1 /*
2  * Copyright (c) 2022 Omar Polo <op@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  *
16  * Apply patches.
17  *
18  * Things that are still missing:
19  *     + "No final newline" handling
20  *
21  * Things that we may want to support:
22  *     + support indented patches?
23  *     + support other kinds of patches?
24  */
25 
26 #include <sys/types.h>
27 #include <sys/queue.h>
28 #include <sys/socket.h>
29 #include <sys/uio.h>
30 
31 #include <limits.h>
32 #include <stdint.h>
33 #include <stdio.h>
34 #include <stdlib.h>
35 #include <string.h>
36 #include <unistd.h>
37 #include <imsg.h>
38 
39 #include "got_error.h"
40 #include "got_object.h"
41 #include "got_path.h"
42 #include "got_reference.h"
43 #include "got_cancel.h"
44 #include "got_worktree.h"
45 #include "got_opentemp.h"
46 #include "got_patch.h"
47 
48 #include "got_lib_delta.h"
49 #include "got_lib_object.h"
50 #include "got_lib_privsep.h"
51 
52 #define MIN(a, b) ((a) < (b) ? (a) : (b))
53 
54 struct got_patch_hunk {
55 	STAILQ_ENTRY(got_patch_hunk) entries;
56 	long	old_from;
57 	long	old_lines;
58 	long	new_from;
59 	long	new_lines;
60 	size_t	len;
61 	size_t	cap;
62 	char	**lines;
63 };
64 
65 struct got_patch {
66 	char	*old;
67 	char	*new;
68 	STAILQ_HEAD(, got_patch_hunk) head;
69 };
70 
71 static const struct got_error *
72 send_patch(struct imsgbuf *ibuf, int fd)
73 {
74 	const struct got_error *err = NULL;
75 
76 	if (imsg_compose(ibuf, GOT_IMSG_PATCH_FILE, 0, 0, fd,
77 	    NULL, 0) == -1) {
78 		err = got_error_from_errno(
79 		    "imsg_compose GOT_IMSG_PATCH_FILE");
80 		close(fd);
81 		return err;
82 	}
83 
84 	if (imsg_flush(ibuf) == -1) {
85 		err = got_error_from_errno("imsg_flush");
86 		imsg_clear(ibuf);
87 	}
88 
89 	return err;
90 }
91 
92 static void
93 patch_free(struct got_patch *p)
94 {
95 	struct got_patch_hunk *h;
96 	size_t i;
97 
98 	while (!STAILQ_EMPTY(&p->head)) {
99 		h = STAILQ_FIRST(&p->head);
100 		STAILQ_REMOVE_HEAD(&p->head, entries);
101 
102 		for (i = 0; i < h->len; ++i)
103 			free(h->lines[i]);
104 		free(h->lines);
105 		free(h);
106 	}
107 
108 	free(p->new);
109 	free(p->old);
110 }
111 
112 static const struct got_error *
113 pushline(struct got_patch_hunk *h, const char *line)
114 {
115 	void 	*t;
116 	size_t	 newcap;
117 
118 	if (h->len == h->cap) {
119 		if ((newcap = h->cap * 1.5) == 0)
120 			newcap = 16;
121 		t = recallocarray(h->lines, h->cap, newcap,
122 		    sizeof(h->lines[0]));
123 		if (t == NULL)
124 			return got_error_from_errno("recallocarray");
125 		h->lines = t;
126 		h->cap = newcap;
127 	}
128 
129 	if ((t = strdup(line)) == NULL)
130 		return got_error_from_errno("strdup");
131 
132 	h->lines[h->len++] = t;
133 	return NULL;
134 }
135 
136 static const struct got_error *
