]>
Commit | Line | Data |
---|---|---|
cee7f245 | 1 | /* |
31653c1a | 2 | * Blame |
cee7f245 | 3 | * |
7e6ac6e4 DK |
4 | * Copyright (c) 2006, 2014 by its authors |
5 | * See COPYING for licensing conditions | |
cee7f245 JH |
6 | */ |
7 | ||
8 | #include "cache.h" | |
fb58c8d5 | 9 | #include "refs.h" |
cee7f245 | 10 | #include "builtin.h" |
cee7f245 JH |
11 | #include "commit.h" |
12 | #include "tag.h" | |
13 | #include "tree-walk.h" | |
14 | #include "diff.h" | |
15 | #include "diffcore.h" | |
16 | #include "revision.h" | |
717d1462 | 17 | #include "quote.h" |
cee7f245 | 18 | #include "xdiff-interface.h" |
c455c87c | 19 | #include "string-list.h" |
f9567384 | 20 | #include "mailmap.h" |
7e6ac6e4 | 21 | #include "mergesort.h" |
5817da01 | 22 | #include "parse-options.h" |
7e6ac6e4 | 23 | #include "prio-queue.h" |
ffaf9cc0 | 24 | #include "utf8.h" |
3b8a12e8 | 25 | #include "userdiff.h" |
25ed3412 | 26 | #include "line-range.h" |
58dbfa2e | 27 | #include "line-log.h" |
dbe44faa | 28 | #include "dir.h" |
aba37f49 | 29 | #include "progress.h" |
dc076ae5 | 30 | #include "blame.h" |
cee7f245 | 31 | |
ce41720c | 32 | static char blame_usage[] = N_("git blame [<options>] [<rev-opts>] [<rev>] [--] <file>"); |
5817da01 PH |
33 | |
34 | static const char *blame_opt_usage[] = { | |
35 | blame_usage, | |
36 | "", | |
9c9b4f2f | 37 | N_("<rev-opts> are documented in git-rev-list(1)"), |
5817da01 PH |
38 | NULL |
39 | }; | |
cee7f245 JH |
40 | |
41 | static int longest_file; | |
42 | static int longest_author; | |
43 | static int max_orig_digits; | |
44 | static int max_digits; | |
5ff62c30 | 45 | static int max_score_digits; |
4c10a5ca | 46 | static int show_root; |
85af7929 | 47 | static int reverse; |
4c10a5ca | 48 | static int blank_boundary; |
717d1462 | 49 | static int incremental; |
582aa00b | 50 | static int xdl_opts; |
84393bfd | 51 | static int abbrev = -1; |
3d1aa566 | 52 | static int no_whole_file_rename; |
aba37f49 | 53 | static int show_progress; |
31653c1a | 54 | |
a5481a6c | 55 | static struct date_mode blame_date_mode = { DATE_ISO8601 }; |
31653c1a EL |
56 | static size_t blame_date_width; |
57 | ||
2721ce21 | 58 | static struct string_list mailmap = STRING_LIST_INIT_NODUP; |
cee7f245 | 59 | |
54a4c617 JH |
60 | #ifndef DEBUG |
61 | #define DEBUG 0 | |
62 | #endif | |
63 | ||
d24bba80 | 64 | #define PICKAXE_BLAME_MOVE 01 |
18abd745 JH |
65 | #define PICKAXE_BLAME_COPY 02 |
66 | #define PICKAXE_BLAME_COPY_HARDER 04 | |
c63777c0 | 67 | #define PICKAXE_BLAME_COPY_HARDEST 010 |
d24bba80 | 68 | |
4a0fc95f JH |
69 | static unsigned blame_move_score; |
70 | static unsigned blame_copy_score; | |
71 | #define BLAME_DEFAULT_MOVE_SCORE 20 | |
72 | #define BLAME_DEFAULT_COPY_SCORE 40 | |
73 | ||
208acbfb | 74 | /* Remember to update object flag allocation in object.h */ |
cee7f245 JH |
75 | #define METAINFO_SHOWN (1u<<12) |
76 | #define MORE_THAN_ONE_PATH (1u<<13) | |
77 | ||
aba37f49 ECA |
78 | struct progress_info { |
79 | struct progress *progress; | |
80 | int blamed_lines; | |
81 | }; | |
82 | ||
17a07e2a | 83 | static int diff_hunks(mmfile_t *file_a, mmfile_t *file_b, |
73e1c299 | 84 | xdl_emit_hunk_consume_func_t hunk_func, void *cb_data, int xdl_opts) |
4b4132fd RS |
85 | { |
86 | xpparam_t xpp = {0}; | |
87 | xdemitconf_t xecfg = {0}; | |
85c20c30 | 88 | xdemitcb_t ecb = {NULL}; |
4b4132fd RS |
89 | |
90 | xpp.flags = xdl_opts; | |
4b4132fd RS |
91 | xecfg.hunk_func = hunk_func; |
92 | ecb.priv = cb_data; | |
93 | return xdi_diff(file_a, file_b, &xpp, &xecfg, &ecb); | |
94 | } | |
95 | ||
1732a1fd JH |
96 | /* |
97 | * Given an origin, prepare mmfile_t structure to be used by the | |
98 | * diff machinery | |
99 | */ | |
3b8a12e8 | 100 | static void fill_origin_blob(struct diff_options *opt, |
8449528d | 101 | struct blame_origin *o, mmfile_t *file, int *num_read_blob) |
c2e525d9 JH |
102 | { |
103 | if (!o->file.ptr) { | |
21666f1a | 104 | enum object_type type; |
3b8a12e8 AB |
105 | unsigned long file_size; |
106 | ||
8449528d | 107 | (*num_read_blob)++; |
3b8a12e8 | 108 | if (DIFF_OPT_TST(opt, ALLOW_TEXTCONV) && |
acad70d1 | 109 | textconv_object(o->path, o->mode, &o->blob_oid, 1, &file->ptr, &file_size)) |
3b8a12e8 AB |
110 | ; |
111 | else | |
a7bcfa12 | 112 | file->ptr = read_sha1_file(o->blob_oid.hash, &type, |
113 | &file_size); | |
3b8a12e8 AB |
114 | file->size = file_size; |
115 | ||
ab43e495 JH |
116 | if (!file->ptr) |
117 | die("Cannot read blob %s for path %s", | |
a7bcfa12 | 118 | oid_to_hex(&o->blob_oid), |
ab43e495 | 119 | o->path); |
c2e525d9 JH |
120 | o->file = *file; |
121 | } | |
122 | else | |
123 | *file = o->file; | |
c2e525d9 JH |
124 | } |
125 | ||
f84afb9c | 126 | static void drop_origin_blob(struct blame_origin *o) |
7c3c7962 JH |
127 | { |
128 | if (o->file.ptr) { | |
129 | free(o->file.ptr); | |
130 | o->file.ptr = NULL; | |
131 | } | |
132 | } | |
133 | ||
7e6ac6e4 DK |
134 | /* |
135 | * Any merge of blames happens on lists of blames that arrived via | |
136 | * different parents in a single suspect. In this case, we want to | |
137 | * sort according to the suspect line numbers as opposed to the final | |
138 | * image line numbers. The function body is somewhat longish because | |
139 | * it avoids unnecessary writes. | |
140 | */ | |
141 | ||
142 | static struct blame_entry *blame_merge(struct blame_entry *list1, | |
143 | struct blame_entry *list2) | |
144 | { | |
145 | struct blame_entry *p1 = list1, *p2 = list2, | |
146 | **tail = &list1; | |
147 | ||
148 | if (!p1) | |
149 | return p2; | |
150 | if (!p2) | |
151 | return p1; | |
152 | ||
153 | if (p1->s_lno <= p2->s_lno) { | |
154 | do { | |
155 | tail = &p1->next; | |
156 | if ((p1 = *tail) == NULL) { | |
157 | *tail = p2; | |
158 | return list1; | |
159 | } | |
160 | } while (p1->s_lno <= p2->s_lno); | |
161 | } | |
162 | for (;;) { | |
163 | *tail = p2; | |
164 | do { | |
165 | tail = &p2->next; | |
166 | if ((p2 = *tail) == NULL) { | |
167 | *tail = p1; | |
168 | return list1; | |
169 | } | |
170 | } while (p1->s_lno > p2->s_lno); | |
171 | *tail = p1; | |
172 | do { | |
173 | tail = &p1->next; | |
174 | if ((p1 = *tail) == NULL) { | |
175 | *tail = p2; | |
176 | return list1; | |
177 | } | |
178 | } while (p1->s_lno <= p2->s_lno); | |
179 | } | |
180 | } | |
181 | ||
182 | static void *get_next_blame(const void *p) | |
183 | { | |
184 | return ((struct blame_entry *)p)->next; | |
185 | } | |
186 | ||
187 | static void set_next_blame(void *p1, void *p2) | |
188 | { | |
189 | ((struct blame_entry *)p1)->next = p2; | |
190 | } | |
191 | ||
192 | /* | |
193 | * Final image line numbers are all different, so we don't need a | |
194 | * three-way comparison here. | |
195 | */ | |
196 | ||
197 | static int compare_blame_final(const void *p1, const void *p2) | |
198 | { | |
199 | return ((struct blame_entry *)p1)->lno > ((struct blame_entry *)p2)->lno | |
200 | ? 1 : -1; | |
201 | } | |
202 | ||
203 | static int compare_blame_suspect(const void *p1, const void *p2) | |
204 | { | |
205 | const struct blame_entry *s1 = p1, *s2 = p2; | |
206 | /* | |
207 | * to allow for collating suspects, we sort according to the | |
208 | * respective pointer value as the primary sorting criterion. | |
209 | * The actual relation is pretty unimportant as long as it | |
210 | * establishes a total order. Comparing as integers gives us | |
211 | * that. | |
212 | */ | |
213 | if (s1->suspect != s2->suspect) | |
214 | return (intptr_t)s1->suspect > (intptr_t)s2->suspect ? 1 : -1; | |
215 | if (s1->s_lno == s2->s_lno) | |
216 | return 0; | |
217 | return s1->s_lno > s2->s_lno ? 1 : -1; | |
218 | } | |
219 | ||
7e6ac6e4 DK |
220 | static int compare_commits_by_reverse_commit_date(const void *a, |
221 | const void *b, | |
222 | void *c) | |
223 | { | |
224 | return -compare_commits_by_commit_date(a, b, c); | |
225 | } | |
226 | ||
78b06e66 JS |
227 | static void blame_sort_final(struct blame_scoreboard *sb) |
228 | { | |
229 | sb->ent = llist_mergesort(sb->ent, get_next_blame, set_next_blame, | |
230 | compare_blame_final); | |
231 | } | |
232 | ||
9807b3d6 | 233 | static void sanity_check_refcnt(struct blame_scoreboard *); |
54a4c617 | 234 | |
1732a1fd JH |
235 | /* |
236 | * If two blame entries that are next to each other came from | |
237 | * contiguous lines in the same origin (i.e. <commit, path> pair), | |
238 | * merge them together. | |
239 | */ | |
c6971362 | 240 | static void blame_coalesce(struct blame_scoreboard *sb) |
cee7f245 JH |
241 | { |
242 | struct blame_entry *ent, *next; | |
243 | ||
244 | for (ent = sb->ent; ent && (next = ent->next); ent = next) { | |
0a88f08e | 245 | if (ent->suspect == next->suspect && |
cee7f245 JH |
246 | ent->s_lno + ent->num_lines == next->s_lno) { |
247 | ent->num_lines += next->num_lines; | |
248 | ent->next = next->next; | |
006a0744 | 249 | blame_origin_decref(next->suspect); |
cee7f245 | 250 | free(next); |
46014766 | 251 | ent->score = 0; |
cee7f245 JH |
252 | next = ent; /* again */ |
253 | } | |
254 | } | |
54a4c617 | 255 | |
4149c186 | 256 | if (sb->debug) /* sanity */ |
54a4c617 | 257 | sanity_check_refcnt(sb); |
cee7f245 JH |
258 | } |
259 | ||
7e6ac6e4 DK |
260 | /* |
261 | * Merge the given sorted list of blames into a preexisting origin. | |
262 | * If there were no previous blames to that commit, it is entered into | |
263 | * the commit priority queue of the score board. | |
264 | */ | |
265 | ||
9807b3d6 | 266 | static void queue_blames(struct blame_scoreboard *sb, struct blame_origin *porigin, |
7e6ac6e4 DK |
267 | struct blame_entry *sorted) |
268 | { | |
269 | if (porigin->suspects) | |
270 | porigin->suspects = blame_merge(porigin->suspects, sorted); | |
271 | else { | |
f84afb9c | 272 | struct blame_origin *o; |
7e6ac6e4 DK |
273 | for (o = porigin->commit->util; o; o = o->next) { |
274 | if (o->suspects) { | |
275 | porigin->suspects = sorted; | |
276 | return; | |
277 | } | |
278 | } | |
279 | porigin->suspects = sorted; | |
280 | prio_queue_put(&sb->commits, porigin->commit); | |
281 | } | |
282 | } | |
283 | ||
1732a1fd JH |
284 | /* |
285 | * Fill the blob_sha1 field of an origin if it hasn't, so that later | |
286 | * call to fill_origin_blob() can use it to locate the data. blob_sha1 | |
287 | * for an origin is also used to pass the blame for the entire file to | |
288 | * the parent to detect the case where a child's blob is identical to | |
289 | * that of its parent's. | |
90064710 KS |
290 | * |
291 | * This also fills origin->mode for corresponding tree path. | |
1732a1fd | 292 | */ |
f84afb9c | 293 | static int fill_blob_sha1_and_mode(struct blame_origin *origin) |
f6c0e191 | 294 | { |
a7bcfa12 | 295 | if (!is_null_oid(&origin->blob_oid)) |
f6c0e191 | 296 | return 0; |
ed1c9977 | 297 | if (get_tree_entry(origin->commit->object.oid.hash, |
f6c0e191 | 298 | origin->path, |
a7bcfa12 | 299 | origin->blob_oid.hash, &origin->mode)) |
f6c0e191 | 300 | goto error_out; |
a7bcfa12 | 301 | if (sha1_object_info(origin->blob_oid.hash, NULL) != OBJ_BLOB) |
f6c0e191 JH |
302 | goto error_out; |
303 | return 0; | |
304 | error_out: | |
a7bcfa12 | 305 | oidclr(&origin->blob_oid); |
90064710 | 306 | origin->mode = S_IFINVALID; |
f6c0e191 JH |
307 | return -1; |
308 | } | |
309 | ||
1732a1fd JH |
310 | /* |
311 | * We have an origin -- check if the same path exists in the | |
312 | * parent and return an origin structure to represent it. | |
313 | */ | |
f84afb9c JS |
314 | static struct blame_origin *find_origin(struct commit *parent, |
315 | struct blame_origin *origin) | |
cee7f245 | 316 | { |
f84afb9c | 317 | struct blame_origin *porigin; |
cee7f245 | 318 | struct diff_options diff_opts; |
f6c0e191 JH |
319 | const char *paths[2]; |
320 | ||
7e6ac6e4 DK |
321 | /* First check any existing origins */ |
322 | for (porigin = parent->util; porigin; porigin = porigin->next) | |
323 | if (!strcmp(porigin->path, origin->path)) { | |
1732a1fd JH |
324 | /* |
325 | * The same path between origin and its parent | |
326 | * without renaming -- the most common case. | |
327 | */ | |
006a0744 | 328 | return blame_origin_incref (porigin); |
854b97f6 | 329 | } |
0d981c67 | 330 | |
f6c0e191 JH |
331 | /* See if the origin->path is different between parent |
332 | * and origin first. Most of the time they are the | |
333 | * same and diff-tree is fairly efficient about this. | |
334 | */ | |
335 | diff_setup(&diff_opts); | |
8f67f8ae | 336 | DIFF_OPT_SET(&diff_opts, RECURSIVE); |
f6c0e191 JH |
337 | diff_opts.detect_rename = 0; |
338 | diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT; | |
339 | paths[0] = origin->path; | |
340 | paths[1] = NULL; | |
341 | ||
4a2d5ae2 NTND |
342 | parse_pathspec(&diff_opts.pathspec, |
343 | PATHSPEC_ALL_MAGIC & ~PATHSPEC_LITERAL, | |
344 | PATHSPEC_LITERAL_PATH, "", paths); | |
28452655 | 345 | diff_setup_done(&diff_opts); |
1cfe7733 | 346 | |
f2fd0760 | 347 | if (is_null_oid(&origin->commit->object.oid)) |
ed1c9977 | 348 | do_diff_cache(parent->tree->object.oid.hash, &diff_opts); |
1cfe7733 | 349 | else |
ed1c9977 | 350 | diff_tree_sha1(parent->tree->object.oid.hash, |
