tree-diff: remove the use of pathspec's raw[] in follow-rename codepath
[git/git.git] / tree-diff.c
1 /*
2 * Helper functions for tree diff generation
3 */
4 #include "cache.h"
5 #include "diff.h"
6 #include "diffcore.h"
7 #include "tree.h"
8
9 static void show_entry(struct diff_options *opt, const char *prefix,
10 struct tree_desc *desc, struct strbuf *base);
11
12 static int compare_tree_entry(struct tree_desc *t1, struct tree_desc *t2,
13 struct strbuf *base, struct diff_options *opt)
14 {
15 unsigned mode1, mode2;
16 const char *path1, *path2;
17 const unsigned char *sha1, *sha2;
18 int cmp, pathlen1, pathlen2;
19 int old_baselen = base->len;
20
21 sha1 = tree_entry_extract(t1, &path1, &mode1);
22 sha2 = tree_entry_extract(t2, &path2, &mode2);
23
24 pathlen1 = tree_entry_len(&t1->entry);
25 pathlen2 = tree_entry_len(&t2->entry);
26 cmp = base_name_compare(path1, pathlen1, mode1, path2, pathlen2, mode2);
27 if (cmp < 0) {
28 show_entry(opt, "-", t1, base);
29 return -1;
30 }
31 if (cmp > 0) {
32 show_entry(opt, "+", t2, base);
33 return 1;
34 }
35 if (!DIFF_OPT_TST(opt, FIND_COPIES_HARDER) && !hashcmp(sha1, sha2) && mode1 == mode2)
36 return 0;
37
38 /*
39 * If the filemode has changed to/from a directory from/to a regular
40 * file, we need to consider it a remove and an add.
41 */
42 if (S_ISDIR(mode1) != S_ISDIR(mode2)) {
43 show_entry(opt, "-", t1, base);
44 show_entry(opt, "+", t2, base);
45 return 0;
46 }
47
48 strbuf_add(base, path1, pathlen1);
49 if (DIFF_OPT_TST(opt, RECURSIVE) && S_ISDIR(mode1)) {
50 if (DIFF_OPT_TST(opt, TREE_IN_RECURSIVE)) {
51 opt->change(opt, mode1, mode2,
52 sha1, sha2, 1, 1, base->buf, 0, 0);
53 }
54 strbuf_addch(base, '/');
55 diff_tree_sha1(sha1, sha2, base->buf, opt);
56 } else {
57 opt->change(opt, mode1, mode2, sha1, sha2, 1, 1, base->buf, 0, 0);
58 }
59 strbuf_setlen(base, old_baselen);
60 return 0;
61 }
62
63 /* A whole sub-tree went away or appeared */
64 static void show_tree(struct diff_options *opt, const char *prefix,
65 struct tree_desc *desc, struct strbuf *base)
66 {
67 enum interesting match = entry_not_interesting;
68 for (; desc->size; update_tree_entry(desc)) {
69 if (match != all_entries_interesting) {
70 match = tree_entry_interesting(&desc->entry, base, 0,
71 &opt->pathspec);
72 if (match == all_entries_not_interesting)
73 break;
74 if (match == entry_not_interesting)
75 continue;
76 }
77 show_entry(opt, prefix, desc, base);
78 }
79 }
80
81 /* A file entry went away or appeared */
82 static void show_entry(struct diff_options *opt, const char *prefix,
83 struct tree_desc *desc, struct strbuf *base)
84 {
85 unsigned mode;
86 const char *path;
87 const unsigned char *sha1 = tree_entry_extract(desc, &path, &mode);
88 int pathlen = tree_entry_len(&desc->entry);
89 int old_baselen = base->len;
90
91 strbuf_add(base, path, pathlen);
92 if (DIFF_OPT_TST(opt, RECURSIVE) && S_ISDIR(mode)) {
93 enum object_type type;
94 struct tree_desc inner;
95 void *tree;
96 unsigned long size;
97
98 tree = read_sha1_file(sha1, &type, &size);
99 if (!tree || type != OBJ_TREE)
100 die("corrupt tree sha %s", sha1_to_hex(sha1));
