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_blob_object *blob = NULL, *pblob = NULL;
190 	struct got_diff_changes *changes = NULL;
191 
192 	err = got_object_open_by_path(&obj, repo, id, path);
193 	if (err)
194 		goto done;
195 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
196 		err = got_error(GOT_ERR_OBJ_TYPE);
197 		goto done;
198 	}
199 
200 	err = got_object_open_by_path(&pobj, repo, pid, path);
201 	if (err) {
202 		if (err->code == GOT_ERR_NO_OBJ) {
203 			/* Blob's history began in previous commit. */
204 			err = got_error(GOT_ERR_ITER_COMPLETED);
205 		}
206 		goto done;
207 	}
208 	if (got_object_get_type(pobj) != GOT_OBJ_TYPE_BLOB) {
209 		/*
210 		 * Encountered a non-blob at the path (probably a tree).
211 		 * Blob's history began in previous commit.
212 		 */
213 		err = got_error(GOT_ERR_ITER_COMPLETED);
214 		goto done;
215 	}
216 
217 	/* If blob hashes match then don't bother with diffing. */
218 	if (got_object_id_cmp(&obj->id, &pobj->id) == 0) {
219 		if (cb)
220 			err = cb(arg, blame->nlines, -1, id);
221 		goto done;
222 	}
223 
224 	err = got_object_blob_open(&blob, repo, obj, 8192);
225 	if (err)
226 		goto done;
227 
228 	err = got_object_blob_open(&pblob, repo, pobj, 8192);
229 	if (err)
230 		goto done;
231 
232 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
233 	if (err)
234 		goto done;
235 
236 	if (changes) {
237 		err = blame_changes(blame, changes, id, cb, arg);
238 		got_diff_free_changes(changes);
239 	} else if (cb)
240 		err = cb(arg, blame->nlines, -1, id);
241 done:
242 	if (obj)
243 		got_object_close(obj);
244 	if (pobj)
245 		got_object_close(pobj);
246 	if (blob)
247 		got_object_blob_close(blob);
248 	if (pblob)
249 		got_object_blob_close(pblob);
250 	return err;
251 }
252 
253 static void
254 blame_close(struct got_blame *blame)
255 {
256 	struct got_blame_diff_offsets *diff_offsets;
257 
258 	if (blame->f)
259 		fclose(blame->f);
260 	free(blame->lines);
261 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
262 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
263 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
264 		free_diff_offsets(diff_offsets);
265 	}
266 	free(blame);
267 }
268 
269 static const struct got_error *
270 blame_open(struct got_blame **blamep, const char *path,
271     struct got_object_id *start_commit_id, struct got_repository *repo,
272     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
273     void *arg)
274 {
275 	const struct got_error *err = NULL;
276 	struct got_object *obj = NULL;
277 	struct got_blob_object *blob = NULL;
278 	struct got_blame *blame = NULL;
279 	struct got_commit_object *commit = NULL;
280 	struct got_object_id *id = NULL;
281 	int lineno;
282 
283 	*blamep = NULL;
284 
285 	err = got_object_open_by_path(&obj, repo, start_commit_id, path);
286 	if (err)
287 		return err;
288 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
289 		err = got_error(GOT_ERR_OBJ_TYPE);
290 		goto done;
291 	}
292 
293 	err = got_object_blob_open(&blob, repo, obj, 8192);
294 	if (err)
295 		goto done;
296 
297 	blame = calloc(1, sizeof(*blame));
298 	if (blame == NULL)
299 		return got_error_from_errno();
300 
301 	blame->f = got_opentemp();
302 	if (blame->f == NULL) {
303 		err = got_error_from_errno();
304 		goto done;
305 	}
306 	err = got_object_blob_dump_to_file(NULL, &blame->nlines, blame->f,
307 	    blob);
308 	if (err)
309 		goto done;
310 
311 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
312 	if (blame->lines == NULL) {
313 		err = got_error_from_errno();
314 		goto done;
315 	}
316 
317 	/* Loop over first-parent history and try to blame commits. */
