]>
Commit | Line | Data |
---|---|---|
cee7f245 JH |
1 | /* |
2 | * Pickaxe | |
3 | * | |
4 | * Copyright (c) 2006, Junio C Hamano | |
5 | */ | |
6 | ||
7 | #include "cache.h" | |
8 | #include "builtin.h" | |
9 | #include "blob.h" | |
10 | #include "commit.h" | |
11 | #include "tag.h" | |
12 | #include "tree-walk.h" | |
13 | #include "diff.h" | |
14 | #include "diffcore.h" | |
15 | #include "revision.h" | |
717d1462 | 16 | #include "quote.h" |
cee7f245 | 17 | #include "xdiff-interface.h" |
1cfe7733 | 18 | #include "cache-tree.h" |
f9567384 JH |
19 | #include "path-list.h" |
20 | #include "mailmap.h" | |
cee7f245 | 21 | |
acca687f | 22 | static char blame_usage[] = |
b82871b3 | 23 | "git-blame [-c] [-b] [-l] [--root] [-t] [-f] [-n] [-s] [-p] [-w] [-L n,m] [-S <revs-file>] [-M] [-C] [-C] [--contents <filename>] [--incremental] [commit] [--] file\n" |
635f4a30 | 24 | " -c Use the same output mode as git-annotate (Default: off)\n" |
4c10a5ca | 25 | " -b Show blank SHA-1 for boundary commits (Default: off)\n" |
635f4a30 | 26 | " -l Show long commit SHA1 (Default: off)\n" |
4c10a5ca | 27 | " --root Do not treat root commits as boundaries (Default: off)\n" |
635f4a30 | 28 | " -t Show raw timestamp (Default: off)\n" |
cee7f245 JH |
29 | " -f, --show-name Show original filename (Default: auto)\n" |
30 | " -n, --show-number Show original linenumber (Default: off)\n" | |
093dc5be | 31 | " -s Suppress author name and timestamp (Default: off)\n" |
cee7f245 | 32 | " -p, --porcelain Show in a format designed for machine consumption\n" |
b82871b3 | 33 | " -w Ignore whitespace differences\n" |
cee7f245 | 34 | " -L n,m Process only line range n,m, counting from 1\n" |
18abd745 | 35 | " -M, -C Find line movements within and across files\n" |
717d1462 | 36 | " --incremental Show blame entries as we find them, incrementally\n" |
1cfe7733 | 37 | " --contents file Use <file>'s contents as the final image\n" |
cee7f245 JH |
38 | " -S revs-file Use revisions from revs-file instead of calling git-rev-list\n"; |
39 | ||
40 | static int longest_file; | |
41 | static int longest_author; | |
42 | static int max_orig_digits; | |
43 | static int max_digits; | |
5ff62c30 | 44 | static int max_score_digits; |
4c10a5ca JH |
45 | static int show_root; |
46 | static int blank_boundary; | |
717d1462 | 47 | static int incremental; |
e68989a7 | 48 | static int cmd_is_annotate; |
b82871b3 | 49 | static int xdl_opts = XDF_NEED_MINIMAL; |
f9567384 | 50 | static struct path_list mailmap; |
cee7f245 | 51 | |
54a4c617 JH |
52 | #ifndef DEBUG |
53 | #define DEBUG 0 | |
54 | #endif | |
55 | ||
c2e525d9 JH |
56 | /* stats */ |
57 | static int num_read_blob; | |
58 | static int num_get_patch; | |
59 | static int num_commits; | |
60 | ||
d24bba80 | 61 | #define PICKAXE_BLAME_MOVE 01 |
18abd745 JH |
62 | #define PICKAXE_BLAME_COPY 02 |
63 | #define PICKAXE_BLAME_COPY_HARDER 04 | |
c63777c0 | 64 | #define PICKAXE_BLAME_COPY_HARDEST 010 |
d24bba80 | 65 | |
4a0fc95f JH |
66 | /* |
67 | * blame for a blame_entry with score lower than these thresholds | |
68 | * is not passed to the parent using move/copy logic. | |
69 | */ | |
70 | static unsigned blame_move_score; | |
71 | static unsigned blame_copy_score; | |
72 | #define BLAME_DEFAULT_MOVE_SCORE 20 | |
73 | #define BLAME_DEFAULT_COPY_SCORE 40 | |
74 | ||
cee7f245 JH |
75 | /* bits #0..7 in revision.h, #8..11 used for merge_bases() in commit.c */ |
76 | #define METAINFO_SHOWN (1u<<12) | |
77 | #define MORE_THAN_ONE_PATH (1u<<13) | |
78 | ||
79 | /* | |
54a4c617 | 80 | * One blob in a commit that is being suspected |
cee7f245 JH |
81 | */ |
82 | struct origin { | |
54a4c617 | 83 | int refcnt; |
cee7f245 | 84 | struct commit *commit; |
c2e525d9 | 85 | mmfile_t file; |
cee7f245 JH |
86 | unsigned char blob_sha1[20]; |
87 | char path[FLEX_ARRAY]; | |
88 | }; | |
89 | ||
1732a1fd JH |
90 | /* |
91 | * Given an origin, prepare mmfile_t structure to be used by the | |
92 | * diff machinery | |
93 | */ | |
c2e525d9 JH |
94 | static char *fill_origin_blob(struct origin *o, mmfile_t *file) |
95 | { | |
96 | if (!o->file.ptr) { | |
21666f1a | 97 | enum object_type type; |
c2e525d9 | 98 | num_read_blob++; |
21666f1a | 99 | file->ptr = read_sha1_file(o->blob_sha1, &type, |
c2e525d9 JH |
100 | (unsigned long *)(&(file->size))); |
101 | o->file = *file; | |
102 | } | |
103 | else | |
104 | *file = o->file; | |
105 | return file->ptr; | |
106 | } | |
107 | ||
1732a1fd JH |
108 | /* |
109 | * Origin is refcounted and usually we keep the blob contents to be | |
110 | * reused. | |
111 | */ | |
54a4c617 JH |
112 | static inline struct origin *origin_incref(struct origin *o) |
113 | { | |
114 | if (o) | |
115 | o->refcnt++; | |
116 | return o; | |
117 | } | |
118 | ||
119 | static void origin_decref(struct origin *o) | |
120 | { | |
121 | if (o && --o->refcnt <= 0) { | |
c2e525d9 JH |
122 | if (o->file.ptr) |
123 | free(o->file.ptr); | |
54a4c617 JH |
124 | memset(o, 0, sizeof(*o)); |
125 | free(o); | |
126 | } | |
127 | } | |
128 | ||
1732a1fd JH |
129 | /* |
130 | * Each group of lines is described by a blame_entry; it can be split | |
131 | * as we pass blame to the parents. They form a linked list in the | |
132 | * scoreboard structure, sorted by the target line number. | |
133 | */ | |
cee7f245 JH |
134 | struct blame_entry { |
135 | struct blame_entry *prev; | |
136 | struct blame_entry *next; | |
137 | ||
138 | /* the first line of this group in the final image; | |
139 | * internally all line numbers are 0 based. | |
140 | */ | |
141 | int lno; | |
142 | ||
143 | /* how many lines this group has */ | |
144 | int num_lines; | |
145 | ||
146 | /* the commit that introduced this group into the final image */ | |
147 | struct origin *suspect; | |
148 | ||
149 | /* true if the suspect is truly guilty; false while we have not | |
150 | * checked if the group came from one of its parents. | |
151 | */ | |
152 | char guilty; | |
153 | ||
154 | /* the line number of the first line of this group in the | |
155 | * suspect's file; internally all line numbers are 0 based. | |
156 | */ | |
157 | int s_lno; | |
5ff62c30 JH |
158 | |
159 | /* how significant this entry is -- cached to avoid | |
1732a1fd | 160 | * scanning the lines over and over. |
5ff62c30 JH |
161 | */ |
162 | unsigned score; | |
cee7f245 JH |
163 | }; |
164 | ||
1732a1fd JH |
165 | /* |
166 | * The current state of the blame assignment. | |
167 | */ | |
cee7f245 JH |
168 | struct scoreboard { |
169 | /* the final commit (i.e. where we started digging from) */ | |
170 | struct commit *final; | |
171 | ||
172 | const char *path; | |
173 | ||
1732a1fd JH |
174 | /* |
175 | * The contents in the final image. | |
176 | * Used by many functions to obtain contents of the nth line, | |
177 | * indexed with scoreboard.lineno[blame_entry.lno]. | |
cee7f245 JH |
178 | */ |
179 | const char *final_buf; | |
180 | unsigned long final_buf_size; | |
181 | ||
182 | /* linked list of blames */ | |
183 | struct blame_entry *ent; | |
184 | ||
612702e8 | 185 | /* look-up a line in the final buffer */ |
cee7f245 JH |
186 | int num_lines; |
187 | int *lineno; | |
188 | }; | |
189 | ||
171dccd5 | 190 | static inline int same_suspect(struct origin *a, struct origin *b) |
46014766 | 191 | { |
171dccd5 | 192 | if (a == b) |
57584d9e | 193 | return 1; |
171dccd5 JH |
194 | if (a->commit != b->commit) |
195 | return 0; | |
196 | return !strcmp(a->path, b->path); | |
46014766 JH |
197 | } |
198 | ||
54a4c617 JH |
199 | static void sanity_check_refcnt(struct scoreboard *); |
200 | ||
1732a1fd JH |
201 | /* |
202 | * If two blame entries that are next to each other came from | |
203 | * contiguous lines in the same origin (i.e. <commit, path> pair), | |
204 | * merge them together. | |
205 | */ | |
cee7f245 JH |
206 | static void coalesce(struct scoreboard *sb) |
207 | { | |
208 | struct blame_entry *ent, *next; | |
209 | ||
210 | for (ent = sb->ent; ent && (next = ent->next); ent = next) { | |
171dccd5 | 211 | if (same_suspect(ent->suspect, next->suspect) && |
cee7f245 JH |
212 | ent->guilty == next->guilty && |
213 | ent->s_lno + ent->num_lines == next->s_lno) { | |
214 | ent->num_lines += next->num_lines; | |
215 | ent->next = next->next; | |
216 | if (ent->next) | |
217 | ent->next->prev = ent; | |
54a4c617 | 218 | origin_decref(next->suspect); |
cee7f245 | 219 | free(next); |
46014766 | 220 | ent->score = 0; |
cee7f245 JH |
221 | next = ent; /* again */ |
222 | } | |
223 | } | |
54a4c617 JH |
224 | |
225 | if (DEBUG) /* sanity */ | |
226 | sanity_check_refcnt(sb); | |
cee7f245 JH |
227 | } |
228 | ||
1732a1fd JH |
229 | /* |
230 | * Given a commit and a path in it, create a new origin structure. | |
231 | * The callers that add blame to the scoreboard should use | |
232 | * get_origin() to obtain shared, refcounted copy instead of calling | |
233 | * this function directly. | |
234 | */ | |
854b97f6 JH |
235 | static struct origin *make_origin(struct commit *commit, const char *path) |
236 | { | |
237 | struct origin *o; | |
238 | o = xcalloc(1, sizeof(*o) + strlen(path) + 1); | |
239 | o->commit = commit; | |
240 | o->refcnt = 1; | |
241 | strcpy(o->path, path); | |
242 | return o; | |
243 | } | |
244 | ||
1732a1fd JH |
245 | /* |
246 | * Locate an existing origin or create a new one. | |
247 | */ | |
f6c0e191 JH |
248 | static struct origin *get_origin(struct scoreboard *sb, |
249 | struct commit *commit, | |
250 | const char *path) | |
cee7f245 | 251 | { |
f6c0e191 | 252 | struct blame_entry *e; |
cee7f245 | 253 | |
f6c0e191 JH |
254 | for (e = sb->ent; e; e = e->next) { |
255 | if (e->suspect->commit == commit && | |
256 | !strcmp(e->suspect->path, path)) | |
54a4c617 | 257 | return origin_incref(e->suspect); |
cee7f245 | 258 | } |
854b97f6 | 259 | return make_origin(commit, path); |
cee7f245 JH |
260 | } |
261 | ||
1732a1fd JH |
262 | /* |
263 | * Fill the blob_sha1 field of an origin if it hasn't, so that later | |
264 | * call to fill_origin_blob() can use it to locate the data. blob_sha1 | |
265 | * for an origin is also used to pass the blame for the entire file to | |
266 | * the parent to detect the case where a child's blob is identical to | |
267 | * that of its parent's. | |
268 | */ | |
f6c0e191 JH |
269 | static int fill_blob_sha1(struct origin *origin) |
270 | { | |
271 | unsigned mode; | |
f6c0e191 JH |
272 | |
273 | if (!is_null_sha1(origin->blob_sha1)) | |
274 | return 0; | |
275 | if (get_tree_entry(origin->commit->object.sha1, | |
276 | origin->path, | |
277 | origin->blob_sha1, &mode)) | |
278 | goto error_out; | |
21666f1a | 279 | if (sha1_object_info(origin->blob_sha1, NULL) != OBJ_BLOB) |
f6c0e191 JH |
280 | goto error_out; |
281 | return 0; | |
282 | error_out: | |
283 | hashclr(origin->blob_sha1); | |
284 | return -1; | |
285 | } | |
286 | ||
1732a1fd JH |
287 | /* |
288 | * We have an origin -- check if the same path exists in the | |
289 | * parent and return an origin structure to represent it. | |
290 | */ | |
f6c0e191 | 291 | static struct origin *find_origin(struct scoreboard *sb, |
cee7f245 JH |
292 | struct commit *parent, |
293 | struct origin *origin) | |
294 | { | |
295 | struct origin *porigin = NULL; | |
296 | struct diff_options diff_opts; | |
