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 	off_t offset;
44 };
45 
46 struct got_blame_diff_offsets {
47 	struct got_diffoffset_chunks *chunks;
48 	struct got_object_id *commit_id;
49 	SLIST_ENTRY(got_blame_diff_offsets) entry;
50 };
51 
52 SLIST_HEAD(got_blame_diff_offsets_list, got_blame_diff_offsets);
53 
54 struct got_blame {
55 	FILE *f;
56 	size_t filesize;
57 	int nlines;
58 	int nannotated;
59 	struct got_blame_line *lines; /* one per line */
60 	off_t *line_offsets;		/* one per line */
61 	int ncommits;
62 	struct got_blame_diff_offsets_list diff_offsets_list;
63 };
64 
65 static void
66 free_diff_offsets(struct got_blame_diff_offsets *diff_offsets)
67 {
68 	if (diff_offsets->chunks)
69 		got_diffoffset_free(diff_offsets->chunks);
70 	free(diff_offsets->commit_id);
71 	free(diff_offsets);
72 }
73 
74 static const struct got_error *
75 alloc_diff_offsets(struct got_blame_diff_offsets **diff_offsets,
76     struct got_object_id *commit_id)
77 {
78 	const struct got_error *err = NULL;
79 
80 	*diff_offsets = calloc(1, sizeof(**diff_offsets));
81 	if (*diff_offsets == NULL)
82 		return got_error_from_errno("calloc");
83 
84 	(*diff_offsets)->commit_id = got_object_id_dup(commit_id);
85 	if ((*diff_offsets)->commit_id == NULL) {
86 		err = got_error_from_errno("got_object_id_dup");
87 		free_diff_offsets(*diff_offsets);
88 		*diff_offsets = NULL;
89 		return err;
90 	}
91 
92 	err = got_diffoffset_alloc(&(*diff_offsets)->chunks);
93 	if (err) {
94 		free_diff_offsets(*diff_offsets);
95 		return err;
96 	}
97 
98 	return NULL;
99 }
100 
101 static const struct got_error *
102 annotate_line(struct got_blame *blame, int lineno, struct got_object_id *id,
103     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
104     void *arg)
105 {
106 	const struct got_error *err = NULL;
107 	struct got_blame_line *line;
108 
109 	if (lineno < 1 || lineno > blame->nlines)
110 		return NULL;
111 	
112 	line = &blame->lines[lineno - 1];
113 	if (line->annotated)
114 		return NULL;
115 
116 	memcpy(&line->id, id, sizeof(line->id));
117 	line->annotated = 1;
118 	blame->nannotated++;
119 	if (cb)
120 		err = cb(arg, blame->nlines, lineno, id);
121 	return err;
122 }
123 
124 static int
125 get_blamed_line(struct got_blame_diff_offsets_list *diff_offsets_list,
126     int lineno)
127 {
128 	struct got_blame_diff_offsets *diff_offsets;
129 
130 	SLIST_FOREACH(diff_offsets, diff_offsets_list, entry)
131 		lineno = got_diffoffset_get(diff_offsets->chunks, lineno);
132 
133 	return lineno;
134 }
135 
136 static const struct got_error *
137 blame_changes(struct got_blame *blame, struct got_diff_changes *changes,
138     struct got_object_id *commit_id,
139     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
140     void *arg)
141 {
142 	const struct got_error *err = NULL;
143 	struct got_diff_change *change;
144 	struct got_blame_diff_offsets *diff_offsets;
145 
146 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
147 		int c = change->cv.c;
148 		int d = change->cv.d;
149 		int new_lineno = c;
150 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
151 		int ln;
152 
153 		for (ln = new_lineno; ln < new_lineno + new_length; ln++) {
154 			err = annotate_line(blame,
155 			    get_blamed_line(&blame->diff_offsets_list, ln),
156 			    commit_id, cb, arg);
157 			if (err)
158 				return err;
159 			if (blame->nlines == blame->nannotated)
160 				return NULL;
161 		}
162 	}
163 
164 	err = alloc_diff_offsets(&diff_offsets, commit_id);
165 	if (err)
166 		return err;
167 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
168 		int a = change->cv.a;
169 		int b = change->cv.b;
170 		int c = change->cv.c;
171 		int d = change->cv.d;
172 		int old_lineno = a;
173 		int old_length = (a < b ? b - a + 1 : (a == b ? 1 : 0));
174 		int new_lineno = c;
175 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
176 
177 		err = got_diffoffset_add(diff_offsets->chunks,
178 		    old_lineno, old_length, new_lineno, new_length);
179 		if (err) {
180 			free_diff_offsets(diff_offsets);
181 			return err;
182 		}
183 	}
184 	SLIST_INSERT_HEAD(&blame->diff_offsets_list, diff_offsets, entry);
185 
186 	return NULL;
187 }
188 
189 static const struct got_error *