318 	err = got_object_open_as_commit(&commit, repo, start_commit_id);
319 	if (err)
320 		goto done;
321 	id = got_object_id_dup(start_commit_id);
322 	if (id == NULL) {
323 		err = got_error_from_errno();
324 		goto done;
325 	}
326 	while (1) {
327 		struct got_object_qid *pid;
328 
329 		pid = SIMPLEQ_FIRST(&commit->parent_ids);
330 		if (pid == NULL)
331 			break;
332 
333 		err = blame_commit(blame, id, pid->id, path, repo, cb, arg);
334 		if (err) {
335 			if (err->code == GOT_ERR_ITER_COMPLETED)
336 				err = NULL;
337 			break;
338 		}
339 
340 		free(id);
341 		id = got_object_id_dup(pid->id);
342 		if (id == NULL) {
343 			err = got_error_from_errno();
344 			goto done;
345 		}
346 		got_object_commit_close(commit);
347 		err = got_object_open_as_commit(&commit, repo, id);
348 		if (err)
349 			goto done;
350 	}
351 
352 	/* Annotate remaining non-annotated lines with last commit. */
353 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
354 		err = annotate_line(blame, lineno, id, cb, arg);
355 		if (err)
356 			goto done;
357 	}
358 
359 done:
360 	free(id);
361 	if (obj)
362 		got_object_close(obj);
363 	if (blob)
364 		got_object_blob_close(blob);
365 	if (commit)
366 		got_object_commit_close(commit);
367 	if (err) {
368 		if (blame)
369 			blame_close(blame);
370 	} else
371 		*blamep = blame;
372 
373 	return err;
374 }
375 
376 static const struct got_error *
377 blame_line(struct got_object_id **id, struct got_blame *blame, int lineno)
378 {
379 	if (lineno < 1 || lineno > blame->nlines)
380 		return got_error(GOT_ERR_RANGE);
381 	*id = &blame->lines[lineno - 1].id;
382 	return NULL;
383 }
384 
385 static char *
386 parse_next_line(FILE *f, size_t *len)
387 {
388 	char *line;
389 	size_t linelen;
390 	size_t lineno;
391 	const char delim[3] = { '\0', '\0', '\0'};
392 
393 	line = fparseln(f, &linelen, &lineno, delim, 0);
394 	if (len)
395 		*len = linelen;
396 	return line;
397 }
398 
399 const struct got_error *
400 got_blame(const char *path, struct got_object_id *start_commit_id,
401     struct got_repository *repo, FILE *outfile)
402 {
403 	const struct got_error *err = NULL;
404 	struct got_blame *blame;
405 	int lineno;
406 	char *abspath;
407 
408 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
409 		return got_error_from_errno();
410 
411 	err = blame_open(&blame, abspath, start_commit_id, repo, NULL, NULL);
412 	if (err) {
413 		free(abspath);
414 		return err;
415 	}
416 
417 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
418 		struct got_object_id *id;
419 		char *line, *id_str;
420 		
421 		line = parse_next_line(blame->f, NULL);
422 		if (line == NULL)
423 			break;
424 
425 		err = blame_line(&id, blame, lineno);
426 		if (err) {
427 			free(line);
428 			break;
429 		}
430 
431 		err = got_object_id_str(&id_str, id);
432 		/* Do not free id; It points into blame->lines. */
433 		if (err) {
434 			free(line);
435 			break;
436 		}
437 
438 		fprintf(outfile, "%.8s %s\n", id_str, line);
439 		free(line);
440 		free(id_str);
441 	}
442 
443 	blame_close(blame);
444 	free(abspath);
445 	return err;
446 }
447 
448 const struct got_error *
449 got_blame_incremental(const char *path, struct got_object_id *commit_id,
450     struct got_repository *repo,
451     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
452     void *arg)
453 {
454 	const struct got_error *err = NULL;
455 	struct got_blame *blame;
456 	char *abspath;
457 
458 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
459 		return got_error_from_errno();
460 
461 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
462 	free(abspath);
463 	if (blame)
464 		blame_close(blame);
465 	return err;
466 }