367:
1.110 schwarze 1368: start = p = buf->buf + pos;
1.42 schwarze 1369:
1.110 schwarze 1370: while (*p != '\0') {
1.42 schwarze 1371: sz = strcspn(p, "-\\");
1372: p += sz;
1373:
1.110 schwarze 1374: if (*p == '\0')
1.42 schwarze 1375: break;
1376:
1.110 schwarze 1377: if (*p == '\\') {
1.42 schwarze 1378: /* Skip over escapes. */
1379: p++;
1.62 schwarze 1380: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1381: if (esc == ESCAPE_ERROR)
1.42 schwarze 1382: break;
1.136 schwarze 1383: while (*p == '-')
1384: p++;
1.42 schwarze 1385: continue;
1386: } else if (p == start) {
1387: p++;
1388: continue;
1389: }
1390:
1.44 schwarze 1391: if (isalpha((unsigned char)p[-1]) &&
1392: isalpha((unsigned char)p[1]))
1.42 schwarze 1393: *p = ASCII_HYPH;
1394: p++;
1.8 schwarze 1395: }
1396:
1.51 schwarze 1397: /* Spring the input line trap. */
1.110 schwarze 1398: if (roffit_lines == 1) {
1399: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1400: free(buf->buf);
1401: buf->buf = p;
1402: buf->sz = isz + 1;
1.51 schwarze 1403: *offs = 0;
1404: free(roffit_macro);
1405: roffit_lines = 0;
1406: return(ROFF_REPARSE);
1.110 schwarze 1407: } else if (roffit_lines > 1)
1.51 schwarze 1408: --roffit_lines;
1.42 schwarze 1409: return(ROFF_CONT);
1.8 schwarze 1410: }
1411:
1.1 schwarze 1412: enum rofferr
1.110 schwarze 1413: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1414: {
1415: enum rofft t;
1.27 schwarze 1416: enum rofferr e;
1.110 schwarze 1417: int pos; /* parse point */
1.115 schwarze 1418: int spos; /* saved parse point for messages */
1.110 schwarze 1419: int ppos; /* original offset in buf->buf */
1420: int ctl; /* macro line (boolean) */
1421:
1422: ppos = pos = *offs;
1.1 schwarze 1423:
1.102 schwarze 1424: /* Handle in-line equation delimiters. */
1425:
1.108 schwarze 1426: if (r->tbl == NULL &&
1427: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1428: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1429: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1430: if (e == ROFF_REPARSE)
1431: return(e);
1432: assert(e == ROFF_CONT);
1433: }
1434:
1435: /* Expand some escape sequences. */
1.8 schwarze 1436:
1.110 schwarze 1437: e = roff_res(r, buf, ln, pos);
1438: if (e == ROFF_IGN)
1.45 schwarze 1439: return(e);
1.110 schwarze 1440: assert(e == ROFF_CONT);
1.8 schwarze 1441:
1.110 schwarze 1442: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1443:
1.8 schwarze 1444: /*
1.2 schwarze 1445: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1446: * text through the macro's filter.
1447: * Equations process all content themselves.
1448: * Tables process almost all content themselves, but we want
1449: * to warn about macros before passing it there.
1.2 schwarze 1450: */
1451:
1.124 schwarze 1452: if (r->last != NULL && ! ctl) {
1.2 schwarze 1453: t = r->last->tok;
1454: assert(roffs[t].text);
1.110 schwarze 1455: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1456: assert(e == ROFF_IGN || e == ROFF_CONT);
1457: if (e != ROFF_CONT)
1.32 schwarze 1458: return(e);
1.54 schwarze 1459: }
1.124 schwarze 1460: if (r->eqn != NULL)
1.110 schwarze 1461: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.124 schwarze 1462: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.131 schwarze 1463: return(tbl_read(r->tbl, ln, buf->buf, ppos));
1.124 schwarze 1464: if ( ! ctl)
1.110 schwarze 1465: return(roff_parsetext(buf, pos, offs));
1.100 schwarze 1466:
1467: /* Skip empty request lines. */
1468:
1.110 schwarze 1469: if (buf->buf[pos] == '"') {
1.100 schwarze 1470: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1471: ln, pos, NULL);
1472: return(ROFF_IGN);
1.110 schwarze 1473: } else if (buf->buf[pos] == '\0')
1.100 schwarze 1474: return(ROFF_IGN);
1.2 schwarze 1475:
1476: /*
1477: * If a scope is open, go to the child handler for that macro,
1478: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1479: * Don't do so if an equation is open.
1.2 schwarze 1480: */
1481:
1482: if (r->last) {
1.1 schwarze 1483: t = r->last->tok;
1484: assert(roffs[t].sub);
1.110 schwarze 1485: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1486: }
1487:
1.115 schwarze 1488: /* No scope is open. This is a new request or macro. */
1489:
1490: spos = pos;
1491: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1492:
1493: /* Tables ignore most macros. */
1494:
1495: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1496: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1497: ln, pos, buf->buf + spos);
1.129 schwarze 1498: if (t == ROFF_TS)
1499: return(ROFF_IGN);
1500: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1501: pos++;
1502: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1503: pos++;
1504: return(tbl_read(r->tbl, ln, buf->buf, pos));
1.115 schwarze 1505: }
1506:
1.2 schwarze 1507: /*
1.115 schwarze 1508: * This is neither a roff request nor a user-defined macro.
1509: * Let the standard macro set parsers handle it.
1.2 schwarze 1510: */
1511:
1.115 schwarze 1512: if (t == ROFF_MAX)
1.1 schwarze 1513: return(ROFF_CONT);
1.115 schwarze 1514:
1515: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1516:
1.2 schwarze 1517: assert(roffs[t].proc);
1.110 schwarze 1518: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1519: }
1520:
1.27 schwarze 1521: void
1.2 schwarze 1522: roff_endparse(struct roff *r)
1523: {
1.1 schwarze 1524:
1.27 schwarze 1525: if (r->last)
1.94 schwarze 1526: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1527: r->last->line, r->last->col,
1528: roffs[r->last->tok].name);
1.27 schwarze 1529:
1.32 schwarze 1530: if (r->eqn) {
1.94 schwarze 1531: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1532: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41 schwarze 1533: eqn_end(&r->eqn);
1.32 schwarze 1534: }
1535:
1.27 schwarze 1536: if (r->tbl) {
1.94 schwarze 1537: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1538: r->tbl->line, r->tbl->pos, "TS");
1.41 schwarze 1539: tbl_end(&r->tbl);
1.27 schwarze 1540: }
1.1 schwarze 1541: }
1542:
1543: /*
1544: * Parse a roff node's type from the input buffer. This must be in the
1545: * form of ".foo xxx" in the usual way.
1546: */
1547: static enum rofft
1.87 schwarze 1548: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1549: {
1.87 schwarze 1550: char *cp;
1.16 schwarze 1551: const char *mac;
1552: size_t maclen;
1.1 schwarze 1553: enum rofft t;
1554:
1.87 schwarze 1555: cp = buf + *pos;
1556:
1557: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.1 schwarze 1558: return(ROFF_MAX);
1559:
1.87 schwarze 1560: mac = cp;
1561: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 1562:
1.16 schwarze 1563: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42 schwarze 1564: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1 schwarze 1565:
1.87 schwarze 1566: if (ROFF_MAX != t)
1567: *pos = cp - buf;
1.1 schwarze 1568:
1569: return(t);
1570: }
1571:
1.138 ! schwarze 1572: /* --- handling of request blocks ----------------------------------------- */
! 1573:
1.1 schwarze 1574: static enum rofferr
1.2 schwarze 1575: roff_cblock(ROFF_ARGS)
1.1 schwarze 1576: {
1577:
1.2 schwarze 1578: /*
1579: * A block-close `..' should only be invoked as a child of an
1580: * ignore macro, otherwise raise a warning and just ignore it.
