op public repos

Blob

Date:: Thu Oct 18 09:45:20 2018 UTC
Message:: make amount of diff context configurable at API level
Actions:: History | Blame | Raw File
1 /*
2  * Copyright (c) 2017 Stefan Sperling <stsp@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 
17 #include <sys/queue.h>
18 #include <sys/stat.h>
19 
20 #include <stdio.h>
21 #include <stdlib.h>
22 #include <string.h>
23 #include <limits.h>
24 #include <sha1.h>
25 #include <zlib.h>
26 
27 #include "got_repository.h"
28 #include "got_object.h"
29 #include "got_error.h"
30 #include "got_diff.h"
31 #include "got_opentemp.h"
32 
33 #include "got_lib_diff.h"
34 #include "got_lib_path.h"
35 
36 static const struct got_error *
37 diff_blobs(struct got_blob_object *blob1, struct got_blob_object *blob2,
38     const char *label1, const char *label2, int diff_context, FILE *outfile,
39     struct got_diff_changes *changes)
40 {
41 	struct got_diff_state ds;
42 	struct got_diff_args args;
43 	const struct got_error *err = NULL;
44 	FILE *f1 = NULL, *f2 = NULL;
45 	char hex1[SHA1_DIGEST_STRING_LENGTH];
46 	char hex2[SHA1_DIGEST_STRING_LENGTH];
47 	char *idstr1 = NULL, *idstr2 = NULL;
48 	size_t size1, size2;
49 	int res, flags = 0;
50 
51 	if (blob1) {
52 		f1 = got_opentemp();
53 		if (f1 == NULL)
54 			return got_error(GOT_ERR_FILE_OPEN);
55 	} else
56 		flags |= D_EMPTY1;
57 
58 	if (blob2) {
59 		f2 = got_opentemp();
60 		if (f2 == NULL) {
61 			fclose(f1);
62 			return got_error(GOT_ERR_FILE_OPEN);
63 		}
64 	} else
65 		flags |= D_EMPTY2;
66 
67 	size1 = 0;
68 	if (blob1) {
69 		idstr1 = got_object_blob_id_str(blob1, hex1, sizeof(hex1));
70 		err = got_object_blob_dump_to_file(&size1, NULL, f1, blob1);
71 		if (err)
72 			goto done;
73 	} else
74 		idstr1 = "/dev/null";
75 
76 	size2 = 0;
77 	if (blob2) {
78 		idstr2 = got_object_blob_id_str(blob2, hex2, sizeof(hex2));
79 		err = got_object_blob_dump_to_file(&size2, NULL, f2, blob2);
80 		if (err)
81 			goto done;
82 	} else
83 		idstr2 = "/dev/null";
84 
85 	memset(&ds, 0, sizeof(ds));
86 	/* XXX should stat buffers be passed in args instead of ds? */
87 	ds.stb1.st_mode = S_IFREG;
88 	if (blob1)
89 		ds.stb1.st_size = size1;
90 	ds.stb1.st_mtime = 0; /* XXX */
91 
92 	ds.stb2.st_mode = S_IFREG;
93 	if (blob2)
94 		ds.stb2.st_size = size2;
95 	ds.stb2.st_mtime = 0; /* XXX */
96 
97 	memset(&args, 0, sizeof(args));
98 	args.diff_format = D_UNIFIED;
99 	args.label[0] = label1 ? label1 : idstr1;
100 	args.label[1] = label2 ? label2 : idstr2;
101 	args.diff_context = diff_context;
102 	flags |= D_PROTOTYPE;
103 
104 	if (label1 && strcmp(label1, idstr1) != 0)
105 		fprintf(outfile, "blob - %s\n", idstr1);
106 	if (label2 && strcmp(label2, idstr2) != 0)
107 		fprintf(outfile, "blob + %s\n", idstr2);
108 
109 	err = got_diffreg(&res, f1, f2, flags, &args, &ds, outfile, changes);
110 done:
111 	if (f1)
112 		fclose(f1);
113 	if (f2)
114 		fclose(f2);
115 	return err;
116 }
117 
118 const struct got_error *
119 got_diff_blob(struct got_blob_object *blob1, struct got_blob_object *blob2,
120     const char *label1, const char *label2, int diff_context, FILE *outfile)
