Gotweb

Blob

Date:: Mon Jun 27 18:47:02 2022 UTC
Message:: got patch: handle git-style diffs for the 3-way merge too tweak and ok stsp@
Actions:: History | Blame | Raw File
1 /*
2  * Copyright 1986, Larry Wall
3  *
4  * Redistribution and use in source and binary forms, with or without
5  * modification, are permitted provided that the following condition is met:
6  * 1. Redistributions of source code must retain the above copyright notice,
7  * this condition and the following disclaimer.
8  *
9  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND ANY
10  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
11  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
12  * DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
13  * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
14  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
15  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
16  * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
17  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
18  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
19  * SUCH DAMAGE.
20  */
21 
22 /*
23  * Copyright (c) 2022 Omar Polo <op@openbsd.org>
24  *
25  * Permission to use, copy, modify, and distribute this software for any
26  * purpose with or without fee is hereby granted, provided that the above
27  * copyright notice and this permission notice appear in all copies.
28  *
29  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
30  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
31  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
32  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
33  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
34  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
35  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
36  */
37 
38 #include <sys/types.h>
39 #include <sys/queue.h>
40 #include <sys/uio.h>
41 
42 #include <ctype.h>
43 #include <limits.h>
44 #include <paths.h>
45 #include <sha1.h>
46 #include <stdint.h>
47 #include <stdio.h>
48 #include <stdlib.h>
49 #include <string.h>
50 #include <unistd.h>
51 #include <imsg.h>
52 
53 #include "got_error.h"
54 #include "got_object.h"
55 
56 #include "got_lib_delta.h"
57 #include "got_lib_object.h"
58 #include "got_lib_privsep.h"
59 #include "got_lib_sha1.h"
60 
61 struct imsgbuf ibuf;
62 
63 static const struct got_error *
64 send_patch(const char *oldname, const char *newname, const char *commitid,
65     const char *blob, int git)
66 {
67 	struct got_imsg_patch p;
68 
69 	memset(&p, 0, sizeof(p));
70 
71 	if (oldname != NULL)
72 		strlcpy(p.old, oldname, sizeof(p.old));
73 
74 	if (newname != NULL)
75 		strlcpy(p.new, newname, sizeof(p.new));
76 
77 	if (commitid != NULL)
78 		strlcpy(p.cid, commitid, sizeof(p.cid));
79 
80 	if (blob != NULL)
81 		strlcpy(p.blob, blob, sizeof(p.blob));
82 
83 	p.git = git;
84 	if (imsg_compose(&ibuf, GOT_IMSG_PATCH, 0, 0, -1, &p, sizeof(p)) == -1)
85 		return got_error_from_errno("imsg_compose GOT_IMSG_PATCH");
86 	return NULL;
87 }
88 
89 static const struct got_error *
90 send_patch_done(void)
91 {
92 	if (imsg_compose(&ibuf, GOT_IMSG_PATCH_DONE, 0, 0, -1,
93 	    NULL, 0) == -1)
94 		return got_error_from_errno("imsg_compose GOT_IMSG_PATCH_EOF");
95 	if (imsg_flush(&ibuf) == -1)
96 		return got_error_from_errno("imsg_flush");
97 	return NULL;
98 }
99 
100 /* based on fetchname from usr.bin/patch/util.c */
101 static const struct got_error *
102 filename(const char *at, char **name)
103 {
104 	char	*tmp, *t;
105 
106 	*name = NULL;
107 	if (*at == '\0')
108 		return NULL;
109 
110 	while (isspace((unsigned char)*at))
111 		at++;
112 
113 	/* files can be created or removed by diffing against /dev/null */
114 	if (!strncmp(at, _PATH_DEVNULL, sizeof(_PATH_DEVNULL) - 1))
115 		return NULL;
116 
117 	tmp = strdup(at);
118 	if (tmp == NULL)
119 		return got_error_from_errno("strdup");
120 	if ((t = strchr(tmp, '\t')) != NULL)
121 		*t = '\0';
122 	if ((t = strchr(tmp, '\n')) != NULL)
123 		*t = '\0';
124 
125 	*name = strdup(tmp);
126 	free(tmp);
127 	if (*name == NULL)
128 		return got_error_from_errno("strdup");
129 	return NULL;
130 }
131 
132 static const struct got_error *
133 blobid(const char *line, char **blob, int git)
134 {
135 	uint8_t digest[SHA1_DIGEST_LENGTH];
136 	size_t len;
137 
138 	*blob = NULL;
139 
140 	len = strspn(line, "0123456789abcdefABCDEF");
141 	if ((*blob = strndup(line, len)) == NULL)
142 		return got_error_from_errno("strndup");
143 
144 	if (!git && !got_parse_sha1_digest(digest, *blob)) {
145 		/* silently ignore invalid blob ids */
146 		free(*blob);
147 		*blob = NULL;
148 	}
149 	return NULL;
150 }
151 
152 static const struct got_error *
153 find_patch(int *done, FILE *fp)
154 {
155 	const struct got_error *err = NULL;
156 	char	*old = NULL, *new = NULL;
157 	char	*commitid = NULL, *blob = NULL;
158 	char	*line = NULL;
159 	size_t	 linesize = 0;
160 	ssize_t	 linelen;
161 	int	 create, rename = 0, git = 0;
162 
163 	while ((linelen = getline(&line, &linesize, fp)) != -1) {
164 		/*
165 		 * Ignore the Index name like GNU and larry' patch,
166 		 * we don't have to follow POSIX.
