xref: /openbsd/usr.bin/cvs/rcsparse.c (revision cca36db2)
1 /*	$OpenBSD: rcsparse.c,v 1.6 2012/02/04 21:22:32 tobias Exp $	*/
2 /*
3  * Copyright (c) 2010 Tobias Stoeckmann <tobias@openbsd.org>
4  *
5  * Permission to use, copy, modify, and distribute this software for any
6  * purpose with or without fee is hereby granted, provided that the above
7  * copyright notice and this permission notice appear in all copies.
8  *
9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
16  */
17 
18 #include <sys/queue.h>
19 
20 #include <ctype.h>
21 #include <err.h>
22 #include <pwd.h>
23 #include <stdarg.h>
24 #include <stdio.h>
25 #include <stdlib.h>
26 #include <string.h>
27 #include <unistd.h>
28 
29 #include "log.h"
30 #include "rcs.h"
31 #include "rcsparse.h"
32 #include "xmalloc.h"
33 
34 #define RCS_BUFSIZE	16384
35 #define RCS_BUFEXTSIZE	8192
36 
37 /* RCS token types */
38 #define RCS_TOK_HEAD		(1 << 0)
39 #define RCS_TOK_BRANCH		(1 << 1)
40 #define RCS_TOK_ACCESS		(1 << 2)
41 #define RCS_TOK_SYMBOLS		(1 << 3)
42 #define RCS_TOK_LOCKS		(1 << 4)
43 #define RCS_TOK_STRICT		(1 << 5)
44 #define RCS_TOK_COMMENT		(1 << 6)
45 #define RCS_TOK_COMMITID	(1 << 7)
46 #define RCS_TOK_EXPAND		(1 << 8)
47 #define RCS_TOK_DESC		(1 << 9)
48 #define RCS_TOK_DATE		(1 << 10)
49 #define RCS_TOK_AUTHOR		(1 << 11)
50 #define RCS_TOK_STATE		(1 << 12)
51 #define RCS_TOK_BRANCHES	(1 << 13)
52 #define RCS_TOK_NEXT		(1 << 14)
53 #define RCS_TOK_LOG		(1 << 15)
54 #define RCS_TOK_TEXT		(1 << 16)
55 #define RCS_TOK_COLON		(1 << 17)
56 #define RCS_TOK_COMMA		(1 << 18)
57 #define RCS_TOK_SCOLON		(1 << 19)
58 
59 #define RCS_TYPE_STRING		(1 << 20)
60 #define RCS_TYPE_NUMBER		(1 << 21)
61 #define RCS_TYPE_BRANCH		(1 << 22)
62 #define RCS_TYPE_REVISION	(1 << 23)
63 #define RCS_TYPE_LOGIN		(1 << 24)
64 #define RCS_TYPE_STATE		(1 << 25)
65 #define RCS_TYPE_SYMBOL		(1 << 26)
66 #define RCS_TYPE_DATE		(1 << 27)
67 #define RCS_TYPE_KEYWORD	(1 << 28)
68 #define RCS_TYPE_COMMITID	(1 << 29)
69 
70 #define MANDATORY	0
71 #define OPTIONAL	1
72 
73 /* opaque parse data */
74 struct rcs_pdata {
75 	char			*rp_buf;
76 	size_t			 rp_blen;
77 	char			*rp_bufend;
78 	size_t			 rp_tlen;
79 
80 	struct rcs_delta	*rp_delta;
81 	int			 rp_lineno;
82 	int			 rp_msglineno;
83 	int			 rp_token;
84 
85 	union {
86 		RCSNUM		*rev;
87 		char		*str;
88 		struct tm	 date;
89 	} rp_value;
90 };
91 
92 struct rcs_keyword {
93 	const char	*k_name;
94 	int		 k_val;
95 };
96 
97 struct rcs_section {
98 	int	token;
99 	int	(*parse)(RCSFILE *, struct rcs_pdata *);
100 	int	opt;
101 };
102 
103 /* this has to be sorted always */
104 static const struct rcs_keyword keywords[] = {
105 	{ "access",		RCS_TOK_ACCESS},
106 	{ "author",		RCS_TOK_AUTHOR},
107 	{ "branch",		RCS_TOK_BRANCH},
108 	{ "branches",		RCS_TOK_BRANCHES},
109 	{ "comment",		RCS_TOK_COMMENT},
110 	{ "date",		RCS_TOK_DATE},
111 	{ "desc",		RCS_TOK_DESC},
112 	{ "expand",		RCS_TOK_EXPAND},
113 	{ "head",		RCS_TOK_HEAD},
114 	{ "locks",		RCS_TOK_LOCKS},
115 	{ "log",		RCS_TOK_LOG},
116 	{ "next",		RCS_TOK_NEXT},
117 	{ "state",		RCS_TOK_STATE},
118 	{ "strict",		RCS_TOK_STRICT},
119 	{ "symbols",		RCS_TOK_SYMBOLS},
120 	{ "text",		RCS_TOK_TEXT}
121 };
122 
123 /* parser functions specified in rcs_section structs */
124 static int	rcsparse_head(RCSFILE *, struct rcs_pdata *);
125 static int	rcsparse_branch(RCSFILE *, struct rcs_pdata *);
126 static int	rcsparse_access(RCSFILE *, struct rcs_pdata *);
127 static int	rcsparse_symbols(RCSFILE *, struct rcs_pdata *);
128 static int	rcsparse_locks(RCSFILE *, struct rcs_pdata *);
129 static int	rcsparse_strict(RCSFILE *, struct rcs_pdata *);
130 static int	rcsparse_comment(RCSFILE *, struct rcs_pdata *);
131 static int	rcsparse_commitid(RCSFILE *, struct rcs_pdata *);
132 static int	rcsparse_expand(RCSFILE *, struct rcs_pdata *);
133 static int	rcsparse_deltarevision(RCSFILE *, struct rcs_pdata *);