f6c0e191 JH |
297 | const char *paths[2]; |
298 | ||
0d981c67 | 299 | if (parent->util) { |
1732a1fd JH |
300 | /* |
301 | * Each commit object can cache one origin in that | |
302 | * commit. This is a freestanding copy of origin and | |
303 | * not refcounted. | |
854b97f6 | 304 | */ |
0d981c67 | 305 | struct origin *cached = parent->util; |
854b97f6 | 306 | if (!strcmp(cached->path, origin->path)) { |
1732a1fd JH |
307 | /* |
308 | * The same path between origin and its parent | |
309 | * without renaming -- the most common case. | |
310 | */ | |
854b97f6 | 311 | porigin = get_origin(sb, parent, cached->path); |
1732a1fd JH |
312 | |
313 | /* | |
314 | * If the origin was newly created (i.e. get_origin | |
315 | * would call make_origin if none is found in the | |
316 | * scoreboard), it does not know the blob_sha1, | |
317 | * so copy it. Otherwise porigin was in the | |
318 | * scoreboard and already knows blob_sha1. | |
319 | */ | |
854b97f6 JH |
320 | if (porigin->refcnt == 1) |
321 | hashcpy(porigin->blob_sha1, cached->blob_sha1); | |
322 | return porigin; | |
323 | } | |
324 | /* otherwise it was not very useful; free it */ | |
325 | free(parent->util); | |
326 | parent->util = NULL; | |
0d981c67 JH |
327 | } |
328 | ||
f6c0e191 JH |
329 | /* See if the origin->path is different between parent |
330 | * and origin first. Most of the time they are the | |
331 | * same and diff-tree is fairly efficient about this. | |
332 | */ | |
333 | diff_setup(&diff_opts); | |
334 | diff_opts.recursive = 1; | |
335 | diff_opts.detect_rename = 0; | |
336 | diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT; | |
337 | paths[0] = origin->path; | |
338 | paths[1] = NULL; | |
339 | ||
340 | diff_tree_setup_paths(paths, &diff_opts); | |
341 | if (diff_setup_done(&diff_opts) < 0) | |
342 | die("diff-setup"); | |
1cfe7733 JH |
343 | |
344 | if (is_null_sha1(origin->commit->object.sha1)) | |
345 | do_diff_cache(parent->tree->object.sha1, &diff_opts); | |
346 | else | |
347 | diff_tree_sha1(parent->tree->object.sha1, | |
348 | origin->commit->tree->object.sha1, | |
349 | "", &diff_opts); | |
f6c0e191 JH |
350 | diffcore_std(&diff_opts); |
351 | ||
352 | /* It is either one entry that says "modified", or "created", | |
353 | * or nothing. | |
354 | */ | |
355 | if (!diff_queued_diff.nr) { | |
356 | /* The path is the same as parent */ | |
357 | porigin = get_origin(sb, parent, origin->path); | |
358 | hashcpy(porigin->blob_sha1, origin->blob_sha1); | |
359 | } | |
360 | else if (diff_queued_diff.nr != 1) | |
acca687f | 361 | die("internal error in blame::find_origin"); |
f6c0e191 JH |
362 | else { |
363 | struct diff_filepair *p = diff_queued_diff.queue[0]; | |
364 | switch (p->status) { | |
365 | default: | |
acca687f | 366 | die("internal error in blame::find_origin (%c)", |
f6c0e191 JH |
367 | p->status); |
368 | case 'M': | |
369 | porigin = get_origin(sb, parent, origin->path); | |
370 | hashcpy(porigin->blob_sha1, p->one->sha1); | |
371 | break; | |
372 | case 'A': | |
373 | case 'T': | |
374 | /* Did not exist in parent, or type changed */ | |
375 | break; | |
376 | } | |
377 | } | |
378 | diff_flush(&diff_opts); | |
0d981c67 | 379 | if (porigin) { |
1732a1fd JH |
380 | /* |
381 | * Create a freestanding copy that is not part of | |
382 | * the refcounted origin found in the scoreboard, and | |
383 | * cache it in the commit. | |
384 | */ | |
854b97f6 | 385 | struct origin *cached; |
1732a1fd | 386 | |
854b97f6 JH |
387 | cached = make_origin(porigin->commit, porigin->path); |
388 | hashcpy(cached->blob_sha1, porigin->blob_sha1); | |
389 | parent->util = cached; | |
0d981c67 | 390 | } |
f69e743d JH |
391 | return porigin; |
392 | } | |
f6c0e191 | 393 | |
1732a1fd JH |
394 | /* |
395 | * We have an origin -- find the path that corresponds to it in its | |
396 | * parent and return an origin structure to represent it. | |
397 | */ | |
f69e743d JH |
398 | static struct origin *find_rename(struct scoreboard *sb, |
399 | struct commit *parent, | |
400 | struct origin *origin) | |
401 | { | |
402 | struct origin *porigin = NULL; | |
403 | struct diff_options diff_opts; | |
404 | int i; | |
405 | const char *paths[2]; | |
cee7f245 JH |
406 | |
407 | diff_setup(&diff_opts); | |
408 | diff_opts.recursive = 1; | |
409 | diff_opts.detect_rename = DIFF_DETECT_RENAME; | |
410 | diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT; | |
2f3f8b21 | 411 | diff_opts.single_follow = origin->path; |
cee7f245 JH |
412 | paths[0] = NULL; |
413 | diff_tree_setup_paths(paths, &diff_opts); | |
414 | if (diff_setup_done(&diff_opts) < 0) | |
415 | die("diff-setup"); | |
1cfe7733 JH |
416 | |
417 | if (is_null_sha1(origin->commit->object.sha1)) | |
418 | do_diff_cache(parent->tree->object.sha1, &diff_opts); | |
419 | else | |
420 | diff_tree_sha1(parent->tree->object.sha1, | |
421 | origin->commit->tree->object.sha1, | |
422 | "", &diff_opts); | |
cee7f245 JH |
423 | diffcore_std(&diff_opts); |
424 | ||
425 | for (i = 0; i < diff_queued_diff.nr; i++) { | |
426 | struct diff_filepair *p = diff_queued_diff.queue[i]; | |
612702e8 | 427 | if ((p->status == 'R' || p->status == 'C') && |
f6c0e191 JH |
428 | !strcmp(p->two->path, origin->path)) { |
429 | porigin = get_origin(sb, parent, p->one->path); | |
430 | hashcpy(porigin->blob_sha1, p->one->sha1); | |
cee7f245 JH |
431 | break; |
432 | } | |
433 | } | |
434 | diff_flush(&diff_opts); | |
435 | return porigin; | |
436 | } | |
437 | ||
1732a1fd JH |
438 | /* |
439 | * Parsing of patch chunks... | |
440 | */ | |
cee7f245 JH |
441 | struct chunk { |
442 | /* line number in postimage; up to but not including this | |
443 | * line is the same as preimage | |
444 | */ | |
445 | int same; | |
446 | ||
447 | /* preimage line number after this chunk */ | |
448 | int p_next; | |
449 | ||
450 | /* postimage line number after this chunk */ | |
451 | int t_next; | |
452 | }; | |
453 | ||
454 | struct patch { | |
455 | struct chunk *chunks; | |
456 | int num; | |
457 | }; | |
458 | ||
459 | struct blame_diff_state { | |
460 | struct xdiff_emit_state xm; | |
461 | struct patch *ret; | |
462 | unsigned hunk_post_context; | |
463 | unsigned hunk_in_pre_context : 1; | |
464 | }; | |
465 | ||
466 | static void process_u_diff(void *state_, char *line, unsigned long len) | |
467 | { | |
468 | struct blame_diff_state *state = state_; | |
469 | struct chunk *chunk; | |
470 | int off1, off2, len1, len2, num; | |
471 | ||
cee7f245 JH |
472 | num = state->ret->num; |
473 | if (len < 4 || line[0] != '@' || line[1] != '@') { | |
474 | if (state->hunk_in_pre_context && line[0] == ' ') | |
475 | state->ret->chunks[num - 1].same++; | |
476 | else { | |
477 | state->hunk_in_pre_context = 0; | |
478 | if (line[0] == ' ') | |
479 | state->hunk_post_context++; | |
480 | else | |
481 | state->hunk_post_context = 0; | |
482 | } | |
483 | return; | |
484 | } | |
485 | ||
486 | if (num && state->hunk_post_context) { | |
487 | chunk = &state->ret->chunks[num - 1]; | |
488 | chunk->p_next -= state->hunk_post_context; | |
489 | chunk->t_next -= state->hunk_post_context; | |
490 | } | |
491 | state->ret->num = ++num; | |
492 | state->ret->chunks = xrealloc(state->ret->chunks, | |
493 | sizeof(struct chunk) * num); | |
494 | chunk = &state->ret->chunks[num - 1]; | |
495 | if (parse_hunk_header(line, len, &off1, &len1, &off2, &len2)) { | |
496 | state->ret->num--; | |
497 | return; | |
498 | } | |
499 | ||
500 | /* Line numbers in patch output are one based. */ | |
501 | off1--; | |
502 | off2--; | |
503 | ||
504 | chunk->same = len2 ? off2 : (off2 + 1); | |
505 | ||
506 | chunk->p_next = off1 + (len1 ? len1 : 1); | |
507 | chunk->t_next = chunk->same + len2; | |
508 | state->hunk_in_pre_context = 1; | |
509 | state->hunk_post_context = 0; | |
510 | } | |
511 | ||
512 | static struct patch *compare_buffer(mmfile_t *file_p, mmfile_t *file_o, | |
513 | int context) | |
514 | { | |
515 | struct blame_diff_state state; | |
516 | xpparam_t xpp; | |
517 | xdemitconf_t xecfg; | |
518 | xdemitcb_t ecb; | |
519 | ||
b82871b3 | 520 | xpp.flags = xdl_opts; |
30b25010 | 521 | memset(&xecfg, 0, sizeof(xecfg)); |
cee7f245 | 522 | xecfg.ctxlen = context; |
cee7f245 JH |
523 | ecb.outf = xdiff_outf; |
524 | ecb.priv = &state; | |
525 | memset(&state, 0, sizeof(state)); | |
526 | state.xm.consume = process_u_diff; | |
527 | state.ret = xmalloc(sizeof(struct patch)); | |
528 | state.ret->chunks = NULL; | |
529 | state.ret->num = 0; | |
530 | ||
531 | xdl_diff(file_p, file_o, &xpp, &xecfg, &ecb); | |
532 | ||
533 | if (state.ret->num) { | |
534 | struct chunk *chunk; | |
535 | chunk = &state.ret->chunks[state.ret->num - 1]; | |
536 | chunk->p_next -= state.hunk_post_context; | |
537 | chunk->t_next -= state.hunk_post_context; | |
538 | } | |
539 | return state.ret; | |
540 | } | |
541 | ||
1732a1fd JH |
542 | /* |
543 | * Run diff between two origins and grab the patch output, so that | |
544 | * we can pass blame for lines origin is currently suspected for | |
545 | * to its parent. | |
546 | */ | |
cee7f245 JH |
547 | static struct patch *get_patch(struct origin *parent, struct origin *origin) |
548 | { | |
549 | mmfile_t file_p, file_o; | |
cee7f245 JH |
550 | struct patch *patch; |
551 | ||
c2e525d9 JH |
552 | fill_origin_blob(parent, &file_p); |
553 | fill_origin_blob(origin, &file_o); | |
554 | if (!file_p.ptr || !file_o.ptr) | |
cee7f245 | 555 | return NULL; |
cee7f245 | 556 | patch = compare_buffer(&file_p, &file_o, 0); |
c2e525d9 | 557 | num_get_patch++; |
cee7f245 JH |
558 | return patch; |
559 | } | |
560 | ||
561 | static void free_patch(struct patch *p) | |
562 | { | |
563 | free(p->chunks); | |
564 | free(p); | |
565 | } | |
566 | ||
1732a1fd | 567 | /* |
3dff5379 | 568 | * Link in a new blame entry to the scoreboard. Entries that cover the |
1732a1fd JH |
569 | * same line range have been removed from the scoreboard previously. |
570 | */ | |
cee7f245 JH |
571 | static void add_blame_entry(struct scoreboard *sb, struct blame_entry *e) |
572 | { | |
573 | struct blame_entry *ent, *prev = NULL; | |
574 | ||
54a4c617 JH |
575 | origin_incref(e->suspect); |
576 | ||
cee7f245 JH |
577 | for (ent = sb->ent; ent && ent->lno < e->lno; ent = ent->next) |
578 | prev = ent; | |
579 | ||
580 | /* prev, if not NULL, is the last one that is below e */ | |
581 | e->prev = prev; | |
582 | if (prev) { | |
583 | e->next = prev->next; | |
584 | prev->next = e; | |
585 | } | |
586 | else { | |
587 | e->next = sb->ent; | |
588 | sb->ent = e; | |
589 | } | |
590 | if (e->next) | |
591 | e->next->prev = e; | |
592 | } | |
593 | ||
1732a1fd JH |
594 | /* |
595 | * src typically is on-stack; we want to copy the information in it to | |
596 | * an malloced blame_entry that is already on the linked list of the | |
597 | * scoreboard. The origin of dst loses a refcnt while the origin of src | |
598 | * gains one. | |
599 | */ | |
cee7f245 JH |
600 | static void dup_entry(struct blame_entry *dst, struct blame_entry *src) |
601 | { | |
602 | struct blame_entry *p, *n; | |
54a4c617 | 603 | |
cee7f245 JH |
604 | p = dst->prev; |
605 | n = dst->next; | |
54a4c617 JH |
606 | origin_incref(src->suspect); |
607 | origin_decref(dst->suspect); | |
