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