121 {
122 	return diff_blobs(blob1, blob2, label1, label2, diff_context, outfile,
123 	    NULL);
124 }
125 
126 const struct got_error *
127 got_diff_blob_lines_changed(struct got_diff_changes **changes,
128     struct got_blob_object *blob1, struct got_blob_object *blob2)
129 {
130 	const struct got_error *err = NULL;
131 
132 	*changes = calloc(1, sizeof(**changes));
133 	if (*changes == NULL)
134 		return got_error_from_errno();
135 	SIMPLEQ_INIT(&(*changes)->entries);
136 
137 	err = diff_blobs(blob1, blob2, NULL, NULL, 3, NULL, *changes);
138 	if (err) {
139 		got_diff_free_changes(*changes);
140 		*changes = NULL;
141 	}
142 	return err;
143 }
144 
145 void
146 got_diff_free_changes(struct got_diff_changes *changes)
147 {
148 	struct got_diff_change *change;
149 	while (!SIMPLEQ_EMPTY(&changes->entries)) {
150 		change = SIMPLEQ_FIRST(&changes->entries);
151 		SIMPLEQ_REMOVE_HEAD(&changes->entries, entry);
152 		free(change);
153 	}
154 	free(changes);
155 }
156 
157 struct got_tree_entry *
158 match_entry_by_name(struct got_tree_entry *te1, struct got_tree_object *tree2)
159 {
160 	struct got_tree_entry *te2;
161 	const struct got_tree_entries *entries2;
162 
163 	entries2 = got_object_tree_get_entries(tree2); 
164 	SIMPLEQ_FOREACH(te2, &entries2->head, entry) {
165 		if (strcmp(te1->name, te2->name) == 0)
166 			return te2;
167 	}
168 	return NULL;
169 }
170 
171 static const struct got_error *
172 diff_added_blob(struct got_object_id *id, const char *label,
173     int diff_context, struct got_repository *repo, FILE *outfile)
174 {
175 	const struct got_error *err;
176 	struct got_blob_object  *blob = NULL;
177 	struct got_object *obj = NULL;
178 
179 	err = got_object_open(&obj, repo, id);
180 	if (err)
181 		return err;
182 
183 	err = got_object_blob_open(&blob, repo, obj, 8192);
184 	if (err)
185 		goto done;
186 	err = got_diff_blob(NULL, blob, NULL, label, diff_context, outfile);
187 done:
188 	got_object_close(obj);
189 	if (blob)
190 		got_object_blob_close(blob);
191 	return err;
192 }
193 
194 static const struct got_error *
195 diff_modified_blob(struct got_object_id *id1, struct got_object_id *id2,
196     const char *label1, const char *label2, int diff_context,
197     struct got_repository *repo, FILE *outfile)
198 {
199 	const struct got_error *err;
200 	struct got_object *obj1 = NULL;
201 	struct got_object *obj2 = NULL;
202 	struct got_blob_object *blob1 = NULL;
203 	struct got_blob_object *blob2 = NULL;
204 
205 	err = got_object_open(&obj1, repo, id1);
206 	if (err)
207 		return err;
208 	if (got_object_get_type(obj1) != GOT_OBJ_TYPE_BLOB) {
209 		err = got_error(GOT_ERR_OBJ_TYPE);
210 		goto done;
211 	}
212 
213 	err = got_object_open(&obj2, repo, id2);
214 	if (err)
215 		goto done;
216 	if (got_object_get_type(obj2) != GOT_OBJ_TYPE_BLOB) {
217 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
218 		goto done;
219 	}
220 
221 	err = got_object_blob_open(&blob1, repo, obj1, 8192);
222 	if (err)
223 		goto done;
224 
225 	err = got_object_blob_open(&blob2, repo, obj2, 8192);
