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 
31 #include "diff.h"
32 
33 static FILE *
34 opentemp(void)
35 {
36 	char name[PATH_MAX];
37 	int fd;
38 	FILE *f;
39 
40 	if (strlcpy(name, "/tmp/got.XXXXXXXX", sizeof(name)) >= sizeof(name))
41 		return NULL;
42 
43 	fd = mkstemp(name);
44 	if (fd < 0)
45 		return NULL;
46 
47 	unlink(name);
48 	f = fdopen(fd, "w+");
49 	if (f == NULL) {
50 		close(fd);
51 		return NULL;
52 	}
53 
54 	return f;
55 }
56 
57 const struct got_error *
58 got_diff_blob(struct got_blob_object *blob1, struct got_blob_object *blob2,
59     const char *label1, const char *label2, FILE *outfile)
60 {
61 	struct got_diff_state ds;
62 	struct got_diff_args args;
63 	const struct got_error *err = NULL;
64 	FILE *f1 = NULL, *f2 = NULL;
65 	char hex1[SHA1_DIGEST_STRING_LENGTH];
66 	char hex2[SHA1_DIGEST_STRING_LENGTH];
67 	size_t len, hdrlen;
68 	int res, flags = 0;
69 
70 	if (blob1) {
71 		f1 = opentemp();
72 		if (f1 == NULL)
73 			return got_error(GOT_ERR_FILE_OPEN);
74 	} else
75 		flags |= D_EMPTY1;
76 
77 	if (blob2) {
78 		f2 = opentemp();
79 		if (f2 == NULL) {
80 			fclose(f1);
81 			return got_error(GOT_ERR_FILE_OPEN);
82 		}
83 	} else
84 		flags |= D_EMPTY2;
85 
86 	if (blob1 == NULL) {
87 		f1 = NULL;
88 	} else {
89 		hdrlen = blob1->hdrlen;
90 		do {
91 			err = got_object_blob_read_block(blob1, &len);
92 			if (err)
93 				goto done;
94 			/* Skip blob object header first time around. */
95 			fwrite(blob1->zb.outbuf + hdrlen, len - hdrlen, 1, f1);
96 			hdrlen = 0;
97 		} while (len != 0);
98 	}
99 
100 	hdrlen = blob2->hdrlen;
101 	do {
102 		err = got_object_blob_read_block(blob2, &len);
103 		if (err)
104 			goto done;
105 		/* Skip blob object header first time around. */
106 		fwrite(blob2->zb.outbuf + hdrlen, len - hdrlen, 1, f2);
107 		hdrlen = 0;
108 	} while (len != 0);
109 
110 	fflush(f1);
111 	fflush(f2);
112 	/* rewind(f1); */
113 	/* rewind(f2);*/
114 
115 	memset(&ds, 0, sizeof(ds));
116 	/* XXX should stat buffers be passed in args instead of ds? */
117 	ds.stb1.st_mode = S_IFREG;
118 	ds.stb1.st_size = blob1->zb.z.total_out;
119 	ds.stb1.st_mtime = 0; /* XXX */
120 
121 	ds.stb2.st_mode = S_IFREG;
122 	ds.stb2.st_size = blob2->zb.z.total_out;
123 	ds.stb2.st_mtime = 0; /* XXX */
124 
125 	memset(&args, 0, sizeof(args));
126 	args.diff_format = D_UNIFIED;
127 	args.label[0] = label1 ?
128 	    label1 : got_object_id_str(&blob1->id, hex1, sizeof(hex1));
129 	args.label[1] = label2 ?
130 	    label2 : got_object_id_str(&blob2->id, hex2, sizeof(hex2));
131 
132 	err = got_diffreg(&res, f1, f2, flags, &args, &ds, outfile);
133 done:
134 	fclose(f1);
135 	fclose(f2);
136 	return err;
137 }
138 
139 static const struct got_error *
140 match_entry_by_name(struct got_tree_entry **te, struct got_tree_entry *te1,
141     struct got_tree_object *tree2)
142 {
143 	*te = NULL;
144 	return NULL;
145 }
146 
147 static int
148 same_id(struct got_object_id *id1, struct got_object_id *id2)
149 {
150 	return (memcmp(id1->sha1, id2->sha1, SHA1_DIGEST_LENGTH) == 0);
151 }
152 
153 static const struct got_error *
154 diff_added_blob(struct got_object_id *id, struct got_repository *repo)
155 {
156 	const struct got_error *err;
157 	struct got_blob_object  *blob;
158 	struct got_object *obj;
159 
160 	err = got_object_open(&obj, repo, id);
161 	if (err)
162 		return err;
163 	err = got_object_blob_open(&blob, repo, obj, 512);
164 	if (err != NULL)
165 		return err;
166 
167 	return got_diff_blob(NULL, blob, NULL, NULL, stdout);
168 }
169 
170 static const struct got_error *
171 diff_modified_blob(struct got_object_id *id1, struct got_object_id *id2,
172     struct got_repository *repo)
173 {
174 	const struct got_error *err;
175 	struct got_object *obj1 = NULL;
176 	struct got_object *obj2 = NULL;
177 	struct got_blob_object *blob1 = NULL;
178 	struct got_blob_object *blob2 = NULL;
179 
180 	err = got_object_open(&obj1, repo, id1);
181 	if (err)
182 		return got_error(GOT_ERR_BAD_OBJ_HDR);
183 	if (obj1->type != GOT_OBJ_TYPE_BLOB) {