137 recv_patch(struct imsgbuf *ibuf, int *done, struct got_patch *p)
138 {
139 	const struct got_error *err = NULL;
140 	struct imsg imsg;
141 	struct got_imsg_patch_hunk hdr;
142 	struct got_imsg_patch patch;
143 	struct got_patch_hunk *h = NULL;
144 	size_t datalen;
145 
146 	memset(p, 0, sizeof(*p));
147 	STAILQ_INIT(&p->head);
148 
149 	err = got_privsep_recv_imsg(&imsg, ibuf, 0);
150 	if (err)
151 		return err;
152 	if (imsg.hdr.type == GOT_IMSG_PATCH_EOF) {
153 		*done = 1;
154 		goto done;
155 	}
156 	if (imsg.hdr.type != GOT_IMSG_PATCH) {
157 		err = got_error(GOT_ERR_PRIVSEP_MSG);
158 		goto done;
159 	}
160 	datalen = imsg.hdr.len - IMSG_HEADER_SIZE;
161 	if (datalen != sizeof(patch)) {
162 		err = got_error(GOT_ERR_PRIVSEP_LEN);
163 		goto done;
164 	}
165 	memcpy(&patch, imsg.data, sizeof(patch));
166 	if (*patch.old != '\0' && (p->old = strdup(patch.old)) == NULL) {
167 		err = got_error_from_errno("strdup");
168 		goto done;
169 	}
170 	if (*patch.new != '\0' && (p->new = strdup(patch.new)) == NULL) {
171 		err = got_error_from_errno("strdup");
172 		goto done;
173 	}
174 
175 	imsg_free(&imsg);
176 
177 	for (;;) {
178 		char *t;
179 
180 		err = got_privsep_recv_imsg(&imsg, ibuf, 0);
181 		if (err)
182 			return err;
183 
184 		switch (imsg.hdr.type) {
185 		case GOT_IMSG_PATCH_DONE:
186 			goto done;
187 		case GOT_IMSG_PATCH_HUNK:
188 			datalen = imsg.hdr.len - IMSG_HEADER_SIZE;
189 			if (datalen != sizeof(hdr)) {
190 				err = got_error(GOT_ERR_PRIVSEP_LEN);
191 				goto done;
192 			}
193 			memcpy(&hdr, imsg.data, sizeof(hdr));
194 			if ((h = calloc(1, sizeof(*h))) == NULL) {
195 				err = got_error_from_errno("calloc");
196 				goto done;
197 			}
198 			h->old_from = hdr.oldfrom;
199 			h->old_lines = hdr.oldlines;
200 			h->new_from = hdr.newfrom;
201 			h->new_lines = hdr.newlines;
202 			STAILQ_INSERT_TAIL(&p->head, h, entries);
203 			break;
204 		case GOT_IMSG_PATCH_LINE:
205 			if (h == NULL) {
206 				err = got_error(GOT_ERR_PRIVSEP_MSG);
207 				goto done;
208 			}
209 			datalen = imsg.hdr.len - IMSG_HEADER_SIZE;
210 			t = imsg.data;
211 			/* at least one char plus newline */
212 			if (datalen < 2 || t[datalen-1] != '\0') {
213 				err = got_error(GOT_ERR_PRIVSEP_MSG);
214 				goto done;
215 			}
216 			if (*t != ' ' && *t != '-' && *t != '+') {
217 				err = got_error(GOT_ERR_PRIVSEP_MSG);
218 				goto done;
219 			}
220 			err = pushline(h, t);
221 			if (err)
222 				goto done;
223 			break;
224 		default:
225 			err = got_error(GOT_ERR_PRIVSEP_MSG);
226 			goto done;
227 		}
228 
229 		imsg_free(&imsg);
230 	}
231 
232 done:
233 	imsg_free(&imsg);
234 	return err;
235 }
236 
237 /*
238  * Copy data from orig starting at copypos until pos into tmp.
239  * If pos is -1, copy until EOF.
