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 filesize;
56 	int nlines;
57 	int nannotated;
58 	struct got_blame_line *lines; /* one per line */
59 	off_t *line_offsets;		/* one per line */
60 	int ncommits;
61 	struct got_blame_diff_offsets_list diff_offsets_list;
62 };
63 
64 static void
65 free_diff_offsets(struct got_blame_diff_offsets *diff_offsets)
66 {
67 	if (diff_offsets->chunks)
68 		got_diffoffset_free(diff_offsets->chunks);
69 	free(diff_offsets->commit_id);
70 	free(diff_offsets);
71 }
72 
73 static const struct got_error *
74 alloc_diff_offsets(struct got_blame_diff_offsets **diff_offsets,
75     struct got_object_id *commit_id)
76 {
77 	const struct got_error *err = NULL;
78 
79 	*diff_offsets = calloc(1, sizeof(**diff_offsets));
80 	if (*diff_offsets == NULL)
81 		return got_error_from_errno("calloc");
82 
83 	(*diff_offsets)->commit_id = got_object_id_dup(commit_id);
84 	if ((*diff_offsets)->commit_id == NULL) {
85 		err = got_error_from_errno("got_object_id_dup");
86 		free_diff_offsets(*diff_offsets);
87 		*diff_offsets = NULL;
88 		return err;
89 	}
90 
91 	err = got_diffoffset_alloc(&(*diff_offsets)->chunks);
92 	if (err) {
93 		free_diff_offsets(*diff_offsets);
94 		return err;
95 	}
96 
97 	return NULL;
98 }
99 
100 static const struct got_error *
101 annotate_line(struct got_blame *blame, int lineno, struct got_object_id *id,
102     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
103     void *arg)
104 {
105 	const struct got_error *err = NULL;
106 	struct got_blame_line *line;
107 
108 	if (lineno < 1 || lineno > blame->nlines)
109 		return NULL;
110 	
111 	line = &blame->lines[lineno - 1];
112 	if (line->annotated)
113 		return NULL;
114 
115 	memcpy(&line->id, id, sizeof(line->id));
116 	line->annotated = 1;
117 	blame->nannotated++;
118 	if (cb)
119 		err = cb(arg, blame->nlines, lineno, id);
120 	return err;
121 }
122 
123 static int
124 get_blamed_line(struct got_blame_diff_offsets_list *diff_offsets_list,
125     int lineno)
126 {
127 	struct got_blame_diff_offsets *diff_offsets;
128 	int offset = 0;
129 
130 	SLIST_FOREACH(diff_offsets, diff_offsets_list, entry)
131 		lineno = got_diffoffset_get(diff_offsets->chunks, lineno);
132 
133 	return lineno + offset;
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     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_commit_object *commit = NULL;
199 	struct got_blob_object *blob = NULL, *pblob = NULL;
200 	struct got_diff_changes *changes = NULL;
201 	struct got_object_qid *pid = NULL;
202 
203 	err = got_object_open_as_commit(&commit, repo, id);
204 	if (err)
205 		return err;
206 
207 	err = got_object_id_by_path(&obj_id, repo, id, path);
208 	if (err)
209 		goto done;
210 
211 	err = got_object_open(&obj, repo, obj_id);
212 	if (err)
213 		goto done;
214 
215 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
216 		err = got_error(GOT_ERR_OBJ_TYPE);
217 		goto done;
218 	}
219 
220 	pid = SIMPLEQ_FIRST(got_object_commit_get_parent_ids(commit));
221 	if (pid) {
222 		err = got_object_id_by_path(&pobj_id, repo, pid->id, path);
223 		if (err) {
224 			if (err->code == GOT_ERR_NO_TREE_ENTRY) {
225 				/* Blob's history began in previous commit. */
226 				err = got_error(GOT_ERR_ITER_COMPLETED);
227 			}
228 			goto done;
229 		}
230 
231 		/* If IDs match then don't bother with diffing. */
232 		if (got_object_id_cmp(obj_id, pobj_id) == 0) {
233 			if (cb)
234 				err = cb(arg, blame->nlines, -1, id);
235 			goto done;
236 		}
237 
238 		err = got_object_open(&pobj, repo, pobj_id);
239 		if (err)
240 			goto done;
241 
242 		if (pobj->type != GOT_OBJ_TYPE_BLOB) {
243 			/*
244 			 * Encountered a non-blob at the path (probably a tree).
245 			 * Blob's history began in previous commit.
