Gotweb

Blob

Date:: Fri Dec 1 21:09:59 2017 UTC
Message:: make tree diffing work
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 
32 #include "diff.h"
33 
34 static FILE *
35 opentemp(void)
36 {
37 	char name[PATH_MAX];
38 	int fd;
39 	FILE *f;
40 
41 	if (strlcpy(name, "/tmp/got.XXXXXXXX", sizeof(name)) >= sizeof(name))
42 		return NULL;
43 
44 	fd = mkstemp(name);
45 	if (fd < 0)
46 		return NULL;
47 
48 	unlink(name);
49 	f = fdopen(fd, "w+");
50 	if (f == NULL) {
51 		close(fd);
52 		return NULL;
53 	}
54 
55 	return f;
56 }
57 
58 const struct got_error *
59 got_diff_blob(struct got_blob_object *blob1, struct got_blob_object *blob2,
60     const char *label1, const char *label2, FILE *outfile)
61 {
62 	struct got_diff_state ds;
63 	struct got_diff_args args;
64 	const struct got_error *err = NULL;
65 	FILE *f1 = NULL, *f2 = NULL;
66 	char hex1[SHA1_DIGEST_STRING_LENGTH];
67 	char hex2[SHA1_DIGEST_STRING_LENGTH];
68 	char *idstr1 = NULL, *idstr2 = NULL;
69 	size_t len, hdrlen;
70 	int res, flags = 0;
71 
72 	if (blob1) {
73 		f1 = opentemp();
74 		if (f1 == NULL)
75 			return got_error(GOT_ERR_FILE_OPEN);
76 	} else
77 		flags |= D_EMPTY1;
78 
79 	if (blob2) {
80 		f2 = opentemp();
81 		if (f2 == NULL) {
82 			fclose(f1);
83 			return got_error(GOT_ERR_FILE_OPEN);
84 		}
85 	} else
86 		flags |= D_EMPTY2;
87 
88 	if (blob1) {
89 		idstr1 = got_object_id_str(&blob1->id, hex1, sizeof(hex1));
90 		hdrlen = blob1->hdrlen;
91 		do {
92 			err = got_object_blob_read_block(blob1, &len);
93 			if (err)
94 				goto done;
95 			/* Skip blob object header first time around. */
96 			fwrite(blob1->zb.outbuf + hdrlen, len - hdrlen, 1, f1);
97 			hdrlen = 0;
98 		} while (len != 0);
99 	} else
100 		idstr1 = "/dev/null";
101 
102 	if (blob2) {
103 		idstr2 = got_object_id_str(&blob2->id, hex2, sizeof(hex2));
104 		hdrlen = blob2->hdrlen;
105 		do {
106 			err = got_object_blob_read_block(blob2, &len);
107 			if (err)
108 				goto done;
109 			/* Skip blob object header first time around. */
110 			fwrite(blob2->zb.outbuf + hdrlen, len - hdrlen, 1, f2);
111 			hdrlen = 0;
112 		} while (len != 0);
113 	} else
114 		idstr2 = "/dev/null";
115 
116 	if (f1)
117 		fflush(f1);
118 	if (f2)
119 		fflush(f2);
120 
121 	memset(&ds, 0, sizeof(ds));
122 	/* XXX should stat buffers be passed in args instead of ds? */
123 	ds.stb1.st_mode = S_IFREG;
124 	if (blob1)
125 		ds.stb1.st_size = blob1->zb.z.total_out;
126 	ds.stb1.st_mtime = 0; /* XXX */
127 
128 	ds.stb2.st_mode = S_IFREG;
129 	if (blob2)
130 		ds.stb2.st_size = blob2->zb.z.total_out;
131 	ds.stb2.st_mtime = 0; /* XXX */
132 
133 	memset(&args, 0, sizeof(args));
134 	args.diff_format = D_UNIFIED;
135 	args.label[0] = label1 ? label1 : idstr1;
136 	args.label[1] = label2 ? label2 : idstr2;
137 
138 	err = got_diffreg(&res, f1, f2, flags, &args, &ds, outfile);
139 done:
140 	if (f1)
141 		fclose(f1);
142 	if (f2)
143 		fclose(f2);
144 	return err;
145 }
146 
147 struct got_tree_entry *
148 match_entry_by_name(struct got_tree_entry *te1, struct got_tree_object *tree2)
149 {
150 	struct got_tree_entry *te2;
151 
152 	SIMPLEQ_FOREACH(te2, &tree2->entries, entry) {
153 		if (strcmp(te1->name, te2->name) == 0)
154 			return te2;
155 	}
156 	return NULL;
157 }
158 
159 static int
160 same_id(struct got_object_id *id1, struct got_object_id *id2)
161 {
162 	return (memcmp(id1->sha1, id2->sha1, SHA1_DIGEST_LENGTH) == 0);
163 }
164 
165 static const struct got_error *
166 diff_added_blob(struct got_object_id *id, struct got_repository *repo)
167 {
168 	const struct got_error *err;
169 	struct got_blob_object  *blob;
