Add mailmap.file as configurational option for mailmap location
[git/git.git] / builtin-blame.c
CommitLineData
cee7f245
JH
1/*
2 * Pickaxe
3 *
4 * Copyright (c) 2006, Junio C Hamano
5 */
6
7#include "cache.h"
8#include "builtin.h"
9#include "blob.h"
10#include "commit.h"
11#include "tag.h"
12#include "tree-walk.h"
13#include "diff.h"
14#include "diffcore.h"
15#include "revision.h"
717d1462 16#include "quote.h"
cee7f245 17#include "xdiff-interface.h"
1cfe7733 18#include "cache-tree.h"
c455c87c 19#include "string-list.h"
f9567384 20#include "mailmap.h"
5817da01 21#include "parse-options.h"
cee7f245 22
588c038a 23static char blame_usage[] = "git blame [options] [rev-opts] [rev] [--] file";
5817da01
PH
24
25static const char *blame_opt_usage[] = {
26 blame_usage,
27 "",
76447d23 28 "[rev-opts] are documented in git-rev-list(1)",
5817da01
PH
29 NULL
30};
cee7f245
JH
31
32static int longest_file;
33static int longest_author;
34static int max_orig_digits;
35static int max_digits;
5ff62c30 36static int max_score_digits;
4c10a5ca 37static int show_root;
85af7929 38static int reverse;
4c10a5ca 39static int blank_boundary;
717d1462 40static int incremental;
b82871b3 41static int xdl_opts = XDF_NEED_MINIMAL;
c455c87c 42static struct string_list mailmap;
cee7f245 43
54a4c617
JH
44#ifndef DEBUG
45#define DEBUG 0
46#endif
47
c2e525d9
JH
48/* stats */
49static int num_read_blob;
50static int num_get_patch;
51static int num_commits;
52
d24bba80 53#define PICKAXE_BLAME_MOVE 01
18abd745
JH
54#define PICKAXE_BLAME_COPY 02
55#define PICKAXE_BLAME_COPY_HARDER 04
c63777c0 56#define PICKAXE_BLAME_COPY_HARDEST 010
d24bba80 57
4a0fc95f
JH
58/*
59 * blame for a blame_entry with score lower than these thresholds
60 * is not passed to the parent using move/copy logic.
61 */
62static unsigned blame_move_score;
63static unsigned blame_copy_score;
64#define BLAME_DEFAULT_MOVE_SCORE 20
65#define BLAME_DEFAULT_COPY_SCORE 40
66
cee7f245
JH
67/* bits #0..7 in revision.h, #8..11 used for merge_bases() in commit.c */
68#define METAINFO_SHOWN (1u<<12)
69#define MORE_THAN_ONE_PATH (1u<<13)
70
71/*
54a4c617 72 * One blob in a commit that is being suspected
cee7f245
JH
73 */
74struct origin {
54a4c617 75 int refcnt;
cee7f245 76 struct commit *commit;
c2e525d9 77 mmfile_t file;
cee7f245
JH
78 unsigned char blob_sha1[20];
79 char path[FLEX_ARRAY];
80};
81
1732a1fd
JH
82/*
83 * Given an origin, prepare mmfile_t structure to be used by the
84 * diff machinery
85 */
f6c07d7d 86static void fill_origin_blob(struct origin *o, mmfile_t *file)
c2e525d9
JH
87{
88 if (!o->file.ptr) {
21666f1a 89 enum object_type type;
c2e525d9 90 num_read_blob++;
21666f1a 91 file->ptr = read_sha1_file(o->blob_sha1, &type,
c2e525d9 92 (unsigned long *)(&(file->size)));
ab43e495
JH
93 if (!file->ptr)
94 die("Cannot read blob %s for path %s",
95 sha1_to_hex(o->blob_sha1),
96 o->path);
c2e525d9
JH
97 o->file = *file;
98 }
99 else
100 *file = o->file;
c2e525d9
JH
101}
102
1732a1fd
JH
103/*
104 * Origin is refcounted and usually we keep the blob contents to be
105 * reused.
106 */
54a4c617
JH
107static inline struct origin *origin_incref(struct origin *o)
108{
109 if (o)
110 o->refcnt++;
111 return o;
112}
113
114static void origin_decref(struct origin *o)
115{
116 if (o && --o->refcnt <= 0) {
8e0f7003 117 free(o->file.ptr);
54a4c617
JH
118 free(o);
119 }
120}
121
7c3c7962
JH
122static void drop_origin_blob(struct origin *o)
123{
124 if (o->file.ptr) {
125 free(o->file.ptr);
126 o->file.ptr = NULL;
127 }
128}
129
1732a1fd
JH
130/*
131 * Each group of lines is described by a blame_entry; it can be split
132 * as we pass blame to the parents. They form a linked list in the
133 * scoreboard structure, sorted by the target line number.