cee7f245 JH |
608 | memcpy(dst, src, sizeof(*src)); |
609 | dst->prev = p; | |
610 | dst->next = n; | |
5ff62c30 | 611 | dst->score = 0; |
cee7f245 JH |
612 | } |
613 | ||
614 | static const char *nth_line(struct scoreboard *sb, int lno) | |
615 | { | |
616 | return sb->final_buf + sb->lineno[lno]; | |
617 | } | |
618 | ||
1732a1fd JH |
619 | /* |
620 | * It is known that lines between tlno to same came from parent, and e | |
621 | * has an overlap with that range. it also is known that parent's | |
622 | * line plno corresponds to e's line tlno. | |
623 | * | |
624 | * <---- e -----> | |
625 | * <------> | |
626 | * <------------> | |
627 | * <------------> | |
628 | * <------------------> | |
629 | * | |
630 | * Split e into potentially three parts; before this chunk, the chunk | |
631 | * to be blamed for the parent, and after that portion. | |
632 | */ | |
54a4c617 | 633 | static void split_overlap(struct blame_entry *split, |
cee7f245 JH |
634 | struct blame_entry *e, |
635 | int tlno, int plno, int same, | |
636 | struct origin *parent) | |
637 | { | |
cee7f245 JH |
638 | int chunk_end_lno; |
639 | memset(split, 0, sizeof(struct blame_entry [3])); | |
640 | ||
641 | if (e->s_lno < tlno) { | |
642 | /* there is a pre-chunk part not blamed on parent */ | |
54a4c617 | 643 | split[0].suspect = origin_incref(e->suspect); |
cee7f245 JH |
644 | split[0].lno = e->lno; |
645 | split[0].s_lno = e->s_lno; | |
646 | split[0].num_lines = tlno - e->s_lno; | |
647 | split[1].lno = e->lno + tlno - e->s_lno; | |
648 | split[1].s_lno = plno; | |
649 | } | |
650 | else { | |
651 | split[1].lno = e->lno; | |
652 | split[1].s_lno = plno + (e->s_lno - tlno); | |
653 | } | |
654 | ||
655 | if (same < e->s_lno + e->num_lines) { | |
656 | /* there is a post-chunk part not blamed on parent */ | |
54a4c617 | 657 | split[2].suspect = origin_incref(e->suspect); |
cee7f245 JH |
658 | split[2].lno = e->lno + (same - e->s_lno); |
659 | split[2].s_lno = e->s_lno + (same - e->s_lno); | |
660 | split[2].num_lines = e->s_lno + e->num_lines - same; | |
661 | chunk_end_lno = split[2].lno; | |
662 | } | |
663 | else | |
664 | chunk_end_lno = e->lno + e->num_lines; | |
665 | split[1].num_lines = chunk_end_lno - split[1].lno; | |
666 | ||
1732a1fd JH |
667 | /* |
668 | * if it turns out there is nothing to blame the parent for, | |
669 | * forget about the splitting. !split[1].suspect signals this. | |
670 | */ | |
cee7f245 JH |
671 | if (split[1].num_lines < 1) |
672 | return; | |
54a4c617 | 673 | split[1].suspect = origin_incref(parent); |
cee7f245 JH |
674 | } |
675 | ||
1732a1fd JH |
676 | /* |
677 | * split_overlap() divided an existing blame e into up to three parts | |
678 | * in split. Adjust the linked list of blames in the scoreboard to | |
679 | * reflect the split. | |
680 | */ | |
cee7f245 | 681 | static void split_blame(struct scoreboard *sb, |
54a4c617 | 682 | struct blame_entry *split, |
cee7f245 JH |
683 | struct blame_entry *e) |
684 | { | |
685 | struct blame_entry *new_entry; | |
686 | ||
687 | if (split[0].suspect && split[2].suspect) { | |
1732a1fd | 688 | /* The first part (reuse storage for the existing entry e) */ |
cee7f245 JH |
689 | dup_entry(e, &split[0]); |
690 | ||
1732a1fd | 691 | /* The last part -- me */ |
cee7f245 JH |
692 | new_entry = xmalloc(sizeof(*new_entry)); |
693 | memcpy(new_entry, &(split[2]), sizeof(struct blame_entry)); | |
694 | add_blame_entry(sb, new_entry); | |
695 | ||
1732a1fd | 696 | /* ... and the middle part -- parent */ |
cee7f245 JH |
697 | new_entry = xmalloc(sizeof(*new_entry)); |
698 | memcpy(new_entry, &(split[1]), sizeof(struct blame_entry)); | |
699 | add_blame_entry(sb, new_entry); | |
700 | } | |
701 | else if (!split[0].suspect && !split[2].suspect) | |
1732a1fd JH |
702 | /* |
703 | * The parent covers the entire area; reuse storage for | |
704 | * e and replace it with the parent. | |
705 | */ | |
cee7f245 JH |
706 | dup_entry(e, &split[1]); |
707 | else if (split[0].suspect) { | |
1732a1fd | 708 | /* me and then parent */ |
cee7f245 JH |
709 | dup_entry(e, &split[0]); |
710 | ||
711 | new_entry = xmalloc(sizeof(*new_entry)); | |
712 | memcpy(new_entry, &(split[1]), sizeof(struct blame_entry)); | |
713 | add_blame_entry(sb, new_entry); | |
714 | } | |
715 | else { | |
1732a1fd | 716 | /* parent and then me */ |
cee7f245 JH |
717 | dup_entry(e, &split[1]); |
718 | ||
719 | new_entry = xmalloc(sizeof(*new_entry)); | |
720 | memcpy(new_entry, &(split[2]), sizeof(struct blame_entry)); | |
721 | add_blame_entry(sb, new_entry); | |
722 | } | |
723 | ||
54a4c617 | 724 | if (DEBUG) { /* sanity */ |
cee7f245 | 725 | struct blame_entry *ent; |
612702e8 | 726 | int lno = sb->ent->lno, corrupt = 0; |
cee7f245 JH |
727 | |
728 | for (ent = sb->ent; ent; ent = ent->next) { | |
729 | if (lno != ent->lno) | |
730 | corrupt = 1; | |
731 | if (ent->s_lno < 0) | |
732 | corrupt = 1; | |
733 | lno += ent->num_lines; | |
734 | } | |
735 | if (corrupt) { | |
612702e8 | 736 | lno = sb->ent->lno; |
cee7f245 JH |
737 | for (ent = sb->ent; ent; ent = ent->next) { |
738 | printf("L %8d l %8d n %8d\n", | |
739 | lno, ent->lno, ent->num_lines); | |
740 | lno = ent->lno + ent->num_lines; | |
741 | } | |
742 | die("oops"); | |
743 | } | |
744 | } | |
745 | } | |
746 | ||
1732a1fd JH |
747 | /* |
748 | * After splitting the blame, the origins used by the | |
749 | * on-stack blame_entry should lose one refcnt each. | |
750 | */ | |
54a4c617 JH |
751 | static void decref_split(struct blame_entry *split) |
752 | { | |
753 | int i; | |
754 | ||
755 | for (i = 0; i < 3; i++) | |
756 | origin_decref(split[i].suspect); | |
757 | } | |
758 | ||
1732a1fd JH |
759 | /* |
760 | * Helper for blame_chunk(). blame_entry e is known to overlap with | |
761 | * the patch hunk; split it and pass blame to the parent. | |
762 | */ | |
cee7f245 JH |
763 | static void blame_overlap(struct scoreboard *sb, struct blame_entry *e, |
764 | int tlno, int plno, int same, | |
765 | struct origin *parent) | |
766 | { | |
767 | struct blame_entry split[3]; | |
768 | ||
769 | split_overlap(split, e, tlno, plno, same, parent); | |
54a4c617 JH |
770 | if (split[1].suspect) |
771 | split_blame(sb, split, e); | |
772 | decref_split(split); | |
cee7f245 JH |
773 | } |
774 | ||
1732a1fd JH |
775 | /* |
776 | * Find the line number of the last line the target is suspected for. | |
777 | */ | |
cee7f245 JH |
778 | static int find_last_in_target(struct scoreboard *sb, struct origin *target) |
779 | { | |
780 | struct blame_entry *e; | |
781 | int last_in_target = -1; | |
782 | ||
783 | for (e = sb->ent; e; e = e->next) { | |
171dccd5 | 784 | if (e->guilty || !same_suspect(e->suspect, target)) |
cee7f245 JH |
785 | continue; |
786 | if (last_in_target < e->s_lno + e->num_lines) | |
787 | last_in_target = e->s_lno + e->num_lines; | |
788 | } | |
789 | return last_in_target; | |
790 | } | |
791 | ||
1732a1fd JH |
792 | /* |
793 | * Process one hunk from the patch between the current suspect for | |
794 | * blame_entry e and its parent. Find and split the overlap, and | |
795 | * pass blame to the overlapping part to the parent. | |
796 | */ | |
cee7f245 JH |
797 | static void blame_chunk(struct scoreboard *sb, |
798 | int tlno, int plno, int same, | |
799 | struct origin *target, struct origin *parent) | |
800 | { | |
612702e8 | 801 | struct blame_entry *e; |
cee7f245 | 802 | |
612702e8 | 803 | for (e = sb->ent; e; e = e->next) { |
171dccd5 | 804 | if (e->guilty || !same_suspect(e->suspect, target)) |
cee7f245 JH |
805 | continue; |
806 | if (same <= e->s_lno) | |
807 | continue; | |
808 | if (tlno < e->s_lno + e->num_lines) | |
809 | blame_overlap(sb, e, tlno, plno, same, parent); | |
810 | } | |
811 | } | |
812 | ||
1732a1fd JH |
813 | /* |
814 | * We are looking at the origin 'target' and aiming to pass blame | |
815 | * for the lines it is suspected to its parent. Run diff to find | |
816 | * which lines came from parent and pass blame for them. | |
817 | */ | |
cee7f245 JH |
818 | static int pass_blame_to_parent(struct scoreboard *sb, |
819 | struct origin *target, | |
820 | struct origin *parent) | |
821 | { | |
822 | int i, last_in_target, plno, tlno; | |
823 | struct patch *patch; | |
824 | ||
825 | last_in_target = find_last_in_target(sb, target); | |
826 | if (last_in_target < 0) | |
827 | return 1; /* nothing remains for this target */ | |
828 | ||
829 | patch = get_patch(parent, target); | |
830 | plno = tlno = 0; | |
831 | for (i = 0; i < patch->num; i++) { | |
832 | struct chunk *chunk = &patch->chunks[i]; | |
833 | ||
cee7f245 JH |
834 | blame_chunk(sb, tlno, plno, chunk->same, target, parent); |
835 | plno = chunk->p_next; | |
836 | tlno = chunk->t_next; | |
837 | } | |
1732a1fd | 838 | /* The rest (i.e. anything after tlno) are the same as the parent */ |
cee7f245 JH |
839 | blame_chunk(sb, tlno, plno, last_in_target, target, parent); |
840 | ||
841 | free_patch(patch); | |
842 | return 0; | |
843 | } | |
844 | ||
1732a1fd JH |
845 | /* |
846 | * The lines in blame_entry after splitting blames many times can become | |
847 | * very small and trivial, and at some point it becomes pointless to | |
848 | * blame the parents. E.g. "\t\t}\n\t}\n\n" appears everywhere in any | |
849 | * ordinary C program, and it is not worth to say it was copied from | |
850 | * totally unrelated file in the parent. | |
851 | * | |
852 | * Compute how trivial the lines in the blame_entry are. | |
853 | */ | |
5ff62c30 JH |
854 | static unsigned ent_score(struct scoreboard *sb, struct blame_entry *e) |
855 | { | |
856 | unsigned score; | |
857 | const char *cp, *ep; | |
858 | ||
859 | if (e->score) | |
860 | return e->score; | |
861 | ||
612702e8 | 862 | score = 1; |
5ff62c30 JH |
863 | cp = nth_line(sb, e->lno); |
864 | ep = nth_line(sb, e->lno + e->num_lines); | |
865 | while (cp < ep) { | |
866 | unsigned ch = *((unsigned char *)cp); | |
867 | if (isalnum(ch)) | |
868 | score++; | |
869 | cp++; | |
870 | } | |
871 | e->score = score; | |
872 | return score; | |
873 | } | |
874 | ||
1732a1fd JH |
875 | /* |
876 | * best_so_far[] and this[] are both a split of an existing blame_entry | |
877 | * that passes blame to the parent. Maintain best_so_far the best split | |
878 | * so far, by comparing this and best_so_far and copying this into | |
879 | * bst_so_far as needed. | |
880 | */ | |
5ff62c30 | 881 | static void copy_split_if_better(struct scoreboard *sb, |
54a4c617 JH |
882 | struct blame_entry *best_so_far, |
883 | struct blame_entry *this) | |
d24bba80 | 884 | { |
54a4c617 JH |
885 | int i; |
886 | ||
d24bba80 JH |
887 | if (!this[1].suspect) |
888 | return; | |
5ff62c30 JH |
889 | if (best_so_far[1].suspect) { |
890 | if (ent_score(sb, &this[1]) < ent_score(sb, &best_so_far[1])) | |
891 | return; | |
892 | } | |
54a4c617 JH |
893 | |
894 | for (i = 0; i < 3; i++) | |
895 | origin_incref(this[i].suspect); | |
896 | decref_split(best_so_far); | |
d24bba80 JH |
897 | memcpy(best_so_far, this, sizeof(struct blame_entry [3])); |
898 | } | |
899 | ||
dd166aa8 JH |
900 | /* |
901 | * We are looking at a part of the final image represented by | |
902 | * ent (tlno and same are offset by ent->s_lno). | |
