Gotweb

Blob

Date:: Tue Mar 27 21:08:07 2018 UTC
Message:: be careful about NULL trees when diffing entries
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 "got_lib_diff.h"
33 #include "got_lib_path.h"
34 
35 const struct got_error *
36 got_diff_blob(struct got_blob_object *blob1, struct got_blob_object *blob2,
37     const char *label1, const char *label2, FILE *outfile)
38 {
39 	struct got_diff_state ds;
40 	struct got_diff_args args;
41 	const struct got_error *err = NULL;
42 	FILE *f1 = NULL, *f2 = NULL;
43 	char hex1[SHA1_DIGEST_STRING_LENGTH];
44 	char hex2[SHA1_DIGEST_STRING_LENGTH];
45 	char *idstr1 = NULL, *idstr2 = NULL;
46 	size_t len, hdrlen;
47 	size_t size1, size2;
48 	int res, flags = 0;
49 
50 	if (blob1) {
51 		f1 = got_opentemp();
52 		if (f1 == NULL)
53 			return got_error(GOT_ERR_FILE_OPEN);
54 	} else
55 		flags |= D_EMPTY1;
56 
57 	if (blob2) {
58 		f2 = got_opentemp();
59 		if (f2 == NULL) {
60 			fclose(f1);
61 			return got_error(GOT_ERR_FILE_OPEN);
62 		}
63 	} else
64 		flags |= D_EMPTY2;
65 
66 	size1 = 0;
67 	if (blob1) {
68 		idstr1 = got_object_blob_id_str(blob1, hex1, sizeof(hex1));
69 		hdrlen = got_object_blob_get_hdrlen(blob1);
70 		do {
71 			err = got_object_blob_read_block(&len, blob1);
72 			if (err)
73 				goto done;
74 			if (len == 0)
75 				break;
76 			size1 += len;
77 			/* Skip blob object header first time around. */
78 			fwrite(got_object_blob_get_read_buf(blob1) + hdrlen, len - hdrlen, 1, f1);
79 			hdrlen = 0;
80 		} while (len != 0);
81 	} else
82 		idstr1 = "/dev/null";
83 
84 	size2 = 0;
85 	if (blob2) {
86 		idstr2 = got_object_blob_id_str(blob2, hex2, sizeof(hex2));
87 		hdrlen = got_object_blob_get_hdrlen(blob2);
88 		do {
89 			err = got_object_blob_read_block(&len, blob2);
90 			if (err)
91 				goto done;
92 			if (len == 0)
93 				break;
94 			size2 += len;
95 			/* Skip blob object header first time around. */
96 			fwrite(got_object_blob_get_read_buf(blob2) + hdrlen, len - hdrlen, 1, f2);
97 			hdrlen = 0;
98 		} while (len != 0);
99 	} else
100 		idstr2 = "/dev/null";
101 
102 	if (f1) {
103 		fflush(f1);
104 		rewind(f1);
105 	}
106 	if (f2) {
107 		fflush(f2);
108 		rewind(f2);
109 	}
110 
111 	memset(&ds, 0, sizeof(ds));
112 	/* XXX should stat buffers be passed in args instead of ds? */
113 	ds.stb1.st_mode = S_IFREG;
114 	if (blob1)
115 		ds.stb1.st_size = size1;
116 	ds.stb1.st_mtime = 0; /* XXX */
117 
118 	ds.stb2.st_mode = S_IFREG;
119 	if (blob2)
120 		ds.stb2.st_size = size2;
121 	ds.stb2.st_mtime = 0; /* XXX */
122 
123 	memset(&args, 0, sizeof(args));
124 	args.diff_format = D_UNIFIED;
125 	args.label[0] = label1 ? label1 : idstr1;
126 	args.label[1] = label2 ? label2 : idstr2;
127 	args.diff_context = 3;
128 
129 	err = got_diffreg(&res, f1, f2, flags, &args, &ds, outfile);
130 done:
131 	if (f1)
132 		fclose(f1);
133 	if (f2)
134 		fclose(f2);
135 	return err;
136 }
137 
138 struct got_tree_entry *
139 match_entry_by_name(struct got_tree_entry *te1, struct got_tree_object *tree2)
140 {
141 	struct got_tree_entry *te2;
142 
143 	SIMPLEQ_FOREACH(te2, &tree2->entries, entry) {
144 		if (strcmp(te1->name, te2->name) == 0)
145 			return te2;
146 	}
147 	return NULL;
148 }
149 
150 static const struct got_error *
151 diff_added_blob(struct got_object_id *id, struct got_repository *repo,
152     FILE *outfile)
153 {
154 	const struct got_error *err;
155 	struct got_blob_object  *blob;
156 	struct got_object *obj;
157 
158 	err = got_object_open(&obj, repo, id);
159 	if (err)
160 		return err;
161 	err = got_object_blob_open(&blob, repo, obj, 8192);
162 	if (err != NULL)
163 		return err;
164 