351 | origin->commit->tree->object.oid.hash, | |
1cfe7733 | 352 | "", &diff_opts); |
f6c0e191 JH |
353 | diffcore_std(&diff_opts); |
354 | ||
f6c0e191 JH |
355 | if (!diff_queued_diff.nr) { |
356 | /* The path is the same as parent */ | |
8265921c | 357 | porigin = get_origin(parent, origin->path); |
a7bcfa12 | 358 | oidcpy(&porigin->blob_oid, &origin->blob_oid); |
90064710 | 359 | porigin->mode = origin->mode; |
a9b2d424 JH |
360 | } else { |
361 | /* | |
362 | * Since origin->path is a pathspec, if the parent | |
363 | * commit had it as a directory, we will see a whole | |
364 | * bunch of deletion of files in the directory that we | |
365 | * do not care about. | |
366 | */ | |
367 | int i; | |
368 | struct diff_filepair *p = NULL; | |
369 | for (i = 0; i < diff_queued_diff.nr; i++) { | |
370 | const char *name; | |
371 | p = diff_queued_diff.queue[i]; | |
372 | name = p->one->path ? p->one->path : p->two->path; | |
373 | if (!strcmp(name, origin->path)) | |
374 | break; | |
375 | } | |
376 | if (!p) | |
377 | die("internal error in blame::find_origin"); | |
f6c0e191 JH |
378 | switch (p->status) { |
379 | default: | |
acca687f | 380 | die("internal error in blame::find_origin (%c)", |
f6c0e191 JH |
381 | p->status); |
382 | case 'M': | |
8265921c | 383 | porigin = get_origin(parent, origin->path); |
a7bcfa12 | 384 | oidcpy(&porigin->blob_oid, &p->one->oid); |
90064710 | 385 | porigin->mode = p->one->mode; |
f6c0e191 JH |
386 | break; |
387 | case 'A': | |
388 | case 'T': | |
389 | /* Did not exist in parent, or type changed */ | |
390 | break; | |
391 | } | |
392 | } | |
393 | diff_flush(&diff_opts); | |
ed6e8038 | 394 | clear_pathspec(&diff_opts.pathspec); |
f69e743d JH |
395 | return porigin; |
396 | } | |
f6c0e191 | 397 | |
1732a1fd JH |
398 | /* |
399 | * We have an origin -- find the path that corresponds to it in its | |
400 | * parent and return an origin structure to represent it. | |
401 | */ | |
f84afb9c JS |
402 | static struct blame_origin *find_rename(struct commit *parent, |
403 | struct blame_origin *origin) | |
f69e743d | 404 | { |
f84afb9c | 405 | struct blame_origin *porigin = NULL; |
f69e743d JH |
406 | struct diff_options diff_opts; |
407 | int i; | |
cee7f245 JH |
408 | |
409 | diff_setup(&diff_opts); | |
8f67f8ae | 410 | DIFF_OPT_SET(&diff_opts, RECURSIVE); |
cee7f245 JH |
411 | diff_opts.detect_rename = DIFF_DETECT_RENAME; |
412 | diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT; | |
2f3f8b21 | 413 | diff_opts.single_follow = origin->path; |
28452655 | 414 | diff_setup_done(&diff_opts); |
1cfe7733 | 415 | |
f2fd0760 | 416 | if (is_null_oid(&origin->commit->object.oid)) |
ed1c9977 | 417 | do_diff_cache(parent->tree->object.oid.hash, &diff_opts); |
1cfe7733 | 418 | else |
ed1c9977 | 419 | diff_tree_sha1(parent->tree->object.oid.hash, |
420 | origin->commit->tree->object.oid.hash, | |
1cfe7733 | 421 | "", &diff_opts); |
cee7f245 JH |
422 | diffcore_std(&diff_opts); |
423 | ||
424 | for (i = 0; i < diff_queued_diff.nr; i++) { | |
425 | struct diff_filepair *p = diff_queued_diff.queue[i]; | |
612702e8 | 426 | if ((p->status == 'R' || p->status == 'C') && |
f6c0e191 | 427 | !strcmp(p->two->path, origin->path)) { |
8265921c | 428 | porigin = get_origin(parent, p->one->path); |
a7bcfa12 | 429 | oidcpy(&porigin->blob_oid, &p->one->oid); |
90064710 | 430 | porigin->mode = p->one->mode; |
cee7f245 JH |
431 | break; |
432 | } | |
433 | } | |
434 | diff_flush(&diff_opts); | |
ed6e8038 | 435 | clear_pathspec(&diff_opts.pathspec); |
cee7f245 JH |
436 | return porigin; |
437 | } | |
438 | ||
1732a1fd | 439 | /* |
7e6ac6e4 | 440 | * Append a new blame entry to a given output queue. |
1732a1fd | 441 | */ |
dfa3ad32 RS |
442 | static void add_blame_entry(struct blame_entry ***queue, |
443 | const struct blame_entry *src) | |
cee7f245 | 444 | { |
dfa3ad32 RS |
445 | struct blame_entry *e = xmalloc(sizeof(*e)); |
446 | memcpy(e, src, sizeof(*e)); | |
006a0744 | 447 | blame_origin_incref(e->suspect); |
54a4c617 | 448 | |
7e6ac6e4 DK |
449 | e->next = **queue; |
450 | **queue = e; | |
451 | *queue = &e->next; | |
cee7f245 JH |
452 | } |
453 | ||
1732a1fd JH |
454 | /* |
455 | * src typically is on-stack; we want to copy the information in it to | |
7e6ac6e4 DK |
456 | * a malloced blame_entry that gets added to the given queue. The |
457 | * origin of dst loses a refcnt. | |
1732a1fd | 458 | */ |
7e6ac6e4 DK |
459 | static void dup_entry(struct blame_entry ***queue, |
460 | struct blame_entry *dst, struct blame_entry *src) | |
cee7f245 | 461 | { |
006a0744 JS |
462 | blame_origin_incref(src->suspect); |
463 | blame_origin_decref(dst->suspect); | |
cee7f245 | 464 | memcpy(dst, src, sizeof(*src)); |
7e6ac6e4 DK |
465 | dst->next = **queue; |
466 | **queue = dst; | |
467 | *queue = &dst->next; | |
cee7f245 JH |
468 | } |
469 | ||
935202bd | 470 | static const char *blame_nth_line(struct blame_scoreboard *sb, long lno) |
cee7f245 JH |
471 | { |
472 | return sb->final_buf + sb->lineno[lno]; | |
473 | } | |
474 | ||
25ed3412 BY |
475 | static const char *nth_line_cb(void *data, long lno) |
476 | { | |
935202bd | 477 | return blame_nth_line((struct blame_scoreboard *)data, lno); |
25ed3412 BY |
478 | } |
479 | ||
1732a1fd JH |
480 | /* |
481 | * It is known that lines between tlno to same came from parent, and e | |
482 | * has an overlap with that range. it also is known that parent's | |
483 | * line plno corresponds to e's line tlno. | |
484 | * | |
485 | * <---- e -----> | |
486 | * <------> | |
487 | * <------------> | |
488 | * <------------> | |
489 | * <------------------> | |
490 | * | |
491 | * Split e into potentially three parts; before this chunk, the chunk | |
492 | * to be blamed for the parent, and after that portion. | |
493 | */ | |
54a4c617 | 494 | static void split_overlap(struct blame_entry *split, |
cee7f245 JH |
495 | struct blame_entry *e, |
496 | int tlno, int plno, int same, | |
f84afb9c | 497 | struct blame_origin *parent) |
cee7f245 | 498 | { |
cee7f245 JH |
499 | int chunk_end_lno; |
500 | memset(split, 0, sizeof(struct blame_entry [3])); | |
501 | ||
502 | if (e->s_lno < tlno) { | |
503 | /* there is a pre-chunk part not blamed on parent */ | |
006a0744 | 504 | split[0].suspect = blame_origin_incref(e->suspect); |
cee7f245 JH |
505 | split[0].lno = e->lno; |
506 | split[0].s_lno = e->s_lno; | |
507 | split[0].num_lines = tlno - e->s_lno; | |
508 | split[1].lno = e->lno + tlno - e->s_lno; | |
509 | split[1].s_lno = plno; | |
510 | } | |
511 | else { | |
512 | split[1].lno = e->lno; | |
513 | split[1].s_lno = plno + (e->s_lno - tlno); | |
514 | } | |
515 | ||
516 | if (same < e->s_lno + e->num_lines) { | |
517 | /* there is a post-chunk part not blamed on parent */ | |
006a0744 | 518 | split[2].suspect = blame_origin_incref(e->suspect); |
cee7f245 JH |
519 | split[2].lno = e->lno + (same - e->s_lno); |
520 | split[2].s_lno = e->s_lno + (same - e->s_lno); | |
521 | split[2].num_lines = e->s_lno + e->num_lines - same; | |
522 | chunk_end_lno = split[2].lno; | |
523 | } | |
524 | else | |
525 | chunk_end_lno = e->lno + e->num_lines; | |
526 | split[1].num_lines = chunk_end_lno - split[1].lno; | |
527 | ||
1732a1fd JH |
528 | /* |
529 | * if it turns out there is nothing to blame the parent for, | |
530 | * forget about the splitting. !split[1].suspect signals this. | |
531 | */ | |
cee7f245 JH |
532 | if (split[1].num_lines < 1) |
533 | return; | |
006a0744 | 534 | split[1].suspect = blame_origin_incref(parent); |
cee7f245 JH |
535 | } |
536 | ||
1732a1fd JH |
537 | /* |
538 | * split_overlap() divided an existing blame e into up to three parts | |
7e6ac6e4 | 539 | * in split. Any assigned blame is moved to queue to |
1732a1fd JH |
540 | * reflect the split. |
541 | */ | |
7e6ac6e4 DK |
542 | static void split_blame(struct blame_entry ***blamed, |
543 | struct blame_entry ***unblamed, | |
54a4c617 | 544 | struct blame_entry *split, |
cee7f245 JH |
545 | struct blame_entry *e) |
546 | { | |
cee7f245 | 547 | if (split[0].suspect && split[2].suspect) { |
1732a1fd | 548 | /* The first part (reuse storage for the existing entry e) */ |
7e6ac6e4 | 549 | dup_entry(unblamed, e, &split[0]); |
cee7f245 | 550 | |
1732a1fd | 551 | /* The last part -- me */ |
dfa3ad32 | 552 | add_blame_entry(unblamed, &split[2]); |
cee7f245 | 553 | |
1732a1fd | 554 | /* ... and the middle part -- parent */ |
dfa3ad32 | 555 | add_blame_entry(blamed, &split[1]); |
cee7f245 JH |
556 | } |
557 | else if (!split[0].suspect && !split[2].suspect) | |
1732a1fd JH |
558 | /* |
559 | * The parent covers the entire area; reuse storage for | |
560 | * e and replace it with the parent. | |
561 | */ | |
7e6ac6e4 | 562 | dup_entry(blamed, e, &split[1]); |
cee7f245 | 563 | else if (split[0].suspect) { |
1732a1fd | 564 | /* me and then parent */ |
7e6ac6e4 | 565 | dup_entry(unblamed, e, &split[0]); |
dfa3ad32 | 566 | add_blame_entry(blamed, &split[1]); |
cee7f245 JH |
567 | } |
568 | else { | |
1732a1fd | 569 | /* parent and then me */ |
7e6ac6e4 | 570 | dup_entry(blamed, e, &split[1]); |
dfa3ad32 | 571 | add_blame_entry(unblamed, &split[2]); |
cee7f245 JH |
572 | } |
573 | } | |
574 | ||
1732a1fd JH |
575 | /* |
576 | * After splitting the blame, the origins used by the | |
577 | * on-stack blame_entry should lose one refcnt each. | |
578 | */ | |
54a4c617 JH |
579 | static void decref_split(struct blame_entry *split) |
580 | { | |
581 | int i; | |
582 | ||
583 | for (i = 0; i < 3; i++) | |
006a0744 | 584 | blame_origin_decref(split[i].suspect); |
54a4c617 JH |
585 | } |
586 | ||
1732a1fd | 587 | /* |
7e6ac6e4 DK |
588 | * reverse_blame reverses the list given in head, appending tail. |
589 | * That allows us to build lists in reverse order, then reverse them | |
590 | * afterwards. This can be faster than building the list in proper | |
591 | * order right away. The reason is that building in proper order | |
592 | * requires writing a link in the _previous_ element, while building | |
593 | * in reverse order just requires placing the list head into the | |
594 | * _current_ element. | |
1732a1fd | 595 | */ |
cee7f245 | 596 | |
7e6ac6e4 DK |
597 | static struct blame_entry *reverse_blame(struct blame_entry *head, |
598 | struct blame_entry *tail) | |
cee7f245 | 599 | { |
7e6ac6e4 DK |
600 | while (head) { |
601 | struct blame_entry *next = head->next; | |
602 | head->next = tail; | |
603 | tail = head; | |
604 | head = next; | |
cee7f245 | 605 | } |
7e6ac6e4 | 606 | return tail; |
cee7f245 JH |
607 | } |
608 | ||
1732a1fd JH |
609 | /* |
610 | * Process one hunk from the patch between the current suspect for | |
7e6ac6e4 DK |
611 | * blame_entry e and its parent. This first blames any unfinished |
612 | * entries before the chunk (which is where target and parent start | |
613 | * differing) on the parent, and then splits blame entries at the | |
614 | * start and at the end of the difference region. Since use of -M and | |
615 | * -C options may lead to overlapping/duplicate source line number | |
616 | * ranges, all we can rely on from sorting/merging is the order of the | |
617 | * first suspect line number. | |
1732a1fd | 618 | */ |
7e6ac6e4 DK |
619 | static void blame_chunk(struct blame_entry ***dstq, struct blame_entry ***srcq, |
620 | int tlno, int offset, int same, | |
f84afb9c | 621 | struct blame_origin *parent) |
cee7f245 | 622 | { |
7e6ac6e4 DK |
623 | struct blame_entry *e = **srcq; |
624 | struct blame_entry *samep = NULL, *diffp = NULL; | |
cee7f245 | 625 | |
7e6ac6e4 DK |
626 | while (e && e->s_lno < tlno) { |
627 | struct blame_entry *next = e->next; | |
628 | /* | |
629 | * current record starts before differing portion. If | |
630 | * it reaches into it, we need to split it up and | |
631 | * examine the second part separately. | |
632 | */ | |
633 | if (e->s_lno + e->num_lines > tlno) { | |
634 | /* Move second half to a new record */ | |
635 | int len = tlno - e->s_lno; | |
636 | struct blame_entry *n = xcalloc(1, sizeof (struct blame_entry)); | |
637 | n->suspect = e->suspect; | |
638 | n->lno = e->lno + len; | |
639 | n->s_lno = e->s_lno + len; | |
640 | n->num_lines = e->num_lines - len; | |
641 | e->num_lines = len; | |
642 | e->score = 0; | |
643 | /* Push new record to diffp */ | |
644 | n->next = diffp; | |
645 | diffp = n; | |
646 | } else | |
006a0744 | 647 | blame_origin_decref(e->suspect); |
7e6ac6e4 DK |
648 | /* Pass blame for everything before the differing |
649 | * chunk to the parent */ | |
006a0744 | 650 | e->suspect = blame_origin_incref(parent); |
7e6ac6e4 DK |
651 | e->s_lno += offset; |
652 | e->next = samep; | |
653 | samep = e; | |
654 | e = next; | |
655 | } | |
656 | /* | |
657 | * As we don't know how much of a common stretch after this | |
658 | * diff will occur, the currently blamed parts are all that we | |
659 | * can assign to the parent for now. | |
660 | */ | |
661 | ||
662 | if (samep) { | |
663 | **dstq = reverse_blame(samep, **dstq); | |
664 | *dstq = &samep->next; | |
cee7f245 | 665 | } |
7e6ac6e4 DK |
666 | /* |
667 | * Prepend the split off portions: everything after e starts | |
