[PATCH] Introducing software archaeologist's tool "pickaxe".
[git/dscho.git] / diff-tree.c
blob233a25066818b49c94112f6f2680c6f6d7c881f8
1 #include <ctype.h>
2 #include "cache.h"
3 #include "diff.h"
5 static int silent = 0;
6 static int show_root_diff = 0;
7 static int verbose_header = 0;
8 static int ignore_merges = 1;
9 static int recursive = 0;
10 static int read_stdin = 0;
11 static int line_termination = '\n';
12 static int generate_patch = 0;
13 static int detect_rename = 0;
14 static int reverse_diff = 0;
15 static int diff_score_opt = 0;
16 static char *pickaxe = 0;
17 static const char *header = NULL;
18 static const char *header_prefix = "";
20 // What paths are we interested in?
21 static int nr_paths = 0;
22 static char **paths = NULL;
23 static int *pathlens = NULL;
25 static int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base);
27 static void update_tree_entry(void **bufp, unsigned long *sizep)
29 void *buf = *bufp;
30 unsigned long size = *sizep;
31 int len = strlen(buf) + 1 + 20;
33 if (size < len)
34 die("corrupt tree file");
35 *bufp = buf + len;
36 *sizep = size - len;
39 static const unsigned char *extract(void *tree, unsigned long size, const char **pathp, unsigned int *modep)
41 int len = strlen(tree)+1;
42 const unsigned char *sha1 = tree + len;
43 const char *path = strchr(tree, ' ');
45 if (!path || size < len + 20 || sscanf(tree, "%o", modep) != 1)
46 die("corrupt tree file");
47 *pathp = path+1;
48 return sha1;
51 static char *malloc_base(const char *base, const char *path, int pathlen)
53 int baselen = strlen(base);
54 char *newbase = xmalloc(baselen + pathlen + 2);
55 memcpy(newbase, base, baselen);
56 memcpy(newbase + baselen, path, pathlen);
57 memcpy(newbase + baselen + pathlen, "/", 2);
58 return newbase;
61 static void show_file(const char *prefix, void *tree, unsigned long size, const char *base);
62 static void show_tree(const char *prefix, void *tree, unsigned long size, const char *base);
64 /* A file entry went away or appeared */
65 static void show_file(const char *prefix, void *tree, unsigned long size, const char *base)
67 unsigned mode;
68 const char *path;
69 const unsigned char *sha1 = extract(tree, size, &path, &mode);
71 if (header) {
72 printf("%s", header);
73 header = NULL;
76 if (silent)
77 return;
79 if (recursive && S_ISDIR(mode)) {
80 char type[20];
81 unsigned long size;
82 char *newbase = malloc_base(base, path, strlen(path));
83 void *tree;
85 tree = read_sha1_file(sha1, type, &size);
86 if (!tree || strcmp(type, "tree"))
87 die("corrupt tree sha %s", sha1_to_hex(sha1));
89 show_tree(prefix, tree, size, newbase);
91 free(tree);
92 free(newbase);
93 return;
96 diff_addremove(prefix[0], mode, sha1, base, path);
99 static int compare_tree_entry(void *tree1, unsigned long size1, void *tree2, unsigned long size2, const char *base)
101 unsigned mode1, mode2;
102 const char *path1, *path2;
103 const unsigned char *sha1, *sha2;
104 int cmp, pathlen1, pathlen2;
106 sha1 = extract(tree1, size1, &path1, &mode1);
107 sha2 = extract(tree2, size2, &path2, &mode2);
109 pathlen1 = strlen(path1);
110 pathlen2 = strlen(path2);
111 cmp = base_name_compare(path1, pathlen1, mode1, path2, pathlen2, mode2);
112 if (cmp < 0) {
113 show_file("-", tree1, size1, base);
114 return -1;
116 if (cmp > 0) {
117 show_file("+", tree2, size2, base);
118 return 1;
120 if (!memcmp(sha1, sha2, 20) && mode1 == mode2)
121 return 0;
124 * If the filemode has changed to/from a directory from/to a regular
125 * file, we need to consider it a remove and an add.
