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 			return err;
173 	}
174 	SLIST_INSERT_HEAD(&blame->diff_offsets_list, diff_offsets, entry);
175 
176 	return NULL;
177 }
178 
179 static const struct got_error *
180 blame_commit(struct got_blame *blame, struct got_object_id *id,
181     struct got_object_id *pid, const char *path, struct got_repository *repo,
182     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
183     void *arg)
184 {
185 	const struct got_error *err = NULL;
186 	struct got_object *obj = NULL, *pobj = NULL;
187 	struct got_blob_object *blob = NULL, *pblob = NULL;
188 	struct got_diff_changes *changes = NULL;
189 
190 	err = got_object_open_by_path(&obj, repo, id, path);
191 	if (err)
192 		goto done;
193 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
194 		err = got_error(GOT_ERR_OBJ_TYPE);
195 		goto done;
196 	}
197 
198 	err = got_object_open_by_path(&pobj, repo, pid, path);
199 	if (err) {
200 		if (err->code == GOT_ERR_NO_OBJ) {
201 			/* Blob's history began in previous commit. */
202 			err = got_error(GOT_ERR_ITER_COMPLETED);
203 		}
204 		goto done;
205 	}
206 	if (got_object_get_type(pobj) != GOT_OBJ_TYPE_BLOB) {
207 		/*
208 		 * Encountered a non-blob at the path (probably a tree).
209 		 * Blob's history began in previous commit.
210 		 */
211 		err = got_error(GOT_ERR_ITER_COMPLETED);
212 		goto done;
213 	}
214 
215 	/* If blob hashes 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_blob_open(&blob, repo, obj, 8192);
223 	if (err)
224 		goto done;
225 
226 	err = got_object_blob_open(&pblob, repo, pobj, 8192);
227 	if (err)
228 		goto done;
229 
230 	err = got_diff_blob_lines_changed(&changes, pblob, blob);
231 	if (err)
232 		goto done;
233 
234 	if (changes) {
235 		err = blame_changes(blame, changes, id, cb, arg);
236 	} else if (cb)
237 		err = cb(arg, blame->nlines, -1, id);
238 done:
239 	if (obj)
240 		got_object_close(obj);
241 	if (pobj)
242 		got_object_close(pobj);
243 	if (blob)
244 		got_object_blob_close(blob);
245 	if (pblob)
246 		got_object_blob_close(pblob);
247 	return err;
248 }
249 
250 static void
251 blame_close(struct got_blame *blame)
252 {
253 	struct got_blame_diff_offsets *diff_offsets;
254 
255 	if (blame->f)
256 		fclose(blame->f);
257 	free(blame->lines);
258 	while (!SLIST_EMPTY(&blame->diff_offsets_list)) {
259 		diff_offsets = SLIST_FIRST(&blame->diff_offsets_list);
260 		SLIST_REMOVE_HEAD(&blame->diff_offsets_list, entry);
261 		free_diff_offsets(diff_offsets);
262 	}
263 	free(blame);
264 }
265 
266 static const struct got_error *
267 blame_open(struct got_blame **blamep, const char *path,
268     struct got_object_id *start_commit_id, struct got_repository *repo,
269     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
270     void *arg)
271 {
272 	const struct got_error *err = NULL;
273 	struct got_object *obj = NULL;
274 	struct got_blob_object *blob = NULL;
275 	struct got_blame *blame = NULL;
276 	struct got_commit_object *commit = NULL;
277 	struct got_object_id *id = NULL;
278 	int lineno;
279 
280 	*blamep = NULL;
281 
282 	err = got_object_open_by_path(&obj, repo, start_commit_id, path);
283 	if (err)
284 		return err;
285 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
286 		err = got_error(GOT_ERR_OBJ_TYPE);
287 		goto done;
288 	}
289 
290 	err = got_object_blob_open(&blob, repo, obj, 8192);
291 	if (err)
292 		goto done;
293 
294 	blame = calloc(1, sizeof(*blame));
295 	if (blame == NULL)
296 		return got_error_from_errno();
297 
298 	blame->f = got_opentemp();
299 	if (blame->f == NULL) {
300 		err = got_error_from_errno();
301 		goto done;
302 	}
303 	err = got_object_blob_dump_to_file(NULL, &blame->nlines, blame->f,
304 	    blob);
305 	if (err)
306 		goto done;
307 
308 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