668 | * after the blameable portion. | |
669 | */ | |
670 | e = reverse_blame(diffp, e); | |
671 | ||
672 | /* | |
673 | * Now retain records on the target while parts are different | |
674 | * from the parent. | |
675 | */ | |
676 | samep = NULL; | |
677 | diffp = NULL; | |
678 | while (e && e->s_lno < same) { | |
679 | struct blame_entry *next = e->next; | |
680 | ||
681 | /* | |
682 | * If current record extends into sameness, need to split. | |
683 | */ | |
684 | if (e->s_lno + e->num_lines > same) { | |
685 | /* | |
686 | * Move second half to a new record to be | |
687 | * processed by later chunks | |
688 | */ | |
689 | int len = same - e->s_lno; | |
690 | struct blame_entry *n = xcalloc(1, sizeof (struct blame_entry)); | |
006a0744 | 691 | n->suspect = blame_origin_incref(e->suspect); |
7e6ac6e4 DK |
692 | n->lno = e->lno + len; |
693 | n->s_lno = e->s_lno + len; | |
694 | n->num_lines = e->num_lines - len; | |
695 | e->num_lines = len; | |
696 | e->score = 0; | |
697 | /* Push new record to samep */ | |
698 | n->next = samep; | |
699 | samep = n; | |
700 | } | |
701 | e->next = diffp; | |
702 | diffp = e; | |
703 | e = next; | |
704 | } | |
705 | **srcq = reverse_blame(diffp, reverse_blame(samep, e)); | |
706 | /* Move across elements that are in the unblamable portion */ | |
707 | if (diffp) | |
708 | *srcq = &diffp->next; | |
cee7f245 JH |
709 | } |
710 | ||
cdcab133 | 711 | struct blame_chunk_cb_data { |
f84afb9c | 712 | struct blame_origin *parent; |
7e6ac6e4 DK |
713 | long offset; |
714 | struct blame_entry **dstq; | |
715 | struct blame_entry **srcq; | |
cdcab133 RS |
716 | }; |
717 | ||
7e6ac6e4 | 718 | /* diff chunks are from parent to target */ |
0af596c6 RS |
719 | static int blame_chunk_cb(long start_a, long count_a, |
720 | long start_b, long count_b, void *data) | |
cdcab133 RS |
721 | { |
722 | struct blame_chunk_cb_data *d = data; | |
7e6ac6e4 DK |
723 | if (start_a - start_b != d->offset) |
724 | die("internal error in blame::blame_chunk_cb"); | |
725 | blame_chunk(&d->dstq, &d->srcq, start_b, start_a - start_b, | |
726 | start_b + count_b, d->parent); | |
727 | d->offset = start_a + count_a - (start_b + count_b); | |
0af596c6 | 728 | return 0; |
cdcab133 RS |
729 | } |
730 | ||
1732a1fd JH |
731 | /* |
732 | * We are looking at the origin 'target' and aiming to pass blame | |
733 | * for the lines it is suspected to its parent. Run diff to find | |
734 | * which lines came from parent and pass blame for them. | |
735 | */ | |
9807b3d6 | 736 | static void pass_blame_to_parent(struct blame_scoreboard *sb, |
f84afb9c JS |
737 | struct blame_origin *target, |
738 | struct blame_origin *parent) | |
cee7f245 | 739 | { |
7c4ed8c8 | 740 | mmfile_t file_p, file_o; |
66dbfd55 | 741 | struct blame_chunk_cb_data d; |
7e6ac6e4 | 742 | struct blame_entry *newdest = NULL; |
0af596c6 | 743 | |
7e6ac6e4 DK |
744 | if (!target->suspects) |
745 | return; /* nothing remains for this target */ | |
746 | ||
747 | d.parent = parent; | |
748 | d.offset = 0; | |
749 | d.dstq = &newdest; d.srcq = &target->suspects; | |
cee7f245 | 750 | |
8449528d JS |
751 | fill_origin_blob(&sb->revs->diffopt, parent, &file_p, &sb->num_read_blob); |
752 | fill_origin_blob(&sb->revs->diffopt, target, &file_o, &sb->num_read_blob); | |
753 | sb->num_get_patch++; | |
cee7f245 | 754 | |
73e1c299 | 755 | if (diff_hunks(&file_p, &file_o, blame_chunk_cb, &d, sb->xdl_opts)) |
3efb9880 | 756 | die("unable to generate diff (%s -> %s)", |
f2fd0760 | 757 | oid_to_hex(&parent->commit->object.oid), |
758 | oid_to_hex(&target->commit->object.oid)); | |
7e6ac6e4 DK |
759 | /* The rest are the same as the parent */ |
760 | blame_chunk(&d.dstq, &d.srcq, INT_MAX, d.offset, INT_MAX, parent); | |
761 | *d.dstq = NULL; | |
762 | queue_blames(sb, parent, newdest); | |
cee7f245 | 763 | |
7e6ac6e4 | 764 | return; |
cee7f245 JH |
765 | } |
766 | ||
1732a1fd JH |
767 | /* |
768 | * The lines in blame_entry after splitting blames many times can become | |
769 | * very small and trivial, and at some point it becomes pointless to | |
770 | * blame the parents. E.g. "\t\t}\n\t}\n\n" appears everywhere in any | |
771 | * ordinary C program, and it is not worth to say it was copied from | |
772 | * totally unrelated file in the parent. | |
773 | * | |
774 | * Compute how trivial the lines in the blame_entry are. | |
775 | */ | |
1a31a2d9 | 776 | static unsigned blame_entry_score(struct blame_scoreboard *sb, struct blame_entry *e) |
5ff62c30 JH |
777 | { |
778 | unsigned score; | |
779 | const char *cp, *ep; | |
780 | ||
781 | if (e->score) | |
782 | return e->score; | |
783 | ||
612702e8 | 784 | score = 1; |
935202bd JS |
785 | cp = blame_nth_line(sb, e->lno); |
786 | ep = blame_nth_line(sb, e->lno + e->num_lines); | |
5ff62c30 JH |
787 | while (cp < ep) { |
788 | unsigned ch = *((unsigned char *)cp); | |
789 | if (isalnum(ch)) | |
790 | score++; | |
791 | cp++; | |
792 | } | |
793 | e->score = score; | |
794 | return score; | |
795 | } | |
796 | ||
1732a1fd JH |
797 | /* |
798 | * best_so_far[] and this[] are both a split of an existing blame_entry | |
799 | * that passes blame to the parent. Maintain best_so_far the best split | |
800 | * so far, by comparing this and best_so_far and copying this into | |
801 | * bst_so_far as needed. | |
802 | */ | |
9807b3d6 | 803 | static void copy_split_if_better(struct blame_scoreboard *sb, |
54a4c617 JH |
804 | struct blame_entry *best_so_far, |
805 | struct blame_entry *this) | |
d24bba80 | 806 | { |
54a4c617 JH |
807 | int i; |
808 | ||
d24bba80 JH |
809 | if (!this[1].suspect) |
810 | return; | |
5ff62c30 | 811 | if (best_so_far[1].suspect) { |
1a31a2d9 | 812 | if (blame_entry_score(sb, &this[1]) < blame_entry_score(sb, &best_so_far[1])) |
5ff62c30 JH |
813 | return; |
814 | } | |
54a4c617 JH |
815 | |
816 | for (i = 0; i < 3; i++) | |
006a0744 | 817 | blame_origin_incref(this[i].suspect); |
54a4c617 | 818 | decref_split(best_so_far); |
d24bba80 JH |
819 | memcpy(best_so_far, this, sizeof(struct blame_entry [3])); |
820 | } | |
821 | ||
dd166aa8 JH |
822 | /* |
823 | * We are looking at a part of the final image represented by | |
824 | * ent (tlno and same are offset by ent->s_lno). | |
825 | * tlno is where we are looking at in the final image. | |
826 | * up to (but not including) same match preimage. | |
827 | * plno is where we are looking at in the preimage. | |
828 | * | |
829 | * <-------------- final image ----------------------> | |
830 | * <------ent------> | |
831 | * ^tlno ^same | |
832 | * <---------preimage-----> | |
833 | * ^plno | |
834 | * | |
835 | * All line numbers are 0-based. | |
836 | */ | |
9807b3d6 | 837 | static void handle_split(struct blame_scoreboard *sb, |
dd166aa8 JH |
838 | struct blame_entry *ent, |
839 | int tlno, int plno, int same, | |
f84afb9c | 840 | struct blame_origin *parent, |
dd166aa8 JH |
841 | struct blame_entry *split) |
842 | { | |
843 | if (ent->num_lines <= tlno) | |
844 | return; | |
845 | if (tlno < same) { | |
846 | struct blame_entry this[3]; | |
847 | tlno += ent->s_lno; | |
848 | same += ent->s_lno; | |
849 | split_overlap(this, ent, tlno, plno, same, parent); | |
850 | copy_split_if_better(sb, split, this); | |
851 | decref_split(this); | |
852 | } | |
853 | } | |
854 | ||
cdcab133 | 855 | struct handle_split_cb_data { |
9807b3d6 | 856 | struct blame_scoreboard *sb; |
cdcab133 | 857 | struct blame_entry *ent; |
f84afb9c | 858 | struct blame_origin *parent; |
cdcab133 RS |
859 | struct blame_entry *split; |
860 | long plno; | |
861 | long tlno; | |
862 | }; | |
863 | ||
5d23ec76 RS |
864 | static int handle_split_cb(long start_a, long count_a, |
865 | long start_b, long count_b, void *data) | |
cdcab133 RS |
866 | { |
867 | struct handle_split_cb_data *d = data; | |
5d23ec76 RS |
868 | handle_split(d->sb, d->ent, d->tlno, d->plno, start_b, d->parent, |
869 | d->split); | |
870 | d->plno = start_a + count_a; | |
871 | d->tlno = start_b + count_b; | |
872 | return 0; | |
cdcab133 RS |
873 | } |
874 | ||
1732a1fd JH |
875 | /* |
876 | * Find the lines from parent that are the same as ent so that | |
877 | * we can pass blames to it. file_p has the blob contents for | |
878 | * the parent. | |
879 | */ | |
9807b3d6 | 880 | static void find_copy_in_blob(struct blame_scoreboard *sb, |
d24bba80 | 881 | struct blame_entry *ent, |
f84afb9c | 882 | struct blame_origin *parent, |
54a4c617 | 883 | struct blame_entry *split, |
d24bba80 JH |
884 | mmfile_t *file_p) |
885 | { | |
886 | const char *cp; | |
d24bba80 | 887 | mmfile_t file_o; |
66dbfd55 | 888 | struct handle_split_cb_data d; |
5d23ec76 | 889 | |
66dbfd55 GV |
890 | memset(&d, 0, sizeof(d)); |
891 | d.sb = sb; d.ent = ent; d.parent = parent; d.split = split; | |
1732a1fd JH |
892 | /* |
893 | * Prepare mmfile that contains only the lines in ent. | |
894 | */ | |
935202bd | 895 | cp = blame_nth_line(sb, ent->lno); |
4b25d091 | 896 | file_o.ptr = (char *) cp; |
935202bd | 897 | file_o.size = blame_nth_line(sb, ent->lno + ent->num_lines) - cp; |
d24bba80 | 898 | |
dd166aa8 JH |
899 | /* |
900 | * file_o is a part of final image we are annotating. | |
901 | * file_p partially may match that image. | |
902 | */ | |
d24bba80 | 903 | memset(split, 0, sizeof(struct blame_entry [3])); |
73e1c299 | 904 | if (diff_hunks(file_p, &file_o, handle_split_cb, &d, sb->xdl_opts)) |
3efb9880 | 905 | die("unable to generate diff (%s)", |
f2fd0760 | 906 | oid_to_hex(&parent->commit->object.oid)); |
dd166aa8 | 907 | /* remainder, if any, all match the preimage */ |
cdcab133 | 908 | handle_split(sb, ent, d.tlno, d.plno, ent->num_lines, parent, split); |
d24bba80 JH |
909 | } |
910 | ||
7e6ac6e4 DK |
911 | /* Move all blame entries from list *source that have a score smaller |
912 | * than score_min to the front of list *small. | |
913 | * Returns a pointer to the link pointing to the old head of the small list. | |
914 | */ | |
915 | ||
9807b3d6 | 916 | static struct blame_entry **filter_small(struct blame_scoreboard *sb, |
7e6ac6e4 DK |
917 | struct blame_entry **small, |
918 | struct blame_entry **source, | |
919 | unsigned score_min) | |
920 | { | |
921 | struct blame_entry *p = *source; | |
922 | struct blame_entry *oldsmall = *small; | |
923 | while (p) { | |
1a31a2d9 | 924 | if (blame_entry_score(sb, p) <= score_min) { |
7e6ac6e4 DK |
925 | *small = p; |
926 | small = &p->next; | |
927 | p = *small; | |
928 | } else { | |
929 | *source = p; | |
930 | source = &p->next; | |
931 | p = *source; | |
932 | } | |
933 | } | |
934 | *small = oldsmall; | |
935 | *source = NULL; | |
936 | return small; | |
937 | } | |
938 | ||
1732a1fd JH |
939 | /* |
940 | * See if lines currently target is suspected for can be attributed to | |
941 | * parent. | |
942 | */ | |
9807b3d6 | 943 | static void find_move_in_parent(struct blame_scoreboard *sb, |
7e6ac6e4 DK |
944 | struct blame_entry ***blamed, |
945 | struct blame_entry **toosmall, | |
f84afb9c JS |
946 | struct blame_origin *target, |
947 | struct blame_origin *parent) | |
d24bba80 | 948 | { |
46014766 | 949 | struct blame_entry *e, split[3]; |
7e6ac6e4 DK |
950 | struct blame_entry *unblamed = target->suspects; |
951 | struct blame_entry *leftover = NULL; | |
d24bba80 | 952 | mmfile_t file_p; |
d24bba80 | 953 | |
7e6ac6e4 DK |
954 | if (!unblamed) |
955 | return; /* nothing remains for this target */ | |
d24bba80 | 956 | |
8449528d | 957 | fill_origin_blob(&sb->revs->diffopt, parent, &file_p, &sb->num_read_blob); |
c2e525d9 | 958 | if (!file_p.ptr) |
7e6ac6e4 | 959 | return; |
d24bba80 | 960 | |
7e6ac6e4 DK |
961 | /* At each iteration, unblamed has a NULL-terminated list of |
962 | * entries that have not yet been tested for blame. leftover | |
963 | * contains the reversed list of entries that have been tested | |
964 | * without being assignable to the parent. | |
965 | */ | |
966 | do { | |
967 | struct blame_entry **unblamedtail = &unblamed; | |
968 | struct blame_entry *next; | |
969 | for (e = unblamed; e; e = next) { | |
970 | next = e->next; | |
650e2f67 JH |
971 | find_copy_in_blob(sb, e, parent, split, &file_p); |
972 | if (split[1].suspect && | |
18ec0d62 | 973 | sb->move_score < blame_entry_score(sb, &split[1])) { |
7e6ac6e4 DK |
974 | split_blame(blamed, &unblamedtail, split, e); |
975 | } else { | |
976 | e->next = leftover; | |
977 | leftover = e; | |
650e2f67 JH |
978 | } |
979 | decref_split(split); | |
980 | } | |
7e6ac6e4 | 981 | *unblamedtail = NULL; |
18ec0d62 | 982 | toosmall = filter_small(sb, toosmall, &unblamed, sb->move_score); |
7e6ac6e4 DK |
983 | } while (unblamed); |
984 | target->suspects = reverse_blame(leftover, NULL); | |
d24bba80 JH |
985 | } |
986 | ||
33494784 JH |
987 | struct blame_list { |
988 | struct blame_entry *ent; | |
989 | struct blame_entry split[3]; | |
990 | }; | |
991 | ||
1732a1fd JH |
992 | /* |
993 | * Count the number of entries the target is suspected for, | |
994 | * and prepare a list of entry and the best split. | |
995 | */ | |
7e6ac6e4 | 996 | static struct blame_list *setup_blame_list(struct blame_entry *unblamed, |
