Annotation of mandoc/roff.c, Revision 1.265
1.265 ! schwarze 1: /* $Id: roff.c,v 1.264 2015/04/04 13:53:42 schwarze Exp $ */
1.1 kristaps 2: /*
1.265 ! schwarze 3: * Copyright (c) 2009-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.247 schwarze 4: * Copyright (c) 2010-2015 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.66 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 kristaps 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66 kristaps 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.66 kristaps 18: #include "config.h"
1.225 schwarze 19:
20: #include <sys/types.h>
1.30 kristaps 21:
1.67 kristaps 22: #include <assert.h>
1.85 kristaps 23: #include <ctype.h>
1.245 schwarze 24: #include <limits.h>
1.178 schwarze 25: #include <stdio.h>
1.1 kristaps 26: #include <stdlib.h>
1.67 kristaps 27: #include <string.h>
1.1 kristaps 28:
1.67 kristaps 29: #include "mandoc.h"
1.201 schwarze 30: #include "mandoc_aux.h"
1.265 ! schwarze 31: #include "roff.h"
1.226 schwarze 32: #include "libmandoc.h"
1.109 kristaps 33: #include "libroff.h"
1.33 kristaps 34:
1.141 kristaps 35: /* Maximum number of nested if-else conditionals. */
1.82 kristaps 36: #define RSTACK_MAX 128
37:
1.170 schwarze 38: /* Maximum number of string expansions per line, to break infinite loops. */
39: #define EXPAND_LIMIT 1000
40:
1.67 kristaps 41: enum rofft {
1.251 schwarze 42: ROFF_ab,
1.103 kristaps 43: ROFF_ad,
1.251 schwarze 44: ROFF_af,
45: ROFF_aln,
46: ROFF_als,
1.80 kristaps 47: ROFF_am,
1.251 schwarze 48: ROFF_am1,
1.80 kristaps 49: ROFF_ami,
1.251 schwarze 50: ROFF_ami1,
1.193 schwarze 51: ROFF_as,
1.251 schwarze 52: ROFF_as1,
53: ROFF_asciify,
54: ROFF_backtrace,
55: ROFF_bd,
56: ROFF_bleedat,
57: ROFF_blm,
58: ROFF_box,
59: ROFF_boxa,
60: ROFF_bp,
61: ROFF_BP,
62: /* MAN_br, MDOC_br */
63: ROFF_break,
64: ROFF_breakchar,
65: ROFF_brnl,
66: ROFF_brp,
67: ROFF_brpnl,
68: ROFF_c2,
1.174 kristaps 69: ROFF_cc,
1.194 schwarze 70: ROFF_ce,
1.251 schwarze 71: ROFF_cf,
72: ROFF_cflags,
73: ROFF_ch,
74: ROFF_char,
75: ROFF_chop,
76: ROFF_class,
77: ROFF_close,
78: ROFF_CL,
79: ROFF_color,
80: ROFF_composite,
81: ROFF_continue,
82: ROFF_cp,
83: ROFF_cropat,
84: ROFF_cs,
85: ROFF_cu,
86: ROFF_da,
87: ROFF_dch,
88: ROFF_Dd,
1.80 kristaps 89: ROFF_de,
1.251 schwarze 90: ROFF_de1,
91: ROFF_defcolor,
1.80 kristaps 92: ROFF_dei,
1.251 schwarze 93: ROFF_dei1,
94: ROFF_device,
95: ROFF_devicem,
96: ROFF_di,
97: ROFF_do,
1.83 schwarze 98: ROFF_ds,
1.251 schwarze 99: ROFF_ds1,
100: ROFF_dwh,
101: ROFF_dt,
102: ROFF_ec,
103: ROFF_ecr,
104: ROFF_ecs,
1.82 kristaps 105: ROFF_el,
1.251 schwarze 106: ROFF_em,
107: ROFF_EN,
108: ROFF_eo,
109: ROFF_EP,
110: ROFF_EQ,
111: ROFF_errprint,
112: ROFF_ev,
113: ROFF_evc,
114: ROFF_ex,
115: ROFF_fallback,
1.185 schwarze 116: ROFF_fam,
1.251 schwarze 117: ROFF_fc,
118: ROFF_fchar,
119: ROFF_fcolor,
120: ROFF_fdeferlig,
121: ROFF_feature,
122: /* MAN_fi; ignored in mdoc(7) */
123: ROFF_fkern,
124: ROFF_fl,
125: ROFF_flig,
126: ROFF_fp,
127: ROFF_fps,
128: ROFF_fschar,
129: ROFF_fspacewidth,
130: ROFF_fspecial,
131: /* MAN_ft; ignored in mdoc(7) */
132: ROFF_ftr,
133: ROFF_fzoom,
134: ROFF_gcolor,
135: ROFF_hc,
136: ROFF_hcode,
137: ROFF_hidechar,
138: ROFF_hla,
139: ROFF_hlm,
140: ROFF_hpf,
141: ROFF_hpfa,
142: ROFF_hpfcode,
1.186 schwarze 143: ROFF_hw,
1.103 kristaps 144: ROFF_hy,
1.251 schwarze 145: ROFF_hylang,
146: ROFF_hylen,
147: ROFF_hym,
148: ROFF_hypp,
149: ROFF_hys,
1.82 kristaps 150: ROFF_ie,
1.75 kristaps 151: ROFF_if,
1.76 kristaps 152: ROFF_ig,
1.251 schwarze 153: /* MAN_in; ignored in mdoc(7) */
154: ROFF_index,
1.123 schwarze 155: ROFF_it,
1.251 schwarze 156: ROFF_itc,
157: ROFF_IX,
158: ROFF_kern,
159: ROFF_kernafter,
160: ROFF_kernbefore,
161: ROFF_kernpair,
162: ROFF_lc,
163: ROFF_lc_ctype,
164: ROFF_lds,
165: ROFF_length,
166: ROFF_letadj,
167: ROFF_lf,
168: ROFF_lg,
169: ROFF_lhang,
170: ROFF_linetabs,
171: /* MAN_ll, MDOC_ll */
172: ROFF_lnr,
173: ROFF_lnrf,
174: ROFF_lpfx,
175: ROFF_ls,
176: ROFF_lsm,
177: ROFF_lt,
178: ROFF_mc,
179: ROFF_mediasize,
180: ROFF_minss,
181: ROFF_mk,
182: ROFF_mso,
1.256 schwarze 183: ROFF_na,
1.103 kristaps 184: ROFF_ne,
1.251 schwarze 185: /* MAN_nf; ignored in mdoc(7) */
1.103 kristaps 186: ROFF_nh,
1.251 schwarze 187: ROFF_nhychar,
188: ROFF_nm,
189: ROFF_nn,
190: ROFF_nop,
1.104 kristaps 191: ROFF_nr,
1.251 schwarze 192: ROFF_nrf,
193: ROFF_nroff,
1.124 schwarze 194: ROFF_ns,
1.251 schwarze 195: ROFF_nx,
196: ROFF_open,
197: ROFF_opena,
198: ROFF_os,
199: ROFF_output,
200: ROFF_padj,
201: ROFF_papersize,
202: ROFF_pc,
203: ROFF_pev,
204: ROFF_pi,
205: ROFF_PI,
1.229 schwarze 206: ROFF_pl,
1.251 schwarze 207: ROFF_pm,
208: ROFF_pn,
209: ROFF_pnr,
210: ROFF_po,
1.124 schwarze 211: ROFF_ps,
1.251 schwarze 212: ROFF_psbb,
213: ROFF_pshape,
214: ROFF_pso,
215: ROFF_ptr,
216: ROFF_pvs,
217: ROFF_rchar,
218: ROFF_rd,
219: ROFF_recursionlimit,
220: ROFF_return,
221: ROFF_rfschar,
222: ROFF_rhang,
223: ROFF_rj,
1.83 schwarze 224: ROFF_rm,
1.251 schwarze 225: ROFF_rn,
226: ROFF_rnn,
1.203 schwarze 227: ROFF_rr,
1.251 schwarze 228: ROFF_rs,
229: ROFF_rt,
230: ROFF_schar,
231: ROFF_sentchar,
232: ROFF_shc,
233: ROFF_shift,
234: ROFF_sizes,
1.105 kristaps 235: ROFF_so,
1.251 schwarze 236: /* MAN_sp, MDOC_sp */
237: ROFF_spacewidth,
238: ROFF_special,
239: ROFF_spreadwarn,
240: ROFF_ss,
241: ROFF_sty,
242: ROFF_substring,
243: ROFF_sv,
244: ROFF_sy,
245: ROFF_T_,
1.124 schwarze 246: ROFF_ta,
1.251 schwarze 247: ROFF_tc,
248: ROFF_TE,
249: ROFF_TH,
250: ROFF_ti,
251: ROFF_tkf,
252: ROFF_tl,
253: ROFF_tm,
254: ROFF_tm1,
255: ROFF_tmc,
1.83 schwarze 256: ROFF_tr,
1.251 schwarze 257: ROFF_track,
258: ROFF_transchar,
259: ROFF_trf,
260: ROFF_trimat,
261: ROFF_trin,
262: ROFF_trnt,
263: ROFF_troff,
1.109 kristaps 264: ROFF_TS,
1.251 schwarze 265: ROFF_uf,
266: ROFF_ul,
267: ROFF_unformat,
268: ROFF_unwatch,
269: ROFF_unwatchn,
270: ROFF_vpt,
271: ROFF_vs,
272: ROFF_warn,
273: ROFF_warnscale,
274: ROFF_watch,
275: ROFF_watchlength,
276: ROFF_watchn,
277: ROFF_wh,
278: ROFF_while,
279: ROFF_write,
280: ROFF_writec,
281: ROFF_writem,
282: ROFF_xflag,
1.76 kristaps 283: ROFF_cblock,
1.106 kristaps 284: ROFF_USERDEF,
1.67 kristaps 285: ROFF_MAX
286: };
287:
1.147 kristaps 288: /*
1.167 kristaps 289: * An incredibly-simple string buffer.
290: */
1.94 kristaps 291: struct roffstr {
1.167 kristaps 292: char *p; /* nil-terminated buffer */
293: size_t sz; /* saved strlen(p) */
1.166 kristaps 294: };
295:
296: /*
1.167 kristaps 297: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 298: */
299: struct roffkv {
300: struct roffstr key;
301: struct roffstr val;
302: struct roffkv *next; /* next in list */
1.94 kristaps 303: };
304:
1.180 schwarze 305: /*
306: * A single number register as part of a singly-linked list.