127 if (S_ISDIR(mode1) != S_ISDIR(mode2)) {
128 show_file("-", tree1, size1, base);
129 show_file("+", tree2, size2, base);
130 return 0;
133 if (recursive && S_ISDIR(mode1)) {
134 int retval;
135 char *newbase = malloc_base(base, path1, pathlen1);
136 retval = diff_tree_sha1(sha1, sha2, newbase);
137 free(newbase);
138 return retval;
141 if (header) {
142 printf("%s", header);
143 header = NULL;
145 if (silent)
146 return 0;
148 diff_change(mode1, mode2, sha1, sha2, base, path1);
149 return 0;
152 static int interesting(void *tree, unsigned long size, const char *base)
154 const char *path;
155 unsigned mode;
156 int i;
157 int baselen, pathlen;
159 if (!nr_paths)
160 return 1;
162 (void)extract(tree, size, &path, &mode);
164 pathlen = strlen(path);
165 baselen = strlen(base);
167 for (i=0; i < nr_paths; i++) {
168 const char *match = paths[i];
169 int matchlen = pathlens[i];
171 if (baselen >= matchlen) {
172 /* If it doesn't match, move along... */
173 if (strncmp(base, match, matchlen))
174 continue;
176 /* The base is a subdirectory of a path which was specified. */
177 return 1;
180 /* Does the base match? */
181 if (strncmp(base, match, baselen))
182 continue;
184 match += baselen;
185 matchlen -= baselen;
187 if (pathlen > matchlen)
188 continue;
190 if (matchlen > pathlen) {
191 if (match[pathlen] != '/')
192 continue;
193 if (!S_ISDIR(mode))
194 continue;
197 if (strncmp(path, match, pathlen))
198 continue;
200 return 1;
202 return 0; /* No matches */
205 /* A whole sub-tree went away or appeared */
206 static void show_tree(const char *prefix, void *tree, unsigned long size, const char *base)
208 while (size) {
209 if (interesting(tree, size, base))
210 show_file(prefix, tree, size, base);
211 update_tree_entry(&tree, &size);
215 static int diff_tree(void *tree1, unsigned long size1, void *tree2, unsigned long size2, const char *base)
217 while (size1 | size2) {
218 if (nr_paths && size1 && !interesting(tree1, size1, base)) {
219 update_tree_entry(&tree1, &size1);
220 continue;
222 if (nr_paths && size2 && !interesting(tree2, size2, base)) {
223 update_tree_entry(&tree2, &size2);
224 continue;
226 if (!size1) {
227 show_file("+", tree2, size2, base);
228 update_tree_entry(&tree2, &size2);
229 continue;
231 if (!size2) {
232 show_file("-", tree1, size1, base);
233 update_tree_entry(&tree1, &size1);
234 continue;
236 switch (compare_tree_entry(tree1, size1, tree2, size2, base)) {
237 case -1:
238 update_tree_entry(&tree1, &size1);
239 continue;
240 case 0:
241 update_tree_entry(&tree1, &size1);
242 /* Fallthrough */
243 case 1:
244 update_tree_entry(&tree2, &size2);
245 continue;
247 die("git-diff-tree: internal error");
249 return 0;
252 static int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base)
254 void *tree1, *tree2;
255 unsigned long size1, size2;
256 int retval;
258 tree1 = read_object_with_reference(old, "tree", &size1, NULL);
259 if (!tree1)
260 die("unable to read source tree (%s)", sha1_to_hex(old));
261 tree2 = read_object_with_reference(new, "tree", &size2, NULL);
262 if (!tree2)
263 die("unable to read destination tree (%s)", sha1_to_hex(new));
264 retval = diff_tree(tree1, size1, tree2, size2, base);
265 free(tree1);
266 free(tree2);
267 return retval;
270 static int diff_tree_sha1_top(const unsigned char *old,
271 const unsigned char *new, const char *base)
273 int ret;
275 diff_setup(detect_rename, diff_score_opt, pickaxe,
276 reverse_diff, (generate_patch ? -1 : line_termination),
277 NULL, 0);
278 ret = diff_tree_sha1(old, new, base);
279 diff_flush();
280 return ret;
283 static int diff_root_tree(const unsigned char *new, const char *base)
285 int retval;
286 void *tree;
287 unsigned long size;
289 diff_setup(detect_rename, diff_score_opt, pickaxe,
290 reverse_diff, (generate_patch ? -1 : line_termination),
291 NULL, 0);
292 tree = read_object_with_reference(new, "tree", &size, NULL);
293 if (!tree)
294 die("unable to read root tree (%s)", sha1_to_hex(new));
295 retval = diff_tree("", 0, tree, size, base);
296 free(tree);
297 diff_flush();
298 return retval;
301 static int get_one_line(const char *msg, unsigned long len)
303 int ret = 0;
305 while (len--) {
306 ret++;
307 if (*msg++ == '\n')
308 break;
310 return ret;
313 static int add_author_info(char *buf, const char *line, int len)
315 char *date;
316 unsigned int namelen;
317 unsigned long time;
318 int tz;
320 line += strlen("author ");
321 date = strchr(line, '>');
322 if (!date)
323 return 0;
324 namelen = ++date - line;
325 time = strtoul(date, &date, 10);
326 tz = strtol(date, NULL, 10);
328 return sprintf(buf, "Author: %.*s\nDate: %s\n",
329 namelen, line,