33494784 JH |
997 | int *num_ents_p) |
998 | { | |
999 | struct blame_entry *e; | |
1000 | int num_ents, i; | |
1001 | struct blame_list *blame_list = NULL; | |
1002 | ||
7e6ac6e4 DK |
1003 | for (e = unblamed, num_ents = 0; e; e = e->next) |
1004 | num_ents++; | |
33494784 JH |
1005 | if (num_ents) { |
1006 | blame_list = xcalloc(num_ents, sizeof(struct blame_list)); | |
7e6ac6e4 DK |
1007 | for (e = unblamed, i = 0; e; e = e->next) |
1008 | blame_list[i++].ent = e; | |
33494784 JH |
1009 | } |
1010 | *num_ents_p = num_ents; | |
1011 | return blame_list; | |
1012 | } | |
1013 | ||
1732a1fd JH |
1014 | /* |
1015 | * For lines target is suspected for, see if we can find code movement | |
1016 | * across file boundary from the parent commit. porigin is the path | |
1017 | * in the parent we already tried. | |
1018 | */ | |
9807b3d6 | 1019 | static void find_copy_in_parent(struct blame_scoreboard *sb, |
7e6ac6e4 DK |
1020 | struct blame_entry ***blamed, |
1021 | struct blame_entry **toosmall, | |
f84afb9c | 1022 | struct blame_origin *target, |
7e6ac6e4 | 1023 | struct commit *parent, |
f84afb9c | 1024 | struct blame_origin *porigin, |
7e6ac6e4 | 1025 | int opt) |
18abd745 JH |
1026 | { |
1027 | struct diff_options diff_opts; | |
aec8fa1f | 1028 | int i, j; |
33494784 | 1029 | struct blame_list *blame_list; |
aec8fa1f | 1030 | int num_ents; |
7e6ac6e4 DK |
1031 | struct blame_entry *unblamed = target->suspects; |
1032 | struct blame_entry *leftover = NULL; | |
18abd745 | 1033 | |
7e6ac6e4 DK |
1034 | if (!unblamed) |
1035 | return; /* nothing remains for this target */ | |
18abd745 JH |
1036 | |
1037 | diff_setup(&diff_opts); | |
8f67f8ae | 1038 | DIFF_OPT_SET(&diff_opts, RECURSIVE); |
18abd745 JH |
1039 | diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT; |
1040 | ||
28452655 | 1041 | diff_setup_done(&diff_opts); |
33494784 JH |
1042 | |
1043 | /* Try "find copies harder" on new path if requested; | |
1044 | * we do not want to use diffcore_rename() actually to | |
1045 | * match things up; find_copies_harder is set only to | |
1046 | * force diff_tree_sha1() to feed all filepairs to diff_queue, | |
1047 | * and this code needs to be after diff_setup_done(), which | |
1048 | * usually makes find-copies-harder imply copy detection. | |
1049 | */ | |
c63777c0 JH |
1050 | if ((opt & PICKAXE_BLAME_COPY_HARDEST) |
1051 | || ((opt & PICKAXE_BLAME_COPY_HARDER) | |
1052 | && (!porigin || strcmp(target->path, porigin->path)))) | |
8f67f8ae | 1053 | DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER); |
33494784 | 1054 | |
f2fd0760 | 1055 | if (is_null_oid(&target->commit->object.oid)) |
ed1c9977 | 1056 | do_diff_cache(parent->tree->object.oid.hash, &diff_opts); |
1cfe7733 | 1057 | else |
ed1c9977 | 1058 | diff_tree_sha1(parent->tree->object.oid.hash, |
1059 | target->commit->tree->object.oid.hash, | |
1cfe7733 | 1060 | "", &diff_opts); |
18abd745 | 1061 | |
8f67f8ae | 1062 | if (!DIFF_OPT_TST(&diff_opts, FIND_COPIES_HARDER)) |
33494784 | 1063 | diffcore_std(&diff_opts); |
18abd745 | 1064 | |
7e6ac6e4 DK |
1065 | do { |
1066 | struct blame_entry **unblamedtail = &unblamed; | |
1067 | blame_list = setup_blame_list(unblamed, &num_ents); | |
33494784 JH |
1068 | |
1069 | for (i = 0; i < diff_queued_diff.nr; i++) { | |
1070 | struct diff_filepair *p = diff_queued_diff.queue[i]; | |
f84afb9c | 1071 | struct blame_origin *norigin; |
33494784 | 1072 | mmfile_t file_p; |
33494784 JH |
1073 | struct blame_entry this[3]; |
1074 | ||
1075 | if (!DIFF_FILE_VALID(p->one)) | |
1076 | continue; /* does not exist in parent */ | |
5209ac4d AG |
1077 | if (S_ISGITLINK(p->one->mode)) |
1078 | continue; /* ignore git links */ | |
33494784 JH |
1079 | if (porigin && !strcmp(p->one->path, porigin->path)) |
1080 | /* find_move already dealt with this path */ | |
1081 | continue; | |
1082 | ||
8265921c | 1083 | norigin = get_origin(parent, p->one->path); |
a7bcfa12 | 1084 | oidcpy(&norigin->blob_oid, &p->one->oid); |
90064710 | 1085 | norigin->mode = p->one->mode; |
8449528d | 1086 | fill_origin_blob(&sb->revs->diffopt, norigin, &file_p, &sb->num_read_blob); |
33494784 JH |
1087 | if (!file_p.ptr) |
1088 | continue; | |
1089 | ||
1090 | for (j = 0; j < num_ents; j++) { | |
1091 | find_copy_in_blob(sb, blame_list[j].ent, | |
1092 | norigin, this, &file_p); | |
1093 | copy_split_if_better(sb, blame_list[j].split, | |
1094 | this); | |
1095 | decref_split(this); | |
1096 | } | |
006a0744 | 1097 | blame_origin_decref(norigin); |
33494784 | 1098 | } |
18abd745 | 1099 | |
aec8fa1f | 1100 | for (j = 0; j < num_ents; j++) { |
33494784 JH |
1101 | struct blame_entry *split = blame_list[j].split; |
1102 | if (split[1].suspect && | |
18ec0d62 | 1103 | sb->copy_score < blame_entry_score(sb, &split[1])) { |
7e6ac6e4 DK |
1104 | split_blame(blamed, &unblamedtail, split, |
1105 | blame_list[j].ent); | |
1106 | } else { | |
1107 | blame_list[j].ent->next = leftover; | |
1108 | leftover = blame_list[j].ent; | |
33494784 JH |
1109 | } |
1110 | decref_split(split); | |
18abd745 | 1111 | } |
33494784 | 1112 | free(blame_list); |
7e6ac6e4 | 1113 | *unblamedtail = NULL; |
18ec0d62 | 1114 | toosmall = filter_small(sb, toosmall, &unblamed, sb->copy_score); |
7e6ac6e4 DK |
1115 | } while (unblamed); |
1116 | target->suspects = reverse_blame(leftover, NULL); | |
33494784 | 1117 | diff_flush(&diff_opts); |
ed6e8038 | 1118 | clear_pathspec(&diff_opts.pathspec); |
18abd745 JH |
1119 | } |
1120 | ||
1732a1fd JH |
1121 | /* |
1122 | * The blobs of origin and porigin exactly match, so everything | |
c2e525d9 JH |
1123 | * origin is suspected for can be blamed on the parent. |
1124 | */ | |
9807b3d6 | 1125 | static void pass_whole_blame(struct blame_scoreboard *sb, |
f84afb9c | 1126 | struct blame_origin *origin, struct blame_origin *porigin) |
c2e525d9 | 1127 | { |
7e6ac6e4 | 1128 | struct blame_entry *e, *suspects; |
c2e525d9 JH |
1129 | |
1130 | if (!porigin->file.ptr && origin->file.ptr) { | |
1131 | /* Steal its file */ | |
1132 | porigin->file = origin->file; | |
1133 | origin->file.ptr = NULL; | |
1134 | } | |
7e6ac6e4 DK |
1135 | suspects = origin->suspects; |
1136 | origin->suspects = NULL; | |
1137 | for (e = suspects; e; e = e->next) { | |
006a0744 JS |
1138 | blame_origin_incref(porigin); |
1139 | blame_origin_decref(e->suspect); | |
c2e525d9 JH |
1140 | e->suspect = porigin; |
1141 | } | |
7e6ac6e4 | 1142 | queue_blames(sb, porigin, suspects); |
c2e525d9 JH |
1143 | } |
1144 | ||
69264f46 JH |
1145 | /* |
1146 | * We pass blame from the current commit to its parents. We keep saying | |
1147 | * "parent" (and "porigin"), but what we mean is to find scapegoat to | |
1148 | * exonerate ourselves. | |
1149 | */ | |
f81d70e9 JS |
1150 | static struct commit_list *first_scapegoat(struct rev_info *revs, struct commit *commit, |
1151 | int reverse) | |
69264f46 | 1152 | { |
95a4fb0e JK |
1153 | if (!reverse) { |
1154 | if (revs->first_parent_only && | |
1155 | commit->parents && | |
1156 | commit->parents->next) { | |
1157 | free_commit_list(commit->parents->next); | |
1158 | commit->parents->next = NULL; | |
1159 | } | |
85af7929 | 1160 | return commit->parents; |
95a4fb0e | 1161 | } |
85af7929 | 1162 | return lookup_decoration(&revs->children, &commit->object); |
69264f46 JH |
1163 | } |
1164 | ||
f81d70e9 | 1165 | static int num_scapegoats(struct rev_info *revs, struct commit *commit, int reverse) |
69264f46 | 1166 | { |
f81d70e9 | 1167 | struct commit_list *l = first_scapegoat(revs, commit, reverse); |
4bbaa1eb | 1168 | return commit_list_count(l); |
69264f46 JH |
1169 | } |
1170 | ||
7e6ac6e4 DK |
1171 | /* Distribute collected unsorted blames to the respected sorted lists |
1172 | * in the various origins. | |
1173 | */ | |
9807b3d6 | 1174 | static void distribute_blame(struct blame_scoreboard *sb, struct blame_entry *blamed) |
7e6ac6e4 | 1175 | { |
78b06e66 JS |
1176 | blamed = llist_mergesort(blamed, get_next_blame, set_next_blame, |
1177 | compare_blame_suspect); | |
7e6ac6e4 DK |
1178 | while (blamed) |
1179 | { | |
f84afb9c | 1180 | struct blame_origin *porigin = blamed->suspect; |
7e6ac6e4 DK |
1181 | struct blame_entry *suspects = NULL; |
1182 | do { | |
1183 | struct blame_entry *next = blamed->next; | |
1184 | blamed->next = suspects; | |
1185 | suspects = blamed; | |
1186 | blamed = next; | |
1187 | } while (blamed && blamed->suspect == porigin); | |
1188 | suspects = reverse_blame(suspects, NULL); | |
1189 | queue_blames(sb, porigin, suspects); | |
1190 | } | |
1191 | } | |
1192 | ||
69264f46 | 1193 | #define MAXSG 16 |
cee7f245 | 1194 | |
9807b3d6 | 1195 | static void pass_blame(struct blame_scoreboard *sb, struct blame_origin *origin, int opt) |
cee7f245 | 1196 | { |
85af7929 | 1197 | struct rev_info *revs = sb->revs; |
69264f46 | 1198 | int i, pass, num_sg; |
cee7f245 | 1199 | struct commit *commit = origin->commit; |
69264f46 | 1200 | struct commit_list *sg; |
f84afb9c JS |
1201 | struct blame_origin *sg_buf[MAXSG]; |
1202 | struct blame_origin *porigin, **sg_origin = sg_buf; | |
7e6ac6e4 DK |
1203 | struct blame_entry *toosmall = NULL; |
1204 | struct blame_entry *blames, **blametail = &blames; | |
69264f46 | 1205 | |
f81d70e9 | 1206 | num_sg = num_scapegoats(revs, commit, sb->reverse); |
69264f46 JH |
1207 | if (!num_sg) |
1208 | goto finish; | |
1209 | else if (num_sg < ARRAY_SIZE(sg_buf)) | |
1210 | memset(sg_buf, 0, sizeof(sg_buf)); | |
1211 | else | |
1212 | sg_origin = xcalloc(num_sg, sizeof(*sg_origin)); | |
cee7f245 | 1213 | |
69264f46 JH |
1214 | /* |
1215 | * The first pass looks for unrenamed path to optimize for | |
f69e743d JH |
1216 | * common cases, then we look for renames in the second pass. |
1217 | */ | |
1f44129b | 1218 | for (pass = 0; pass < 2 - sb->no_whole_file_rename; pass++) { |
f84afb9c | 1219 | struct blame_origin *(*find)(struct commit *, struct blame_origin *); |
f69e743d JH |
1220 | find = pass ? find_rename : find_origin; |
1221 | ||
f81d70e9 | 1222 | for (i = 0, sg = first_scapegoat(revs, commit, sb->reverse); |
69264f46 JH |
1223 | i < num_sg && sg; |
1224 | sg = sg->next, i++) { | |
1225 | struct commit *p = sg->item; | |
0421d9f8 | 1226 | int j, same; |
f69e743d | 1227 | |
69264f46 | 1228 | if (sg_origin[i]) |
f69e743d JH |
1229 | continue; |
1230 | if (parse_commit(p)) | |
1231 | continue; | |
8265921c | 1232 | porigin = find(p, origin); |
f69e743d JH |
1233 | if (!porigin) |
1234 | continue; | |
a7bcfa12 | 1235 | if (!oidcmp(&porigin->blob_oid, &origin->blob_oid)) { |
c2e525d9 | 1236 | pass_whole_blame(sb, origin, porigin); |
006a0744 | 1237 | blame_origin_decref(porigin); |
f69e743d JH |
1238 | goto finish; |
1239 | } | |
0421d9f8 | 1240 | for (j = same = 0; j < i; j++) |
69264f46 | 1241 | if (sg_origin[j] && |
a7bcfa12 | 1242 | !oidcmp(&sg_origin[j]->blob_oid, &porigin->blob_oid)) { |
0421d9f8 JH |
1243 | same = 1; |
1244 | break; | |
1245 | } | |
1246 | if (!same) | |
69264f46 | 1247 | sg_origin[i] = porigin; |
0421d9f8 | 1248 | else |
006a0744 | 1249 | blame_origin_decref(porigin); |
cee7f245 | 1250 | } |
cee7f245 JH |
1251 | } |
1252 | ||
8449528d | 1253 | sb->num_commits++; |
f81d70e9 | 1254 | for (i = 0, sg = first_scapegoat(revs, commit, sb->reverse); |
69264f46 JH |
1255 | i < num_sg && sg; |
1256 | sg = sg->next, i++) { | |
f84afb9c | 1257 | struct blame_origin *porigin = sg_origin[i]; |
cee7f245 JH |
1258 | if (!porigin) |
1259 | continue; | |
96e11709 | 1260 | if (!origin->previous) { |
006a0744 | 1261 | blame_origin_incref(porigin); |
96e11709 JH |
1262 | origin->previous = porigin; |
1263 | } | |
7e6ac6e4 DK |
1264 | pass_blame_to_parent(sb, origin, porigin); |
1265 | if (!origin->suspects) | |
54a4c617 | 1266 | goto finish; |
cee7f245 | 1267 | } |
d24bba80 JH |
1268 | |
1269 | /* | |
1732a1fd | 1270 | * Optionally find moves in parents' files. |
d24bba80 | 1271 | */ |
7e6ac6e4 | 1272 | if (opt & PICKAXE_BLAME_MOVE) { |
18ec0d62 | 1273 | filter_small(sb, &toosmall, &origin->suspects, sb->move_score); |
7e6ac6e4 | 1274 | if (origin->suspects) { |
f81d70e9 | 1275 | for (i = 0, sg = first_scapegoat(revs, commit, sb->reverse); |
7e6ac6e4 DK |
1276 | i < num_sg && sg; |
1277 | sg = sg->next, i++) { | |
f84afb9c | 1278 | struct blame_origin *porigin = sg_origin[i]; |
7e6ac6e4 DK |
1279 | if (!porigin) |
1280 | continue; | |
1281 | find_move_in_parent(sb, &blametail, &toosmall, origin, porigin); | |
1282 | if (!origin->suspects) | |
1283 | break; | |
1284 | } | |
d24bba80 | 1285 | } |
7e6ac6e4 | 1286 | } |
d24bba80 | 1287 | |
18abd745 | 1288 | /* |
1732a1fd | 1289 | * Optionally find copies from parents' files. |
18abd745 | 1290 | */ |
7e6ac6e4 | 1291 | if (opt & PICKAXE_BLAME_COPY) { |
18ec0d62 JS |
1292 | if (sb->copy_score > sb->move_score) |
1293 | filter_small(sb, &toosmall, &origin->suspects, sb->copy_score); | |
1294 | else if (sb->copy_score < sb->move_score) { | |
7e6ac6e4 DK |
1295 | origin->suspects = blame_merge(origin->suspects, toosmall); |
1296 | toosmall = NULL; | |
18ec0d62 | 1297 | filter_small(sb, &toosmall, &origin->suspects, sb->copy_score); |
7e6ac6e4 DK |