307: */
308: struct roffreg {
309: struct roffstr key;
1.181 schwarze 310: int val;
1.180 schwarze 311: struct roffreg *next;
312: };
313:
1.67 kristaps 314: struct roff {
1.128 kristaps 315: struct mparse *parse; /* parse point */
1.237 schwarze 316: const struct mchars *mchars; /* character table */
1.67 kristaps 317: struct roffnode *last; /* leaf of stack */
1.223 schwarze 318: int *rstack; /* stack of inverted `ie' values */
1.180 schwarze 319: struct roffreg *regtab; /* number registers */
1.166 kristaps 320: struct roffkv *strtab; /* user-defined strings & macros */
1.167 kristaps 321: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
322: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 323: const char *current_string; /* value of last called user macro */
1.118 kristaps 324: struct tbl_node *first_tbl; /* first table parsed */
325: struct tbl_node *last_tbl; /* last table parsed */
326: struct tbl_node *tbl; /* current table being parsed */
1.125 kristaps 327: struct eqn_node *last_eqn; /* last equation parsed */
328: struct eqn_node *first_eqn; /* first equation parsed */
329: struct eqn_node *eqn; /* current equation being parsed */
1.230 schwarze 330: int eqn_inline; /* current equation is inline */
1.223 schwarze 331: int options; /* parse options */
332: int rstacksz; /* current size limit of rstack */
333: int rstackpos; /* position in rstack */
1.227 schwarze 334: int format; /* current file in mdoc or man format */
1.223 schwarze 335: char control; /* control character */
1.79 kristaps 336: };
337:
1.67 kristaps 338: struct roffnode {
339: enum rofft tok; /* type of node */
340: struct roffnode *parent; /* up one in stack */
341: int line; /* parse line */
342: int col; /* parse col */
1.106 kristaps 343: char *name; /* node name, e.g. macro name */
1.79 kristaps 344: char *end; /* end-rules: custom token */
345: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 346: int rule; /* current evaluation rule */
1.67 kristaps 347: };
348:
349: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.72 kristaps 350: enum rofft tok, /* tok of macro */ \
1.238 schwarze 351: struct buf *buf, /* input buffer */ \
1.67 kristaps 352: int ln, /* parse line */ \
1.75 kristaps 353: int ppos, /* original pos in buffer */ \
354: int pos, /* current pos in buffer */ \
1.74 kristaps 355: int *offs /* reset offset of buffer data */
1.67 kristaps 356:
357: typedef enum rofferr (*roffproc)(ROFF_ARGS);
358:
359: struct roffmac {
360: const char *name; /* macro name */
1.79 kristaps 361: roffproc proc; /* process new macro */
362: roffproc text; /* process as child text of macro */
363: roffproc sub; /* process as child of macro */
364: int flags;
365: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.85 kristaps 366: struct roffmac *next;
1.67 kristaps 367: };
368:
1.141 kristaps 369: struct predef {
370: const char *name; /* predefined input name */
371: const char *str; /* replacement symbol */
372: };
373:
374: #define PREDEF(__name, __str) \
375: { (__name), (__str) },
376:
1.155 kristaps 377: static enum rofft roffhash_find(const char *, size_t);
378: static void roffhash_init(void);
379: static void roffnode_cleanscope(struct roff *);
380: static void roffnode_pop(struct roff *);
381: static void roffnode_push(struct roff *, enum rofft,
382: const char *, int, int);
1.80 kristaps 383: static enum rofferr roff_block(ROFF_ARGS);
384: static enum rofferr roff_block_text(ROFF_ARGS);
385: static enum rofferr roff_block_sub(ROFF_ARGS);
1.251 schwarze 386: static enum rofferr roff_brp(ROFF_ARGS);
1.80 kristaps 387: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 388: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 389: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 390: static enum rofferr roff_cond(ROFF_ARGS);
391: static enum rofferr roff_cond_text(ROFF_ARGS);
392: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 393: static enum rofferr roff_ds(ROFF_ARGS);
1.238 schwarze 394: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.234 kristaps 395: static int roff_evalcond(struct roff *r, int,
396: const char *, int *);
397: static int roff_evalnum(struct roff *, int,
398: const char *, int *, int *, int);
399: static int roff_evalpar(struct roff *, int,
1.261 schwarze 400: const char *, int *, int *, int);
1.198 schwarze 401: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 402: static void roff_free1(struct roff *);
1.180 schwarze 403: static void roff_freereg(struct roffreg *);
1.167 kristaps 404: static void roff_freestr(struct roffkv *);
1.212 schwarze 405: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 406: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 407: static int roff_getop(const char *, int *, char *);
1.181 schwarze 408: static int roff_getregn(const struct roff *,
409: const char *, size_t);
1.192 schwarze 410: static int roff_getregro(const char *name);
1.207 schwarze 411: static const char *roff_getstrn(const struct roff *,
1.94 kristaps 412: const char *, size_t);
1.251 schwarze 413: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 414: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 415: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 ! schwarze 416: static void roff_man_alloc1(struct roff_man *);
! 417: static void roff_man_free1(struct roff_man *);
1.89 kristaps 418: static enum rofferr roff_nr(ROFF_ARGS);
1.214 schwarze 419: static enum rofft roff_parse(struct roff *, char *, int *,
420: int, int);
1.238 schwarze 421: static enum rofferr roff_parsetext(struct buf *, int, int *);
422: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 423: static enum rofferr roff_rm(ROFF_ARGS);
1.203 schwarze 424: static enum rofferr roff_rr(ROFF_ARGS);
1.94 kristaps 425: static void roff_setstr(struct roff *,
1.106 kristaps 426: const char *, const char *, int);
1.207 schwarze 427: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 428: size_t, const char *, size_t, int);
1.105 kristaps 429: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 430: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 431: static enum rofferr roff_Dd(ROFF_ARGS);
432: static enum rofferr roff_TH(ROFF_ARGS);
1.109 kristaps 433: static enum rofferr roff_TE(ROFF_ARGS);
434: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 435: static enum rofferr roff_EQ(ROFF_ARGS);
436: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 437: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 438: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 439: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 440:
1.155 kristaps 441: /* See roffhash_find() */
1.85 kristaps 442:
443: #define ASCII_HI 126
444: #define ASCII_LO 33
445: #define HASHWIDTH (ASCII_HI - ASCII_LO + 1)
446:
1.261 schwarze 447: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
448: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
449:
1.85 kristaps 450: static struct roffmac *hash[HASHWIDTH];
451:
452: static struct roffmac roffs[ROFF_MAX] = {
1.251 schwarze 453: { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.103 kristaps 454: { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 455: { "af", roff_line_ignore, NULL, NULL, 0, NULL },
456: { "aln", roff_unsupp, NULL, NULL, 0, NULL },
457: { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 458: { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 459: { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.85 kristaps 460: { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 461: { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.193 schwarze 462: { "as", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 463: { "as1", roff_ds, NULL, NULL, 0, NULL },
464: { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
465: { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
466: { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
467: { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
468: { "blm", roff_unsupp, NULL, NULL, 0, NULL },
469: { "box", roff_unsupp, NULL, NULL, 0, NULL },
470: { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
471: { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
472: { "BP", roff_unsupp, NULL, NULL, 0, NULL },
473: { "break", roff_unsupp, NULL, NULL, 0, NULL },
474: { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
475: { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
476: { "brp", roff_brp, NULL, NULL, 0, NULL },
477: { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
478: { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.174 kristaps 479: { "cc", roff_cc, NULL, NULL, 0, NULL },
1.194 schwarze 480: { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 481: { "cf", roff_insec, NULL, NULL, 0, NULL },
482: { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
483: { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
484: { "char", roff_unsupp, NULL, NULL, 0, NULL },
485: { "chop", roff_unsupp, NULL, NULL, 0, NULL },
486: { "class", roff_line_ignore, NULL, NULL, 0, NULL },
487: { "close", roff_insec, NULL, NULL, 0, NULL },
488: { "CL", roff_unsupp, NULL, NULL, 0, NULL },
489: { "color", roff_line_ignore, NULL, NULL, 0, NULL },
490: { "composite", roff_unsupp, NULL, NULL, 0, NULL },
491: { "continue", roff_unsupp, NULL, NULL, 0, NULL },
492: { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
493: { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
494: { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
495: { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
496: { "da", roff_unsupp, NULL, NULL, 0, NULL },
497: { "dch", roff_unsupp, NULL, NULL, 0, NULL },
498: { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.85 kristaps 499: { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 500: { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
501: { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 502: { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 503: { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
504: { "device", roff_unsupp, NULL, NULL, 0, NULL },
505: { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
506: { "di", roff_unsupp, NULL, NULL, 0, NULL },
507: { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.92 schwarze 508: { "ds", roff_ds, NULL, NULL, 0, NULL },
1.251 schwarze 509: { "ds1", roff_ds, NULL, NULL, 0, NULL },
510: { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
511: { "dt", roff_unsupp, NULL, NULL, 0, NULL },
512: { "ec", roff_unsupp, NULL, NULL, 0, NULL },
513: { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
514: { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.85 kristaps 515: { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.251 schwarze 516: { "em", roff_unsupp, NULL, NULL, 0, NULL },
517: { "EN", roff_EN, NULL, NULL, 0, NULL },
518: { "eo", roff_unsupp, NULL, NULL, 0, NULL },
519: { "EP", roff_unsupp, NULL, NULL, 0, NULL },
520: { "EQ", roff_EQ, NULL, NULL, 0, NULL },
521: { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
522: { "ev", roff_unsupp, NULL, NULL, 0, NULL },
523: { "evc", roff_unsupp, NULL, NULL, 0, NULL },
524: { "ex", roff_unsupp, NULL, NULL, 0, NULL },
525: { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.185 schwarze 526: { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 527: { "fc", roff_unsupp, NULL, NULL, 0, NULL },
528: { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
529: { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
530: { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
531: { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
532: { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
533: { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
534: { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
535: { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
536: { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
537: { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
538: { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
539: { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
540: { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
541: { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
542: { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
543: { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
544: { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
545: { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
546: { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
547: { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
548: { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
549: { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
550: { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.186 schwarze 551: { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 552: { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 553: { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
554: { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
555: { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
556: { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
557: { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 558: { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
559: { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
560: { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.251 schwarze 561: { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.178 schwarze 562: { "it", roff_it, NULL, NULL, 0, NULL },
1.251 schwarze 563: { "itc", roff_unsupp, NULL, NULL, 0, NULL },
564: { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
565: { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
566: { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
567: { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
568: { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
569: { "lc", roff_unsupp, NULL, NULL, 0, NULL },
570: { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
571: { "lds", roff_unsupp, NULL, NULL, 0, NULL },
572: { "length", roff_unsupp, NULL, NULL, 0, NULL },
573: { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
574: { "lf", roff_insec, NULL, NULL, 0, NULL },
575: { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
576: { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
577: { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
578: { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
579: { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
580: { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
581: { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
582: { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
583: { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
584: { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
585: { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
586: { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
587: { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
588: { "mso", roff_insec, NULL, NULL, 0, NULL },
1.256 schwarze 589: { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.103 kristaps 590: { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
591: { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 592: { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
593: { "nm", roff_unsupp, NULL, NULL, 0, NULL },
594: { "nn", roff_unsupp, NULL, NULL, 0, NULL },
595: { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.104 kristaps 596: { "nr", roff_nr, NULL, NULL, 0, NULL },
1.251 schwarze 597: { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
598: { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 599: { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 600: { "nx", roff_insec, NULL, NULL, 0, NULL },
601: { "open", roff_insec, NULL, NULL, 0, NULL },
602: { "opena", roff_insec, NULL, NULL, 0, NULL },
603: { "os", roff_line_ignore, NULL, NULL, 0, NULL },
604: { "output", roff_unsupp, NULL, NULL, 0, NULL },
605: { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
606: { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
607: { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
608: { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
609: { "pi", roff_insec, NULL, NULL, 0, NULL },
610: { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.229 schwarze 611: { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 612: { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
613: { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
614: { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
615: { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.124 schwarze 616: { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.251 schwarze 617: { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.255 schwarze 618: { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 619: { "pso", roff_insec, NULL, NULL, 0, NULL },
620: { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
621: { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
622: { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
623: { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
624: { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
625: { "return", roff_unsupp, NULL, NULL, 0, NULL },
626: { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
627: { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
628: { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.122 schwarze 629: { "rm", roff_rm, NULL, NULL, 0, NULL },
1.251 schwarze 630: { "rn", roff_unsupp, NULL, NULL, 0, NULL },
631: { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.203 schwarze 632: { "rr", roff_rr, NULL, NULL, 0, NULL },
1.251 schwarze 633: { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
634: { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
635: { "schar", roff_unsupp, NULL, NULL, 0, NULL },
636: { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
637: { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
638: { "shift", roff_unsupp, NULL, NULL, 0, NULL },
639: { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.105 kristaps 640: { "so", roff_so, NULL, NULL, 0, NULL },
1.251 schwarze 641: { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
642: { "special", roff_line_ignore, NULL, NULL, 0, NULL },
643: { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
644: { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
645: { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
646: { "substring", roff_unsupp, NULL, NULL, 0, NULL },
647: { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
648: { "sy", roff_insec, NULL, NULL, 0, NULL },
649: { "T&", roff_T_, NULL, NULL, 0, NULL },
1.255 schwarze 650: { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 651: { "tc", roff_unsupp, NULL, NULL, 0, NULL },
652: { "TE", roff_TE, NULL, NULL, 0, NULL },
653: { "TH", roff_TH, NULL, NULL, 0, NULL },
1.255 schwarze 654: { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.251 schwarze 655: { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
656: { "tl", roff_unsupp, NULL, NULL, 0, NULL },
657: { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
658: { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
659: { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.164 kristaps 660: { "tr", roff_tr, NULL, NULL, 0, NULL },
1.251 schwarze 661: { "track", roff_line_ignore, NULL, NULL, 0, NULL },
662: { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
663: { "trf", roff_insec, NULL, NULL, 0, NULL },
664: { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
665: { "trin", roff_unsupp, NULL, NULL, 0, NULL },
666: { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
667: { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.109 kristaps 668: { "TS", roff_TS, NULL, NULL, 0, NULL },
1.251 schwarze 669: { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
670: { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
671: { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
672: { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
673: { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
674: { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
675: { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
676: { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
677: { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
678: { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
679: { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
680: { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
681: { "wh", roff_unsupp, NULL, NULL, 0, NULL },
682: { "while", roff_unsupp, NULL, NULL, 0, NULL },
683: { "write", roff_insec, NULL, NULL, 0, NULL },
684: { "writec", roff_insec, NULL, NULL, 0, NULL },
685: { "writem", roff_insec, NULL, NULL, 0, NULL },
686: { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.85 kristaps 687: { ".", roff_cblock, NULL, NULL, 0, NULL },
1.106 kristaps 688: { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.67 kristaps 689: };
690:
1.200 schwarze 691: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175 schwarze 692: const char *const __mdoc_reserved[] = {
693: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
694: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
695: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
696: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200 schwarze 697: "Dt", "Dv", "Dx", "D1",
698: "Ec", "Ed", "Ef", "Ek", "El", "Em",
699: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175 schwarze 700: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200 schwarze 701: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
702: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175 schwarze 703: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200 schwarze 704: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
705: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
706: "Sc", "Sh", "Sm", "So", "Sq",
1.175 schwarze 707: "Ss", "St", "Sx", "Sy",
708: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200 schwarze 709: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175 schwarze 710: "%P", "%Q", "%R", "%T", "%U", "%V",
711: NULL
712: };
713:
1.200 schwarze 714: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175 schwarze 715: const char *const __man_reserved[] = {
1.200 schwarze 716: "AT", "B", "BI", "BR", "DT",
717: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
718: "LP", "OP", "P", "PD", "PP",
719: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
720: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175 schwarze 721: NULL
722: };
723:
1.141 kristaps 724: /* Array of injected predefined strings. */
725: #define PREDEFS_MAX 38
726: static const struct predef predefs[PREDEFS_MAX] = {
727: #include "predefs.in"
728: };
729:
1.155 kristaps 730: /* See roffhash_find() */
1.85 kristaps 731: #define ROFF_HASH(p) (p[0] - ASCII_LO)
732:
1.178 schwarze 733: static int roffit_lines; /* number of lines to delay */
734: static char *roffit_macro; /* nil-terminated macro line */
735:
1.207 schwarze 736:
1.85 kristaps 737: static void
1.155 kristaps 738: roffhash_init(void)
1.85 kristaps 739: {
740: struct roffmac *n;
741: int buc, i;
742:
1.106 kristaps 743: for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.85 kristaps 744: assert(roffs[i].name[0] >= ASCII_LO);
745: assert(roffs[i].name[0] <= ASCII_HI);
746:
747: buc = ROFF_HASH(roffs[i].name);
748:
749: if (NULL != (n = hash[buc])) {
750: for ( ; n->next; n = n->next)
751: /* Do nothing. */ ;
752: n->next = &roffs[i];
753: } else
754: hash[buc] = &roffs[i];
755: }
756: }
757:
1.67 kristaps 758: /*
759: * Look up a roff token by its name. Returns ROFF_MAX if no macro by
760: * the nil-terminated string name could be found.
