1 /*
2  * Copyright (c) 2018, 2019, 2020 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 <limits.h>
26 #include <util.h>
27 #include <zlib.h>
28 
29 #include "got_error.h"
30 #include "got_object.h"
31 #include "got_cancel.h"
32 #include "got_blame.h"
33 #include "got_commit_graph.h"
34 #include "got_opentemp.h"
35 
36 #include "got_lib_inflate.h"
37 #include "got_lib_delta.h"
38 #include "got_lib_object.h"
39 #include "got_lib_diff.h"
40 
41 struct got_blame_line {
42 	int annotated;
43 	struct got_object_id id;
44 };
45 
46 struct got_blame {
47 	FILE *f;
48 	size_t filesize;
49 	int nlines;
50 	int nannotated;
51 	struct got_blame_line *lines; /* one per line */
52 	off_t *line_offsets;		/* one per line */
53 	int ncommits;
54 };
55 
56 static const struct got_error *
57 annotate_line(struct got_blame *blame, int lineno, struct got_object_id *id,
58     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
59     void *arg)
60 {
61 	const struct got_error *err = NULL;
62 	struct got_blame_line *line;
63 
64 	if (lineno < 1 || lineno > blame->nlines)
65 		return NULL;
66 	
67 	line = &blame->lines[lineno - 1];
68 	if (line->annotated)
69 		return NULL;
70 
71 	memcpy(&line->id, id, sizeof(line->id));
72 	line->annotated = 1;
73 	blame->nannotated++;
74 	if (cb)
75 		err = cb(arg, blame->nlines, lineno, id);
76 	return err;
77 }
78 
79 static const struct got_error *
80 blame_changes(struct got_blame *blame, struct got_diff_changes *changes,
81     struct got_object_id *commit_id,
82     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
83     void *arg)
84 {
85 	const struct got_error *err = NULL;
86 	struct got_diff_change *change;
87 
88 	SIMPLEQ_FOREACH(change, &changes->entries, entry) {
89 		int c = change->cv.c;
90 		int d = change->cv.d;
91 		int new_lineno = (c < d ? c : d);
92 		int new_length = (c < d ? d - c + 1 : (c == d ? 1 : 0));
93 		int ln;
94 
95 		for (ln = new_lineno; ln < new_lineno + new_length; ln++) {
96 			err = annotate_line(blame, ln, commit_id, cb, arg);
97 			if (err)
98 				return err;
99 			if (blame->nlines == blame->nannotated)
100 				break;
101 		}
102 	}
103 
104 	return NULL;
105 }
106 
107 static const struct got_error *
108 blame_commit(struct got_blame *blame, struct got_object_id *parent_id,
109     struct got_object_id *id, const char *path, struct got_repository *repo,
110     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
111     void *arg)
112 {
113 	const struct got_error *err = NULL;
114 	struct got_object *obj = NULL;
115 	struct got_object_id *obj_id = NULL;
116 	struct got_commit_object *commit = NULL;
117 	struct got_blob_object *blob = NULL;
118 	struct got_diff_changes *changes = NULL;
119 
120 	err = got_object_open_as_commit(&commit, repo, parent_id);
121 	if (err)
122 		return err;
123 
124 	err = got_object_id_by_path(&obj_id, repo, parent_id, path);
125 	if (err) {
126 		if (err->code == GOT_ERR_NO_TREE_ENTRY)
127 			err = NULL;
128 		goto done;
129 	}
130 
131 	err = got_object_open(&obj, repo, obj_id);
132 	if (err)
133 		goto done;
134 
135 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
136 		err = got_error(GOT_ERR_OBJ_TYPE);
137 		goto done;
138 	}
139 
140 	err = got_object_blob_open(&blob, repo, obj, 8192);
141 	if (err)
142 		goto done;
143 
144 	if (fseek(blame->f, 0L, SEEK_SET) == -1) {
145 		err = got_ferror(blame->f, GOT_ERR_IO);
146 		goto done;
147 	}
148 
149 	err = got_diff_blob_file_lines_changed(&changes, blob, blame->f,
150 	    blame->filesize);
151 	if (err)
152 		goto done;
153 
154 	if (changes) {
155 		err = blame_changes(blame, changes, id, cb, arg);
156 		got_diff_free_changes(changes);
157 	} else if (cb)
158 		err = cb(arg, blame->nlines, -1, id);
159 done:
160 	if (commit)
161 		got_object_commit_close(commit);
162 	free(obj_id);
163 	if (obj)
164 		got_object_close(obj);
165 	if (blob)
166 		got_object_blob_close(blob);
167 	return err;
168 }
169 
170 static const struct got_error *
171 blame_close(struct got_blame *blame)
172 {
173 	const struct got_error *err = NULL;
