1 /*
2  * Copyright (c) 2018 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 <sha1.h>
21 #include <string.h>
22 #include <stdio.h>
23 #include <stdlib.h>
24 #include <time.h>
25 #include <util.h>
26 #include <zlib.h>
27 
28 #include "got_error.h"
29 #include "got_object.h"
30 #include "got_blame.h"
31 #include "got_opentemp.h"
32 
33 #include "got_lib_inflate.h"
34 #include "got_lib_delta.h"
35 #include "got_lib_object.h"
36 #include "got_lib_diff.h"
37 #include "got_lib_diffoffset.h"
38 #include "got_commit_graph.h"
39 
40 struct got_blame_line {
41 	int annotated;
42 	struct got_object_id id;
43 };
44 
45 struct got_blame_diff_offsets {
46 	struct got_diffoffset_chunks *chunks;
47 	struct got_object_id *commit_id;
48 	SLIST_ENTRY(got_blame_diff_offsets) entry;
49 };
50 
51 SLIST_HEAD(got_blame_diff_offsets_list, got_blame_diff_offsets);
52 
53 struct got_blame {
54 	FILE *f;
55 	size_t nlines;
56 	struct got_blame_line *lines; /* one per line */
57 	int ncommits;
58 	struct got_blame_diff_offsets_list diff_offsets_list;
59 };
60 
61 static void
62 free_diff_offsets(struct got_blame_diff_offsets *diff_offsets)
63 {
64 	if (diff_offsets->chunks)
65 		got_diffoffset_free(diff_offsets->chunks);
66 	free(diff_offsets->commit_id);
67 	free(diff_offsets);
68 }
69 
70 static const struct got_error *
71 alloc_diff_offsets(struct got_blame_diff_offsets **diff_offsets,
72     struct got_object_id *commit_id)
73 {
74 	const struct got_error *err = NULL;
75 
76 	*diff_offsets = calloc(1, sizeof(**diff_offsets));
77 	if (*diff_offsets == NULL)
78 		return got_error_from_errno();
79 
80 	(*diff_offsets)->commit_id = got_object_id_dup(commit_id);
81 	if ((*diff_offsets)->commit_id == NULL) {
82 		err = got_error_from_errno();
83 		free_diff_offsets(*diff_offsets);
84 		*diff_offsets = NULL;
85 		return err;
86 	}
87 
88 	err = got_diffoffset_alloc(&(*diff_offsets)->chunks);
89 	if (err) {
90 		free_diff_offsets(*diff_offsets);
91 		return err;
92 	}
93 
94 	return NULL;
95 }
96 
97 static const struct got_error *
98 annotate_line(struct got_blame *blame, int lineno, struct got_object_id *id,
99     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
100     void *arg)
101 {
102 	const struct got_error *err = NULL;
103 	struct got_blame_line *line;
104 
105 	if (lineno < 1 || lineno > blame->nlines)
106 		return got_error(GOT_ERR_RANGE);
107 	
108 	line = &blame->lines[lineno - 1];
109 	if (line->annotated)
110 		return NULL;
111 
112 	memcpy(&line->id, id, sizeof(line->id));
113 	line->annotated = 1;
114 	if (cb)
115 		err = cb(arg, blame->nlines, lineno, id);
116 	return err;
117 }
118 
119 static int
120 get_blamed_line(struct got_blame_diff_offsets_list *diff_offsets_list,
121     int lineno)
122 {
123 	struct got_blame_diff_offsets *diff_offsets;
124 
125 	SLIST_FOREACH(diff_offsets, diff_offsets_list, entry)
126 		lineno = got_diffoffset_get(diff_offsets->chunks, lineno);
127 
128 	return lineno;
129 }
130 
131 static const struct got_error *
132 blame_changes(struct got_blame *blame, struct got_diff_changes *changes,
133     struct got_object_id *commit_id,
134     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
135     void *arg)
136 {
137 	const struct got_error *err = NULL;
138 	struct got_diff_change *change;
139 	struct got_blame_diff_offsets *diff_offsets;
140 
141 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
142 		int c = change->cv.c;
143 		int d = change->cv.d;
144 		int new_lineno = c;
145 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
146 		int ln;
147 
148 		for (ln = new_lineno; ln < new_lineno + new_length; ln++) {
149 			err = annotate_line(blame,
150 			    get_blamed_line(&blame->diff_offsets_list, ln),
151 			    commit_id, cb, arg);
152 			if (err)
153 				return err;
154 		}
155 	}
156 
157 	err = alloc_diff_offsets(&diff_offsets, commit_id);
158 	if (err)
159 		return err;
160 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
161 		int a = change->cv.a;
162 		int b = change->cv.b;
163 		int c = change->cv.c;
164 		int d = change->cv.d;
165 		int old_lineno = a;
166 		int old_length = (a < b ? b - a + 1 : (a == b ? 1 : 0));
167 		int new_lineno = c;
168 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