761: */
762: static enum rofft
1.155 kristaps 763: roffhash_find(const char *p, size_t s)
1.67 kristaps 764: {
1.85 kristaps 765: int buc;
766: struct roffmac *n;
1.67 kristaps 767:
1.85 kristaps 768: /*
769: * libroff has an extremely simple hashtable, for the time
770: * being, which simply keys on the first character, which must
771: * be printable, then walks a chain. It works well enough until
772: * optimised.
773: */
774:
775: if (p[0] < ASCII_LO || p[0] > ASCII_HI)
776: return(ROFF_MAX);
777:
778: buc = ROFF_HASH(p);
779:
780: if (NULL == (n = hash[buc]))
781: return(ROFF_MAX);
782: for ( ; n; n = n->next)
1.106 kristaps 783: if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.85 kristaps 784: return((enum rofft)(n - roffs));
1.67 kristaps 785:
786: return(ROFF_MAX);
787: }
788:
789: /*
790: * Pop the current node off of the stack of roff instructions currently
791: * pending.
792: */
793: static void
794: roffnode_pop(struct roff *r)
795: {
796: struct roffnode *p;
797:
1.75 kristaps 798: assert(r->last);
1.207 schwarze 799: p = r->last;
1.82 kristaps 800:
1.75 kristaps 801: r->last = r->last->parent;
1.106 kristaps 802: free(p->name);
803: free(p->end);
1.67 kristaps 804: free(p);
805: }
806:
807: /*
808: * Push a roff node onto the instruction stack. This must later be
809: * removed with roffnode_pop().
810: */
1.98 schwarze 811: static void
1.106 kristaps 812: roffnode_push(struct roff *r, enum rofft tok, const char *name,
813: int line, int col)
1.67 kristaps 814: {
815: struct roffnode *p;
816:
1.98 schwarze 817: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 818: p->tok = tok;
1.106 kristaps 819: if (name)
820: p->name = mandoc_strdup(name);
1.67 kristaps 821: p->parent = r->last;
822: p->line = line;
823: p->col = col;
1.198 schwarze 824: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 825:
826: r->last = p;
827: }
828:
829: static void
830: roff_free1(struct roff *r)
831: {
1.176 schwarze 832: struct tbl_node *tbl;
1.125 kristaps 833: struct eqn_node *e;
1.167 kristaps 834: int i;
1.67 kristaps 835:
1.176 schwarze 836: while (NULL != (tbl = r->first_tbl)) {
837: r->first_tbl = tbl->next;
838: tbl_free(tbl);
1.109 kristaps 839: }
1.113 kristaps 840: r->first_tbl = r->last_tbl = r->tbl = NULL;
841:
1.125 kristaps 842: while (NULL != (e = r->first_eqn)) {
843: r->first_eqn = e->next;
844: eqn_free(e);
845: }
846: r->first_eqn = r->last_eqn = r->eqn = NULL;
847:
1.67 kristaps 848: while (r->last)
849: roffnode_pop(r);
1.109 kristaps 850:
1.223 schwarze 851: free (r->rstack);
852: r->rstack = NULL;
853: r->rstacksz = 0;
854: r->rstackpos = -1;
855:
856: roff_freereg(r->regtab);
857: r->regtab = NULL;
858:
1.167 kristaps 859: roff_freestr(r->strtab);
860: roff_freestr(r->xmbtab);
861: r->strtab = r->xmbtab = NULL;
862:
863: if (r->xtab)
864: for (i = 0; i < 128; i++)
865: free(r->xtab[i].p);
866: free(r->xtab);
867: r->xtab = NULL;
1.67 kristaps 868: }
869:
870: void
871: roff_reset(struct roff *r)
872: {
873:
874: roff_free1(r);
1.227 schwarze 875: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.174 kristaps 876: r->control = 0;
1.67 kristaps 877: }
878:
879: void
880: roff_free(struct roff *r)
881: {
882:
883: roff_free1(r);
884: free(r);
885: }
886:
887: struct roff *
1.237 schwarze 888: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.67 kristaps 889: {
890: struct roff *r;
891:
1.98 schwarze 892: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 893: r->parse = parse;
1.237 schwarze 894: r->mchars = mchars;
1.199 schwarze 895: r->options = options;
1.227 schwarze 896: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 897: r->rstackpos = -1;
1.207 schwarze 898:
1.155 kristaps 899: roffhash_init();
1.141 kristaps 900:
1.67 kristaps 901: return(r);
1.265 ! schwarze 902: }
! 903:
! 904: static void
! 905: roff_man_free1(struct roff_man *man)
! 906: {
! 907:
! 908: if (man->first != NULL) {
! 909: if (man->macroset == MACROSET_MDOC)
! 910: mdoc_node_delete(man, man->first);
! 911: else
! 912: man_node_delete(man, man->first);
! 913: }
! 914: free(man->meta.msec);
! 915: free(man->meta.vol);
! 916: free(man->meta.os);
! 917: free(man->meta.arch);
! 918: free(man->meta.title);
! 919: free(man->meta.name);
! 920: free(man->meta.date);
! 921: }
! 922:
! 923: static void
! 924: roff_man_alloc1(struct roff_man *man)
! 925: {
! 926:
! 927: memset(&man->meta, 0, sizeof(man->meta));
! 928: man->first = mandoc_calloc(1, sizeof(*man->first));
! 929: man->first->type = ROFFT_ROOT;
! 930: man->last = man->first;
! 931: man->last_es = NULL;
! 932: man->flags = 0;
! 933: man->macroset = MACROSET_NONE;
! 934: man->lastsec = man->lastnamed = SEC_NONE;
! 935: man->next = ROFF_NEXT_CHILD;
! 936: }
! 937:
! 938: void
! 939: roff_man_reset(struct roff_man *man)
! 940: {
! 941:
! 942: roff_man_free1(man);
! 943: roff_man_alloc1(man);
! 944: }
! 945:
! 946: void
! 947: roff_man_free(struct roff_man *man)
! 948: {
! 949:
! 950: roff_man_free1(man);
! 951: free(man);
! 952: }
! 953:
! 954: struct roff_man *
! 955: roff_man_alloc(struct roff *roff, struct mparse *parse,
! 956: const char *defos, int quick)
! 957: {
! 958: struct roff_man *man;
! 959:
! 960: man = mandoc_calloc(1, sizeof(*man));
! 961: man->parse = parse;
! 962: man->roff = roff;
! 963: man->defos = defos;
! 964: man->quick = quick;
! 965: roff_man_alloc1(man);
! 966: return(man);
1.67 kristaps 967: }
968:
1.94 kristaps 969: /*
1.206 schwarze 970: * In the current line, expand escape sequences that tend to get
971: * used in numerical expressions and conditional requests.
972: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 973: */
1.172 schwarze 974: static enum rofferr
1.238 schwarze 975: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 976: {
1.208 schwarze 977: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 978: const char *start; /* start of the string to process */
1.209 schwarze 979: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 980: const char *stnam; /* start of the name, after "[(*" */
981: const char *cp; /* end of the name, e.g. before ']' */
982: const char *res; /* the string to be substituted */
1.238 schwarze 983: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 984: size_t maxl; /* expected length of the escape name */
985: size_t naml; /* actual length of the escape name */
1.237 schwarze 986: enum mandoc_esc esc; /* type of the escape sequence */
987: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 988: int expand_count; /* to avoid infinite loops */
1.206 schwarze 989: int npos; /* position in numeric expression */
1.218 schwarze 990: int arg_complete; /* argument not interrupted by eol */
1.206 schwarze 991: char term; /* character terminating the escape */
1.94 kristaps 992:
1.170 schwarze 993: expand_count = 0;
1.238 schwarze 994: start = buf->buf + pos;
1.205 schwarze 995: stesc = strchr(start, '\0') - 1;
996: while (stesc-- > start) {
1.170 schwarze 997:
1.205 schwarze 998: /* Search backwards for the next backslash. */
999:
1.238 schwarze 1000: if (*stesc != '\\')
1.205 schwarze 1001: continue;
1002:
1003: /* If it is escaped, skip it. */
1004:
1005: for (cp = stesc - 1; cp >= start; cp--)
1.238 schwarze 1006: if (*cp != '\\')
1.205 schwarze 1007: break;
1008:
1.238 schwarze 1009: if ((stesc - cp) % 2 == 0) {
1.209 schwarze 1010: stesc = (char *)cp;
1.205 schwarze 1011: continue;
1012: }
1.108 schwarze 1013:
1.206 schwarze 1014: /* Decide whether to expand or to check only. */
1.108 schwarze 1015:
1.206 schwarze 1016: term = '\0';
1.205 schwarze 1017: cp = stesc + 1;
1.181 schwarze 1018: switch (*cp) {
1.207 schwarze 1019: case '*':
1.181 schwarze 1020: res = NULL;
1021: break;
1.207 schwarze 1022: case 'B':
1.206 schwarze 1023: /* FALLTHROUGH */
1.207 schwarze 1024: case 'w':
1.206 schwarze 1025: term = cp[1];
1026: /* FALLTHROUGH */
1.207 schwarze 1027: case 'n':
1.181 schwarze 1028: res = ubuf;
1029: break;
1030: default:
1.237 schwarze 1031: esc = mandoc_escape(&cp, &stnam, &inaml);
1032: if (esc == ESCAPE_ERROR ||
1033: (esc == ESCAPE_SPECIAL &&
1034: mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.219 schwarze 1035: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1036: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1037: "%.*s", (int)(cp - stesc), stesc);
1.205 schwarze 1038: continue;
1.152 kristaps 1039: }
1040:
1.205 schwarze 1041: if (EXPAND_LIMIT < ++expand_count) {
1042: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1043: ln, (int)(stesc - buf->buf), NULL);
1.205 schwarze 1044: return(ROFF_IGN);
1045: }
1.108 schwarze 1046:
1047: /*
1048: * The third character decides the length
1.181 schwarze 1049: * of the name of the string or register.
1.108 schwarze 1050: * Save a pointer to the name.
1051: */
1052:
1.238 schwarze 1053: if (term == '\0') {
1.206 schwarze 1054: switch (*++cp) {
1.207 schwarze 1055: case '\0':
1.206 schwarze 1056: maxl = 0;
1057: break;
1.207 schwarze 1058: case '(':
1.206 schwarze 1059: cp++;
1060: maxl = 2;
1061: break;
1.207 schwarze 1062: case '[':
1.206 schwarze 1063: cp++;
1064: term = ']';
1065: maxl = 0;
1066: break;
1067: default:
1068: maxl = 1;
1069: break;
1070: }
1071: } else {
1072: cp += 2;
1.94 kristaps 1073: maxl = 0;
1074: }
1.108 schwarze 1075: stnam = cp;
1.94 kristaps 1076:
1.108 schwarze 1077: /* Advance to the end of the name. */
1.94 kristaps 1078:
1.253 schwarze 1079: naml = 0;
1.218 schwarze 1080: arg_complete = 1;
1.253 schwarze 1081: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1082: if (*cp == '\0') {
1.219 schwarze 1083: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1084: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1085: arg_complete = 0;
1.206 schwarze 1086: break;
1.153 kristaps 1087: }
1.238 schwarze 1088: if (maxl == 0 && *cp == term) {
1.206 schwarze 1089: cp++;
1.253 schwarze 1090: break;
1091: }
1092: if (*cp++ != '\\' || stesc[1] != 'w') {
1093: naml++;
1094: continue;
1095: }
1096: switch (mandoc_escape(&cp, NULL, NULL)) {
1097: case ESCAPE_SPECIAL:
1098: /* FALLTHROUGH */
1099: case ESCAPE_UNICODE:
1100: /* FALLTHROUGH */
1101: case ESCAPE_NUMBERED:
1102: /* FALLTHROUGH */
1103: case ESCAPE_OVERSTRIKE:
1104: naml++;
1105: break;
1106: default:
1.94 kristaps 1107: break;
1.206 schwarze 1108: }
1.94 kristaps 1109: }
1110:
1.108 schwarze 1111: /*
1112: * Retrieve the replacement string; if it is
1113: * undefined, resume searching for escapes.
1114: */
1115:
1.206 schwarze 1116: switch (stesc[1]) {
1.207 schwarze 1117: case '*':
1.218 schwarze 1118: if (arg_complete)
1119: res = roff_getstrn(r, stnam, naml);
1.206 schwarze 1120: break;
1.207 schwarze 1121: case 'B':
1.206 schwarze 1122: npos = 0;
1.218 schwarze 1123: ubuf[0] = arg_complete &&
1.261 schwarze 1124: roff_evalnum(r, ln, stnam, &npos,
1125: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1126: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1127: ubuf[1] = '\0';
1128: break;
1.207 schwarze 1129: case 'n':
1.218 schwarze 1130: if (arg_complete)
1131: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1132: roff_getregn(r, stnam, naml));
1133: else
1134: ubuf[0] = '\0';
1.206 schwarze 1135: break;
1.207 schwarze 1136: case 'w':
1.218 schwarze 1137: /* use even incomplete args */
1.208 schwarze 1138: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1139: 24 * (int)naml);
1140: break;
1141: }
1.94 kristaps 1142:
1.238 schwarze 1143: if (res == NULL) {
1.219 schwarze 1144: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1145: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1146: "%.*s", (int)naml, stnam);
1.142 kristaps 1147: res = "";
1.246 schwarze 1148: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1149: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1150: ln, (int)(stesc - buf->buf), NULL);
1151: return(ROFF_IGN);
1.94 kristaps 1152: }
1153:
1.108 schwarze 1154: /* Replace the escape sequence by the string. */
1155:
1.209 schwarze 1156: *stesc = '\0';
1.238 schwarze 1157: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1158: buf->buf, res, cp) + 1;
1.94 kristaps 1159:
1.205 schwarze 1160: /* Prepare for the next replacement. */
1.94 kristaps 1161:
1.205 schwarze 1162: start = nbuf + pos;
1.238 schwarze 1163: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1164: free(buf->buf);
1165: buf->buf = nbuf;
1.154 kristaps 1166: }
1.172 schwarze 1167: return(ROFF_CONT);
1.154 kristaps 1168: }
1169:
1170: /*
1.178 schwarze 1171: * Process text streams:
1172: * Convert all breakable hyphens into ASCII_HYPH.
