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_diff_priv.h"
33 #include "got_path_priv.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 
128 	err = got_diffreg(&res, f1, f2, flags, &args, &ds, outfile);
129 done:
130 	if (f1)
131 		fclose(f1);
132 	if (f2)
133 		fclose(f2);
134 	return err;
135 }
136 
137 struct got_tree_entry *
138 match_entry_by_name(struct got_tree_entry *te1, struct got_tree_object *tree2)
139 {
140 	struct got_tree_entry *te2;
141 
142 	SIMPLEQ_FOREACH(te2, &tree2->entries, entry) {
143 		if (strcmp(te1->name, te2->name) == 0)
144 			return te2;
145 	}
146 	return NULL;
147 }
148 
149 static const struct got_error *
150 diff_added_blob(struct got_object_id *id, struct got_repository *repo,
151     FILE *outfile)
152 {
153 	const struct got_error *err;
154 	struct got_blob_object  *blob;
155 	struct got_object *obj;
156 
157 	err = got_object_open(&obj, repo, id);
158 	if (err)
159 		return err;
160 	err = got_object_blob_open(&blob, repo, obj, 8192);
161 	if (err != NULL)
162 		return err;
163 
164 	return got_diff_blob(NULL, blob, NULL, NULL, outfile);
165 }
166 
167 static const struct got_error *
168 diff_modified_blob(struct got_object_id *id1, struct got_object_id *id2,
169     struct got_repository *repo, FILE *outfile)
170 {
171 	const struct got_error *err;
172 	struct got_object *obj1 = NULL;
173 	struct got_object *obj2 = NULL;
174 	struct got_blob_object *blob1 = NULL;
175 	struct got_blob_object *blob2 = NULL;
176 
177 	err = got_object_open(&obj1, repo, id1);
178 	if (err)
179 		return got_error(GOT_ERR_BAD_OBJ_HDR);
180 	if (got_object_get_type(obj1) != GOT_OBJ_TYPE_BLOB) {
181 		err = got_error(GOT_ERR_OBJ_TYPE);
182 		goto done;
183 	}
184 
185 	err = got_object_open(&obj2, repo, id2);
186 	if (err) {
187 		err= got_error(GOT_ERR_BAD_OBJ_HDR);
188 		goto done;
189 	}
190 	if (got_object_get_type(obj2) != GOT_OBJ_TYPE_BLOB) {
191 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
192 		goto done;
193 	}
194 
195 	err = got_object_blob_open(&blob1, repo, obj1, 8192);
196 	if (err != NULL) {
197 		err = got_error(GOT_ERR_FILE_OPEN);
198 		goto done;
199 	}
200 
201 	err = got_object_blob_open(&blob2, repo, obj2, 8192);
202 	if (err != NULL) {
203 		err = got_error(GOT_ERR_FILE_OPEN);
204 		goto done;
205 	}
206 
207 	err = got_diff_blob(blob1, blob2, NULL, NULL, outfile);
208 
209 done:
210 	if (obj1)
211 		got_object_close(obj1);
212 	if (obj2)
213 		got_object_close(obj2);
214 	if (blob1)
215 		got_object_blob_close(blob1);
216 	if (blob2)
217 		got_object_blob_close(blob2);
218 	return err;
219 }
220 
221 static const struct got_error *
222 diff_deleted_blob(struct got_object_id *id, struct got_repository *repo,
223     FILE *outfile)
224 {
225 	const struct got_error *err;
226 	struct got_blob_object  *blob;
227 	struct got_object *obj;
228 
229 	err = got_object_open(&obj, repo, id);
230 	if (err)
231 		return err;
232 	err = got_object_blob_open(&blob, repo, obj, 8192);
233 	if (err != NULL)
234 		return err;
235 
236 	return got_diff_blob(blob, NULL, NULL, NULL, outfile);
237 }
238 
239 static const struct got_error *
240 diff_added_tree(struct got_object_id *id, struct got_repository *repo,
241     FILE *outfile)
242 {
243 	const struct got_error *err = NULL;
244 	struct got_object *treeobj = NULL;
245 	struct got_tree_object *tree = NULL;
246 
247 	err = got_object_open(&treeobj, repo, id);
248 	if (err)
249 		goto done;
250 
251 	if (got_object_get_type(treeobj) != GOT_OBJ_TYPE_TREE) {
252 		err = got_error(GOT_ERR_OBJ_TYPE);
253 		goto done;
254 	}
255 
256 	err = got_object_tree_open(&tree, repo, treeobj);
257 	if (err)
258 		goto done;
259 
260 	err = got_diff_tree(NULL, tree, repo, outfile);
261 
262 done:
263 	if (tree)
264 		got_object_tree_close(tree);
265 	if (treeobj)
266 		got_object_close(treeobj);