170 	struct got_object *obj;
171 
172 	err = got_object_open(&obj, repo, id);
173 	if (err)
174 		return err;
175 	err = got_object_blob_open(&blob, repo, obj, 8192);
176 	if (err != NULL)
177 		return err;
178 
179 	return got_diff_blob(NULL, blob, NULL, NULL, stdout);
180 }
181 
182 static const struct got_error *
183 diff_modified_blob(struct got_object_id *id1, struct got_object_id *id2,
184     struct got_repository *repo)
185 {
186 	const struct got_error *err;
187 	struct got_object *obj1 = NULL;
188 	struct got_object *obj2 = NULL;
189 	struct got_blob_object *blob1 = NULL;
190 	struct got_blob_object *blob2 = NULL;
191 
192 	err = got_object_open(&obj1, repo, id1);
193 	if (err)
194 		return got_error(GOT_ERR_BAD_OBJ_HDR);
195 	if (obj1->type != GOT_OBJ_TYPE_BLOB) {
196 		err = got_error(GOT_ERR_OBJ_TYPE);
197 		goto done;
198 	}
199 
200 	err = got_object_open(&obj2, repo, id2);
201 	if (err) {
202 		err= got_error(GOT_ERR_BAD_OBJ_HDR);
203 		goto done;
204 	}
205 	if (obj2->type != GOT_OBJ_TYPE_BLOB) {
206 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
207 		goto done;
208 	}
209 
210 	err = got_object_blob_open(&blob1, repo, obj1, 8192);
211 	if (err != NULL) {
212 		err = got_error(GOT_ERR_FILE_OPEN);
213 		goto done;
214 	}
215 
216 	err = got_object_blob_open(&blob2, repo, obj2, 8192);
217 	if (err != NULL) {
218 		err = got_error(GOT_ERR_FILE_OPEN);
219 		goto done;
220 	}
221 
222 	err = got_diff_blob(blob1, blob2, NULL, NULL, stdout);
223 
224 done:
225 	if (obj1)
226 		got_object_close(obj1);
227 	if (obj2)
228 		got_object_close(obj2);
229 	if (blob1)
230 		got_object_blob_close(blob1);
231 	if (blob2)
232 		got_object_blob_close(blob2);
233 	return err;
234 }
235 
236 static const struct got_error *
237 diff_deleted_blob(struct got_object_id *id, struct got_repository *repo)
238 {
239 	const struct got_error *err;
240 	struct got_blob_object  *blob;
241 	struct got_object *obj;
242 
243 	err = got_object_open(&obj, repo, id);
244 	if (err)
245 		return err;
246 	err = got_object_blob_open(&blob, repo, obj, 8192);
247 	if (err != NULL)
248 		return err;
249 
250 	return got_diff_blob(blob, NULL, NULL, NULL, stdout);
251 }
252 
253 static const struct got_error *
254 diff_added_tree(struct got_object_id *id, struct got_repository *repo)
255 {
256 	const struct got_error *err = NULL;
257 	struct got_object *treeobj = NULL;
258 	struct got_tree_object *tree = NULL;
259 
260 	err = got_object_open(&treeobj, repo, id);
261 	if (err)
262 		goto done;
263 
264 	if (treeobj->type != GOT_OBJ_TYPE_TREE) {
265 		err = got_error(GOT_ERR_OBJ_TYPE);
266 		goto done;
267 	}
268 
269 	err = got_object_tree_open(&tree, repo, treeobj);
270 	if (err)
271 		goto done;
272 
273 	err = got_diff_tree(NULL, tree, repo);
274 
275 done:
276 	if (tree)
277 		got_object_tree_close(tree);
278 	if (treeobj)
279 		got_object_close(treeobj);
280 	return err;
281 }
282 
283 static const struct got_error *
284 diff_modified_tree(struct got_object_id *id1, struct got_object_id *id2,
285     struct got_repository *repo)
286 {
287 	const struct got_error *err = NULL;
288 	struct got_object *treeobj1 = NULL;
289 	struct got_object *treeobj2 = NULL;
290 	struct got_tree_object *tree1 = NULL;
291 	struct got_tree_object *tree2 = NULL;
292 
293 	err = got_object_open(&treeobj1, repo, id1);
294 	if (err)
295 		goto done;
296 
297 	if (treeobj1->type != GOT_OBJ_TYPE_TREE) {
298 		err = got_error(GOT_ERR_OBJ_TYPE);
299 		goto done;
300 	}
301 
302 	err = got_object_open(&treeobj2, repo, id2);
303 	if (err)
304 		goto done;
305 
306 	if (treeobj2->type != GOT_OBJ_TYPE_TREE) {