226 	if (err)
227 		goto done;
228 
229 	err = got_diff_blob(blob1, blob2, label1, label2, diff_context,
230 	    outfile);
231 
232 done:
233 	if (obj1)
234 		got_object_close(obj1);
235 	if (obj2)
236 		got_object_close(obj2);
237 	if (blob1)
238 		got_object_blob_close(blob1);
239 	if (blob2)
240 		got_object_blob_close(blob2);
241 	return err;
242 }
243 
244 static const struct got_error *
245 diff_deleted_blob(struct got_object_id *id, const char *label,
246     int diff_context, struct got_repository *repo, FILE *outfile)
247 {
248 	const struct got_error *err;
249 	struct got_blob_object  *blob = NULL;
250 	struct got_object *obj = NULL;
251 
252 	err = got_object_open(&obj, repo, id);
253 	if (err)
254 		return err;
255 
256 	err = got_object_blob_open(&blob, repo, obj, 8192);
257 	if (err)
258 		goto done;
259 	err = got_diff_blob(blob, NULL, label, NULL, diff_context, outfile);
260 done:
261 	got_object_close(obj);
262 	if (blob)
263 		got_object_blob_close(blob);
264 	return err;
265 }
266 
267 static const struct got_error *
268 diff_added_tree(struct got_object_id *id, const char *label,
269     int diff_context, struct got_repository *repo, FILE *outfile)
270 {
271 	const struct got_error *err = NULL;
272 	struct got_object *treeobj = NULL;
273 	struct got_tree_object *tree = NULL;
274 
275 	err = got_object_open(&treeobj, repo, id);
276 	if (err)
277 		goto done;
278 
279 	if (got_object_get_type(treeobj) != GOT_OBJ_TYPE_TREE) {
280 		err = got_error(GOT_ERR_OBJ_TYPE);
281 		goto done;
282 	}
283 
284 	err = got_object_tree_open(&tree, repo, treeobj);
285 	if (err)
286 		goto done;
287 
288 	err = got_diff_tree(NULL, tree, NULL, label, diff_context, repo,
289 	    outfile);
290 
291 done:
292 	if (tree)
293 		got_object_tree_close(tree);
294 	if (treeobj)
295 		got_object_close(treeobj);
296 	return err;
297 }
298 
299 static const struct got_error *
300 diff_modified_tree(struct got_object_id *id1, struct got_object_id *id2,
301     const char *label1, const char *label2, int diff_context,
302     struct got_repository *repo, FILE *outfile)
303 {
304 	const struct got_error *err;
305 	struct got_object *treeobj1 = NULL;
306 	struct got_object *treeobj2 = NULL;
307 	struct got_tree_object *tree1 = NULL;
308 	struct got_tree_object *tree2 = NULL;
309 
310 	err = got_object_open(&treeobj1, repo, id1);
311 	if (err)
312 		goto done;
313 
314 	if (got_object_get_type(treeobj1) != GOT_OBJ_TYPE_TREE) {
315 		err = got_error(GOT_ERR_OBJ_TYPE);
316 		goto done;
317 	}
318 
319 	err = got_object_open(&treeobj2, repo, id2);
320 	if (err)
321 		goto done;
322 
323 	if (got_object_get_type(treeobj2) != GOT_OBJ_TYPE_TREE) {
324 		err = got_error(GOT_ERR_OBJ_TYPE);
325 		goto done;
326 	}
327 
328 	err = got_object_tree_open(&tree1, repo, treeobj1);
329 	if (err)
330 		goto done;
331 
332 	err = got_object_tree_open(&tree2, repo, treeobj2);
333 	if (err)
334 		goto done;
335 
336 	err = got_diff_tree(tree1, tree2, label1, label2, diff_context, repo,
337 	    outfile);
338 
339 done:
340 	if (tree1)
341 		got_object_tree_close(tree1);
