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 	if (pid) {
215 		err = got_object_id_by_path(&pobj_id, repo, pid, path);
216 		if (err) {
217 			if (err->code == GOT_ERR_NO_TREE_ENTRY) {
218 				/* Blob's history began in previous commit. */
219 				err = got_error(GOT_ERR_ITER_COMPLETED);
220 			}
221 			goto done;
222 		}
223 
224 		/* If IDs match then don't bother with diffing. */
225 		if (got_object_id_cmp(obj_id, pobj_id) == 0) {
226 			if (cb)
227 				err = cb(arg, blame->nlines, -1, id);
228 			goto done;
229 		}
230 
231 		err = got_object_open(&pobj, repo, pobj_id);
232 		if (err)
233 			goto done;
234 
235 		if (pobj->type != GOT_OBJ_TYPE_BLOB) {
236 			/*
237 			 * Encountered a non-blob at the path (probably a tree).
238 			 * Blob's history began in previous commit.
239 			 */
240 			err = got_error(GOT_ERR_ITER_COMPLETED);
241 			goto done;
242 		}
243 
244 		err = got_object_blob_open(&pblob, repo, pobj, 8192);
245 		if (err)
246 			goto done;
247 	}
248 
249 	err = got_object_blob_open(&blob, repo, obj, 8192);
250 	if (err)
251 		goto done;
252 
253 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
254 	if (err)
255 		goto done;
256 
257 	if (changes) {
258 		err = blame_changes(blame, changes, id, cb, arg);
259 		got_diff_free_changes(changes);
260 	} else if (cb)
261 		err = cb(arg, blame->nlines, -1, id);
262 done:
263 	free(obj_id);
264 	free(pobj_id);
265 	if (obj)
266 		got_object_close(obj);
267 	if (pobj)
268 		got_object_close(pobj);
269 	if (blob)
270 		got_object_blob_close(blob);
271 	if (pblob)
272 		got_object_blob_close(pblob);
273 	return err;
274 }
275 
276 static const struct got_error *
277 blame_close(struct got_blame *blame)
278 {
279 	const struct got_error *err = NULL;
280 	struct got_blame_diff_offsets *diff_offsets;
281 
282 	if (blame->f && fclose(blame->f) != 0)
283 		err = got_error_from_errno("fclose");
284 	free(blame->lines);
285 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
286 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
287 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
288 		free_diff_offsets(diff_offsets);
289 	}
290 	free(blame);
291 	return err;
292 }
293 
294 static const struct got_error *
295 blame_open(struct got_blame **blamep, const char *path,
296     struct got_object_id *start_commit_id, struct got_repository *repo,
297     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
298     void *arg)
299 {
300 	const struct got_error *err = NULL;
301 	struct got_object *obj = NULL;
302 	struct got_object_id *obj_id = NULL;
303 	struct got_blob_object *blob = NULL;
304 	struct got_blame *blame = NULL;
305 	struct got_object_id *id = NULL, *parent_id = NULL;
306 	int lineno;
307 	struct got_commit_graph *graph = NULL;
308 
309 	*blamep = NULL;
310 
311 	err = got_object_id_by_path(&obj_id, repo, start_commit_id, path);
312 	if (err)
313 		return err;
314 
315 	err = got_object_open(&obj, repo, obj_id);
316 	if (err)
317 		goto done;
318 
319 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
320 		err = got_error(GOT_ERR_OBJ_TYPE);
321 		goto done;
322 	}
323 
324 	err = got_object_blob_open(&blob, repo, obj, 8192);
325 	if (err)
326 		goto done;
327 
328 	blame = calloc(1, sizeof(*blame));
329 	if (blame == NULL)
330 		return got_error_from_errno("calloc");
331 
332 	blame->f = got_opentemp();
333 	if (blame->f == NULL) {
334 		err = got_error_from_errno("got_opentemp");
335 		goto done;
336 	}
337 	err = got_object_blob_dump_to_file(&blame->filesize, &blame->nlines,
338 	   &blame->line_offsets, blame->f, blob);
339 	if (err)
340 		goto done;
341 
342 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
343 	if (blame->lines == NULL) {
344 		err = got_error_from_errno("calloc");
345 		goto done;
346 	}
347 
348 	err = got_commit_graph_open(&graph, start_commit_id, path, 0, repo);
349 	if (err)
350 		return err;
351 	err = got_commit_graph_iter_start(graph, start_commit_id, repo);