134 static int	rcsparse_date(RCSFILE *, struct rcs_pdata *);
135 static int	rcsparse_author(RCSFILE *, struct rcs_pdata *);
136 static int	rcsparse_state(RCSFILE *, struct rcs_pdata *);
137 static int	rcsparse_branches(RCSFILE *, struct rcs_pdata *);
138 static int	rcsparse_next(RCSFILE *, struct rcs_pdata *);
139 static int	rcsparse_textrevision(RCSFILE *, struct rcs_pdata *);
140 static int	rcsparse_log(RCSFILE *, struct rcs_pdata *);
141 static int	rcsparse_text(RCSFILE *, struct rcs_pdata *);
142 
143 static int	rcsparse_delta(RCSFILE *);
144 static int	rcsparse_deltatext(RCSFILE *);
145 static int	rcsparse_desc(RCSFILE *);
146 
147 static int	kw_cmp(const void *, const void *);
148 static int	rcsparse(RCSFILE *, struct rcs_section *);
149 static void	rcsparse_growbuf(RCSFILE *);
150 static int	rcsparse_string(RCSFILE *, int);
151 static int	rcsparse_token(RCSFILE *, int);
152 static void	rcsparse_warnx(RCSFILE *, char *, ...);
153 static int	valid_login(char *);
154 
155 /*
156  * head [REVISION];
157  * [branch BRANCH];
158  * access [LOGIN ...];
159  * symbols [SYMBOL:REVISION ...];
160  * locks [LOGIN:REVISION ...];
161  * [strict;]
162  * [comment [@[...]@];]
163  * [expand [@[...]@];]
164  */
165 static struct rcs_section sec_admin[] = {
166 	{ RCS_TOK_HEAD, rcsparse_head, MANDATORY },
167 	{ RCS_TOK_BRANCH, rcsparse_branch, OPTIONAL },
168 	{ RCS_TOK_ACCESS, rcsparse_access, MANDATORY },
169 	{ RCS_TOK_SYMBOLS, rcsparse_symbols, MANDATORY },
170 	{ RCS_TOK_LOCKS, rcsparse_locks, MANDATORY },
171 	{ RCS_TOK_STRICT, rcsparse_strict, OPTIONAL },
172 	{ RCS_TOK_COMMENT, rcsparse_comment, OPTIONAL },
173 	{ RCS_TOK_EXPAND, rcsparse_expand, OPTIONAL },
174 	{ 0, NULL, 0 }
175 };
176 
177 /*
178  * REVISION
179  * date [YY]YY.MM.DD.HH.MM.SS;
180  * author LOGIN;
181  * state STATE;
182  * branches [REVISION ...];
183  * next [REVISION];
184  * [commitid ID;]
185  */
186 static struct rcs_section sec_delta[] = {
187 	{ RCS_TYPE_REVISION, rcsparse_deltarevision, MANDATORY },
188 	{ RCS_TOK_DATE, rcsparse_date, MANDATORY },
189 	{ RCS_TOK_AUTHOR, rcsparse_author, MANDATORY },
190 	{ RCS_TOK_STATE, rcsparse_state, MANDATORY },
191 	{ RCS_TOK_BRANCHES, rcsparse_branches, MANDATORY },
192 	{ RCS_TOK_NEXT, rcsparse_next, MANDATORY },
193 	{ RCS_TOK_COMMITID, rcsparse_commitid, OPTIONAL },
194 	{ 0, NULL, 0 }
195 };
196 
197 /*
198  * REVISION
199  * log @[...]@
200  * text @[...]@
201  */
202 static struct rcs_section sec_deltatext[] = {
203 	{ RCS_TYPE_REVISION, rcsparse_textrevision, MANDATORY },
204 	{ RCS_TOK_LOG, rcsparse_log, MANDATORY },
205 	{ RCS_TOK_TEXT, rcsparse_text, MANDATORY },
206 	{ 0, NULL, 0 }
207 };
208 
209 /*
210  * rcsparse_init()
211  *
212  * Initializes the parsing data structure and parses the admin section of
213  * RCS file <rfp>.
214  *
215  * Returns 0 on success or 1 on failure.
216  */
217 int
218 rcsparse_init(RCSFILE *rfp)
219 {
220 	struct rcs_pdata *pdp;
221 
222 	if (rfp->rf_flags & RCS_PARSED)
223 		return (0);
224 
225 	pdp = xmalloc(sizeof(*pdp));
226 	pdp->rp_buf = xmalloc(RCS_BUFSIZE);
227 	pdp->rp_blen = RCS_BUFSIZE;
228 	pdp->rp_bufend = pdp->rp_buf + pdp->rp_blen - 1;
229 	pdp->rp_token = -1;
230 	pdp->rp_lineno = 1;
231 	pdp->rp_msglineno = 1;
232 
233 	/* ditch the strict lock */
234 	rfp->rf_flags &= ~RCS_SLOCK;
235 	rfp->rf_pdata = pdp;
236 
237 	if (rcsparse(rfp, sec_admin)) {
238 		rcsparse_free(rfp);
239 		return (1);
240 	}
241 
242 	if ((rfp->rf_flags & RCS_PARSE_FULLY) &&
243 	    rcsparse_deltatexts(rfp, NULL)) {
244 		rcsparse_free(rfp);
245 		return (1);
246 	}
247 
248 	rfp->rf_flags |= RCS_SYNCED;
249 	return (0);
250 }
251 
252 /*
253  * rcsparse_deltas()
254  *
255  * Parse deltas. If <rev> is not NULL, parse only as far as that
256  * revision. If <rev> is NULL, parse all deltas.
257  *
258  * Returns 0 on success or 1 on error.