1581: */
1582:
1.110 schwarze 1583: if (r->last == NULL) {
1.94 schwarze 1584: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1585: ln, ppos, "..");
1.2 schwarze 1586: return(ROFF_IGN);
1587: }
1.1 schwarze 1588:
1.2 schwarze 1589: switch (r->last->tok) {
1.80 schwarze 1590: case ROFF_am:
1.93 schwarze 1591: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.2 schwarze 1592: /* FALLTHROUGH */
1.80 schwarze 1593: case ROFF_ami:
1.2 schwarze 1594: /* FALLTHROUGH */
1.80 schwarze 1595: case ROFF_de:
1.23 schwarze 1596: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2 schwarze 1597: /* FALLTHROUGH */
1.80 schwarze 1598: case ROFF_dei:
1.2 schwarze 1599: /* FALLTHROUGH */
1.80 schwarze 1600: case ROFF_ig:
1.2 schwarze 1601: break;
1602: default:
1.94 schwarze 1603: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1604: ln, ppos, "..");
1.1 schwarze 1605: return(ROFF_IGN);
1.2 schwarze 1606: }
1607:
1.110 schwarze 1608: if (buf->buf[pos] != '\0')
1.90 schwarze 1609: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 1610: ".. %s", buf->buf + pos);
1.2 schwarze 1611:
1612: roffnode_pop(r);
1613: roffnode_cleanscope(r);
1614: return(ROFF_IGN);
1615:
1616: }
1.1 schwarze 1617:
1.2 schwarze 1618: static void
1619: roffnode_cleanscope(struct roff *r)
1620: {
1.1 schwarze 1621:
1.2 schwarze 1622: while (r->last) {
1.46 schwarze 1623: if (--r->last->endspan != 0)
1.2 schwarze 1624: break;
1625: roffnode_pop(r);
1626: }
1627: }
1.1 schwarze 1628:
1.68 schwarze 1629: static void
1630: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 1631: {
1.1 schwarze 1632:
1.2 schwarze 1633: if (NULL == r->last) {
1.94 schwarze 1634: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1635: ln, ppos, "\\}");
1.68 schwarze 1636: return;
1.2 schwarze 1637: }
1.1 schwarze 1638:
1.2 schwarze 1639: switch (r->last->tok) {
1.80 schwarze 1640: case ROFF_el:
1.2 schwarze 1641: /* FALLTHROUGH */
1.80 schwarze 1642: case ROFF_ie:
1.2 schwarze 1643: /* FALLTHROUGH */
1.80 schwarze 1644: case ROFF_if:
1.2 schwarze 1645: break;
1646: default:
1.94 schwarze 1647: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1648: ln, ppos, "\\}");
1.68 schwarze 1649: return;
1.2 schwarze 1650: }
1.1 schwarze 1651:
1.2 schwarze 1652: if (r->last->endspan > -1) {
1.94 schwarze 1653: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1654: ln, ppos, "\\}");
1.68 schwarze 1655: return;
1.2 schwarze 1656: }
1657:
1658: roffnode_pop(r);
1659: roffnode_cleanscope(r);
1.68 schwarze 1660: return;
1.1 schwarze 1661: }
1662:
1663: static enum rofferr
1.2 schwarze 1664: roff_block(ROFF_ARGS)
1.1 schwarze 1665: {
1.93 schwarze 1666: const char *name;
1667: char *iname, *cp;
1.86 schwarze 1668: size_t namesz;
1.16 schwarze 1669:
1.93 schwarze 1670: /* Ignore groff compatibility mode for now. */
1.2 schwarze 1671:
1.110 schwarze 1672: if (tok == ROFF_de1)
1.93 schwarze 1673: tok = ROFF_de;
1.123 schwarze 1674: else if (tok == ROFF_dei1)
1675: tok = ROFF_dei;
1.110 schwarze 1676: else if (tok == ROFF_am1)
1.93 schwarze 1677: tok = ROFF_am;
1.123 schwarze 1678: else if (tok == ROFF_ami1)
1679: tok = ROFF_ami;
1.93 schwarze 1680:
1681: /* Parse the macro name argument. */
1682:
1.110 schwarze 1683: cp = buf->buf + pos;
1684: if (tok == ROFF_ig) {
1.93 schwarze 1685: iname = NULL;
1686: namesz = 0;
1687: } else {
1688: iname = cp;
1689: namesz = roff_getname(r, &cp, ln, ppos);
1690: iname[namesz] = '\0';
1691: }
1.22 schwarze 1692:
1.93 schwarze 1693: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 1694:
1.110 schwarze 1695: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1696: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1697: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1698: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1699: "%.*s", (int)namesz, iname);
1700: namesz = 0;
1701: } else
1702: namesz = strlen(name);
1703: } else
1704: name = iname;
1.22 schwarze 1705:
1.110 schwarze 1706: if (namesz == 0 && tok != ROFF_ig) {
1.93 schwarze 1707: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1708: ln, ppos, roffs[tok].name);
1709: return(ROFF_IGN);
1710: }
1.2 schwarze 1711:
1.16 schwarze 1712: roffnode_push(r, tok, name, ln, ppos);
1713:
1714: /*
1715: * At the beginning of a `de' macro, clear the existing string
1716: * with the same name, if there is one. New content will be
1.66 schwarze 1717: * appended from roff_block_text() in multiline mode.
1.16 schwarze 1718: */
1.22 schwarze 1719:
1.110 schwarze 1720: if (tok == ROFF_de || tok == ROFF_dei)
1.86 schwarze 1721: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2 schwarze 1722:
1.110 schwarze 1723: if (*cp == '\0')
1.2 schwarze 1724: return(ROFF_IGN);
1.1 schwarze 1725:
1.93 schwarze 1726: /* Get the custom end marker. */
1.22 schwarze 1727:
1.93 schwarze 1728: iname = cp;
1.86 schwarze 1729: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 1730:
1731: /* Resolve the end marker if it is indirect. */
1732:
1.110 schwarze 1733: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1734: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93 schwarze 1735: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1736: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1737: "%.*s", (int)namesz, iname);
1738: namesz = 0;
1739: } else
1740: namesz = strlen(name);
1741: } else
1742: name = iname;
1743:
1.86 schwarze 1744: if (namesz)
1745: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 1746:
1.110 schwarze 1747: if (*cp != '\0')
1.90 schwarze 1748: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1749: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.1 schwarze 1750:
1751: return(ROFF_IGN);
1752: }
1753:
1754: static enum rofferr
1.2 schwarze 1755: roff_block_sub(ROFF_ARGS)
1.1 schwarze 1756: {
1.2 schwarze 1757: enum rofft t;
1758: int i, j;
1759:
1760: /*
1761: * First check whether a custom macro exists at this level. If
1762: * it does, then check against it. This is some of groff's
1763: * stranger behaviours. If we encountered a custom end-scope
1764: * tag and that tag also happens to be a "real" macro, then we
1765: * need to try interpreting it again as a real macro. If it's
1766: * not, then return ignore. Else continue.
1767: */
1768:
1769: if (r->last->end) {
1.35 schwarze 1770: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 1771: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 1772: break;
1.1 schwarze 1773:
1.110 schwarze 1774: if (r->last->end[j] == '\0' &&
1775: (buf->buf[i] == '\0' ||
1776: buf->buf[i] == ' ' ||
1777: buf->buf[i] == '\t')) {
1.2 schwarze 1778: roffnode_pop(r);
1779: roffnode_cleanscope(r);
1.1 schwarze 1780:
1.110 schwarze 1781: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 1782: i++;
1783:
1784: pos = i;
1.110 schwarze 1785: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1786: ROFF_MAX)
1.2 schwarze 1787: return(ROFF_RERUN);
1788: return(ROFF_IGN);
1789: }
1.1 schwarze 1790: }
1791:
1.2 schwarze 1792: /*
1793: * If we have no custom end-query or lookup failed, then try
1794: * pulling it out of the hashtable.
1795: */
1.1 schwarze 1796:
1.110 schwarze 1797: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 1798:
1.110 schwarze 1799: if (t != ROFF_cblock) {
1800: if (tok != ROFF_ig)
1801: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.1 schwarze 1802: return(ROFF_IGN);
1.16 schwarze 1803: }
1.1 schwarze 1804:
1.2 schwarze 1805: assert(roffs[t].proc);
1.110 schwarze 1806: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2 schwarze 1807: }
1808:
1809: static enum rofferr
1810: roff_block_text(ROFF_ARGS)
1811: {
1812:
1.110 schwarze 1813: if (tok != ROFF_ig)
1814: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 1815:
1.2 schwarze 1816: return(ROFF_IGN);
1817: }
1818:
1819: static enum rofferr
1820: roff_cond_sub(ROFF_ARGS)
1821: {
1822: enum rofft t;
1.37 schwarze 1823: char *ep;
1.71 schwarze 1824: int rr;
1.2 schwarze 1825:
1826: rr = r->last->rule;
1.37 schwarze 1827: roffnode_cleanscope(r);
1.110 schwarze 1828: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2 schwarze 1829:
1.37 schwarze 1830: /*
1.50 schwarze 1831: * Fully handle known macros when they are structurally
1832: * required or when the conditional evaluated to true.
1.5 schwarze 1833: */
1834:
1.110 schwarze 1835: if ((t != ROFF_MAX) &&
1836: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.50 schwarze 1837: assert(roffs[t].proc);
1.110 schwarze 1838: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.50 schwarze 1839: }
1.39 schwarze 1840:
1.69 schwarze 1841: /*
1842: * If `\}' occurs on a macro line without a preceding macro,
1843: * drop the line completely.