184 		err = got_error(GOT_ERR_OBJ_TYPE);
185 		goto done;
186 	}
187 
188 	err = got_object_open(&obj2, repo, id2);
189 	if (err) {
190 		err= got_error(GOT_ERR_BAD_OBJ_HDR);
191 		goto done;
192 	}
193 	if (obj2->type != GOT_OBJ_TYPE_BLOB) {
194 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
195 		goto done;
196 	}
197 
198 	err = got_object_blob_open(&blob1, repo, obj1, 512);
199 	if (err != NULL) {
200 		err = got_error(GOT_ERR_FILE_OPEN);
201 		goto done;
202 	}
203 
204 	err = got_object_blob_open(&blob2, repo, obj2, 512);
205 	if (err != NULL) {
206 		err = got_error(GOT_ERR_FILE_OPEN);
207 		goto done;
208 	}
209 
210 	err = got_diff_blob(blob1, blob2, NULL, NULL, stdout);
211 
212 done:
213 	got_object_close(obj1);
214 	got_object_close(obj2);
215 	got_object_blob_close(blob1);
216 	got_object_blob_close(blob2);
217 	return err;
218 }
219 
220 static const struct got_error *
221 diff_deleted_blob(struct got_object_id *id, struct got_repository *repo)
222 {
223 	const struct got_error *err;
224 	struct got_blob_object  *blob;
225 	struct got_object *obj;
226 
227 	err = got_object_open(&obj, repo, id);
228 	if (err)
229 		return err;
230 	err = got_object_blob_open(&blob, repo, obj, 512);
231 	if (err != NULL)
232 		return err;
233 
234 	return got_diff_blob(blob, NULL, NULL, NULL, stdout);
235 }
236 
237 static const struct got_error *
238 diff_added_tree(struct got_object_id *id)
239 {
240 	return NULL;
241 }
242 
243 static const struct got_error *
244 diff_modified_tree(struct got_object_id *id1, struct got_object_id *id2)
245 {
246 	return NULL;
247 }
248 
249 static const struct got_error *
250 diff_deleted_tree(struct got_object_id *id)
251 {
252 	return NULL;
253 }
254 
255 static const struct got_error *
256 diff_kind_mismatch(struct got_object_id *id1, struct got_object_id *id2)
257 {
258 	return NULL;
259 }
260 
261 static const struct got_error *
262 diff_entry_old_new(struct got_tree_entry *te1, struct got_tree_object *tree2,
263     struct got_repository *repo)
264 {
265 	const struct got_error *err;
266 	struct got_tree_entry *te2;
267 
268 	err = match_entry_by_name(&te2, te1, tree2);
269 	if (err)
270 		return err;
271 	if (te2 == NULL) {
272 		if (S_ISDIR(te1->mode))
273 			return diff_deleted_tree(&te1->id);
274 		return diff_deleted_blob(&te1->id, repo);
275 	}
276 
277 	if (S_ISDIR(te1->mode) && S_ISDIR(te2->mode)) {
278 		if (!same_id(&te1->id, &te2->id))
279 			return diff_modified_tree(&te1->id, &te2->id);
280 	} else if (S_ISREG(te1->mode) && S_ISREG(te2->mode)) {
281 		if (!same_id(&te1->id, &te2->id))
282 			return diff_modified_blob(&te1->id, &te2->id, repo);
283 	}
284 
285 	return diff_kind_mismatch(&te1->id, &te2->id);
286 }
287 
288 static const struct got_error *
289 diff_entry_new_old(struct got_tree_entry *te2, struct got_tree_object *tree1,
290     struct got_repository *repo)
291 {
292 	const struct got_error *err;
293 	struct got_tree_entry *te1;
294 
295 	err = match_entry_by_name(&te1, te2, tree1);
296 	if (err)
297 		return err;
298 	if (te1 != NULL) /* handled by diff_entry_old_new() */
299 		return NULL;
300 
301 	if (S_ISDIR(te2->mode))
302 		return diff_added_tree(&te2->id);
303 	return diff_added_blob(&te2->id, repo);
304 }
305 
306 const struct got_error *
307 got_diff_tree(struct got_tree_object *tree1, struct got_tree_object *tree2,
308     struct got_repository *repo)
309 {
310 	const struct got_error *err = NULL;
311 	struct got_tree_entry *te1;
312 	struct got_tree_entry *te2;
313 
314 	if (tree1->nentries == 0 && tree2->nentries == 0)
315 		return NULL;
316 
317 	te1 = SIMPLEQ_FIRST(&tree1->entries);
318 	te2 = SIMPLEQ_FIRST(&tree2->entries);
319 
320 	do {
321 		if (te1) {
322 			err = diff_entry_old_new(te1, tree2, repo);
323 			if (err)
324 				break;
325 		}
326 
327 		if (te2) {
328 			err = diff_entry_new_old(te2, tree1, repo);
329 			if (err)
330 				break;
331 		}
332 
333 		if (te1)
334 			te1 = SIMPLEQ_NEXT(te1, entry);
335 		if (te2)
336 			te2 = SIMPLEQ_NEXT(te2, entry);
337 	} while (te1 || te2);
338 
339 	return err;
340 }