352 	if (err)
353 		goto done;
354 
355 	id = NULL;
356 	for (;;) {
357 		err = got_commit_graph_iter_next(&parent_id, graph);
358 		if (err) {
359 			if (err->code == GOT_ERR_ITER_COMPLETED) {
360 				if (id)
361 					err = blame_commit(blame, id,
362 					    parent_id, path, repo, cb, arg);
363 				else
364 					err = NULL;
365 				break;
366 			}
367 			if (err->code != GOT_ERR_ITER_NEED_MORE)
368 				break;
369 			err = got_commit_graph_fetch_commits(graph, 1, repo);
370 			if (err)
371 				break;
372 			continue;
373 		}
374 		if (id) {
375 			err = blame_commit(blame, id, parent_id, path, repo,
376 			    cb, arg);
377 			if (err) {
378 				if (err->code == GOT_ERR_ITER_COMPLETED)
379 					err = NULL;
380 				break;
381 			}
382 			if (blame->nannotated == blame->nlines)
383 				break;
384 		}
385 		id = parent_id;
386 	}
387 
388 	if (id && blame->nannotated < blame->nlines) {
389 		/* Annotate remaining non-annotated lines with last commit. */
390 		for (lineno = 1; lineno <= blame->nlines; lineno++) {
391 			err = annotate_line(blame, lineno, id, cb, arg);
392 			if (err)
393 				goto done;
394 		}
395 	}
396 
397 done:
398 	if (graph)
399 		got_commit_graph_close(graph);
400 	free(obj_id);
401 	if (obj)
402 		got_object_close(obj);
403 	if (blob)
404 		got_object_blob_close(blob);
405 	if (err) {
406 		if (blame)
407 			blame_close(blame);
408 	} else
409 		*blamep = blame;
410 
411 	return err;
412 }
413 
414 static const struct got_error *
415 blame_line(struct got_object_id **id, struct got_blame *blame, int lineno)
416 {
417 	if (lineno < 1 || lineno > blame->nlines) {
418 		*id = NULL;
419 		return got_error(GOT_ERR_RANGE);
420 	}
421 	*id = &blame->lines[lineno - 1].id;
422 	return NULL;
423 }
424 
425 static char *
426 parse_next_line(FILE *f, size_t *len)
427 {
428 	char *line;
429 	size_t linelen;
430 	size_t lineno;
431 	const char delim[3] = { '\0', '\0', '\0'};
432 
433 	line = fparseln(f, &linelen, &lineno, delim, 0);
434 	if (len)
435 		*len = linelen;
436 	return line;
437 }
438 
439 const struct got_error *
440 got_blame(const char *path, struct got_object_id *start_commit_id,
441     struct got_repository *repo, FILE *outfile)
442 {
443 	const struct got_error *err = NULL, *close_err = NULL;
444 	struct got_blame *blame;
445 	int lineno;
446 	char *abspath;
447 
448 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
449 		return got_error_from_errno2("asprintf", path);
450 
451 	err = blame_open(&blame, abspath, start_commit_id, repo, NULL, NULL);
452 	if (err) {
453 		free(abspath);
454 		return err;
455 	}
456 
457 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
458 		struct got_object_id *id = NULL;
459 		char *line, *id_str;
460 
461 		line = parse_next_line(blame->f, NULL);
462 		if (line == NULL)
463 			break;
464 
465 		err = blame_line(&id, blame, lineno);
466 		if (err) {
467 			free(line);
468 			break;
469 		}
470 
471 		err = got_object_id_str(&id_str, id);
472 		/* Do not free id; It points into blame->lines. */
473 		if (err) {
474 			free(line);
475 			break;
476 		}
477 
478 		fprintf(outfile, "%.8s %s\n", id_str, line);
479 		free(line);
480 		free(id_str);
481 	}
482 
483 	close_err = blame_close(blame);
484 	free(abspath);
485 	return err ? err : close_err;
486 }
487 
488 const struct got_error *
489 got_blame_incremental(const char *path, struct got_object_id *commit_id,
490     struct got_repository *repo,
491     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
492     void *arg)
493 {
494 	const struct got_error *err = NULL, *close_err = NULL;
495 	struct got_blame *blame;
496 	char *abspath;
497 
498 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
499 		return got_error_from_errno2("asprintf", path);
500 
501 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
502 	free(abspath);
503 	if (blame)
504 		close_err = blame_close(blame);
505 	return err ? err : close_err;
506 }