134 */
cee7f245
JH
135struct blame_entry {
136 struct blame_entry *prev;
137 struct blame_entry *next;
138
139 /* the first line of this group in the final image;
140 * internally all line numbers are 0 based.
141 */
142 int lno;
143
144 /* how many lines this group has */
145 int num_lines;
146
147 /* the commit that introduced this group into the final image */
148 struct origin *suspect;
149
150 /* true if the suspect is truly guilty; false while we have not
151 * checked if the group came from one of its parents.
152 */
153 char guilty;
154
eb46328e
AG
155 /* true if the entry has been scanned for copies in the current parent
156 */
157 char scanned;
158
cee7f245
JH
159 /* the line number of the first line of this group in the
160 * suspect's file; internally all line numbers are 0 based.
161 */
162 int s_lno;
5ff62c30
JH
163
164 /* how significant this entry is -- cached to avoid
1732a1fd 165 * scanning the lines over and over.
5ff62c30
JH
166 */
167 unsigned score;
cee7f245
JH
168};
169
1732a1fd
JH
170/*
171 * The current state of the blame assignment.
172 */
cee7f245
JH
173struct scoreboard {
174 /* the final commit (i.e. where we started digging from) */
175 struct commit *final;
85af7929 176 struct rev_info *revs;
cee7f245
JH
177 const char *path;
178
1732a1fd
JH
179 /*
180 * The contents in the final image.
181 * Used by many functions to obtain contents of the nth line,
182 * indexed with scoreboard.lineno[blame_entry.lno].
cee7f245
JH
183 */
184 const char *final_buf;
185 unsigned long final_buf_size;
186
187 /* linked list of blames */
188 struct blame_entry *ent;
189
612702e8 190 /* look-up a line in the final buffer */
cee7f245
JH
191 int num_lines;
192 int *lineno;
193};
194
171dccd5 195static inline int same_suspect(struct origin *a, struct origin *b)
46014766 196{
171dccd5 197 if (a == b)
57584d9e 198 return 1;
171dccd5
JH
199 if (a->commit != b->commit)
200 return 0;
201 return !strcmp(a->path, b->path);
46014766
JH
202}
203
54a4c617
JH
204static void sanity_check_refcnt(struct scoreboard *);
205
1732a1fd
JH
206/*
207 * If two blame entries that are next to each other came from
208 * contiguous lines in the same origin (i.e. <commit, path> pair),
209 * merge them together.
210 */
cee7f245
JH
211static void coalesce(struct scoreboard *sb)
212{
213 struct blame_entry *ent, *next;
214
215 for (ent = sb->ent; ent && (next = ent->next); ent = next) {
171dccd5 216 if (same_suspect(ent->suspect, next->suspect) &&
cee7f245
JH
217 ent->guilty == next->guilty &&
218 ent->s_lno + ent->num_lines == next->s_lno) {
219 ent->num_lines += next->num_lines;
220 ent->next = next->next;
221 if (ent->next)
222 ent->next->prev = ent;
54a4c617 223 origin_decref(next->suspect);
cee7f245 224 free(next);
46014766 225 ent->score = 0;
cee7f245
JH
226 next = ent; /* again */
227 }
228 }
54a4c617
JH
229
230 if (DEBUG) /* sanity */
231 sanity_check_refcnt(sb);
cee7f245
JH
232}
233
1732a1fd
JH
234/*
235 * Given a commit and a path in it, create a new origin structure.
236 * The callers that add blame to the scoreboard should use
237 * get_origin() to obtain shared, refcounted copy instead of calling
238 * this function directly.
239 */
854b97f6
JH
240static struct origin *make_origin(struct commit *commit, const char *path)
241{
242 struct origin *o;
243 o = xcalloc(1, sizeof(*o) + strlen(path) + 1);
244 o->commit = commit;
245 o->refcnt = 1;
246 strcpy(o->path, path);
247 return o;
248}
249
1732a1fd
JH
250/*
251 * Locate an existing origin or create a new one.
252 */
f6c0e191
JH
253static struct origin *get_origin(struct scoreboard *sb,
254 struct commit *commit,
255 const char *path)
cee7f245 256{
f6c0e191 257 struct blame_entry *e;
cee7f245 258
f6c0e191
JH
259 for (e = sb->ent; e; e = e->next) {
260 if (e->suspect->commit == commit &&
261 !strcmp(e->suspect->path, path))
54a4c617 262 return origin_incref(e->suspect);
cee7f245 263 }
854b97f6 264 return make_origin(commit, path);
cee7f245
JH
265}
266
1732a1fd
JH
267/*
268 * Fill the blob_sha1 field of an origin if it hasn't, so that later
269 * call to fill_origin_blob() can use it to locate the data. blob_sha1
270 * for an origin is also used to pass the blame for the entire file to
271 * the parent to detect the case where a child's blob is identical to
272 * that of its parent's.
