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 
39 struct got_blame_line {
40 	int annotated;
41 	struct got_object_id id;
42 };
43 
44 struct got_blame_diff_offsets {
45 	struct got_diffoffset_chunks *chunks;
46 	struct got_object_id *commit_id;
47 	SLIST_ENTRY(got_blame_diff_offsets) entry;
48 };
49 
50 SLIST_HEAD(got_blame_diff_offsets_list, got_blame_diff_offsets);
51 
52 struct got_blame {
53 	FILE *f;
54 	size_t nlines;
55 	struct got_blame_line *lines; /* one per line */
56 	int ncommits;
57 	struct got_blame_diff_offsets_list diff_offsets_list;
58 };
59 
60 static void
61 free_diff_offsets(struct got_blame_diff_offsets *diff_offsets)
62 {
63 	if (diff_offsets->chunks)
64 		got_diffoffset_free(diff_offsets->chunks);
65 	free(diff_offsets->commit_id);
66 	free(diff_offsets);
67 }
68 
69 static const struct got_error *
70 alloc_diff_offsets(struct got_blame_diff_offsets **diff_offsets,
71     struct got_object_id *commit_id)
72 {
73 	const struct got_error *err = NULL;
74 
75 	*diff_offsets = calloc(1, sizeof(**diff_offsets));
76 	if (*diff_offsets == NULL)
77 		return got_error_from_errno();
78 
79 	(*diff_offsets)->commit_id = got_object_id_dup(commit_id);
80 	if ((*diff_offsets)->commit_id == NULL) {
81 		err = got_error_from_errno();
82 		free_diff_offsets(*diff_offsets);
83 		*diff_offsets = NULL;
84 		return err;
85 	}
86 
87 	err = got_diffoffset_alloc(&(*diff_offsets)->chunks);
88 	if (err) {
89 		free_diff_offsets(*diff_offsets);
90 		return err;
91 	}
92 
93 	return NULL;
94 }
95 
96 static const struct got_error *
97 annotate_line(struct got_blame *blame, int lineno, struct got_object_id *id,
98     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
99     void *arg)
100 {
101 	const struct got_error *err = NULL;
102 	struct got_blame_line *line;
103 
104 	if (lineno < 1 || lineno > blame->nlines)
105 		return got_error(GOT_ERR_RANGE);
106 	
107 	line = &blame->lines[lineno - 1];
108 	if (line->annotated)
109 		return NULL;
110 
111 	memcpy(&line->id, id, sizeof(line->id));
112 	line->annotated = 1;
113 	if (cb)
114 		err = cb(arg, blame->nlines, lineno, id);
115 	return err;
116 }
117 
118 static int
119 get_blamed_line(struct got_blame_diff_offsets_list *diff_offsets_list,
120     int lineno)
121 {
122 	struct got_blame_diff_offsets *diff_offsets;
123 
124 	SLIST_FOREACH(diff_offsets, diff_offsets_list, entry)
125 		lineno = got_diffoffset_get(diff_offsets->chunks, lineno);
126 
127 	return lineno;
128 }
129 
130 static const struct got_error *
131 blame_changes(struct got_blame *blame, struct got_diff_changes *changes,
132     struct got_object_id *commit_id,
133     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
134     void *arg)
135 {
136 	const struct got_error *err = NULL;
137 	struct got_diff_change *change;
138 	struct got_blame_diff_offsets *diff_offsets;
139 
140 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
141 		int c = change->cv.c;
142 		int d = change->cv.d;
143 		int new_lineno = c;
144 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
145 		int ln;
146 
147 		for (ln = new_lineno; ln < new_lineno + new_length; ln++) {
148 			err = annotate_line(blame,
149 			    get_blamed_line(&blame->diff_offsets_list, ln),
150 			    commit_id, cb, arg);
151 			if (err)
152 				return err;
153 		}
154 	}
155 
156 	err = alloc_diff_offsets(&diff_offsets, commit_id);
157 	if (err)
158 		return err;
159 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
160 		int a = change->cv.a;
161 		int b = change->cv.b;
162 		int c = change->cv.c;
163 		int d = change->cv.d;
164 		int old_lineno = a;
165 		int old_length = (a < b ? b - a + 1 : (a == b ? 1 : 0));
166 		int new_lineno = c;
167 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
168 
169 		err = got_diffoffset_add(diff_offsets->chunks,
170 		    old_lineno, old_length, new_lineno, new_length);
171 		if (err) {
172 			free_diff_offsets(diff_offsets);
173 			return err;
174 		}
175 	}
176 	SLIST_INSERT_HEAD(&blame->diff_offsets_list, diff_offsets, entry);
177 
178 	return NULL;
179 }
180 