259  */
260 int
261 rcsparse_deltas(RCSFILE *rfp, RCSNUM *rev)
262 {
263 	int ret;
264 	struct rcs_delta *enddelta;
265 
266 	if ((rfp->rf_flags & PARSED_DELTAS) || (rfp->rf_flags & RCS_CREATE))
267 		return (0);
268 
269 	for (;;) {
270 		ret = rcsparse_delta(rfp);
271 		if (rev != NULL) {
272 			enddelta = TAILQ_LAST(&(rfp->rf_delta), rcs_dlist);
273 			if (enddelta == NULL)
274 				return (1);
275 
276 			if (rcsnum_cmp(enddelta->rd_num, rev, 0) == 0)
277 				break;
278 		}
279 
280 		if (ret == 0) {
281 			rfp->rf_flags |= PARSED_DELTAS;
282 			break;
283 		}
284 		else if (ret == -1)
285 			return (1);
286 	}
287 
288 	return (0);
289 }
290 
291 /*
292  * rcsparse_deltatexts()
293  *
294  * Parse deltatexts. If <rev> is not NULL, parse only as far as that
295  * revision. If <rev> is NULL, parse everything.
296  *
297  * Returns 0 on success or 1 on error.
298  */
299 int
300 rcsparse_deltatexts(RCSFILE *rfp, RCSNUM *rev)
301 {
302 	int ret;
303 	struct rcs_delta *rdp;
304 
305 	if ((rfp->rf_flags & PARSED_DELTATEXTS) ||
306 	    (rfp->rf_flags & RCS_CREATE))
307 		return (0);
308 
309 	if (!(rfp->rf_flags & PARSED_DESC))
310 		if (rcsparse_desc(rfp))
311 			return (1);
312 
313 	rdp = (rev != NULL) ? rcs_findrev(rfp, rev) : NULL;
314 
315 	for (;;) {
316 		if (rdp != NULL && rdp->rd_text != NULL)
317 			break;
318 		ret = rcsparse_deltatext(rfp);
319 		if (ret == 0) {
320 			rfp->rf_flags |= PARSED_DELTATEXTS;
321 			break;
322 		}
323 		else if (ret == -1)
324 			return (1);
325 	}
326 
327 	return (0);
328 }
329 
330 /*
331  * rcsparse_free()
332  *
333  * Free the contents of the <rfp>'s parser data structure.
334  */
335 void
336 rcsparse_free(RCSFILE *rfp)
337 {
338 	struct rcs_pdata *pdp;
339 
340 	pdp = rfp->rf_pdata;
341 
342 	if (pdp->rp_buf != NULL)
343 		xfree(pdp->rp_buf);
344 	if (pdp->rp_token == RCS_TYPE_REVISION)
345 		rcsnum_free(pdp->rp_value.rev);
346 	xfree(pdp);
347 }
348 
349 /*
350  * rcsparse_desc()
351  *
352  * Parse desc of the RCS file <rfp>.  By calling rcsparse_desc, all deltas
353  * will be parsed in order to proceed the reading cursor to the desc keyword.
354  *
355  * desc @[...]@;
356  *
357  * Returns 0 on success or 1 on error.
358  */
359 static int
360 rcsparse_desc(RCSFILE *rfp)
361 {
362 	struct rcs_pdata *pdp;
363 
364 	if (rfp->rf_flags & PARSED_DESC)
365 		return (0);
366 
367 	if (!(rfp->rf_flags & PARSED_DELTAS) && rcsparse_deltas(rfp, NULL))
368 		return (1);
369 
370 	pdp = (struct rcs_pdata *)rfp->rf_pdata;
371 
372 	if (rcsparse_token(rfp, RCS_TOK_DESC) != RCS_TOK_DESC ||
373 	    rcsparse_token(rfp, RCS_TYPE_STRING) != RCS_TYPE_STRING)
374 		return (1);
375 
376 	rfp->rf_desc = pdp->rp_value.str;
377 	rfp->rf_flags |= PARSED_DESC;
378 
379 	return (0);
380 }
381 
382 /*
383  * rcsparse_deltarevision()
384  *
385  * Called upon reaching a new REVISION entry in the delta section.
386  * A new rcs_delta structure will be prepared in pdp->rp_delta for further
387  * parsing.
388  *
389  * REVISION
390  *
391  * Always returns 0.
392  */
393 static int
394 rcsparse_deltarevision(RCSFILE *rfp, struct rcs_pdata *pdp)
395 {
396 	struct rcs_delta *rdp;
397 
398 	rdp = xcalloc(1, sizeof(*rdp));
399 	TAILQ_INIT(&rdp->rd_branches);
400 	rdp->rd_num = pdp->rp_value.rev;
401 	pdp->rp_delta = rdp;
402 
403 	return (0);
404 }
405 
406 /*
407  * rcsparse_date()
408  *
409  * Parses the specified date of current delta pdp->rp_delta.
410  *
411  * date YYYY.MM.DD.HH.MM.SS;
412  *
413  * Returns 0 on success or 1 on failure.
414  */
415 static int
416 rcsparse_date(RCSFILE *rfp, struct rcs_pdata *pdp)
417 {
418 	if (rcsparse_token(rfp, RCS_TYPE_DATE) != RCS_TYPE_DATE)
419 		return (1);
420 
421 	pdp->rp_delta->rd_date = pdp->rp_value.date;
422 
423 	return (rcsparse_token(rfp, RCS_TOK_SCOLON) != RCS_TOK_SCOLON);
424 }
425 
426 /*
427  * rcsparse_author()
428  *
429  * Parses the specified author of current delta pdp->rp_delta.
430  *
431  * author LOGIN;
432  *
433  * Returns 0 on success or 1 on failure.
434  */
435 static int
436 rcsparse_author(RCSFILE *rfp, struct rcs_pdata *pdp)
437 {
438 	if (rcsparse_token(rfp, RCS_TYPE_LOGIN) != RCS_TYPE_LOGIN)
439 		return (1);
440 
441 	pdp->rp_delta->rd_author = pdp->rp_value.str;
442 
443 	return (rcsparse_token(rfp, RCS_TOK_SCOLON) != RCS_TOK_SCOLON);
444 }
445 
446 /*
447  * rcsparse_state()
448  *
449  * Parses the specified state of current delta pdp->rp_delta.