273 */
f6c0e191
JH
274static int fill_blob_sha1(struct origin *origin)
275{
276 unsigned mode;
f6c0e191
JH
277
278 if (!is_null_sha1(origin->blob_sha1))
279 return 0;
280 if (get_tree_entry(origin->commit->object.sha1,
281 origin->path,
282 origin->blob_sha1, &mode))
283 goto error_out;
21666f1a 284 if (sha1_object_info(origin->blob_sha1, NULL) != OBJ_BLOB)
f6c0e191
JH
285 goto error_out;
286 return 0;
287 error_out:
288 hashclr(origin->blob_sha1);
289 return -1;
290}
291
1732a1fd
JH
292/*
293 * We have an origin -- check if the same path exists in the
294 * parent and return an origin structure to represent it.
295 */
f6c0e191 296static struct origin *find_origin(struct scoreboard *sb,
cee7f245
JH
297 struct commit *parent,
298 struct origin *origin)
299{
300 struct origin *porigin = NULL;
301 struct diff_options diff_opts;
f6c0e191
JH
302 const char *paths[2];
303
0d981c67 304 if (parent->util) {
1732a1fd
JH
305 /*
306 * Each commit object can cache one origin in that
307 * commit. This is a freestanding copy of origin and
308 * not refcounted.
854b97f6 309 */
0d981c67 310 struct origin *cached = parent->util;
854b97f6 311 if (!strcmp(cached->path, origin->path)) {
1732a1fd
JH
312 /*
313 * The same path between origin and its parent
314 * without renaming -- the most common case.
315 */
854b97f6 316 porigin = get_origin(sb, parent, cached->path);
1732a1fd
JH
317
318 /*
319 * If the origin was newly created (i.e. get_origin
320 * would call make_origin if none is found in the
321 * scoreboard), it does not know the blob_sha1,
322 * so copy it. Otherwise porigin was in the
323 * scoreboard and already knows blob_sha1.
324 */
854b97f6
JH
325 if (porigin->refcnt == 1)
326 hashcpy(porigin->blob_sha1, cached->blob_sha1);
327 return porigin;
328 }
329 /* otherwise it was not very useful; free it */
330 free(parent->util);
331 parent->util = NULL;
0d981c67
JH
332 }
333
f6c0e191
JH
334 /* See if the origin->path is different between parent
335 * and origin first. Most of the time they are the
336 * same and diff-tree is fairly efficient about this.
337 */
338 diff_setup(&diff_opts);
8f67f8ae 339 DIFF_OPT_SET(&diff_opts, RECURSIVE);
f6c0e191
JH
340 diff_opts.detect_rename = 0;
341 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
342 paths[0] = origin->path;
343 paths[1] = NULL;
344
345 diff_tree_setup_paths(paths, &diff_opts);
346 if (diff_setup_done(&diff_opts) < 0)
347 die("diff-setup");
1cfe7733
JH
348
349 if (is_null_sha1(origin->commit->object.sha1))
350 do_diff_cache(parent->tree->object.sha1, &diff_opts);
351 else
352 diff_tree_sha1(parent->tree->object.sha1,
353 origin->commit->tree->object.sha1,
354 "", &diff_opts);
f6c0e191
JH
355 diffcore_std(&diff_opts);
356
357 /* It is either one entry that says "modified", or "created",
358 * or nothing.
359 */
360 if (!diff_queued_diff.nr) {
361 /* The path is the same as parent */
362 porigin = get_origin(sb, parent, origin->path);
363 hashcpy(porigin->blob_sha1, origin->blob_sha1);
364 }
365 else if (diff_queued_diff.nr != 1)
acca687f 366 die("internal error in blame::find_origin");
f6c0e191
JH
367 else {
368 struct diff_filepair *p = diff_queued_diff.queue[0];
369 switch (p->status) {
370 default:
acca687f 371 die("internal error in blame::find_origin (%c)",
f6c0e191
JH
372 p->status);
373 case 'M':
374 porigin = get_origin(sb, parent, origin->path);
375 hashcpy(porigin->blob_sha1, p->one->sha1);
376 break;
377 case 'A':
378 case 'T':
379 /* Did not exist in parent, or type changed */
380 break;
381 }
382 }
383 diff_flush(&diff_opts);
03b69c76 384 diff_tree_release_paths(&diff_opts);
0d981c67 385 if (porigin) {
1732a1fd
JH
386 /*
387 * Create a freestanding copy that is not part of
388 * the refcounted origin found in the scoreboard, and
389 * cache it in the commit.
390 */
854b97f6 391 struct origin *cached;
1732a1fd 392
854b97f6
JH
393 cached = make_origin(porigin->commit, porigin->path);
394 hashcpy(cached->blob_sha1, porigin->blob_sha1);
395 parent->util = cached;
0d981c67 396 }
f69e743d
JH
397 return porigin;
398}
f6c0e191 399
1732a1fd
JH
400/*
401 * We have an origin -- find the path that corresponds to it in its
402 * parent and return an origin structure to represent it.