240  */
241 static const struct got_error *
242 copy(FILE *tmp, FILE *orig, off_t copypos, off_t pos)
243 {
244 	char buf[BUFSIZ];
245 	size_t len, r, w;
246 
247 	if (fseek(orig, copypos, SEEK_SET) == -1)
248 		return got_error_from_errno("fseek");
249 
250 	while (pos == -1 || copypos < pos) {
251 		len = sizeof(buf);
252 		if (pos > 0)
253 			len = MIN(len, (size_t)pos - copypos);
254 		r = fread(buf, 1, len, orig);
255 		if (r != len && ferror(orig))
256 			return got_error_from_errno("fread");
257 		w = fwrite(buf, 1, r, tmp);
258 		if (w != r)
259 			return got_error_from_errno("fwrite");
260 		copypos += len;
261 		if (r != len && feof(orig)) {
262 			if (pos == -1)
263 				return NULL;
264 			return got_error(GOT_ERR_PATCH_DONT_APPLY);
265 		}
266 	}
267 	return NULL;
268 }
269 
270 static const struct got_error *
271 locate_hunk(FILE *orig, struct got_patch_hunk *h, long *lineno)
272 {
273 	const struct got_error *err = NULL;
274 	char *line = NULL;
275 	char mode = *h->lines[0];
276 	size_t linesize = 0;
277 	ssize_t linelen;
278 	off_t match = -1;
279 	long match_lineno = -1;
280 
281 	for (;;) {
282 		linelen = getline(&line, &linesize, orig);
283 		if (linelen == -1) {
284 			if (ferror(orig))
285 				err = got_error_from_errno("getline");
286 			else if (match == -1)
287 				err = got_error(GOT_ERR_PATCH_DONT_APPLY);
288 			break;
289 		}
290 		(*lineno)++;
291 
292 		if ((mode == ' ' && !strcmp(h->lines[0]+1, line)) ||
293 		    (mode == '-' && !strcmp(h->lines[0]+1, line)) ||
294 		    (mode == '+' && *lineno == h->old_from)) {
295 			match = ftello(orig);
296 			if (match == -1) {
297 				err = got_error_from_errno("ftello");
298 				break;
299 			}
300 			match -= linelen;
301 			match_lineno = (*lineno)-1;
302 		}
303 
304 		if (*lineno >= h->old_from && match != -1)
305 			break;
306 	}
307 
308 	if (err == NULL) {
309 		*lineno = match_lineno;
310 		if (fseek(orig, match, SEEK_SET) == -1)
311 			err = got_error_from_errno("fseek");
312 	}
313 
314 	free(line);
315 	return err;
316 }
317 
318 static const struct got_error *
319 test_hunk(FILE *orig, struct got_patch_hunk *h)
320 {
321 	const struct got_error *err = NULL;
322 	char *line = NULL;
323 	size_t linesize = 0, i = 0;
324 	ssize_t linelen;
325 
326 	for (i = 0; i < h->len; ++i) {
327 		switch (*h->lines[i]) {
328 		case '+':
329 			continue;
330 		case ' ':
331 		case '-':
332 			linelen = getline(&line, &linesize, orig);
333 			if (linelen == -1) {
334 				if (ferror(orig))
335 					err = got_error_from_errno("getline");
336 				else
337 					err = got_error(
338 					    GOT_ERR_PATCH_DONT_APPLY);
339 				goto done;
340 			}
341 			if (strcmp(h->lines[i]+1, line)) {
342 				err = got_error(GOT_ERR_PATCH_DONT_APPLY);
343 				goto done;
344 			}
345 			break;
346 		}
347 	}
348 
349 done:
350 	free(line);
351 	return err;
352 }
353 
354 static const struct got_error *
355 apply_hunk(FILE *tmp, struct got_patch_hunk *h, long *lineno)
356 {
357 	size_t i = 0;
358 
359 	for (i = 0; i < h->len; ++i) {
