op public repos

Blob

Date:: Wed Apr 20 14:00:12 2022 UTC
Message:: inline struct got_object_id in struct got_object_qid Saves us from doing a malloc/free call for every item on the list. ok op@
Actions:: History | Blame | Raw File
1 /*
2  * Copyright (c) 2018, 2019, 2020 Stefan Sperling <stsp@openbsd.org>
3  * Copyright (c) 2020 Neels Hofmeyr <neels@hofmeyr.de>
4  *
5  * Permission to use, copy, modify, and distribute this software for any
6  * purpose with or without fee is hereby granted, provided that the above
7  * copyright notice and this permission notice appear in all copies.
8  *
9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
16  */
17 
18 #include <sys/queue.h>
19 #include <sys/mman.h>
20 #include <sys/stat.h>
21 
22 #include <errno.h>
23 #include <sha1.h>
24 #include <string.h>
25 #include <stdio.h>
26 #include <stdlib.h>
27 #include <time.h>
28 #include <limits.h>
29 #include <util.h>
30 #include <zlib.h>
31 
32 #include "got_error.h"
33 #include "got_object.h"
34 #include "got_cancel.h"
35 #include "got_blame.h"
36 #include "got_commit_graph.h"
37 #include "got_opentemp.h"
38 
39 #include "got_lib_inflate.h"
40 #include "got_lib_delta.h"
41 #include "got_lib_object.h"
42 #include "got_lib_diff.h"
43 
44 #ifndef MAX
45 #define	MAX(_a,_b) ((_a) > (_b) ? (_a) : (_b))
46 #endif
47 
48 struct got_blame_line {
49 	int annotated;
50 	struct got_object_id id;
51 };
52 
53 struct got_blame {
54 	struct diff_config *cfg;
55 	int nlines;	/* number of lines in file being blamed */
56 	int nannotated;	/* number of lines already annotated */
57 	struct got_blame_line *lines; /* one per line */
58 	int ncommits;
59 
60 	/*
61 	 * These change with every traversed commit. After diffing
62 	 * commits N:N-1, in preparation for diffing commits N-1:N-2,
63 	 * data for commit N is retained and flipped into data for N-1.
64 	 * 
65 	 */
66 	FILE *f1; /* older version from commit N-1. */
67 	FILE *f2; /* newer version from commit N. */
68 	unsigned char *map1;
69 	unsigned char *map2;
70 	off_t size1;
71 	off_t size2;
72 	int nlines1;
73 	int nlines2;
74 	off_t *line_offsets1;
75 	off_t *line_offsets2;
76 
77 	/*
78 	 * Map line numbers of an older version of the file to valid line
79 	 * numbers in the version of the file being blamed. This map is
80 	 * updated with each commit we traverse throughout the file's history.
81 	 * Lines mapped to -1 do not correspond to any line in the version
82 	 * being blamed.
83 	 */
84 	int *linemap1;
85 	int *linemap2;
86 
87 	struct diff_data *data1;
88 	struct diff_data *data2;
89 };
90 
91 static const struct got_error *
92 annotate_line(struct got_blame *blame, int lineno,
93     struct got_commit_object *commit, struct got_object_id *id,
94     got_blame_cb cb, void *arg)
95 {
96 	const struct got_error *err = NULL;
97 	struct got_blame_line *line;
98 
99 	if (lineno < 0 || lineno >= blame->nlines)
100 		return NULL;
101 
102 	line = &blame->lines[lineno];
103 	if (line->annotated)
104 		return NULL;
105 
106 	memcpy(&line->id, id, sizeof(line->id));
107 	line->annotated = 1;
108 	blame->nannotated++;
109 	if (cb)
110 		err = cb(arg, blame->nlines, lineno + 1, commit, id);
111 	return err;
112 }
113 
114 static const struct got_error *
115 blame_changes(struct got_blame *blame, struct diff_result *diff_result,
116     struct got_commit_object *commit, struct got_object_id *commit_id,
117     got_blame_cb cb, void *arg)
118 {
119 	const struct got_error *err = NULL;
120 	int i;
121 	int idx1 = 0, idx2 = 0;
122 
123 	for (i = 0; i < diff_result->chunks.len &&
124 	    blame->nannotated < blame->nlines; i++) {
125 		struct diff_chunk *c = diff_chunk_get(diff_result, i);
126 		unsigned int left_count, right_count;
127 		int j;
128 
129 		/*
130 		 * We do not need to worry about idx1/idx2 growing out
131 		 * of bounds because the diff implementation ensures
132 		 * that chunk ranges never exceed the number of lines
133 		 * in the left/right input files.