403 */
f69e743d
JH
404static struct origin *find_rename(struct scoreboard *sb,
405 struct commit *parent,
406 struct origin *origin)
407{
408 struct origin *porigin = NULL;
409 struct diff_options diff_opts;
410 int i;
411 const char *paths[2];
cee7f245
JH
412
413 diff_setup(&diff_opts);
8f67f8ae 414 DIFF_OPT_SET(&diff_opts, RECURSIVE);
cee7f245
JH
415 diff_opts.detect_rename = DIFF_DETECT_RENAME;
416 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
2f3f8b21 417 diff_opts.single_follow = origin->path;
cee7f245
JH
418 paths[0] = NULL;
419 diff_tree_setup_paths(paths, &diff_opts);
420 if (diff_setup_done(&diff_opts) < 0)
421 die("diff-setup");
1cfe7733
JH
422
423 if (is_null_sha1(origin->commit->object.sha1))
424 do_diff_cache(parent->tree->object.sha1, &diff_opts);
425 else
426 diff_tree_sha1(parent->tree->object.sha1,
427 origin->commit->tree->object.sha1,
428 "", &diff_opts);
cee7f245
JH
429 diffcore_std(&diff_opts);
430
431 for (i = 0; i < diff_queued_diff.nr; i++) {
432 struct diff_filepair *p = diff_queued_diff.queue[i];
612702e8 433 if ((p->status == 'R' || p->status == 'C') &&
f6c0e191
JH
434 !strcmp(p->two->path, origin->path)) {
435 porigin = get_origin(sb, parent, p->one->path);
436 hashcpy(porigin->blob_sha1, p->one->sha1);
cee7f245
JH
437 break;
438 }
439 }
440 diff_flush(&diff_opts);
03b69c76 441 diff_tree_release_paths(&diff_opts);
cee7f245
JH
442 return porigin;
443}
444
1732a1fd 445/*
3dff5379 446 * Link in a new blame entry to the scoreboard. Entries that cover the
1732a1fd
JH
447 * same line range have been removed from the scoreboard previously.
448 */
cee7f245
JH
449static void add_blame_entry(struct scoreboard *sb, struct blame_entry *e)
450{
451 struct blame_entry *ent, *prev = NULL;
452
54a4c617
JH
453 origin_incref(e->suspect);
454
cee7f245
JH
455 for (ent = sb->ent; ent && ent->lno < e->lno; ent = ent->next)
456 prev = ent;
457
458 /* prev, if not NULL, is the last one that is below e */
459 e->prev = prev;
460 if (prev) {
461 e->next = prev->next;
462 prev->next = e;
463 }
464 else {
465 e->next = sb->ent;
466 sb->ent = e;
467 }
468 if (e->next)
469 e->next->prev = e;
470}
471
1732a1fd
JH
472/*
473 * src typically is on-stack; we want to copy the information in it to
790296fd 474 * a malloced blame_entry that is already on the linked list of the
1732a1fd
JH
475 * scoreboard. The origin of dst loses a refcnt while the origin of src
476 * gains one.
477 */
cee7f245
JH
478static void dup_entry(struct blame_entry *dst, struct blame_entry *src)
479{
480 struct blame_entry *p, *n;
54a4c617 481
cee7f245
JH
482 p = dst->prev;
483 n = dst->next;
54a4c617
JH
484 origin_incref(src->suspect);
485 origin_decref(dst->suspect);
cee7f245
JH
486 memcpy(dst, src, sizeof(*src));
487 dst->prev = p;
488 dst->next = n;
5ff62c30 489 dst->score = 0;
cee7f245
JH
490}
491
492static const char *nth_line(struct scoreboard *sb, int lno)
493{
494 return sb->final_buf + sb->lineno[lno];
495}
496
1732a1fd
JH
497/*
498 * It is known that lines between tlno to same came from parent, and e
499 * has an overlap with that range. it also is known that parent's
500 * line plno corresponds to e's line tlno.
501 *
502 * <---- e ----->
503 * <------>
504 * <------------>
505 * <------------>
506 * <------------------>
507 *
508 * Split e into potentially three parts; before this chunk, the chunk
509 * to be blamed for the parent, and after that portion.