1844: */
1845:
1.110 schwarze 1846: ep = buf->buf + pos;
1847: if (ep[0] == '\\' && ep[1] == '}')
1.71 schwarze 1848: rr = 0;
1.69 schwarze 1849:
1.50 schwarze 1850: /* Always check for the closing delimiter `\}'. */
1.39 schwarze 1851:
1.110 schwarze 1852: while ((ep = strchr(ep, '\\')) != NULL) {
1853: if (*(++ep) == '}') {
1.70 schwarze 1854: *ep = '&';
1.110 schwarze 1855: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1856: }
1.119 schwarze 1857: if (*ep != '\0')
1858: ++ep;
1.50 schwarze 1859: }
1.71 schwarze 1860: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1861: }
1862:
1863: static enum rofferr
1864: roff_cond_text(ROFF_ARGS)
1865: {
1.37 schwarze 1866: char *ep;
1.71 schwarze 1867: int rr;
1.2 schwarze 1868:
1869: rr = r->last->rule;
1.37 schwarze 1870: roffnode_cleanscope(r);
1.1 schwarze 1871:
1.110 schwarze 1872: ep = buf->buf + pos;
1873: while ((ep = strchr(ep, '\\')) != NULL) {
1874: if (*(++ep) == '}') {
1.70 schwarze 1875: *ep = '&';
1.110 schwarze 1876: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 1877: }
1.119 schwarze 1878: if (*ep != '\0')
1879: ++ep;
1.2 schwarze 1880: }
1.71 schwarze 1881: return(rr ? ROFF_CONT : ROFF_IGN);
1.2 schwarze 1882: }
1883:
1.138 ! schwarze 1884: /* --- handling of numeric and conditional expressions -------------------- */
! 1885:
1.77 schwarze 1886: /*
1887: * Parse a single signed integer number. Stop at the first non-digit.
1888: * If there is at least one digit, return success and advance the
1889: * parse point, else return failure and let the parse point unchanged.
1890: * Ignore overflows, treat them just like the C language.
1891: */
1.56 schwarze 1892: static int
1.133 schwarze 1893: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56 schwarze 1894: {
1.133 schwarze 1895: int myres, scaled, n, p;
1.79 schwarze 1896:
1897: if (NULL == res)
1898: res = &myres;
1.56 schwarze 1899:
1900: p = *pos;
1901: n = v[p] == '-';
1.133 schwarze 1902: if (n || v[p] == '+')
1.56 schwarze 1903: p++;
1904:
1.133 schwarze 1905: if (flags & ROFFNUM_WHITE)
1906: while (isspace((unsigned char)v[p]))
1907: p++;
1908:
1.56 schwarze 1909: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 1910: *res = 10 * *res + v[p] - '0';
1.56 schwarze 1911: if (p == *pos + n)
1912: return 0;
1913:
1914: if (n)
1915: *res = -*res;
1916:
1.126 schwarze 1917: /* Each number may be followed by one optional scaling unit. */
1918:
1919: switch (v[p]) {
1920: case 'f':
1.133 schwarze 1921: scaled = *res * 65536;
1.126 schwarze 1922: break;
1923: case 'i':
1.133 schwarze 1924: scaled = *res * 240;
1.126 schwarze 1925: break;
1926: case 'c':
1.133 schwarze 1927: scaled = *res * 240 / 2.54;
1.126 schwarze 1928: break;
1929: case 'v':
1930: /* FALLTROUGH */
1931: case 'P':
1.133 schwarze 1932: scaled = *res * 40;
1.126 schwarze 1933: break;
1934: case 'm':
1935: /* FALLTROUGH */
1936: case 'n':
1.133 schwarze 1937: scaled = *res * 24;
1.126 schwarze 1938: break;
1939: case 'p':
1.133 schwarze 1940: scaled = *res * 10 / 3;
1.126 schwarze 1941: break;
1942: case 'u':
1.133 schwarze 1943: scaled = *res;
1.126 schwarze 1944: break;
1945: case 'M':
1.133 schwarze 1946: scaled = *res * 6 / 25;
1.126 schwarze 1947: break;
1948: default:
1.133 schwarze 1949: scaled = *res;
1.126 schwarze 1950: p--;
1951: break;
1952: }
1.133 schwarze 1953: if (flags & ROFFNUM_SCALE)
1954: *res = scaled;
1.126 schwarze 1955:
1956: *pos = p + 1;
1957: return(1);
1.56 schwarze 1958: }
1959:
1.71 schwarze 1960: /*
1961: * Evaluate a string comparison condition.
1962: * The first character is the delimiter.
1963: * Succeed if the string up to its second occurrence
1964: * matches the string up to its third occurence.
1965: * Advance the cursor after the third occurrence
1966: * or lacking that, to the end of the line.
1967: */
1968: static int
1969: roff_evalstrcond(const char *v, int *pos)
1970: {
1971: const char *s1, *s2, *s3;
1972: int match;
1973:
1974: match = 0;
1975: s1 = v + *pos; /* initial delimiter */
1976: s2 = s1 + 1; /* for scanning the first string */
1977: s3 = strchr(s2, *s1); /* for scanning the second string */
1978:
1979: if (NULL == s3) /* found no middle delimiter */
1980: goto out;
1981:
1982: while ('\0' != *++s3) {
1983: if (*s2 != *s3) { /* mismatch */
1984: s3 = strchr(s3, *s1);
1985: break;
1986: }
1987: if (*s3 == *s1) { /* found the final delimiter */
1988: match = 1;
1989: break;
1990: }
1991: s2++;
1992: }
1993:
1994: out:
1995: if (NULL == s3)
1996: s3 = strchr(s2, '\0');
1.114 schwarze 1997: else if (*s3 != '\0')
1.71 schwarze 1998: s3++;
1999: *pos = s3 - v;
2000: return(match);
2001: }
2002:
1.77 schwarze 2003: /*
2004: * Evaluate an optionally negated single character, numerical,
2005: * or string condition.
2006: */
1.71 schwarze 2007: static int
1.106 schwarze 2008: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.5 schwarze 2009: {
1.113 schwarze 2010: int number, savepos, wanttrue;
1.5 schwarze 2011:
1.71 schwarze 2012: if ('!' == v[*pos]) {
2013: wanttrue = 0;
2014: (*pos)++;
2015: } else
2016: wanttrue = 1;
2017:
1.5 schwarze 2018: switch (v[*pos]) {
1.112 schwarze 2019: case '\0':
2020: return(0);
1.80 schwarze 2021: case 'n':
1.71 schwarze 2022: /* FALLTHROUGH */
1.80 schwarze 2023: case 'o':
1.5 schwarze 2024: (*pos)++;
1.71 schwarze 2025: return(wanttrue);
1.80 schwarze 2026: case 'c':
1.71 schwarze 2027: /* FALLTHROUGH */
1.80 schwarze 2028: case 'd':
1.71 schwarze 2029: /* FALLTHROUGH */
1.80 schwarze 2030: case 'e':
1.5 schwarze 2031: /* FALLTHROUGH */
1.80 schwarze 2032: case 'r':
1.5 schwarze 2033: /* FALLTHROUGH */
1.80 schwarze 2034: case 't':
1.111 schwarze 2035: /* FALLTHROUGH */
2036: case 'v':
1.5 schwarze 2037: (*pos)++;
1.71 schwarze 2038: return(!wanttrue);
1.5 schwarze 2039: default:
2040: break;
2041: }
2042:
1.113 schwarze 2043: savepos = *pos;
1.133 schwarze 2044: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.77 schwarze 2045: return((number > 0) == wanttrue);
1.113 schwarze 2046: else if (*pos == savepos)
2047: return(roff_evalstrcond(v, pos) == wanttrue);
1.77 schwarze 2048: else
1.113 schwarze 2049: return (0);
1.5 schwarze 2050: }
2051:
1.2 schwarze 2052: static enum rofferr
1.21 schwarze 2053: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 2054: {
1.30 schwarze 2055:
1.21 schwarze 2056: return(ROFF_IGN);
2057: }
2058:
2059: static enum rofferr
1.123 schwarze 2060: roff_insec(ROFF_ARGS)
2061: {
2062:
2063: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
2064: ln, ppos, roffs[tok].name);
2065: return(ROFF_IGN);
2066: }
2067:
2068: static enum rofferr
2069: roff_unsupp(ROFF_ARGS)
2070: {
2071:
2072: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
2073: ln, ppos, roffs[tok].name);
2074: return(ROFF_IGN);
2075: }
2076:
2077: static enum rofferr
1.2 schwarze 2078: roff_cond(ROFF_ARGS)
2079: {
1.46 schwarze 2080:
2081: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 2082:
1.80 schwarze 2083: /*
1.35 schwarze 2084: * An `.el' has no conditional body: it will consume the value
2085: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 2086: * defaults to DENY.
1.35 schwarze 2087: *
2088: * If we're not an `el', however, then evaluate the conditional.
2089: */
1.1 schwarze 2090:
1.110 schwarze 2091: r->last->rule = tok == ROFF_el ?