1298 | } |
1299 | if (!origin->suspects) | |
1300 | goto finish; | |
1301 | ||
f81d70e9 | 1302 | for (i = 0, sg = first_scapegoat(revs, commit, sb->reverse); |
69264f46 JH |
1303 | i < num_sg && sg; |
1304 | sg = sg->next, i++) { | |
f84afb9c | 1305 | struct blame_origin *porigin = sg_origin[i]; |
7e6ac6e4 DK |
1306 | find_copy_in_parent(sb, &blametail, &toosmall, |
1307 | origin, sg->item, porigin, opt); | |
1308 | if (!origin->suspects) | |
54a4c617 | 1309 | goto finish; |
18abd745 | 1310 | } |
7e6ac6e4 | 1311 | } |
54a4c617 | 1312 | |
7e6ac6e4 DK |
1313 | finish: |
1314 | *blametail = NULL; | |
1315 | distribute_blame(sb, blames); | |
1316 | /* | |
1317 | * prepend toosmall to origin->suspects | |
1318 | * | |
1319 | * There is no point in sorting: this ends up on a big | |
1320 | * unsorted list in the caller anyway. | |
1321 | */ | |
1322 | if (toosmall) { | |
1323 | struct blame_entry **tail = &toosmall; | |
1324 | while (*tail) | |
1325 | tail = &(*tail)->next; | |
1326 | *tail = origin->suspects; | |
1327 | origin->suspects = toosmall; | |
1328 | } | |
69264f46 JH |
1329 | for (i = 0; i < num_sg; i++) { |
1330 | if (sg_origin[i]) { | |
1331 | drop_origin_blob(sg_origin[i]); | |
006a0744 | 1332 | blame_origin_decref(sg_origin[i]); |
7c3c7962 JH |
1333 | } |
1334 | } | |
1335 | drop_origin_blob(origin); | |
69264f46 JH |
1336 | if (sg_buf != sg_origin) |
1337 | free(sg_origin); | |
cee7f245 JH |
1338 | } |
1339 | ||
1732a1fd JH |
1340 | /* |
1341 | * Information on commits, used for output. | |
1342 | */ | |
9cba13ca | 1343 | struct commit_info { |
ea02ffa3 AP |
1344 | struct strbuf author; |
1345 | struct strbuf author_mail; | |
dddbad72 | 1346 | timestamp_t author_time; |
ea02ffa3 | 1347 | struct strbuf author_tz; |
cee7f245 JH |
1348 | |
1349 | /* filled only when asked for details */ | |
ea02ffa3 AP |
1350 | struct strbuf committer; |
1351 | struct strbuf committer_mail; | |
dddbad72 | 1352 | timestamp_t committer_time; |
ea02ffa3 | 1353 | struct strbuf committer_tz; |
cee7f245 | 1354 | |
ea02ffa3 | 1355 | struct strbuf summary; |
cee7f245 JH |
1356 | }; |
1357 | ||
1732a1fd JH |
1358 | /* |
1359 | * Parse author/committer line in the commit object buffer | |
1360 | */ | |
cee7f245 | 1361 | static void get_ac_line(const char *inbuf, const char *what, |
ea02ffa3 | 1362 | struct strbuf *name, struct strbuf *mail, |
dddbad72 | 1363 | timestamp_t *time, struct strbuf *tz) |
cee7f245 | 1364 | { |
3c020bd5 | 1365 | struct ident_split ident; |
ea02ffa3 AP |
1366 | size_t len, maillen, namelen; |
1367 | char *tmp, *endp; | |
1368 | const char *namebuf, *mailbuf; | |
cee7f245 JH |
1369 | |
1370 | tmp = strstr(inbuf, what); | |
1371 | if (!tmp) | |
1372 | goto error_out; | |
1373 | tmp += strlen(what); | |
1374 | endp = strchr(tmp, '\n'); | |
1375 | if (!endp) | |
1376 | len = strlen(tmp); | |
1377 | else | |
1378 | len = endp - tmp; | |
3c020bd5 AP |
1379 | |
1380 | if (split_ident_line(&ident, tmp, len)) { | |
cee7f245 JH |
1381 | error_out: |
1382 | /* Ugh */ | |
ea02ffa3 AP |
1383 | tmp = "(unknown)"; |
1384 | strbuf_addstr(name, tmp); | |
1385 | strbuf_addstr(mail, tmp); | |
1386 | strbuf_addstr(tz, tmp); | |
cee7f245 JH |
1387 | *time = 0; |
1388 | return; | |
1389 | } | |
cee7f245 | 1390 | |
3c020bd5 | 1391 | namelen = ident.name_end - ident.name_begin; |
ea02ffa3 | 1392 | namebuf = ident.name_begin; |
cee7f245 | 1393 | |
ea02ffa3 AP |
1394 | maillen = ident.mail_end - ident.mail_begin; |
1395 | mailbuf = ident.mail_begin; | |
f9567384 | 1396 | |
de5abe9f RS |
1397 | if (ident.date_begin && ident.date_end) |
1398 | *time = strtoul(ident.date_begin, NULL, 10); | |
1399 | else | |
1400 | *time = 0; | |
f9567384 | 1401 | |
de5abe9f RS |
1402 | if (ident.tz_begin && ident.tz_end) |
1403 | strbuf_add(tz, ident.tz_begin, ident.tz_end - ident.tz_begin); | |
1404 | else | |
1405 | strbuf_addstr(tz, "(unknown)"); | |
f9567384 | 1406 | |
f9567384 | 1407 | /* |
d20d654f | 1408 | * Now, convert both name and e-mail using mailmap |
f9567384 | 1409 | */ |
ea02ffa3 AP |
1410 | map_user(&mailmap, &mailbuf, &maillen, |
1411 | &namebuf, &namelen); | |
1412 | ||
1413 | strbuf_addf(mail, "<%.*s>", (int)maillen, mailbuf); | |
1414 | strbuf_add(name, namebuf, namelen); | |
1415 | } | |
1416 | ||
1417 | static void commit_info_init(struct commit_info *ci) | |
1418 | { | |
1419 | ||
1420 | strbuf_init(&ci->author, 0); | |
1421 | strbuf_init(&ci->author_mail, 0); | |
1422 | strbuf_init(&ci->author_tz, 0); | |
1423 | strbuf_init(&ci->committer, 0); | |
1424 | strbuf_init(&ci->committer_mail, 0); | |
1425 | strbuf_init(&ci->committer_tz, 0); | |
1426 | strbuf_init(&ci->summary, 0); | |
1427 | } | |
1428 | ||
1429 | static void commit_info_destroy(struct commit_info *ci) | |
1430 | { | |
1431 | ||
1432 | strbuf_release(&ci->author); | |
1433 | strbuf_release(&ci->author_mail); | |
1434 | strbuf_release(&ci->author_tz); | |
1435 | strbuf_release(&ci->committer); | |
1436 | strbuf_release(&ci->committer_mail); | |
1437 | strbuf_release(&ci->committer_tz); | |
1438 | strbuf_release(&ci->summary); | |
cee7f245 JH |
1439 | } |
1440 | ||
1441 | static void get_commit_info(struct commit *commit, | |
1442 | struct commit_info *ret, | |
1443 | int detailed) | |
1444 | { | |
1445 | int len; | |
e297cf5a | 1446 | const char *subject, *encoding; |
b000c59b | 1447 | const char *message; |
ea02ffa3 AP |
1448 | |
1449 | commit_info_init(ret); | |
cee7f245 | 1450 | |
e297cf5a | 1451 | encoding = get_log_output_encoding(); |
5a10d236 | 1452 | message = logmsg_reencode(commit, NULL, encoding); |
69cd8f63 | 1453 | get_ac_line(message, "\nauthor ", |
ea02ffa3 | 1454 | &ret->author, &ret->author_mail, |
cee7f245 JH |
1455 | &ret->author_time, &ret->author_tz); |
1456 | ||
69cd8f63 | 1457 | if (!detailed) { |
b66103c3 | 1458 | unuse_commit_buffer(commit, message); |
cee7f245 | 1459 | return; |
69cd8f63 | 1460 | } |
cee7f245 | 1461 | |
69cd8f63 | 1462 | get_ac_line(message, "\ncommitter ", |
ea02ffa3 | 1463 | &ret->committer, &ret->committer_mail, |
cee7f245 JH |
1464 | &ret->committer_time, &ret->committer_tz); |
1465 | ||
ad98a58b | 1466 | len = find_commit_subject(message, &subject); |
ea02ffa3 AP |
1467 | if (len) |
1468 | strbuf_add(&ret->summary, subject, len); | |
1469 | else | |
f2fd0760 | 1470 | strbuf_addf(&ret->summary, "(%s)", oid_to_hex(&commit->object.oid)); |
ea02ffa3 | 1471 | |
b66103c3 | 1472 | unuse_commit_buffer(commit, message); |
cee7f245 JH |
1473 | } |
1474 | ||
1732a1fd | 1475 | /* |
4e768329 JK |
1476 | * Write out any suspect information which depends on the path. This must be |
1477 | * handled separately from emit_one_suspect_detail(), because a given commit | |
1478 | * may have changes in multiple paths. So this needs to appear each time | |
1479 | * we mention a new group. | |
1480 | * | |
1732a1fd JH |
1481 | * To allow LF and other nonportable characters in pathnames, |
1482 | * they are c-style quoted as needed. | |
1483 | */ | |
f84afb9c | 1484 | static void write_filename_info(struct blame_origin *suspect) |
46e5e69d | 1485 | { |
4e768329 | 1486 | if (suspect->previous) { |
f84afb9c | 1487 | struct blame_origin *prev = suspect->previous; |
4e768329 JK |
1488 | printf("previous %s ", oid_to_hex(&prev->commit->object.oid)); |
1489 | write_name_quoted(prev->path, stdout, '\n'); | |
1490 | } | |
46e5e69d | 1491 | printf("filename "); |
4e768329 | 1492 | write_name_quoted(suspect->path, stdout, '\n'); |
46e5e69d JH |
1493 | } |
1494 | ||
9991030c JH |
1495 | /* |
1496 | * Porcelain/Incremental format wants to show a lot of details per | |
1497 | * commit. Instead of repeating this every line, emit it only once, | |
e86226e3 JK |
1498 | * the first time each commit appears in the output (unless the |
1499 | * user has specifically asked for us to repeat). | |
9991030c | 1500 | */ |
f84afb9c | 1501 | static int emit_one_suspect_detail(struct blame_origin *suspect, int repeat) |
9991030c JH |
1502 | { |
1503 | struct commit_info ci; | |
1504 | ||
e86226e3 | 1505 | if (!repeat && (suspect->commit->object.flags & METAINFO_SHOWN)) |
9991030c JH |
1506 | return 0; |
1507 | ||
1508 | suspect->commit->object.flags |= METAINFO_SHOWN; | |
1509 | get_commit_info(suspect->commit, &ci, 1); | |
ea02ffa3 AP |
1510 | printf("author %s\n", ci.author.buf); |
1511 | printf("author-mail %s\n", ci.author_mail.buf); | |
cb71f8bd | 1512 | printf("author-time %"PRItime"\n", ci.author_time); |
ea02ffa3 AP |
1513 | printf("author-tz %s\n", ci.author_tz.buf); |
1514 | printf("committer %s\n", ci.committer.buf); | |
1515 | printf("committer-mail %s\n", ci.committer_mail.buf); | |
cb71f8bd | 1516 | printf("committer-time %"PRItime"\n", ci.committer_time); |
ea02ffa3 AP |
1517 | printf("committer-tz %s\n", ci.committer_tz.buf); |
1518 | printf("summary %s\n", ci.summary.buf); | |
9991030c JH |
1519 | if (suspect->commit->object.flags & UNINTERESTING) |
1520 | printf("boundary\n"); | |
ea02ffa3 AP |
1521 | |
1522 | commit_info_destroy(&ci); | |
1523 | ||
9991030c JH |
1524 | return 1; |
1525 | } | |
1526 | ||
1732a1fd | 1527 | /* |
7e6ac6e4 DK |
1528 | * The blame_entry is found to be guilty for the range. |
1529 | * Show it in incremental output. | |
1732a1fd | 1530 | */ |
8c59921d | 1531 | static void found_guilty_entry(struct blame_entry *ent, void *data) |
717d1462 | 1532 | { |
8c59921d JS |
1533 | struct progress_info *pi = (struct progress_info *)data; |
1534 | ||
717d1462 | 1535 | if (incremental) { |
f84afb9c | 1536 | struct blame_origin *suspect = ent->suspect; |
717d1462 LT |
1537 | |
1538 | printf("%s %d %d %d\n", | |
f2fd0760 | 1539 | oid_to_hex(&suspect->commit->object.oid), |
717d1462 | 1540 | ent->s_lno + 1, ent->lno + 1, ent->num_lines); |
e86226e3 | 1541 | emit_one_suspect_detail(suspect, 0); |
4e768329 | 1542 | write_filename_info(suspect); |
06f59e9f | 1543 | maybe_flush_or_die(stdout, "stdout"); |
717d1462 | 1544 | } |
aba37f49 ECA |
1545 | pi->blamed_lines += ent->num_lines; |
1546 | display_progress(pi->progress, pi->blamed_lines); | |
717d1462 LT |
1547 | } |
1548 | ||
1732a1fd | 1549 | /* |
7e6ac6e4 DK |
1550 | * The main loop -- while we have blobs with lines whose true origin |
1551 | * is still unknown, pick one blob, and allow its lines to pass blames | |
1552 | * to its parents. */ | |
9807b3d6 | 1553 | static void assign_blame(struct blame_scoreboard *sb, int opt) |
717d1462 | 1554 | { |
85af7929 | 1555 | struct rev_info *revs = sb->revs; |
7e6ac6e4 | 1556 | struct commit *commit = prio_queue_get(&sb->commits); |
85af7929 | 1557 | |
7e6ac6e4 | 1558 | while (commit) { |
717d1462 | 1559 | struct blame_entry *ent; |
f84afb9c | 1560 | struct blame_origin *suspect = commit->util; |
717d1462 LT |
1561 | |
1562 | /* find one suspect to break down */ | |
7e6ac6e4 DK |
1563 | while (suspect && !suspect->suspects) |
1564 | suspect = suspect->next; | |
1565 | ||
1566 | if (!suspect) { | |
1567 | commit = prio_queue_get(&sb->commits); | |
1568 | continue; | |
1569 | } | |
1570 | ||
1571 | assert(commit == suspect->commit); | |
717d1462 | 1572 | |
1732a1fd JH |
1573 | /* |
1574 | * We will use this suspect later in the loop, | |
1575 | * so hold onto it in the meantime. | |
1576 | */ | |
006a0744 | 1577 | blame_origin_incref(suspect); |
0064053b | 1578 | parse_commit(commit); |
f81d70e9 | 1579 | if (sb->reverse || |
85af7929 JH |
1580 | (!(commit->object.flags & UNINTERESTING) && |
1581 | !(revs->max_age != -1 && commit->date < revs->max_age))) | |
717d1462 LT |
1582 | pass_blame(sb, suspect, opt); |
1583 | else { | |
1584 | commit->object.flags |= UNINTERESTING; | |
1585 | if (commit->object.parsed) | |
1586 | mark_parents_uninteresting(commit); | |
1587 | } | |
1588 | /* treat root commit as boundary */ | |
2cf83374 | 1589 | if (!commit->parents && !sb->show_root) |
717d1462 LT |
1590 | commit->object.flags |= UNINTERESTING; |
1591 | ||
1592 | /* Take responsibility for the remaining entries */ | |
7e6ac6e4 DK |
1593 | ent = suspect->suspects; |
1594 | if (ent) { | |
1595 | suspect->guilty = 1; | |
1596 | for (;;) { | |
1597 | struct blame_entry *next = ent->next; | |
8c59921d JS |
1598 | if (sb->found_guilty_entry) |
1599 | sb->found_guilty_entry(ent, sb->found_guilty_entry_data); | |
7e6ac6e4 DK |
1600 | if (next) { |
1601 | ent = next; | |
1602 | continue; | |
1603 | } | |
1604 | ent->next = sb->ent; | |
1605 | sb->ent = suspect->suspects; | |
1606 | suspect->suspects = NULL; | |
1607 | break; | |
1608 | } | |
1609 | } | |
006a0744 | 1610 | blame_origin_decref(suspect); |
717d1462 | 1611 | |
4149c186 | 1612 | if (sb->debug) /* sanity */ |
717d1462 LT |
1613 | sanity_check_refcnt(sb); |
1614 | } | |
1615 | } | |
1616 | ||
dddbad72 | 1617 | static const char *format_time(timestamp_t time, const char *tz_str, |
717d1462 LT |
1618 | int show_raw_time) |
1619 | { | |
bccce0f8 | 1620 | static struct strbuf time_buf = STRBUF_INIT; |
717d1462 | 1621 | |
bccce0f8 | 1622 | strbuf_reset(&time_buf); |
717d1462 | 1623 | if (show_raw_time) { |
cb71f8bd | 1624 | strbuf_addf(&time_buf, "%"PRItime" %s", time, tz_str); |
717d1462 | 1625 | } |