510 */
54a4c617 511static void split_overlap(struct blame_entry *split,
cee7f245
JH
512 struct blame_entry *e,
513 int tlno, int plno, int same,
514 struct origin *parent)
515{
cee7f245
JH
516 int chunk_end_lno;
517 memset(split, 0, sizeof(struct blame_entry [3]));
518
519 if (e->s_lno < tlno) {
520 /* there is a pre-chunk part not blamed on parent */
54a4c617 521 split[0].suspect = origin_incref(e->suspect);
cee7f245
JH
522 split[0].lno = e->lno;
523 split[0].s_lno = e->s_lno;
524 split[0].num_lines = tlno - e->s_lno;
525 split[1].lno = e->lno + tlno - e->s_lno;
526 split[1].s_lno = plno;
527 }
528 else {
529 split[1].lno = e->lno;
530 split[1].s_lno = plno + (e->s_lno - tlno);
531 }
532
533 if (same < e->s_lno + e->num_lines) {
534 /* there is a post-chunk part not blamed on parent */
54a4c617 535 split[2].suspect = origin_incref(e->suspect);
cee7f245
JH
536 split[2].lno = e->lno + (same - e->s_lno);
537 split[2].s_lno = e->s_lno + (same - e->s_lno);
538 split[2].num_lines = e->s_lno + e->num_lines - same;
539 chunk_end_lno = split[2].lno;
540 }
541 else
542 chunk_end_lno = e->lno + e->num_lines;
543 split[1].num_lines = chunk_end_lno - split[1].lno;
544
1732a1fd
JH
545 /*
546 * if it turns out there is nothing to blame the parent for,
547 * forget about the splitting. !split[1].suspect signals this.
548 */
cee7f245
JH
549 if (split[1].num_lines < 1)
550 return;
54a4c617 551 split[1].suspect = origin_incref(parent);
cee7f245
JH
552}
553
1732a1fd
JH
554/*
555 * split_overlap() divided an existing blame e into up to three parts
556 * in split. Adjust the linked list of blames in the scoreboard to
557 * reflect the split.
558 */
cee7f245 559static void split_blame(struct scoreboard *sb,
54a4c617 560 struct blame_entry *split,
cee7f245
JH
561 struct blame_entry *e)
562{
563 struct blame_entry *new_entry;
564
565 if (split[0].suspect && split[2].suspect) {
1732a1fd 566 /* The first part (reuse storage for the existing entry e) */
cee7f245
JH
567 dup_entry(e, &split[0]);
568
1732a1fd 569 /* The last part -- me */
cee7f245
JH
570 new_entry = xmalloc(sizeof(*new_entry));
571 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
572 add_blame_entry(sb, new_entry);
573
1732a1fd 574 /* ... and the middle part -- parent */
cee7f245
JH
575 new_entry = xmalloc(sizeof(*new_entry));
576 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
577 add_blame_entry(sb, new_entry);
578 }
579 else if (!split[0].suspect && !split[2].suspect)
1732a1fd
JH
580 /*
581 * The parent covers the entire area; reuse storage for
582 * e and replace it with the parent.
583 */
cee7f245
JH
584 dup_entry(e, &split[1]);
585 else if (split[0].suspect) {
1732a1fd 586 /* me and then parent */
cee7f245
JH
587 dup_entry(e, &split[0]);
588
589 new_entry = xmalloc(sizeof(*new_entry));
590 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
591 add_blame_entry(sb, new_entry);
592 }
593 else {
1732a1fd 594 /* parent and then me */
cee7f245
JH
595 dup_entry(e, &split[1]);
596
597 new_entry = xmalloc(sizeof(*new_entry));
598 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
599 add_blame_entry(sb, new_entry);
600 }
601
54a4c617 602 if (DEBUG) { /* sanity */
cee7f245 603 struct blame_entry *ent;
612702e8 604 int lno = sb->ent->lno, corrupt = 0;
cee7f245
JH
605
606 for (ent = sb->ent; ent; ent = ent->next) {
607 if (lno != ent->lno)
608 corrupt = 1;
609 if (ent->s_lno < 0)
610 corrupt = 1;
611 lno += ent->num_lines;
612 }
613 if (corrupt) {
612702e8 614 lno = sb->ent->lno;
cee7f245
JH
615 for (ent = sb->ent; ent; ent = ent->next) {
616 printf("L %8d l %8d n %8d\n",
617 lno, ent->lno, ent->num_lines);
618 lno = ent->lno + ent->num_lines;
619 }
620 die("oops");
621 }
622 }
623}
624
1732a1fd
JH
625/*
626 * After splitting the blame, the origins used by the
627 * on-stack blame_entry should lose one refcnt each.
628 */
54a4c617
JH
629static void decref_split(struct blame_entry *split)
630{
631 int i;
632
633 for (i = 0; i < 3; i++)
634 origin_decref(split[i].suspect);
635}
636
1732a1fd
JH
637/*
638 * Helper for blame_chunk(). blame_entry e is known to overlap with
639 * the patch hunk; split it and pass blame to the parent.
640 */
cee7f245
JH
641static void blame_overlap(struct scoreboard *sb, struct blame_entry *e,
642 int tlno, int plno, int same,
643 struct origin *parent)
644{
645 struct blame_entry split[3];
646
647 split_overlap(split, e, tlno, plno, same, parent);
54a4c617
JH
648 if (split[1].suspect)
649 split_blame(sb, split, e);
650 decref_split(split);
cee7f245
JH
651}
652
1732a1fd
JH
653/*
654 * Find the line number of the last line the target is suspected for.