1.80 schwarze 2092: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 2093: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 2094:
1.35 schwarze 2095: /*
2096: * An if-else will put the NEGATION of the current evaluated
2097: * conditional into the stack of rules.
2098: */
2099:
1.110 schwarze 2100: if (tok == ROFF_ie) {
1.96 schwarze 2101: if (r->rstackpos + 1 == r->rstacksz) {
2102: r->rstacksz += 16;
2103: r->rstack = mandoc_reallocarray(r->rstack,
2104: r->rstacksz, sizeof(int));
1.35 schwarze 2105: }
1.71 schwarze 2106: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 2107: }
1.5 schwarze 2108:
2109: /* If the parent has false as its rule, then so do we. */
2110:
1.71 schwarze 2111: if (r->last->parent && !r->last->parent->rule)
2112: r->last->rule = 0;
1.5 schwarze 2113:
2114: /*
1.46 schwarze 2115: * Determine scope.
2116: * If there is nothing on the line after the conditional,
2117: * not even whitespace, use next-line scope.
1.5 schwarze 2118: */
1.2 schwarze 2119:
1.110 schwarze 2120: if (buf->buf[pos] == '\0') {
1.46 schwarze 2121: r->last->endspan = 2;
2122: goto out;
2123: }
2124:
1.110 schwarze 2125: while (buf->buf[pos] == ' ')
1.46 schwarze 2126: pos++;
2127:
2128: /* An opening brace requests multiline scope. */
1.2 schwarze 2129:
1.110 schwarze 2130: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 2131: r->last->endspan = -1;
2132: pos += 2;
1.46 schwarze 2133: goto out;
1.80 schwarze 2134: }
1.2 schwarze 2135:
2136: /*
1.46 schwarze 2137: * Anything else following the conditional causes
2138: * single-line scope. Warn if the scope contains
2139: * nothing but trailing whitespace.
1.2 schwarze 2140: */
2141:
1.110 schwarze 2142: if (buf->buf[pos] == '\0')
1.89 schwarze 2143: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
2144: ln, ppos, roffs[tok].name);
1.2 schwarze 2145:
1.46 schwarze 2146: r->last->endspan = 1;
1.1 schwarze 2147:
1.46 schwarze 2148: out:
1.2 schwarze 2149: *offs = pos;
2150: return(ROFF_RERUN);
1.1 schwarze 2151: }
2152:
1.2 schwarze 2153: static enum rofferr
1.7 schwarze 2154: roff_ds(ROFF_ARGS)
2155: {
1.85 schwarze 2156: char *string;
2157: const char *name;
2158: size_t namesz;
1.10 schwarze 2159:
1.123 schwarze 2160: /* Ignore groff compatibility mode for now. */
2161:
2162: if (tok == ROFF_ds1)
2163: tok = ROFF_ds;
2164: else if (tok == ROFF_as1)
2165: tok = ROFF_as;
2166:
1.10 schwarze 2167: /*
1.85 schwarze 2168: * The first word is the name of the string.
2169: * If it is empty or terminated by an escape sequence,
2170: * abort the `ds' request without defining anything.
1.10 schwarze 2171: */
1.7 schwarze 2172:
1.110 schwarze 2173: name = string = buf->buf + pos;
2174: if (*name == '\0')
1.7 schwarze 2175: return(ROFF_IGN);
2176:
1.85 schwarze 2177: namesz = roff_getname(r, &string, ln, pos);
1.110 schwarze 2178: if (name[namesz] == '\\')
1.85 schwarze 2179: return(ROFF_IGN);
2180:
2181: /* Read past the initial double-quote, if any. */
1.110 schwarze 2182: if (*string == '"')
1.7 schwarze 2183: string++;
2184:
1.10 schwarze 2185: /* The rest is the value. */
1.85 schwarze 2186: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2187: ROFF_as == tok);
1.7 schwarze 2188: return(ROFF_IGN);
2189: }
2190:
1.77 schwarze 2191: /*
2192: * Parse a single operator, one or two characters long.
2193: * If the operator is recognized, return success and advance the
2194: * parse point, else return failure and let the parse point unchanged.
2195: */
2196: static int
2197: roff_getop(const char *v, int *pos, char *res)
2198: {
2199:
2200: *res = v[*pos];
2201:
2202: switch (*res) {
1.80 schwarze 2203: case '+':
1.77 schwarze 2204: /* FALLTHROUGH */
1.80 schwarze 2205: case '-':
1.77 schwarze 2206: /* FALLTHROUGH */
1.80 schwarze 2207: case '*':
1.77 schwarze 2208: /* FALLTHROUGH */
1.80 schwarze 2209: case '/':
1.77 schwarze 2210: /* FALLTHROUGH */
1.80 schwarze 2211: case '%':
1.77 schwarze 2212: /* FALLTHROUGH */
1.80 schwarze 2213: case '&':
1.77 schwarze 2214: /* FALLTHROUGH */
1.80 schwarze 2215: case ':':
1.77 schwarze 2216: break;
2217: case '<':
2218: switch (v[*pos + 1]) {
1.80 schwarze 2219: case '=':
1.77 schwarze 2220: *res = 'l';
2221: (*pos)++;
2222: break;
1.80 schwarze 2223: case '>':
1.77 schwarze 2224: *res = '!';
2225: (*pos)++;
2226: break;
1.80 schwarze 2227: case '?':
1.77 schwarze 2228: *res = 'i';
2229: (*pos)++;
2230: break;
2231: default:
2232: break;
2233: }
2234: break;
2235: case '>':
2236: switch (v[*pos + 1]) {
1.80 schwarze 2237: case '=':
1.77 schwarze 2238: *res = 'g';
2239: (*pos)++;
2240: break;
1.80 schwarze 2241: case '?':
1.77 schwarze 2242: *res = 'a';
2243: (*pos)++;
2244: break;
2245: default:
2246: break;
2247: }
2248: break;
2249: case '=':
2250: if ('=' == v[*pos + 1])
2251: (*pos)++;
2252: break;
2253: default:
2254: return(0);
2255: }
2256: (*pos)++;
2257:
2258: return(*res);
2259: }
2260:
2261: /*
2262: * Evaluate either a parenthesized numeric expression
2263: * or a single signed integer number.
2264: */
2265: static int
1.106 schwarze 2266: roff_evalpar(struct roff *r, int ln,
1.133 schwarze 2267: const char *v, int *pos, int *res, int flags)
1.77 schwarze 2268: {
2269:
2270: if ('(' != v[*pos])
1.133 schwarze 2271: return(roff_getnum(v, pos, res, flags));
1.77 schwarze 2272:
2273: (*pos)++;
1.133 schwarze 2274: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.77 schwarze 2275: return(0);
2276:
1.79 schwarze 2277: /*
2278: * Omission of the closing parenthesis
2279: * is an error in validation mode,
2280: * but ignored in evaluation mode.
2281: */
2282:
1.77 schwarze 2283: if (')' == v[*pos])
2284: (*pos)++;
1.79 schwarze 2285: else if (NULL == res)
2286: return(0);
1.77 schwarze 2287:
2288: return(1);
2289: }
2290:
2291: /*
2292: * Evaluate a complete numeric expression.
2293: * Proceed left to right, there is no concept of precedence.