903 | * tlno is where we are looking at in the final image. | |
904 | * up to (but not including) same match preimage. | |
905 | * plno is where we are looking at in the preimage. | |
906 | * | |
907 | * <-------------- final image ----------------------> | |
908 | * <------ent------> | |
909 | * ^tlno ^same | |
910 | * <---------preimage-----> | |
911 | * ^plno | |
912 | * | |
913 | * All line numbers are 0-based. | |
914 | */ | |
915 | static void handle_split(struct scoreboard *sb, | |
916 | struct blame_entry *ent, | |
917 | int tlno, int plno, int same, | |
918 | struct origin *parent, | |
919 | struct blame_entry *split) | |
920 | { | |
921 | if (ent->num_lines <= tlno) | |
922 | return; | |
923 | if (tlno < same) { | |
924 | struct blame_entry this[3]; | |
925 | tlno += ent->s_lno; | |
926 | same += ent->s_lno; | |
927 | split_overlap(this, ent, tlno, plno, same, parent); | |
928 | copy_split_if_better(sb, split, this); | |
929 | decref_split(this); | |
930 | } | |
931 | } | |
932 | ||
1732a1fd JH |
933 | /* |
934 | * Find the lines from parent that are the same as ent so that | |
935 | * we can pass blames to it. file_p has the blob contents for | |
936 | * the parent. | |
937 | */ | |
d24bba80 JH |
938 | static void find_copy_in_blob(struct scoreboard *sb, |
939 | struct blame_entry *ent, | |
940 | struct origin *parent, | |
54a4c617 | 941 | struct blame_entry *split, |
d24bba80 JH |
942 | mmfile_t *file_p) |
943 | { | |
944 | const char *cp; | |
945 | int cnt; | |
946 | mmfile_t file_o; | |
947 | struct patch *patch; | |
948 | int i, plno, tlno; | |
949 | ||
1732a1fd JH |
950 | /* |
951 | * Prepare mmfile that contains only the lines in ent. | |
952 | */ | |
d24bba80 JH |
953 | cp = nth_line(sb, ent->lno); |
954 | file_o.ptr = (char*) cp; | |
955 | cnt = ent->num_lines; | |
956 | ||
957 | while (cnt && cp < sb->final_buf + sb->final_buf_size) { | |
958 | if (*cp++ == '\n') | |
959 | cnt--; | |
960 | } | |
961 | file_o.size = cp - file_o.ptr; | |
962 | ||
963 | patch = compare_buffer(file_p, &file_o, 1); | |
964 | ||
dd166aa8 JH |
965 | /* |
966 | * file_o is a part of final image we are annotating. | |
967 | * file_p partially may match that image. | |
968 | */ | |
d24bba80 JH |
969 | memset(split, 0, sizeof(struct blame_entry [3])); |
970 | plno = tlno = 0; | |
971 | for (i = 0; i < patch->num; i++) { | |
972 | struct chunk *chunk = &patch->chunks[i]; | |
973 | ||
dd166aa8 | 974 | handle_split(sb, ent, tlno, plno, chunk->same, parent, split); |
d24bba80 JH |
975 | plno = chunk->p_next; |
976 | tlno = chunk->t_next; | |
977 | } | |
dd166aa8 JH |
978 | /* remainder, if any, all match the preimage */ |
979 | handle_split(sb, ent, tlno, plno, ent->num_lines, parent, split); | |
d24bba80 JH |
980 | free_patch(patch); |
981 | } | |
982 | ||
1732a1fd JH |
983 | /* |
984 | * See if lines currently target is suspected for can be attributed to | |
985 | * parent. | |
986 | */ | |
d24bba80 JH |
987 | static int find_move_in_parent(struct scoreboard *sb, |
988 | struct origin *target, | |
989 | struct origin *parent) | |
990 | { | |
650e2f67 | 991 | int last_in_target, made_progress; |
46014766 | 992 | struct blame_entry *e, split[3]; |
d24bba80 | 993 | mmfile_t file_p; |
d24bba80 JH |
994 | |
995 | last_in_target = find_last_in_target(sb, target); | |
996 | if (last_in_target < 0) | |
997 | return 1; /* nothing remains for this target */ | |
998 | ||
c2e525d9 JH |
999 | fill_origin_blob(parent, &file_p); |
1000 | if (!file_p.ptr) | |
d24bba80 | 1001 | return 0; |
d24bba80 | 1002 | |
650e2f67 JH |
1003 | made_progress = 1; |
1004 | while (made_progress) { | |
1005 | made_progress = 0; | |
1006 | for (e = sb->ent; e; e = e->next) { | |
171dccd5 | 1007 | if (e->guilty || !same_suspect(e->suspect, target)) |
650e2f67 JH |
1008 | continue; |
1009 | find_copy_in_blob(sb, e, parent, split, &file_p); | |
1010 | if (split[1].suspect && | |
1011 | blame_move_score < ent_score(sb, &split[1])) { | |
1012 | split_blame(sb, split, e); | |
1013 | made_progress = 1; | |
1014 | } | |
1015 | decref_split(split); | |
1016 | } | |
d24bba80 | 1017 | } |
d24bba80 JH |
1018 | return 0; |
1019 | } | |
1020 | ||
33494784 JH |
1021 | struct blame_list { |
1022 | struct blame_entry *ent; | |
1023 | struct blame_entry split[3]; | |
1024 | }; | |
1025 | ||
1732a1fd JH |
1026 | /* |
1027 | * Count the number of entries the target is suspected for, | |
1028 | * and prepare a list of entry and the best split. | |
1029 | */ | |
33494784 JH |
1030 | static struct blame_list *setup_blame_list(struct scoreboard *sb, |
1031 | struct origin *target, | |
1032 | int *num_ents_p) | |
1033 | { | |
1034 | struct blame_entry *e; | |
1035 | int num_ents, i; | |
1036 | struct blame_list *blame_list = NULL; | |
1037 | ||
33494784 | 1038 | for (e = sb->ent, num_ents = 0; e; e = e->next) |
171dccd5 | 1039 | if (!e->guilty && same_suspect(e->suspect, target)) |
33494784 JH |
1040 | num_ents++; |
1041 | if (num_ents) { | |
1042 | blame_list = xcalloc(num_ents, sizeof(struct blame_list)); | |
1043 | for (e = sb->ent, i = 0; e; e = e->next) | |
171dccd5 | 1044 | if (!e->guilty && same_suspect(e->suspect, target)) |
33494784 JH |
1045 | blame_list[i++].ent = e; |
1046 | } | |
1047 | *num_ents_p = num_ents; | |
1048 | return blame_list; | |
1049 | } | |
1050 | ||
1732a1fd JH |
1051 | /* |
1052 | * For lines target is suspected for, see if we can find code movement | |
1053 | * across file boundary from the parent commit. porigin is the path | |
1054 | * in the parent we already tried. | |
1055 | */ | |
18abd745 JH |
1056 | static int find_copy_in_parent(struct scoreboard *sb, |
1057 | struct origin *target, | |
1058 | struct commit *parent, | |
1059 | struct origin *porigin, | |
1060 | int opt) | |
1061 | { | |
1062 | struct diff_options diff_opts; | |
1063 | const char *paths[1]; | |
aec8fa1f | 1064 | int i, j; |
33494784 JH |
1065 | int retval; |
1066 | struct blame_list *blame_list; | |
aec8fa1f | 1067 | int num_ents; |
18abd745 | 1068 | |
33494784 JH |
1069 | blame_list = setup_blame_list(sb, target, &num_ents); |
1070 | if (!blame_list) | |
18abd745 JH |
1071 | return 1; /* nothing remains for this target */ |
1072 | ||
1073 | diff_setup(&diff_opts); | |
1074 | diff_opts.recursive = 1; | |
1075 | diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT; | |
1076 | ||
18abd745 JH |
1077 | paths[0] = NULL; |
1078 | diff_tree_setup_paths(paths, &diff_opts); | |
1079 | if (diff_setup_done(&diff_opts) < 0) | |
1080 | die("diff-setup"); | |
33494784 JH |
1081 | |
1082 | /* Try "find copies harder" on new path if requested; | |
1083 | * we do not want to use diffcore_rename() actually to | |
1084 | * match things up; find_copies_harder is set only to | |
1085 | * force diff_tree_sha1() to feed all filepairs to diff_queue, | |
1086 | * and this code needs to be after diff_setup_done(), which | |
1087 | * usually makes find-copies-harder imply copy detection. | |
1088 | */ | |
c63777c0 JH |
1089 | if ((opt & PICKAXE_BLAME_COPY_HARDEST) |
1090 | || ((opt & PICKAXE_BLAME_COPY_HARDER) | |
1091 | && (!porigin || strcmp(target->path, porigin->path)))) | |
33494784 JH |
1092 | diff_opts.find_copies_harder = 1; |
1093 | ||
1cfe7733 JH |
1094 | if (is_null_sha1(target->commit->object.sha1)) |
1095 | do_diff_cache(parent->tree->object.sha1, &diff_opts); | |
1096 | else | |
1097 | diff_tree_sha1(parent->tree->object.sha1, | |
1098 | target->commit->tree->object.sha1, | |
1099 | "", &diff_opts); | |
18abd745 | 1100 | |
33494784 JH |
1101 | if (!diff_opts.find_copies_harder) |
1102 | diffcore_std(&diff_opts); | |
18abd745 | 1103 | |
33494784 JH |
1104 | retval = 0; |
1105 | while (1) { | |
1106 | int made_progress = 0; | |
1107 | ||
1108 | for (i = 0; i < diff_queued_diff.nr; i++) { | |
1109 | struct diff_filepair *p = diff_queued_diff.queue[i]; | |
1110 | struct origin *norigin; | |
1111 | mmfile_t file_p; | |
33494784 JH |
1112 | struct blame_entry this[3]; |
1113 | ||
1114 | if (!DIFF_FILE_VALID(p->one)) | |
1115 | continue; /* does not exist in parent */ | |
1116 | if (porigin && !strcmp(p->one->path, porigin->path)) | |
1117 | /* find_move already dealt with this path */ | |
1118 | continue; | |
1119 | ||
1120 | norigin = get_origin(sb, parent, p->one->path); | |
1121 | hashcpy(norigin->blob_sha1, p->one->sha1); | |
c2e525d9 | 1122 | fill_origin_blob(norigin, &file_p); |
33494784 JH |
1123 | if (!file_p.ptr) |
1124 | continue; | |
1125 | ||
1126 | for (j = 0; j < num_ents; j++) { | |
1127 | find_copy_in_blob(sb, blame_list[j].ent, | |
1128 | norigin, this, &file_p); | |
1129 | copy_split_if_better(sb, blame_list[j].split, | |
1130 | this); | |
1131 | decref_split(this); | |
1132 | } | |
33494784 JH |
1133 | origin_decref(norigin); |
1134 | } | |
18abd745 | 1135 | |
aec8fa1f | 1136 | for (j = 0; j < num_ents; j++) { |
33494784 JH |
1137 | struct blame_entry *split = blame_list[j].split; |
1138 | if (split[1].suspect && | |
1139 | blame_copy_score < ent_score(sb, &split[1])) { | |
1140 | split_blame(sb, split, blame_list[j].ent); | |
1141 | made_progress = 1; | |
1142 | } | |
1143 | decref_split(split); | |
18abd745 | 1144 | } |
33494784 | 1145 | free(blame_list); |
aec8fa1f | 1146 | |
33494784 JH |
1147 | if (!made_progress) |
1148 | break; | |
1149 | blame_list = setup_blame_list(sb, target, &num_ents); | |
1150 | if (!blame_list) { | |
1151 | retval = 1; | |
1152 | break; | |
1153 | } | |
18abd745 | 1154 | } |
33494784 | 1155 | diff_flush(&diff_opts); |
18abd745 | 1156 | |
33494784 | 1157 | return retval; |
18abd745 JH |
1158 | } |
1159 | ||
1732a1fd JH |
1160 | /* |
1161 | * The blobs of origin and porigin exactly match, so everything | |
c2e525d9 JH |
1162 | * origin is suspected for can be blamed on the parent. |
1163 | */ | |
1164 | static void pass_whole_blame(struct scoreboard *sb, | |
1165 | struct origin *origin, struct origin *porigin) | |
1166 | { | |
1167 | struct blame_entry *e; | |
1168 | ||
1169 | if (!porigin->file.ptr && origin->file.ptr) { | |
1170 | /* Steal its file */ | |
1171 | porigin->file = origin->file; | |
1172 | origin->file.ptr = NULL; | |
1173 | } | |
1174 | for (e = sb->ent; e; e = e->next) { | |
171dccd5 | 1175 | if (!same_suspect(e->suspect, origin)) |
c2e525d9 JH |
1176 | continue; |
1177 | origin_incref(porigin); | |
1178 | origin_decref(e->suspect); | |
1179 | e->suspect = porigin; | |
1180 | } | |
1181 | } | |
1182 | ||
cee7f245 JH |
1183 | #define MAXPARENT 16 |
1184 | ||
d24bba80 | 1185 | static void pass_blame(struct scoreboard *sb, struct origin *origin, int opt) |
cee7f245 | 1186 | { |
f69e743d | 1187 | int i, pass; |
cee7f245 JH |
1188 | struct commit *commit = origin->commit; |
1189 | struct commit_list *parent; | |
1190 | struct origin *parent_origin[MAXPARENT], *porigin; | |
1191 | ||
1192 | memset(parent_origin, 0, sizeof(parent_origin)); | |
cee7f245 | 1193 | |
f69e743d JH |
1194 | /* The first pass looks for unrenamed path to optimize for |
1195 | * common cases, then we look for renames in the second pass. | |
1196 | */ | |
1197 | for (pass = 0; pass < 2; pass++) { | |
1198 | struct origin *(*find)(struct scoreboard *, | |
1199 | struct commit *, struct origin *); | |
1200 | find = pass ? find_rename : find_origin; | |
1201 | ||