165 	return got_diff_blob(NULL, blob, NULL, NULL, outfile);
166 }
167 
168 static const struct got_error *
169 diff_modified_blob(struct got_object_id *id1, struct got_object_id *id2,
170     struct got_repository *repo, FILE *outfile)
171 {
172 	const struct got_error *err;
173 	struct got_object *obj1 = NULL;
174 	struct got_object *obj2 = NULL;
175 	struct got_blob_object *blob1 = NULL;
176 	struct got_blob_object *blob2 = NULL;
177 
178 	err = got_object_open(&obj1, repo, id1);
179 	if (err)
180 		return got_error(GOT_ERR_BAD_OBJ_HDR);
181 	if (got_object_get_type(obj1) != GOT_OBJ_TYPE_BLOB) {
182 		err = got_error(GOT_ERR_OBJ_TYPE);
183 		goto done;
184 	}
185 
186 	err = got_object_open(&obj2, repo, id2);
187 	if (err) {
188 		err= got_error(GOT_ERR_BAD_OBJ_HDR);
189 		goto done;
190 	}
191 	if (got_object_get_type(obj2) != GOT_OBJ_TYPE_BLOB) {
192 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
193 		goto done;
194 	}
195 
196 	err = got_object_blob_open(&blob1, repo, obj1, 8192);
197 	if (err != NULL) {
198 		err = got_error(GOT_ERR_FILE_OPEN);
199 		goto done;
200 	}
201 
202 	err = got_object_blob_open(&blob2, repo, obj2, 8192);
203 	if (err != NULL) {
204 		err = got_error(GOT_ERR_FILE_OPEN);
205 		goto done;
206 	}
207 
208 	err = got_diff_blob(blob1, blob2, NULL, NULL, outfile);
209 
210 done:
211 	if (obj1)
212 		got_object_close(obj1);
213 	if (obj2)
214 		got_object_close(obj2);
215 	if (blob1)
216 		got_object_blob_close(blob1);
217 	if (blob2)
218 		got_object_blob_close(blob2);
219 	return err;
220 }
221 
222 static const struct got_error *
223 diff_deleted_blob(struct got_object_id *id, struct got_repository *repo,
224     FILE *outfile)
225 {
226 	const struct got_error *err;
227 	struct got_blob_object  *blob;
228 	struct got_object *obj;
229 
230 	err = got_object_open(&obj, repo, id);
231 	if (err)
232 		return err;
233 	err = got_object_blob_open(&blob, repo, obj, 8192);
234 	if (err != NULL)
235 		return err;
236 
237 	return got_diff_blob(blob, NULL, NULL, NULL, outfile);
238 }
239 
240 static const struct got_error *
241 diff_added_tree(struct got_object_id *id, struct got_repository *repo,
242     FILE *outfile)
243 {
244 	const struct got_error *err = NULL;
245 	struct got_object *treeobj = NULL;
246 	struct got_tree_object *tree = NULL;
247 
248 	err = got_object_open(&treeobj, repo, id);
249 	if (err)
250 		goto done;
251 
252 	if (got_object_get_type(treeobj) != GOT_OBJ_TYPE_TREE) {
253 		err = got_error(GOT_ERR_OBJ_TYPE);
254 		goto done;
255 	}
256 
257 	err = got_object_tree_open(&tree, repo, treeobj);
258 	if (err)
259 		goto done;
260 
261 	err = got_diff_tree(NULL, tree, repo, outfile);
262 
263 done:
264 	if (tree)
265 		got_object_tree_close(tree);
266 	if (treeobj)
267 		got_object_close(treeobj);
268 	return err;
269 }
270 
271 static const struct got_error *
272 diff_modified_tree(struct got_object_id *id1, struct got_object_id *id2,
273     struct got_repository *repo, FILE *outfile)
274 {
275 	const struct got_error *err = NULL;
276 	struct got_object *treeobj1 = NULL;
277 	struct got_object *treeobj2 = NULL;
278 	struct got_tree_object *tree1 = NULL;
279 	struct got_tree_object *tree2 = NULL;
280 
281 	err = got_object_open(&treeobj1, repo, id1);
282 	if (err)
283 		goto done;
284 
285 	if (got_object_get_type(treeobj1) != GOT_OBJ_TYPE_TREE) {
286 		err = got_error(GOT_ERR_OBJ_TYPE);
287 		goto done;
288 	}
289 
290 	err = got_object_open(&treeobj2, repo, id2);
291 	if (err)
292 		goto done;
293 
294 	if (got_object_get_type(treeobj2) != GOT_OBJ_TYPE_TREE) {
295 		err = got_error(GOT_ERR_OBJ_TYPE);
296 		goto done;
297 	}
298 
299 	err = got_object_tree_open(&tree1, repo, treeobj1);