655 */
cee7f245
JH
656static int find_last_in_target(struct scoreboard *sb, struct origin *target)
657{
658 struct blame_entry *e;
659 int last_in_target = -1;
660
661 for (e = sb->ent; e; e = e->next) {
171dccd5 662 if (e->guilty || !same_suspect(e->suspect, target))
cee7f245
JH
663 continue;
664 if (last_in_target < e->s_lno + e->num_lines)
665 last_in_target = e->s_lno + e->num_lines;
666 }
667 return last_in_target;
668}
669
1732a1fd
JH
670/*
671 * Process one hunk from the patch between the current suspect for
672 * blame_entry e and its parent. Find and split the overlap, and
673 * pass blame to the overlapping part to the parent.
674 */
cee7f245
JH
675static void blame_chunk(struct scoreboard *sb,
676 int tlno, int plno, int same,
677 struct origin *target, struct origin *parent)
678{
612702e8 679 struct blame_entry *e;
cee7f245 680
612702e8 681 for (e = sb->ent; e; e = e->next) {
171dccd5 682 if (e->guilty || !same_suspect(e->suspect, target))
cee7f245
JH
683 continue;
684 if (same <= e->s_lno)
685 continue;
686 if (tlno < e->s_lno + e->num_lines)
687 blame_overlap(sb, e, tlno, plno, same, parent);
688 }
689}
690
cdcab133
RS
691struct blame_chunk_cb_data {
692 struct scoreboard *sb;
693 struct origin *target;
694 struct origin *parent;
695 long plno;
696 long tlno;
697};
698
699static void blame_chunk_cb(void *data, long same, long p_next, long t_next)
700{
701 struct blame_chunk_cb_data *d = data;
702 blame_chunk(d->sb, d->tlno, d->plno, same, d->target, d->parent);
703 d->plno = p_next;
704 d->tlno = t_next;
705}
706
1732a1fd
JH
707/*
708 * We are looking at the origin 'target' and aiming to pass blame
709 * for the lines it is suspected to its parent. Run diff to find
710 * which lines came from parent and pass blame for them.
711 */
cee7f245
JH
712static int pass_blame_to_parent(struct scoreboard *sb,
713 struct origin *target,
714 struct origin *parent)
715{
cdcab133 716 int last_in_target;
7c4ed8c8 717 mmfile_t file_p, file_o;
cdcab133
RS
718 struct blame_chunk_cb_data d = { sb, target, parent, 0, 0 };
719 xpparam_t xpp;
720 xdemitconf_t xecfg;
cee7f245
JH
721
722 last_in_target = find_last_in_target(sb, target);
723 if (last_in_target < 0)
724 return 1; /* nothing remains for this target */
725
7c4ed8c8
RS
726 fill_origin_blob(parent, &file_p);
727 fill_origin_blob(target, &file_o);
7c4ed8c8 728 num_get_patch++;
cee7f245 729
cdcab133
RS
730 memset(&xpp, 0, sizeof(xpp));
731 xpp.flags = xdl_opts;
732 memset(&xecfg, 0, sizeof(xecfg));
733 xecfg.ctxlen = 0;
734 xdi_diff_hunks(&file_p, &file_o, blame_chunk_cb, &d, &xpp, &xecfg);
1732a1fd 735 /* The rest (i.e. anything after tlno) are the same as the parent */
cdcab133 736 blame_chunk(sb, d.tlno, d.plno, last_in_target, target, parent);
cee7f245 737
cee7f245
JH
738 return 0;
739}
740
1732a1fd
JH
741/*
742 * The lines in blame_entry after splitting blames many times can become
743 * very small and trivial, and at some point it becomes pointless to
744 * blame the parents. E.g. "\t\t}\n\t}\n\n" appears everywhere in any
745 * ordinary C program, and it is not worth to say it was copied from
746 * totally unrelated file in the parent.
747 *
748 * Compute how trivial the lines in the blame_entry are.
749 */
5ff62c30
JH
750static unsigned ent_score(struct scoreboard *sb, struct blame_entry *e)
751{
752 unsigned score;
753 const char *cp, *ep;
754
755 if (e->score)
756 return e->score;
757
612702e8 758 score = 1;
5ff62c30
JH
759 cp = nth_line(sb, e->lno);
760 ep = nth_line(sb, e->lno + e->num_lines);
761 while (cp < ep) {
762 unsigned ch = *((unsigned char *)cp);
763 if (isalnum(ch))
764 score++;
765 cp++;
766 }
767 e->score = score;
768 return score;
769}
770
1732a1fd
JH
771/*
772 * best_so_far[] and this[] are both a split of an existing blame_entry
773 * that passes blame to the parent. Maintain best_so_far the best split
774 * so far, by comparing this and best_so_far and copying this into
775 * bst_so_far as needed.