1202 | for (i = 0, parent = commit->parents; | |
1203 | i < MAXPARENT && parent; | |
1204 | parent = parent->next, i++) { | |
1205 | struct commit *p = parent->item; | |
0421d9f8 | 1206 | int j, same; |
f69e743d JH |
1207 | |
1208 | if (parent_origin[i]) | |
1209 | continue; | |
1210 | if (parse_commit(p)) | |
1211 | continue; | |
0d981c67 | 1212 | porigin = find(sb, p, origin); |
f69e743d JH |
1213 | if (!porigin) |
1214 | continue; | |
1215 | if (!hashcmp(porigin->blob_sha1, origin->blob_sha1)) { | |
c2e525d9 | 1216 | pass_whole_blame(sb, origin, porigin); |
f69e743d JH |
1217 | origin_decref(porigin); |
1218 | goto finish; | |
1219 | } | |
0421d9f8 | 1220 | for (j = same = 0; j < i; j++) |
33494784 JH |
1221 | if (parent_origin[j] && |
1222 | !hashcmp(parent_origin[j]->blob_sha1, | |
0421d9f8 JH |
1223 | porigin->blob_sha1)) { |
1224 | same = 1; | |
1225 | break; | |
1226 | } | |
1227 | if (!same) | |
1228 | parent_origin[i] = porigin; | |
1229 | else | |
1230 | origin_decref(porigin); | |
cee7f245 | 1231 | } |
cee7f245 JH |
1232 | } |
1233 | ||
c2e525d9 | 1234 | num_commits++; |
cee7f245 JH |
1235 | for (i = 0, parent = commit->parents; |
1236 | i < MAXPARENT && parent; | |
1237 | parent = parent->next, i++) { | |
1238 | struct origin *porigin = parent_origin[i]; | |
1239 | if (!porigin) | |
1240 | continue; | |
1241 | if (pass_blame_to_parent(sb, origin, porigin)) | |
54a4c617 | 1242 | goto finish; |
cee7f245 | 1243 | } |
d24bba80 JH |
1244 | |
1245 | /* | |
1732a1fd | 1246 | * Optionally find moves in parents' files. |
d24bba80 JH |
1247 | */ |
1248 | if (opt & PICKAXE_BLAME_MOVE) | |
1249 | for (i = 0, parent = commit->parents; | |
1250 | i < MAXPARENT && parent; | |
1251 | parent = parent->next, i++) { | |
1252 | struct origin *porigin = parent_origin[i]; | |
1253 | if (!porigin) | |
1254 | continue; | |
1255 | if (find_move_in_parent(sb, origin, porigin)) | |
54a4c617 | 1256 | goto finish; |
d24bba80 JH |
1257 | } |
1258 | ||
18abd745 | 1259 | /* |
1732a1fd | 1260 | * Optionally find copies from parents' files. |
18abd745 JH |
1261 | */ |
1262 | if (opt & PICKAXE_BLAME_COPY) | |
1263 | for (i = 0, parent = commit->parents; | |
1264 | i < MAXPARENT && parent; | |
1265 | parent = parent->next, i++) { | |
1266 | struct origin *porigin = parent_origin[i]; | |
1267 | if (find_copy_in_parent(sb, origin, parent->item, | |
1268 | porigin, opt)) | |
54a4c617 | 1269 | goto finish; |
18abd745 | 1270 | } |
54a4c617 JH |
1271 | |
1272 | finish: | |
1273 | for (i = 0; i < MAXPARENT; i++) | |
1274 | origin_decref(parent_origin[i]); | |
cee7f245 JH |
1275 | } |
1276 | ||
1732a1fd JH |
1277 | /* |
1278 | * Information on commits, used for output. | |
1279 | */ | |
cee7f245 JH |
1280 | struct commit_info |
1281 | { | |
3a55602e SP |
1282 | const char *author; |
1283 | const char *author_mail; | |
cee7f245 | 1284 | unsigned long author_time; |
3a55602e | 1285 | const char *author_tz; |
cee7f245 JH |
1286 | |
1287 | /* filled only when asked for details */ | |
3a55602e SP |
1288 | const char *committer; |
1289 | const char *committer_mail; | |
cee7f245 | 1290 | unsigned long committer_time; |
3a55602e | 1291 | const char *committer_tz; |
cee7f245 | 1292 | |
3a55602e | 1293 | const char *summary; |
cee7f245 JH |
1294 | }; |
1295 | ||
1732a1fd JH |
1296 | /* |
1297 | * Parse author/committer line in the commit object buffer | |
1298 | */ | |
cee7f245 | 1299 | static void get_ac_line(const char *inbuf, const char *what, |
3a55602e SP |
1300 | int bufsz, char *person, const char **mail, |
1301 | unsigned long *time, const char **tz) | |
cee7f245 | 1302 | { |
f9567384 JH |
1303 | int len, tzlen, maillen; |
1304 | char *tmp, *endp, *timepos; | |
cee7f245 JH |
1305 | |
1306 | tmp = strstr(inbuf, what); | |
1307 | if (!tmp) | |
1308 | goto error_out; | |
1309 | tmp += strlen(what); | |
1310 | endp = strchr(tmp, '\n'); | |
1311 | if (!endp) | |
1312 | len = strlen(tmp); | |
1313 | else | |
1314 | len = endp - tmp; | |
1315 | if (bufsz <= len) { | |
1316 | error_out: | |
1317 | /* Ugh */ | |
3a55602e | 1318 | *mail = *tz = "(unknown)"; |
cee7f245 JH |
1319 | *time = 0; |
1320 | return; | |
1321 | } | |
1322 | memcpy(person, tmp, len); | |
1323 | ||
1324 | tmp = person; | |
1325 | tmp += len; | |
1326 | *tmp = 0; | |
1327 | while (*tmp != ' ') | |
1328 | tmp--; | |
1329 | *tz = tmp+1; | |
f9567384 | 1330 | tzlen = (person+len)-(tmp+1); |
cee7f245 JH |
1331 | |
1332 | *tmp = 0; | |
1333 | while (*tmp != ' ') | |
1334 | tmp--; | |
1335 | *time = strtoul(tmp, NULL, 10); | |
f9567384 | 1336 | timepos = tmp; |
cee7f245 JH |
1337 | |
1338 | *tmp = 0; | |
1339 | while (*tmp != ' ') | |
1340 | tmp--; | |
1341 | *mail = tmp + 1; | |
1342 | *tmp = 0; | |
f9567384 JH |
1343 | maillen = timepos - tmp; |
1344 | ||
1345 | if (!mailmap.nr) | |
1346 | return; | |
1347 | ||
1348 | /* | |
1349 | * mailmap expansion may make the name longer. | |
1350 | * make room by pushing stuff down. | |
1351 | */ | |
1352 | tmp = person + bufsz - (tzlen + 1); | |
1353 | memmove(tmp, *tz, tzlen); | |
1354 | tmp[tzlen] = 0; | |
1355 | *tz = tmp; | |
1356 | ||
1357 | tmp = tmp - (maillen + 1); | |
1358 | memmove(tmp, *mail, maillen); | |
1359 | tmp[maillen] = 0; | |
1360 | *mail = tmp; | |
1361 | ||
1362 | /* | |
1363 | * Now, convert e-mail using mailmap | |
1364 | */ | |
1365 | map_email(&mailmap, tmp + 1, person, tmp-person-1); | |
cee7f245 JH |
1366 | } |
1367 | ||
1368 | static void get_commit_info(struct commit *commit, | |
1369 | struct commit_info *ret, | |
1370 | int detailed) | |
1371 | { | |
1372 | int len; | |
1373 | char *tmp, *endp; | |
1374 | static char author_buf[1024]; | |
1375 | static char committer_buf[1024]; | |
1376 | static char summary_buf[1024]; | |
1377 | ||
1732a1fd JH |
1378 | /* |
1379 | * We've operated without save_commit_buffer, so | |
612702e8 JH |
1380 | * we now need to populate them for output. |
1381 | */ | |
1382 | if (!commit->buffer) { | |
21666f1a | 1383 | enum object_type type; |
612702e8 JH |
1384 | unsigned long size; |
1385 | commit->buffer = | |
21666f1a | 1386 | read_sha1_file(commit->object.sha1, &type, &size); |
612702e8 | 1387 | } |
cee7f245 JH |
1388 | ret->author = author_buf; |
1389 | get_ac_line(commit->buffer, "\nauthor ", | |
1390 | sizeof(author_buf), author_buf, &ret->author_mail, | |
1391 | &ret->author_time, &ret->author_tz); | |
1392 | ||
1393 | if (!detailed) | |
1394 | return; | |
1395 | ||
1396 | ret->committer = committer_buf; | |
1397 | get_ac_line(commit->buffer, "\ncommitter ", | |
1398 | sizeof(committer_buf), committer_buf, &ret->committer_mail, | |
1399 | &ret->committer_time, &ret->committer_tz); | |
1400 | ||
1401 | ret->summary = summary_buf; | |
1402 | tmp = strstr(commit->buffer, "\n\n"); | |
1403 | if (!tmp) { | |
1404 | error_out: | |
1405 | sprintf(summary_buf, "(%s)", sha1_to_hex(commit->object.sha1)); | |
1406 | return; | |
1407 | } | |
1408 | tmp += 2; | |
1409 | endp = strchr(tmp, '\n'); | |
1410 | if (!endp) | |
1cfe7733 | 1411 | endp = tmp + strlen(tmp); |
cee7f245 | 1412 | len = endp - tmp; |
1cfe7733 | 1413 | if (len >= sizeof(summary_buf) || len == 0) |
cee7f245 JH |
1414 | goto error_out; |
1415 | memcpy(summary_buf, tmp, len); | |
1416 | summary_buf[len] = 0; | |
1417 | } | |
1418 | ||
1732a1fd JH |
1419 | /* |
1420 | * To allow LF and other nonportable characters in pathnames, | |
1421 | * they are c-style quoted as needed. | |
1422 | */ | |
46e5e69d JH |
1423 | static void write_filename_info(const char *path) |
1424 | { | |
1425 | printf("filename "); | |
1426 | write_name_quoted(NULL, 0, path, 1, stdout); | |
1427 | putchar('\n'); | |
1428 | } | |
1429 | ||
1732a1fd JH |
1430 | /* |
1431 | * The blame_entry is found to be guilty for the range. Mark it | |
1432 | * as such, and show it in incremental output. | |
1433 | */ | |
717d1462 LT |
1434 | static void found_guilty_entry(struct blame_entry *ent) |
1435 | { | |
1436 | if (ent->guilty) | |
1437 | return; | |
1438 | ent->guilty = 1; | |
1439 | if (incremental) { | |
1440 | struct origin *suspect = ent->suspect; | |
1441 | ||
1442 | printf("%s %d %d %d\n", | |
1443 | sha1_to_hex(suspect->commit->object.sha1), | |
1444 | ent->s_lno + 1, ent->lno + 1, ent->num_lines); | |
1445 | if (!(suspect->commit->object.flags & METAINFO_SHOWN)) { | |
1446 | struct commit_info ci; | |
1447 | suspect->commit->object.flags |= METAINFO_SHOWN; | |
1448 | get_commit_info(suspect->commit, &ci, 1); | |
1449 | printf("author %s\n", ci.author); | |
1450 | printf("author-mail %s\n", ci.author_mail); | |
1451 | printf("author-time %lu\n", ci.author_time); | |
1452 | printf("author-tz %s\n", ci.author_tz); | |
1453 | printf("committer %s\n", ci.committer); | |
1454 | printf("committer-mail %s\n", ci.committer_mail); | |
1455 | printf("committer-time %lu\n", ci.committer_time); | |
1456 | printf("committer-tz %s\n", ci.committer_tz); | |
1457 | printf("summary %s\n", ci.summary); | |
1458 | if (suspect->commit->object.flags & UNINTERESTING) | |
1459 | printf("boundary\n"); | |
1460 | } | |
46e5e69d | 1461 | write_filename_info(suspect->path); |
06f59e9f | 1462 | maybe_flush_or_die(stdout, "stdout"); |
717d1462 LT |
1463 | } |
1464 | } | |
1465 | ||
1732a1fd JH |
1466 | /* |
1467 | * The main loop -- while the scoreboard has lines whose true origin | |
3dff5379 | 1468 | * is still unknown, pick one blame_entry, and allow its current |
1732a1fd JH |
1469 | * suspect to pass blames to its parents. |
1470 | */ | |
717d1462 LT |
1471 | static void assign_blame(struct scoreboard *sb, struct rev_info *revs, int opt) |
1472 | { | |
1473 | while (1) { | |
1474 | struct blame_entry *ent; | |
1475 | struct commit *commit; | |
1476 | struct origin *suspect = NULL; | |
1477 | ||
1478 | /* find one suspect to break down */ | |
1479 | for (ent = sb->ent; !suspect && ent; ent = ent->next) | |
1480 | if (!ent->guilty) | |
1481 | suspect = ent->suspect; | |
1482 | if (!suspect) | |
1483 | return; /* all done */ | |
1484 | ||
1732a1fd JH |
1485 | /* |
1486 | * We will use this suspect later in the loop, | |
1487 | * so hold onto it in the meantime. | |
1488 | */ | |
717d1462 LT |
1489 | origin_incref(suspect); |
1490 | commit = suspect->commit; | |
1491 | if (!commit->object.parsed) | |
1492 | parse_commit(commit); | |
1493 | if (!(commit->object.flags & UNINTERESTING) && | |
1732a1fd | 1494 | !(revs->max_age != -1 && commit->date < revs->max_age)) |
717d1462 LT |
1495 | pass_blame(sb, suspect, opt); |
1496 | else { | |
1497 | commit->object.flags |= UNINTERESTING; | |
1498 | if (commit->object.parsed) | |
1499 | mark_parents_uninteresting(commit); | |
1500 | } | |
1501 | /* treat root commit as boundary */ | |
1502 | if (!commit->parents && !show_root) | |
1503 | commit->object.flags |= UNINTERESTING; | |
1504 | ||
1505 | /* Take responsibility for the remaining entries */ | |
1506 | for (ent = sb->ent; ent; ent = ent->next) | |
171dccd5 | 1507 | if (same_suspect(ent->suspect, suspect)) |
717d1462 LT |
1508 | found_guilty_entry(ent); |
1509 | origin_decref(suspect); | |
1510 | ||
1511 | if (DEBUG) /* sanity */ | |
1512 | sanity_check_refcnt(sb); | |
1513 | } | |
1514 | } | |
1515 | ||
1516 | static const char *format_time(unsigned long time, const char *tz_str, | |