267 	return err;
268 }
269 
270 static const struct got_error *
271 diff_modified_tree(struct got_object_id *id1, struct got_object_id *id2,
272     struct got_repository *repo, FILE *outfile)
273 {
274 	const struct got_error *err = NULL;
275 	struct got_object *treeobj1 = NULL;
276 	struct got_object *treeobj2 = NULL;
277 	struct got_tree_object *tree1 = NULL;
278 	struct got_tree_object *tree2 = NULL;
279 
280 	err = got_object_open(&treeobj1, repo, id1);
281 	if (err)
282 		goto done;
283 
284 	if (got_object_get_type(treeobj1) != GOT_OBJ_TYPE_TREE) {
285 		err = got_error(GOT_ERR_OBJ_TYPE);
286 		goto done;
287 	}
288 
289 	err = got_object_open(&treeobj2, repo, id2);
290 	if (err)
291 		goto done;
292 
293 	if (got_object_get_type(treeobj2) != GOT_OBJ_TYPE_TREE) {
294 		err = got_error(GOT_ERR_OBJ_TYPE);
295 		goto done;
296 	}
297 
298 	err = got_object_tree_open(&tree1, repo, treeobj1);
299 	if (err)
300 		goto done;
301 
302 	err = got_object_tree_open(&tree2, repo, treeobj2);
303 	if (err)
304 		goto done;
305 
306 	err = got_diff_tree(tree1, tree2, repo, outfile);
307 
308 done:
309 	if (tree1)
310 		got_object_tree_close(tree1);
311 	if (tree2)
312 		got_object_tree_close(tree2);
313 	if (treeobj1)
314 		got_object_close(treeobj1);
315 	if (treeobj2)
316 		got_object_close(treeobj2);
317 	return err;
318 }
319 
320 static const struct got_error *
321 diff_deleted_tree(struct got_object_id *id, struct got_repository *repo, FILE *outfile)
322 {
323 	const struct got_error *err = NULL;
324 	struct got_object *treeobj = NULL;
325 	struct got_tree_object *tree = NULL;
326 
327 	err = got_object_open(&treeobj, repo, id);
328 	if (err)
329 		goto done;
330 
331 	if (got_object_get_type(treeobj) != GOT_OBJ_TYPE_TREE) {
332 		err = got_error(GOT_ERR_OBJ_TYPE);
333 		goto done;
334 	}
335 
336 	err = got_object_tree_open(&tree, repo, treeobj);
337 	if (err)
338 		goto done;
339 
340 	err = got_diff_tree(tree, NULL, repo, outfile);
341 
342 done:
343 	if (tree)
344 		got_object_tree_close(tree);
345 	if (treeobj)
346 		got_object_close(treeobj);
347 	return err;
348 }
349 
350 static const struct got_error *
351 diff_kind_mismatch(struct got_object_id *id1, struct got_object_id *id2,
352     FILE *outfile)
353 {
354 	/* XXX TODO */
355 	return NULL;
356 }
357 
358 static const struct got_error *
359 diff_entry_old_new(struct got_tree_entry *te1, struct got_tree_object *tree2,
360     struct got_repository *repo, FILE *outfile)
361 {
362 	const struct got_error *err;
363 	struct got_tree_entry *te2;
364 	char hex[SHA1_DIGEST_STRING_LENGTH];
365 
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 	const struct got_error *err;
391 	struct got_tree_entry *te1;
392 
393 	te1 = match_entry_by_name(te2, tree1);
394 	if (te1 != NULL) /* handled by diff_entry_old_new() */
395 		return NULL;
396 
397 	if (S_ISDIR(te2->mode))
398 		return diff_added_tree(te2->id, repo, outfile);
399 	return diff_added_blob(te2->id, repo, outfile);
400 }
401 
402 const struct got_error *
403 got_diff_tree(struct got_tree_object *tree1, struct got_tree_object *tree2,
404     struct got_repository *repo, FILE *outfile)
405 {
406 	const struct got_error *err = NULL;
407 	struct got_tree_entry *te1 = NULL;
408 	struct got_tree_entry *te2 = NULL;
409 
410 	if (tree1)
411 		te1 = SIMPLEQ_FIRST(&tree1->entries);
412 	if (tree2)
413 		te2 = SIMPLEQ_FIRST(&tree2->entries);
414 
415 	do {
416 		if (te1) {
417 			err = diff_entry_old_new(te1, tree2, repo, outfile);
418 			if (err)
419 				break;
420 		}
421 
422 		if (te2) {
423 			err = diff_entry_new_old(te2, tree1, repo, outfile);
424 			if (err)
425 				break;
426 		}
427 
428 		if (te1)
429 			te1 = SIMPLEQ_NEXT(te1, entry);
430 		if (te2)
431 			te2 = SIMPLEQ_NEXT(te2, entry);
432 	} while (te1 || te2);
433 
434 	return err;
435 }