1173: * Decrement and spring input line trap.
1.154 kristaps 1174: */
1175: static enum rofferr
1.238 schwarze 1176: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154 kristaps 1177: {
1178: size_t sz;
1179: const char *start;
1.178 schwarze 1180: char *p;
1181: int isz;
1.154 kristaps 1182: enum mandoc_esc esc;
1183:
1.238 schwarze 1184: start = p = buf->buf + pos;
1.154 kristaps 1185:
1.238 schwarze 1186: while (*p != '\0') {
1.154 kristaps 1187: sz = strcspn(p, "-\\");
1188: p += sz;
1189:
1.238 schwarze 1190: if (*p == '\0')
1.159 kristaps 1191: break;
1192:
1.238 schwarze 1193: if (*p == '\\') {
1.154 kristaps 1194: /* Skip over escapes. */
1195: p++;
1.189 schwarze 1196: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1197: if (esc == ESCAPE_ERROR)
1.154 kristaps 1198: break;
1.264 schwarze 1199: while (*p == '-')
1200: p++;
1.155 kristaps 1201: continue;
1.159 kristaps 1202: } else if (p == start) {
1.158 kristaps 1203: p++;
1.155 kristaps 1204: continue;
1.158 kristaps 1205: }
1.155 kristaps 1206:
1.171 schwarze 1207: if (isalpha((unsigned char)p[-1]) &&
1208: isalpha((unsigned char)p[1]))
1.155 kristaps 1209: *p = ASCII_HYPH;
1210: p++;
1.94 kristaps 1211: }
1212:
1.178 schwarze 1213: /* Spring the input line trap. */
1.238 schwarze 1214: if (roffit_lines == 1) {
1215: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1216: free(buf->buf);
1217: buf->buf = p;
1218: buf->sz = isz + 1;
1.178 schwarze 1219: *offs = 0;
1220: free(roffit_macro);
1221: roffit_lines = 0;
1222: return(ROFF_REPARSE);
1.238 schwarze 1223: } else if (roffit_lines > 1)
1.178 schwarze 1224: --roffit_lines;
1.154 kristaps 1225: return(ROFF_CONT);
1.94 kristaps 1226: }
1227:
1.67 kristaps 1228: enum rofferr
1.238 schwarze 1229: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1230: {
1231: enum rofft t;
1.109 kristaps 1232: enum rofferr e;
1.238 schwarze 1233: int pos; /* parse point */
1.243 schwarze 1234: int spos; /* saved parse point for messages */
1.238 schwarze 1235: int ppos; /* original offset in buf->buf */
1236: int ctl; /* macro line (boolean) */
1237:
1238: ppos = pos = *offs;
1.79 kristaps 1239:
1.230 schwarze 1240: /* Handle in-line equation delimiters. */
1241:
1.236 schwarze 1242: if (r->tbl == NULL &&
1243: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1244: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1245: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1246: if (e == ROFF_REPARSE)
1247: return(e);
1248: assert(e == ROFF_CONT);
1249: }
1250:
1251: /* Expand some escape sequences. */
1.94 kristaps 1252:
1.238 schwarze 1253: e = roff_res(r, buf, ln, pos);
1254: if (e == ROFF_IGN)
1.172 schwarze 1255: return(e);
1.238 schwarze 1256: assert(e == ROFF_CONT);
1.94 kristaps 1257:
1.238 schwarze 1258: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1259:
1.94 kristaps 1260: /*
1.79 kristaps 1261: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1262: * text through the macro's filter.
1263: * Equations process all content themselves.
1264: * Tables process almost all content themselves, but we want
1265: * to warn about macros before passing it there.
1.79 kristaps 1266: */
1.74 kristaps 1267:
1.252 schwarze 1268: if (r->last != NULL && ! ctl) {
1.78 kristaps 1269: t = r->last->tok;
1270: assert(roffs[t].text);
1.238 schwarze 1271: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1272: assert(e == ROFF_IGN || e == ROFF_CONT);
1273: if (e != ROFF_CONT)
1.125 kristaps 1274: return(e);
1.182 schwarze 1275: }
1.252 schwarze 1276: if (r->eqn != NULL)
1.238 schwarze 1277: return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.252 schwarze 1278: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.259 schwarze 1279: return(tbl_read(r->tbl, ln, buf->buf, ppos));
1.252 schwarze 1280: if ( ! ctl)
1.238 schwarze 1281: return(roff_parsetext(buf, pos, offs));
1.228 schwarze 1282:
1283: /* Skip empty request lines. */
1284:
1.238 schwarze 1285: if (buf->buf[pos] == '"') {
1.228 schwarze 1286: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1287: ln, pos, NULL);
1288: return(ROFF_IGN);
1.238 schwarze 1289: } else if (buf->buf[pos] == '\0')
1.228 schwarze 1290: return(ROFF_IGN);
1.67 kristaps 1291:
1.79 kristaps 1292: /*
1293: * If a scope is open, go to the child handler for that macro,
1294: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1295: * Don't do so if an equation is open.
1.79 kristaps 1296: */
1.78 kristaps 1297:
1.79 kristaps 1298: if (r->last) {
1299: t = r->last->tok;
1300: assert(roffs[t].sub);
1.238 schwarze 1301: return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1302: }
1.78 kristaps 1303:
1.243 schwarze 1304: /* No scope is open. This is a new request or macro. */
1305:
1306: spos = pos;
1307: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1308:
1309: /* Tables ignore most macros. */
1310:
1311: if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
1312: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1313: ln, pos, buf->buf + spos);
1.257 schwarze 1314: if (t == ROFF_TS)
1315: return(ROFF_IGN);
1316: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1317: pos++;
1318: while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
1319: pos++;
1320: return(tbl_read(r->tbl, ln, buf->buf, pos));
1.243 schwarze 1321: }
1322:
1.79 kristaps 1323: /*
1.243 schwarze 1324: * This is neither a roff request nor a user-defined macro.
1325: * Let the standard macro set parsers handle it.
1.79 kristaps 1326: */
1.67 kristaps 1327:
1.243 schwarze 1328: if (t == ROFF_MAX)
1.79 kristaps 1329: return(ROFF_CONT);
1.243 schwarze 1330:
1331: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1332:
1.75 kristaps 1333: assert(roffs[t].proc);
1.238 schwarze 1334: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.74 kristaps 1335: }
1336:
1.117 kristaps 1337: void
1.74 kristaps 1338: roff_endparse(struct roff *r)
1339: {
1340:
1.110 kristaps 1341: if (r->last)
1.221 schwarze 1342: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1343: r->last->line, r->last->col,
1344: roffs[r->last->tok].name);
1.117 kristaps 1345:
1.125 kristaps 1346: if (r->eqn) {
1.221 schwarze 1347: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1348: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151 kristaps 1349: eqn_end(&r->eqn);
1.125 kristaps 1350: }
1351:
1.117 kristaps 1352: if (r->tbl) {
1.221 schwarze 1353: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1354: r->tbl->line, r->tbl->pos, "TS");
1.151 kristaps 1355: tbl_end(&r->tbl);
1.117 kristaps 1356: }
1.67 kristaps 1357: }
1358:
1359: /*
1360: * Parse a roff node's type from the input buffer. This must be in the
1361: * form of ".foo xxx" in the usual way.
1362: */
1363: static enum rofft
1.214 schwarze 1364: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1365: {
1.214 schwarze 1366: char *cp;
1.106 kristaps 1367: const char *mac;
1368: size_t maclen;
1.67 kristaps 1369: enum rofft t;
1370:
1.214 schwarze 1371: cp = buf + *pos;
1372:
1373: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.67 kristaps 1374: return(ROFF_MAX);
1375:
1.214 schwarze 1376: mac = cp;
1377: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1378:
1.106 kristaps 1379: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.155 kristaps 1380: ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.67 kristaps 1381:
1.214 schwarze 1382: if (ROFF_MAX != t)
1383: *pos = cp - buf;
1.67 kristaps 1384:
1385: return(t);
1386: }
1387:
1388: static enum rofferr
1.76 kristaps 1389: roff_cblock(ROFF_ARGS)
1.67 kristaps 1390: {
1391:
1.79 kristaps 1392: /*
1393: * A block-close `..' should only be invoked as a child of an
1394: * ignore macro, otherwise raise a warning and just ignore it.
1395: */
1396:
1.238 schwarze 1397: if (r->last == NULL) {
1.221 schwarze 1398: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1399: ln, ppos, "..");
1.76 kristaps 1400: return(ROFF_IGN);
1401: }
1.67 kristaps 1402:
1.81 kristaps 1403: switch (r->last->tok) {
1.207 schwarze 1404: case ROFF_am:
1.220 schwarze 1405: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.81 kristaps 1406: /* FALLTHROUGH */
1.207 schwarze 1407: case ROFF_ami:
1.81 kristaps 1408: /* FALLTHROUGH */
1.207 schwarze 1409: case ROFF_de:
1.108 schwarze 1410: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.81 kristaps 1411: /* FALLTHROUGH */
1.207 schwarze 1412: case ROFF_dei:
1.81 kristaps 1413: /* FALLTHROUGH */
1.207 schwarze 1414: case ROFF_ig:
1.81 kristaps 1415: break;
1416: default:
1.221 schwarze 1417: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1418: ln, ppos, "..");
1.67 kristaps 1419: return(ROFF_IGN);
1.76 kristaps 1420: }
1.67 kristaps 1421:
1.238 schwarze 1422: if (buf->buf[pos] != '\0')
1.217 schwarze 1423: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1424: ".. %s", buf->buf + pos);
1.71 kristaps 1425:
1426: roffnode_pop(r);
1.76 kristaps 1427: roffnode_cleanscope(r);
1428: return(ROFF_IGN);
1.71 kristaps 1429:
1.67 kristaps 1430: }
1431:
1.76 kristaps 1432: static void
1433: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1434: {
1435:
1.76 kristaps 1436: while (r->last) {
1.173 schwarze 1437: if (--r->last->endspan != 0)
1.76 kristaps 1438: break;
1439: roffnode_pop(r);
1440: }
1.67 kristaps 1441: }
1442:
1.195 schwarze 1443: static void
1444: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1445: {
1446:
1.76 kristaps 1447: if (NULL == r->last) {
1.221 schwarze 1448: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1449: ln, ppos, "\\}");
1.195 schwarze 1450: return;
1.76 kristaps 1451: }
1452:
1.82 kristaps 1453: switch (r->last->tok) {
1.207 schwarze 1454: case ROFF_el:
1.82 kristaps 1455: /* FALLTHROUGH */
1.207 schwarze 1456: case ROFF_ie:
1.82 kristaps 1457: /* FALLTHROUGH */
1.207 schwarze 1458: case ROFF_if:
1.82 kristaps 1459: break;
1460: default:
1.221 schwarze 1461: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1462: ln, ppos, "\\}");
1.195 schwarze 1463: return;
1.75 kristaps 1464: }
1465:
1.76 kristaps 1466: if (r->last->endspan > -1) {
1.221 schwarze 1467: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1468: ln, ppos, "\\}");
1.195 schwarze 1469: return;
1.76 kristaps 1470: }
1471:
1.75 kristaps 1472: roffnode_pop(r);
1.76 kristaps 1473: roffnode_cleanscope(r);
1.195 schwarze 1474: return;
1.76 kristaps 1475: }
1476:
1477: static enum rofferr
1.80 kristaps 1478: roff_block(ROFF_ARGS)
1.76 kristaps 1479: {
1.220 schwarze 1480: const char *name;
1481: char *iname, *cp;
1.213 schwarze 1482: size_t namesz;
1.106 kristaps 1483:
1.220 schwarze 1484: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1485:
1.238 schwarze 1486: if (tok == ROFF_de1)
1.220 schwarze 1487: tok = ROFF_de;
1.251 schwarze 1488: else if (tok == ROFF_dei1)
1489: tok = ROFF_dei;
1.238 schwarze 1490: else if (tok == ROFF_am1)
1.220 schwarze 1491: tok = ROFF_am;
1.251 schwarze 1492: else if (tok == ROFF_ami1)
1493: tok = ROFF_ami;
1.220 schwarze 1494:
1495: /* Parse the macro name argument. */
1496:
1.238 schwarze 1497: cp = buf->buf + pos;
1498: if (tok == ROFF_ig) {
1.220 schwarze 1499: iname = NULL;
1500: namesz = 0;
1501: } else {
1502: iname = cp;
1503: namesz = roff_getname(r, &cp, ln, ppos);
1504: iname[namesz] = '\0';
1505: }
1.107 kristaps 1506:
1.220 schwarze 1507: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1508:
1.238 schwarze 1509: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1510: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1511: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1512: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1513: "%.*s", (int)namesz, iname);
1514: namesz = 0;
1515: } else
1516: namesz = strlen(name);
1517: } else
1518: name = iname;
1.107 kristaps 1519:
1.238 schwarze 1520: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1521: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1522: ln, ppos, roffs[tok].name);
1523: return(ROFF_IGN);
1524: }
1.80 kristaps 1525:
1.106 kristaps 1526: roffnode_push(r, tok, name, ln, ppos);
1527:
1528: /*
1529: * At the beginning of a `de' macro, clear the existing string
1530: * with the same name, if there is one. New content will be
1.193 schwarze 1531: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1532: */
1.107 kristaps 1533:
1.238 schwarze 1534: if (tok == ROFF_de || tok == ROFF_dei)
1.213 schwarze 1535: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76 kristaps 1536:
1.238 schwarze 1537: if (*cp == '\0')
1.78 kristaps 1538: return(ROFF_IGN);
1539:
1.220 schwarze 1540: /* Get the custom end marker. */
1.107 kristaps 1541:
1.220 schwarze 1542: iname = cp;
1.213 schwarze 1543: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1544:
1545: /* Resolve the end marker if it is indirect. */
1546:
1.238 schwarze 1547: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1548: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1549: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1550: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1551: "%.*s", (int)namesz, iname);
1552: namesz = 0;
1553: } else
1554: namesz = strlen(name);
1555: } else
1556: name = iname;
1557:
1.213 schwarze 1558: if (namesz)
1559: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1560:
1.238 schwarze 1561: if (*cp != '\0')
1.217 schwarze 1562: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1563: ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.74 kristaps 1564:
1.78 kristaps 1565: return(ROFF_IGN);
1566: }
1567:
1568: static enum rofferr
1.80 kristaps 1569: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1570: {
1571: enum rofft t;
1572: int i, j;
1573:
1574: /*
1575: * First check whether a custom macro exists at this level. If
1576: * it does, then check against it. This is some of groff's
1577: * stranger behaviours. If we encountered a custom end-scope
1578: * tag and that tag also happens to be a "real" macro, then we
1579: * need to try interpreting it again as a real macro. If it's
1580: * not, then return ignore. Else continue.