1517 | int show_raw_time) | |
1518 | { | |
1519 | static char time_buf[128]; | |
1520 | time_t t = time; | |
1521 | int minutes, tz; | |
1522 | struct tm *tm; | |
1523 | ||
1524 | if (show_raw_time) { | |
1525 | sprintf(time_buf, "%lu %s", time, tz_str); | |
1526 | return time_buf; | |
1527 | } | |
1528 | ||
1529 | tz = atoi(tz_str); | |
1530 | minutes = tz < 0 ? -tz : tz; | |
1531 | minutes = (minutes / 100)*60 + (minutes % 100); | |
1532 | minutes = tz < 0 ? -minutes : minutes; | |
1533 | t = time + minutes * 60; | |
1534 | tm = gmtime(&t); | |
1535 | ||
1536 | strftime(time_buf, sizeof(time_buf), "%Y-%m-%d %H:%M:%S ", tm); | |
1537 | strcat(time_buf, tz_str); | |
1538 | return time_buf; | |
1539 | } | |
1540 | ||
cee7f245 JH |
1541 | #define OUTPUT_ANNOTATE_COMPAT 001 |
1542 | #define OUTPUT_LONG_OBJECT_NAME 002 | |
1543 | #define OUTPUT_RAW_TIMESTAMP 004 | |
1544 | #define OUTPUT_PORCELAIN 010 | |
1545 | #define OUTPUT_SHOW_NAME 020 | |
1546 | #define OUTPUT_SHOW_NUMBER 040 | |
5ff62c30 | 1547 | #define OUTPUT_SHOW_SCORE 0100 |
093dc5be | 1548 | #define OUTPUT_NO_AUTHOR 0200 |
cee7f245 JH |
1549 | |
1550 | static void emit_porcelain(struct scoreboard *sb, struct blame_entry *ent) | |
1551 | { | |
1552 | int cnt; | |
1553 | const char *cp; | |
1554 | struct origin *suspect = ent->suspect; | |
1555 | char hex[41]; | |
1556 | ||
1557 | strcpy(hex, sha1_to_hex(suspect->commit->object.sha1)); | |
1558 | printf("%s%c%d %d %d\n", | |
1559 | hex, | |
1560 | ent->guilty ? ' ' : '*', // purely for debugging | |
1561 | ent->s_lno + 1, | |
1562 | ent->lno + 1, | |
1563 | ent->num_lines); | |
1564 | if (!(suspect->commit->object.flags & METAINFO_SHOWN)) { | |
1565 | struct commit_info ci; | |
1566 | suspect->commit->object.flags |= METAINFO_SHOWN; | |
1567 | get_commit_info(suspect->commit, &ci, 1); | |
1568 | printf("author %s\n", ci.author); | |
1569 | printf("author-mail %s\n", ci.author_mail); | |
1570 | printf("author-time %lu\n", ci.author_time); | |
1571 | printf("author-tz %s\n", ci.author_tz); | |
1572 | printf("committer %s\n", ci.committer); | |
1573 | printf("committer-mail %s\n", ci.committer_mail); | |
1574 | printf("committer-time %lu\n", ci.committer_time); | |
1575 | printf("committer-tz %s\n", ci.committer_tz); | |
46e5e69d | 1576 | write_filename_info(suspect->path); |
cee7f245 | 1577 | printf("summary %s\n", ci.summary); |
b11121d9 JH |
1578 | if (suspect->commit->object.flags & UNINTERESTING) |
1579 | printf("boundary\n"); | |
cee7f245 JH |
1580 | } |
1581 | else if (suspect->commit->object.flags & MORE_THAN_ONE_PATH) | |
46e5e69d | 1582 | write_filename_info(suspect->path); |
cee7f245 JH |
1583 | |
1584 | cp = nth_line(sb, ent->lno); | |
1585 | for (cnt = 0; cnt < ent->num_lines; cnt++) { | |
1586 | char ch; | |
1587 | if (cnt) | |
1588 | printf("%s %d %d\n", hex, | |
1589 | ent->s_lno + 1 + cnt, | |
1590 | ent->lno + 1 + cnt); | |
1591 | putchar('\t'); | |
1592 | do { | |
1593 | ch = *cp++; | |
1594 | putchar(ch); | |
1595 | } while (ch != '\n' && | |
1596 | cp < sb->final_buf + sb->final_buf_size); | |
1597 | } | |
1598 | } | |
1599 | ||
1600 | static void emit_other(struct scoreboard *sb, struct blame_entry *ent, int opt) | |
1601 | { | |
1602 | int cnt; | |
1603 | const char *cp; | |
1604 | struct origin *suspect = ent->suspect; | |
1605 | struct commit_info ci; | |
1606 | char hex[41]; | |
1607 | int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP); | |
1608 | ||
1609 | get_commit_info(suspect->commit, &ci, 1); | |
1610 | strcpy(hex, sha1_to_hex(suspect->commit->object.sha1)); | |
1611 | ||
1612 | cp = nth_line(sb, ent->lno); | |
1613 | for (cnt = 0; cnt < ent->num_lines; cnt++) { | |
1614 | char ch; | |
b11121d9 JH |
1615 | int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? 40 : 8; |
1616 | ||
1617 | if (suspect->commit->object.flags & UNINTERESTING) { | |
e68989a7 JH |
1618 | if (blank_boundary) |
1619 | memset(hex, ' ', length); | |
1620 | else if (!cmd_is_annotate) { | |
4c10a5ca JH |
1621 | length--; |
1622 | putchar('^'); | |
1623 | } | |
b11121d9 | 1624 | } |
cee7f245 | 1625 | |
b11121d9 | 1626 | printf("%.*s", length, hex); |
cee7f245 JH |
1627 | if (opt & OUTPUT_ANNOTATE_COMPAT) |
1628 | printf("\t(%10s\t%10s\t%d)", ci.author, | |
1629 | format_time(ci.author_time, ci.author_tz, | |
1630 | show_raw_time), | |
1631 | ent->lno + 1 + cnt); | |
1632 | else { | |
5ff62c30 | 1633 | if (opt & OUTPUT_SHOW_SCORE) |
54a4c617 JH |
1634 | printf(" %*d %02d", |
1635 | max_score_digits, ent->score, | |
1636 | ent->suspect->refcnt); | |
cee7f245 JH |
1637 | if (opt & OUTPUT_SHOW_NAME) |
1638 | printf(" %-*.*s", longest_file, longest_file, | |
1639 | suspect->path); | |
1640 | if (opt & OUTPUT_SHOW_NUMBER) | |
1641 | printf(" %*d", max_orig_digits, | |
1642 | ent->s_lno + 1 + cnt); | |
093dc5be JH |
1643 | |
1644 | if (!(opt & OUTPUT_NO_AUTHOR)) | |
1645 | printf(" (%-*.*s %10s", | |
1646 | longest_author, longest_author, | |
1647 | ci.author, | |
1648 | format_time(ci.author_time, | |
1649 | ci.author_tz, | |
1650 | show_raw_time)); | |
1651 | printf(" %*d) ", | |
cee7f245 JH |
1652 | max_digits, ent->lno + 1 + cnt); |
1653 | } | |
1654 | do { | |
1655 | ch = *cp++; | |
1656 | putchar(ch); | |
1657 | } while (ch != '\n' && | |
1658 | cp < sb->final_buf + sb->final_buf_size); | |
1659 | } | |
1660 | } | |
1661 | ||
1662 | static void output(struct scoreboard *sb, int option) | |
1663 | { | |
1664 | struct blame_entry *ent; | |
1665 | ||
1666 | if (option & OUTPUT_PORCELAIN) { | |
1667 | for (ent = sb->ent; ent; ent = ent->next) { | |
1668 | struct blame_entry *oth; | |
1669 | struct origin *suspect = ent->suspect; | |
1670 | struct commit *commit = suspect->commit; | |
1671 | if (commit->object.flags & MORE_THAN_ONE_PATH) | |
1672 | continue; | |
1673 | for (oth = ent->next; oth; oth = oth->next) { | |
1674 | if ((oth->suspect->commit != commit) || | |
1675 | !strcmp(oth->suspect->path, suspect->path)) | |
1676 | continue; | |
1677 | commit->object.flags |= MORE_THAN_ONE_PATH; | |
1678 | break; | |
1679 | } | |
1680 | } | |
1681 | } | |
1682 | ||
1683 | for (ent = sb->ent; ent; ent = ent->next) { | |
1684 | if (option & OUTPUT_PORCELAIN) | |
1685 | emit_porcelain(sb, ent); | |
5ff62c30 | 1686 | else { |
cee7f245 | 1687 | emit_other(sb, ent, option); |
5ff62c30 | 1688 | } |
cee7f245 JH |
1689 | } |
1690 | } | |
1691 | ||
1732a1fd JH |
1692 | /* |
1693 | * To allow quick access to the contents of nth line in the | |
1694 | * final image, prepare an index in the scoreboard. | |
1695 | */ | |
cee7f245 JH |
1696 | static int prepare_lines(struct scoreboard *sb) |
1697 | { | |
1698 | const char *buf = sb->final_buf; | |
1699 | unsigned long len = sb->final_buf_size; | |
1700 | int num = 0, incomplete = 0, bol = 1; | |
1701 | ||
1702 | if (len && buf[len-1] != '\n') | |
1703 | incomplete++; /* incomplete line at the end */ | |
1704 | while (len--) { | |
1705 | if (bol) { | |
1706 | sb->lineno = xrealloc(sb->lineno, | |
1707 | sizeof(int* ) * (num + 1)); | |
1708 | sb->lineno[num] = buf - sb->final_buf; | |
1709 | bol = 0; | |
1710 | } | |
1711 | if (*buf++ == '\n') { | |
1712 | num++; | |
1713 | bol = 1; | |
1714 | } | |
1715 | } | |
1ca6ca87 JH |
1716 | sb->lineno = xrealloc(sb->lineno, |
1717 | sizeof(int* ) * (num + incomplete + 1)); | |
1718 | sb->lineno[num + incomplete] = buf - sb->final_buf; | |
cee7f245 JH |
1719 | sb->num_lines = num + incomplete; |
1720 | return sb->num_lines; | |
1721 | } | |
1722 | ||
1732a1fd JH |
1723 | /* |
1724 | * Add phony grafts for use with -S; this is primarily to | |
1725 | * support git-cvsserver that wants to give a linear history | |
1726 | * to its clients. | |
1727 | */ | |
cee7f245 JH |
1728 | static int read_ancestry(const char *graft_file) |
1729 | { | |
1730 | FILE *fp = fopen(graft_file, "r"); | |
1731 | char buf[1024]; | |
1732 | if (!fp) | |
1733 | return -1; | |
1734 | while (fgets(buf, sizeof(buf), fp)) { | |
1735 | /* The format is just "Commit Parent1 Parent2 ...\n" */ | |
1736 | int len = strlen(buf); | |
1737 | struct commit_graft *graft = read_graft_line(buf, len); | |
8eaf7986 JH |
1738 | if (graft) |
1739 | register_commit_graft(graft, 0); | |
cee7f245 JH |
1740 | } |
1741 | fclose(fp); | |
1742 | return 0; | |
1743 | } | |
1744 | ||
1732a1fd JH |
1745 | /* |
1746 | * How many columns do we need to show line numbers in decimal? | |
1747 | */ | |
cee7f245 JH |
1748 | static int lineno_width(int lines) |
1749 | { | |
f2915045 | 1750 | int i, width; |
cee7f245 | 1751 | |
f2915045 JH |
1752 | for (width = 1, i = 10; i <= lines + 1; width++) |
1753 | i *= 10; | |
1754 | return width; | |
cee7f245 JH |
1755 | } |
1756 | ||
1732a1fd JH |
1757 | /* |
1758 | * How many columns do we need to show line numbers, authors, | |
1759 | * and filenames? | |
1760 | */ | |
cee7f245 JH |
1761 | static void find_alignment(struct scoreboard *sb, int *option) |
1762 | { | |
1763 | int longest_src_lines = 0; | |
1764 | int longest_dst_lines = 0; | |
5ff62c30 | 1765 | unsigned largest_score = 0; |
cee7f245 JH |
1766 | struct blame_entry *e; |
1767 | ||
1768 | for (e = sb->ent; e; e = e->next) { | |
1769 | struct origin *suspect = e->suspect; | |
1770 | struct commit_info ci; | |
1771 | int num; | |
1772 | ||
ab3bb800 JH |
1773 | if (strcmp(suspect->path, sb->path)) |
1774 | *option |= OUTPUT_SHOW_NAME; | |
1775 | num = strlen(suspect->path); | |
1776 | if (longest_file < num) | |
1777 | longest_file = num; | |
cee7f245 JH |
1778 | if (!(suspect->commit->object.flags & METAINFO_SHOWN)) { |
1779 | suspect->commit->object.flags |= METAINFO_SHOWN; | |
1780 | get_commit_info(suspect->commit, &ci, 1); | |
cee7f245 JH |
1781 | num = strlen(ci.author); |
1782 | if (longest_author < num) | |
1783 | longest_author = num; | |
1784 | } | |
1785 | num = e->s_lno + e->num_lines; | |
1786 | if (longest_src_lines < num) | |
1787 | longest_src_lines = num; | |
1788 | num = e->lno + e->num_lines; | |
1789 | if (longest_dst_lines < num) | |
1790 | longest_dst_lines = num; | |
5ff62c30 JH |
1791 | if (largest_score < ent_score(sb, e)) |
1792 | largest_score = ent_score(sb, e); | |
cee7f245 JH |
1793 | } |
1794 | max_orig_digits = lineno_width(longest_src_lines); | |
1795 | max_digits = lineno_width(longest_dst_lines); | |
5ff62c30 | 1796 | max_score_digits = lineno_width(largest_score); |
cee7f245 JH |
1797 | } |
1798 | ||
1732a1fd JH |
1799 | /* |
1800 | * For debugging -- origin is refcounted, and this asserts that | |
1801 | * we do not underflow. | |
1802 | */ | |
54a4c617 JH |
1803 | static void sanity_check_refcnt(struct scoreboard *sb) |
1804 | { | |
1805 | int baa = 0; | |
1806 | struct blame_entry *ent; | |
1807 | ||
1808 | for (ent = sb->ent; ent; ent = ent->next) { | |
ae86ad65 | 1809 | /* Nobody should have zero or negative refcnt */ |
854b97f6 JH |
1810 | if (ent->suspect->refcnt <= 0) { |
1811 | fprintf(stderr, "%s in %s has negative refcnt %d\n", | |
1812 | ent->suspect->path, | |
1813 | sha1_to_hex(ent->suspect->commit->object.sha1), | |
1814 | ent->suspect->refcnt); | |
ae86ad65 | 1815 | baa = 1; |
854b97f6 | 1816 | } |
ae86ad65 JH |
1817 | } |
1818 | for (ent = sb->ent; ent; ent = ent->next) { | |