169 
170 		err = got_diffoffset_add(diff_offsets->chunks,
171 		    old_lineno, old_length, new_lineno, new_length);
172 		if (err) {
173 			free_diff_offsets(diff_offsets);
174 			return err;
175 		}
176 	}
177 	SLIST_INSERT_HEAD(&blame->diff_offsets_list, diff_offsets, entry);
178 
179 	return NULL;
180 }
181 
182 static const struct got_error *
183 blame_commit(struct got_blame *blame, struct got_object_id *id,
184     struct got_object_id *pid, const char *path, struct got_repository *repo,
185     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
186     void *arg)
187 {
188 	const struct got_error *err = NULL;
189 	struct got_object *obj = NULL, *pobj = NULL;
190 	struct got_object_id *obj_id = NULL, *pobj_id = NULL;
191 	struct got_blob_object *blob = NULL, *pblob = NULL;
192 	struct got_diff_changes *changes = NULL;
193 
194 	err = got_object_id_by_path(&obj_id, repo, id, path);
195 	if (err)
196 		goto done;
197 
198 	err = got_object_open(&obj, repo, obj_id);
199 	if (err)
200 		goto done;
201 
202 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
203 		err = got_error(GOT_ERR_OBJ_TYPE);
204 		goto done;
205 	}
206 
207 	err = got_object_id_by_path(&pobj_id, repo, pid, path);
208 	if (err) {
209 		if (err->code == GOT_ERR_NO_OBJ) {
210 			/* Blob's history began in previous commit. */
211 			err = got_error(GOT_ERR_ITER_COMPLETED);
212 		}
213 		goto done;
214 	}
215 
216 	/* If IDs match then don't bother with diffing. */
217 	if (got_object_id_cmp(obj_id, pobj_id) == 0) {
218 		if (cb)
219 			err = cb(arg, blame->nlines, -1, id);
220 		goto done;
221 	}
222 
223 	err = got_object_open(&pobj, repo, pobj_id);
224 	if (err)
225 		goto done;
226 
227 	if (got_object_get_type(pobj) != GOT_OBJ_TYPE_BLOB) {
228 		/*
229 		 * Encountered a non-blob at the path (probably a tree).
230 		 * Blob's history began in previous commit.
231 		 */
232 		err = got_error(GOT_ERR_ITER_COMPLETED);
233 		goto done;
234 	}
235 
236 	err = got_object_blob_open(&blob, repo, obj, 8192);
237 	if (err)
238 		goto done;
239 
240 	err = got_object_blob_open(&pblob, repo, pobj, 8192);
241 	if (err)
242 		goto done;
243 
244 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
245 	if (err)
246 		goto done;
247 
248 	if (changes) {
249 		err = blame_changes(blame, changes, id, cb, arg);
250 		got_diff_free_changes(changes);
251 	} else if (cb)
252 		err = cb(arg, blame->nlines, -1, id);
253 done:
254 	free(obj_id);
255 	free(pobj_id);
256 	if (obj)
257 		got_object_close(obj);
258 	if (pobj)
259 		got_object_close(pobj);
260 	if (blob)
261 		got_object_blob_close(blob);
262 	if (pblob)
263 		got_object_blob_close(pblob);
264 	return err;
265 }
266 
267 static void
268 blame_close(struct got_blame *blame)
269 {
270 	struct got_blame_diff_offsets *diff_offsets;
271 
272 	if (blame->f)
273 		fclose(blame->f);
274 	free(blame->lines);
275 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
276 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
277 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
278 		free_diff_offsets(diff_offsets);
279 	}
280 	free(blame);
281 }
282 
283 static const struct got_error *
284 blame_open(struct got_blame **blamep, const char *path,
285     struct got_object_id *start_commit_id, struct got_repository *repo,
286     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
287     void *arg)
288 {
289 	const struct got_error *err = NULL;
290 	struct got_object *obj = NULL;
291 	struct got_object_id *obj_id = NULL;
292 	struct got_blob_object *blob = NULL;
293 	struct got_blame *blame = NULL;
294 	struct got_object_id *id = NULL;
295 	int lineno;
296 	struct got_commit_graph *graph = NULL;
297 
298 	*blamep = NULL;
299 
300 	err = got_object_id_by_path(&obj_id, repo, start_commit_id, path);
301 	if (err)
302 		return err;
303 
304 	err = got_object_open(&obj, repo, obj_id);
305 	if (err)
306 		goto done;
307 
308 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
309 		err = got_error(GOT_ERR_OBJ_TYPE);
310 		goto done;
311 	}
312 
313 	err = got_object_blob_open(&blob, repo, obj, 8192);
314 	if (err)
315 		goto done;
316 
317 	blame = calloc(1, sizeof(*blame));
318 	if (blame == NULL)
319 		return got_error_from_errno();