300 	if (err)
301 		goto done;
302 
303 	err = got_object_tree_open(&tree2, repo, treeobj2);
304 	if (err)
305 		goto done;
306 
307 	err = got_diff_tree(tree1, tree2, repo, outfile);
308 
309 done:
310 	if (tree1)
311 		got_object_tree_close(tree1);
312 	if (tree2)
313 		got_object_tree_close(tree2);
314 	if (treeobj1)
315 		got_object_close(treeobj1);
316 	if (treeobj2)
317 		got_object_close(treeobj2);
318 	return err;
319 }
320 
321 static const struct got_error *
322 diff_deleted_tree(struct got_object_id *id, struct got_repository *repo, FILE *outfile)
323 {
324 	const struct got_error *err = NULL;
325 	struct got_object *treeobj = NULL;
326 	struct got_tree_object *tree = NULL;
327 
328 	err = got_object_open(&treeobj, repo, id);
329 	if (err)
330 		goto done;
331 
332 	if (got_object_get_type(treeobj) != GOT_OBJ_TYPE_TREE) {
333 		err = got_error(GOT_ERR_OBJ_TYPE);
334 		goto done;
335 	}
336 
337 	err = got_object_tree_open(&tree, repo, treeobj);
338 	if (err)
339 		goto done;
340 
341 	err = got_diff_tree(tree, NULL, repo, outfile);
342 
343 done:
344 	if (tree)
345 		got_object_tree_close(tree);
346 	if (treeobj)
347 		got_object_close(treeobj);
348 	return err;
349 }
350 
351 static const struct got_error *
352 diff_kind_mismatch(struct got_object_id *id1, struct got_object_id *id2,
353     FILE *outfile)
354 {
355 	/* XXX TODO */
356 	return NULL;
357 }
358 
359 static const struct got_error *
360 diff_entry_old_new(struct got_tree_entry *te1, struct got_tree_object *tree2,
361     struct got_repository *repo, FILE *outfile)
362 {
363 	struct got_tree_entry *te2 = NULL;
364 
365 	if (tree2)
366 		te2 = match_entry_by_name(te1, tree2);
367 	if (te2 == NULL) {
368 		if (S_ISDIR(te1->mode))
369 			return diff_deleted_tree(te1->id, repo, outfile);
370 		return diff_deleted_blob(te1->id, repo, outfile);
371 	}
372 
373 	if (S_ISDIR(te1->mode) && S_ISDIR(te2->mode)) {
374 		if (got_object_id_cmp(te1->id, te2->id) != 0)
375 			return diff_modified_tree(te1->id, te2->id, repo,
376 			    outfile);
377 	} else if (S_ISREG(te1->mode) && S_ISREG(te2->mode)) {
378 		if (got_object_id_cmp(te1->id, te2->id) != 0)
379 			return diff_modified_blob(te1->id, te2->id, repo,
380 			    outfile);
381 	}
382 
383 	return diff_kind_mismatch(te1->id, te2->id, outfile);
384 }
385 
386 static const struct got_error *
387 diff_entry_new_old(struct got_tree_entry *te2, struct got_tree_object *tree1,
388     struct got_repository *repo, FILE *outfile)
389 {
390 	if (tree1) {
391 		struct got_tree_entry *te1 = match_entry_by_name(te2, tree1);
392 		if (te1 != NULL) /* handled by diff_entry_old_new() */
393 			return NULL;
394 	}
395 
396 	if (S_ISDIR(te2->mode))
397 		return diff_added_tree(te2->id, repo, outfile);
398 	return diff_added_blob(te2->id, repo, outfile);
399 }
400 
401 const struct got_error *
402 got_diff_tree(struct got_tree_object *tree1, struct got_tree_object *tree2,
403     struct got_repository *repo, FILE *outfile)
404 {
405 	const struct got_error *err = NULL;
406 	struct got_tree_entry *te1 = NULL;
407 	struct got_tree_entry *te2 = NULL;
408 
409 	if (tree1)
410 		te1 = SIMPLEQ_FIRST(&tree1->entries);
411 	if (tree2)
412 		te2 = SIMPLEQ_FIRST(&tree2->entries);
413 
414 	do {
415 		if (te1) {
416 			err = diff_entry_old_new(te1, tree2, repo, outfile);
417 			if (err)
418 				break;
419 		}
420 
421 		if (te2) {
422 			err = diff_entry_new_old(te2, tree1, repo, outfile);
423 			if (err)
424 				break;
425 		}
426 
427 		if (te1)
428 			te1 = SIMPLEQ_NEXT(te1, entry);
429 		if (te2)
430 			te2 = SIMPLEQ_NEXT(te2, entry);
431 	} while (te1 || te2);
432 
433 	return err;
434 }