450  *
451  * state STATE;
452  *
453  * Returns 0 on success or 1 on failure.
454  */
455 static int
456 rcsparse_state(RCSFILE *rfp, struct rcs_pdata *pdp)
457 {
458 	if (rcsparse_token(rfp, RCS_TYPE_STATE) != RCS_TYPE_STATE)
459 		return (1);
460 
461 	pdp->rp_delta->rd_state = pdp->rp_value.str;
462 
463 	return (rcsparse_token(rfp, RCS_TOK_SCOLON) != RCS_TOK_SCOLON);
464 }
465 
466 /*
467  * rcsparse_branches()
468  *
469  * Parses the specified branches of current delta pdp->rp_delta.
470  *
471  * branches [REVISION ...];
472  *
473  * Returns 0 on success or 1 on failure.
474  */
475 static int
476 rcsparse_branches(RCSFILE *rfp, struct rcs_pdata *pdp)
477 {
478 	struct rcs_branch *rb;
479 	int type;
480 
481 	while ((type = rcsparse_token(rfp, RCS_TOK_SCOLON|RCS_TYPE_REVISION))
482 	    == RCS_TYPE_REVISION) {
483 		rb = xmalloc(sizeof(*rb));
484 		rb->rb_num = pdp->rp_value.rev;
485 		TAILQ_INSERT_TAIL(&(pdp->rp_delta->rd_branches), rb, rb_list);
486 	}
487 
488 	return (type != RCS_TOK_SCOLON);
489 }
490 
491 /*
492  * rcsparse_next()
493  *
494  * Parses the specified next revision of current delta pdp->rp_delta.
495  *
496  * next [REVISION];
497  *
498  * Returns 0 on success or 1 on failure.
499  */
500 static int
501 rcsparse_next(RCSFILE *rfp, struct rcs_pdata *pdp)
502 {
503 	int type;
504 
505 	type = rcsparse_token(rfp, RCS_TYPE_REVISION|RCS_TOK_SCOLON);
506 	if (type == RCS_TYPE_REVISION) {
507 		pdp->rp_delta->rd_next = pdp->rp_value.rev;
508 		type = rcsparse_token(rfp, RCS_TOK_SCOLON);
509 	} else
510 		pdp->rp_delta->rd_next = rcsnum_alloc();
511 
512 	return (type != RCS_TOK_SCOLON);
513 }
514 
515 /*
516  * rcsparse_commitid()
517  *
518  * Parses the specified commit id of current delta pdp->rp_delta. The
519  * commitid keyword is optional and can be omitted.
520  *
521  * [commitid ID;]
522  *
523  * Returns 0 on success or 1 on failure.
524  */
525 static int
526 rcsparse_commitid(RCSFILE *rfp, struct rcs_pdata *pdp)
527 {
528 	if (rcsparse_token(rfp, RCS_TYPE_COMMITID) != RCS_TYPE_COMMITID)
529 		return (1);
530 
531 	/* XXX - do something with commitid */
532 
533 	return (rcsparse_token(rfp, RCS_TOK_SCOLON) != RCS_TOK_SCOLON);
534 }
535 
536 /*
537  * rcsparse_textrevision()
538  *
539  * Called upon reaching a new REVISION entry in the delta text section.
540  * pdp->rp_delta will be set to REVISION's delta (created in delta section)
541  * for further parsing.
542  *
543  * REVISION
544  *
545  * Returns 0 on success or 1 on failure.
546  */
547 static int
548 rcsparse_textrevision(RCSFILE *rfp, struct rcs_pdata *pdp)
549 {
550 	struct rcs_delta *rdp;
551 
552 	TAILQ_FOREACH(rdp, &rfp->rf_delta, rd_list) {
553 		if (rcsnum_cmp(rdp->rd_num, pdp->rp_value.rev, 0) == 0)
554 			break;
555 	}
556 	if (rdp == NULL) {
557 		rcsparse_warnx(rfp, "delta for revision \"%s\" not found",
558 		    pdp->rp_buf);
559 		rcsnum_free(pdp->rp_value.rev);
560 		return (1);
561 	}
562 	pdp->rp_delta = rdp;
563 
564 	rcsnum_free(pdp->rp_value.rev);
565 	return (0);
566 }
567 
568 /*
569  * rcsparse_log()
570  *
571  * Parses the specified log of current deltatext pdp->rp_delta.
572  *
573  * log @[...]@
574  *
575  * Returns 0 on success or 1 on failure.
576  */
577 static int
578 rcsparse_log(RCSFILE *rfp, struct rcs_pdata *pdp)
579 {
580 	if (rcsparse_token(rfp, RCS_TYPE_STRING) != RCS_TYPE_STRING)
581 		return (1);
582 
583 	pdp->rp_delta->rd_log = pdp->rp_value.str;
584 
585 	return (0);
586 }
587 
588 /*
589  * rcsparse_text()
590  *
591  * Parses the specified text of current deltatext pdp->rp_delta.
592  *
593  * text @[...]@
594  *
595  * Returns 0 on success or 1 on failure.
596  */
597 static int
598 rcsparse_text(RCSFILE *rfp, struct rcs_pdata *pdp)
599 {
600 	if (rcsparse_token(rfp, RCS_TYPE_STRING) != RCS_TYPE_STRING)
601 		return (1);
602 
603 	pdp->rp_delta->rd_tlen = pdp->rp_tlen - 1;
604 	if (pdp->rp_delta->rd_tlen == 0) {
605 		pdp->rp_delta->rd_text = xstrdup("");
606 	} else {
607 		pdp->rp_delta->rd_text = xmalloc(pdp->rp_delta->rd_tlen);
608 		memcpy(pdp->rp_delta->rd_text, pdp->rp_buf,
609 		    pdp->rp_delta->rd_tlen);
610 	}
611 	xfree(pdp->rp_value.str);
612 
613 	return (0);
614 }
615 
616 /*
617  * rcsparse_head()
618  *
619  * Parses the head revision of RCS file <rfp>.