360 		switch (*h->lines[i]) {
361 		case ' ':
362 			if (fprintf(tmp, "%s", h->lines[i]+1) < 0)
363 				return got_error_from_errno("fprintf");
364 			/* fallthrough */
365 		case '-':
366 			(*lineno)++;
367 			break;
368 		case '+':
369 			if (fprintf(tmp, "%s", h->lines[i]+1) < 0)
370 				return got_error_from_errno("fprintf");
371 			break;
372 		}
373 	}
374 	return NULL;
375 }
376 
377 static const struct got_error *
378 apply_patch(struct got_worktree *worktree, struct got_repository *repo,
379     struct got_patch *p, got_worktree_delete_cb delete_cb,
380     got_worktree_checkout_cb add_cb)
381 {
382 	const struct got_error *err = NULL;
383 	struct got_pathlist_head paths;
384 	struct got_pathlist_entry *pe;
385 	char *path = NULL, *tmppath = NULL;
386 	FILE *orig = NULL, *tmp = NULL;
387 	struct got_patch_hunk *h;
388 	size_t i;
389 	long lineno = 0;
390 	off_t copypos, pos;
391 	char *line = NULL;
392 	size_t linesize = 0;
393 	ssize_t linelen;
394 
395 	TAILQ_INIT(&paths);
396 
397 	if (p->old == NULL && p->new == NULL)
398 		return got_error(GOT_ERR_PATCH_MALFORMED);
399 
400 	err = got_worktree_resolve_path(&path, worktree,
401 	    p->new != NULL ? p->new : p->old);
402 	if (err)
403 		return err;
404 	err = got_pathlist_insert(&pe, &paths, path, NULL);
405 	if (err)
406 		goto done;
407 
408 	if (p->old != NULL && p->new == NULL) {
409 		/*
410 		 * special case: delete a file.  don't try to match
411 		 * the lines but just schedule the removal.
412 		 */
413 		err = got_worktree_schedule_delete(worktree, &paths,
414 		    0, NULL, delete_cb, NULL, repo, 0, 0);
415 		goto done;
416 	} else if (p->old != NULL && strcmp(p->old, p->new)) {
417 		err = got_error(GOT_ERR_PATCH_PATHS_DIFFER);
418 		goto done;
419 	}
420 
421 	err = got_opentemp_named(&tmppath, &tmp,
422 	    got_worktree_get_root_path(worktree));
423 	if (err)
424 		goto done;
425 
426 	if (p->old == NULL) {				/* create */
427 		h = STAILQ_FIRST(&p->head);
428 		if (h == NULL || STAILQ_NEXT(h, entries) != NULL) {
429 			err = got_error(GOT_ERR_PATCH_MALFORMED);
430 			goto done;
431 		}
432 		for (i = 0; i < h->len; ++i) {
433 			if (fprintf(tmp, "%s", h->lines[i]+1) < 0) {
434 				err = got_error_from_errno("fprintf");
435 				goto done;
436 			}
437 		}
438 		goto rename;
439 	}
440 
441 	if ((orig = fopen(path, "r")) == NULL) {
442 		err = got_error_from_errno2("fopen", path);
443 		goto done;
444 	}
445 
446 	copypos = 0;
447 	STAILQ_FOREACH(h, &p->head, entries) {
448 	tryagain:
449 		err = locate_hunk(orig, h, &lineno);
450 		if (err != NULL)
451 			goto done;
452 		if ((pos = ftello(orig)) == -1) {
453 			err = got_error_from_errno("ftello");
454 			goto done;
455 		}
456 		err = copy(tmp, orig, copypos, pos);
457 		if (err != NULL)
458 			goto done;
459 		copypos = pos;
460 
461 		err = test_hunk(orig, h);
462 		if (err != NULL && err->code == GOT_ERR_PATCH_DONT_APPLY) {
463 			/*
464 			 * try to apply the hunk again starting the search
465 			 * after the previous partial match.