174 
175 	if (blame->f && fclose(blame->f) != 0)
176 		err = got_error_from_errno("fclose");
177 	free(blame->lines);
178 	free(blame);
179 	return err;
180 }
181 
182 static const struct got_error *
183 blame_open(struct got_blame **blamep, const char *path,
184     struct got_object_id *start_commit_id, struct got_repository *repo,
185     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
186     void *arg, got_cancel_cb cancel_cb, void *cancel_arg)
187 {
188 	const struct got_error *err = NULL;
189 	struct got_object *obj = NULL;
190 	struct got_object_id *obj_id = NULL;
191 	struct got_blob_object *blob = NULL;
192 	struct got_blame *blame = NULL;
193 	struct got_object_id *id = NULL, *pid = NULL;
194 	int lineno;
195 	struct got_commit_graph *graph = NULL;
196 
197 	*blamep = NULL;
198 
199 	err = got_object_id_by_path(&obj_id, repo, start_commit_id, path);
200 	if (err)
201 		return err;
202 
203 	err = got_object_open(&obj, repo, obj_id);
204 	if (err)
205 		goto done;
206 
207 	if (obj->type != GOT_OBJ_TYPE_BLOB) {
208 		err = got_error(GOT_ERR_OBJ_TYPE);
209 		goto done;
210 	}
211 
212 	err = got_object_blob_open(&blob, repo, obj, 8192);
213 	if (err)
214 		goto done;
215 
216 	blame = calloc(1, sizeof(*blame));
217 	if (blame == NULL)
218 		return got_error_from_errno("calloc");
219 
220 	blame->f = got_opentemp();
221 	if (blame->f == NULL) {
222 		err = got_error_from_errno("got_opentemp");
223 		goto done;
224 	}
225 	err = got_object_blob_dump_to_file(&blame->filesize, &blame->nlines,
226 	   &blame->line_offsets, blame->f, blob);
227 	if (err || blame->nlines == 0)
228 		goto done;
229 
230 	/* Don't include \n at EOF in the blame line count. */
231 	if (blame->line_offsets[blame->nlines - 1] == blame->filesize)
232 		blame->nlines--;
233 
234 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
235 	if (blame->lines == NULL) {
236 		err = got_error_from_errno("calloc");
237 		goto done;
238 	}
239 
240 	err = got_commit_graph_open(&graph, path, 1);
241 	if (err)
242 		return err;
243 	err = got_commit_graph_iter_start(graph, start_commit_id, repo,
244 	    cancel_cb, cancel_arg);
245 	if (err)
246 		goto done;
247 	id = start_commit_id;
248 	for (;;) {
249 		err = got_commit_graph_iter_next(&pid, graph, repo,
250 		    cancel_cb, cancel_arg);
251 		if (err) {
252 			if (err->code == GOT_ERR_ITER_COMPLETED)
253 				err = NULL;
254 			break;
255 		}
256 		if (pid) {
257 			err = blame_commit(blame, pid, id, path, repo, cb, arg);
258 			if (err) {
259 				if (err->code == GOT_ERR_ITER_COMPLETED)
260 					err = NULL;
261 				break;
262 			}
263 			if (blame->nannotated == blame->nlines)
264 				break;
265 		}
266 		id = pid;
267 	}
268 
269 	if (id && blame->nannotated < blame->nlines) {
270 		/* Annotate remaining non-annotated lines with last commit. */
271 		for (lineno = 1; lineno <= blame->nlines; lineno++) {
272 			err = annotate_line(blame, lineno, id, cb, arg);
273 			if (err)
274 				goto done;
275 		}
276 	}
277 
278 done:
279 	if (graph)
280 		got_commit_graph_close(graph);
281 	free(obj_id);
282 	if (obj)
283 		got_object_close(obj);
284 	if (blob)
285 		got_object_blob_close(blob);
286 	if (err) {
287 		if (blame)
288 			blame_close(blame);
289 	} else
290 		*blamep = blame;
291 
292 	return err;
293 }
294 
295 const struct got_error *
296 got_blame(const char *path, struct got_object_id *commit_id,
297     struct got_repository *repo,
298     const struct got_error *(*cb)(void *, int, int, struct got_object_id *),
299     void *arg, got_cancel_cb cancel_cb, void* cancel_arg)
300 {
301 	const struct got_error *err = NULL, *close_err = NULL;
302 	struct got_blame *blame;
303 	char *abspath;
304 
305 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
306 		return got_error_from_errno2("asprintf", path);
307 
308 	err = blame_open(&blame, abspath, commit_id, repo, cb, arg,
309 	    cancel_cb, cancel_arg);
310 	free(abspath);
311 	if (blame)
312 		close_err = blame_close(blame);
313 	return err ? err : close_err;
314 }