620  *
621  * head [REVISION];
622  *
623  * Returns 0 on success or 1 on failure.
624  */
625 static int
626 rcsparse_head(RCSFILE *rfp, struct rcs_pdata *pdp)
627 {
628 	int type;
629 
630 	type = rcsparse_token(rfp, RCS_TYPE_REVISION|RCS_TOK_SCOLON);
631 	if (type == RCS_TYPE_REVISION) {
632 		rfp->rf_head = pdp->rp_value.rev;
633 		type = rcsparse_token(rfp, RCS_TOK_SCOLON);
634 	}
635 
636 	return (type != RCS_TOK_SCOLON);
637 }
638 
639 /*
640  * rcsparse_branch()
641  *
642  * Parses the default branch of RCS file <rfp>. The branch keyword is
643  * optional and can be omitted.
644  *
645  * [branch BRANCH;]
646  *
647  * Returns 0 on success or 1 on failure.
648  */
649 static int
650 rcsparse_branch(RCSFILE *rfp, struct rcs_pdata *pdp)
651 {
652 	int type;
653 
654 	type = rcsparse_token(rfp, RCS_TYPE_BRANCH|RCS_TOK_SCOLON);
655 	if (type == RCS_TYPE_BRANCH) {
656 		rfp->rf_branch = pdp->rp_value.rev;
657 		type = rcsparse_token(rfp, RCS_TOK_SCOLON);
658 	}
659 
660 	return (type != RCS_TOK_SCOLON);
661 }
662 
663 /*
664  * rcsparse_access()
665  *
666  * Parses the access list of RCS file <rfp>.
667  *
668  * access [LOGIN ...];
669  *
670  * Returns 0 on success or 1 on failure.
671  */
672 static int
673 rcsparse_access(RCSFILE *rfp, struct rcs_pdata *pdp)
674 {
675 	struct rcs_access *ap;
676 	int type;
677 
678 	while ((type = rcsparse_token(rfp, RCS_TOK_SCOLON|RCS_TYPE_LOGIN))
679 	    == RCS_TYPE_LOGIN) {
680 		ap = xmalloc(sizeof(*ap));
681 		ap->ra_name = pdp->rp_value.str;
682 		TAILQ_INSERT_TAIL(&(rfp->rf_access), ap, ra_list);
683 	}
684 
685 	return (type != RCS_TOK_SCOLON);
686 }
687 
688 /*
689  * rcsparse_symbols()
690  *
691  * Parses the symbol list of RCS file <rfp>.
692  *
693  * symbols [SYMBOL:REVISION ...];
694  *
695  * Returns 0 on success or 1 on failure.
696  */
697 static int
698 rcsparse_symbols(RCSFILE *rfp, struct rcs_pdata *pdp)
699 {
700 	struct rcs_sym *symp;
701 	char *name;
702 	int type;
703 
704 	while ((type = rcsparse_token(rfp, RCS_TOK_SCOLON|RCS_TYPE_SYMBOL)) ==
705 	    RCS_TYPE_SYMBOL) {
706 		name = pdp->rp_value.str;
707 		if (rcsparse_token(rfp, RCS_TOK_COLON) != RCS_TOK_COLON ||
708 		    rcsparse_token(rfp, RCS_TYPE_NUMBER) != RCS_TYPE_NUMBER) {
709 			xfree(name);
710 			return (1);
711 		}
712 		symp = xmalloc(sizeof(*symp));
713 		symp->rs_name = name;
714 		symp->rs_num = pdp->rp_value.rev;
715 		TAILQ_INSERT_TAIL(&(rfp->rf_symbols), symp, rs_list);
716 	}
717 
718 	return (type != RCS_TOK_SCOLON);
719 }
720 
721 /*
722  * rcsparse_locks()
723  *
724  * Parses the lock list of RCS file <rfp>.
725  *
726  * locks [SYMBOL:REVISION ...];
727  *
728  * Returns 0 on success or 1 on failure.
729  */
730 static int
731 rcsparse_locks(RCSFILE *rfp, struct rcs_pdata *pdp)
732 {
733 	struct rcs_lock *lkp;
734 	char *name;
735 	int type;
736 
737 	while ((type = rcsparse_token(rfp, RCS_TOK_SCOLON|RCS_TYPE_LOGIN)) ==
738 	    RCS_TYPE_LOGIN) {
739 		name = pdp->rp_value.str;
740 		if (rcsparse_token(rfp, RCS_TOK_COLON) != RCS_TOK_COLON ||
741 		    rcsparse_token(rfp, RCS_TYPE_REVISION) !=
742 		    RCS_TYPE_REVISION) {
743 			xfree(name);
744 			return (1);
745 		}
746 		lkp = xmalloc(sizeof(*lkp));
747 		lkp->rl_name = name;
748 		lkp->rl_num = pdp->rp_value.rev;
749 		TAILQ_INSERT_TAIL(&(rfp->rf_locks), lkp, rl_list);
750 	}
751 
752 	return (type != RCS_TOK_SCOLON);
753 }
754 
755 /*
756  * rcsparse_locks()
757  *
758  * Parses the strict keyword of RCS file <rfp>. The strict keyword is
759  * optional and can be omitted.
760  *
761  * [strict;]
762  *
763  * Returns 0 on success or 1 on failure.
764  */
765 static int
766 rcsparse_strict(RCSFILE *rfp, struct rcs_pdata *pdp)
767 {
768 	rfp->rf_flags |= RCS_SLOCK;
769 
770 	return (rcsparse_token(rfp, RCS_TOK_SCOLON) != RCS_TOK_SCOLON);
771 }
772 
773 /*
774  * rcsparse_comment()
775  *
776  * Parses the comment of RCS file <rfp>.  The comment keyword is optional
777  * and can be omitted.