1819 | /* Mark the ones that haven't been checked */ | |
54a4c617 JH |
1820 | if (0 < ent->suspect->refcnt) |
1821 | ent->suspect->refcnt = -ent->suspect->refcnt; | |
54a4c617 JH |
1822 | } |
1823 | for (ent = sb->ent; ent; ent = ent->next) { | |
1732a1fd JH |
1824 | /* |
1825 | * ... then pick each and see if they have the the | |
1826 | * correct refcnt. | |
54a4c617 JH |
1827 | */ |
1828 | int found; | |
1829 | struct blame_entry *e; | |
1830 | struct origin *suspect = ent->suspect; | |
1831 | ||
1832 | if (0 < suspect->refcnt) | |
1833 | continue; | |
ae86ad65 | 1834 | suspect->refcnt = -suspect->refcnt; /* Unmark */ |
54a4c617 JH |
1835 | for (found = 0, e = sb->ent; e; e = e->next) { |
1836 | if (e->suspect != suspect) | |
1837 | continue; | |
1838 | found++; | |
1839 | } | |
854b97f6 JH |
1840 | if (suspect->refcnt != found) { |
1841 | fprintf(stderr, "%s in %s has refcnt %d, not %d\n", | |
1842 | ent->suspect->path, | |
1843 | sha1_to_hex(ent->suspect->commit->object.sha1), | |
1844 | ent->suspect->refcnt, found); | |
1845 | baa = 2; | |
1846 | } | |
54a4c617 JH |
1847 | } |
1848 | if (baa) { | |
1849 | int opt = 0160; | |
1850 | find_alignment(sb, &opt); | |
1851 | output(sb, opt); | |
854b97f6 | 1852 | die("Baa %d!", baa); |
54a4c617 JH |
1853 | } |
1854 | } | |
1855 | ||
1732a1fd JH |
1856 | /* |
1857 | * Used for the command line parsing; check if the path exists | |
1858 | * in the working tree. | |
1859 | */ | |
cee7f245 JH |
1860 | static int has_path_in_work_tree(const char *path) |
1861 | { | |
1862 | struct stat st; | |
1863 | return !lstat(path, &st); | |
1864 | } | |
1865 | ||
4a0fc95f JH |
1866 | static unsigned parse_score(const char *arg) |
1867 | { | |
1868 | char *end; | |
1869 | unsigned long score = strtoul(arg, &end, 10); | |
1870 | if (*end) | |
1871 | return 0; | |
1872 | return score; | |
1873 | } | |
1874 | ||
20239bae JK |
1875 | static const char *add_prefix(const char *prefix, const char *path) |
1876 | { | |
1877 | if (!prefix || !prefix[0]) | |
1878 | return path; | |
1879 | return prefix_path(prefix, strlen(prefix), path); | |
1880 | } | |
1881 | ||
1732a1fd JH |
1882 | /* |
1883 | * Parsing of (comma separated) one item in the -L option | |
1884 | */ | |
931233bc JH |
1885 | static const char *parse_loc(const char *spec, |
1886 | struct scoreboard *sb, long lno, | |
1887 | long begin, long *ret) | |
1888 | { | |
1889 | char *term; | |
1890 | const char *line; | |
1891 | long num; | |
1892 | int reg_error; | |
1893 | regex_t regexp; | |
1894 | regmatch_t match[1]; | |
1895 | ||
7bd9641d JH |
1896 | /* Allow "-L <something>,+20" to mean starting at <something> |
1897 | * for 20 lines, or "-L <something>,-5" for 5 lines ending at | |
1898 | * <something>. | |
1899 | */ | |
1900 | if (1 < begin && (spec[0] == '+' || spec[0] == '-')) { | |
1901 | num = strtol(spec + 1, &term, 10); | |
1902 | if (term != spec + 1) { | |
1903 | if (spec[0] == '-') | |
1904 | num = 0 - num; | |
1905 | if (0 < num) | |
1906 | *ret = begin + num - 2; | |
1907 | else if (!num) | |
1908 | *ret = begin; | |
1909 | else | |
1910 | *ret = begin + num; | |
1911 | return term; | |
1912 | } | |
1913 | return spec; | |
1914 | } | |
931233bc JH |
1915 | num = strtol(spec, &term, 10); |
1916 | if (term != spec) { | |
1917 | *ret = num; | |
1918 | return term; | |
1919 | } | |
1920 | if (spec[0] != '/') | |
1921 | return spec; | |
1922 | ||
1923 | /* it could be a regexp of form /.../ */ | |
1924 | for (term = (char*) spec + 1; *term && *term != '/'; term++) { | |
1925 | if (*term == '\\') | |
1926 | term++; | |
1927 | } | |
1928 | if (*term != '/') | |
1929 | return spec; | |
1930 | ||
1931 | /* try [spec+1 .. term-1] as regexp */ | |
1932 | *term = 0; | |
1933 | begin--; /* input is in human terms */ | |
1934 | line = nth_line(sb, begin); | |
1935 | ||
1936 | if (!(reg_error = regcomp(®exp, spec + 1, REG_NEWLINE)) && | |
1937 | !(reg_error = regexec(®exp, line, 1, match, 0))) { | |
1938 | const char *cp = line + match[0].rm_so; | |
1939 | const char *nline; | |
1940 | ||
1941 | while (begin++ < lno) { | |
1942 | nline = nth_line(sb, begin); | |
1943 | if (line <= cp && cp < nline) | |
1944 | break; | |
1945 | line = nline; | |
1946 | } | |
1947 | *ret = begin; | |
1948 | regfree(®exp); | |
1949 | *term++ = '/'; | |
1950 | return term; | |
1951 | } | |
1952 | else { | |
1953 | char errbuf[1024]; | |
1954 | regerror(reg_error, ®exp, errbuf, 1024); | |
1955 | die("-L parameter '%s': %s", spec + 1, errbuf); | |
1956 | } | |
1957 | } | |
1958 | ||
1732a1fd JH |
1959 | /* |
1960 | * Parsing of -L option | |
1961 | */ | |
931233bc JH |
1962 | static void prepare_blame_range(struct scoreboard *sb, |
1963 | const char *bottomtop, | |
1964 | long lno, | |
1965 | long *bottom, long *top) | |
1966 | { | |
1967 | const char *term; | |
1968 | ||
1969 | term = parse_loc(bottomtop, sb, lno, 1, bottom); | |
1970 | if (*term == ',') { | |
1971 | term = parse_loc(term + 1, sb, lno, *bottom + 1, top); | |
1972 | if (*term) | |
acca687f | 1973 | usage(blame_usage); |
931233bc JH |
1974 | } |
1975 | if (*term) | |
acca687f | 1976 | usage(blame_usage); |
931233bc JH |
1977 | } |
1978 | ||
4c10a5ca JH |
1979 | static int git_blame_config(const char *var, const char *value) |
1980 | { | |
1981 | if (!strcmp(var, "blame.showroot")) { | |
1982 | show_root = git_config_bool(var, value); | |
1983 | return 0; | |
1984 | } | |
1985 | if (!strcmp(var, "blame.blankboundary")) { | |
1986 | blank_boundary = git_config_bool(var, value); | |
1987 | return 0; | |
1988 | } | |
1989 | return git_default_config(var, value); | |
1990 | } | |
1991 | ||
1cfe7733 JH |
1992 | static struct commit *fake_working_tree_commit(const char *path, const char *contents_from) |
1993 | { | |
1994 | struct commit *commit; | |
1995 | struct origin *origin; | |
1996 | unsigned char head_sha1[20]; | |
1997 | char *buf; | |
1998 | const char *ident; | |
1999 | int fd; | |
2000 | time_t now; | |
2001 | unsigned long fin_size; | |
2002 | int size, len; | |
2003 | struct cache_entry *ce; | |
2004 | unsigned mode; | |
2005 | ||
2006 | if (get_sha1("HEAD", head_sha1)) | |
2007 | die("No such ref: HEAD"); | |
2008 | ||
2009 | time(&now); | |
2010 | commit = xcalloc(1, sizeof(*commit)); | |
2011 | commit->parents = xcalloc(1, sizeof(*commit->parents)); | |
2012 | commit->parents->item = lookup_commit_reference(head_sha1); | |
2013 | commit->object.parsed = 1; | |
2014 | commit->date = now; | |
2015 | commit->object.type = OBJ_COMMIT; | |
2016 | ||
2017 | origin = make_origin(commit, path); | |
2018 | ||
2019 | if (!contents_from || strcmp("-", contents_from)) { | |
2020 | struct stat st; | |
2021 | const char *read_from; | |
2022 | ||
2023 | if (contents_from) { | |
2024 | if (stat(contents_from, &st) < 0) | |
2025 | die("Cannot stat %s", contents_from); | |
2026 | read_from = contents_from; | |
2027 | } | |
2028 | else { | |
2029 | if (lstat(path, &st) < 0) | |
2030 | die("Cannot lstat %s", path); | |
2031 | read_from = path; | |
2032 | } | |
dc49cd76 | 2033 | fin_size = xsize_t(st.st_size); |
1cfe7733 JH |
2034 | buf = xmalloc(fin_size+1); |
2035 | mode = canon_mode(st.st_mode); | |
2036 | switch (st.st_mode & S_IFMT) { | |
2037 | case S_IFREG: | |
2038 | fd = open(read_from, O_RDONLY); | |
2039 | if (fd < 0) | |
2040 | die("cannot open %s", read_from); | |
2041 | if (read_in_full(fd, buf, fin_size) != fin_size) | |
2042 | die("cannot read %s", read_from); | |
2043 | break; | |
2044 | case S_IFLNK: | |
2045 | if (readlink(read_from, buf, fin_size+1) != fin_size) | |
2046 | die("cannot readlink %s", read_from); | |
2047 | break; | |
2048 | default: | |
2049 | die("unsupported file type %s", read_from); | |
2050 | } | |
2051 | } | |
2052 | else { | |
2053 | /* Reading from stdin */ | |
2054 | contents_from = "standard input"; | |
2055 | buf = NULL; | |
2056 | fin_size = 0; | |
2057 | mode = 0; | |
2058 | while (1) { | |
2059 | ssize_t cnt = 8192; | |
2060 | buf = xrealloc(buf, fin_size + cnt); | |
2061 | cnt = xread(0, buf + fin_size, cnt); | |
2062 | if (cnt < 0) | |
2063 | die("read error %s from stdin", | |
2064 | strerror(errno)); | |
2065 | if (!cnt) | |
2066 | break; | |
2067 | fin_size += cnt; | |
2068 | } | |
2069 | buf = xrealloc(buf, fin_size + 1); | |
2070 | } | |
2071 | buf[fin_size] = 0; | |
2072 | origin->file.ptr = buf; | |
2073 | origin->file.size = fin_size; | |
21666f1a | 2074 | pretend_sha1_file(buf, fin_size, OBJ_BLOB, origin->blob_sha1); |
1cfe7733 JH |
2075 | commit->util = origin; |
2076 | ||
2077 | /* | |
2078 | * Read the current index, replace the path entry with | |
2079 | * origin->blob_sha1 without mucking with its mode or type | |
2080 | * bits; we are not going to write this index out -- we just | |
2081 | * want to run "diff-index --cached". | |
2082 | */ | |
2083 | discard_cache(); | |
2084 | read_cache(); | |
2085 | ||
2086 | len = strlen(path); | |
2087 | if (!mode) { | |
2088 | int pos = cache_name_pos(path, len); | |
2089 | if (0 <= pos) | |
2090 | mode = ntohl(active_cache[pos]->ce_mode); | |
2091 | else | |
2092 | /* Let's not bother reading from HEAD tree */ | |
2093 | mode = S_IFREG | 0644; | |
2094 | } | |
2095 | size = cache_entry_size(len); | |
2096 | ce = xcalloc(1, size); | |
2097 | hashcpy(ce->sha1, origin->blob_sha1); | |
2098 | memcpy(ce->name, path, len); | |
2099 | ce->ce_flags = create_ce_flags(len, 0); | |
2100 | ce->ce_mode = create_ce_mode(mode); | |
2101 | add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE); | |
2102 | ||
2103 | /* | |
2104 | * We are not going to write this out, so this does not matter | |
2105 | * right now, but someday we might optimize diff-index --cached | |
2106 | * with cache-tree information. | |
2107 | */ | |
2108 | cache_tree_invalidate_path(active_cache_tree, path); | |
2109 | ||
2110 | commit->buffer = xmalloc(400); | |
2111 | ident = fmt_ident("Not Committed Yet", "not.committed.yet", NULL, 0); | |
1bb88be9 | 2112 | snprintf(commit->buffer, 400, |
1cfe7733 JH |
2113 | "tree 0000000000000000000000000000000000000000\n" |
2114 | "parent %s\n" | |
2115 | "author %s\n" | |
2116 | "committer %s\n\n" | |
2117 | "Version of %s from %s\n", | |
2118 | sha1_to_hex(head_sha1), | |
2119 | ident, ident, path, contents_from ? contents_from : path); | |
2120 | return commit; | |
2121 | } | |
2122 | ||
acca687f | 2123 | int cmd_blame(int argc, const char **argv, const char *prefix) |
cee7f245 JH |
2124 | { |
2125 | struct rev_info revs; | |
2126 | const char *path; | |
2127 | struct scoreboard sb; | |
2128 | struct origin *o; | |
2129 | struct blame_entry *ent; | |
d24bba80 | 2130 | int i, seen_dashdash, unk, opt; |
cee7f245 JH |
2131 | long bottom, top, lno; |
2132 | int output_option = 0; | |
870b39c1 | 2133 | int show_stats = 0; |
cee7f245 JH |
2134 | const char *revs_file = NULL; |
2135 | const char *final_commit_name = NULL; | |
21666f1a | 2136 | enum object_type type; |
931233bc | 2137 | const char *bottomtop = NULL; |
1cfe7733 | 2138 | const char *contents_from = NULL; |
cee7f245 | 2139 | |
e68989a7 JH |
2140 | cmd_is_annotate = !strcmp(argv[0], "annotate"); |
2141 | ||
4c10a5ca | 2142 | git_config(git_blame_config); |
612702e8 JH |
2143 | save_commit_buffer = 0; |
2144 | ||
d24bba80 | 2145 | opt = 0; |