342 	if (tree2)
343 		got_object_tree_close(tree2);
344 	if (treeobj1)
345 		got_object_close(treeobj1);
346 	if (treeobj2)
347 		got_object_close(treeobj2);
348 	return err;
349 }
350 
351 static const struct got_error *
352 diff_deleted_tree(struct got_object_id *id, const char *label,
353     int diff_context, struct got_repository *repo, FILE *outfile)
354 {
355 	const struct got_error *err;
356 	struct got_object *treeobj = NULL;
357 	struct got_tree_object *tree = NULL;
358 
359 	err = got_object_open(&treeobj, repo, id);
360 	if (err)
361 		goto done;
362 
363 	if (got_object_get_type(treeobj) != GOT_OBJ_TYPE_TREE) {
364 		err = got_error(GOT_ERR_OBJ_TYPE);
365 		goto done;
366 	}
367 
368 	err = got_object_tree_open(&tree, repo, treeobj);
369 	if (err)
370 		goto done;
371 
372 	err = got_diff_tree(tree, NULL, label, NULL, diff_context, repo,
373 	    outfile);
374 done:
375 	if (tree)
376 		got_object_tree_close(tree);
377 	if (treeobj)
378 		got_object_close(treeobj);
379 	return err;
380 }
381 
382 static const struct got_error *
383 diff_kind_mismatch(struct got_object_id *id1, struct got_object_id *id2,
384     const char *label1, const char *label2, FILE *outfile)
385 {
386 	/* XXX TODO */
387 	return NULL;
388 }
389 
390 static const struct got_error *
391 diff_entry_old_new(struct got_tree_entry *te1, struct got_tree_entry *te2,
392     const char *label1, const char *label2, int diff_context,
393     struct got_repository *repo, FILE *outfile)
394 {
395 	const struct got_error *err = NULL;
396 
397 	if (te2 == NULL) {
398 		if (S_ISDIR(te1->mode))
399 			err = diff_deleted_tree(te1->id, label1, diff_context,
400 			    repo, outfile);
401 		else
402 			err = diff_deleted_blob(te1->id, label1, diff_context,
403 			    repo, outfile);
404 		return err;
405 	}
406 
407 	if (S_ISDIR(te1->mode) && S_ISDIR(te2->mode)) {
408 		if (got_object_id_cmp(te1->id, te2->id) != 0)
409 			return diff_modified_tree(te1->id, te2->id,
410 			    label1, label2, diff_context, repo, outfile);
411 	} else if (S_ISREG(te1->mode) && S_ISREG(te2->mode)) {
412 		if (got_object_id_cmp(te1->id, te2->id) != 0)
413 			return diff_modified_blob(te1->id, te2->id,
414 			    label1, label2, diff_context, repo, outfile);
415 	}
416 
417 	if (got_object_id_cmp(te1->id, te2->id) == 0)
418 		return NULL;
419 
420 	return diff_kind_mismatch(te1->id, te2->id, label1, label2, outfile);
421 }
422 
423 static const struct got_error *
424 diff_entry_new_old(struct got_tree_entry *te2, struct got_tree_entry *te1,
425     const char *label2, int diff_context, struct got_repository *repo,
426     FILE *outfile)
427 {
428 	if (te1 != NULL) /* handled by diff_entry_old_new() */
429 		return NULL;
430 
431 	if (S_ISDIR(te2->mode))
432 		return diff_added_tree(te2->id, label2, diff_context, repo,
433 		    outfile);
434 
435 	return diff_added_blob(te2->id, label2, diff_context, repo, outfile);
436 }
437 
438 const struct got_error *
439 got_diff_tree(struct got_tree_object *tree1, struct got_tree_object *tree2,
440     const char *label1, const char *label2, int diff_context,