190 blame_commit(struct got_blame *blame, struct got_object_id *id,
191     struct got_object_id *pid, const char *path, struct got_repository *repo,
192     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
193     void *arg)
194 {
195 	const struct got_error *err = NULL;
196 	struct got_object *obj = NULL, *pobj = NULL;
197 	struct got_object_id *obj_id = NULL, *pobj_id = NULL;
198 	struct got_blob_object *blob = NULL, *pblob = NULL;
199 	struct got_diff_changes *changes = NULL;
200 
201 	err = got_object_id_by_path(&obj_id, repo, id, path);
202 	if (err)
203 		goto done;
204 
205 	err = got_object_open(&obj, repo, obj_id);
206 	if (err)
207 		goto done;
208 
209 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
210 		err = got_error(GOT_ERR_OBJ_TYPE);
211 		goto done;
212 	}
213 
214 	err = got_object_id_by_path(&pobj_id, repo, pid, path);
215 	if (err) {
216 		if (err->code == GOT_ERR_NO_TREE_ENTRY) {
217 			/* Blob's history began in previous commit. */
218 			err = got_error(GOT_ERR_ITER_COMPLETED);
219 		}
220 		goto done;
221 	}
222 
223 	/* If IDs match then don't bother with diffing. */
224 	if (got_object_id_cmp(obj_id, pobj_id) == 0) {
225 		if (cb)
226 			err = cb(arg, blame->nlines, -1, id);
227 		goto done;
228 	}
229 
230 	err = got_object_open(&pobj, repo, pobj_id);
231 	if (err)
232 		goto done;
233 
234 	if (pobj->type != GOT_OBJ_TYPE_BLOB) {
235 		/*
236 		 * Encountered a non-blob at the path (probably a tree).
237 		 * Blob's history began in previous commit.
238 		 */
239 		err = got_error(GOT_ERR_ITER_COMPLETED);
240 		goto done;
241 	}
242 
243 	err = got_object_blob_open(&blob, repo, obj, 8192);
244 	if (err)
245 		goto done;
246 
247 	err = got_object_blob_open(&pblob, repo, pobj, 8192);
248 	if (err)
249 		goto done;
250 
251 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
252 	if (err)
253 		goto done;
254 
255 	if (changes) {
256 		err = blame_changes(blame, changes, id, cb, arg);
257 		got_diff_free_changes(changes);
258 	} else if (cb)
259 		err = cb(arg, blame->nlines, -1, id);
260 done:
261 	free(obj_id);
262 	free(pobj_id);
263 	if (obj)
264 		got_object_close(obj);
265 	if (pobj)
266 		got_object_close(pobj);
267 	if (blob)
268 		got_object_blob_close(blob);
269 	if (pblob)
270 		got_object_blob_close(pblob);
271 	return err;
272 }
273 
274 static const struct got_error *
275 blame_close(struct got_blame *blame)
276 {
277 	const struct got_error *err = NULL;
278 	struct got_blame_diff_offsets *diff_offsets;
279 
280 	if (blame->f && fclose(blame->f) != 0)
281 		err = got_error_from_errno("fclose");
282 	free(blame->lines);
283 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
284 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
285 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
286 		free_diff_offsets(diff_offsets);
287 	}
288 	free(blame);
289 	return err;
290 }
291 
292 static const struct got_error *
293 blame_open(struct got_blame **blamep, const char *path,
294     struct got_object_id *start_commit_id, struct got_repository *repo,
295     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
296     void *arg)
297 {
298 	const struct got_error *err = NULL;
299 	struct got_object *obj = NULL;
300 	struct got_object_id *obj_id = NULL;
301 	struct got_blob_object *blob = NULL;
302 	struct got_blame *blame = NULL;
303 	struct got_object_id *id = NULL;
304 	int lineno;
305 	struct got_commit_graph *graph = NULL;
306 
307 	*blamep = NULL;
308 
309 	err = got_object_id_by_path(&obj_id, repo, start_commit_id, path);
310 	if (err)
311 		return err;
312 
313 	err = got_object_open(&obj, repo, obj_id);
314 	if (err)
315 		goto done;
316 
317 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
318 		err = got_error(GOT_ERR_OBJ_TYPE);
319 		goto done;
320 	}
321 
322 	err = got_object_blob_open(&blob, repo, obj, 8192);
323 	if (err)
324 		goto done;
325 
326 	blame = calloc(1, sizeof(*blame));
327 	if (blame == NULL)
328 		return got_error_from_errno("calloc");
329 
330 	blame->f = got_opentemp();
331 	if (blame->f == NULL) {
332 		err = got_error_from_errno("got_opentemp");
333 		goto done;
334 	}
335 	err = got_object_blob_dump_to_file(&blame->filesize, &blame->nlines,
336 	   &blame->line_offsets, blame->f, blob);