cee7f245 JH |
2146 | seen_dashdash = 0; |
2147 | for (unk = i = 1; i < argc; i++) { | |
2148 | const char *arg = argv[i]; | |
2149 | if (*arg != '-') | |
2150 | break; | |
4c10a5ca JH |
2151 | else if (!strcmp("-b", arg)) |
2152 | blank_boundary = 1; | |
2153 | else if (!strcmp("--root", arg)) | |
2154 | show_root = 1; | |
870b39c1 JH |
2155 | else if (!strcmp(arg, "--show-stats")) |
2156 | show_stats = 1; | |
cee7f245 JH |
2157 | else if (!strcmp("-c", arg)) |
2158 | output_option |= OUTPUT_ANNOTATE_COMPAT; | |
2159 | else if (!strcmp("-t", arg)) | |
2160 | output_option |= OUTPUT_RAW_TIMESTAMP; | |
2161 | else if (!strcmp("-l", arg)) | |
2162 | output_option |= OUTPUT_LONG_OBJECT_NAME; | |
093dc5be JH |
2163 | else if (!strcmp("-s", arg)) |
2164 | output_option |= OUTPUT_NO_AUTHOR; | |
b82871b3 JH |
2165 | else if (!strcmp("-w", arg)) |
2166 | xdl_opts |= XDF_IGNORE_WHITESPACE; | |
cee7f245 JH |
2167 | else if (!strcmp("-S", arg) && ++i < argc) |
2168 | revs_file = argv[i]; | |
599065a3 | 2169 | else if (!prefixcmp(arg, "-M")) { |
d24bba80 | 2170 | opt |= PICKAXE_BLAME_MOVE; |
4a0fc95f JH |
2171 | blame_move_score = parse_score(arg+2); |
2172 | } | |
599065a3 | 2173 | else if (!prefixcmp(arg, "-C")) { |
c63777c0 JH |
2174 | /* |
2175 | * -C enables copy from removed files; | |
2176 | * -C -C enables copy from existing files, but only | |
2177 | * when blaming a new file; | |
2178 | * -C -C -C enables copy from existing files for | |
2179 | * everybody | |
2180 | */ | |
2181 | if (opt & PICKAXE_BLAME_COPY_HARDER) | |
2182 | opt |= PICKAXE_BLAME_COPY_HARDEST; | |
18abd745 JH |
2183 | if (opt & PICKAXE_BLAME_COPY) |
2184 | opt |= PICKAXE_BLAME_COPY_HARDER; | |
2185 | opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE; | |
4a0fc95f | 2186 | blame_copy_score = parse_score(arg+2); |
18abd745 | 2187 | } |
599065a3 | 2188 | else if (!prefixcmp(arg, "-L")) { |
2c40f984 JH |
2189 | if (!arg[2]) { |
2190 | if (++i >= argc) | |
acca687f | 2191 | usage(blame_usage); |
2c40f984 JH |
2192 | arg = argv[i]; |
2193 | } | |
2194 | else | |
2195 | arg += 2; | |
931233bc | 2196 | if (bottomtop) |
cee7f245 | 2197 | die("More than one '-L n,m' option given"); |
931233bc | 2198 | bottomtop = arg; |
cee7f245 | 2199 | } |
1cfe7733 JH |
2200 | else if (!strcmp("--contents", arg)) { |
2201 | if (++i >= argc) | |
2202 | usage(blame_usage); | |
2203 | contents_from = argv[i]; | |
2204 | } | |
717d1462 LT |
2205 | else if (!strcmp("--incremental", arg)) |
2206 | incremental = 1; | |
5ff62c30 JH |
2207 | else if (!strcmp("--score-debug", arg)) |
2208 | output_option |= OUTPUT_SHOW_SCORE; | |
cee7f245 JH |
2209 | else if (!strcmp("-f", arg) || |
2210 | !strcmp("--show-name", arg)) | |
2211 | output_option |= OUTPUT_SHOW_NAME; | |
2212 | else if (!strcmp("-n", arg) || | |
2213 | !strcmp("--show-number", arg)) | |
2214 | output_option |= OUTPUT_SHOW_NUMBER; | |
2215 | else if (!strcmp("-p", arg) || | |
2216 | !strcmp("--porcelain", arg)) | |
2217 | output_option |= OUTPUT_PORCELAIN; | |
2218 | else if (!strcmp("--", arg)) { | |
2219 | seen_dashdash = 1; | |
2220 | i++; | |
2221 | break; | |
2222 | } | |
2223 | else | |
2224 | argv[unk++] = arg; | |
2225 | } | |
2226 | ||
4f0219a4 RS |
2227 | if (!incremental) |
2228 | setup_pager(); | |
2229 | ||
4a0fc95f JH |
2230 | if (!blame_move_score) |
2231 | blame_move_score = BLAME_DEFAULT_MOVE_SCORE; | |
2232 | if (!blame_copy_score) | |
2233 | blame_copy_score = BLAME_DEFAULT_COPY_SCORE; | |
2234 | ||
1732a1fd JH |
2235 | /* |
2236 | * We have collected options unknown to us in argv[1..unk] | |
cee7f245 | 2237 | * which are to be passed to revision machinery if we are |
3dff5379 | 2238 | * going to do the "bottom" processing. |
cee7f245 JH |
2239 | * |
2240 | * The remaining are: | |
2241 | * | |
2242 | * (1) if seen_dashdash, its either | |
2243 | * "-options -- <path>" or | |
2244 | * "-options -- <path> <rev>". | |
2245 | * but the latter is allowed only if there is no | |
2246 | * options that we passed to revision machinery. | |
2247 | * | |
2248 | * (2) otherwise, we may have "--" somewhere later and | |
2249 | * might be looking at the first one of multiple 'rev' | |
2250 | * parameters (e.g. " master ^next ^maint -- path"). | |
2251 | * See if there is a dashdash first, and give the | |
2252 | * arguments before that to revision machinery. | |
2253 | * After that there must be one 'path'. | |
2254 | * | |
2255 | * (3) otherwise, its one of the three: | |
2256 | * "-options <path> <rev>" | |
2257 | * "-options <rev> <path>" | |
2258 | * "-options <path>" | |
2259 | * but again the first one is allowed only if | |
2260 | * there is no options that we passed to revision | |
2261 | * machinery. | |
2262 | */ | |
2263 | ||
2264 | if (seen_dashdash) { | |
2265 | /* (1) */ | |
2266 | if (argc <= i) | |
acca687f | 2267 | usage(blame_usage); |
20239bae | 2268 | path = add_prefix(prefix, argv[i]); |
cee7f245 JH |
2269 | if (i + 1 == argc - 1) { |
2270 | if (unk != 1) | |
acca687f | 2271 | usage(blame_usage); |
cee7f245 JH |
2272 | argv[unk++] = argv[i + 1]; |
2273 | } | |
2274 | else if (i + 1 != argc) | |
2275 | /* garbage at end */ | |
acca687f | 2276 | usage(blame_usage); |
cee7f245 JH |
2277 | } |
2278 | else { | |
2279 | int j; | |
2280 | for (j = i; !seen_dashdash && j < argc; j++) | |
2281 | if (!strcmp(argv[j], "--")) | |
2282 | seen_dashdash = j; | |
2283 | if (seen_dashdash) { | |
f4421325 | 2284 | /* (2) */ |
cee7f245 | 2285 | if (seen_dashdash + 1 != argc - 1) |
acca687f | 2286 | usage(blame_usage); |
20239bae | 2287 | path = add_prefix(prefix, argv[seen_dashdash + 1]); |
cee7f245 JH |
2288 | for (j = i; j < seen_dashdash; j++) |
2289 | argv[unk++] = argv[j]; | |
2290 | } | |
2291 | else { | |
2292 | /* (3) */ | |
f4421325 TK |
2293 | if (argc <= i) |
2294 | usage(blame_usage); | |
20239bae | 2295 | path = add_prefix(prefix, argv[i]); |
cee7f245 JH |
2296 | if (i + 1 == argc - 1) { |
2297 | final_commit_name = argv[i + 1]; | |
2298 | ||
2299 | /* if (unk == 1) we could be getting | |
2300 | * old-style | |
2301 | */ | |
2302 | if (unk == 1 && !has_path_in_work_tree(path)) { | |
20239bae | 2303 | path = add_prefix(prefix, argv[i + 1]); |
cee7f245 JH |
2304 | final_commit_name = argv[i]; |
2305 | } | |
2306 | } | |
2307 | else if (i != argc - 1) | |
acca687f | 2308 | usage(blame_usage); /* garbage at end */ |
cee7f245 JH |
2309 | |
2310 | if (!has_path_in_work_tree(path)) | |
2311 | die("cannot stat path %s: %s", | |
2312 | path, strerror(errno)); | |
2313 | } | |
2314 | } | |
2315 | ||
2316 | if (final_commit_name) | |
2317 | argv[unk++] = final_commit_name; | |
2318 | ||
1732a1fd JH |
2319 | /* |
2320 | * Now we got rev and path. We do not want the path pruning | |
cee7f245 JH |
2321 | * but we may want "bottom" processing. |
2322 | */ | |
6bee4e40 | 2323 | argv[unk++] = "--"; /* terminate the rev name */ |
cee7f245 JH |
2324 | argv[unk] = NULL; |
2325 | ||
2326 | init_revisions(&revs, NULL); | |
1cfe7733 | 2327 | setup_revisions(unk, argv, &revs, NULL); |
cee7f245 JH |
2328 | memset(&sb, 0, sizeof(sb)); |
2329 | ||
1732a1fd JH |
2330 | /* |
2331 | * There must be one and only one positive commit in the | |
cee7f245 JH |
2332 | * revs->pending array. |
2333 | */ | |
2334 | for (i = 0; i < revs.pending.nr; i++) { | |
2335 | struct object *obj = revs.pending.objects[i].item; | |
2336 | if (obj->flags & UNINTERESTING) | |
2337 | continue; | |
2338 | while (obj->type == OBJ_TAG) | |
2339 | obj = deref_tag(obj, NULL, 0); | |
2340 | if (obj->type != OBJ_COMMIT) | |
2341 | die("Non commit %s?", | |
2342 | revs.pending.objects[i].name); | |
2343 | if (sb.final) | |
2344 | die("More than one commit to dig from %s and %s?", | |
2345 | revs.pending.objects[i].name, | |
2346 | final_commit_name); | |
2347 | sb.final = (struct commit *) obj; | |
2348 | final_commit_name = revs.pending.objects[i].name; | |
2349 | } | |
2350 | ||
2351 | if (!sb.final) { | |
1732a1fd JH |
2352 | /* |
2353 | * "--not A B -- path" without anything positive; | |
1cfe7733 JH |
2354 | * do not default to HEAD, but use the working tree |
2355 | * or "--contents". | |
1732a1fd | 2356 | */ |
1cfe7733 JH |
2357 | sb.final = fake_working_tree_commit(path, contents_from); |
2358 | add_pending_object(&revs, &(sb.final->object), ":"); | |
cee7f245 | 2359 | } |
1cfe7733 JH |
2360 | else if (contents_from) |
2361 | die("Cannot use --contents with final commit object name"); | |
cee7f245 | 2362 | |
1732a1fd JH |
2363 | /* |
2364 | * If we have bottom, this will mark the ancestors of the | |
cee7f245 JH |
2365 | * bottom commits we would reach while traversing as |
2366 | * uninteresting. | |
2367 | */ | |
2368 | prepare_revision_walk(&revs); | |
2369 | ||
1cfe7733 JH |
2370 | if (is_null_sha1(sb.final->object.sha1)) { |
2371 | char *buf; | |
2372 | o = sb.final->util; | |
2373 | buf = xmalloc(o->file.size + 1); | |
2374 | memcpy(buf, o->file.ptr, o->file.size + 1); | |
2375 | sb.final_buf = buf; | |
2376 | sb.final_buf_size = o->file.size; | |
2377 | } | |
2378 | else { | |
2379 | o = get_origin(&sb, sb.final, path); | |
2380 | if (fill_blob_sha1(o)) | |
2381 | die("no such path %s in %s", path, final_commit_name); | |
cee7f245 | 2382 | |
21666f1a | 2383 | sb.final_buf = read_sha1_file(o->blob_sha1, &type, |
1cfe7733 JH |
2384 | &sb.final_buf_size); |
2385 | } | |
c2e525d9 | 2386 | num_read_blob++; |
cee7f245 JH |
2387 | lno = prepare_lines(&sb); |
2388 | ||
931233bc JH |
2389 | bottom = top = 0; |
2390 | if (bottomtop) | |
2391 | prepare_blame_range(&sb, bottomtop, lno, &bottom, &top); | |
2392 | if (bottom && top && top < bottom) { | |
2393 | long tmp; | |
2394 | tmp = top; top = bottom; bottom = tmp; | |
2395 | } | |
cee7f245 JH |
2396 | if (bottom < 1) |
2397 | bottom = 1; | |
2398 | if (top < 1) | |
2399 | top = lno; | |
2400 | bottom--; | |
2401 | if (lno < top) | |
2402 | die("file %s has only %lu lines", path, lno); | |
2403 | ||
2404 | ent = xcalloc(1, sizeof(*ent)); | |
2405 | ent->lno = bottom; | |
2406 | ent->num_lines = top - bottom; | |
2407 | ent->suspect = o; | |
2408 | ent->s_lno = bottom; | |
2409 | ||
2410 | sb.ent = ent; | |
2411 | sb.path = path; | |
2412 | ||
2413 | if (revs_file && read_ancestry(revs_file)) | |
2414 | die("reading graft file %s failed: %s", | |
2415 | revs_file, strerror(errno)); | |
2416 | ||
50acc589 | 2417 | read_mailmap(&mailmap, ".mailmap", NULL); |
f9567384 | 2418 | |
d24bba80 | 2419 | assign_blame(&sb, &revs, opt); |
cee7f245 | 2420 | |
717d1462 LT |
2421 | if (incremental) |
2422 | return 0; | |
2423 | ||
cee7f245 JH |
2424 | coalesce(&sb); |
2425 | ||
2426 | if (!(output_option & OUTPUT_PORCELAIN)) | |
2427 | find_alignment(&sb, &output_option); | |
2428 | ||
2429 | output(&sb, output_option); | |
2430 | free((void *)sb.final_buf); | |
2431 | for (ent = sb.ent; ent; ) { | |
2432 | struct blame_entry *e = ent->next; | |
2433 | free(ent); | |
2434 | ent = e; | |
2435 | } | |
c2e525d9 | 2436 | |
870b39c1 | 2437 | if (show_stats) { |
c2e525d9 JH |
2438 | printf("num read blob: %d\n", num_read_blob); |
2439 | printf("num get patch: %d\n", num_get_patch); | |
2440 | printf("num commits: %d\n", num_commits); | |
2441 | } | |
cee7f245 JH |
2442 | return 0; |
2443 | } |