441     struct got_repository *repo, FILE *outfile)
442 {
443 	const struct got_error *err = NULL;
444 	struct got_tree_entry *te1 = NULL;
445 	struct got_tree_entry *te2 = NULL;
446 	char *l1 = NULL, *l2 = NULL;
447 
448 	if (tree1) {
449 		const struct got_tree_entries *entries;
450 		entries = got_object_tree_get_entries(tree1);
451 		te1 = SIMPLEQ_FIRST(&entries->head);
452 		if (te1 && asprintf(&l1, "%s%s%s", label1, label1[0] ? "/" : "",
453 		    te1->name) == -1)
454 			return got_error_from_errno();
455 	}
456 	if (tree2) {
457 		const struct got_tree_entries *entries;
458 		entries = got_object_tree_get_entries(tree2);
459 		te2 = SIMPLEQ_FIRST(&entries->head);
460 		if (te2 && asprintf(&l2, "%s%s%s", label2, label2[0] ? "/" : "",
461 		    te2->name) == -1)
462 			return got_error_from_errno();
463 	}
464 
465 	do {
466 		if (te1) {
467 			struct got_tree_entry *te = NULL;
468 			if (tree2)
469 				te = match_entry_by_name(te1, tree2);
470 			if (te) {
471 				free(l2);
472 				l2 = NULL;
473 				if (te && asprintf(&l2, "%s%s%s", label2,
474 				    label2[0] ? "/" : "", te->name) == -1)
475 					return got_error_from_errno();
476 			}
477 			err = diff_entry_old_new(te1, te, l1, l2, diff_context,
478 			    repo, outfile);
479 			if (err)
480 				break;
481 		}
482 
483 		if (te2) {
484 			struct got_tree_entry *te = NULL;
485 			if (tree1)
486 				te = match_entry_by_name(te2, tree1);
487 			err = diff_entry_new_old(te2, te, l2, diff_context,
488 			    repo, outfile);
489 			if (err)
490 				break;
491 		}
492 
493 		free(l1);
494 		l1 = NULL;
495 		if (te1) {
496 			te1 = SIMPLEQ_NEXT(te1, entry);
497 			if (te1 &&
498 			    asprintf(&l1, "%s%s%s", label1,
499 			    label1[0] ? "/" : "", te1->name) == -1)
500 				return got_error_from_errno();
501 		}
502 		free(l2);
503 		l2 = NULL;
504 		if (te2) {
505 			te2 = SIMPLEQ_NEXT(te2, entry);
506 			if (te2 &&
507 			    asprintf(&l2, "%s%s%s", label2,
508 			        label2[0] ? "/" : "", te2->name) == -1)
509 				return got_error_from_errno();
510 		}
511 	} while (te1 || te2);
512 
513 	return err;
514 }
515 
516 const struct got_error *
517 got_diff_objects_as_blobs(struct got_object *obj1, struct got_object *obj2,
518     const char *label1, const char *label2, int diff_context,
519     struct got_repository *repo, FILE *outfile)
520 {
521 	const struct got_error *err;
522 	struct got_blob_object *blob1 = NULL, *blob2 = NULL;
523 
524 	if (obj1 == NULL && obj2 == NULL)
525 		return got_error(GOT_ERR_NO_OBJ);
526 
527 	if (obj1) {
528 		err = got_object_blob_open(&blob1, repo, obj1, 8192);
529 		if (err)
530 			goto done;
531 	}
532 	if (obj2) {
533 		err = got_object_blob_open(&blob2, repo, obj2, 8192);
534 		if (err)
535 			goto done;
536 	}
537 	err = got_diff_blob(blob1, blob2, label1, label2, diff_context,
538 	    outfile);
539 done:
540 	if (blob1)
541 		got_object_blob_close(blob1);
542 	if (blob2)
543 		got_object_blob_close(blob2);
544 	return err;
545 }
546 
547 const struct got_error *
548 got_diff_objects_as_trees(struct got_object *obj1, struct got_object *obj2,