2294: */
2295: static int
1.106 schwarze 2296: roff_evalnum(struct roff *r, int ln, const char *v,
1.133 schwarze 2297: int *pos, int *res, int flags)
1.77 schwarze 2298: {
2299: int mypos, operand2;
2300: char operator;
2301:
2302: if (NULL == pos) {
2303: mypos = 0;
2304: pos = &mypos;
2305: }
2306:
1.133 schwarze 2307: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2308: while (isspace((unsigned char)v[*pos]))
2309: (*pos)++;
2310:
1.133 schwarze 2311: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.77 schwarze 2312: return(0);
2313:
2314: while (1) {
1.133 schwarze 2315: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2316: while (isspace((unsigned char)v[*pos]))
2317: (*pos)++;
2318:
2319: if ( ! roff_getop(v, pos, &operator))
2320: break;
2321:
1.133 schwarze 2322: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2323: while (isspace((unsigned char)v[*pos]))
2324: (*pos)++;
2325:
1.133 schwarze 2326: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.77 schwarze 2327: return(0);
2328:
1.133 schwarze 2329: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2330: while (isspace((unsigned char)v[*pos]))
2331: (*pos)++;
1.79 schwarze 2332:
2333: if (NULL == res)
2334: continue;
1.77 schwarze 2335:
2336: switch (operator) {
1.80 schwarze 2337: case '+':
1.77 schwarze 2338: *res += operand2;
2339: break;
1.80 schwarze 2340: case '-':
1.77 schwarze 2341: *res -= operand2;
2342: break;
1.80 schwarze 2343: case '*':
1.77 schwarze 2344: *res *= operand2;
2345: break;
1.80 schwarze 2346: case '/':
1.116 schwarze 2347: if (operand2 == 0) {
1.106 schwarze 2348: mandoc_msg(MANDOCERR_DIVZERO,
2349: r->parse, ln, *pos, v);
2350: *res = 0;
2351: break;
2352: }
1.77 schwarze 2353: *res /= operand2;
2354: break;
1.80 schwarze 2355: case '%':
1.116 schwarze 2356: if (operand2 == 0) {
2357: mandoc_msg(MANDOCERR_DIVZERO,
2358: r->parse, ln, *pos, v);
2359: *res = 0;
2360: break;
2361: }
1.77 schwarze 2362: *res %= operand2;
2363: break;
1.80 schwarze 2364: case '<':
1.77 schwarze 2365: *res = *res < operand2;
2366: break;
1.80 schwarze 2367: case '>':
1.77 schwarze 2368: *res = *res > operand2;
2369: break;
1.80 schwarze 2370: case 'l':
1.77 schwarze 2371: *res = *res <= operand2;
2372: break;
1.80 schwarze 2373: case 'g':
1.77 schwarze 2374: *res = *res >= operand2;
2375: break;
1.80 schwarze 2376: case '=':
1.77 schwarze 2377: *res = *res == operand2;
2378: break;
1.80 schwarze 2379: case '!':
1.77 schwarze 2380: *res = *res != operand2;
2381: break;
1.80 schwarze 2382: case '&':
1.77 schwarze 2383: *res = *res && operand2;
2384: break;
1.80 schwarze 2385: case ':':
1.77 schwarze 2386: *res = *res || operand2;
2387: break;
1.80 schwarze 2388: case 'i':
1.77 schwarze 2389: if (operand2 < *res)
2390: *res = operand2;
2391: break;
1.80 schwarze 2392: case 'a':
1.77 schwarze 2393: if (operand2 > *res)
2394: *res = operand2;
2395: break;
2396: default:
2397: abort();
2398: }
2399: }
2400: return(1);
2401: }
2402:
1.138 ! schwarze 2403: /* --- register management ------------------------------------------------ */
! 2404:
1.52 schwarze 2405: void
1.60 schwarze 2406: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 2407: {
1.52 schwarze 2408: struct roffreg *reg;
2409:
2410: /* Search for an existing register with the same name. */
2411: reg = r->regtab;
2412:
2413: while (reg && strcmp(name, reg->key.p))
2414: reg = reg->next;
1.41 schwarze 2415:
1.52 schwarze 2416: if (NULL == reg) {
2417: /* Create a new register. */
2418: reg = mandoc_malloc(sizeof(struct roffreg));
2419: reg->key.p = mandoc_strdup(name);
2420: reg->key.sz = strlen(name);
1.60 schwarze 2421: reg->val = 0;
1.52 schwarze 2422: reg->next = r->regtab;
2423: r->regtab = reg;
2424: }
2425:
1.60 schwarze 2426: if ('+' == sign)
2427: reg->val += val;
2428: else if ('-' == sign)
2429: reg->val -= val;
2430: else
2431: reg->val = val;
1.41 schwarze 2432: }
2433:
1.65 schwarze 2434: /*
2435: * Handle some predefined read-only number registers.
2436: * For now, return -1 if the requested register is not predefined;
2437: * in case a predefined read-only register having the value -1
2438: * were to turn up, another special value would have to be chosen.
2439: */
2440: static int
2441: roff_getregro(const char *name)
2442: {
2443:
2444: switch (*name) {
1.80 schwarze 2445: case 'A': /* ASCII approximation mode is always off. */
1.65 schwarze 2446: return(0);
1.80 schwarze 2447: case 'g': /* Groff compatibility mode is always on. */
1.65 schwarze 2448: return(1);
1.80 schwarze 2449: case 'H': /* Fixed horizontal resolution. */
1.65 schwarze 2450: return (24);
1.80 schwarze 2451: case 'j': /* Always adjust left margin only. */
1.65 schwarze 2452: return(0);
1.80 schwarze 2453: case 'T': /* Some output device is always defined. */
1.65 schwarze 2454: return(1);
1.80 schwarze 2455: case 'V': /* Fixed vertical resolution. */
1.65 schwarze 2456: return (40);
2457: default:
2458: return (-1);
2459: }
2460: }
2461:
1.53 schwarze 2462: int
1.52 schwarze 2463: roff_getreg(const struct roff *r, const char *name)
1.41 schwarze 2464: {
1.52 schwarze 2465: struct roffreg *reg;
1.65 schwarze 2466: int val;
2467:
2468: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
2469: val = roff_getregro(name + 1);
2470: if (-1 != val)
2471: return (val);
2472: }
1.52 schwarze 2473:
2474: for (reg = r->regtab; reg; reg = reg->next)
2475: if (0 == strcmp(name, reg->key.p))
1.53 schwarze 2476: return(reg->val);
2477:
2478: return(0);
2479: }
2480:
2481: static int
2482: roff_getregn(const struct roff *r, const char *name, size_t len)
2483: {
2484: struct roffreg *reg;
1.65 schwarze 2485: int val;
2486:
2487: if ('.' == name[0] && 2 == len) {
2488: val = roff_getregro(name + 1);
2489: if (-1 != val)
2490: return (val);
2491: }
1.53 schwarze 2492:
2493: for (reg = r->regtab; reg; reg = reg->next)
2494: if (len == reg->key.sz &&
2495: 0 == strncmp(name, reg->key.p, len))
2496: return(reg->val);
1.41 schwarze 2497:
1.52 schwarze 2498: return(0);
1.41 schwarze 2499: }
2500:
1.52 schwarze 2501: static void
2502: roff_freereg(struct roffreg *reg)
1.41 schwarze 2503: {
1.52 schwarze 2504: struct roffreg *old_reg;
1.41 schwarze 2505:
1.52 schwarze 2506: while (NULL != reg) {
2507: free(reg->key.p);
2508: old_reg = reg;
2509: reg = reg->next;
2510: free(old_reg);
2511: }
1.41 schwarze 2512: }
1.7 schwarze 2513:
2514: static enum rofferr
1.6 schwarze 2515: roff_nr(ROFF_ARGS)
1.1 schwarze 2516: {
1.85 schwarze 2517: char *key, *val;
2518: size_t keysz;
1.37 schwarze 2519: int iv;
1.60 schwarze 2520: char sign;
1.6 schwarze 2521:
1.110 schwarze 2522: key = val = buf->buf + pos;
2523: if (*key == '\0')
1.85 schwarze 2524: return(ROFF_IGN);
2525:
2526: keysz = roff_getname(r, &val, ln, pos);
1.110 schwarze 2527: if (key[keysz] == '\\')
1.85 schwarze 2528: return(ROFF_IGN);
2529: key[keysz] = '\0';
1.6 schwarze 2530:
1.60 schwarze 2531: sign = *val;
1.110 schwarze 2532: if (sign == '+' || sign == '-')
1.60 schwarze 2533: val++;
2534:
1.133 schwarze 2535: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77 schwarze 2536: roff_setreg(r, key, iv, sign);
1.1 schwarze 2537:
1.76 schwarze 2538: return(ROFF_IGN);
2539: }
2540:
2541: static enum rofferr
2542: roff_rr(ROFF_ARGS)
2543: {
2544: struct roffreg *reg, **prev;
1.85 schwarze 2545: char *name, *cp;
2546: size_t namesz;
1.76 schwarze 2547:
1.110 schwarze 2548: name = cp = buf->buf + pos;
2549: if (*name == '\0')
1.85 schwarze 2550: return(ROFF_IGN);
2551: namesz = roff_getname(r, &cp, ln, pos);
2552: name[namesz] = '\0';
1.76 schwarze 2553:
2554: prev = &r->regtab;
2555: while (1) {
2556: reg = *prev;
1.110 schwarze 2557: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 2558: break;
2559: prev = ®->next;
2560: }
1.110 schwarze 2561: if (reg != NULL) {
1.76 schwarze 2562: *prev = reg->next;
2563: free(reg->key.p);
2564: free(reg);
2565: }
1.29 schwarze 2566: return(ROFF_IGN);
2567: }
2568:
1.138 ! schwarze 2569: /* --- handler functions for roff requests -------------------------------- */
! 2570:
1.29 schwarze 2571: static enum rofferr
2572: roff_rm(ROFF_ARGS)
2573: {
2574: const char *name;
2575: char *cp;
1.85 schwarze 2576: size_t namesz;
1.29 schwarze 2577:
1.110 schwarze 2578: cp = buf->buf + pos;
2579: while (*cp != '\0') {
1.85 schwarze 2580: name = cp;
1.110 schwarze 2581: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 2582: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110 schwarze 2583: if (name[namesz] == '\\')
1.85 schwarze 2584: break;
1.29 schwarze 2585: }
1.51 schwarze 2586: return(ROFF_IGN);
2587: }
2588:
2589: static enum rofferr
2590: roff_it(ROFF_ARGS)
2591: {
2592: int iv;
2593:
2594: /* Parse the number of lines. */
1.133 schwarze 2595:
2596: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95 schwarze 2597: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110 schwarze 2598: ln, ppos, buf->buf + 1);
1.51 schwarze 2599: return(ROFF_IGN);
2600: }
2601:
1.134 schwarze 2602: while (isspace((unsigned char)buf->buf[pos]))
2603: pos++;
2604:
2605: /*
2606: * Arm the input line trap.