101
102 if (DIFF_OPT_TST(opt, TREE_IN_RECURSIVE))
103 opt->add_remove(opt, *prefix, mode, sha1, 1, base->buf, 0);
104
105 strbuf_addch(base, '/');
106
107 init_tree_desc(&inner, tree, size);
108 show_tree(opt, prefix, &inner, base);
109 free(tree);
110 } else
111 opt->add_remove(opt, prefix[0], mode, sha1, 1, base->buf, 0);
112
113 strbuf_setlen(base, old_baselen);
114 }
115
116 static void skip_uninteresting(struct tree_desc *t, struct strbuf *base,
117 struct diff_options *opt,
118 enum interesting *match)
119 {
120 while (t->size) {
121 *match = tree_entry_interesting(&t->entry, base, 0, &opt->pathspec);
122 if (*match) {
123 if (*match == all_entries_not_interesting)
124 t->size = 0;
125 break;
126 }
127 update_tree_entry(t);
128 }
129 }
130
131 int diff_tree(struct tree_desc *t1, struct tree_desc *t2,
132 const char *base_str, struct diff_options *opt)
133 {
134 struct strbuf base;
135 int baselen = strlen(base_str);
136 enum interesting t1_match = entry_not_interesting;
137 enum interesting t2_match = entry_not_interesting;
138
139 /* Enable recursion indefinitely */
140 opt->pathspec.recursive = DIFF_OPT_TST(opt, RECURSIVE);
141
142 strbuf_init(&base, PATH_MAX);
143 strbuf_add(&base, base_str, baselen);
144
145 for (;;) {
146 if (diff_can_quit_early(opt))
147 break;
148 if (opt->pathspec.nr) {
149 skip_uninteresting(t1, &base, opt, &t1_match);
150 skip_uninteresting(t2, &base, opt, &t2_match);
151 }
152 if (!t1->size) {
153 if (!t2->size)
154 break;
155 show_entry(opt, "+", t2, &base);
156 update_tree_entry(t2);
157 continue;
158 }
159 if (!t2->size) {
160 show_entry(opt, "-", t1, &base);
161 update_tree_entry(t1);
162 continue;
163 }
164 switch (compare_tree_entry(t1, t2, &base, opt)) {
165 case -1:
166 update_tree_entry(t1);
167 continue;
168 case 0:
169 update_tree_entry(t1);
170 /* Fallthrough */
171 case 1:
172 update_tree_entry(t2);
173 continue;
174 }
175 die("git diff-tree: internal error");
176 }
177
178 strbuf_release(&base);
179 return 0;
180 }
181
182 /*
183 * Does it look like the resulting diff might be due to a rename?
184 * - single entry
185 * - not a valid previous file
186 */
187 static inline int diff_might_be_rename(void)
188 {
189 return diff_queued_diff.nr == 1 &&
190 !DIFF_FILE_VALID(diff_queued_diff.queue[0]->one);
191 }
192
193 static void try_to_follow_renames(struct tree_desc *t1, struct tree_desc *t2, const char *base, struct diff_options *opt)
194 {
195 struct diff_options diff_opts;
196 struct diff_queue_struct *q = &diff_queued_diff;
197 struct diff_filepair *choice;
198 int i;
199
200 /*
201 * follow-rename code is very specific, we need exactly one
202 * path. Magic that matches more than one path is not
203 * supported.
204 */
205 GUARD_PATHSPEC(&opt->pathspec, PATHSPEC_FROMTOP);
206 #if 0
207 /*
208 * We should reject wildcards as well. Unfortunately we
209 * haven't got a reliable way to detect that 'foo\*bar' in
210 * fact has no wildcards. nowildcard_len is merely a hint for
211 * optimization. Let it slip for now until wildmatch is taught
212 * about dry-run mode and returns wildcard info.