776 */
5ff62c30 777static void copy_split_if_better(struct scoreboard *sb,
54a4c617
JH
778 struct blame_entry *best_so_far,
779 struct blame_entry *this)
d24bba80 780{
54a4c617
JH
781 int i;
782
d24bba80
JH
783 if (!this[1].suspect)
784 return;
5ff62c30
JH
785 if (best_so_far[1].suspect) {
786 if (ent_score(sb, &this[1]) < ent_score(sb, &best_so_far[1]))
787 return;
788 }
54a4c617
JH
789
790 for (i = 0; i < 3; i++)
791 origin_incref(this[i].suspect);
792 decref_split(best_so_far);
d24bba80
JH
793 memcpy(best_so_far, this, sizeof(struct blame_entry [3]));
794}
795
dd166aa8
JH
796/*
797 * We are looking at a part of the final image represented by
798 * ent (tlno and same are offset by ent->s_lno).
799 * tlno is where we are looking at in the final image.
800 * up to (but not including) same match preimage.
801 * plno is where we are looking at in the preimage.
802 *
803 * <-------------- final image ---------------------->
804 * <------ent------>
805 * ^tlno ^same
806 * <---------preimage----->
807 * ^plno
808 *
809 * All line numbers are 0-based.
810 */
811static void handle_split(struct scoreboard *sb,
812 struct blame_entry *ent,
813 int tlno, int plno, int same,
814 struct origin *parent,
815 struct blame_entry *split)
816{
817 if (ent->num_lines <= tlno)
818 return;
819 if (tlno < same) {
820 struct blame_entry this[3];
821 tlno += ent->s_lno;
822 same += ent->s_lno;
823 split_overlap(this, ent, tlno, plno, same, parent);
824 copy_split_if_better(sb, split, this);
825 decref_split(this);
826 }
827}
828
cdcab133
RS
829struct handle_split_cb_data {
830 struct scoreboard *sb;
831 struct blame_entry *ent;
832 struct origin *parent;
833 struct blame_entry *split;
834 long plno;
835 long tlno;
836};
837
838static void handle_split_cb(void *data, long same, long p_next, long t_next)
839{
840 struct handle_split_cb_data *d = data;
841 handle_split(d->sb, d->ent, d->tlno, d->plno, same, d->parent, d->split);
842 d->plno = p_next;
843 d->tlno = t_next;
844}
845
1732a1fd
JH
846/*
847 * Find the lines from parent that are the same as ent so that
848 * we can pass blames to it. file_p has the blob contents for
849 * the parent.
850 */
d24bba80
JH
851static void find_copy_in_blob(struct scoreboard *sb,
852 struct blame_entry *ent,
853 struct origin *parent,
54a4c617 854 struct blame_entry *split,
d24bba80
JH
855 mmfile_t *file_p)
856{
857 const char *cp;
858 int cnt;
859 mmfile_t file_o;
cdcab133
RS
860 struct handle_split_cb_data d = { sb, ent, parent, split, 0, 0 };
861 xpparam_t xpp;
862 xdemitconf_t xecfg;
d24bba80 863
1732a1fd
JH
864 /*
865 * Prepare mmfile that contains only the lines in ent.
866 */
d24bba80
JH
867 cp = nth_line(sb, ent->lno);
868 file_o.ptr = (char*) cp;
869 cnt = ent->num_lines;
870
871 while (cnt && cp < sb->final_buf + sb->final_buf_size) {
872 if (*cp++ == '\n')
873 cnt--;
874 }
875 file_o.size = cp - file_o.ptr;
876
dd166aa8
JH
877 /*
878 * file_o is a part of final image we are annotating.
879 * file_p partially may match that image.
880 */
cdcab133
RS
881 memset(&xpp, 0, sizeof(xpp));
882 xpp.flags = xdl_opts;
883 memset(&xecfg, 0, sizeof(xecfg));
884 xecfg.ctxlen = 1;
d24bba80 885 memset(split, 0, sizeof(struct blame_entry [3]));
cdcab133 886 xdi_diff_hunks(file_p, &file_o, handle_split_cb, &d, &xpp, &xecfg);
dd166aa8 887 /* remainder, if any, all match the preimage */
cdcab133 888 handle_split(sb, ent, d.tlno, d.plno, ent->num_lines, parent, split);
d24bba80
JH
889}
890
1732a1fd
JH
891/*
892 * See if lines currently target is suspected for can be attributed to
893 * parent.