167 		 */
168 
169 		if (!strncmp(line, "--- ", 4)) {
170 			free(old);
171 			err = filename(line+4, &old);
172 		} else if (rename && !strncmp(line, "rename from ", 12)) {
173 			free(old);
174 			err = filename(line+12, &old);
175 		} else if (!strncmp(line, "+++ ", 4)) {
176 			free(new);
177 			err = filename(line+4, &new);
178 		} else if (!git && !strncmp(line, "blob - ", 7)) {
179 			free(blob);
180 			err = blobid(line + 7, &blob, git);
181 		} else if (rename && !strncmp(line, "rename to ", 10)) {
182 			free(new);
183 			err = filename(line + 10, &new);
184 		} else if (git && !strncmp(line, "similarity index 100%", 21))
185 			rename = 1;
186 		else if (git && !strncmp(line, "index ", 6)) {
187 			free(blob);
188 			err = blobid(line + 6, &blob, git);
189 		} else if (!strncmp(line, "diff --git a/", 13)) {
190 			git = 1;
191 			free(commitid);
192 			commitid = NULL;
193 			free(blob);
194 			blob = NULL;
195 		} else if (!git && !strncmp(line, "diff ", 5)) {
196 			free(commitid);
197 			err = blobid(line + 5, &commitid, git);
198 		} else if (!git && !strncmp(line, "commit - ", 9)) {
199 			free(commitid);
200 			err = blobid(line + 9, &commitid, git);
201 		}
202 
203 		if (err)
204 			break;
205 
206 		/*
207 		 * Git-style diffs with "similarity index 100%" don't
208 		 * have any hunks and ends with the "rename to foobar"
209 		 * line.
210 		 */
211 		if (rename && old != NULL && new != NULL) {
212 			*done = 1;
213 			err = send_patch(old, new, commitid,
214 			    blob, git);
215 			break;
216 		}
217 
218 		if (!strncmp(line, "@@ -", 4)) {
219 			create = !strncmp(line+4, "0,0", 3);
220 			if ((old == NULL && new == NULL) ||
221 			    (!create && old == NULL))
222 				err = got_error(GOT_ERR_PATCH_MALFORMED);
223 			else
224 				err = send_patch(old, new, commitid,
225 				    blob, git);
226 
227 			if (err)
228 				break;
229 
230 			/* rewind to previous line */
231 			if (fseeko(fp, -linelen, SEEK_CUR) == -1)
232 				err = got_error_from_errno("fseeko");
233 			break;
234 		}
235 	}
236 
237 	free(old);
238 	free(new);
239 	free(commitid);
240 	free(blob);
241 	free(line);
242 	if (ferror(fp) && err == NULL)
243 		err = got_error_from_errno("getline");
244 	if (feof(fp) && err == NULL)
245 		err = got_error(GOT_ERR_NO_PATCH);
246 	return err;
247 }
248 
249 static const struct got_error *
250 strtolnum(char **str, int *n)
251 {
252 	char		*p, c;
253 	const char	*errstr;
254 
255 	for (p = *str; isdigit((unsigned char)*p); ++p)
256 		/* nop */;
257 
258 	c = *p;
259 	*p = '\0';
260 
261 	*n = strtonum(*str, 0, INT_MAX, &errstr);
262 	if (errstr != NULL)
263 		return got_error(GOT_ERR_PATCH_MALFORMED);
264 
265 	*p = c;
266 	*str = p;
267 	return NULL;
268 }
269 
270 static const struct got_error *
271 parse_hdr(char *s, int *done, struct got_imsg_patch_hunk *hdr)
272 {
273 	static const struct got_error *err = NULL;
274 
275 	if (strncmp(s, "@@ -", 4)) {
276 		*done = 1;
277 		return NULL;
278 	}
279 
280 	s += 4;
281 	if (!*s)
282 		return NULL;
283 	err = strtolnum(&s, &hdr->oldfrom);
284 	if (err)
285 		return err;
286 	if (*s == ',') {
287 		s++;
288 		err = strtolnum(&s, &hdr->oldlines);
289 		if (err)
290 			return err;
291 	} else
292 		hdr->oldlines = 1;