134 		 */
135 		left_count = diff_chunk_get_left_count(c);
136 		right_count = diff_chunk_get_right_count(c);
137 
138 		if (left_count == right_count) {
139 			for (j = 0; j < left_count; j++) {
140 				blame->linemap1[idx1++] =
141 				    blame->linemap2[idx2++];
142 			}
143 			continue;
144 		}
145 
146 		if (right_count == 0) {
147 			for (j = 0; j < left_count; j++) {
148 				blame->linemap1[idx1++] = -1;
149 			}
150 			continue;
151 		}
152 
153 		for (j = 0; j < right_count; j++) {
154 			int ln = blame->linemap2[idx2++];
155 			err = annotate_line(blame, ln, commit, commit_id,
156 			    cb, arg);
157 			if (err)
158 				return err;
159 			if (blame->nlines == blame->nannotated)
160 				break;
161 		}
162 	}
163 
164 	return NULL;
165 }
166 
167 static const struct got_error *
168 blame_prepare_file(FILE *f, unsigned char **p, off_t *size,
169     int *nlines, off_t **line_offsets, struct diff_data *diff_data,
170     const struct diff_config *cfg, struct got_blob_object *blob)
171 {
172 	const struct got_error *err = NULL;
173 	int diff_flags = 0, rc;
174 
175 	err = got_object_blob_dump_to_file(size, nlines, line_offsets,
176 	    f, blob);
177 	if (err)
178 		return err;
179 
180 #ifndef GOT_DIFF_NO_MMAP
181 	*p = mmap(NULL, *size, PROT_READ, MAP_PRIVATE, fileno(f), 0);
182 	if (*p == MAP_FAILED)
183 #endif
184 		*p = NULL; /* fall back on file I/O */
185 
186 	/* Allow blaming lines in binary files even though it's useless. */
187 	diff_flags |= DIFF_FLAG_FORCE_TEXT_DATA;
188 
189 	rc = diff_atomize_file(diff_data, cfg, f, *p, *size, diff_flags);
190 	if (rc)
191 		return got_error_set_errno(rc, "diff_atomize_file");
192 
193 	return NULL;
194 }
195 
196 static const struct got_error *
197 blame_commit(struct got_blame *blame, struct got_object_id *id,
198     const char *path, struct got_repository *repo,
199     got_blame_cb cb, void *arg)
200 {
201 	const struct got_error *err = NULL;
202 	struct got_commit_object *commit = NULL, *pcommit = NULL;
203 	struct got_object_qid *pid = NULL;
204 	struct got_object_id *pblob_id = NULL;
205 	struct got_blob_object *pblob = NULL;
206 	struct diff_result *diff_result = NULL;
207 
208 	err = got_object_open_as_commit(&commit, repo, id);
209 	if (err)
210 		return err;
211 
212 	pid = STAILQ_FIRST(got_object_commit_get_parent_ids(commit));
213 	if (pid == NULL) {
214 		got_object_commit_close(commit);
215 		return NULL;
216 	}
217 
218 	err = got_object_open_as_commit(&pcommit, repo, &pid->id);
219 	if (err)
220 		goto done;
221 
222 	err = got_object_id_by_path(&pblob_id, repo, pcommit, path);
223 	if (err) {
224 		if (err->code == GOT_ERR_NO_TREE_ENTRY)
225 			err = NULL;
226 		goto done;
227 	}
228 
229 	err = got_object_open_as_blob(&pblob, repo, pblob_id, 8192);
230 	if (err)
231 		goto done;
232 
233 	blame->f1 = got_opentemp();
234 	if (blame->f1 == NULL) {
235 		err = got_error_from_errno("got_opentemp");
236 		goto done;
237 	}
238 
239 	err = blame_prepare_file(blame->f1, &blame->map1, &blame->size1,
240 	    &blame->nlines1, &blame->line_offsets1, blame->data1,
241 	    blame->cfg, pblob);
242 	if (err)
243 		goto done;
244 
245 	diff_result = diff_main(blame->cfg, blame->data1, blame->data2);
246 	if (diff_result == NULL) {
247 		err = got_error_set_errno(ENOMEM, "malloc");
248 		goto done;
249 	}
250 	if (diff_result->rc != DIFF_RC_OK) {
251 		err = got_error_set_errno(diff_result->rc, "diff");
252 		goto done;
253 	}
254 	if (diff_result->chunks.len > 0) {
255 		if (blame->nlines1 > 0) {
256 			blame->linemap1 = calloc(blame->nlines1,
257 			    sizeof(*blame->linemap1));
258 			if (blame->linemap1 == NULL) {
259 				err = got_error_from_errno("malloc");