181 static const struct got_error *
182 blame_commit(struct got_blame *blame, struct got_object_id *id,
183     struct got_object_id *pid, const char *path, struct got_repository *repo,
184     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
185     void *arg)
186 {
187 	const struct got_error *err = NULL;
188 	struct got_object *obj = NULL, *pobj = NULL;
189 	struct got_object_id *obj_id = NULL, *pobj_id = NULL;
190 	struct got_blob_object *blob = NULL, *pblob = NULL;
191 	struct got_diff_changes *changes = NULL;
192 
193 	err = got_object_id_by_path(&obj_id, repo, id, path);
194 	if (err)
195 		goto done;
196 
197 	err = got_object_open(&obj, repo, obj_id);
198 	if (err)
199 		goto done;
200 
201 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
202 		err = got_error(GOT_ERR_OBJ_TYPE);
203 		goto done;
204 	}
205 
206 	err = got_object_id_by_path(&pobj_id, repo, pid, path);
207 	if (err) {
208 		if (err->code == GOT_ERR_NO_OBJ) {
209 			/* Blob's history began in previous commit. */
210 			err = got_error(GOT_ERR_ITER_COMPLETED);
211 		}
212 		goto done;
213 	}
214 
215 	/* If IDs match then don't bother with diffing. */
216 	if (got_object_id_cmp(obj_id, pobj_id) == 0) {
217 		if (cb)
218 			err = cb(arg, blame->nlines, -1, id);
219 		goto done;
220 	}
221 
222 	err = got_object_open(&pobj, repo, pobj_id);
223 	if (err)
224 		goto done;
225 
226 	if (got_object_get_type(pobj) != GOT_OBJ_TYPE_BLOB) {
227 		/*
228 		 * Encountered a non-blob at the path (probably a tree).
229 		 * Blob's history began in previous commit.
230 		 */
231 		err = got_error(GOT_ERR_ITER_COMPLETED);
232 		goto done;
233 	}
234 
235 	err = got_object_blob_open(&blob, repo, obj, 8192);
236 	if (err)
237 		goto done;
238 
239 	err = got_object_blob_open(&pblob, repo, pobj, 8192);
240 	if (err)
241 		goto done;
242 
243 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
244 	if (err)
245 		goto done;
246 
247 	if (changes) {
248 		err = blame_changes(blame, changes, id, cb, arg);
249 		got_diff_free_changes(changes);
250 	} else if (cb)
251 		err = cb(arg, blame->nlines, -1, id);
252 done:
253 	free(obj_id);
254 	free(pobj_id);
255 	if (obj)
256 		got_object_close(obj);
257 	if (pobj)
258 		got_object_close(pobj);
259 	if (blob)
260 		got_object_blob_close(blob);
261 	if (pblob)
262 		got_object_blob_close(pblob);
263 	return err;
264 }
265 
266 static void
267 blame_close(struct got_blame *blame)
268 {
269 	struct got_blame_diff_offsets *diff_offsets;
270 
271 	if (blame->f)
272 		fclose(blame->f);
273 	free(blame->lines);
274 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
275 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
276 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
277 		free_diff_offsets(diff_offsets);
278 	}
279 	free(blame);
280 }
281 
282 static const struct got_error *
283 blame_open(struct got_blame **blamep, const char *path,
284     struct got_object_id *start_commit_id, struct got_repository *repo,
285     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
286     void *arg)
287 {
288 	const struct got_error *err = NULL;
289 	struct got_object *obj = NULL;
290 	struct got_object_id *obj_id = NULL;
291 	struct got_blob_object *blob = NULL;
292 	struct got_blame *blame = NULL;
293 	struct got_commit_object *commit = NULL;
294 	struct got_object_id *id = NULL;
295 	int lineno;
296 
297 	*blamep = NULL;
298 
299 	err = got_object_id_by_path(&obj_id, repo, start_commit_id, path);
300 	if (err)
301 		return err;
302 
303 	err = got_object_open(&obj, repo, obj_id);
304 	if (err)
305 		goto done;
306 
307 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
308 		err = got_error(GOT_ERR_OBJ_TYPE);
309 		goto done;
310 	}
311 
312 	err = got_object_blob_open(&blob, repo, obj, 8192);
313 	if (err)
314 		goto done;
315 
316 	blame = calloc(1, sizeof(*blame));
317 	if (blame == NULL)
318 		return got_error_from_errno();
319 
320 	blame->f = got_opentemp();
321 	if (blame->f == NULL) {
322 		err = got_error_from_errno();
323 		goto done;
324 	}