31653c1a | 1626 | else { |
ac39b277 | 1627 | const char *time_str; |
bccce0f8 | 1628 | size_t time_width; |
ac39b277 | 1629 | int tz; |
31653c1a | 1630 | tz = atoi(tz_str); |
a5481a6c | 1631 | time_str = show_date(time, tz, &blame_date_mode); |
bccce0f8 JX |
1632 | strbuf_addstr(&time_buf, time_str); |
1633 | /* | |
1634 | * Add space paddings to time_buf to display a fixed width | |
1635 | * string, and use time_width for display width calibration. | |
1636 | */ | |
1637 | for (time_width = utf8_strwidth(time_str); | |
1638 | time_width < blame_date_width; | |
1639 | time_width++) | |
1640 | strbuf_addch(&time_buf, ' '); | |
31653c1a | 1641 | } |
bccce0f8 | 1642 | return time_buf.buf; |
717d1462 LT |
1643 | } |
1644 | ||
cee7f245 JH |
1645 | #define OUTPUT_ANNOTATE_COMPAT 001 |
1646 | #define OUTPUT_LONG_OBJECT_NAME 002 | |
1647 | #define OUTPUT_RAW_TIMESTAMP 004 | |
1648 | #define OUTPUT_PORCELAIN 010 | |
1649 | #define OUTPUT_SHOW_NAME 020 | |
1650 | #define OUTPUT_SHOW_NUMBER 040 | |
5ff62c30 | 1651 | #define OUTPUT_SHOW_SCORE 0100 |
093dc5be | 1652 | #define OUTPUT_NO_AUTHOR 0200 |
1b8cdce9 | 1653 | #define OUTPUT_SHOW_EMAIL 0400 |
ed747dd5 | 1654 | #define OUTPUT_LINE_PORCELAIN 01000 |
cee7f245 | 1655 | |
f84afb9c | 1656 | static void emit_porcelain_details(struct blame_origin *suspect, int repeat) |
e86226e3 JK |
1657 | { |
1658 | if (emit_one_suspect_detail(suspect, repeat) || | |
1659 | (suspect->commit->object.flags & MORE_THAN_ONE_PATH)) | |
4e768329 | 1660 | write_filename_info(suspect); |
e86226e3 JK |
1661 | } |
1662 | ||
9807b3d6 | 1663 | static void emit_porcelain(struct blame_scoreboard *sb, struct blame_entry *ent, |
e86226e3 | 1664 | int opt) |
cee7f245 | 1665 | { |
ed747dd5 | 1666 | int repeat = opt & OUTPUT_LINE_PORCELAIN; |
cee7f245 JH |
1667 | int cnt; |
1668 | const char *cp; | |
f84afb9c | 1669 | struct blame_origin *suspect = ent->suspect; |
dc01505f | 1670 | char hex[GIT_MAX_HEXSZ + 1]; |
cee7f245 | 1671 | |
2490574d | 1672 | oid_to_hex_r(hex, &suspect->commit->object.oid); |
7e6ac6e4 | 1673 | printf("%s %d %d %d\n", |
cee7f245 | 1674 | hex, |
cee7f245 JH |
1675 | ent->s_lno + 1, |
1676 | ent->lno + 1, | |
1677 | ent->num_lines); | |
ed747dd5 | 1678 | emit_porcelain_details(suspect, repeat); |
cee7f245 | 1679 | |
935202bd | 1680 | cp = blame_nth_line(sb, ent->lno); |
cee7f245 JH |
1681 | for (cnt = 0; cnt < ent->num_lines; cnt++) { |
1682 | char ch; | |
ed747dd5 | 1683 | if (cnt) { |
cee7f245 JH |
1684 | printf("%s %d %d\n", hex, |
1685 | ent->s_lno + 1 + cnt, | |
1686 | ent->lno + 1 + cnt); | |
ed747dd5 JK |
1687 | if (repeat) |
1688 | emit_porcelain_details(suspect, 1); | |
1689 | } | |
cee7f245 JH |
1690 | putchar('\t'); |
1691 | do { | |
1692 | ch = *cp++; | |
1693 | putchar(ch); | |
1694 | } while (ch != '\n' && | |
1695 | cp < sb->final_buf + sb->final_buf_size); | |
1696 | } | |
a5ca8367 JS |
1697 | |
1698 | if (sb->final_buf_size && cp[-1] != '\n') | |
1699 | putchar('\n'); | |
cee7f245 JH |
1700 | } |
1701 | ||
9807b3d6 | 1702 | static void emit_other(struct blame_scoreboard *sb, struct blame_entry *ent, int opt) |
cee7f245 JH |
1703 | { |
1704 | int cnt; | |
1705 | const char *cp; | |
f84afb9c | 1706 | struct blame_origin *suspect = ent->suspect; |
cee7f245 | 1707 | struct commit_info ci; |
dc01505f | 1708 | char hex[GIT_MAX_HEXSZ + 1]; |
cee7f245 JH |
1709 | int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP); |
1710 | ||
1711 | get_commit_info(suspect->commit, &ci, 1); | |
2490574d | 1712 | oid_to_hex_r(hex, &suspect->commit->object.oid); |
cee7f245 | 1713 | |
935202bd | 1714 | cp = blame_nth_line(sb, ent->lno); |
cee7f245 JH |
1715 | for (cnt = 0; cnt < ent->num_lines; cnt++) { |
1716 | char ch; | |
110d26fc | 1717 | int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? GIT_SHA1_HEXSZ : abbrev; |
b11121d9 JH |
1718 | |
1719 | if (suspect->commit->object.flags & UNINTERESTING) { | |
e68989a7 JH |
1720 | if (blank_boundary) |
1721 | memset(hex, ' ', length); | |
7ceacdff | 1722 | else if (!(opt & OUTPUT_ANNOTATE_COMPAT)) { |
4c10a5ca JH |
1723 | length--; |
1724 | putchar('^'); | |
1725 | } | |
b11121d9 | 1726 | } |
cee7f245 | 1727 | |
b11121d9 | 1728 | printf("%.*s", length, hex); |
1b8cdce9 KB |
1729 | if (opt & OUTPUT_ANNOTATE_COMPAT) { |
1730 | const char *name; | |
1731 | if (opt & OUTPUT_SHOW_EMAIL) | |
ea02ffa3 | 1732 | name = ci.author_mail.buf; |
1b8cdce9 | 1733 | else |
ea02ffa3 | 1734 | name = ci.author.buf; |
1b8cdce9 | 1735 | printf("\t(%10s\t%10s\t%d)", name, |
ea02ffa3 | 1736 | format_time(ci.author_time, ci.author_tz.buf, |
cee7f245 JH |
1737 | show_raw_time), |
1738 | ent->lno + 1 + cnt); | |
1b8cdce9 | 1739 | } else { |
5ff62c30 | 1740 | if (opt & OUTPUT_SHOW_SCORE) |
54a4c617 JH |
1741 | printf(" %*d %02d", |
1742 | max_score_digits, ent->score, | |
1743 | ent->suspect->refcnt); | |
cee7f245 JH |
1744 | if (opt & OUTPUT_SHOW_NAME) |
1745 | printf(" %-*.*s", longest_file, longest_file, | |
1746 | suspect->path); | |
1747 | if (opt & OUTPUT_SHOW_NUMBER) | |
1748 | printf(" %*d", max_orig_digits, | |
1749 | ent->s_lno + 1 + cnt); | |
093dc5be | 1750 | |
ffaf9cc0 | 1751 | if (!(opt & OUTPUT_NO_AUTHOR)) { |
1b8cdce9 KB |
1752 | const char *name; |
1753 | int pad; | |
1754 | if (opt & OUTPUT_SHOW_EMAIL) | |
ea02ffa3 | 1755 | name = ci.author_mail.buf; |
1b8cdce9 | 1756 | else |
ea02ffa3 | 1757 | name = ci.author.buf; |
1b8cdce9 | 1758 | pad = longest_author - utf8_strwidth(name); |
ffaf9cc0 | 1759 | printf(" (%s%*s %10s", |
1b8cdce9 | 1760 | name, pad, "", |
093dc5be | 1761 | format_time(ci.author_time, |
ea02ffa3 | 1762 | ci.author_tz.buf, |
093dc5be | 1763 | show_raw_time)); |
ffaf9cc0 | 1764 | } |
093dc5be | 1765 | printf(" %*d) ", |
cee7f245 JH |
1766 | max_digits, ent->lno + 1 + cnt); |
1767 | } | |
1768 | do { | |
1769 | ch = *cp++; | |
1770 | putchar(ch); | |
1771 | } while (ch != '\n' && | |
1772 | cp < sb->final_buf + sb->final_buf_size); | |
1773 | } | |
a5ca8367 JS |
1774 | |
1775 | if (sb->final_buf_size && cp[-1] != '\n') | |
1776 | putchar('\n'); | |
ea02ffa3 AP |
1777 | |
1778 | commit_info_destroy(&ci); | |
cee7f245 JH |
1779 | } |
1780 | ||
9807b3d6 | 1781 | static void output(struct blame_scoreboard *sb, int option) |
cee7f245 JH |
1782 | { |
1783 | struct blame_entry *ent; | |
1784 | ||
1785 | if (option & OUTPUT_PORCELAIN) { | |
1786 | for (ent = sb->ent; ent; ent = ent->next) { | |
7e6ac6e4 | 1787 | int count = 0; |
f84afb9c | 1788 | struct blame_origin *suspect; |
7e6ac6e4 | 1789 | struct commit *commit = ent->suspect->commit; |
cee7f245 JH |
1790 | if (commit->object.flags & MORE_THAN_ONE_PATH) |
1791 | continue; | |
7e6ac6e4 DK |
1792 | for (suspect = commit->util; suspect; suspect = suspect->next) { |
1793 | if (suspect->guilty && count++) { | |
1794 | commit->object.flags |= MORE_THAN_ONE_PATH; | |
1795 | break; | |
1796 | } | |
cee7f245 JH |
1797 | } |
1798 | } | |
1799 | } | |
1800 | ||
1801 | for (ent = sb->ent; ent; ent = ent->next) { | |
1802 | if (option & OUTPUT_PORCELAIN) | |
e86226e3 | 1803 | emit_porcelain(sb, ent, option); |
5ff62c30 | 1804 | else { |
cee7f245 | 1805 | emit_other(sb, ent, option); |
5ff62c30 | 1806 | } |
cee7f245 JH |
1807 | } |
1808 | } | |
1809 | ||
29aa0b20 RS |
1810 | static const char *get_next_line(const char *start, const char *end) |
1811 | { | |
1812 | const char *nl = memchr(start, '\n', end - start); | |
60d85e11 | 1813 | return nl ? nl + 1 : end; |
29aa0b20 RS |
1814 | } |
1815 | ||
1732a1fd JH |
1816 | /* |
1817 | * To allow quick access to the contents of nth line in the | |
1818 | * final image, prepare an index in the scoreboard. | |
1819 | */ | |
9807b3d6 | 1820 | static int prepare_lines(struct blame_scoreboard *sb) |
cee7f245 JH |
1821 | { |
1822 | const char *buf = sb->final_buf; | |
1823 | unsigned long len = sb->final_buf_size; | |
352bbbd9 DK |
1824 | const char *end = buf + len; |
1825 | const char *p; | |
1826 | int *lineno; | |
60d85e11 | 1827 | int num = 0; |
352bbbd9 | 1828 | |
60d85e11 | 1829 | for (p = buf; p < end; p = get_next_line(p, end)) |
29aa0b20 | 1830 | num++; |
352bbbd9 | 1831 | |
b32fa95f JK |
1832 | ALLOC_ARRAY(sb->lineno, num + 1); |
1833 | lineno = sb->lineno; | |
352bbbd9 | 1834 | |
60d85e11 | 1835 | for (p = buf; p < end; p = get_next_line(p, end)) |
29aa0b20 | 1836 | *lineno++ = p - buf; |
352bbbd9 | 1837 | |
60d85e11 | 1838 | *lineno = len; |
352bbbd9 | 1839 | |
60d85e11 | 1840 | sb->num_lines = num; |
cee7f245 JH |
1841 | return sb->num_lines; |
1842 | } | |
1843 | ||
1732a1fd JH |
1844 | /* |
1845 | * Add phony grafts for use with -S; this is primarily to | |
34baebce | 1846 | * support git's cvsserver that wants to give a linear history |
1732a1fd JH |
1847 | * to its clients. |
1848 | */ | |
cee7f245 JH |
1849 | static int read_ancestry(const char *graft_file) |
1850 | { | |
1851 | FILE *fp = fopen(graft_file, "r"); | |
e228c173 | 1852 | struct strbuf buf = STRBUF_INIT; |
cee7f245 JH |
1853 | if (!fp) |
1854 | return -1; | |
e228c173 | 1855 | while (!strbuf_getwholeline(&buf, fp, '\n')) { |
cee7f245 | 1856 | /* The format is just "Commit Parent1 Parent2 ...\n" */ |
e228c173 | 1857 | struct commit_graft *graft = read_graft_line(buf.buf, buf.len); |
8eaf7986 JH |
1858 | if (graft) |
1859 | register_commit_graft(graft, 0); | |
cee7f245 JH |
1860 | } |
1861 | fclose(fp); | |
e228c173 | 1862 | strbuf_release(&buf); |
cee7f245 JH |
1863 | return 0; |
1864 | } | |
1865 | ||
f84afb9c | 1866 | static int update_auto_abbrev(int auto_abbrev, struct blame_origin *suspect) |
b31272f7 | 1867 | { |
ed1c9977 | 1868 | const char *uniq = find_unique_abbrev(suspect->commit->object.oid.hash, |
b31272f7 JH |
1869 | auto_abbrev); |
1870 | int len = strlen(uniq); | |
1871 | if (auto_abbrev < len) | |
1872 | return len; | |
1873 | return auto_abbrev; | |
1874 | } | |
1875 | ||
1732a1fd JH |
1876 | /* |
1877 | * How many columns do we need to show line numbers, authors, | |
1878 | * and filenames? | |
1879 | */ | |
9807b3d6 | 1880 | static void find_alignment(struct blame_scoreboard *sb, int *option) |
cee7f245 JH |
1881 | { |
1882 | int longest_src_lines = 0; | |
1883 | int longest_dst_lines = 0; | |
5ff62c30 | 1884 | unsigned largest_score = 0; |
cee7f245 | 1885 | struct blame_entry *e; |
b31272f7 | 1886 | int compute_auto_abbrev = (abbrev < 0); |
5293284b | 1887 | int auto_abbrev = DEFAULT_ABBREV; |
cee7f245 JH |
1888 | |
1889 | for (e = sb->ent; e; e = e->next) { | |
f84afb9c | 1890 | struct blame_origin *suspect = e->suspect; |
cee7f245 JH |
1891 | int num; |
1892 | ||
b31272f7 JH |
1893 | if (compute_auto_abbrev) |
1894 | auto_abbrev = update_auto_abbrev(auto_abbrev, suspect); | |
ab3bb800 JH |
1895 | if (strcmp(suspect->path, sb->path)) |
1896 | *option |= OUTPUT_SHOW_NAME; | |
1897 | num = strlen(suspect->path); | |
1898 | if (longest_file < num) | |
1899 | longest_file = num; | |
cee7f245 | 1900 | if (!(suspect->commit->object.flags & METAINFO_SHOWN)) { |
e6005927 | 1901 | struct commit_info ci; |
cee7f245 JH |
1902 | suspect->commit->object.flags |= METAINFO_SHOWN; |
1903 | get_commit_info(suspect->commit, &ci, 1); | |
1b8cdce9 | 1904 | if (*option & OUTPUT_SHOW_EMAIL) |
ea02ffa3 | 1905 | num = utf8_strwidth(ci.author_mail.buf); |
1b8cdce9 | 1906 | else |
ea02ffa3 | 1907 | num = utf8_strwidth(ci.author.buf); |
cee7f245 JH |
1908 | if (longest_author < num) |
1909 | longest_author = num; | |
e6005927 | 1910 | commit_info_destroy(&ci); |
cee7f245 JH |
1911 | } |
1912 | num = e->s_lno + e->num_lines; | |
1913 | if (longest_src_lines < num) | |
1914 | longest_src_lines = num; | |
1915 | num = e->lno + e->num_lines; | |
1916 | if (longest_dst_lines < num) | |
1917 | longest_dst_lines = num; | |
1a31a2d9 JS |
1918 | if (largest_score < blame_entry_score(sb, e)) |
1919 | largest_score = blame_entry_score(sb, e); | |
cee7f245 | 1920 | } |
ec7ff5ba ZJS |
1921 | max_orig_digits = decimal_width(longest_src_lines); |
1922 | max_digits = decimal_width(longest_dst_lines); | |
1923 | max_score_digits = decimal_width(largest_score); | |
b31272f7 JH |
1924 | |
1925 | if (compute_auto_abbrev) | |
1926 | /* one more abbrev length is needed for the boundary commit */ | |
1927 | abbrev = auto_abbrev + 1; | |
cee7f245 JH |
1928 | } |
1929 | ||
1732a1fd JH |
1930 | /* |
1931 | * For debugging -- origin is refcounted, and this asserts that | |
1932 | * we do not underflow. | |
1933 | */ | |
9807b3d6 | 1934 | static void sanity_check_refcnt(struct blame_scoreboard *sb) |
54a4c617 JH |
1935 | { |
1936 | int baa = 0; | |
1937 | struct blame_entry *ent; | |
1938 | ||
1939 | for (ent = sb->ent; ent; ent = ent->next) { | |
ae86ad65 | 1940 | /* Nobody should have zero or negative refcnt */ |
854b97f6 JH |
1941 | if (ent->suspect->refcnt <= 0) { |
1942 | fprintf(stderr, "%s in %s has negative refcnt %d\n", | |
1943 | ent->suspect->path, | |
f2fd0760 | 1944 | oid_to_hex(&ent->suspect->commit->object.oid), |
854b97f6 | 1945 | ent->suspect->refcnt); |
ae86ad65 | 1946 | baa = 1; |
854b97f6 | 1947 | } |
ae86ad65 | 1948 | } |
4149c186 JS |
1949 | if (baa) |
1950 | sb->on_sanity_fail(sb, baa); | |
1951 | } | |