246 			 */
247 			err = got_error(GOT_ERR_ITER_COMPLETED);
248 			goto done;
249 		}
250 
251 		err = got_object_blob_open(&pblob, repo, pobj, 8192);
252 		if (err)
253 			goto done;
254 	}
255 
256 	err = got_object_blob_open(&blob, repo, obj, 8192);
257 	if (err)
258 		goto done;
259 
260 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
261 	if (err)
262 		goto done;
263 
264 	if (changes) {
265 		err = blame_changes(blame, changes, id, cb, arg);
266 		got_diff_free_changes(changes);
267 	} else if (cb)
268 		err = cb(arg, blame->nlines, -1, id);
269 done:
270 	if (commit)
271 		got_object_commit_close(commit);
272 	free(obj_id);
273 	free(pobj_id);
274 	if (obj)
275 		got_object_close(obj);
276 	if (pobj)
277 		got_object_close(pobj);
278 	if (blob)
279 		got_object_blob_close(blob);
280 	if (pblob)
281 		got_object_blob_close(pblob);
282 	return err;
283 }
284 
285 static const struct got_error *
286 blame_close(struct got_blame *blame)
287 {
288 	const struct got_error *err = NULL;
289 	struct got_blame_diff_offsets *diff_offsets;
290 
291 	if (blame->f && fclose(blame->f) != 0)
292 		err = got_error_from_errno("fclose");
293 	free(blame->lines);
294 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
295 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
296 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
297 		free_diff_offsets(diff_offsets);
298 	}
299 	free(blame);
300 	return err;
301 }
302 
303 static const struct got_error *
304 blame_open(struct got_blame **blamep, const char *path,
305     struct got_object_id *start_commit_id, struct got_repository *repo,
306     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
307     void *arg)
308 {
309 	const struct got_error *err = NULL;
310 	struct got_object *obj = NULL;
311 	struct got_object_id *obj_id = NULL;
312 	struct got_blob_object *blob = NULL;
313 	struct got_blame *blame = NULL;
314 	struct got_object_id *id = NULL, *next_id = NULL;
315 	int lineno;
316 	struct got_commit_graph *graph = NULL;
317 
318 	*blamep = NULL;
319 
320 	err = got_object_id_by_path(&obj_id, repo, start_commit_id, path);
321 	if (err)
322 		return err;
323 
324 	err = got_object_open(&obj, repo, obj_id);
325 	if (err)
326 		goto done;
327 
328 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
329 		err = got_error(GOT_ERR_OBJ_TYPE);
330 		goto done;
331 	}
332 
333 	err = got_object_blob_open(&blob, repo, obj, 8192);
334 	if (err)
335 		goto done;
336 
337 	blame = calloc(1, sizeof(*blame));
338 	if (blame == NULL)
339 		return got_error_from_errno("calloc");
340 
341 	blame->f = got_opentemp();
342 	if (blame->f == NULL) {
343 		err = got_error_from_errno("got_opentemp");
344 		goto done;
345 	}
346 	err = got_object_blob_dump_to_file(&blame->filesize, &blame->nlines,
347 	   &blame->line_offsets, blame->f, blob);
348 	if (err || blame->nlines == 0)
349 		goto done;
350 
351 	/* Don't include \n at EOF in the blame line count. */
352 	if (blame->line_offsets[blame->nlines - 1] == blame->filesize)
353 		blame->nlines--;
354 
355 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
356 	if (blame->lines == NULL) {
357 		err = got_error_from_errno("calloc");
358 		goto done;
359 	}
360 
361 	err = got_commit_graph_open(&graph, start_commit_id, path, 1, repo);
362 	if (err)
363 		return err;
364 	err = got_commit_graph_iter_start(graph, start_commit_id, repo);
365 	if (err)
366 		goto done;
367 
368 	id = NULL;
369 	for (;;) {
370 		err = got_commit_graph_iter_next(&next_id, graph);
371 		if (err) {
372 			if (err->code == GOT_ERR_ITER_COMPLETED) {
373 				if (id)
374 					err = blame_commit(blame, id,
375 					    path, repo, cb, arg);
376 				else
377 					err = NULL;
378 				break;
379 			}
380 			if (err->code != GOT_ERR_ITER_NEED_MORE)
381 				break;
382 			err = got_commit_graph_fetch_commits(graph, 1, repo);
383 			if (err)
384 				break;
385 			continue;
386 		}
387 		if (id) {
388 			err = blame_commit(blame, id, path, repo,
389 			    cb, arg);
390 			if (err) {
391 				if (err->code == GOT_ERR_ITER_COMPLETED)
392 					err = NULL;
393 				break;
394 			}
395 			if (blame->nannotated == blame->nlines)
396 				break;
397 		}
398 		id = next_id;
399 	}
400 
401 	if (id && blame->nannotated < blame->nlines) {
402 		/* Annotate remaining non-annotated lines with last commit. */
403 		for (lineno = 1; lineno <= blame->nlines; lineno++) {
404 			err = annotate_line(blame, lineno, id, cb, arg);
405 			if (err)
406 				goto done;
407 		}
408 	}
409 
410 done:
411 	if (graph)
412 		got_commit_graph_close(graph);
413 	free(obj_id);
414 	if (obj)
415 		got_object_close(obj);
416 	if (blob)
417 		got_object_blob_close(blob);
418 	if (err) {
419 		if (blame)
420 			blame_close(blame);
421 	} else
422 		*blamep = blame;
423 
424 	return err;
425 }
426 
427 const struct got_error *
428 got_blame(const char *path, struct got_object_id *commit_id,
429     struct got_repository *repo,
430     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
431     void *arg)
432 {
433 	const struct got_error *err = NULL, *close_err = NULL;
434 	struct got_blame *blame;
435 	char *abspath;
436 
437 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
438 		return got_error_from_errno2("asprintf", path);
439 
440 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
441 	free(abspath);
442 	if (blame)
443 		close_err = blame_close(blame);
444 	return err ? err : close_err;
445 }