293 
294 	if (*s == ' ')
295 		s++;
296 
297 	if (*s != '+' || !*++s)
298 		return got_error(GOT_ERR_PATCH_MALFORMED);
299 	err = strtolnum(&s, &hdr->newfrom);
300 	if (err)
301 		return err;
302 	if (*s == ',') {
303 		s++;
304 		err = strtolnum(&s, &hdr->newlines);
305 		if (err)
306 			return err;
307 	} else
308 		hdr->newlines = 1;
309 
310 	if (*s == ' ')
311 		s++;
312 
313 	if (*s != '@')
314 		return got_error(GOT_ERR_PATCH_MALFORMED);
315 
316 	if (hdr->oldfrom >= INT_MAX - hdr->oldlines ||
317 	    hdr->newfrom >= INT_MAX - hdr->newlines ||
318 	    /* not so sure about this one */
319 	    hdr->oldlines >= INT_MAX - hdr->newlines - 1 ||
320 	    (hdr->oldlines == 0 && hdr->newlines == 0))
321 		return got_error(GOT_ERR_PATCH_MALFORMED);
322 
323 	if (hdr->oldlines == 0) {
324 		/* larry says to "do append rather than insert"; I don't
325 		 * quite get it, but i trust him.
326 		 */
327 		hdr->oldfrom++;
328 	}
329 
330 	if (imsg_compose(&ibuf, GOT_IMSG_PATCH_HUNK, 0, 0, -1,
331 	    hdr, sizeof(*hdr)) == -1)
332 		return got_error_from_errno(
333 		    "imsg_compose GOT_IMSG_PATCH_HUNK");
334 	return NULL;
335 }
336 
337 static const struct got_error *
338 send_line(const char *line)
339 {
340 	static const struct got_error *err = NULL;
341 	char *p = NULL;
342 
343 	if (*line != '+' && *line != '-' && *line != ' ' && *line != '\\') {
344 		if (asprintf(&p, " %s", line) == -1)
345 			return got_error_from_errno("asprintf");
346 		line = p;
347 	}
348 
349 	if (imsg_compose(&ibuf, GOT_IMSG_PATCH_LINE, 0, 0, -1,
350 	    line, strlen(line) + 1) == -1)
351 		err = got_error_from_errno(
352 		    "imsg_compose GOT_IMSG_PATCH_LINE");
353 
354 	free(p);
355 	return err;
356 }
357 
358 static const struct got_error *
359 peek_special_line(FILE *fp)
360 {
361 	const struct got_error *err;
362 	int ch;
363 
364 	ch = fgetc(fp);
365 	if (ch != EOF && ch != '\\') {
366 		ungetc(ch, fp);
367 		return NULL;
368 	}
369 
370 	if (ch == '\\') {
371 		err = send_line("\\");
372 		if (err)
373 			return err;
374 	}
375 
376 	while (ch != EOF && ch != '\n')
377 		ch = fgetc(fp);
378 
379 	if (ch != EOF || feof(fp))
380 		return NULL;
381 	return got_error(GOT_ERR_IO);
382 }
383 
384 static const struct got_error *
385 parse_hunk(FILE *fp, int *done)
386 {
387 	static const struct got_error *err = NULL;
388 	struct got_imsg_patch_hunk hdr;
389 	char	*line = NULL, ch;
390 	size_t	 linesize = 0;
391 	ssize_t	 linelen;
392 	int	 leftold, leftnew;
393 
394 	linelen = getline(&line, &linesize, fp);
395 	if (linelen == -1) {
396 		*done = 1;
397 		goto done;
398 	}
399 
400 	err = parse_hdr(line, done, &hdr);
401 	if (err)
402 		goto done;
403 	if (*done) {
404 		if (fseeko(fp, -linelen, SEEK_CUR) == -1)
405 			err = got_error_from_errno("fseeko");
406 		goto done;
407 	}
408 
409 	leftold = hdr.oldlines;
410 	leftnew = hdr.newlines;
411 
412 	while (leftold > 0 || leftnew > 0) {
413 		linelen = getline(&line, &linesize, fp);
414 		if (linelen == -1) {
415 			if (ferror(fp)) {
416 				err = got_error_from_errno("getline");
417 				goto done;
418 			}
419 
420 			/* trailing newlines may be chopped */
421 			if (leftold < 3 && leftnew < 3) {
422 				*done = 1;
423 				break;
424 			}