330 show_date(time, tz));
333 static char *generate_header(const char *commit, const char *parent, const char *msg, unsigned long len)
335 static char this_header[1000];
336 int offset;
338 offset = sprintf(this_header, "%s%s (from %s)\n", header_prefix, commit, parent);
339 if (verbose_header) {
340 int hdr = 1;
342 for (;;) {
343 const char *line = msg;
344 int linelen = get_one_line(msg, len);
346 if (!linelen)
347 break;
348 if (offset + linelen + 10 > sizeof(this_header))
349 break;
351 msg += linelen;
352 len -= linelen;
353 if (linelen == 1)
354 hdr = 0;
355 if (hdr) {
356 if (!memcmp(line, "author ", 7))
357 offset += add_author_info(this_header + offset, line, linelen);
358 continue;
360 memset(this_header + offset, ' ', 4);
361 memcpy(this_header + offset + 4, line, linelen);
362 offset += linelen + 4;
364 this_header[offset++] = '\n';
365 this_header[offset] = 0;
368 return this_header;
371 static int diff_tree_commit(const unsigned char *commit, const char *name)
373 unsigned long size, offset;
374 char *buf = read_object_with_reference(commit, "commit", &size, NULL);
376 if (!buf)
377 return -1;
379 if (!name) {
380 static char commit_name[60];
381 strcpy(commit_name, sha1_to_hex(commit));
382 name = commit_name;
385 /* Root commit? */
386 if (show_root_diff && memcmp(buf + 46, "parent ", 7)) {
387 header = generate_header(name, "root", buf, size);
388 diff_root_tree(commit, "");
391 /* More than one parent? */
392 if (ignore_merges) {
393 if (!memcmp(buf + 46 + 48, "parent ", 7))
394 return 0;
397 offset = 46;
398 while (offset + 48 < size && !memcmp(buf + offset, "parent ", 7)) {
399 unsigned char parent[20];
400 if (get_sha1_hex(buf + offset + 7, parent))
401 return -1;
402 header = generate_header(name, sha1_to_hex(parent), buf, size);
403 diff_tree_sha1_top(parent, commit, "");
404 if (!header && verbose_header)
405 header_prefix = "\ndiff-tree ";
406 offset += 48;
408 return 0;
411 static int diff_tree_stdin(char *line)
413 int len = strlen(line);
414 unsigned char commit[20], parent[20];
415 static char this_header[1000];
417 if (!len || line[len-1] != '\n')
418 return -1;
419 line[len-1] = 0;
420 if (get_sha1_hex(line, commit))
421 return -1;
422 if (isspace(line[40]) && !get_sha1_hex(line+41, parent)) {
423 line[40] = 0;
424 line[81] = 0;
425 sprintf(this_header, "%s (from %s)\n", line, line+41);
426 header = this_header;
427 return diff_tree_sha1_top(parent, commit, "");
429 line[40] = 0;
430 return diff_tree_commit(commit, line);
433 static char *diff_tree_usage =
434 "git-diff-tree [-p] [-r] [-z] [--stdin] [-M] [-C] [-R] [-S<string>] [-m] [-s] [-v] <tree-ish> <tree-ish>";
436 int main(int argc, char **argv)
438 int nr_sha1;
439 char line[1000];
440 unsigned char sha1[2][20];
442 nr_sha1 = 0;
443 for (;;) {
444 char *arg;
446 argv++;
447 argc--;
448 arg = *argv;
449 if (!arg)
450 break;
452 if (*arg != '-') {
453 if (nr_sha1 < 2 && !get_sha1(arg, sha1[nr_sha1])) {
454 nr_sha1++;
455 continue;
457 break;
460 if (!strcmp(arg, "--")) {
461 argv++;
462 argc--;
463 break;
465 if (!strcmp(arg, "-r")) {
466 recursive = 1;
467 continue;
469 if (!strcmp(arg, "-R")) {
470 reverse_diff = 1;
471 continue;
473 if (!strcmp(arg, "-p")) {
474 recursive = generate_patch = 1;
475 continue;
477 if (!strncmp(arg, "-S", 2)) {
478 pickaxe = arg + 2;
479 continue;
481 if (!strncmp(arg, "-M", 2)) {
482 detect_rename = recursive = generate_patch = 1;
483 diff_score_opt = diff_scoreopt_parse(arg);
484 continue;
486 if (!strncmp(arg, "-C", 2)) {
487 detect_rename = 2;
488 recursive = generate_patch = 1;
489 diff_score_opt = diff_scoreopt_parse(arg);
490 continue;
492 if (!strcmp(arg, "-z")) {
493 line_termination = '\0';
494 continue;
496 if (!strcmp(arg, "-m")) {
497 ignore_merges = 0;
498 continue;
500 if (!strcmp(arg, "-s")) {
501 silent = 1;
502 continue;
504 if (!strcmp(arg, "-v")) {
505 verbose_header = 1;
506 header_prefix = "diff-tree ";
507 continue;
509 if (!strcmp(arg, "--stdin")) {
510 read_stdin = 1;
511 continue;
513 if (!strcmp(arg, "--root")) {
514 show_root_diff = 1;
515 continue;
517 usage(diff_tree_usage);
520 if (argc > 0) {
521 int i;
523 paths = argv;
524 nr_paths = argc;
525 pathlens = xmalloc(nr_paths * sizeof(int));
526 for (i=0; i<nr_paths; i++)
527 pathlens[i] = strlen(paths[i]);
530 switch (nr_sha1) {
531 case 0:
532 if (!read_stdin)
533 usage(diff_tree_usage);
534 break;
535 case 1:
536 diff_tree_commit(sha1[0], NULL);
537 break;
538 case 2:
539 diff_tree_sha1_top(sha1[0], sha1[1], "");
540 break;
543 if (!read_stdin)
544 return 0;
546 while (fgets(line, sizeof(line), stdin))
547 diff_tree_stdin(line);
549 return 0;