778  *
779  * [comment [@[...]@];]
780  *
781  * Returns 0 on success or 1 on failure.
782  */
783 static int
784 rcsparse_comment(RCSFILE *rfp, struct rcs_pdata *pdp)
785 {
786 	int type;
787 
788 	type = rcsparse_token(rfp, RCS_TYPE_STRING|RCS_TOK_SCOLON);
789 	if (type == RCS_TYPE_STRING) {
790 		rfp->rf_comment = pdp->rp_value.str;
791 		type = rcsparse_token(rfp, RCS_TOK_SCOLON);
792 	}
793 
794 	return (type != RCS_TOK_SCOLON);
795 }
796 
797 /*
798  * rcsparse_expand()
799  *
800  * Parses expand of RCS file <rfp>.  The expand keyword is optional and
801  * can be omitted.
802  *
803  * [expand [@[...]@];]
804  *
805  * Returns 0 on success or 1 on failure.
806  */
807 static int
808 rcsparse_expand(RCSFILE *rfp, struct rcs_pdata *pdp)
809 {
810 	int type;
811 
812 	type = rcsparse_token(rfp, RCS_TYPE_STRING|RCS_TOK_SCOLON);
813 	if (type == RCS_TYPE_STRING) {
814 		rfp->rf_expand = pdp->rp_value.str;
815 		type = rcsparse_token(rfp, RCS_TOK_SCOLON);
816 	}
817 
818 	return (type != RCS_TOK_SCOLON);
819 }
820 
821 #define RBUF_PUTC(ch) \
822 do { \
823 	if (bp == pdp->rp_bufend - 1) { \
824 		len = bp - pdp->rp_buf; \
825 		rcsparse_growbuf(rfp); \
826 		bp = pdp->rp_buf + len; \
827 	} \
828 	*(bp++) = (ch); \
829 	pdp->rp_tlen++; \
830 } while (0);
831 
832 static int
833 rcsparse_string(RCSFILE *rfp, int allowed)
834 {
835 	struct rcs_pdata *pdp;
836 	int c;
837 	size_t len;
838 	char *bp;
839 
840 	pdp = (struct rcs_pdata *)rfp->rf_pdata;
841 
842 	bp = pdp->rp_buf;
843 	pdp->rp_tlen = 0;
844 	*bp = '\0';
845 
846 	for (;;) {
847 		c = getc(rfp->rf_file);
848 		if (c == '@') {
849 			c = getc(rfp->rf_file);
850 			if (c == EOF) {
851 				return (EOF);
852 			} else if (c != '@') {
853 				ungetc(c, rfp->rf_file);
854 				break;
855 			}
856 		}
857 
858 		if (c == EOF) {
859 			return (EOF);
860 		} else if (c == '\n')
861 			pdp->rp_lineno++;
862 
863 		RBUF_PUTC(c);
864 	}
865 
866 	bp = pdp->rp_buf + pdp->rp_tlen;
867 	RBUF_PUTC('\0');
868 
869 	if (!(allowed & RCS_TYPE_STRING)) {
870 		rcsparse_warnx(rfp, "unexpected RCS string");
871 		return (0);
872 	}
873 
874 	pdp->rp_value.str = xstrdup(pdp->rp_buf);
875 
876 	return (RCS_TYPE_STRING);
877 }
878 
879 static int
880 rcsparse_token(RCSFILE *rfp, int allowed)
881 {
882 	const struct rcs_keyword *p;
883 	struct rcs_pdata *pdp;
884 	int c, pre, ret, type;
885 	char *bp;
886 	size_t len;
887 	RCSNUM *datenum;
888 
889 	pdp = (struct rcs_pdata *)rfp->rf_pdata;
890 
891 	if (pdp->rp_token != -1) {
892 		/* no need to check for allowed here */
893 		type = pdp->rp_token;
894 		pdp->rp_token = -1;
895 		return (type);
896 	}
897 
898 	/* skip whitespaces */
899 	c = EOF;
900 	do {
901 		pre = c;
902 		c = getc(rfp->rf_file);
903 		if (c == EOF) {
904 			if (ferror(rfp->rf_file)) {
905 				rcsparse_warnx(rfp, "error during parsing");
906 				return (0);
907 			}
908 			if (pre != '\n')
909 				rcsparse_warnx(rfp,
910 				    "no newline at end of file");
911 			return (EOF);
912 		} else if (c == '\n')
913 			pdp->rp_lineno++;
914 	} while (isspace(c));
915 
916 	pdp->rp_msglineno = pdp->rp_lineno;
917 	type = 0;
918 	switch (c) {
919 	case '@':
920 		ret = rcsparse_string(rfp, allowed);
921 		if (ret == EOF && ferror(rfp->rf_file)) {
922 			rcsparse_warnx(rfp, "error during parsing");
923 			return (0);
924 		}
925 		return (ret);
926 		/* NOTREACHED */
927 	case ':':
928 		type = RCS_TOK_COLON;
929 		if (type & allowed)
930 			return (type);
931 		rcsparse_warnx(rfp, "unexpected token \"%c\"", c);
932 		return (0);
933 		/* NOTREACHED */
934 	case ';':
935 		type = RCS_TOK_SCOLON;
936 		if (type & allowed)
937 			return (type);
938 		rcsparse_warnx(rfp, "unexpected token \"%c\"", c);
939 		return (0);
940 		/* NOTREACHED */
941 	case ',':
942 		type = RCS_TOK_COMMA;
943 		if (type & allowed)
944 			return (type);
945 		rcsparse_warnx(rfp, "unexpected token \"%c\"", c);
946 		return (0);
947 		/* NOTREACHED */
948 	default:
949 		if (!isgraph(c)) {
950 			rcsparse_warnx(rfp, "unexpected character 0x%.2X", c);