1581: */
1582:
1583: if (r->last->end) {
1.130 kristaps 1584: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1585: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1586: break;
1587:
1.238 schwarze 1588: if (r->last->end[j] == '\0' &&
1589: (buf->buf[i] == '\0' ||
1590: buf->buf[i] == ' ' ||
1591: buf->buf[i] == '\t')) {
1.79 kristaps 1592: roffnode_pop(r);
1593: roffnode_cleanscope(r);
1594:
1.238 schwarze 1595: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1596: i++;
1597:
1598: pos = i;
1.238 schwarze 1599: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1600: ROFF_MAX)
1.79 kristaps 1601: return(ROFF_RERUN);
1602: return(ROFF_IGN);
1603: }
1604: }
1605:
1606: /*
1607: * If we have no custom end-query or lookup failed, then try
1608: * pulling it out of the hashtable.
1609: */
1610:
1.238 schwarze 1611: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1612:
1.238 schwarze 1613: if (t != ROFF_cblock) {
1614: if (tok != ROFF_ig)
1615: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.79 kristaps 1616: return(ROFF_IGN);
1.106 kristaps 1617: }
1.79 kristaps 1618:
1619: assert(roffs[t].proc);
1.238 schwarze 1620: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.79 kristaps 1621: }
1622:
1623: static enum rofferr
1.80 kristaps 1624: roff_block_text(ROFF_ARGS)
1.78 kristaps 1625: {
1626:
1.238 schwarze 1627: if (tok != ROFF_ig)
1628: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1629:
1.78 kristaps 1630: return(ROFF_IGN);
1631: }
1632:
1633: static enum rofferr
1.82 kristaps 1634: roff_cond_sub(ROFF_ARGS)
1635: {
1636: enum rofft t;
1.139 kristaps 1637: char *ep;
1.198 schwarze 1638: int rr;
1.82 kristaps 1639:
1640: rr = r->last->rule;
1.139 kristaps 1641: roffnode_cleanscope(r);
1.238 schwarze 1642: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82 kristaps 1643:
1.139 kristaps 1644: /*
1.177 schwarze 1645: * Fully handle known macros when they are structurally
1646: * required or when the conditional evaluated to true.
1.87 kristaps 1647: */
1648:
1.238 schwarze 1649: if ((t != ROFF_MAX) &&
1650: (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.177 schwarze 1651: assert(roffs[t].proc);
1.238 schwarze 1652: return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.177 schwarze 1653: }
1.144 kristaps 1654:
1.196 schwarze 1655: /*
1656: * If `\}' occurs on a macro line without a preceding macro,
1657: * drop the line completely.
1658: */
1659:
1.238 schwarze 1660: ep = buf->buf + pos;
1661: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1662: rr = 0;
1.196 schwarze 1663:
1.177 schwarze 1664: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1665:
1.238 schwarze 1666: while ((ep = strchr(ep, '\\')) != NULL) {
1667: if (*(++ep) == '}') {
1.197 schwarze 1668: *ep = '&';
1.238 schwarze 1669: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1670: }
1.247 schwarze 1671: if (*ep != '\0')
1672: ++ep;
1.177 schwarze 1673: }
1.198 schwarze 1674: return(rr ? ROFF_CONT : ROFF_IGN);
1.82 kristaps 1675: }
1676:
1677: static enum rofferr
1678: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1679: {
1.140 kristaps 1680: char *ep;
1.198 schwarze 1681: int rr;
1.82 kristaps 1682:
1683: rr = r->last->rule;
1.140 kristaps 1684: roffnode_cleanscope(r);
1.82 kristaps 1685:
1.238 schwarze 1686: ep = buf->buf + pos;
1687: while ((ep = strchr(ep, '\\')) != NULL) {
1688: if (*(++ep) == '}') {
1.197 schwarze 1689: *ep = '&';
1.238 schwarze 1690: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1691: }
1.247 schwarze 1692: if (*ep != '\0')
1693: ++ep;
1.78 kristaps 1694: }
1.198 schwarze 1695: return(rr ? ROFF_CONT : ROFF_IGN);
1.74 kristaps 1696: }
1697:
1.204 schwarze 1698: /*
1699: * Parse a single signed integer number. Stop at the first non-digit.
1700: * If there is at least one digit, return success and advance the
1701: * parse point, else return failure and let the parse point unchanged.
1702: * Ignore overflows, treat them just like the C language.
1703: */
1.184 schwarze 1704: static int
1.261 schwarze 1705: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 1706: {
1.261 schwarze 1707: int myres, scaled, n, p;
1.206 schwarze 1708:
1709: if (NULL == res)
1710: res = &myres;
1.184 schwarze 1711:
1712: p = *pos;
1713: n = v[p] == '-';
1.261 schwarze 1714: if (n || v[p] == '+')
1.184 schwarze 1715: p++;
1716:
1.261 schwarze 1717: if (flags & ROFFNUM_WHITE)
1718: while (isspace((unsigned char)v[p]))
1719: p++;
1720:
1.184 schwarze 1721: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 1722: *res = 10 * *res + v[p] - '0';
1.184 schwarze 1723: if (p == *pos + n)
1724: return 0;
1725:
1726: if (n)
1727: *res = -*res;
1728:
1.254 schwarze 1729: /* Each number may be followed by one optional scaling unit. */
1730:
1731: switch (v[p]) {
1732: case 'f':
1.261 schwarze 1733: scaled = *res * 65536;
1.254 schwarze 1734: break;
1735: case 'i':
1.261 schwarze 1736: scaled = *res * 240;
1.254 schwarze 1737: break;
1738: case 'c':
1.261 schwarze 1739: scaled = *res * 240 / 2.54;
1.254 schwarze 1740: break;
1741: case 'v':
1742: /* FALLTROUGH */
1743: case 'P':
1.261 schwarze 1744: scaled = *res * 40;
1.254 schwarze 1745: break;
1746: case 'm':
1747: /* FALLTROUGH */
1748: case 'n':
1.261 schwarze 1749: scaled = *res * 24;
1.254 schwarze 1750: break;
1751: case 'p':
1.261 schwarze 1752: scaled = *res * 10 / 3;
1.254 schwarze 1753: break;
1754: case 'u':
1.261 schwarze 1755: scaled = *res;
1.254 schwarze 1756: break;
1757: case 'M':
1.261 schwarze 1758: scaled = *res * 6 / 25;
1.254 schwarze 1759: break;
1760: default:
1.261 schwarze 1761: scaled = *res;
1.254 schwarze 1762: p--;
1763: break;
1764: }
1.261 schwarze 1765: if (flags & ROFFNUM_SCALE)
1766: *res = scaled;
1.254 schwarze 1767:
1768: *pos = p + 1;
1769: return(1);
1.184 schwarze 1770: }
1771:
1.198 schwarze 1772: /*
1773: * Evaluate a string comparison condition.
1774: * The first character is the delimiter.
1775: * Succeed if the string up to its second occurrence
1776: * matches the string up to its third occurence.
1777: * Advance the cursor after the third occurrence
1778: * or lacking that, to the end of the line.
1779: */
1780: static int
1781: roff_evalstrcond(const char *v, int *pos)
1782: {
1783: const char *s1, *s2, *s3;
1784: int match;
1785:
1786: match = 0;
1787: s1 = v + *pos; /* initial delimiter */
1788: s2 = s1 + 1; /* for scanning the first string */
1789: s3 = strchr(s2, *s1); /* for scanning the second string */
1790:
1791: if (NULL == s3) /* found no middle delimiter */
1792: goto out;
1793:
1794: while ('\0' != *++s3) {
1795: if (*s2 != *s3) { /* mismatch */
1796: s3 = strchr(s3, *s1);
1797: break;
1798: }
1799: if (*s3 == *s1) { /* found the final delimiter */
1800: match = 1;
1801: break;
1802: }
1803: s2++;
1804: }
1805:
1806: out:
1807: if (NULL == s3)
1808: s3 = strchr(s2, '\0');
1.242 schwarze 1809: else if (*s3 != '\0')
1.198 schwarze 1810: s3++;
1811: *pos = s3 - v;
1812: return(match);
1813: }
1814:
1.204 schwarze 1815: /*
1816: * Evaluate an optionally negated single character, numerical,
1817: * or string condition.
1818: */
1.198 schwarze 1819: static int
1.234 kristaps 1820: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.88 kristaps 1821: {
1.241 schwarze 1822: int number, savepos, wanttrue;
1.88 kristaps 1823:
1.198 schwarze 1824: if ('!' == v[*pos]) {
1825: wanttrue = 0;
1826: (*pos)++;
1827: } else
1828: wanttrue = 1;
1829:
1.88 kristaps 1830: switch (v[*pos]) {
1.240 schwarze 1831: case '\0':
1832: return(0);
1.207 schwarze 1833: case 'n':
1.198 schwarze 1834: /* FALLTHROUGH */
1.207 schwarze 1835: case 'o':
1.88 kristaps 1836: (*pos)++;
1.198 schwarze 1837: return(wanttrue);
1.207 schwarze 1838: case 'c':
1.198 schwarze 1839: /* FALLTHROUGH */
1.207 schwarze 1840: case 'd':
1.198 schwarze 1841: /* FALLTHROUGH */
1.207 schwarze 1842: case 'e':
1.88 kristaps 1843: /* FALLTHROUGH */
1.207 schwarze 1844: case 'r':
1.88 kristaps 1845: /* FALLTHROUGH */
1.207 schwarze 1846: case 't':
1.239 schwarze 1847: /* FALLTHROUGH */
1848: case 'v':
1.88 kristaps 1849: (*pos)++;
1.198 schwarze 1850: return(!wanttrue);
1.88 kristaps 1851: default:
1852: break;
1853: }
1854:
1.241 schwarze 1855: savepos = *pos;
1.261 schwarze 1856: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.204 schwarze 1857: return((number > 0) == wanttrue);
1.241 schwarze 1858: else if (*pos == savepos)
1859: return(roff_evalstrcond(v, pos) == wanttrue);
1.204 schwarze 1860: else
1.241 schwarze 1861: return (0);
1.88 kristaps 1862: }
1863:
1.74 kristaps 1864: static enum rofferr
1.103 kristaps 1865: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 1866: {
1.123 schwarze 1867:
1.89 kristaps 1868: return(ROFF_IGN);
1869: }
1870:
1.104 kristaps 1871: static enum rofferr
1.251 schwarze 1872: roff_insec(ROFF_ARGS)
1873: {
1874:
1875: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1876: ln, ppos, roffs[tok].name);
1877: return(ROFF_IGN);
1878: }
1879:
1880: static enum rofferr
1881: roff_unsupp(ROFF_ARGS)
1882: {
1883:
1884: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1885: ln, ppos, roffs[tok].name);
1886: return(ROFF_IGN);
1887: }
1888:
1889: static enum rofferr
1.82 kristaps 1890: roff_cond(ROFF_ARGS)
1.74 kristaps 1891: {
1.173 schwarze 1892:
1893: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 1894:
1.207 schwarze 1895: /*
1.134 kristaps 1896: * An `.el' has no conditional body: it will consume the value
1897: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 1898: * defaults to DENY.
1.134 kristaps 1899: *
1900: * If we're not an `el', however, then evaluate the conditional.
1901: */
1.133 kristaps 1902:
1.238 schwarze 1903: r->last->rule = tok == ROFF_el ?
1.207 schwarze 1904: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 1905: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 1906:
1.134 kristaps 1907: /*
1908: * An if-else will put the NEGATION of the current evaluated
1909: * conditional into the stack of rules.
1910: */
1911:
1.238 schwarze 1912: if (tok == ROFF_ie) {
1.223 schwarze 1913: if (r->rstackpos + 1 == r->rstacksz) {
1914: r->rstacksz += 16;
1915: r->rstack = mandoc_reallocarray(r->rstack,
1916: r->rstacksz, sizeof(int));
1.134 kristaps 1917: }
1.198 schwarze 1918: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 1919: }
1.88 kristaps 1920:
1921: /* If the parent has false as its rule, then so do we. */
1922:
1.198 schwarze 1923: if (r->last->parent && !r->last->parent->rule)
1924: r->last->rule = 0;
1.88 kristaps 1925:
1926: /*
1.173 schwarze 1927: * Determine scope.
1928: * If there is nothing on the line after the conditional,
1929: * not even whitespace, use next-line scope.