2607: * Special-casing "an-trap" is an ugly workaround to cope
2608: * with DocBook stupidly fiddling with man(7) internals.
2609: */
1.133 schwarze 2610:
1.51 schwarze 2611: roffit_lines = iv;
1.134 schwarze 2612: roffit_macro = mandoc_strdup(iv != 1 ||
2613: strcmp(buf->buf + pos, "an-trap") ?
2614: buf->buf + pos : "br");
1.2 schwarze 2615: return(ROFF_IGN);
1.47 schwarze 2616: }
2617:
2618: static enum rofferr
2619: roff_Dd(ROFF_ARGS)
2620: {
2621: const char *const *cp;
2622:
1.99 schwarze 2623: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47 schwarze 2624: for (cp = __mdoc_reserved; *cp; cp++)
2625: roff_setstr(r, *cp, NULL, 0);
2626:
1.99 schwarze 2627: if (r->format == 0)
2628: r->format = MPARSE_MDOC;
2629:
1.47 schwarze 2630: return(ROFF_CONT);
2631: }
2632:
2633: static enum rofferr
2634: roff_TH(ROFF_ARGS)
2635: {
2636: const char *const *cp;
2637:
1.99 schwarze 2638: if ((r->options & MPARSE_QUICK) == 0)
1.47 schwarze 2639: for (cp = __man_reserved; *cp; cp++)
2640: roff_setstr(r, *cp, NULL, 0);
2641:
1.99 schwarze 2642: if (r->format == 0)
2643: r->format = MPARSE_MAN;
2644:
1.47 schwarze 2645: return(ROFF_CONT);
1.14 schwarze 2646: }
2647:
2648: static enum rofferr
1.27 schwarze 2649: roff_TE(ROFF_ARGS)
2650: {
2651:
2652: if (NULL == r->tbl)
1.94 schwarze 2653: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2654: ln, ppos, "TE");
1.130 schwarze 2655: else if ( ! tbl_end(&r->tbl)) {
2656: free(buf->buf);
2657: buf->buf = mandoc_strdup(".sp");
2658: buf->sz = 4;
2659: return(ROFF_REPARSE);
2660: }
1.27 schwarze 2661: return(ROFF_IGN);
2662: }
2663:
2664: static enum rofferr
2665: roff_T_(ROFF_ARGS)
2666: {
2667:
2668: if (NULL == r->tbl)
1.94 schwarze 2669: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2670: ln, ppos, "T&");
1.27 schwarze 2671: else
2672: tbl_restart(ppos, ln, r->tbl);
2673:
2674: return(ROFF_IGN);
2675: }
2676:
1.102 schwarze 2677: /*
2678: * Handle in-line equation delimiters.
2679: */
2680: static enum rofferr
1.110 schwarze 2681: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 2682: {
1.105 schwarze 2683: char *cp1, *cp2;
2684: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 2685:
1.102 schwarze 2686: /*
2687: * Outside equations, look for an opening delimiter.
2688: * If we are inside an equation, we already know it is
2689: * in-line, or this function wouldn't have been called;
2690: * so look for a closing delimiter.
2691: */
2692:
1.110 schwarze 2693: cp1 = buf->buf + pos;
1.102 schwarze 2694: cp2 = strchr(cp1, r->eqn == NULL ?
2695: r->last_eqn->odelim : r->last_eqn->cdelim);
2696: if (cp2 == NULL)
2697: return(ROFF_CONT);
2698:
1.105 schwarze 2699: *cp2++ = '\0';
2700: bef_pr = bef_nl = aft_nl = aft_pr = "";
2701:
2702: /* Handle preceding text, protecting whitespace. */
2703:
1.110 schwarze 2704: if (*buf->buf != '\0') {
1.105 schwarze 2705: if (r->eqn == NULL)
2706: bef_pr = "\\&";
2707: bef_nl = "\n";
2708: }
2709:
2710: /*
2711: * Prepare replacing the delimiter with an equation macro
2712: * and drop leading white space from the equation.
2713: */
1.102 schwarze 2714:
1.105 schwarze 2715: if (r->eqn == NULL) {
2716: while (*cp2 == ' ')
2717: cp2++;
2718: mac = ".EQ";
2719: } else
2720: mac = ".EN";
2721:
2722: /* Handle following text, protecting whitespace. */
2723:
2724: if (*cp2 != '\0') {
2725: aft_nl = "\n";
2726: if (r->eqn != NULL)
2727: aft_pr = "\\&";
2728: }
2729:
2730: /* Do the actual replacement. */
2731:
1.110 schwarze 2732: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 2733: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 2734: free(buf->buf);
2735: buf->buf = cp1;
1.102 schwarze 2736:
2737: /* Toggle the in-line state of the eqn subsystem. */
2738:
2739: r->eqn_inline = r->eqn == NULL;
2740: return(ROFF_REPARSE);
1.41 schwarze 2741: }
2742:
1.107 schwarze 2743: static enum rofferr
2744: roff_EQ(ROFF_ARGS)
1.32 schwarze 2745: {
1.41 schwarze 2746: struct eqn_node *e;
1.32 schwarze 2747:
1.110 schwarze 2748: assert(r->eqn == NULL);
1.107 schwarze 2749: e = eqn_alloc(ppos, ln, r->parse);
1.32 schwarze 2750:
1.102 schwarze 2751: if (r->last_eqn) {
1.32 schwarze 2752: r->last_eqn->next = e;
1.102 schwarze 2753: e->delim = r->last_eqn->delim;
2754: e->odelim = r->last_eqn->odelim;
2755: e->cdelim = r->last_eqn->cdelim;
2756: } else
1.32 schwarze 2757: r->first_eqn = r->last_eqn = e;
2758:
2759: r->eqn = r->last_eqn = e;
1.41 schwarze 2760:
1.110 schwarze 2761: if (buf->buf[pos] != '\0')
1.107 schwarze 2762: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 2763: ".EQ %s", buf->buf + pos);
1.41 schwarze 2764:
1.32 schwarze 2765: return(ROFF_IGN);
2766: }
2767:
2768: static enum rofferr
2769: roff_EN(ROFF_ARGS)
2770: {
2771:
1.94 schwarze 2772: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.32 schwarze 2773: return(ROFF_IGN);
2774: }
2775:
2776: static enum rofferr
1.27 schwarze 2777: roff_TS(ROFF_ARGS)
2778: {
1.49 schwarze 2779: struct tbl_node *tbl;
1.27 schwarze 2780:
2781: if (r->tbl) {
1.94 schwarze 2782: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2783: ln, ppos, "TS breaks TS");
1.41 schwarze 2784: tbl_end(&r->tbl);
1.27 schwarze 2785: }
2786:
1.49 schwarze 2787: tbl = tbl_alloc(ppos, ln, r->parse);
1.27 schwarze 2788:
2789: if (r->last_tbl)
1.49 schwarze 2790: r->last_tbl->next = tbl;
1.27 schwarze 2791: else
1.49 schwarze 2792: r->first_tbl = r->last_tbl = tbl;
1.27 schwarze 2793:
1.49 schwarze 2794: r->tbl = r->last_tbl = tbl;
1.27 schwarze 2795: return(ROFF_IGN);
1.123 schwarze 2796: }
2797:
2798: static enum rofferr
2799: roff_brp(ROFF_ARGS)
2800: {
2801:
2802: buf->buf[pos - 1] = '\0';
2803: return(ROFF_CONT);
1.27 schwarze 2804: }
2805:
2806: static enum rofferr
1.48 schwarze 2807: roff_cc(ROFF_ARGS)
2808: {
2809: const char *p;
2810:
1.110 schwarze 2811: p = buf->buf + pos;
1.48 schwarze 2812:
1.110 schwarze 2813: if (*p == '\0' || (r->control = *p++) == '.')