307 		err = got_error(GOT_ERR_OBJ_TYPE);
308 		goto done;
309 	}
310 
311 	err = got_object_tree_open(&tree1, repo, treeobj1);
312 	if (err)
313 		goto done;
314 
315 	err = got_object_tree_open(&tree2, repo, treeobj2);
316 	if (err)
317 		goto done;
318 
319 	err = got_diff_tree(tree1, tree2, repo);
320 
321 done:
322 	if (tree1)
323 		got_object_tree_close(tree1);
324 	if (tree2)
325 		got_object_tree_close(tree2);
326 	if (treeobj1)
327 		got_object_close(treeobj1);
328 	if (treeobj2)
329 		got_object_close(treeobj2);
330 	return err;
331 }
332 
333 static const struct got_error *
334 diff_deleted_tree(struct got_object_id *id, struct got_repository *repo)
335 {
336 	const struct got_error *err = NULL;
337 	struct got_object *treeobj = NULL;
338 	struct got_tree_object *tree = NULL;
339 
340 	err = got_object_open(&treeobj, repo, id);
341 	if (err)
342 		goto done;
343 
344 	if (treeobj->type != GOT_OBJ_TYPE_TREE) {
345 		err = got_error(GOT_ERR_OBJ_TYPE);
346 		goto done;
347 	}
348 
349 	err = got_object_tree_open(&tree, repo, treeobj);
350 	if (err)
351 		goto done;
352 
353 	err = got_diff_tree(tree, NULL, repo);
354 
355 done:
356 	if (tree)
357 		got_object_tree_close(tree);
358 	if (treeobj)
359 		got_object_close(treeobj);
360 	return err;
361 }
362 
363 static const struct got_error *
364 diff_kind_mismatch(struct got_object_id *id1, struct got_object_id *id2)
365 {
366 	/* XXX TODO */
367 	return NULL;
368 }
369 
370 static const struct got_error *
371 diff_entry_old_new(struct got_tree_entry *te1, struct got_tree_object *tree2,
372     struct got_repository *repo)
373 {
374 	const struct got_error *err;
375 	struct got_tree_entry *te2;
376 	char hex[SHA1_DIGEST_STRING_LENGTH];
377 
378 	te2 = match_entry_by_name(te1, tree2);
379 	if (te2 == NULL) {
380 		if (S_ISDIR(te1->mode))
381 			return diff_deleted_tree(&te1->id, repo);
382 		return diff_deleted_blob(&te1->id, repo);
383 	}
384 
385 	if (S_ISDIR(te1->mode) && S_ISDIR(te2->mode)) {
386 		if (!same_id(&te1->id, &te2->id))
387 			return diff_modified_tree(&te1->id, &te2->id, repo);
388 	} else if (S_ISREG(te1->mode) && S_ISREG(te2->mode)) {
389 		if (!same_id(&te1->id, &te2->id))
390 			return diff_modified_blob(&te1->id, &te2->id, repo);
391 
392 	}
393 
394 	return diff_kind_mismatch(&te1->id, &te2->id);
395 }
396 
397 static const struct got_error *
398 diff_entry_new_old(struct got_tree_entry *te2, struct got_tree_object *tree1,
399     struct got_repository *repo)
400 {
401 	const struct got_error *err;
402 	struct got_tree_entry *te1;
403 
404 	te1 = match_entry_by_name(te2, tree1);
405 	if (te1 != NULL) /* handled by diff_entry_old_new() */
406 		return NULL;
407 
408 	if (S_ISDIR(te2->mode))
409 		return diff_added_tree(&te2->id, repo);
410 	return diff_added_blob(&te2->id, repo);
411 }
412 
413 const struct got_error *
414 got_diff_tree(struct got_tree_object *tree1, struct got_tree_object *tree2,
415     struct got_repository *repo)
416 {
417 	const struct got_error *err = NULL;
418 	struct got_tree_entry *te1 = NULL;
419 	struct got_tree_entry *te2 = NULL;
420 
421 	if (tree1)
422 		te1 = SIMPLEQ_FIRST(&tree1->entries);
423 	if (tree2)
424 		te2 = SIMPLEQ_FIRST(&tree2->entries);
425 
426 	do {
427 		if (te1) {
428 			err = diff_entry_old_new(te1, tree2, repo);
429 			if (err)
430 				break;
431 		}
432 
433 		if (te2) {
434 			err = diff_entry_new_old(te2, tree1, repo);
435 			if (err)
436 				break;
437 		}
438 
439 		if (te1)
440 			te1 = SIMPLEQ_NEXT(te1, entry);
441 		if (te2)
442 			te2 = SIMPLEQ_NEXT(te2, entry);
443 	} while (te1 || te2);
444 
445 	return err;
446 }