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_zbuf.h"
34 #include "got_lib_delta.h"
35 #include "got_lib_object.h"
36 #include "got_lib_diff.h"
37 
38 struct got_blame_line {
39 	int annotated;
40 	struct got_object_id id;
41 };
42 
43 struct got_blame {
44 	FILE *f;
45 	size_t nlines;
46 	struct got_blame_line *lines; /* one per line */
47 };
48 
49 static const struct got_error *
50 dump_blob_and_count_lines(size_t *nlines, FILE *outfile,
51     struct got_blob_object *blob)
52 {
53 	const struct got_error *err = NULL;
54 	size_t len, hdrlen;
55 	const uint8_t *buf;
56 	int i;
57 
58 	hdrlen = got_object_blob_get_hdrlen(blob);
59 	*nlines = 0;
60 	do {
61 		err = got_object_blob_read_block(&len, blob);
62 		if (err)
63 			return err;
64 		if (len == 0)
65 			break;
66 		buf = got_object_blob_get_read_buf(blob);
67 		for (i = 0; i < len; i++) {
68 			if (buf[i] == '\n')
69 				(*nlines)++;
70 		}
71 		/* Skip blob object header first time around. */
72 		fwrite(buf + hdrlen, len - hdrlen, 1, outfile);
73 		hdrlen = 0;
74 	} while (len != 0);
75 
76 
77 	fflush(outfile);
78 	rewind(outfile);
79 
80 	return NULL;
81 }
82 
83 static void
84 annotate_line(struct got_blame *blame, int lineno, struct got_object_id *id)
85 {
86 	struct got_blame_line *line;
87 
88 	if (lineno < 1 || lineno > blame->nlines)
89 		return;
90 	
91 	line = &blame->lines[lineno - 1];
92 	if (line->annotated)
93 		return;
94 
95 	memcpy(&line->id, id, sizeof(line->id));
96 	line->annotated = 1;
97 }
98 
99 static const struct got_error *
100 blame_commit(struct got_blame *blame, struct got_object_id *id,
101     struct got_object_id *pid, const char *path, struct got_repository *repo)
102 {
103 	const struct got_error *err = NULL;
104 	struct got_object *obj = NULL, *pobj = NULL;
105 	struct got_blob_object *blob = NULL, *pblob = NULL;
106 	struct got_diff_changes *changes = NULL;
107 
108 	err = got_object_open_by_path(&obj, repo, id, path);
109 	if (err)
110 		goto done;
111 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
112 		err = got_error(GOT_ERR_OBJ_TYPE);
113 		goto done;
114 	}
115 
116 	err = got_object_open_by_path(&pobj, repo, pid, path);
117 	if (err) {
118 		if (err->code == GOT_ERR_NO_OBJ) {
119 			/* Blob's history began in previous commit. */
120 			err = got_error(GOT_ERR_ITER_COMPLETED);
121 		}
122 		goto done;
123 	}
124 	if (got_object_get_type(pobj) != GOT_OBJ_TYPE_BLOB) {
125 		/*
126 		 * Encountered a non-blob at the path (probably a tree).
127 		 * Blob's history began in previous commit.
128 		 */
129 		err = got_error(GOT_ERR_ITER_COMPLETED);
130 		goto done;
131 	}
132 
133 	/* If blob hashes match then don't bother with diffing. */
134 	if (got_object_id_cmp(&obj->id, &pobj->id) == 0)
135 		goto done;
136 
137 	err = got_object_blob_open(&blob, repo, obj, 8192);
138 	if (err)
139 		goto done;
140 
141 	err = got_object_blob_open(&pblob, repo, pobj, 8192);
142 	if (err)
143 		goto done;
144 
145 	err = got_diff_blob_lines_changed(&changes, blob, pblob);
146 	if (err)
147 		goto done;
148 
149 	if (changes) {
150 		struct got_diff_change *change;
151 		SIMPLEQ_FOREACH(change, &changes->entries, entry) {
152 			int a = change->cv.a;
153 			int b = change->cv.b;
154 			int lineno;
155 			for (lineno = a; lineno <= b; lineno++)
156 				annotate_line(blame, lineno, id);
157 		}
158 	}
159 done:
160 	if (obj)
161 		got_object_close(obj);
162 	if (pobj)
163 		got_object_close(pobj);
164 	if (blob)
165 		got_object_blob_close(blob);
166 	if (pblob)
167 		got_object_blob_close(pblob);
168 	return err;
169 }
170 
171 static void
172 blame_close(struct got_blame *blame)
173 {
174 	if (blame->f)
175 		fclose(blame->f);
176 	free(blame->lines);
177 	free(blame);
178 }
179 
180 static const struct got_error *
181 blame_open(struct got_blame **blamep, const char *path,
182     struct got_object_id *start_commit_id, struct got_repository *repo)
183 {
184 	const struct got_error *err = NULL;
185 	struct got_object *obj = NULL;
186 	struct got_blob_object *blob = NULL;