1952 | ||
1953 | static void sanity_check_on_fail(struct blame_scoreboard *sb, int baa) | |
1954 | { | |
1955 | int opt = OUTPUT_SHOW_SCORE | OUTPUT_SHOW_NUMBER | OUTPUT_SHOW_NAME; | |
1956 | find_alignment(sb, &opt); | |
1957 | output(sb, opt); | |
1958 | die("Baa %d!", baa); | |
54a4c617 JH |
1959 | } |
1960 | ||
4a0fc95f JH |
1961 | static unsigned parse_score(const char *arg) |
1962 | { | |
1963 | char *end; | |
1964 | unsigned long score = strtoul(arg, &end, 10); | |
1965 | if (*end) | |
1966 | return 0; | |
1967 | return score; | |
1968 | } | |
1969 | ||
20239bae JK |
1970 | static const char *add_prefix(const char *prefix, const char *path) |
1971 | { | |
097971f5 | 1972 | return prefix_path(prefix, prefix ? strlen(prefix) : 0, path); |
20239bae JK |
1973 | } |
1974 | ||
ef90d6d4 | 1975 | static int git_blame_config(const char *var, const char *value, void *cb) |
4c10a5ca JH |
1976 | { |
1977 | if (!strcmp(var, "blame.showroot")) { | |
1978 | show_root = git_config_bool(var, value); | |
1979 | return 0; | |
1980 | } | |
1981 | if (!strcmp(var, "blame.blankboundary")) { | |
1982 | blank_boundary = git_config_bool(var, value); | |
1983 | return 0; | |
1984 | } | |
8b504db3 QN |
1985 | if (!strcmp(var, "blame.showemail")) { |
1986 | int *output_option = cb; | |
1987 | if (git_config_bool(var, value)) | |
1988 | *output_option |= OUTPUT_SHOW_EMAIL; | |
1989 | else | |
1990 | *output_option &= ~OUTPUT_SHOW_EMAIL; | |
1991 | return 0; | |
1992 | } | |
31653c1a EL |
1993 | if (!strcmp(var, "blame.date")) { |
1994 | if (!value) | |
1995 | return config_error_nonbool(var); | |
a5481a6c | 1996 | parse_date_format(value, &blame_date_mode); |
31653c1a EL |
1997 | return 0; |
1998 | } | |
3b8a12e8 | 1999 | |
5b162879 MH |
2000 | if (git_diff_heuristic_config(var, value, cb) < 0) |
2001 | return -1; | |
6680a087 | 2002 | if (userdiff_config(var, value) < 0) |
3b8a12e8 | 2003 | return -1; |
3b8a12e8 | 2004 | |
ef90d6d4 | 2005 | return git_default_config(var, value, cb); |
4c10a5ca JH |
2006 | } |
2007 | ||
7cb5f7c4 JK |
2008 | static struct commit *find_single_final(struct rev_info *revs, |
2009 | const char **name_p) | |
f6c07d7d JH |
2010 | { |
2011 | int i; | |
7cb5f7c4 JK |
2012 | struct commit *found = NULL; |
2013 | const char *name = NULL; | |
f6c07d7d | 2014 | |
f6c07d7d JH |
2015 | for (i = 0; i < revs->pending.nr; i++) { |
2016 | struct object *obj = revs->pending.objects[i].item; | |
2017 | if (obj->flags & UNINTERESTING) | |
2018 | continue; | |
31da121f | 2019 | obj = deref_tag(obj, NULL, 0); |
f6c07d7d JH |
2020 | if (obj->type != OBJ_COMMIT) |
2021 | die("Non commit %s?", revs->pending.objects[i].name); | |
1b0d4000 | 2022 | if (found) |
f6c07d7d | 2023 | die("More than one commit to dig from %s and %s?", |
7cb5f7c4 JK |
2024 | revs->pending.objects[i].name, name); |
2025 | found = (struct commit *)obj; | |
2026 | name = revs->pending.objects[i].name; | |
1b0d4000 | 2027 | } |
7cb5f7c4 JK |
2028 | if (name_p) |
2029 | *name_p = name; | |
1b0d4000 MK |
2030 | return found; |
2031 | } | |
2032 | ||
835c49f7 JS |
2033 | static struct commit *dwim_reverse_initial(struct rev_info *revs, |
2034 | const char **name_p) | |
e1d09701 JH |
2035 | { |
2036 | /* | |
2037 | * DWIM "git blame --reverse ONE -- PATH" as | |
2038 | * "git blame --reverse ONE..HEAD -- PATH" but only do so | |
2039 | * when it makes sense. | |
2040 | */ | |
2041 | struct object *obj; | |
2042 | struct commit *head_commit; | |
2043 | unsigned char head_sha1[20]; | |
2044 | ||
835c49f7 | 2045 | if (revs->pending.nr != 1) |
e1d09701 JH |
2046 | return NULL; |
2047 | ||
2048 | /* Is that sole rev a committish? */ | |
835c49f7 | 2049 | obj = revs->pending.objects[0].item; |
e1d09701 JH |
2050 | obj = deref_tag(obj, NULL, 0); |
2051 | if (obj->type != OBJ_COMMIT) | |
2052 | return NULL; | |
2053 | ||
2054 | /* Do we have HEAD? */ | |
2055 | if (!resolve_ref_unsafe("HEAD", RESOLVE_REF_READING, head_sha1, NULL)) | |
2056 | return NULL; | |
2057 | head_commit = lookup_commit_reference_gently(head_sha1, 1); | |
2058 | if (!head_commit) | |
2059 | return NULL; | |
2060 | ||
2061 | /* Turn "ONE" into "ONE..HEAD" then */ | |
2062 | obj->flags |= UNINTERESTING; | |
835c49f7 | 2063 | add_pending_object(revs, &head_commit->object, "HEAD"); |
e1d09701 | 2064 | |
835c49f7 JS |
2065 | if (name_p) |
2066 | *name_p = revs->pending.objects[0].name; | |
2067 | return (struct commit *)obj; | |
e1d09701 JH |
2068 | } |
2069 | ||
835c49f7 JS |
2070 | static struct commit *find_single_initial(struct rev_info *revs, |
2071 | const char **name_p) | |
85af7929 JH |
2072 | { |
2073 | int i; | |
2074 | const char *final_commit_name = NULL; | |
835c49f7 | 2075 | struct commit *found = NULL; |
85af7929 JH |
2076 | |
2077 | /* | |
2078 | * There must be one and only one negative commit, and it must be | |
2079 | * the boundary. | |
2080 | */ | |
2081 | for (i = 0; i < revs->pending.nr; i++) { | |
2082 | struct object *obj = revs->pending.objects[i].item; | |
2083 | if (!(obj->flags & UNINTERESTING)) | |
2084 | continue; | |
31da121f | 2085 | obj = deref_tag(obj, NULL, 0); |
85af7929 JH |
2086 | if (obj->type != OBJ_COMMIT) |
2087 | die("Non commit %s?", revs->pending.objects[i].name); | |
835c49f7 | 2088 | if (found) |
d993ce1e | 2089 | die("More than one commit to dig up from, %s and %s?", |
85af7929 JH |
2090 | revs->pending.objects[i].name, |
2091 | final_commit_name); | |
835c49f7 | 2092 | found = (struct commit *) obj; |
85af7929 JH |
2093 | final_commit_name = revs->pending.objects[i].name; |
2094 | } | |
e1d09701 JH |
2095 | |
2096 | if (!final_commit_name) | |
835c49f7 | 2097 | found = dwim_reverse_initial(revs, &final_commit_name); |
85af7929 | 2098 | if (!final_commit_name) |
d993ce1e | 2099 | die("No commit to dig up from?"); |
835c49f7 JS |
2100 | |
2101 | if (name_p) | |
2102 | *name_p = final_commit_name; | |
2103 | return found; | |
85af7929 JH |
2104 | } |
2105 | ||
5817da01 PH |
2106 | static int blame_copy_callback(const struct option *option, const char *arg, int unset) |
2107 | { | |
2108 | int *opt = option->value; | |
2109 | ||
2110 | /* | |
2111 | * -C enables copy from removed files; | |
2112 | * -C -C enables copy from existing files, but only | |
2113 | * when blaming a new file; | |
2114 | * -C -C -C enables copy from existing files for | |
2115 | * everybody | |
2116 | */ | |
2117 | if (*opt & PICKAXE_BLAME_COPY_HARDER) | |
2118 | *opt |= PICKAXE_BLAME_COPY_HARDEST; | |
2119 | if (*opt & PICKAXE_BLAME_COPY) | |
2120 | *opt |= PICKAXE_BLAME_COPY_HARDER; | |
2121 | *opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE; | |
2122 | ||
2123 | if (arg) | |
2124 | blame_copy_score = parse_score(arg); | |
2125 | return 0; | |
2126 | } | |
2127 | ||
2128 | static int blame_move_callback(const struct option *option, const char *arg, int unset) | |
2129 | { | |
2130 | int *opt = option->value; | |
2131 | ||
2132 | *opt |= PICKAXE_BLAME_MOVE; | |
2133 | ||
2134 | if (arg) | |
2135 | blame_move_score = parse_score(arg); | |
2136 | return 0; | |
2137 | } | |
2138 | ||
6e4c9b5b JS |
2139 | void init_scoreboard(struct blame_scoreboard *sb) |
2140 | { | |
2141 | memset(sb, 0, sizeof(struct blame_scoreboard)); | |
2142 | sb->move_score = BLAME_DEFAULT_MOVE_SCORE; | |
2143 | sb->copy_score = BLAME_DEFAULT_COPY_SCORE; | |
2144 | } | |
2145 | ||
d0d0ef1f JS |
2146 | void setup_scoreboard(struct blame_scoreboard *sb, const char *path, struct blame_origin **orig) |
2147 | { | |
2148 | const char *final_commit_name = NULL; | |
2149 | struct blame_origin *o; | |
2150 | struct commit *final_commit = NULL; | |
2151 | enum object_type type; | |
2152 | ||
2153 | if (sb->reverse && sb->contents_from) | |
2154 | die(_("--contents and --reverse do not blend well.")); | |
2155 | ||
2156 | if (!sb->reverse) { | |
2157 | sb->final = find_single_final(sb->revs, &final_commit_name); | |
2158 | sb->commits.compare = compare_commits_by_commit_date; | |
2159 | } else { | |
2160 | sb->final = find_single_initial(sb->revs, &final_commit_name); | |
2161 | sb->commits.compare = compare_commits_by_reverse_commit_date; | |
2162 | } | |
2163 | ||
2164 | if (sb->final && sb->contents_from) | |
2165 | die(_("cannot use --contents with final commit object name")); | |
2166 | ||
2167 | if (sb->reverse && sb->revs->first_parent_only) | |
2168 | sb->revs->children.name = NULL; | |
2169 | ||
2170 | if (!sb->final) { | |
2171 | /* | |
2172 | * "--not A B -- path" without anything positive; | |
2173 | * do not default to HEAD, but use the working tree | |
2174 | * or "--contents". | |
2175 | */ | |
2176 | setup_work_tree(); | |
2177 | sb->final = fake_working_tree_commit(&sb->revs->diffopt, | |
2178 | path, sb->contents_from); | |
2179 | add_pending_object(sb->revs, &(sb->final->object), ":"); | |
2180 | } | |
2181 | ||
2182 | if (sb->reverse && sb->revs->first_parent_only) { | |
2183 | final_commit = find_single_final(sb->revs, NULL); | |
2184 | if (!final_commit) | |
2185 | die(_("--reverse and --first-parent together require specified latest commit")); | |
2186 | } | |
2187 | ||
2188 | /* | |
2189 | * If we have bottom, this will mark the ancestors of the | |
2190 | * bottom commits we would reach while traversing as | |
2191 | * uninteresting. | |
2192 | */ | |
2193 | if (prepare_revision_walk(sb->revs)) | |
2194 | die(_("revision walk setup failed")); | |
2195 | ||
2196 | if (sb->reverse && sb->revs->first_parent_only) { | |
2197 | struct commit *c = final_commit; | |
2198 | ||
2199 | sb->revs->children.name = "children"; | |
2200 | while (c->parents && | |
2201 | oidcmp(&c->object.oid, &sb->final->object.oid)) { | |
2202 | struct commit_list *l = xcalloc(1, sizeof(*l)); | |
2203 | ||
2204 | l->item = c; | |
2205 | if (add_decoration(&sb->revs->children, | |
2206 | &c->parents->item->object, l)) | |
2207 | die("BUG: not unique item in first-parent chain"); | |
2208 | c = c->parents->item; | |
2209 | } | |
2210 | ||
2211 | if (oidcmp(&c->object.oid, &sb->final->object.oid)) | |
2212 | die(_("--reverse --first-parent together require range along first-parent chain")); | |
2213 | } | |
2214 | ||
2215 | if (is_null_oid(&sb->final->object.oid)) { | |
2216 | o = sb->final->util; | |
2217 | sb->final_buf = xmemdupz(o->file.ptr, o->file.size); | |
2218 | sb->final_buf_size = o->file.size; | |
2219 | } | |
2220 | else { | |
2221 | o = get_origin(sb->final, path); | |
2222 | if (fill_blob_sha1_and_mode(o)) | |
2223 | die(_("no such path %s in %s"), path, final_commit_name); | |
2224 | ||
2225 | if (DIFF_OPT_TST(&sb->revs->diffopt, ALLOW_TEXTCONV) && | |
2226 | textconv_object(path, o->mode, &o->blob_oid, 1, (char **) &sb->final_buf, | |
2227 | &sb->final_buf_size)) | |
2228 | ; | |
2229 | else | |
2230 | sb->final_buf = read_sha1_file(o->blob_oid.hash, &type, | |
2231 | &sb->final_buf_size); | |
2232 | ||
2233 | if (!sb->final_buf) | |
2234 | die(_("cannot read blob %s for path %s"), | |
2235 | oid_to_hex(&o->blob_oid), | |
2236 | path); | |
2237 | } | |
2238 | sb->num_read_blob++; | |
2239 | prepare_lines(sb); | |
2240 | ||
2241 | if (orig) | |
2242 | *orig = o; | |
2243 | } | |
2244 | ||
e94f77f0 JS |
2245 | struct blame_entry *blame_entry_prepend(struct blame_entry *head, |
2246 | long start, long end, | |
2247 | struct blame_origin *o) | |
2248 | { | |
2249 | struct blame_entry *new_head = xcalloc(1, sizeof(struct blame_entry)); | |
2250 | new_head->lno = start; | |
2251 | new_head->num_lines = end - start; | |
2252 | new_head->suspect = o; | |
2253 | new_head->s_lno = start; | |
2254 | new_head->next = head; | |
2255 | blame_origin_incref(o); | |
2256 | return new_head; | |
2257 | } | |
2258 | ||
acca687f | 2259 | int cmd_blame(int argc, const char **argv, const char *prefix) |
cee7f245 JH |
2260 | { |
2261 | struct rev_info revs; | |
2262 | const char *path; | |
9807b3d6 | 2263 | struct blame_scoreboard sb; |
f84afb9c | 2264 | struct blame_origin *o; |
58dbfa2e ES |
2265 | struct blame_entry *ent = NULL; |
2266 | long dashdash_pos, lno; | |
8c59921d | 2267 | struct progress_info pi = { NULL, 0 }; |
5817da01 | 2268 | |
64093fc0 JK |
2269 | struct string_list range_list = STRING_LIST_INIT_NODUP; |
2270 | int output_option = 0, opt = 0; | |
2271 | int show_stats = 0; | |
2272 | const char *revs_file = NULL; | |
2273 | const char *contents_from = NULL; | |
2274 | const struct option options[] = { | |
d5d09d47 SB |
2275 | OPT_BOOL(0, "incremental", &incremental, N_("Show blame entries as we find them, incrementally")), |
2276 | OPT_BOOL('b', NULL, &blank_boundary, N_("Show blank SHA-1 for boundary commits (Default: off)")), | |
2277 | OPT_BOOL(0, "root", &show_root, N_("Do not treat root commits as boundaries (Default: off)")), | |
2278 | OPT_BOOL(0, "show-stats", &show_stats, N_("Show work cost statistics")), | |
aba37f49 | 2279 | OPT_BOOL(0, "progress", &show_progress, N_("Force progress reporting")), |
efd2a8bd NTND |
2280 | OPT_BIT(0, "score-debug", &output_option, N_("Show output score for blame entries"), OUTPUT_SHOW_SCORE), |
2281 | OPT_BIT('f', "show-name", &output_option, N_("Show original filename (Default: auto)"), OUTPUT_SHOW_NAME), | |
