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 
129 	SLIST_FOREACH(diff_offsets, diff_offsets_list, entry)
130 		lineno = got_diffoffset_get(diff_offsets->chunks, lineno);
131 
132 	return lineno;
133 }
134 
135 static const struct got_error *
136 blame_changes(struct got_blame *blame, struct got_diff_changes *changes,
137     struct got_object_id *commit_id,
138     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
139     void *arg)
140 {
141 	const struct got_error *err = NULL;
142 	struct got_diff_change *change;
143 	struct got_blame_diff_offsets *diff_offsets;
144 
145 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
146 		int c = change->cv.c;
147 		int d = change->cv.d;
148 		int new_lineno = c;
149 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
150 		int ln;
151 
152 		for (ln = new_lineno; ln < new_lineno + new_length; ln++) {
153 			err = annotate_line(blame,
154 			    get_blamed_line(&blame->diff_offsets_list, ln),
155 			    commit_id, cb, arg);
156 			if (err)
157 				return err;
158 			if (blame->nlines == blame->nannotated)
159 				return NULL;
160 		}
161 	}
162 
163 	err = alloc_diff_offsets(&diff_offsets, commit_id);
164 	if (err)
165 		return err;
166 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
167 		int a = change->cv.a;
168 		int b = change->cv.b;
169 		int c = change->cv.c;
170 		int d = change->cv.d;
171 		int old_lineno = a;
172 		int old_length = (a < b ? b - a + 1 : (a == b ? 1 : 0));
173 		int new_lineno = c;
174 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
175 
176 		err = got_diffoffset_add(diff_offsets->chunks,
177 		    old_lineno, old_length, new_lineno, new_length);
178 		if (err) {
179 			free_diff_offsets(diff_offsets);
180 			return err;
181 		}
182 	}
183 	SLIST_INSERT_HEAD(&blame->diff_offsets_list, diff_offsets, entry);
184 
185 	return NULL;
186 }
187 
188 static const struct got_error *
189 blame_commit(struct got_blame *blame, struct got_object_id *id,
190     const char *path, struct got_repository *repo,
191     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
192     void *arg)
193 {
194 	const struct got_error *err = NULL;
195 	struct got_object *obj = NULL, *pobj = NULL;
196 	struct got_object_id *obj_id = NULL, *pobj_id = NULL;
197 	struct got_commit_object *commit = NULL;
198 	struct got_blob_object *blob = NULL, *pblob = NULL;
199 	struct got_diff_changes *changes = NULL;
200 	struct got_object_qid *pid = NULL;
201 
202 	err = got_object_open_as_commit(&commit, repo, id);
203 	if (err)
204 		return err;
205 
206 	err = got_object_id_by_path(&obj_id, repo, id, path);
207 	if (err)
208 		goto done;
209 
210 	err = got_object_open(&obj, repo, obj_id);
211 	if (err)
212 		goto done;
213 
214 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
215 		err = got_error(GOT_ERR_OBJ_TYPE);
216 		goto done;
217 	}
218 
219 	pid = SIMPLEQ_FIRST(got_object_commit_get_parent_ids(commit));
220 	if (pid) {
221 		err = got_object_id_by_path(&pobj_id, repo, pid->id, path);
222 		if (err) {
223 			if (err->code == GOT_ERR_NO_TREE_ENTRY) {
224 				/* Blob's history began in previous commit. */
225 				err = got_error(GOT_ERR_ITER_COMPLETED);
226 			}
227 			goto done;
228 		}
229 
230 		/* If IDs match then don't bother with diffing. */
231 		if (got_object_id_cmp(obj_id, pobj_id) == 0) {
232 			if (cb)
233 				err = cb(arg, blame->nlines, -1, id);
234 			goto done;
235 		}
236 
237 		err = got_object_open(&pobj, repo, pobj_id);
238 		if (err)
239 			goto done;
240 
241 		if (pobj->type != GOT_OBJ_TYPE_BLOB) {
242 			/*
243 			 * Encountered a non-blob at the path (probably a tree).
244 			 * Blob's history began in previous commit.