466 			 */
467 			if (fseek(orig, pos, SEEK_SET) == -1) {
468 				err = got_error_from_errno("fseek");
469 				goto done;
470 			}
471 			linelen = getline(&line, &linesize, orig);
472 			if (linelen == -1) {
473 				err = got_error_from_errno("getline");
474 				goto done;
475 			}
476 			lineno++;
477 			goto tryagain;
478 		}
479 		if (err != NULL)
480 			goto done;
481 
482 		err = apply_hunk(tmp, h, &lineno);
483 		if (err != NULL)
484 			goto done;
485 		
486 		copypos = ftello(orig);
487 		if (copypos == -1) {
488 			err = got_error_from_errno("ftello");
489 			goto done;
490 		}
491 	}
492 
493 	if (!feof(orig)) {
494 		err = copy(tmp, orig, copypos, -1);
495 		if (err)
496 			goto done;
497 	}
498 
499 rename:
500 	if (rename(tmppath, path) == -1) {
501 		err = got_error_from_errno3("rename", tmppath, path);
502 		goto done;
503 	}
504 
505 	if (p->old == NULL)
506 		err = got_worktree_schedule_add(worktree, &paths,
507 		    add_cb, NULL, repo, 1);
508 	else
509 		printf("M  %s\n", path); /* XXX */
510 done:
511 	if (err != NULL && p->old == NULL && path != NULL)
512 		unlink(path);
513 	if (tmp != NULL)
514 		fclose(tmp);
515 	if (tmppath != NULL)
516 		unlink(tmppath);
517 	free(tmppath);
518 	if (orig != NULL) {
519 		if (p->old == NULL && err != NULL)
520 			unlink(path);
521 		fclose(orig);
522 	}
523 	free(path);
524 	free(line);
525 	got_pathlist_free(&paths);
526 	return err;
527 }
528 
529 const struct got_error *
530 got_patch(int fd, struct got_worktree *worktree, struct got_repository *repo,
531     got_worktree_delete_cb delete_cb, got_worktree_checkout_cb add_cb)
532 {
533 	const struct got_error *err = NULL;
534 	struct imsgbuf *ibuf;
535 	int imsg_fds[2] = {-1, -1};
536 	int done = 0;
537 	pid_t pid;
538 
539 	ibuf = calloc(1, sizeof(*ibuf));
540 	if (ibuf == NULL) {
541 		err = got_error_from_errno("calloc");
542 		goto done;
543 	}
544 
545 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fds) == -1) {
546 		err = got_error_from_errno("socketpair");
547 		goto done;
548 	}
549 
550 	pid = fork();
551 	if (pid == -1) {
552 		err = got_error_from_errno("fork");
553 		goto done;
554 	} else if (pid == 0) {
555 		got_privsep_exec_child(imsg_fds, GOT_PATH_PROG_READ_PATCH,
556 		    NULL);
557 		/* not reached */
558 	}
559 
560 	if (close(imsg_fds[1]) == -1) {
561 		err = got_error_from_errno("close");
562 		goto done;
563 	}
564 	imsg_fds[1] = -1;
565 	imsg_init(ibuf, imsg_fds[0]);
566 
567 	err = send_patch(ibuf, fd);
568 	fd = -1;
569 	if (err)
570 		goto done;
571 
572 	while (!done && err == NULL) {
573 		struct got_patch p;
574 
575 		err = recv_patch(ibuf, &done, &p);
576 		if (err || done)
577 			break;
578 
579 		err = apply_patch(worktree, repo, &p, delete_cb, add_cb);
580 		patch_free(&p);
581 		if (err)
582 			break;
583 	}
584 
585 done:
586 	if (fd != -1 && close(fd) == -1 && err == NULL)
587 		err = got_error_from_errno("close");
588 	if (ibuf != NULL)
589 		imsg_clear(ibuf);
590 	if (imsg_fds[0] != -1 && close(imsg_fds[0]) == -1 && err == NULL)
591 		err = got_error_from_errno("close");
592 	if (imsg_fds[1] != -1 && close(imsg_fds[1]) == -1 && err == NULL)
593 		err = got_error_from_errno("close");
594 	return err;
595 }