2282 | OPT_BIT('n', "show-number", &output_option, N_("Show original linenumber (Default: off)"), OUTPUT_SHOW_NUMBER), | |
2283 | OPT_BIT('p', "porcelain", &output_option, N_("Show in a format designed for machine consumption"), OUTPUT_PORCELAIN), | |
2284 | OPT_BIT(0, "line-porcelain", &output_option, N_("Show porcelain format with per-line commit information"), OUTPUT_PORCELAIN|OUTPUT_LINE_PORCELAIN), | |
2285 | OPT_BIT('c', NULL, &output_option, N_("Use the same output mode as git-annotate (Default: off)"), OUTPUT_ANNOTATE_COMPAT), | |
2286 | OPT_BIT('t', NULL, &output_option, N_("Show raw timestamp (Default: off)"), OUTPUT_RAW_TIMESTAMP), | |
2287 | OPT_BIT('l', NULL, &output_option, N_("Show long commit SHA1 (Default: off)"), OUTPUT_LONG_OBJECT_NAME), | |
2288 | OPT_BIT('s', NULL, &output_option, N_("Suppress author name and timestamp (Default: off)"), OUTPUT_NO_AUTHOR), | |
2289 | OPT_BIT('e', "show-email", &output_option, N_("Show author email instead of name (Default: off)"), OUTPUT_SHOW_EMAIL), | |
2290 | OPT_BIT('w', NULL, &xdl_opts, N_("Ignore whitespace differences"), XDF_IGNORE_WHITESPACE), | |
5b162879 MH |
2291 | |
2292 | /* | |
2293 | * The following two options are parsed by parse_revision_opt() | |
2294 | * and are only included here to get included in the "-h" | |
2295 | * output: | |
2296 | */ | |
3cde4e02 | 2297 | { OPTION_LOWLEVEL_CALLBACK, 0, "indent-heuristic", NULL, NULL, N_("Use an experimental heuristic to improve diffs"), PARSE_OPT_NOARG, parse_opt_unknown_cb }, |
5b162879 | 2298 | |
efd2a8bd NTND |
2299 | OPT_BIT(0, "minimal", &xdl_opts, N_("Spend extra cycles to find better match"), XDF_NEED_MINIMAL), |
2300 | OPT_STRING('S', NULL, &revs_file, N_("file"), N_("Use revisions from <file> instead of calling git-rev-list")), | |
2301 | OPT_STRING(0, "contents", &contents_from, N_("file"), N_("Use <file>'s contents as the final image")), | |
2302 | { OPTION_CALLBACK, 'C', NULL, &opt, N_("score"), N_("Find line copies within and across files"), PARSE_OPT_OPTARG, blame_copy_callback }, | |
2303 | { OPTION_CALLBACK, 'M', NULL, &opt, N_("score"), N_("Find line movements within and across files"), PARSE_OPT_OPTARG, blame_move_callback }, | |
58dbfa2e | 2304 | OPT_STRING_LIST('L', NULL, &range_list, N_("n,m"), N_("Process only line range n,m, counting from 1")), |
84393bfd | 2305 | OPT__ABBREV(&abbrev), |
5817da01 PH |
2306 | OPT_END() |
2307 | }; | |
2308 | ||
2309 | struct parse_opt_ctx_t ctx; | |
7ceacdff | 2310 | int cmd_is_annotate = !strcmp(argv[0], "annotate"); |
58dbfa2e ES |
2311 | struct range_set ranges; |
2312 | unsigned int range_i; | |
52f4d126 | 2313 | long anchor; |
e68989a7 | 2314 | |
8b504db3 | 2315 | git_config(git_blame_config, &output_option); |
5817da01 | 2316 | init_revisions(&revs, NULL); |
31653c1a | 2317 | revs.date_mode = blame_date_mode; |
3b8a12e8 | 2318 | DIFF_OPT_SET(&revs.diffopt, ALLOW_TEXTCONV); |
3d1aa566 | 2319 | DIFF_OPT_SET(&revs.diffopt, FOLLOW_RENAMES); |
31653c1a | 2320 | |
612702e8 | 2321 | save_commit_buffer = 0; |
3f8d5204 | 2322 | dashdash_pos = 0; |
aba37f49 | 2323 | show_progress = -1; |
612702e8 | 2324 | |
9ca1169f SB |
2325 | parse_options_start(&ctx, argc, argv, prefix, options, |
2326 | PARSE_OPT_KEEP_DASHDASH | PARSE_OPT_KEEP_ARGV0); | |
5817da01 | 2327 | for (;;) { |
5817da01 PH |
2328 | switch (parse_options_step(&ctx, options, blame_opt_usage)) { |
2329 | case PARSE_OPT_HELP: | |
2330 | exit(129); | |
2331 | case PARSE_OPT_DONE: | |
3f8d5204 PH |
2332 | if (ctx.argv[0]) |
2333 | dashdash_pos = ctx.cpidx; | |
5817da01 PH |
2334 | goto parse_done; |
2335 | } | |
2336 | ||
2337 | if (!strcmp(ctx.argv[0], "--reverse")) { | |
2338 | ctx.argv[0] = "--children"; | |
2339 | reverse = 1; | |
2340 | } | |
6b61ec05 | 2341 | parse_revision_opt(&revs, &ctx, options, blame_opt_usage); |
5817da01 PH |
2342 | } |
2343 | parse_done: | |
3d1aa566 | 2344 | no_whole_file_rename = !DIFF_OPT_TST(&revs.diffopt, FOLLOW_RENAMES); |
3cde4e02 | 2345 | xdl_opts |= revs.diffopt.xdl_opts & XDF_INDENT_HEURISTIC; |
3d1aa566 | 2346 | DIFF_OPT_CLR(&revs.diffopt, FOLLOW_RENAMES); |
5817da01 PH |
2347 | argc = parse_options_end(&ctx); |
2348 | ||
aba37f49 ECA |
2349 | if (incremental || (output_option & OUTPUT_PORCELAIN)) { |
2350 | if (show_progress > 0) | |
e3f54bff | 2351 | die(_("--progress can't be used with --incremental or porcelain formats")); |
aba37f49 ECA |
2352 | show_progress = 0; |
2353 | } else if (show_progress < 0) | |
2354 | show_progress = isatty(2); | |
2355 | ||
91229834 | 2356 | if (0 < abbrev && abbrev < GIT_SHA1_HEXSZ) |
b31272f7 JH |
2357 | /* one more abbrev length is needed for the boundary commit */ |
2358 | abbrev++; | |
ed58d808 JK |
2359 | else if (!abbrev) |
2360 | abbrev = GIT_SHA1_HEXSZ; | |
84393bfd | 2361 | |
aa9ea77d | 2362 | if (revs_file && read_ancestry(revs_file)) |
d824cbba | 2363 | die_errno("reading graft file '%s' failed", revs_file); |
aa9ea77d | 2364 | |
31653c1a | 2365 | if (cmd_is_annotate) { |
7ceacdff | 2366 | output_option |= OUTPUT_ANNOTATE_COMPAT; |
a5481a6c | 2367 | blame_date_mode.type = DATE_ISO8601; |
31653c1a EL |
2368 | } else { |
2369 | blame_date_mode = revs.date_mode; | |
2370 | } | |
2371 | ||
2372 | /* The maximum width used to show the dates */ | |
a5481a6c | 2373 | switch (blame_date_mode.type) { |
31653c1a EL |
2374 | case DATE_RFC2822: |
2375 | blame_date_width = sizeof("Thu, 19 Oct 2006 16:00:04 -0700"); | |
2376 | break; | |
466fb674 BB |
2377 | case DATE_ISO8601_STRICT: |
2378 | blame_date_width = sizeof("2006-10-19T16:00:04-07:00"); | |
2379 | break; | |
31653c1a EL |
2380 | case DATE_ISO8601: |
2381 | blame_date_width = sizeof("2006-10-19 16:00:04 -0700"); | |
2382 | break; | |
2383 | case DATE_RAW: | |
2384 | blame_date_width = sizeof("1161298804 -0700"); | |
2385 | break; | |
642833db JK |
2386 | case DATE_UNIX: |
2387 | blame_date_width = sizeof("1161298804"); | |
2388 | break; | |
31653c1a EL |
2389 | case DATE_SHORT: |
2390 | blame_date_width = sizeof("2006-10-19"); | |
2391 | break; | |
2392 | case DATE_RELATIVE: | |
dd75553b JX |
2393 | /* TRANSLATORS: This string is used to tell us the maximum |
2394 | display width for a relative timestamp in "git blame" | |
2395 | output. For C locale, "4 years, 11 months ago", which | |
2396 | takes 22 places, is the longest among various forms of | |
2397 | relative timestamps, but your language may need more or | |
2398 | fewer display columns. */ | |
2399 | blame_date_width = utf8_strwidth(_("4 years, 11 months ago")) + 1; /* add the null */ | |
2400 | break; | |
31653c1a EL |
2401 | case DATE_NORMAL: |
2402 | blame_date_width = sizeof("Thu Oct 19 16:00:04 2006 -0700"); | |
2403 | break; | |
aa1462cc JK |
2404 | case DATE_STRFTIME: |
2405 | blame_date_width = strlen(show_date(0, 0, &blame_date_mode)) + 1; /* add the null */ | |
2406 | break; | |
31653c1a EL |
2407 | } |
2408 | blame_date_width -= 1; /* strip the null */ | |
7ceacdff | 2409 | |
b3123f98 JH |
2410 | if (DIFF_OPT_TST(&revs.diffopt, FIND_COPIES_HARDER)) |
2411 | opt |= (PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE | | |
2412 | PICKAXE_BLAME_COPY_HARDER); | |
2413 | ||
1732a1fd JH |
2414 | /* |
2415 | * We have collected options unknown to us in argv[1..unk] | |
cee7f245 | 2416 | * which are to be passed to revision machinery if we are |
3dff5379 | 2417 | * going to do the "bottom" processing. |
cee7f245 JH |
2418 | * |
2419 | * The remaining are: | |
2420 | * | |
22e5e58a | 2421 | * (1) if dashdash_pos != 0, it is either |
3f8d5204 PH |
2422 | * "blame [revisions] -- <path>" or |
2423 | * "blame -- <path> <rev>" | |
cee7f245 | 2424 | * |
22e5e58a | 2425 | * (2) otherwise, it is one of the two: |
3f8d5204 PH |
2426 | * "blame [revisions] <path>" |
2427 | * "blame <path> <rev>" | |
cee7f245 | 2428 | * |
3f8d5204 PH |
2429 | * Note that we must strip out <path> from the arguments: we do not |
2430 | * want the path pruning but we may want "bottom" processing. | |
cee7f245 | 2431 | */ |
3f8d5204 PH |
2432 | if (dashdash_pos) { |
2433 | switch (argc - dashdash_pos - 1) { | |
2434 | case 2: /* (1b) */ | |
2435 | if (argc != 4) | |
5817da01 | 2436 | usage_with_options(blame_opt_usage, options); |
3f8d5204 PH |
2437 | /* reorder for the new way: <rev> -- <path> */ |
2438 | argv[1] = argv[3]; | |
2439 | argv[3] = argv[2]; | |
2440 | argv[2] = "--"; | |
2441 | /* FALLTHROUGH */ | |
2442 | case 1: /* (1a) */ | |
2443 | path = add_prefix(prefix, argv[--argc]); | |
2444 | argv[argc] = NULL; | |
2445 | break; | |
2446 | default: | |
2447 | usage_with_options(blame_opt_usage, options); | |
cee7f245 | 2448 | } |
3f8d5204 PH |
2449 | } else { |
2450 | if (argc < 2) | |
5817da01 | 2451 | usage_with_options(blame_opt_usage, options); |
3f8d5204 | 2452 | path = add_prefix(prefix, argv[argc - 1]); |
dbe44faa | 2453 | if (argc == 3 && !file_exists(path)) { /* (2b) */ |
3f8d5204 PH |
2454 | path = add_prefix(prefix, argv[1]); |
2455 | argv[1] = argv[2]; | |
cee7f245 | 2456 | } |
3f8d5204 | 2457 | argv[argc - 1] = "--"; |
cee7f245 | 2458 | |
3f8d5204 | 2459 | setup_work_tree(); |
dbe44faa | 2460 | if (!file_exists(path)) |
d824cbba | 2461 | die_errno("cannot stat path '%s'", path); |
cee7f245 JH |
2462 | } |
2463 | ||
8b3dce56 | 2464 | revs.disable_stdin = 1; |
3f8d5204 | 2465 | setup_revisions(argc, argv, &revs, NULL); |
cee7f245 | 2466 | |
6e4c9b5b | 2467 | init_scoreboard(&sb); |
85af7929 | 2468 | sb.revs = &revs; |
84be875e | 2469 | sb.contents_from = contents_from; |
f81d70e9 | 2470 | sb.reverse = reverse; |
d0d0ef1f JS |
2471 | setup_scoreboard(&sb, path, &o); |
2472 | lno = sb.num_lines; | |
cee7f245 | 2473 | |
58dbfa2e | 2474 | if (lno && !range_list.nr) |
aa59e14b | 2475 | string_list_append(&range_list, "1"); |
58dbfa2e | 2476 | |
52f4d126 | 2477 | anchor = 1; |
58dbfa2e ES |
2478 | range_set_init(&ranges, range_list.nr); |
2479 | for (range_i = 0; range_i < range_list.nr; ++range_i) { | |
2480 | long bottom, top; | |
2481 | if (parse_range_arg(range_list.items[range_i].string, | |
52f4d126 | 2482 | nth_line_cb, &sb, lno, anchor, |
58dbfa2e ES |
2483 | &bottom, &top, sb.path)) |
2484 | usage(blame_usage); | |
2485 | if (lno < top || ((lno || bottom) && lno < bottom)) | |
e3f54bff VA |
2486 | die(Q_("file %s has only %lu line", |
2487 | "file %s has only %lu lines", | |
2488 | lno), path, lno); | |
58dbfa2e ES |
2489 | if (bottom < 1) |
2490 | bottom = 1; | |
2491 | if (top < 1) | |
2492 | top = lno; | |
2493 | bottom--; | |
2494 | range_set_append_unsafe(&ranges, bottom, top); | |
52f4d126 | 2495 | anchor = top + 1; |
58dbfa2e ES |
2496 | } |
2497 | sort_and_merge_range_set(&ranges); | |
2498 | ||
2499 | for (range_i = ranges.nr; range_i > 0; --range_i) { | |
2500 | const struct range *r = &ranges.ranges[range_i - 1]; | |
e94f77f0 | 2501 | ent = blame_entry_prepend(ent, r->start, r->end, o); |
58dbfa2e | 2502 | } |
7e6ac6e4 DK |
2503 | |
2504 | o->suspects = ent; | |
2505 | prio_queue_put(&sb.commits, o->commit); | |
2506 | ||
006a0744 | 2507 | blame_origin_decref(o); |
58dbfa2e ES |
2508 | |
2509 | range_set_release(&ranges); | |
2510 | string_list_clear(&range_list, 0); | |
cee7f245 | 2511 | |
7e6ac6e4 | 2512 | sb.ent = NULL; |
cee7f245 JH |
2513 | sb.path = path; |
2514 | ||
18ec0d62 JS |
2515 | if (blame_move_score) |
2516 | sb.move_score = blame_move_score; | |
2517 | if (blame_copy_score) | |
2518 | sb.copy_score = blame_copy_score; | |
2519 | ||
4149c186 JS |
2520 | sb.debug = DEBUG; |
2521 | sb.on_sanity_fail = &sanity_check_on_fail; | |
2522 | ||
2cf83374 | 2523 | sb.show_root = show_root; |
73e1c299 | 2524 | sb.xdl_opts = xdl_opts; |
1f44129b | 2525 | sb.no_whole_file_rename = no_whole_file_rename; |
2cf83374 | 2526 | |
d551a488 | 2527 | read_mailmap(&mailmap, NULL); |
f9567384 | 2528 | |
8c59921d JS |
2529 | sb.found_guilty_entry = &found_guilty_entry; |
2530 | sb.found_guilty_entry_data = π | |
2531 | if (show_progress) | |
2532 | pi.progress = start_progress_delay(_("Blaming lines"), | |
2533 | sb.num_lines, 50, 1); | |
2534 | ||
aba37f49 ECA |
2535 | assign_blame(&sb, opt); |
2536 | ||
8c59921d JS |
2537 | stop_progress(&pi.progress); |
2538 | ||
b92565dc MH |
2539 | if (!incremental) |
2540 | setup_pager(); | |
835c49f7 | 2541 | else |
717d1462 LT |
2542 | return 0; |
2543 | ||
78b06e66 | 2544 | blame_sort_final(&sb); |
7e6ac6e4 | 2545 | |
c6971362 | 2546 | blame_coalesce(&sb); |
cee7f245 JH |
2547 | |
2548 | if (!(output_option & OUTPUT_PORCELAIN)) | |
2549 | find_alignment(&sb, &output_option); | |
2550 | ||
2551 | output(&sb, output_option); | |
2552 | free((void *)sb.final_buf); | |
2553 | for (ent = sb.ent; ent; ) { | |
2554 | struct blame_entry *e = ent->next; | |
2555 | free(ent); | |
2556 | ent = e; | |
2557 | } | |
c2e525d9 | 2558 | |
870b39c1 | 2559 | if (show_stats) { |
8449528d JS |
2560 | printf("num read blob: %d\n", sb.num_read_blob); |
2561 | printf("num get patch: %d\n", sb.num_get_patch); | |
2562 | printf("num commits: %d\n", sb.num_commits); | |
c2e525d9 | 2563 | } |
cee7f245 JH |
2564 | return 0; |
2565 | } |