213 */
214 if (opt->pathspec.has_wildcard)
215 die("BUG:%s:%d: wildcards are not supported",
216 __FILE__, __LINE__);
217 #endif
218
219 /* Remove the file creation entry from the diff queue, and remember it */
220 choice = q->queue[0];
221 q->nr = 0;
222
223 diff_setup(&diff_opts);
224 DIFF_OPT_SET(&diff_opts, RECURSIVE);
225 DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
226 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
227 diff_opts.single_follow = opt->pathspec.items[0].match;
228 diff_opts.break_opt = opt->break_opt;
229 diff_opts.rename_score = opt->rename_score;
230 diff_setup_done(&diff_opts);
231 diff_tree(t1, t2, base, &diff_opts);
232 diffcore_std(&diff_opts);
233 free_pathspec(&diff_opts.pathspec);
234
235 /* Go through the new set of filepairing, and see if we find a more interesting one */
236 opt->found_follow = 0;
237 for (i = 0; i < q->nr; i++) {
238 struct diff_filepair *p = q->queue[i];
239
240 /*
241 * Found a source? Not only do we use that for the new
242 * diff_queued_diff, we will also use that as the path in
243 * the future!
244 */
245 if ((p->status == 'R' || p->status == 'C') &&
246 !strcmp(p->two->path, opt->pathspec.items[0].match)) {
247 const char *path[2];
248
249 /* Switch the file-pairs around */
250 q->queue[i] = choice;
251 choice = p;
252
253 /* Update the path we use from now on.. */
254 path[0] = p->one->path;
255 path[1] = NULL;
256 free_pathspec(&opt->pathspec);
257 parse_pathspec(&opt->pathspec, PATHSPEC_ALL_MAGIC, 0, "", path);
258
259 /*
260 * The caller expects us to return a set of vanilla
261 * filepairs to let a later call to diffcore_std()
262 * it makes to sort the renames out (among other
263 * things), but we already have found renames
264 * ourselves; signal diffcore_std() not to muck with
265 * rename information.
266 */
267 opt->found_follow = 1;
268 break;
269 }
270 }
271
272 /*
273 * Then, discard all the non-relevant file pairs...
274 */
275 for (i = 0; i < q->nr; i++) {
276 struct diff_filepair *p = q->queue[i];
277 diff_free_filepair(p);
278 }
279
280 /*
281 * .. and re-instate the one we want (which might be either the
282 * original one, or the rename/copy we found)
283 */
284 q->queue[0] = choice;
285 q->nr = 1;
286 }
287
288 int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base, struct diff_options *opt)
289 {
290 void *tree1, *tree2;
291 struct tree_desc t1, t2;
292 unsigned long size1, size2;
293 int retval;
294
295 tree1 = read_object_with_reference(old, tree_type, &size1, NULL);
296 if (!tree1)
297 die("unable to read source tree (%s)", sha1_to_hex(old));
298 tree2 = read_object_with_reference(new, tree_type, &size2, NULL);
299 if (!tree2)
300 die("unable to read destination tree (%s)", sha1_to_hex(new));
301 init_tree_desc(&t1, tree1, size1);
302 init_tree_desc(&t2, tree2, size2);
303 retval = diff_tree(&t1, &t2, base, opt);
304 if (!*base && DIFF_OPT_TST(opt, FOLLOW_RENAMES) && diff_might_be_rename()) {
305 init_tree_desc(&t1, tree1, size1);
306 init_tree_desc(&t2, tree2, size2);
307 try_to_follow_renames(&t1, &t2, base, opt);
308 }
309 free(tree1);
310 free(tree2);
311 return retval;
312 }
313
314 int diff_root_tree_sha1(const unsigned char *new, const char *base, struct diff_options *opt)
315 {
316 int retval;
317 void *tree;
318 unsigned long size;
319 struct tree_desc empty, real;
320
321 tree = read_object_with_reference(new, tree_type, &size, NULL);
322 if (!tree)
323 die("unable to read root tree (%s)", sha1_to_hex(new));
324 init_tree_desc(&real, tree, size);
325
326 init_tree_desc(&empty, "", 0);
327 retval = diff_tree(&empty, &real, base, opt);
328 free(tree);
329 return retval;
330 }