1.88 kristaps 1930: */
1.74 kristaps 1931:
1.238 schwarze 1932: if (buf->buf[pos] == '\0') {
1.173 schwarze 1933: r->last->endspan = 2;
1934: goto out;
1935: }
1936:
1.238 schwarze 1937: while (buf->buf[pos] == ' ')
1.173 schwarze 1938: pos++;
1939:
1940: /* An opening brace requests multiline scope. */
1.75 kristaps 1941:
1.238 schwarze 1942: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 1943: r->last->endspan = -1;
1944: pos += 2;
1.173 schwarze 1945: goto out;
1.207 schwarze 1946: }
1.74 kristaps 1947:
1.77 kristaps 1948: /*
1.173 schwarze 1949: * Anything else following the conditional causes
1950: * single-line scope. Warn if the scope contains
1951: * nothing but trailing whitespace.
1.77 kristaps 1952: */
1953:
1.238 schwarze 1954: if (buf->buf[pos] == '\0')
1.216 schwarze 1955: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1956: ln, ppos, roffs[tok].name);
1.77 kristaps 1957:
1.173 schwarze 1958: r->last->endspan = 1;
1.74 kristaps 1959:
1.173 schwarze 1960: out:
1.75 kristaps 1961: *offs = pos;
1962: return(ROFF_RERUN);
1.83 schwarze 1963: }
1964:
1965: static enum rofferr
1.92 schwarze 1966: roff_ds(ROFF_ARGS)
1967: {
1.212 schwarze 1968: char *string;
1969: const char *name;
1970: size_t namesz;
1.96 kristaps 1971:
1.251 schwarze 1972: /* Ignore groff compatibility mode for now. */
1973:
1974: if (tok == ROFF_ds1)
1975: tok = ROFF_ds;
1976: else if (tok == ROFF_as1)
1977: tok = ROFF_as;
1978:
1.96 kristaps 1979: /*
1.212 schwarze 1980: * The first word is the name of the string.
1981: * If it is empty or terminated by an escape sequence,
1982: * abort the `ds' request without defining anything.
1.96 kristaps 1983: */
1.92 schwarze 1984:
1.238 schwarze 1985: name = string = buf->buf + pos;
1986: if (*name == '\0')
1.92 schwarze 1987: return(ROFF_IGN);
1988:
1.212 schwarze 1989: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 1990: if (name[namesz] == '\\')
1.212 schwarze 1991: return(ROFF_IGN);
1992:
1993: /* Read past the initial double-quote, if any. */
1.238 schwarze 1994: if (*string == '"')
1.92 schwarze 1995: string++;
1996:
1.96 kristaps 1997: /* The rest is the value. */
1.212 schwarze 1998: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
1999: ROFF_as == tok);
1.92 schwarze 2000: return(ROFF_IGN);
2001: }
2002:
1.204 schwarze 2003: /*
2004: * Parse a single operator, one or two characters long.
2005: * If the operator is recognized, return success and advance the
2006: * parse point, else return failure and let the parse point unchanged.
2007: */
2008: static int
2009: roff_getop(const char *v, int *pos, char *res)
2010: {
2011:
2012: *res = v[*pos];
2013:
2014: switch (*res) {
1.207 schwarze 2015: case '+':
1.204 schwarze 2016: /* FALLTHROUGH */
1.207 schwarze 2017: case '-':
1.204 schwarze 2018: /* FALLTHROUGH */
1.207 schwarze 2019: case '*':
1.204 schwarze 2020: /* FALLTHROUGH */
1.207 schwarze 2021: case '/':
1.204 schwarze 2022: /* FALLTHROUGH */
1.207 schwarze 2023: case '%':
1.204 schwarze 2024: /* FALLTHROUGH */
1.207 schwarze 2025: case '&':
1.204 schwarze 2026: /* FALLTHROUGH */
1.207 schwarze 2027: case ':':
1.204 schwarze 2028: break;
2029: case '<':
2030: switch (v[*pos + 1]) {
1.207 schwarze 2031: case '=':
1.204 schwarze 2032: *res = 'l';
2033: (*pos)++;
2034: break;
1.207 schwarze 2035: case '>':
1.204 schwarze 2036: *res = '!';
2037: (*pos)++;
2038: break;
1.207 schwarze 2039: case '?':
1.204 schwarze 2040: *res = 'i';
2041: (*pos)++;
2042: break;
2043: default:
2044: break;
2045: }
2046: break;
2047: case '>':
2048: switch (v[*pos + 1]) {
1.207 schwarze 2049: case '=':
1.204 schwarze 2050: *res = 'g';
2051: (*pos)++;
2052: break;
1.207 schwarze 2053: case '?':
1.204 schwarze 2054: *res = 'a';
2055: (*pos)++;
2056: break;
2057: default:
2058: break;
2059: }
2060: break;
2061: case '=':
2062: if ('=' == v[*pos + 1])
2063: (*pos)++;
2064: break;
2065: default:
2066: return(0);
2067: }
2068: (*pos)++;
2069:
2070: return(*res);
2071: }
2072:
2073: /*
2074: * Evaluate either a parenthesized numeric expression
2075: * or a single signed integer number.
2076: */
2077: static int
1.235 schwarze 2078: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2079: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2080: {
2081:
2082: if ('(' != v[*pos])
1.261 schwarze 2083: return(roff_getnum(v, pos, res, flags));
1.204 schwarze 2084:
2085: (*pos)++;
1.261 schwarze 2086: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.204 schwarze 2087: return(0);
2088:
1.206 schwarze 2089: /*
2090: * Omission of the closing parenthesis
2091: * is an error in validation mode,
2092: * but ignored in evaluation mode.
2093: */
2094:
1.204 schwarze 2095: if (')' == v[*pos])
2096: (*pos)++;
1.206 schwarze 2097: else if (NULL == res)
2098: return(0);
1.204 schwarze 2099:
2100: return(1);
2101: }
2102:
2103: /*
2104: * Evaluate a complete numeric expression.
2105: * Proceed left to right, there is no concept of precedence.
2106: */
2107: static int
1.235 schwarze 2108: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2109: int *pos, int *res, int flags)
1.204 schwarze 2110: {
2111: int mypos, operand2;
2112: char operator;
2113:
2114: if (NULL == pos) {
2115: mypos = 0;
2116: pos = &mypos;
2117: }
2118:
1.261 schwarze 2119: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2120: while (isspace((unsigned char)v[*pos]))
2121: (*pos)++;
2122:
1.261 schwarze 2123: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.204 schwarze 2124: return(0);
2125:
2126: while (1) {
1.261 schwarze 2127: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2128: while (isspace((unsigned char)v[*pos]))
2129: (*pos)++;
2130:
2131: if ( ! roff_getop(v, pos, &operator))
2132: break;
2133:
1.261 schwarze 2134: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2135: while (isspace((unsigned char)v[*pos]))
2136: (*pos)++;
2137:
1.261 schwarze 2138: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.204 schwarze 2139: return(0);
2140:
1.261 schwarze 2141: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2142: while (isspace((unsigned char)v[*pos]))
2143: (*pos)++;
1.206 schwarze 2144:
2145: if (NULL == res)
2146: continue;
1.204 schwarze 2147:
2148: switch (operator) {
1.207 schwarze 2149: case '+':
1.204 schwarze 2150: *res += operand2;
2151: break;
1.207 schwarze 2152: case '-':
1.204 schwarze 2153: *res -= operand2;
2154: break;
1.207 schwarze 2155: case '*':
1.204 schwarze 2156: *res *= operand2;
2157: break;
1.207 schwarze 2158: case '/':
1.244 schwarze 2159: if (operand2 == 0) {
1.235 schwarze 2160: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2161: r->parse, ln, *pos, v);
2162: *res = 0;
2163: break;
2164: }
1.204 schwarze 2165: *res /= operand2;
2166: break;
1.207 schwarze 2167: case '%':
1.244 schwarze 2168: if (operand2 == 0) {
2169: mandoc_msg(MANDOCERR_DIVZERO,
2170: r->parse, ln, *pos, v);
2171: *res = 0;
2172: break;
2173: }
1.204 schwarze 2174: *res %= operand2;
2175: break;
1.207 schwarze 2176: case '<':
1.204 schwarze 2177: *res = *res < operand2;
2178: break;
1.207 schwarze 2179: case '>':
1.204 schwarze 2180: *res = *res > operand2;
2181: break;
1.207 schwarze 2182: case 'l':
1.204 schwarze 2183: *res = *res <= operand2;
2184: break;
1.207 schwarze 2185: case 'g':
1.204 schwarze 2186: *res = *res >= operand2;
2187: break;
1.207 schwarze 2188: case '=':
1.204 schwarze 2189: *res = *res == operand2;
2190: break;
1.207 schwarze 2191: case '!':
1.204 schwarze 2192: *res = *res != operand2;
2193: break;
1.207 schwarze 2194: case '&':
1.204 schwarze 2195: *res = *res && operand2;
2196: break;
1.207 schwarze 2197: case ':':
1.204 schwarze 2198: *res = *res || operand2;
2199: break;
1.207 schwarze 2200: case 'i':
1.204 schwarze 2201: if (operand2 < *res)
2202: *res = operand2;
2203: break;
1.207 schwarze 2204: case 'a':
1.204 schwarze 2205: if (operand2 > *res)
2206: *res = operand2;
2207: break;
2208: default:
2209: abort();
2210: }
2211: }
2212: return(1);
2213: }
2214:
1.180 schwarze 2215: void
1.187 schwarze 2216: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2217: {
1.180 schwarze 2218: struct roffreg *reg;
2219:
2220: /* Search for an existing register with the same name. */
2221: reg = r->regtab;
2222:
2223: while (reg && strcmp(name, reg->key.p))
2224: reg = reg->next;
1.147 kristaps 2225:
1.180 schwarze 2226: if (NULL == reg) {
2227: /* Create a new register. */
2228: reg = mandoc_malloc(sizeof(struct roffreg));
2229: reg->key.p = mandoc_strdup(name);
2230: reg->key.sz = strlen(name);
1.187 schwarze 2231: reg->val = 0;
1.180 schwarze 2232: reg->next = r->regtab;
2233: r->regtab = reg;
2234: }
2235:
1.187 schwarze 2236: if ('+' == sign)
2237: reg->val += val;
2238: else if ('-' == sign)
2239: reg->val -= val;
2240: else
2241: reg->val = val;
1.147 kristaps 2242: }
2243:
1.192 schwarze 2244: /*
2245: * Handle some predefined read-only number registers.
2246: * For now, return -1 if the requested register is not predefined;
2247: * in case a predefined read-only register having the value -1
2248: * were to turn up, another special value would have to be chosen.
2249: */
2250: static int
2251: roff_getregro(const char *name)
2252: {
2253:
2254: switch (*name) {
1.207 schwarze 2255: case 'A': /* ASCII approximation mode is always off. */
1.192 schwarze 2256: return(0);
1.207 schwarze 2257: case 'g': /* Groff compatibility mode is always on. */
1.192 schwarze 2258: return(1);
1.207 schwarze 2259: case 'H': /* Fixed horizontal resolution. */
1.192 schwarze 2260: return (24);
1.207 schwarze 2261: case 'j': /* Always adjust left margin only. */
1.192 schwarze 2262: return(0);
1.207 schwarze 2263: case 'T': /* Some output device is always defined. */
1.192 schwarze 2264: return(1);
1.207 schwarze 2265: case 'V': /* Fixed vertical resolution. */
1.192 schwarze 2266: return (40);
2267: default:
2268: return (-1);
2269: }
2270: }
2271:
1.181 schwarze 2272: int
1.180 schwarze 2273: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2274: {
1.180 schwarze 2275: struct roffreg *reg;
1.192 schwarze 2276: int val;
2277:
2278: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
2279: val = roff_getregro(name + 1);
2280: if (-1 != val)
2281: return (val);
2282: }
1.180 schwarze 2283:
2284: for (reg = r->regtab; reg; reg = reg->next)
2285: if (0 == strcmp(name, reg->key.p))
1.181 schwarze 2286: return(reg->val);
2287:
2288: return(0);
2289: }
2290:
2291: static int
2292: roff_getregn(const struct roff *r, const char *name, size_t len)
2293: {
2294: struct roffreg *reg;
1.192 schwarze 2295: int val;
2296:
2297: if ('.' == name[0] && 2 == len) {
2298: val = roff_getregro(name + 1);
2299: if (-1 != val)
2300: return (val);
2301: }
1.181 schwarze 2302:
2303: for (reg = r->regtab; reg; reg = reg->next)
2304: if (len == reg->key.sz &&
2305: 0 == strncmp(name, reg->key.p, len))
2306: return(reg->val);
1.147 kristaps 2307:
1.180 schwarze 2308: return(0);
1.147 kristaps 2309: }
2310:
1.180 schwarze 2311: static void
2312: roff_freereg(struct roffreg *reg)
1.147 kristaps 2313: {
1.180 schwarze 2314: struct roffreg *old_reg;
1.147 kristaps 2315:
1.180 schwarze 2316: while (NULL != reg) {
2317: free(reg->key.p);
2318: old_reg = reg;
2319: reg = reg->next;
2320: free(old_reg);
2321: }
1.147 kristaps 2322: }
1.92 schwarze 2323:
2324: static enum rofferr
1.89 kristaps 2325: roff_nr(ROFF_ARGS)
1.83 schwarze 2326: {
1.212 schwarze 2327: char *key, *val;
2328: size_t keysz;
1.138 kristaps 2329: int iv;
1.187 schwarze 2330: char sign;
1.89 kristaps 2331:
1.238 schwarze 2332: key = val = buf->buf + pos;
2333: if (*key == '\0')
1.212 schwarze 2334: return(ROFF_IGN);
2335:
2336: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2337: if (key[keysz] == '\\')
1.212 schwarze 2338: return(ROFF_IGN);
2339: key[keysz] = '\0';
1.89 kristaps 2340:
1.187 schwarze 2341: sign = *val;
1.238 schwarze 2342: if (sign == '+' || sign == '-')
1.187 schwarze 2343: val++;
2344:
1.261 schwarze 2345: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204 schwarze 2346: roff_setreg(r, key, iv, sign);
1.109 kristaps 2347:
1.203 schwarze 2348: return(ROFF_IGN);
2349: }
2350:
2351: static enum rofferr
2352: roff_rr(ROFF_ARGS)
2353: {
2354: struct roffreg *reg, **prev;
1.212 schwarze 2355: char *name, *cp;
2356: size_t namesz;
1.203 schwarze 2357:
1.238 schwarze 2358: name = cp = buf->buf + pos;
2359: if (*name == '\0')
1.212 schwarze 2360: return(ROFF_IGN);
2361: namesz = roff_getname(r, &cp, ln, pos);
2362: name[namesz] = '\0';
1.203 schwarze 2363:
2364: prev = &r->regtab;
2365: while (1) {
2366: reg = *prev;
1.238 schwarze 2367: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2368: break;
2369: prev = ®->next;
2370: }
1.238 schwarze 2371: if (reg != NULL) {
1.203 schwarze 2372: *prev = reg->next;
2373: free(reg->key.p);
2374: free(reg);
2375: }
1.122 schwarze 2376: return(ROFF_IGN);
2377: }
2378:
2379: static enum rofferr
2380: roff_rm(ROFF_ARGS)
2381: {
2382: const char *name;
2383: char *cp;
1.212 schwarze 2384: size_t namesz;
1.122 schwarze 2385:
1.238 schwarze 2386: cp = buf->buf + pos;
2387: while (*cp != '\0') {
1.212 schwarze 2388: name = cp;
1.238 schwarze 2389: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2390: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238 schwarze 2391: if (name[namesz] == '\\')
1.212 schwarze 2392: break;
1.122 schwarze 2393: }
1.178 schwarze 2394: return(ROFF_IGN);
2395: }
2396:
2397: static enum rofferr
2398: roff_it(ROFF_ARGS)
2399: {
2400: int iv;
2401:
2402: /* Parse the number of lines. */
1.261 schwarze 2403:
2404: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2405: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2406: ln, ppos, buf->buf + 1);
1.178 schwarze 2407: return(ROFF_IGN);
2408: }
2409:
1.262 schwarze 2410: while (isspace((unsigned char)buf->buf[pos]))
2411: pos++;
2412:
2413: /*
2414: * Arm the input line trap.