260 				goto done;
261 			}
262 		}
263 		err = blame_changes(blame, diff_result, commit, id, cb, arg);
264 		if (err)
265 			goto done;
266 	} else if (cb)
267 		err = cb(arg, blame->nlines, -1, commit, id);
268 done:
269 	if (diff_result)
270 		diff_result_free(diff_result);
271 	if (commit)
272 		got_object_commit_close(commit);
273 	if (pcommit)
274 		got_object_commit_close(pcommit);
275 	free(pblob_id);
276 	if (pblob)
277 		got_object_blob_close(pblob);
278 	return err;
279 }
280 
281 static const struct got_error *
282 blame_close(struct got_blame *blame)
283 {
284 	const struct got_error *err = NULL;
285 
286 	diff_data_free(blame->data1);
287 	free(blame->data1);
288 	diff_data_free(blame->data2);
289 	free(blame->data2);
290 	if (blame->map1) {
291 		if (munmap(blame->map1, blame->size1) == -1 && err == NULL)
292 			err = got_error_from_errno("munmap");
293 	}
294 	if (blame->map2) {
295 		if (munmap(blame->map2, blame->size2) == -1 && err == NULL)
296 			err = got_error_from_errno("munmap");
297 	}
298 	if (blame->f1 && fclose(blame->f1) == EOF && err == NULL)
299 		err = got_error_from_errno("fclose");
300 	if (blame->f2 && fclose(blame->f2) == EOF && err == NULL)
301 		err = got_error_from_errno("fclose");
302 	free(blame->lines);
303 	free(blame->line_offsets1);
304 	free(blame->line_offsets2);
305 	free(blame->linemap1);
306 	free(blame->linemap2);
307 	free(blame->cfg);
308 	free(blame);
309 	return err;
310 }
311 
312 static int
313 atomize_file(struct diff_data *d, FILE *f, off_t filesize, int nlines,
314     off_t *line_offsets)
315 {
316 	int i, rc = DIFF_RC_OK;
317 	int embedded_nul = 0;
318 
319 	ARRAYLIST_INIT(d->atoms, nlines);
320 
321 	for (i = 0; i < nlines; i++) {
322 		struct diff_atom *atom;
323 		off_t len, pos = line_offsets[i];
324 		unsigned int hash = 0;
325 		int j;
326 
327 		ARRAYLIST_ADD(atom, d->atoms);
328 		if (atom == NULL) {
329 			rc = errno;
330 			break;
331 		}
332 
333 		if (i < nlines - 1)
334 			len = line_offsets[i + 1] - pos;
335 		else
336 			len = filesize - pos;
337 
338 		if (fseeko(f, pos, SEEK_SET) == -1) {
339 			rc = errno;
340 			break;
341 		}
342 		for (j = 0; j < len; j++) {
343 			int c = fgetc(f);
344 			if (c == EOF) {
345 				if (feof(f))
346 					rc = EIO; /* unexpected EOF */
347 				else
348 					rc = errno;
349 				goto done;
350 			}
351 
352 			hash = diff_atom_hash_update(hash, (unsigned char)c);
353 
354 			if (c == '\0')
355 				embedded_nul = 1;
356 
357 		}
358 		*atom = (struct diff_atom){
359 			.root = d,
360 			.pos = pos,
361 			.at = NULL,	/* atom data is not memory-mapped */
362 			.len = len,
363 			.hash = hash,
364 		};
365 	}
366 
367 	/* File are considered binary if they contain embedded '\0' bytes. */
368 	if (embedded_nul)
369 		d->atomizer_flags |= DIFF_ATOMIZER_FOUND_BINARY_DATA;
370 done:
371 	if (rc)
372 		ARRAYLIST_FREE(d->atoms);
373 
374 	return rc;
375 }
376 
377 static int
378 atomize_file_mmap(struct diff_data *d, unsigned char *p,
379     off_t filesize, int nlines, off_t *line_offsets)
380 {
381 	int i, rc = DIFF_RC_OK;
382 	int embedded_nul = 0;
383 
384 	ARRAYLIST_INIT(d->atoms, nlines);
385 
386 	for (i = 0; i < nlines; i++) {
387 		struct diff_atom *atom;
388 		off_t len, pos = line_offsets[i];
389 		unsigned int hash = 0;
390 		int j;
391 
392 		ARRAYLIST_ADD(atom, d->atoms);
393 		if (atom == NULL) {
394 			rc = errno;
395 			break;
396 		}
397 
398 		if (i < nlines - 1)
399 			len = line_offsets[i + 1] - pos;
400 		else
401 			len = filesize - pos;
402 
403 		for (j = 0; j < len; j++)
404 			hash = diff_atom_hash_update(hash, p[pos + j]);
405 
406 		if (!embedded_nul && memchr(&p[pos], '\0', len) != NULL)