245 			 */
246 			err = got_error(GOT_ERR_ITER_COMPLETED);
247 			goto done;
248 		}
249 
250 		err = got_object_blob_open(&pblob, repo, pobj, 8192);
251 		if (err)
252 			goto done;
253 	}
254 
255 	err = got_object_blob_open(&blob, repo, obj, 8192);
256 	if (err)
257 		goto done;
258 
259 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
260 	if (err)
261 		goto done;
262 
263 	if (changes) {
264 		err = blame_changes(blame, changes, id, cb, arg);
265 		got_diff_free_changes(changes);
266 	} else if (cb)
267 		err = cb(arg, blame->nlines, -1, id);
268 done:
269 	if (commit)
270 		got_object_commit_close(commit);
271 	free(obj_id);
272 	free(pobj_id);
273 	if (obj)
274 		got_object_close(obj);
275 	if (pobj)
276 		got_object_close(pobj);
277 	if (blob)
278 		got_object_blob_close(blob);
279 	if (pblob)
280 		got_object_blob_close(pblob);
281 	return err;
282 }
283 
284 static const struct got_error *
285 blame_close(struct got_blame *blame)
286 {
287 	const struct got_error *err = NULL;
288 	struct got_blame_diff_offsets *diff_offsets;
289 
290 	if (blame->f && fclose(blame->f) != 0)
291 		err = got_error_from_errno("fclose");
292 	free(blame->lines);
293 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
294 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
295 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
296 		free_diff_offsets(diff_offsets);
297 	}
298 	free(blame);
299 	return err;
300 }
301 
302 static const struct got_error *
303 blame_open(struct got_blame **blamep, const char *path,
304     struct got_object_id *start_commit_id, struct got_repository *repo,
305     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
306     void *arg)
307 {
308 	const struct got_error *err = NULL;
309 	struct got_object *obj = NULL;
310 	struct got_object_id *obj_id = NULL;
311 	struct got_blob_object *blob = NULL;
312 	struct got_blame *blame = NULL;
313 	struct got_object_id *id = NULL, *next_id = NULL;
314 	int lineno;
315 	struct got_commit_graph *graph = NULL;
316 
317 	*blamep = NULL;
318 
319 	err = got_object_id_by_path(&obj_id, repo, start_commit_id, path);
320 	if (err)
321 		return err;
322 
323 	err = got_object_open(&obj, repo, obj_id);
324 	if (err)
325 		goto done;
326 
327 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
328 		err = got_error(GOT_ERR_OBJ_TYPE);
329 		goto done;
330 	}
331 
332 	err = got_object_blob_open(&blob, repo, obj, 8192);
333 	if (err)
334 		goto done;
335 
336 	blame = calloc(1, sizeof(*blame));
337 	if (blame == NULL)
338 		return got_error_from_errno("calloc");
339 
340 	blame->f = got_opentemp();
341 	if (blame->f == NULL) {
342 		err = got_error_from_errno("got_opentemp");
343 		goto done;
344 	}
345 	err = got_object_blob_dump_to_file(&blame->filesize, &blame->nlines,
346 	   &blame->line_offsets, blame->f, blob);
347 	if (err)
348 		goto done;
349 
350 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
351 	if (blame->lines == NULL) {
352 		err = got_error_from_errno("calloc");
353 		goto done;
354 	}
355 
356 	err = got_commit_graph_open(&graph, start_commit_id, path, 1, repo);
357 	if (err)
358 		return err;
359 	err = got_commit_graph_iter_start(graph, start_commit_id, repo);
360 	if (err)
361 		goto done;
362 
363 	id = NULL;
364 	for (;;) {
365 		err = got_commit_graph_iter_next(&next_id, graph);
366 		if (err) {
367 			if (err->code == GOT_ERR_ITER_COMPLETED) {
368 				if (id)
369 					err = blame_commit(blame, id,
370 					    path, repo, cb, arg);
371 				else
372 					err = NULL;
373 				break;
374 			}
375 			if (err->code != GOT_ERR_ITER_NEED_MORE)
376 				break;
377 			err = got_commit_graph_fetch_commits(graph, 1, repo);
378 			if (err)
379 				break;
380 			continue;
381 		}
382 		if (id) {
383 			err = blame_commit(blame, id, path, repo,
384 			    cb, arg);
385 			if (err) {
386 				if (err->code == GOT_ERR_ITER_COMPLETED)
387 					err = NULL;
388 				break;
389 			}
390 			if (blame->nannotated == blame->nlines)
391 				break;
392 		}
393 		id = next_id;
394 	}
395 
396 	if (id && blame->nannotated < blame->nlines) {
397 		/* Annotate remaining non-annotated lines with last commit. */
398 		for (lineno = 1; lineno <= blame->nlines; lineno++) {
399 			err = annotate_line(blame, lineno, id, cb, arg);
400 			if (err)
401 				goto done;
402 		}
403 	}
404 
405 done:
406 	if (graph)
407 		got_commit_graph_close(graph);
408 	free(obj_id);
409 	if (obj)
410 		got_object_close(obj);
411 	if (blob)
412 		got_object_blob_close(blob);
413 	if (err) {
414 		if (blame)
415 			blame_close(blame);
416 	} else
417 		*blamep = blame;
418 
419 	return err;
420 }
421 
422 const struct got_error *
423 got_blame(const char *path, struct got_object_id *commit_id,
424     struct got_repository *repo,
425     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
426     void *arg)
427 {
428 	const struct got_error *err = NULL, *close_err = NULL;
429 	struct got_blame *blame;
430 	char *abspath;
431 
432 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
433 		return got_error_from_errno2("asprintf", path);
434 
435 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
436 	free(abspath);
437 	if (blame)
438 		close_err = blame_close(blame);
439 	return err ? err : close_err;
440 }