2415: * Special-casing "an-trap" is an ugly workaround to cope
2416: * with DocBook stupidly fiddling with man(7) internals.
2417: */
1.261 schwarze 2418:
1.178 schwarze 2419: roffit_lines = iv;
1.262 schwarze 2420: roffit_macro = mandoc_strdup(iv != 1 ||
2421: strcmp(buf->buf + pos, "an-trap") ?
2422: buf->buf + pos : "br");
1.109 kristaps 2423: return(ROFF_IGN);
1.175 schwarze 2424: }
2425:
2426: static enum rofferr
2427: roff_Dd(ROFF_ARGS)
2428: {
2429: const char *const *cp;
2430:
1.227 schwarze 2431: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175 schwarze 2432: for (cp = __mdoc_reserved; *cp; cp++)
2433: roff_setstr(r, *cp, NULL, 0);
2434:
1.227 schwarze 2435: if (r->format == 0)
2436: r->format = MPARSE_MDOC;
2437:
1.175 schwarze 2438: return(ROFF_CONT);
2439: }
2440:
2441: static enum rofferr
2442: roff_TH(ROFF_ARGS)
2443: {
2444: const char *const *cp;
2445:
1.227 schwarze 2446: if ((r->options & MPARSE_QUICK) == 0)
1.175 schwarze 2447: for (cp = __man_reserved; *cp; cp++)
2448: roff_setstr(r, *cp, NULL, 0);
2449:
1.227 schwarze 2450: if (r->format == 0)
2451: r->format = MPARSE_MAN;
2452:
1.175 schwarze 2453: return(ROFF_CONT);
1.109 kristaps 2454: }
2455:
2456: static enum rofferr
2457: roff_TE(ROFF_ARGS)
2458: {
2459:
2460: if (NULL == r->tbl)
1.221 schwarze 2461: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2462: ln, ppos, "TE");
1.258 schwarze 2463: else if ( ! tbl_end(&r->tbl)) {
2464: free(buf->buf);
2465: buf->buf = mandoc_strdup(".sp");
2466: buf->sz = 4;
2467: return(ROFF_REPARSE);
2468: }
1.112 kristaps 2469: return(ROFF_IGN);
2470: }
2471:
2472: static enum rofferr
2473: roff_T_(ROFF_ARGS)
2474: {
2475:
2476: if (NULL == r->tbl)
1.221 schwarze 2477: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2478: ln, ppos, "T&");
1.112 kristaps 2479: else
1.116 kristaps 2480: tbl_restart(ppos, ln, r->tbl);
1.112 kristaps 2481:
1.109 kristaps 2482: return(ROFF_IGN);
2483: }
2484:
1.230 schwarze 2485: /*
2486: * Handle in-line equation delimiters.
2487: */
2488: static enum rofferr
1.238 schwarze 2489: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2490: {
1.233 schwarze 2491: char *cp1, *cp2;
2492: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2493:
1.230 schwarze 2494: /*
2495: * Outside equations, look for an opening delimiter.
2496: * If we are inside an equation, we already know it is
2497: * in-line, or this function wouldn't have been called;
2498: * so look for a closing delimiter.
2499: */
2500:
1.238 schwarze 2501: cp1 = buf->buf + pos;
1.230 schwarze 2502: cp2 = strchr(cp1, r->eqn == NULL ?
2503: r->last_eqn->odelim : r->last_eqn->cdelim);
2504: if (cp2 == NULL)
2505: return(ROFF_CONT);
2506:
1.233 schwarze 2507: *cp2++ = '\0';
2508: bef_pr = bef_nl = aft_nl = aft_pr = "";
2509:
2510: /* Handle preceding text, protecting whitespace. */
2511:
1.238 schwarze 2512: if (*buf->buf != '\0') {
1.233 schwarze 2513: if (r->eqn == NULL)
2514: bef_pr = "\\&";
2515: bef_nl = "\n";
2516: }
2517:
2518: /*
2519: * Prepare replacing the delimiter with an equation macro
2520: * and drop leading white space from the equation.
2521: */
1.230 schwarze 2522:
1.233 schwarze 2523: if (r->eqn == NULL) {
2524: while (*cp2 == ' ')
2525: cp2++;
2526: mac = ".EQ";
2527: } else
2528: mac = ".EN";
2529:
2530: /* Handle following text, protecting whitespace. */
2531:
2532: if (*cp2 != '\0') {
2533: aft_nl = "\n";
2534: if (r->eqn != NULL)
2535: aft_pr = "\\&";
2536: }
2537:
2538: /* Do the actual replacement. */
2539:
1.238 schwarze 2540: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2541: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2542: free(buf->buf);
2543: buf->buf = cp1;
1.230 schwarze 2544:
2545: /* Toggle the in-line state of the eqn subsystem. */
2546:
2547: r->eqn_inline = r->eqn == NULL;
2548: return(ROFF_REPARSE);
1.151 kristaps 2549: }
2550:
1.235 schwarze 2551: static enum rofferr
2552: roff_EQ(ROFF_ARGS)
1.125 kristaps 2553: {
1.151 kristaps 2554: struct eqn_node *e;
1.125 kristaps 2555:
1.238 schwarze 2556: assert(r->eqn == NULL);
1.235 schwarze 2557: e = eqn_alloc(ppos, ln, r->parse);
1.125 kristaps 2558:
1.230 schwarze 2559: if (r->last_eqn) {
1.125 kristaps 2560: r->last_eqn->next = e;
1.230 schwarze 2561: e->delim = r->last_eqn->delim;
2562: e->odelim = r->last_eqn->odelim;
2563: e->cdelim = r->last_eqn->cdelim;
2564: } else
1.125 kristaps 2565: r->first_eqn = r->last_eqn = e;
2566:
2567: r->eqn = r->last_eqn = e;
1.151 kristaps 2568:
1.238 schwarze 2569: if (buf->buf[pos] != '\0')
1.235 schwarze 2570: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2571: ".EQ %s", buf->buf + pos);
1.151 kristaps 2572:
1.125 kristaps 2573: return(ROFF_IGN);
2574: }
2575:
2576: static enum rofferr
2577: roff_EN(ROFF_ARGS)
2578: {
2579:
1.221 schwarze 2580: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.125 kristaps 2581: return(ROFF_IGN);
2582: }
2583:
2584: static enum rofferr
1.109 kristaps 2585: roff_TS(ROFF_ARGS)
2586: {
1.176 schwarze 2587: struct tbl_node *tbl;
1.89 kristaps 2588:
1.115 kristaps 2589: if (r->tbl) {
1.221 schwarze 2590: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2591: ln, ppos, "TS breaks TS");
1.151 kristaps 2592: tbl_end(&r->tbl);
1.115 kristaps 2593: }
1.83 schwarze 2594:
1.176 schwarze 2595: tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2596:
2597: if (r->last_tbl)
1.176 schwarze 2598: r->last_tbl->next = tbl;
1.113 kristaps 2599: else
1.176 schwarze 2600: r->first_tbl = r->last_tbl = tbl;
1.113 kristaps 2601:
1.176 schwarze 2602: r->tbl = r->last_tbl = tbl;
1.83 schwarze 2603: return(ROFF_IGN);
1.251 schwarze 2604: }
2605:
2606: static enum rofferr
2607: roff_brp(ROFF_ARGS)
2608: {
2609:
2610: buf->buf[pos - 1] = '\0';
2611: return(ROFF_CONT);
1.92 schwarze 2612: }
2613:
1.105 kristaps 2614: static enum rofferr
1.174 kristaps 2615: roff_cc(ROFF_ARGS)
2616: {
2617: const char *p;
2618:
1.238 schwarze 2619: p = buf->buf + pos;
1.174 kristaps 2620:
1.238 schwarze 2621: if (*p == '\0' || (r->control = *p++) == '.')
1.174 kristaps 2622: r->control = 0;
2623:
1.238 schwarze 2624: if (*p != '\0')
1.260 schwarze 2625: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2626: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 2627:
2628: return(ROFF_IGN);
2629: }
2630:
2631: static enum rofferr
1.164 kristaps 2632: roff_tr(ROFF_ARGS)
2633: {
2634: const char *p, *first, *second;
2635: size_t fsz, ssz;
2636: enum mandoc_esc esc;
2637:
1.238 schwarze 2638: p = buf->buf + pos;
1.164 kristaps 2639:
1.238 schwarze 2640: if (*p == '\0') {
1.260 schwarze 2641: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.164 kristaps 2642: return(ROFF_IGN);
2643: }
2644:
1.238 schwarze 2645: while (*p != '\0') {
1.164 kristaps 2646: fsz = ssz = 1;
2647:
2648: first = p++;
1.238 schwarze 2649: if (*first == '\\') {
1.164 kristaps 2650: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2651: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2652: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2653: ln, (int)(p - buf->buf), first);
1.164 kristaps 2654: return(ROFF_IGN);
2655: }
2656: fsz = (size_t)(p - first);
2657: }
2658:
2659: second = p++;
1.238 schwarze 2660: if (*second == '\\') {
1.164 kristaps 2661: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2662: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2663: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2664: ln, (int)(p - buf->buf), second);
1.164 kristaps 2665: return(ROFF_IGN);
2666: }
2667: ssz = (size_t)(p - second);
1.238 schwarze 2668: } else if (*second == '\0') {
1.260 schwarze 2669: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2670: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 2671: second = " ";
1.165 kristaps 2672: p--;
1.164 kristaps 2673: }
2674:
1.167 kristaps 2675: if (fsz > 1) {
1.207 schwarze 2676: roff_setstrn(&r->xmbtab, first, fsz,
2677: second, ssz, 0);
1.167 kristaps 2678: continue;
2679: }
2680:
1.238 schwarze 2681: if (r->xtab == NULL)
1.207 schwarze 2682: r->xtab = mandoc_calloc(128,
2683: sizeof(struct roffstr));
1.167 kristaps 2684:
2685: free(r->xtab[(int)*first].p);
2686: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
2687: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 2688: }
2689:
2690: return(ROFF_IGN);
2691: }
2692:
2693: static enum rofferr
1.105 kristaps 2694: roff_so(ROFF_ARGS)
2695: {
1.249 schwarze 2696: char *name, *cp;
1.105 kristaps 2697:
1.238 schwarze 2698: name = buf->buf + pos;
1.224 schwarze 2699: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 2700:
2701: /*
2702: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
2703: * opening anything that's not in our cwd or anything beneath
2704: * it. Thus, explicitly disallow traversing up the file-system
2705: * or using absolute paths.
2706: */
2707:
1.238 schwarze 2708: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 2709: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
2710: ".so %s", name);
1.249 schwarze 2711: buf->sz = mandoc_asprintf(&cp,
2712: ".sp\nSee the file %s.\n.sp", name) + 1;
2713: free(buf->buf);
2714: buf->buf = cp;
2715: *offs = 0;
2716: return(ROFF_REPARSE);
1.105 kristaps 2717: }
2718:
2719: *offs = pos;
2720: return(ROFF_SO);
2721: }
1.92 schwarze 2722:
1.106 kristaps 2723: static enum rofferr
2724: roff_userdef(ROFF_ARGS)
1.99 kristaps 2725: {
1.263 schwarze 2726: const char *arg[9], *ap;
1.106 kristaps 2727: char *cp, *n1, *n2;
1.119 schwarze 2728: int i;
1.263 schwarze 2729: size_t asz, rsz;
1.106 kristaps 2730:
2731: /*
2732: * Collect pointers to macro argument strings
1.188 schwarze 2733: * and NUL-terminate them.