549     char *label1, char *label2, int diff_context, struct got_repository *repo,
550     FILE *outfile)
551 {
552 	const struct got_error *err;
553 	struct got_tree_object *tree1 = NULL, *tree2 = NULL;
554 
555 	if (obj1 == NULL && obj2 == NULL)
556 		return got_error(GOT_ERR_NO_OBJ);
557 
558 	if (obj1) {
559 		err = got_object_tree_open(&tree1, repo, obj1);
560 		if (err)
561 			goto done;
562 	}
563 	if (obj2) {
564 		err = got_object_tree_open(&tree2, repo, obj2);
565 		if (err)
566 			goto done;
567 	}
568 	err = got_diff_tree(tree1, tree2, label1, label2, diff_context,
569 	   repo, outfile);
570 done:
571 	if (tree1)
572 		got_object_tree_close(tree1);
573 	if (tree2)
574 		got_object_tree_close(tree2);
575 	return err;
576 }
577 
578 static char *
579 get_datestr(time_t *time, char *datebuf)
580 {
581 	char *p, *s = ctime_r(time, datebuf);
582 	p = strchr(s, '\n');
583 	if (p)
584 		*p = '\0';
585 	return s;
586 }
587 
588 const struct got_error *
589 got_diff_objects_as_commits(struct got_object *obj1, struct got_object *obj2,
590     int diff_context, struct got_repository *repo, FILE *outfile)
591 {
592 	const struct got_error *err;
593 	struct got_commit_object *commit1 = NULL, *commit2 = NULL;
594 	struct got_object *tree_obj1  = NULL, *tree_obj2 = NULL;
595 	char *id_str;
596 	char datebuf[26];
597 	time_t time;
598 
599 	if (obj2 == NULL)
600 		return got_error(GOT_ERR_NO_OBJ);
601 
602 	if (obj1) {
603 		err = got_object_commit_open(&commit1, repo, obj1);
604 		if (err)
605 			goto done;
606 		err = got_object_open(&tree_obj1, repo, commit1->tree_id);
607 		if (err)
608 			goto done;
609 	}
610 
611 	err = got_object_commit_open(&commit2, repo, obj2);
612 	if (err)
613 		goto done;
614 	err = got_object_open(&tree_obj2, repo, commit2->tree_id);
615 	if (err)
616 		goto done;
617 	err = got_object_get_id_str(&id_str, obj2);
618 	if (err)
619 		goto done;
620 	if (fprintf(outfile, "commit: %s\n", id_str) < 0) {
621 		err = got_error_from_errno();
622 		free(id_str);
623 		goto done;
624 	}
625 	free(id_str);
626 	if (fprintf(outfile, "from: %s\n", commit2->author) < 0) {
627 		err = got_error_from_errno();
628 		goto done;
629 	}
630 	time = mktime(&commit2->tm_committer);
631 	if (fprintf(outfile, "date: %s UTC\n",
632 	    get_datestr(&time, datebuf)) < 0) {
633 		err = got_error_from_errno();
634 		goto done;
635 	}
636 	if (strcmp(commit2->author, commit2->committer) != 0 &&
637 	    fprintf(outfile, "via: %s\n", commit2->committer) < 0) {
638 		err = got_error_from_errno();
639 		goto done;
640 	}
641 	if (fprintf(outfile, "%s\n", commit2->logmsg) < 0) {
642 		err = got_error_from_errno();
643 		goto done;
644 	}
645 
646 	err = got_diff_objects_as_trees(tree_obj1, tree_obj2, "", "",
647 	    diff_context, repo, outfile);
648 done:
649 	if (tree_obj1)
650 		got_object_close(tree_obj1);
651 	if (tree_obj2)
652 		got_object_close(tree_obj2);
653 	if (commit1)
654 		got_object_commit_close(commit1);
655 	if (commit2)
656 		got_object_commit_close(commit2);
657 	return err;
658 }