325 	err = got_object_blob_dump_to_file(NULL, &blame->nlines, blame->f,
326 	    blob);
327 	if (err)
328 		goto done;
329 
330 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
331 	if (blame->lines == NULL) {
332 		err = got_error_from_errno();
333 		goto done;
334 	}
335 
336 	/* Loop over first-parent history and try to blame commits. */
337 	/* TODO: Iterate commits via commit graph instead. */
338 	err = got_object_open_as_commit(&commit, repo, start_commit_id);
339 	if (err)
340 		goto done;
341 	id = got_object_id_dup(start_commit_id);
342 	if (id == NULL) {
343 		err = got_error_from_errno();
344 		goto done;
345 	}
346 	while (1) {
347 		struct got_object_qid *pid;
348 
349 		pid = SIMPLEQ_FIRST(&commit->parent_ids);
350 		if (pid == NULL)
351 			break;
352 
353 		err = blame_commit(blame, id, pid->id, path, repo, cb, arg);
354 		if (err) {
355 			if (err->code == GOT_ERR_ITER_COMPLETED)
356 				err = NULL;
357 			break;
358 		}
359 
360 		free(id);
361 		id = got_object_id_dup(pid->id);
362 		if (id == NULL) {
363 			err = got_error_from_errno();
364 			goto done;
365 		}
366 		got_object_commit_close(commit);
367 		err = got_object_open_as_commit(&commit, repo, id);
368 		if (err)
369 			goto done;
370 	}
371 
372 	/* Annotate remaining non-annotated lines with last commit. */
373 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
374 		err = annotate_line(blame, lineno, id, cb, arg);
375 		if (err)
376 			goto done;
377 	}
378 
379 done:
380 	free(obj_id);
381 	if (obj)
382 		got_object_close(obj);
383 	if (blob)
384 		got_object_blob_close(blob);
385 	if (commit)
386 		got_object_commit_close(commit);
387 	if (err) {
388 		if (blame)
389 			blame_close(blame);
390 	} else
391 		*blamep = blame;
392 
393 	return err;
394 }
395 
396 static const struct got_error *
397 blame_line(struct got_object_id **id, struct got_blame *blame, int lineno)
398 {
399 	if (lineno < 1 || lineno > blame->nlines)
400 		return got_error(GOT_ERR_RANGE);
401 	*id = &blame->lines[lineno - 1].id;
402 	return NULL;
403 }
404 
405 static char *
406 parse_next_line(FILE *f, size_t *len)
407 {
408 	char *line;
409 	size_t linelen;
410 	size_t lineno;
411 	const char delim[3] = { '\0', '\0', '\0'};
412 
413 	line = fparseln(f, &linelen, &lineno, delim, 0);
414 	if (len)
415 		*len = linelen;
416 	return line;
417 }
418 
419 const struct got_error *
420 got_blame(const char *path, struct got_object_id *start_commit_id,
421     struct got_repository *repo, FILE *outfile)
422 {
423 	const struct got_error *err = NULL;
424 	struct got_blame *blame;
425 	int lineno;
426 	char *abspath;
427 
428 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
429 		return got_error_from_errno();
430 
431 	err = blame_open(&blame, abspath, start_commit_id, repo, NULL, NULL);
432 	if (err) {
433 		free(abspath);
434 		return err;
435 	}
436 
437 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
438 		struct got_object_id *id;
439 		char *line, *id_str;
440 		
441 		line = parse_next_line(blame->f, NULL);
442 		if (line == NULL)
443 			break;
444 
445 		err = blame_line(&id, blame, lineno);
446 		if (err) {
447 			free(line);
448 			break;
449 		}
450 
451 		err = got_object_id_str(&id_str, id);
452 		/* Do not free id; It points into blame->lines. */
453 		if (err) {
454 			free(line);
455 			break;
456 		}
457 
458 		fprintf(outfile, "%.8s %s\n", id_str, line);
459 		free(line);
460 		free(id_str);
461 	}
462 
463 	blame_close(blame);
464 	free(abspath);
465 	return err;
466 }
467 
468 const struct got_error *
469 got_blame_incremental(const char *path, struct got_object_id *commit_id,
470     struct got_repository *repo,
471     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
472     void *arg)
473 {
474 	const struct got_error *err = NULL;
475 	struct got_blame *blame;
476 	char *abspath;
477 
478 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
479 		return got_error_from_errno();
480 
481 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
482 	free(abspath);
483 	if (blame)
484 		blame_close(blame);
485 	return err;
486 }