320 
321 	blame->f = got_opentemp();
322 	if (blame->f == NULL) {
323 		err = got_error_from_errno();
324 		goto done;
325 	}
326 	err = got_object_blob_dump_to_file(NULL, &blame->nlines, blame->f,
327 	    blob);
328 	if (err)
329 		goto done;
330 
331 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
332 	if (blame->lines == NULL) {
333 		err = got_error_from_errno();
334 		goto done;
335 	}
336 
337 	err = got_commit_graph_open(&graph, start_commit_id, path, 0, repo);
338 	if (err)
339 		return err;
340 	err = got_commit_graph_iter_start(graph, start_commit_id, repo);
341 	if (err)
342 		goto done;
343 
344 	id = NULL;
345 	while (1) {
346 		struct got_object_id *next_id;
347 
348 		err = got_commit_graph_iter_next(&next_id, graph);
349 		if (err) {
350 			if (err->code == GOT_ERR_ITER_COMPLETED) {
351 				err = NULL;
352 				break;
353 			}
354 			if (err->code != GOT_ERR_ITER_NEED_MORE)
355 				break;
356 			err = got_commit_graph_fetch_commits(graph, 1, repo);
357 			if (err)
358 				break;
359 			else
360 				continue;
361 		}
362 		if (next_id == NULL)
363 			break;
364 		if (id) {
365 			err = blame_commit(blame, id, next_id, path, repo,
366 			    cb, arg);
367 			if (err) {
368 				if (err->code == GOT_ERR_ITER_COMPLETED)
369 					err = NULL;
370 				break;
371 			}
372 		}
373 		id = next_id;
374 	}
375 
376 	if (id) {
377 		/* Annotate remaining non-annotated lines with last commit. */
378 		for (lineno = 1; lineno <= blame->nlines; lineno++) {
379 			err = annotate_line(blame, lineno, id, cb, arg);
380 			if (err)
381 				goto done;
382 		}
383 	}
384 
385 done:
386 	if (graph)
387 		got_commit_graph_close(graph);
388 	free(obj_id);
389 	if (obj)
390 		got_object_close(obj);
391 	if (blob)
392 		got_object_blob_close(blob);
393 	if (err) {
394 		if (blame)
395 			blame_close(blame);
396 	} else
397 		*blamep = blame;
398 
399 	return err;
400 }
401 
402 static const struct got_error *
403 blame_line(struct got_object_id **id, struct got_blame *blame, int lineno)
404 {
405 	if (lineno < 1 || lineno > blame->nlines)
406 		return got_error(GOT_ERR_RANGE);
407 	*id = &blame->lines[lineno - 1].id;
408 	return NULL;
409 }
410 
411 static char *
412 parse_next_line(FILE *f, size_t *len)
413 {
414 	char *line;
415 	size_t linelen;
416 	size_t lineno;
417 	const char delim[3] = { '\0', '\0', '\0'};
418 
419 	line = fparseln(f, &linelen, &lineno, delim, 0);
420 	if (len)
421 		*len = linelen;
422 	return line;
423 }
424 
425 const struct got_error *
426 got_blame(const char *path, struct got_object_id *start_commit_id,
427     struct got_repository *repo, FILE *outfile)
428 {
429 	const struct got_error *err = NULL;
430 	struct got_blame *blame;
431 	int lineno;
432 	char *abspath;
433 
434 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
435 		return got_error_from_errno();
436 
437 	err = blame_open(&blame, abspath, start_commit_id, repo, NULL, NULL);
438 	if (err) {
439 		free(abspath);
440 		return err;
441 	}
442 
443 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
444 		struct got_object_id *id;
445 		char *line, *id_str;
446 		
447 		line = parse_next_line(blame->f, NULL);
448 		if (line == NULL)
449 			break;
450 
451 		err = blame_line(&id, blame, lineno);
452 		if (err) {
453 			free(line);
454 			break;
455 		}
456 
457 		err = got_object_id_str(&id_str, id);
458 		/* Do not free id; It points into blame->lines. */
459 		if (err) {
460 			free(line);
461 			break;
462 		}
463 
464 		fprintf(outfile, "%.8s %s\n", id_str, line);
465 		free(line);
466 		free(id_str);
467 	}
468 
469 	blame_close(blame);
470 	free(abspath);
471 	return err;
472 }
473 
474 const struct got_error *
475 got_blame_incremental(const char *path, struct got_object_id *commit_id,
476     struct got_repository *repo,
477     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
478     void *arg)
479 {
480 	const struct got_error *err = NULL;
481 	struct got_blame *blame;
482 	char *abspath;
483 
484 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
485 		return got_error_from_errno();
486 
487 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
488 	free(abspath);
489 	if (blame)
490 		blame_close(blame);
491 	return err;
492 }