187 	struct got_blame *blame = NULL;
188 	struct got_commit_object *commit = NULL;
189 	struct got_object_id *id = NULL;
190 	int lineno;
191 
192 	*blamep = NULL;
193 
194 	err = got_object_open_by_path(&obj, repo, start_commit_id, path);
195 	if (err)
196 		return err;
197 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_BLOB) {
198 		err = got_error(GOT_ERR_OBJ_TYPE);
199 		goto done;
200 	}
201 
202 	err = got_object_blob_open(&blob, repo, obj, 8192);
203 	if (err)
204 		goto done;
205 
206 	blame = calloc(1, sizeof(*blame));
207 	if (blame == NULL)
208 		return got_error_from_errno();
209 
210 	blame->f = got_opentemp();
211 	if (blame->f == NULL) {
212 		err = got_error_from_errno();
213 		goto done;
214 	}
215 	err = dump_blob_and_count_lines(&blame->nlines, blame->f, blob);
216 	if (err)
217 		goto done;
218 
219 	blame->lines = calloc(blame->nlines, sizeof(*blame->lines));
220 	if (blame->lines == NULL) {
221 		err = got_error_from_errno();
222 		goto done;
223 	}
224 
225 	/* Loop over first-parent history and try to blame commits. */
226 	err = got_object_open_as_commit(&commit, repo, start_commit_id);
227 	if (err)
228 		goto done;
229 	id = got_object_id_dup(start_commit_id);
230 	if (id == NULL) {
231 		err = got_error_from_errno();
232 		goto done;
233 	}
234 	while (1) {
235 		struct got_object_qid *pid;
236 
237 		pid = SIMPLEQ_FIRST(&commit->parent_ids);
238 		if (pid == NULL)
239 			break;
240 
241 		err = blame_commit(blame, id, pid->id, path, repo);
242 		if (err) {
243 			if (err->code == GOT_ERR_ITER_COMPLETED)
244 				err = NULL;
245 			break;
246 		}
247 
248 		free(id);
249 		id = got_object_id_dup(pid->id);
250 		if (id == NULL) {
251 			err = got_error_from_errno();
252 			goto done;
253 		}
254 		got_object_commit_close(commit);
255 		err = got_object_open_as_commit(&commit, repo, id);
256 		if (err)
257 			break;
258 	}
259 
260 	/* Annotate remaining non-annotated lines with last commit. */
261 	for (lineno = 1; lineno <= blame->nlines; lineno++)
262 		annotate_line(blame, lineno, id);
263 
264 done:
265 	free(id);
266 	if (obj)
267 		got_object_close(obj);
268 	if (blob)
269 		got_object_blob_close(blob);
270 	if (commit)
271 		got_object_commit_close(commit);
272 	if (err)
273 		blame_close(blame);
274 	else
275 		*blamep = blame;
276 
277 	return err;
278 }
279 
280 static const struct got_error *
281 blame_line(struct got_object_id **id, struct got_blame *blame, int lineno)
282 {
283 	if (lineno < 1 || lineno > blame->nlines)
284 		return got_error(GOT_ERR_RANGE);
285 	*id = &blame->lines[lineno - 1].id;
286 	return NULL;
287 }
288 
289 static char *
290 parse_next_line(FILE *f, size_t *len)
291 {
292 	char *line;
293 	size_t linelen;
294 	size_t lineno;
295 	const char delim[3] = { '\0', '\0', '\0'};
296 
297 	line = fparseln(f, &linelen, &lineno, delim, 0);
298 	if (len)
299 		*len = linelen;
300 	return line;
301 }
302 
303 const struct got_error *
304 got_blame(const char *path, struct got_object_id *start_commit_id,
305     struct got_repository *repo, FILE *outfile)
306 {
307 	const struct got_error *err = NULL;
308 	struct got_blame *blame;
309 	int lineno;
310 	char *abspath;
311 
312 	if (asprintf(&abspath, "%s%s", path[0] == '/' ? "" : "/", path) == -1)
313 		return got_error_from_errno();
314 
315 	err = blame_open(&blame, abspath, start_commit_id, repo);
316 	if (err) {
317 		free(abspath);
318 		return err;
319 	}
320 
321 	for (lineno = 1; lineno <= blame->nlines; lineno++) {
322 		struct got_object_id *id;
323 		char *line, *id_str;
324 		
325 		line = parse_next_line(blame->f, NULL);
326 		if (line == NULL)
327 			break;
328 
329 		err = blame_line(&id, blame, lineno);
330 		if (err)
331 			break;
332 
333 		err = got_object_id_str(&id_str, id);
334 		if (err) {
335 			free(line);
336 			break;
337 		}
338 
339 		fprintf(outfile, "%.8s %s\n", id_str, line);
340 		free(line);
341 		free(id_str);
342 	}
343 
344 	blame_close(blame);
345 	free(abspath);
346 	return err;
347 }