407 			embedded_nul = 1;
408 
409 		*atom = (struct diff_atom){
410 			.root = d,
411 			.pos = pos,
412 			.at = &p[pos],
413 			.len = len,
414 			.hash = hash,
415 		};
416 	}
417 
418 	/* File are considered binary if they contain embedded '\0' bytes. */
419 	if (embedded_nul)
420 		d->atomizer_flags |= DIFF_ATOMIZER_FOUND_BINARY_DATA;
421 
422 	if (rc)
423 		ARRAYLIST_FREE(d->atoms);
424 
425 	return rc;
426 }
427 
428 /* Implements diff_atomize_func_t */
429 static int
430 blame_atomize_file(void *arg, struct diff_data *d)
431 {
432 	struct got_blame *blame = arg;
433 
434 	if (d->f == blame->f1) {
435 		if (blame->map1)
436 			return atomize_file_mmap(d, blame->map1,
437 			    blame->size1, blame->nlines1,
438 			    blame->line_offsets1);
439 		else
440 			return atomize_file(d, blame->f1, blame->size1,
441 			    blame->nlines1, blame->line_offsets1);
442 	} else if (d->f == blame->f2) {
443 		if (d->atoms.len > 0) {
444 			/* Re-use data from previous commit. */
445 			return DIFF_RC_OK;
446 		}
447 		if (blame->map2)
448 			return atomize_file_mmap(d, blame->map2,
449 			    blame->size2, blame->nlines2,
450 			    blame->line_offsets2);
451 		else
452 			return atomize_file(d, blame->f2, blame->size2,
453 			    blame->nlines2, blame->line_offsets2);
454 	}
455 
456 	return DIFF_RC_OK;
457 }
458 
459 static const struct got_error *
460 close_file2_and_reuse_file1(struct got_blame *blame)
461 {
462 	struct diff_data *d;
463 
464 	free(blame->line_offsets2);
465 	blame->line_offsets2 = blame->line_offsets1;
466 	blame->line_offsets1 = NULL;
467 
468 	free(blame->linemap2);
469 	blame->linemap2 = blame->linemap1;
470 	blame->linemap1 = NULL;
471 
472 	if (blame->map2) {
473 		if (munmap(blame->map2, blame->size2) == -1)
474 			return got_error_from_errno("munmap");
475 		blame->map2 = blame->map1;
476 		blame->map1 = NULL;
477 
478 	}
479 	blame->size2 = blame->size1;
480 	blame->size1 = 0;
481 
482 	if (fclose(blame->f2) == EOF)
483 		return got_error_from_errno("fclose");
484 	blame->f2 = blame->f1;
485 	blame->f1 = NULL;
486 
487 	blame->nlines2 = blame->nlines1;
488 	blame->nlines1 = 0;
489 
490 	diff_data_free(blame->data2); /* does not free pointer itself */
491 	memset(blame->data2, 0, sizeof(*blame->data2));
492 	d = blame->data2;
493 	blame->data2 = blame->data1;
494 	blame->data1 = d;
495 
496 	return NULL;
497 }
498 
499 static const struct got_error *
500 blame_open(struct got_blame **blamep, const char *path,
501     struct got_object_id *start_commit_id, struct got_repository *repo,
502     got_blame_cb cb, void *arg, got_cancel_cb cancel_cb, void *cancel_arg)
503 {
504 	const struct got_error *err = NULL;
505 	struct got_commit_object *start_commit = NULL, *last_commit = NULL;
506 	struct got_object_id *obj_id = NULL;
507 	struct got_blob_object *blob = NULL;
508 	struct got_blame *blame = NULL;
509 	struct got_object_id *id = NULL;
510 	int lineno;
511 	struct got_commit_graph *graph = NULL;
512 
513 	*blamep = NULL;
514 
515 	err = got_object_open_as_commit(&start_commit, repo, start_commit_id);
516 	if (err)
517 		goto done;
518 
519 	err = got_object_id_by_path(&obj_id, repo, start_commit, path);
520 	if (err)
521 		goto done;
522 
523 	err = got_object_open_as_blob(&blob, repo, obj_id, 8192);
524 	if (err)
525 		goto done;
526 
527 	blame = calloc(1, sizeof(*blame));
528 	if (blame == NULL) {
529 		err = got_error_from_errno("calloc");
530 		goto done;
531 	}
532 
533 	blame->data1 = calloc(1, sizeof(*blame->data1));
534 	if (blame->data1 == NULL) {
535 		err = got_error_from_errno("calloc");
536 		goto done;
537 	}
538 	blame->data2 = calloc(1, sizeof(*blame->data2));