951 			return (0);
952 		}
953 		break;
954 	}
955 	allowed &= ~(RCS_TOK_COLON|RCS_TOK_SCOLON|RCS_TOK_COMMA);
956 
957 	bp = pdp->rp_buf;
958 	pdp->rp_tlen = 0;
959 	*bp = '\0';
960 
961 	for (;;) {
962 		if (c == EOF) {
963 			if (ferror(rfp->rf_file))
964 				rcsparse_warnx(rfp, "error during parsing");
965 			else
966 				rcsparse_warnx(rfp, "unexpected end of file");
967 			return (0);
968 		} else if (c == '\n')
969 			pdp->rp_lineno++;
970 
971 		RBUF_PUTC(c);
972 
973 		c = getc(rfp->rf_file);
974 
975 		if (isspace(c)) {
976 			if (c == '\n')
977 				pdp->rp_lineno++;
978 			RBUF_PUTC('\0');
979 			break;
980 		} else if (c == ';' || c == ':' || c == ',') {
981 			ungetc(c, rfp->rf_file);
982 			RBUF_PUTC('\0');
983 			break;
984 		} else if (!isgraph(c)) {
985 			rcsparse_warnx(rfp, "unexpected character 0x%.2X", c);
986 			return (0);
987 		}
988 	}
989 
990 	switch (allowed) {
991 	case RCS_TYPE_COMMITID:
992 		/* XXX validate commitid */
993 		break;
994 	case RCS_TYPE_LOGIN:
995 		if (!valid_login(pdp->rp_buf)) {
996 			rcsparse_warnx(rfp, "invalid login \"%s\"",
997 			    pdp->rp_buf);
998 			return (0);
999 		}
1000 		pdp->rp_value.str = xstrdup(pdp->rp_buf);
1001 		break;
1002 	case RCS_TYPE_SYMBOL:
1003 		if (!rcs_sym_check(pdp->rp_buf)) {
1004 			rcsparse_warnx(rfp, "invalid symbol \"%s\"",
1005 			    pdp->rp_buf);
1006 			return (0);
1007 		}
1008 		pdp->rp_value.str = xstrdup(pdp->rp_buf);
1009 		break;
1010 		/* FALLTHROUGH */
1011 	case RCS_TYPE_STATE:
1012 		if (rcs_state_check(pdp->rp_buf)) {
1013 			rcsparse_warnx(rfp, "invalid state \"%s\"",
1014 			    pdp->rp_buf);
1015 			return (0);
1016 		}
1017 		pdp->rp_value.str = xstrdup(pdp->rp_buf);
1018 		break;
1019 	case RCS_TYPE_DATE:
1020 		if ((datenum = rcsnum_parse(pdp->rp_buf)) == NULL) {
1021 			rcsparse_warnx(rfp, "invalid date \"%s\"", pdp->rp_buf);
1022 			return (0);
1023 		}
1024 		if (datenum->rn_len != 6) {
1025 			rcsnum_free(datenum);
1026 			rcsparse_warnx(rfp, "invalid date \"%s\"", pdp->rp_buf);
1027 			return (0);
1028 		}
1029 		pdp->rp_value.date.tm_year = datenum->rn_id[0];
1030 		if (pdp->rp_value.date.tm_year >= 1900)
1031 			pdp->rp_value.date.tm_year -= 1900;
1032 		pdp->rp_value.date.tm_mon = datenum->rn_id[1] - 1;
1033 		pdp->rp_value.date.tm_mday = datenum->rn_id[2];
1034 		pdp->rp_value.date.tm_hour = datenum->rn_id[3];
1035 		pdp->rp_value.date.tm_min = datenum->rn_id[4];
1036 		pdp->rp_value.date.tm_sec = datenum->rn_id[5];
1037 		rcsnum_free(datenum);
1038 		break;
1039 	case RCS_TYPE_NUMBER:
1040 		pdp->rp_value.rev = rcsnum_parse(pdp->rp_buf);
1041 		if (pdp->rp_value.rev == NULL) {
1042 			rcsparse_warnx(rfp, "invalid number \"%s\"",
1043 			    pdp->rp_buf);
1044 			return (0);
1045 		}
1046 		break;
1047 	case RCS_TYPE_BRANCH:
1048 		pdp->rp_value.rev = rcsnum_parse(pdp->rp_buf);
1049 		if (pdp->rp_value.rev == NULL) {
1050 			rcsparse_warnx(rfp, "invalid branch \"%s\"",
1051 			    pdp->rp_buf);
1052 			return (0);
1053 		}
1054 		if (!RCSNUM_ISBRANCH(pdp->rp_value.rev)) {
1055 			rcsnum_free(pdp->rp_value.rev);
1056 			rcsparse_warnx(rfp, "expected branch, got \"%s\"",
1057 			    pdp->rp_buf);
1058 			return (0);
1059 		}
1060 		break;
1061 	case RCS_TYPE_KEYWORD:
1062 		if (islower(*pdp->rp_buf)) {
1063 			p = bsearch(pdp->rp_buf, keywords,
1064 			    sizeof(keywords) / sizeof(keywords[0]),
1065 			    sizeof(keywords[0]), kw_cmp);
1066 			if (p != NULL)
1067 				return (p->k_val);
1068 		}
1069 		allowed = RCS_TYPE_REVISION;
1070 		/* FALLTHROUGH */
1071 	case RCS_TYPE_REVISION:
1072 		pdp->rp_value.rev = rcsnum_parse(pdp->rp_buf);
1073 		if (pdp->rp_value.rev != NULL) {
1074 			if (RCSNUM_ISBRANCH(pdp->rp_value.rev)) {
1075 				rcsnum_free(pdp->rp_value.rev);
1076 				rcsparse_warnx(rfp,
1077 				    "expected revision, got \"%s\"",
1078 				    pdp->rp_buf);
1079 				return (0);
1080 			}
1081 			break;
1082 		}
1083 		/* FALLTHROUGH */
1084 	default:
1085 		RBUF_PUTC('\0');
1086 		rcsparse_warnx(rfp, "unexpected token \"%s\"", pdp->rp_buf);