309 	if (blame->lines == NULL) {
310 		err = got_error_from_errno();
311 		goto done;
312 	}
313 
314 	/* Loop over first-parent history and try to blame commits. */
315 	err = got_object_open_as_commit(&commit, repo, start_commit_id);
316 	if (err)
317 		goto done;
318 	id = got_object_id_dup(start_commit_id);
319 	if (id == NULL) {
320 		err = got_error_from_errno();
321 		goto done;
322 	}
323 	while (1) {
324 		struct got_object_qid *pid;
325 
326 		pid = SIMPLEQ_FIRST(&commit->parent_ids);
327 		if (pid == NULL)
328 			break;
329 
330 		err = blame_commit(blame, id, pid->id, path, repo, cb, arg);
331 		if (err) {
332 			if (err->code == GOT_ERR_ITER_COMPLETED)
333 				err = NULL;
334 			break;
335 		}
336 
337 		free(id);
338 		id = got_object_id_dup(pid->id);
339 		if (id == NULL) {
340 			err = got_error_from_errno();
341 			goto done;
342 		}
343 		got_object_commit_close(commit);
344 		err = got_object_open_as_commit(&commit, repo, id);
345 		if (err)
346 			goto done;
347 	}
348 
349 	/* Annotate remaining non-annotated lines with last commit. */
350 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
351 		err = annotate_line(blame, lineno, id, cb, arg);
352 		if (err)
353 			goto done;
354 	}
355 
356 done:
357 	free(id);
358 	if (obj)
359 		got_object_close(obj);
360 	if (blob)
361 		got_object_blob_close(blob);
362 	if (commit)
363 		got_object_commit_close(commit);
364 	if (err) {
365 		if (blame)
366 			blame_close(blame);
367 	} else
368 		*blamep = blame;
369 
370 	return err;
371 }
372 
373 static const struct got_error *
374 blame_line(struct got_object_id **id, struct got_blame *blame, int lineno)
375 {
376 	if (lineno < 1 || lineno > blame->nlines)
377 		return got_error(GOT_ERR_RANGE);
378 	*id = &blame->lines[lineno - 1].id;
379 	return NULL;
380 }
381 
382 static char *
383 parse_next_line(FILE *f, size_t *len)
384 {
385 	char *line;
386 	size_t linelen;
387 	size_t lineno;
388 	const char delim[3] = { '\0', '\0', '\0'};
389 
390 	line = fparseln(f, &linelen, &lineno, delim, 0);
391 	if (len)
392 		*len = linelen;
393 	return line;
394 }
395 
396 const struct got_error *
397 got_blame(const char *path, struct got_object_id *start_commit_id,
398     struct got_repository *repo, FILE *outfile)
399 {
400 	const struct got_error *err = NULL;
401 	struct got_blame *blame;
402 	int lineno;
403 	char *abspath;
404 
405 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
406 		return got_error_from_errno();
407 
408 	err = blame_open(&blame, abspath, start_commit_id, repo, NULL, NULL);
409 	if (err) {
410 		free(abspath);
411 		return err;
412 	}
413 
414 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
415 		struct got_object_id *id;
416 		char *line, *id_str;
417 		
418 		line = parse_next_line(blame->f, NULL);
419 		if (line == NULL)
420 			break;
421 
422 		err = blame_line(&id, blame, lineno);
423 		if (err)
424 			break;
425 
426 		err = got_object_id_str(&id_str, id);
427 		if (err) {
428 			free(line);
429 			break;
430 		}
431 
432 		fprintf(outfile, "%.8s %s\n", id_str, line);
433 		free(line);
434 		free(id_str);
435 	}
436 
437 	blame_close(blame);
438 	free(abspath);
439 	return err;
440 }
441 
442 const struct got_error *
443 got_blame_incremental(const char *path, struct got_object_id *commit_id,
444     struct got_repository *repo,
445     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
446     void *arg)
447 {
448 	const struct got_error *err = NULL;
449 	struct got_blame *blame;
450 	char *abspath;
451 
452 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
453 		return got_error_from_errno();
454 
455 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg);
456 	free(abspath);
457 	if (blame)
458 		blame_close(blame);
459 	return err;
460 }