1.48 schwarze 2814: r->control = 0;
2815:
1.110 schwarze 2816: if (*p != '\0')
1.132 schwarze 2817: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2818: ln, p - buf->buf, "cc ... %s", p);
1.48 schwarze 2819:
2820: return(ROFF_IGN);
2821: }
2822:
2823: static enum rofferr
1.42 schwarze 2824: roff_tr(ROFF_ARGS)
2825: {
2826: const char *p, *first, *second;
2827: size_t fsz, ssz;
2828: enum mandoc_esc esc;
2829:
1.110 schwarze 2830: p = buf->buf + pos;
1.42 schwarze 2831:
1.110 schwarze 2832: if (*p == '\0') {
1.132 schwarze 2833: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.42 schwarze 2834: return(ROFF_IGN);
2835: }
2836:
1.110 schwarze 2837: while (*p != '\0') {
1.42 schwarze 2838: fsz = ssz = 1;
2839:
2840: first = p++;
1.110 schwarze 2841: if (*first == '\\') {
1.42 schwarze 2842: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2843: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2844: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2845: ln, (int)(p - buf->buf), first);
1.42 schwarze 2846: return(ROFF_IGN);
2847: }
2848: fsz = (size_t)(p - first);
2849: }
2850:
2851: second = p++;
1.110 schwarze 2852: if (*second == '\\') {
1.42 schwarze 2853: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 2854: if (esc == ESCAPE_ERROR) {
1.92 schwarze 2855: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 2856: ln, (int)(p - buf->buf), second);
1.42 schwarze 2857: return(ROFF_IGN);
2858: }
2859: ssz = (size_t)(p - second);
1.110 schwarze 2860: } else if (*second == '\0') {
1.132 schwarze 2861: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2862: ln, first - buf->buf, "tr %s", first);
1.42 schwarze 2863: second = " ";
2864: p--;
2865: }
2866:
2867: if (fsz > 1) {
1.80 schwarze 2868: roff_setstrn(&r->xmbtab, first, fsz,
2869: second, ssz, 0);
1.42 schwarze 2870: continue;
2871: }
2872:
1.110 schwarze 2873: if (r->xtab == NULL)
1.80 schwarze 2874: r->xtab = mandoc_calloc(128,
2875: sizeof(struct roffstr));
1.42 schwarze 2876:
2877: free(r->xtab[(int)*first].p);
2878: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2879: r->xtab[(int)*first].sz = ssz;
2880: }
2881:
2882: return(ROFF_IGN);
2883: }
2884:
2885: static enum rofferr
1.14 schwarze 2886: roff_so(ROFF_ARGS)
2887: {
1.121 schwarze 2888: char *name, *cp;
1.15 schwarze 2889:
1.110 schwarze 2890: name = buf->buf + pos;
1.97 schwarze 2891: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14 schwarze 2892:
1.22 schwarze 2893: /*
2894: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2895: * opening anything that's not in our cwd or anything beneath
2896: * it. Thus, explicitly disallow traversing up the file-system
2897: * or using absolute paths.
2898: */
2899:
1.110 schwarze 2900: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83 schwarze 2901: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2902: ".so %s", name);
1.121 schwarze 2903: buf->sz = mandoc_asprintf(&cp,
2904: ".sp\nSee the file %s.\n.sp", name) + 1;
2905: free(buf->buf);
2906: buf->buf = cp;
2907: *offs = 0;
2908: return(ROFF_REPARSE);
1.14 schwarze 2909: }
2910:
2911: *offs = pos;
2912: return(ROFF_SO);
1.7 schwarze 2913: }
2914:
1.138 ! schwarze 2915: /* --- user defined strings and macros ------------------------------------ */
! 2916:
1.16 schwarze 2917: static enum rofferr
2918: roff_userdef(ROFF_ARGS)
1.12 schwarze 2919: {
1.135 schwarze 2920: const char *arg[9], *ap;
1.16 schwarze 2921: char *cp, *n1, *n2;
1.25 schwarze 2922: int i;
1.135 schwarze 2923: size_t asz, rsz;
1.12 schwarze 2924:
1.16 schwarze 2925: /*
2926: * Collect pointers to macro argument strings
1.61 schwarze 2927: * and NUL-terminate them.
1.16 schwarze 2928: */
1.135 schwarze 2929:
1.110 schwarze 2930: cp = buf->buf + pos;
1.25 schwarze 2931: for (i = 0; i < 9; i++)
1.110 schwarze 2932: arg[i] = *cp == '\0' ? "" :
1.35 schwarze 2933: mandoc_getarg(r->parse, &cp, ln, &pos);
1.16 schwarze 2934:
2935: /*
2936: * Expand macro arguments.
1.12 schwarze 2937: */
1.135 schwarze 2938:
2939: buf->sz = strlen(r->current_string) + 1;
2940: n1 = cp = mandoc_malloc(buf->sz);
2941: memcpy(n1, r->current_string, buf->sz);
2942: while (*cp != '\0') {
2943:
2944: /* Scan ahead for the next argument invocation. */
2945:
2946: if (*cp++ != '\\')
2947: continue;
2948: if (*cp++ != '$')
2949: continue;
2950: i = *cp - '1';
2951: if (0 > i || 8 < i)
1.16 schwarze 2952: continue;
1.135 schwarze 2953: cp -= 2;
2954:
2955: /*
2956: * Determine the size of the expanded argument,
2957: * taking escaping of quotes into account.
2958: */
2959:
2960: asz = 0;
2961: for (ap = arg[i]; *ap != '\0'; ap++) {
2962: asz++;
2963: if (*ap == '"')
2964: asz += 3;
2965: }
2966: if (asz != 3) {
2967:
2968: /*
2969: * Determine the size of the rest of the
2970: * unexpanded macro, including the NUL.
2971: */
2972:
2973: rsz = buf->sz - (cp - n1) - 3;
2974:
2975: /*
2976: * When shrinking, move before
2977: * releasing the storage.
2978: */
2979:
2980: if (asz < 3)
2981: memmove(cp + asz, cp + 3, rsz);
2982:
2983: /*
2984: * Resize the storage for the macro
2985: * and readjust the parse pointer.
2986: */
2987:
2988: buf->sz += asz - 3;
2989: n2 = mandoc_realloc(n1, buf->sz);
2990: cp = n2 + (cp - n1);
2991: n1 = n2;
2992:
2993: /*
2994: * When growing, make room
2995: * for the expanded argument.
2996: */
2997:
2998: if (asz > 3)
2999: memmove(cp + asz, cp + 3, rsz);
3000: }
3001:
3002: /* Copy the expanded argument, escaping quotes. */
3003:
3004: n2 = cp;
3005: for (ap = arg[i]; *ap != '\0'; ap++) {
3006: if (*ap == '"') {
3007: memcpy(n2, "\\(dq", 4);
3008: n2 += 4;
3009: } else
3010: *n2++ = *ap;
1.16 schwarze 3011: }
1.12 schwarze 3012: }
3013:
1.16 schwarze 3014: /*
3015: * Replace the macro invocation
3016: * by the expanded macro.
3017: */
1.135 schwarze 3018:
1.110 schwarze 3019: free(buf->buf);
3020: buf->buf = n1;
1.120 schwarze 3021: *offs = 0;
1.16 schwarze 3022:
1.110 schwarze 3023: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.16 schwarze 3024: ROFF_REPARSE : ROFF_APPEND);
1.12 schwarze 3025: }
1.28 schwarze 3026:
1.85 schwarze 3027: static size_t
1.28 schwarze 3028: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3029: {
3030: char *name, *cp;
1.85 schwarze 3031: size_t namesz;
1.28 schwarze 3032:
3033: name = *cpp;
3034: if ('\0' == *name)
1.85 schwarze 3035: return(0);
1.28 schwarze 3036:
1.85 schwarze 3037: /* Read until end of name and terminate it with NUL. */
3038: for (cp = name; 1; cp++) {
3039: if ('\0' == *cp || ' ' == *cp) {
3040: namesz = cp - name;
3041: break;
3042: }
1.28 schwarze 3043: if ('\\' != *cp)
3044: continue;
1.88 schwarze 3045: namesz = cp - name;
3046: if ('{' == cp[1] || '}' == cp[1])
3047: break;
1.28 schwarze 3048: cp++;
3049: if ('\\' == *cp)
3050: continue;
1.97 schwarze 3051: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3052: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 3053: mandoc_escape((const char **)&cp, NULL, NULL);
3054: break;
1.28 schwarze 3055: }
3056:
3057: /* Read past spaces. */
3058: while (' ' == *cp)
3059: cp++;
3060:
3061: *cpp = cp;
1.85 schwarze 3062: return(namesz);
1.28 schwarze 3063: }
3064:
1.16 schwarze 3065: /*
3066: * Store *string into the user-defined string called *name.