337 	if (err)
338 		goto done;
339 
340 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
341 	if (blame->lines == NULL) {
342 		err = got_error_from_errno("calloc");
343 		goto done;
344 	}
345 
346 	err = got_commit_graph_open(&graph, start_commit_id, path, 0, repo);
347 	if (err)
348 		return err;
349 	err = got_commit_graph_iter_start(graph, start_commit_id, repo);
350 	if (err)
351 		goto done;
352 
353 	id = NULL;
354 	for (;;) {
355 		struct got_object_id *next_id;
356 
357 		err = got_commit_graph_iter_next(&next_id, graph);
358 		if (err) {
359 			if (err->code == GOT_ERR_ITER_COMPLETED) {
360 				err = NULL;
361 				break;
362 			}
363 			if (err->code != GOT_ERR_ITER_NEED_MORE)
364 				break;
365 			err = got_commit_graph_fetch_commits(graph, 1, repo);
366 			if (err)
367 				break;
368 			else
369 				continue;
370 		}
371 		if (next_id == NULL)
372 			break;
373 		if (id) {
374 			err = blame_commit(blame, id, next_id, path, repo,
375 			    cb, arg);
376 			if (err) {
377 				if (err->code == GOT_ERR_ITER_COMPLETED)
378 					err = NULL;
379 				break;
380 			}
381 			if (blame->nannotated == blame->nlines)
382 				break;
383 		}
384 		id = next_id;
385 	}
386 
387 	if (id && blame->nannotated < blame->nlines) {
388 		/* Annotate remaining non-annotated lines with last commit. */
389 		for (lineno = 1; lineno <= blame->nlines; lineno++) {
390 			err = annotate_line(blame, lineno, id, cb, arg);
391 			if (err)
392 				goto done;
393 		}
394 	}
395 
396 done:
397 	if (graph)
398 		got_commit_graph_close(graph);
399 	free(obj_id);
400 	if (obj)
401 		got_object_close(obj);
402 	if (blob)
403 		got_object_blob_close(blob);
404 	if (err) {
405 		if (blame)
406 			blame_close(blame);
407 	} else
408 		*blamep = blame;
409 
410 	return err;
411 }
412 
413 static const struct got_error *
414 blame_line(struct got_object_id **id, struct got_blame *blame, int lineno)
415 {
416 	if (lineno < 1 || lineno > blame->nlines)
417 		return got_error(GOT_ERR_RANGE);
418 	*id = &blame->lines[lineno - 1].id;
419 	return NULL;
420 }
421 
422 static char *
423 parse_next_line(FILE *f, size_t *len)
424 {
425 	char *line;
426 	size_t linelen;
427 	size_t lineno;
428 	const char delim[3] = { '\0', '\0', '\0'};
429 
430 	line = fparseln(f, &linelen, &lineno, delim, 0);
431 	if (len)
432 		*len = linelen;
433 	return line;
434 }
435 
436 const struct got_error *
437 got_blame(const char *path, struct got_object_id *start_commit_id,
438     struct got_repository *repo, FILE *outfile)
439 {
440 	const struct got_error *err = NULL, *close_err = NULL;
441 	struct got_blame *blame;
442 	int lineno;
443 	char *abspath;
444 
445 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
446 		return got_error_from_errno2("asprintf", path);
447 
448 	err = blame_open(&blame, abspath, start_commit_id, repo, NULL, NULL);
449 	if (err) {
450 		free(abspath);
451 		return err;
452 	}
453 
454 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
455 		struct got_object_id *id;
456 		char *line, *id_str;
457 
458 		line = parse_next_line(blame->f, NULL);
459 		if (line == NULL)
460 			break;
461 
462 		err = blame_line(&id, blame, lineno);
463 		if (err) {
464 			free(line);
465 			break;
466 		}
467 
468 		err = got_object_id_str(&id_str, id);
469 		/* Do not free id; It points into blame->lines. */
470 		if (err) {
471 			free(line);
472 			break;
473 		}
474 
475 		fprintf(outfile, "%.8s %s\n", id_str, line);
476 		free(line);
477 		free(id_str);
478 	}
479 
480 	close_err = blame_close(blame);
481 	free(abspath);
482 	return err ? err : close_err;
483 }
484 
485 const struct got_error *
486 got_blame_incremental(const char *path, struct got_object_id *commit_id,
487     struct got_repository *repo,
488     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
489     void *arg)
490 {
491 	const struct got_error *err = NULL, *close_err = NULL;
492 	struct got_blame *blame;
493 	char *abspath;
494 
495 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
496 		return got_error_from_errno2("asprintf", path);
497 
498 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
499 	free(abspath);
500 	if (blame)
501 		close_err = blame_close(blame);
502 	return err ? err : close_err;
503 }