539 	if (blame->data2 == NULL) {
540 		err = got_error_from_errno("calloc");
541 		goto done;
542 	}
543 
544 	blame->f2 = got_opentemp();
545 	if (blame->f2 == NULL) {
546 		err = got_error_from_errno("got_opentemp");
547 		goto done;
548 	}
549 	err = got_diff_get_config(&blame->cfg, GOT_DIFF_ALGORITHM_PATIENCE,
550 	    blame_atomize_file, blame);
551 	if (err)
552 		goto done;
553 
554 	err = blame_prepare_file(blame->f2, &blame->map2, &blame->size2,
555 	    &blame->nlines2, &blame->line_offsets2, blame->data2,
556 	    blame->cfg, blob);
557 	blame->nlines = blame->nlines2;
558 	if (err || blame->nlines == 0)
559 		goto done;
560 
561 	got_object_blob_close(blob);
562 	blob = NULL;
563 
564 	/* Don't include \n at EOF in the blame line count. */
565 	if (blame->line_offsets2[blame->nlines - 1] == blame->size2)
566 		blame->nlines--;
567 
568 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
569 	if (blame->lines == NULL) {
570 		err = got_error_from_errno("calloc");
571 		goto done;
572 	}
573 
574 	blame->linemap2 = calloc(blame->nlines2, sizeof(*blame->linemap2));
575 	if (blame->linemap2 == NULL) {
576 		err = got_error_from_errno("calloc");
577 		goto done;
578 	}
579 	for (lineno = 0; lineno < blame->nlines2; lineno++)
580 		blame->linemap2[lineno] = lineno;
581 
582 	err = got_commit_graph_open(&graph, path, 1);
583 	if (err)
584 		goto done;
585 
586 	err = got_commit_graph_iter_start(graph, start_commit_id, repo,
587 	    cancel_cb, cancel_arg);
588 	if (err)
589 		goto done;
590 	for (;;) {
591 		struct got_object_id *next_id;
592 		err = got_commit_graph_iter_next(&next_id, graph, repo,
593 		    cancel_cb, cancel_arg);
594 		if (err) {
595 			if (err->code == GOT_ERR_ITER_COMPLETED) {
596 				err = NULL;
597 				break;
598 			}
599 			goto done;
600 		}
601 		if (next_id) {
602 			id = next_id;
603 			err = blame_commit(blame, id, path, repo, cb, arg);
604 			if (err) {
605 				if (err->code == GOT_ERR_ITER_COMPLETED)
606 					err = NULL;
607 				goto done;
608 			}
609 			if (blame->nannotated == blame->nlines)
610 				break;
611 
612 			err = close_file2_and_reuse_file1(blame);
613 			if (err)
614 				goto done;
615 		}
616 	}
617 
618 	if (id && blame->nannotated < blame->nlines) {
619 		/* Annotate remaining non-annotated lines with last commit. */
620 		err = got_object_open_as_commit(&last_commit, repo, id);
621 		if (err)
622 			goto done;
623 		for (lineno = 0; lineno < blame->nlines; lineno++) {
624 			err = annotate_line(blame, lineno, last_commit, id,
625 			    cb, arg);
626 			if (err)
627 				goto done;
628 		}
629 	}
630 
631 done:
632 	if (graph)
633 		got_commit_graph_close(graph);
634 	free(obj_id);
635 	if (blob)
636 		got_object_blob_close(blob);
637 	if (start_commit)
638 		got_object_commit_close(start_commit);
639 	if (last_commit)
640 		got_object_commit_close(last_commit);
641 	if (err) {
642 		if (blame)
643 			blame_close(blame);
644 	} else
645 		*blamep = blame;
646 
647 	return err;
648 }
649 
650 const struct got_error *
651 got_blame(const char *path, struct got_object_id *commit_id,
652     struct got_repository *repo, got_blame_cb cb, void *arg,
653     got_cancel_cb cancel_cb, void* cancel_arg)
654 {
655 	const struct got_error *err = NULL, *close_err = NULL;
656 	struct got_blame *blame;
657 	char *abspath;
658 
659 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
660 		return got_error_from_errno2("asprintf", path);
661 
662 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg,
663 	    cancel_cb, cancel_arg);
664 	free(abspath);
665 	if (blame)
666 		close_err = blame_close(blame);
667 	return err ? err : close_err;
668 }