3067: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 3068: * append == 0: replace mode
3069: * append == 1: single-line append mode
3070: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 3071: */
1.8 schwarze 3072: static void
1.16 schwarze 3073: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 3074: int append)
1.7 schwarze 3075: {
1.42 schwarze 3076:
3077: roff_setstrn(&r->strtab, name, strlen(name), string,
1.80 schwarze 3078: string ? strlen(string) : 0, append);
1.42 schwarze 3079: }
3080:
3081: static void
3082: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 3083: const char *string, size_t stringsz, int append)
1.42 schwarze 3084: {
3085: struct roffkv *n;
3086: char *c;
3087: int i;
3088: size_t oldch, newch;
1.7 schwarze 3089:
1.16 schwarze 3090: /* Search for an existing string with the same name. */
1.42 schwarze 3091: n = *r;
3092:
1.84 schwarze 3093: while (n && (namesz != n->key.sz ||
3094: strncmp(n->key.p, name, namesz)))
1.7 schwarze 3095: n = n->next;
1.8 schwarze 3096:
3097: if (NULL == n) {
1.16 schwarze 3098: /* Create a new string table entry. */
1.42 schwarze 3099: n = mandoc_malloc(sizeof(struct roffkv));
3100: n->key.p = mandoc_strndup(name, namesz);
3101: n->key.sz = namesz;
3102: n->val.p = NULL;
3103: n->val.sz = 0;
3104: n->next = *r;
3105: *r = n;
1.66 schwarze 3106: } else if (0 == append) {
1.42 schwarze 3107: free(n->val.p);
3108: n->val.p = NULL;
3109: n->val.sz = 0;
1.16 schwarze 3110: }
3111:
3112: if (NULL == string)
3113: return;
3114:
3115: /*
3116: * One additional byte for the '\n' in multiline mode,
3117: * and one for the terminating '\0'.
3118: */
1.66 schwarze 3119: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 3120:
3121: if (NULL == n->val.p) {
3122: n->val.p = mandoc_malloc(newch);
3123: *n->val.p = '\0';
1.16 schwarze 3124: oldch = 0;
3125: } else {
1.42 schwarze 3126: oldch = n->val.sz;
3127: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 3128: }
3129:
3130: /* Skip existing content in the destination buffer. */
1.42 schwarze 3131: c = n->val.p + (int)oldch;
1.16 schwarze 3132:
3133: /* Append new content to the destination buffer. */
1.42 schwarze 3134: i = 0;
3135: while (i < (int)stringsz) {
1.16 schwarze 3136: /*
3137: * Rudimentary roff copy mode:
3138: * Handle escaped backslashes.
3139: */
1.42 schwarze 3140: if ('\\' == string[i] && '\\' == string[i + 1])
3141: i++;
3142: *c++ = string[i++];
1.16 schwarze 3143: }
1.8 schwarze 3144:
1.16 schwarze 3145: /* Append terminating bytes. */
1.66 schwarze 3146: if (1 < append)
1.16 schwarze 3147: *c++ = '\n';
1.42 schwarze 3148:
1.16 schwarze 3149: *c = '\0';
1.42 schwarze 3150: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 3151: }
3152:
1.8 schwarze 3153: static const char *
3154: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7 schwarze 3155: {
1.42 schwarze 3156: const struct roffkv *n;
1.64 schwarze 3157: int i;
1.7 schwarze 3158:
1.42 schwarze 3159: for (n = r->strtab; n; n = n->next)
1.80 schwarze 3160: if (0 == strncmp(name, n->key.p, len) &&
3161: '\0' == n->key.p[(int)len])
1.42 schwarze 3162: return(n->val.p);
1.64 schwarze 3163:
3164: for (i = 0; i < PREDEFS_MAX; i++)
3165: if (0 == strncmp(name, predefs[i].name, len) &&
3166: '\0' == predefs[i].name[(int)len])
3167: return(predefs[i].str);
1.8 schwarze 3168:
1.42 schwarze 3169: return(NULL);
1.7 schwarze 3170: }
3171:
1.8 schwarze 3172: static void
1.42 schwarze 3173: roff_freestr(struct roffkv *r)
1.7 schwarze 3174: {
1.42 schwarze 3175: struct roffkv *n, *nn;
1.7 schwarze 3176:
1.42 schwarze 3177: for (n = r; n; n = nn) {
3178: free(n->key.p);
3179: free(n->val.p);
1.7 schwarze 3180: nn = n->next;
3181: free(n);
3182: }
1.27 schwarze 3183: }
1.138 ! schwarze 3184:
! 3185: /* --- accessors and utility functions ------------------------------------ */
1.27 schwarze 3186:
3187: const struct tbl_span *
3188: roff_span(const struct roff *r)
3189: {
1.80 schwarze 3190:
1.27 schwarze 3191: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32 schwarze 3192: }
3193:
3194: const struct eqn *
3195: roff_eqn(const struct roff *r)
3196: {
1.80 schwarze 3197:
1.32 schwarze 3198: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.42 schwarze 3199: }
3200:
3201: /*
3202: * Duplicate an input string, making the appropriate character
3203: * conversations (as stipulated by `tr') along the way.
3204: * Returns a heap-allocated string with all the replacements made.
3205: */
3206: char *
3207: roff_strdup(const struct roff *r, const char *p)
3208: {
3209: const struct roffkv *cp;
3210: char *res;
3211: const char *pp;
3212: size_t ssz, sz;
3213: enum mandoc_esc esc;
3214:
3215: if (NULL == r->xmbtab && NULL == r->xtab)
3216: return(mandoc_strdup(p));
3217: else if ('\0' == *p)
3218: return(mandoc_strdup(""));
3219:
3220: /*
3221: * Step through each character looking for term matches
3222: * (remember that a `tr' can be invoked with an escape, which is
3223: * a glyph but the escape is multi-character).
3224: * We only do this if the character hash has been initialised
3225: * and the string is >0 length.
3226: */
3227:
3228: res = NULL;
3229: ssz = 0;
3230:
3231: while ('\0' != *p) {
3232: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3233: sz = r->xtab[(int)*p].sz;
3234: res = mandoc_realloc(res, ssz + sz + 1);
3235: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3236: ssz += sz;
3237: p++;
3238: continue;
3239: } else if ('\\' != *p) {
3240: res = mandoc_realloc(res, ssz + 2);
3241: res[ssz++] = *p++;
3242: continue;
3243: }
3244:
3245: /* Search for term matches. */
3246: for (cp = r->xmbtab; cp; cp = cp->next)
3247: if (0 == strncmp(p, cp->key.p, cp->key.sz))
3248: break;
3249:
3250: if (NULL != cp) {
3251: /*
3252: * A match has been found.
3253: * Append the match to the array and move
3254: * forward by its keysize.
3255: */
1.80 schwarze 3256: res = mandoc_realloc(res,
3257: ssz + cp->val.sz + 1);
1.42 schwarze 3258: memcpy(res + ssz, cp->val.p, cp->val.sz);
3259: ssz += cp->val.sz;
3260: p += (int)cp->key.sz;
3261: continue;
3262: }
3263:
3264: /*
3265: * Handle escapes carefully: we need to copy
3266: * over just the escape itself, or else we might
3267: * do replacements within the escape itself.
3268: * Make sure to pass along the bogus string.
3269: */
3270: pp = p++;
3271: esc = mandoc_escape(&p, NULL, NULL);
3272: if (ESCAPE_ERROR == esc) {
3273: sz = strlen(pp);
3274: res = mandoc_realloc(res, ssz + sz + 1);
3275: memcpy(res + ssz, pp, sz);
3276: break;
3277: }
1.80 schwarze 3278: /*
3279: * We bail out on bad escapes.
1.42 schwarze 3280: * No need to warn: we already did so when
3281: * roff_res() was called.
3282: */
3283: sz = (int)(p - pp);
3284: res = mandoc_realloc(res, ssz + sz + 1);
3285: memcpy(res + ssz, pp, sz);
3286: ssz += sz;
3287: }
3288:
3289: res[(int)ssz] = '\0';
3290: return(res);
1.99 schwarze 3291: }
3292:
3293: int
3294: roff_getformat(const struct roff *r)
3295: {
3296:
3297: return(r->format);
1.48 schwarze 3298: }
3299:
3300: /*
1.80 schwarze 3301: * Find out whether a line is a macro line or not.
1.48 schwarze 3302: * If it is, adjust the current position and return one; if it isn't,
3303: * return zero and don't change the current position.
3304: * If the control character has been set with `.cc', then let that grain
3305: * precedence.
3306: * This is slighly contrary to groff, where using the non-breaking
3307: * control character when `cc' has been invoked will cause the
3308: * non-breaking macro contents to be printed verbatim.
3309: */
3310: int
3311: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3312: {
3313: int pos;
3314:
3315: pos = *ppos;
3316:
3317: if (0 != r->control && cp[pos] == r->control)
3318: pos++;
3319: else if (0 != r->control)
3320: return(0);
3321: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3322: pos += 2;
3323: else if ('.' == cp[pos] || '\'' == cp[pos])
3324: pos++;
3325: else
3326: return(0);
3327:
3328: while (' ' == cp[pos] || '\t' == cp[pos])
3329: pos++;
3330:
3331: *ppos = pos;
3332: return(1);
1.1 schwarze 3333: }
![[BACK]](/icons/back.gif){kind=icon}
Return to roff.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / mandoc