1.106 kristaps 2734: */
1.263 schwarze 2735:
1.238 schwarze 2736: cp = buf->buf + pos;
1.119 schwarze 2737: for (i = 0; i < 9; i++)
1.238 schwarze 2738: arg[i] = *cp == '\0' ? "" :
1.136 kristaps 2739: mandoc_getarg(r->parse, &cp, ln, &pos);
1.99 kristaps 2740:
1.106 kristaps 2741: /*
2742: * Expand macro arguments.
1.99 kristaps 2743: */
1.263 schwarze 2744:
2745: buf->sz = strlen(r->current_string) + 1;
2746: n1 = cp = mandoc_malloc(buf->sz);
2747: memcpy(n1, r->current_string, buf->sz);
2748: while (*cp != '\0') {
2749:
2750: /* Scan ahead for the next argument invocation. */
2751:
2752: if (*cp++ != '\\')
2753: continue;
2754: if (*cp++ != '$')
2755: continue;
2756: i = *cp - '1';
2757: if (0 > i || 8 < i)
1.106 kristaps 2758: continue;
1.263 schwarze 2759: cp -= 2;
2760:
2761: /*
2762: * Determine the size of the expanded argument,
2763: * taking escaping of quotes into account.
2764: */
2765:
2766: asz = 0;
2767: for (ap = arg[i]; *ap != '\0'; ap++) {
2768: asz++;
2769: if (*ap == '"')
2770: asz += 3;
2771: }
2772: if (asz != 3) {
2773:
2774: /*
2775: * Determine the size of the rest of the
2776: * unexpanded macro, including the NUL.
2777: */
2778:
2779: rsz = buf->sz - (cp - n1) - 3;
2780:
2781: /*
2782: * When shrinking, move before
2783: * releasing the storage.
2784: */
2785:
2786: if (asz < 3)
2787: memmove(cp + asz, cp + 3, rsz);
2788:
2789: /*
2790: * Resize the storage for the macro
2791: * and readjust the parse pointer.
2792: */
2793:
2794: buf->sz += asz - 3;
2795: n2 = mandoc_realloc(n1, buf->sz);
2796: cp = n2 + (cp - n1);
2797: n1 = n2;
2798:
2799: /*
2800: * When growing, make room
2801: * for the expanded argument.
2802: */
2803:
2804: if (asz > 3)
2805: memmove(cp + asz, cp + 3, rsz);
2806: }
2807:
2808: /* Copy the expanded argument, escaping quotes. */
2809:
2810: n2 = cp;
2811: for (ap = arg[i]; *ap != '\0'; ap++) {
2812: if (*ap == '"') {
2813: memcpy(n2, "\\(dq", 4);
2814: n2 += 4;
2815: } else
2816: *n2++ = *ap;
1.106 kristaps 2817: }
1.99 kristaps 2818: }
2819:
1.106 kristaps 2820: /*
2821: * Replace the macro invocation
2822: * by the expanded macro.
2823: */
1.263 schwarze 2824:
1.238 schwarze 2825: free(buf->buf);
2826: buf->buf = n1;
1.248 schwarze 2827: *offs = 0;
1.106 kristaps 2828:
1.238 schwarze 2829: return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.106 kristaps 2830: ROFF_REPARSE : ROFF_APPEND);
1.99 kristaps 2831: }
1.121 schwarze 2832:
1.212 schwarze 2833: static size_t
1.121 schwarze 2834: roff_getname(struct roff *r, char **cpp, int ln, int pos)
2835: {
2836: char *name, *cp;
1.212 schwarze 2837: size_t namesz;
1.121 schwarze 2838:
2839: name = *cpp;
2840: if ('\0' == *name)
1.212 schwarze 2841: return(0);
1.121 schwarze 2842:
1.212 schwarze 2843: /* Read until end of name and terminate it with NUL. */
2844: for (cp = name; 1; cp++) {
2845: if ('\0' == *cp || ' ' == *cp) {
2846: namesz = cp - name;
2847: break;
2848: }
1.121 schwarze 2849: if ('\\' != *cp)
2850: continue;
1.215 schwarze 2851: namesz = cp - name;
2852: if ('{' == cp[1] || '}' == cp[1])
2853: break;
1.121 schwarze 2854: cp++;
2855: if ('\\' == *cp)
2856: continue;
1.224 schwarze 2857: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
2858: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 2859: mandoc_escape((const char **)&cp, NULL, NULL);
2860: break;
1.121 schwarze 2861: }
2862:
2863: /* Read past spaces. */
2864: while (' ' == *cp)
2865: cp++;
2866:
2867: *cpp = cp;
1.212 schwarze 2868: return(namesz);
1.121 schwarze 2869: }
2870:
1.106 kristaps 2871: /*
2872: * Store *string into the user-defined string called *name.
2873: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 2874: * append == 0: replace mode
2875: * append == 1: single-line append mode
2876: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 2877: */
1.94 kristaps 2878: static void
1.106 kristaps 2879: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 2880: int append)
1.92 schwarze 2881: {
1.164 kristaps 2882:
2883: roff_setstrn(&r->strtab, name, strlen(name), string,
1.207 schwarze 2884: string ? strlen(string) : 0, append);
1.164 kristaps 2885: }
2886:
2887: static void
1.166 kristaps 2888: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 2889: const char *string, size_t stringsz, int append)
1.164 kristaps 2890: {
1.166 kristaps 2891: struct roffkv *n;
1.164 kristaps 2892: char *c;
2893: int i;
2894: size_t oldch, newch;
1.92 schwarze 2895:
1.106 kristaps 2896: /* Search for an existing string with the same name. */
1.164 kristaps 2897: n = *r;
2898:
1.211 schwarze 2899: while (n && (namesz != n->key.sz ||
2900: strncmp(n->key.p, name, namesz)))
1.92 schwarze 2901: n = n->next;
1.94 kristaps 2902:
2903: if (NULL == n) {
1.106 kristaps 2904: /* Create a new string table entry. */
1.166 kristaps 2905: n = mandoc_malloc(sizeof(struct roffkv));
2906: n->key.p = mandoc_strndup(name, namesz);
2907: n->key.sz = namesz;
2908: n->val.p = NULL;
2909: n->val.sz = 0;
1.164 kristaps 2910: n->next = *r;
2911: *r = n;
1.193 schwarze 2912: } else if (0 == append) {
1.166 kristaps 2913: free(n->val.p);
2914: n->val.p = NULL;
2915: n->val.sz = 0;
1.106 kristaps 2916: }
2917:
2918: if (NULL == string)
2919: return;
2920:
2921: /*
2922: * One additional byte for the '\n' in multiline mode,
2923: * and one for the terminating '\0'.
2924: */
1.193 schwarze 2925: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 2926:
1.166 kristaps 2927: if (NULL == n->val.p) {
2928: n->val.p = mandoc_malloc(newch);
2929: *n->val.p = '\0';
1.106 kristaps 2930: oldch = 0;
2931: } else {
1.166 kristaps 2932: oldch = n->val.sz;
2933: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 2934: }
2935:
2936: /* Skip existing content in the destination buffer. */
1.166 kristaps 2937: c = n->val.p + (int)oldch;
1.106 kristaps 2938:
2939: /* Append new content to the destination buffer. */
1.164 kristaps 2940: i = 0;
2941: while (i < (int)stringsz) {
1.106 kristaps 2942: /*
2943: * Rudimentary roff copy mode:
2944: * Handle escaped backslashes.
2945: */
1.164 kristaps 2946: if ('\\' == string[i] && '\\' == string[i + 1])
2947: i++;
2948: *c++ = string[i++];
1.106 kristaps 2949: }
1.94 kristaps 2950:
1.106 kristaps 2951: /* Append terminating bytes. */
1.193 schwarze 2952: if (1 < append)
1.106 kristaps 2953: *c++ = '\n';
1.163 kristaps 2954:
1.106 kristaps 2955: *c = '\0';
1.166 kristaps 2956: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 2957: }
2958:
1.94 kristaps 2959: static const char *
2960: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92 schwarze 2961: {
1.166 kristaps 2962: const struct roffkv *n;
1.191 schwarze 2963: int i;
1.92 schwarze 2964:
1.164 kristaps 2965: for (n = r->strtab; n; n = n->next)
1.207 schwarze 2966: if (0 == strncmp(name, n->key.p, len) &&
2967: '\0' == n->key.p[(int)len])
1.166 kristaps 2968: return(n->val.p);
1.191 schwarze 2969:
2970: for (i = 0; i < PREDEFS_MAX; i++)
2971: if (0 == strncmp(name, predefs[i].name, len) &&
2972: '\0' == predefs[i].name[(int)len])
2973: return(predefs[i].str);
1.94 kristaps 2974:
1.157 kristaps 2975: return(NULL);
1.92 schwarze 2976: }
2977:
1.94 kristaps 2978: static void
1.167 kristaps 2979: roff_freestr(struct roffkv *r)
1.92 schwarze 2980: {
1.166 kristaps 2981: struct roffkv *n, *nn;
1.92 schwarze 2982:
1.167 kristaps 2983: for (n = r; n; n = nn) {
1.166 kristaps 2984: free(n->key.p);
2985: free(n->val.p);
1.92 schwarze 2986: nn = n->next;
2987: free(n);
2988: }
1.114 kristaps 2989: }
2990:
2991: const struct tbl_span *
2992: roff_span(const struct roff *r)
2993: {
1.207 schwarze 2994:
1.114 kristaps 2995: return(r->tbl ? tbl_span(r->tbl) : NULL);
1.125 kristaps 2996: }
2997:
2998: const struct eqn *
2999: roff_eqn(const struct roff *r)
3000: {
1.207 schwarze 3001:
1.125 kristaps 3002: return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.164 kristaps 3003: }
3004:
3005: /*
3006: * Duplicate an input string, making the appropriate character
3007: * conversations (as stipulated by `tr') along the way.
3008: * Returns a heap-allocated string with all the replacements made.
3009: */
3010: char *
3011: roff_strdup(const struct roff *r, const char *p)
3012: {
1.166 kristaps 3013: const struct roffkv *cp;
1.164 kristaps 3014: char *res;
3015: const char *pp;
3016: size_t ssz, sz;
3017: enum mandoc_esc esc;
3018:
1.167 kristaps 3019: if (NULL == r->xmbtab && NULL == r->xtab)
1.164 kristaps 3020: return(mandoc_strdup(p));
3021: else if ('\0' == *p)
3022: return(mandoc_strdup(""));
3023:
3024: /*
3025: * Step through each character looking for term matches
3026: * (remember that a `tr' can be invoked with an escape, which is
3027: * a glyph but the escape is multi-character).
3028: * We only do this if the character hash has been initialised
3029: * and the string is >0 length.
3030: */
3031:
3032: res = NULL;
3033: ssz = 0;
3034:
3035: while ('\0' != *p) {
1.167 kristaps 3036: if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
3037: sz = r->xtab[(int)*p].sz;
3038: res = mandoc_realloc(res, ssz + sz + 1);
3039: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3040: ssz += sz;
3041: p++;
3042: continue;
3043: } else if ('\\' != *p) {
3044: res = mandoc_realloc(res, ssz + 2);
3045: res[ssz++] = *p++;
3046: continue;
3047: }
3048:
1.164 kristaps 3049: /* Search for term matches. */
1.167 kristaps 3050: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3051: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3052: break;
3053:
3054: if (NULL != cp) {
3055: /*
3056: * A match has been found.
3057: * Append the match to the array and move
3058: * forward by its keysize.
3059: */
1.207 schwarze 3060: res = mandoc_realloc(res,
3061: ssz + cp->val.sz + 1);
1.166 kristaps 3062: memcpy(res + ssz, cp->val.p, cp->val.sz);
3063: ssz += cp->val.sz;
3064: p += (int)cp->key.sz;
1.164 kristaps 3065: continue;
3066: }
3067:
1.167 kristaps 3068: /*
3069: * Handle escapes carefully: we need to copy
3070: * over just the escape itself, or else we might
3071: * do replacements within the escape itself.
3072: * Make sure to pass along the bogus string.
3073: */
3074: pp = p++;
3075: esc = mandoc_escape(&p, NULL, NULL);
3076: if (ESCAPE_ERROR == esc) {
3077: sz = strlen(pp);
1.164 kristaps 3078: res = mandoc_realloc(res, ssz + sz + 1);
3079: memcpy(res + ssz, pp, sz);
1.167 kristaps 3080: break;
1.164 kristaps 3081: }
1.207 schwarze 3082: /*
3083: * We bail out on bad escapes.
1.167 kristaps 3084: * No need to warn: we already did so when
3085: * roff_res() was called.
3086: */
3087: sz = (int)(p - pp);
3088: res = mandoc_realloc(res, ssz + sz + 1);
3089: memcpy(res + ssz, pp, sz);
3090: ssz += sz;
1.164 kristaps 3091: }
3092:
3093: res[(int)ssz] = '\0';
3094: return(res);
1.227 schwarze 3095: }
3096:
3097: int
3098: roff_getformat(const struct roff *r)
3099: {
3100:
3101: return(r->format);
1.174 kristaps 3102: }
3103:
3104: /*
1.207 schwarze 3105: * Find out whether a line is a macro line or not.
1.174 kristaps 3106: * If it is, adjust the current position and return one; if it isn't,
3107: * return zero and don't change the current position.
3108: * If the control character has been set with `.cc', then let that grain
3109: * precedence.
3110: * This is slighly contrary to groff, where using the non-breaking
3111: * control character when `cc' has been invoked will cause the
3112: * non-breaking macro contents to be printed verbatim.
3113: */
3114: int
3115: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3116: {
3117: int pos;
3118:
3119: pos = *ppos;
3120:
3121: if (0 != r->control && cp[pos] == r->control)
3122: pos++;
3123: else if (0 != r->control)
3124: return(0);
3125: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3126: pos += 2;
3127: else if ('.' == cp[pos] || '\'' == cp[pos])
3128: pos++;
CVSweb