1087 		return (0);
1088 		/* NOTREACHED */
1089 	}
1090 
1091 	return (allowed);
1092 }
1093 
1094 static int
1095 rcsparse(RCSFILE *rfp, struct rcs_section *sec)
1096 {
1097 	struct rcs_pdata *pdp;
1098 	int i, token;
1099 
1100 	pdp = (struct rcs_pdata *)rfp->rf_pdata;
1101 	i = 0;
1102 
1103 	token = 0;
1104 	for (i = 0; sec[i].token != 0; i++) {
1105 		token = rcsparse_token(rfp, RCS_TYPE_KEYWORD);
1106 		if (token == 0)
1107 			return (1);
1108 
1109 		while (token != sec[i].token) {
1110 			if (sec[i].parse == NULL)
1111 				goto end;
1112 			if (sec[i].opt) {
1113 				i++;
1114 				continue;
1115 			}
1116 			if (token == EOF || (!(rfp->rf_flags & PARSED_DELTAS) &&
1117 			    token == RCS_TOK_DESC))
1118 				goto end;
1119 			rcsparse_warnx(rfp, "unexpected token \"%s\"",
1120 			    pdp->rp_buf);
1121 			return (1);
1122 		}
1123 
1124 		if (sec[i].parse(rfp, pdp))
1125 			return (1);
1126 	}
1127 end:
1128 	if (token == RCS_TYPE_REVISION)
1129 		pdp->rp_token = token;
1130 	else if (token == RCS_TOK_DESC)
1131 		pdp->rp_token = RCS_TOK_DESC;
1132 	else if (token == EOF)
1133 		rfp->rf_flags |= RCS_PARSED;
1134 
1135 	return (0);
1136 }
1137 
1138 static int
1139 rcsparse_deltatext(RCSFILE *rfp)
1140 {
1141 	int ret;
1142 
1143 	if (rfp->rf_flags & PARSED_DELTATEXTS)
1144 		return (0);
1145 
1146 	if (!(rfp->rf_flags & PARSED_DESC))
1147 		if ((ret = rcsparse_desc(rfp)))
1148 			return (ret);
1149 
1150 	if (rcsparse(rfp, sec_deltatext))
1151 		return (-1);
1152 
1153 	if (rfp->rf_flags & RCS_PARSED)
1154 		rfp->rf_flags |= PARSED_DELTATEXTS;
1155 
1156 	return (1);
1157 }
1158 
1159 static int
1160 rcsparse_delta(RCSFILE *rfp)
1161 {
1162 	struct rcs_pdata *pdp;
1163 
1164 	if (rfp->rf_flags & PARSED_DELTAS)
1165 		return (0);
1166 
1167 	pdp = (struct rcs_pdata *)rfp->rf_pdata;
1168 	if (pdp->rp_token == RCS_TOK_DESC) {
1169 		rfp->rf_flags |= PARSED_DELTAS;
1170 		return (0);
1171 	}
1172 
1173 	if (rcsparse(rfp, sec_delta))
1174 		return (-1);
1175 
1176 	if (pdp->rp_delta != NULL) {
1177 		TAILQ_INSERT_TAIL(&rfp->rf_delta, pdp->rp_delta, rd_list);
1178 		pdp->rp_delta = NULL;
1179 		rfp->rf_ndelta++;
1180 		return (1);
1181 	}
1182 
1183 	return (0);
1184 }
1185 
1186 /*
1187  * rcsparse_growbuf()
1188  *
1189  * Attempt to grow the internal parse buffer for the RCS file <rf> by
1190  * RCS_BUFEXTSIZE.
1191  * In case of failure, the original buffer is left unmodified.
1192  */
1193 static void
1194 rcsparse_growbuf(RCSFILE *rfp)
1195 {
1196 	struct rcs_pdata *pdp = (struct rcs_pdata *)rfp->rf_pdata;
1197 
1198 	pdp->rp_buf = xrealloc(pdp->rp_buf, 1,
1199 		pdp->rp_blen + RCS_BUFEXTSIZE);
1200 	pdp->rp_blen += RCS_BUFEXTSIZE;
1201 	pdp->rp_bufend = pdp->rp_buf + pdp->rp_blen - 1;
1202 }
1203 
1204 /*
1205  * Borrowed from src/usr.sbin/user/user.c:
1206  * return 1 if `login' is a valid login name
1207  */
1208 static int
1209 valid_login(char *login_name)
1210 {
1211 	unsigned char *cp;
1212 
1213 	/* The first character cannot be a hyphen */
1214 	if (*login_name == '-')
1215 		return 0;
1216 
1217 	for (cp = login_name ; *cp ; cp++) {
1218 		/* We allow '$' as the last character for samba */
1219 		if (!isalnum(*cp) && *cp != '.' && *cp != '_' && *cp != '-' &&
1220 		    !(*cp == '$' && *(cp + 1) == '\0')) {
1221 			return 0;
1222 		}
1223 	}
1224 	if ((char *)cp - login_name > _PW_NAME_LEN)
1225 		return 0;
1226 	return 1;
1227 }
1228 
1229 static int
1230 kw_cmp(const void *k, const void *e)
1231 {
1232 	return (strcmp(k, ((const struct rcs_keyword *)e)->k_name));
1233 }
1234 
1235 static void
1236 rcsparse_warnx(RCSFILE *rfp, char *fmt, ...)
1237 {
1238 	struct rcs_pdata *pdp;
1239 	va_list ap;
1240 	char *nfmt;
1241 
1242 	pdp = (struct rcs_pdata *)rfp->rf_pdata;
1243 	va_start(ap, fmt);
1244 	if (asprintf(&nfmt, "%s:%d: %s", rfp->rf_path, pdp->rp_msglineno, fmt)
1245 	    == -1)
1246 		nfmt = fmt;
1247 	cvs_vlog(LP_ERR, nfmt, ap);
1248 	va_end(ap);
1249 	if (nfmt != fmt)
1250 		free(nfmt);
1251 }
1252