425 
426 			err = got_error(GOT_ERR_PATCH_TRUNCATED);
427 			goto done;
428 		}
429 		if (line[linelen - 1] == '\n')
430 			line[linelen - 1] = '\0';
431 
432 		/* usr.bin/patch allows '=' as context char */
433 		if (*line == '=')
434 			*line = ' ';
435 
436 		ch = *line;
437 		if (ch == '\t' || ch == '\0')
438 			ch = ' ';	/* the space got eaten */
439 
440 		switch (ch) {
441 		case '-':
442 			leftold--;
443 			break;
444 		case ' ':
445 			leftold--;
446 			leftnew--;
447 			break;
448 		case '+':
449 			leftnew--;
450 			break;
451 		default:
452 			err = got_error(GOT_ERR_PATCH_MALFORMED);
453 			goto done;
454 		}
455 
456 		if (leftold < 0 || leftnew < 0) {
457 			err = got_error(GOT_ERR_PATCH_MALFORMED);
458 			goto done;
459 		}
460 
461 		err = send_line(line);
462 		if (err)
463 			goto done;
464 
465 		if ((ch == '-' && leftold == 0) ||
466 		    (ch == '+' && leftnew == 0)) {
467 			err = peek_special_line(fp);
468 			if (err)
469 				goto done;
470 		}
471 	}
472 
473 done:
474 	free(line);
475 	return err;
476 }
477 
478 static const struct got_error *
479 read_patch(struct imsgbuf *ibuf, int fd)
480 {
481 	const struct got_error *err = NULL;
482 	FILE *fp;
483 	int patch_found = 0;
484 
485 	if ((fp = fdopen(fd, "r")) == NULL) {
486 		err = got_error_from_errno("fdopen");
487 		close(fd);
488 		return err;
489 	}
490 
491 	while (!feof(fp)) {
492 		int done = 0;
493 
494 		err = find_patch(&done, fp);
495 		if (err)
496 			goto done;
497 
498 		patch_found = 1;
499 
500 		while (!done) {
501 			err = parse_hunk(fp, &done);
502 			if (err)
503 				goto done;
504 		}
505 
506 		err = send_patch_done();
507 		if (err)
508 			goto done;
509 	}
510 
511 done:
512 	fclose(fp);
513 
514 	/* ignore trailing gibberish */
515 	if (err != NULL && err->code == GOT_ERR_NO_PATCH && patch_found)
516 		err = NULL;
517 
518 	return err;
519 }
520 
521 int
522 main(int argc, char **argv)
523 {
524 	const struct got_error *err = NULL;
525 	struct imsg imsg;
526 #if 0
527 	static int attached;
528 	while (!attached)
529 		sleep(1);
530 #endif
531 
532 	imsg_init(&ibuf, GOT_IMSG_FD_CHILD);
533 #ifndef PROFILE
534 	/* revoke access to most system calls */
535 	if (pledge("stdio recvfd", NULL) == -1) {
536 		err = got_error_from_errno("pledge");
537 		got_privsep_send_error(&ibuf, err);
538 		return 1;
539 	}
540 #endif
541 
542 	err = got_privsep_recv_imsg(&imsg, &ibuf, 0);
543 	if (err)
544 		goto done;
545 	if (imsg.hdr.type != GOT_IMSG_PATCH_FILE || imsg.fd == -1) {
546 		err = got_error(GOT_ERR_PRIVSEP_MSG);
547 		goto done;
548 	}
549 
550 	err = read_patch(&ibuf, imsg.fd);
551 	if (err)
552 		goto done;
553 	if (imsg_compose(&ibuf, GOT_IMSG_PATCH_EOF, 0, 0, -1,
554 	    NULL, 0) == -1) {
555 		err = got_error_from_errno("imsg_compose GOT_IMSG_PATCH_EOF");
556 		goto done;
557 	}
558 	err = got_privsep_flush_imsg(&ibuf);
559 done:
560 	imsg_free(&imsg);
561 	if (err != NULL) {
562 		got_privsep_send_error(&ibuf, err);
563 		err = NULL;
564 	}
565 	if (close(GOT_IMSG_FD_CHILD) == -1 && err == NULL)
566 		err = got_error_from_errno("close");
567 	if (err && err->code != GOT_ERR_PRIVSEP_PIPE)
568 		fprintf(stderr, "%s: %s\n", getprogname(), err->msg);
569 	return err ? 1 : 0;
570 }