894 */
d24bba80
JH
895static int find_move_in_parent(struct scoreboard *sb,
896 struct origin *target,
897 struct origin *parent)
898{
650e2f67 899 int last_in_target, made_progress;
46014766 900 struct blame_entry *e, split[3];
d24bba80 901 mmfile_t file_p;
d24bba80
JH
902
903 last_in_target = find_last_in_target(sb, target);
904 if (last_in_target < 0)
905 return 1; /* nothing remains for this target */
906
c2e525d9
JH
907 fill_origin_blob(parent, &file_p);
908 if (!file_p.ptr)
d24bba80 909 return 0;
d24bba80 910
650e2f67
JH
911 made_progress = 1;
912 while (made_progress) {
913 made_progress = 0;
914 for (e = sb->ent; e; e = e->next) {
1617fe98
AG
915 if (e->guilty || !same_suspect(e->suspect, target) ||
916 ent_score(sb, e) < blame_move_score)
650e2f67
JH
917 continue;
918 find_copy_in_blob(sb, e, parent, split, &file_p);
919 if (split[1].suspect &&
920 blame_move_score < ent_score(sb, &split[1])) {
921 split_blame(sb, split, e);
922 made_progress = 1;
923 }
924 decref_split(split);
925 }
d24bba80 926 }
d24bba80
JH
927 return 0;
928}
929
33494784
JH
930struct blame_list {
931 struct blame_entry *ent;
932 struct blame_entry split[3];
933};
934
1732a1fd
JH
935/*
936 * Count the number of entries the target is suspected for,
937 * and prepare a list of entry and the best split.
938 */
33494784
JH
939static struct blame_list *setup_blame_list(struct scoreboard *sb,
940 struct origin *target,
1617fe98 941 int min_score,
33494784
JH
942 int *num_ents_p)
943{
944 struct blame_entry *e;
945 int num_ents, i;
946 struct blame_list *blame_list = NULL;
947
33494784 948 for (e = sb->ent, num_ents = 0; e; e = e->next)
1617fe98
AG
949 if (!e->scanned && !e->guilty &&
950 same_suspect(e->suspect, target) &&
951 min_score < ent_score(sb, e))
33494784
JH
952 num_ents++;
953 if (num_ents) {
954 blame_list = xcalloc(num_ents, sizeof(struct blame_list));
955 for (e = sb->ent, i = 0; e; e = e->next)
1617fe98
AG
956 if (!e->scanned && !e->guilty &&
957 same_suspect(e->suspect, target) &&
958 min_score < ent_score(sb, e))
33494784
JH
959 blame_list[i++].ent = e;
960 }
961 *num_ents_p = num_ents;
962 return blame_list;
963}
964
eb46328e
AG
965/*
966 * Reset the scanned status on all entries.
967 */
968static void reset_scanned_flag(struct scoreboard *sb)
969{
970 struct blame_entry *e;
971 for (e = sb->ent; e; e = e->next)
972 e->scanned = 0;
973}
974
1732a1fd
JH
975/*
976 * For lines target is suspected for, see if we can find code movement
977 * across file boundary from the parent commit. porigin is the path
978 * in the parent we already tried.
979 */
18abd745
JH
980static int find_copy_in_parent(struct scoreboard *sb,
981 struct origin *target,
982 struct commit *parent,
983 struct origin *porigin,
984 int opt)
985{
986 struct diff_options diff_opts;
987 const char *paths[1];
aec8fa1f 988 int i, j;
33494784
JH
989 int retval;
990 struct blame_list *blame_list;
aec8fa1f 991 int num_ents;
18abd745 992
1617fe98 993 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784 994 if (!blame_list)
18abd745
JH
995 return 1; /* nothing remains for this target */
996
997 diff_setup(&diff_opts);
8f67f8ae 998 DIFF_OPT_SET(&diff_opts, RECURSIVE);
18abd745
JH
999 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
1000
18abd745
JH
1001 paths[0] = NULL;
1002 diff_tree_setup_paths(paths, &diff_opts);
1003 if (diff_setup_done(&diff_opts) < 0)
1004 die("diff-setup");
33494784
JH
1005
1006 /* Try "find copies harder" on new path if requested;
1007 * we do not want to use diffcore_rename() actually to
1008 * match things up; find_copies_harder is set only to
1009 * force diff_tree_sha1() to feed all filepairs to diff_queue,
1010 * and this code needs to be after diff_setup_done(), which
1011 * usually makes find-copies-harder imply copy detection.
1012 */
c63777c0
JH
1013 if ((opt & PICKAXE_BLAME_COPY_HARDEST)
1014 || ((opt & PICKAXE_BLAME_COPY_HARDER)
1015 && (!porigin || strcmp(target->path, porigin->path))))
8f67f8ae 1016 DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
33494784 1017
1cfe7733
JH
1018 if (is_null_sha1(target->commit->object.sha1))
1019 do_diff_cache(parent->tree->object.sha1, &diff_opts);
1020 else
1021 diff_tree_sha1(parent->tree->object.sha1,
1022 target->commit->tree->object.sha1,
1023 "", &diff_opts);
18abd745 1024
8f67f8ae 1025 if (!DIFF_OPT_TST(&diff_opts, FIND_COPIES_HARDER))
33494784 1026 diffcore_std(&diff_opts);
18abd745 1027
33494784
JH
1028 retval = 0;
1029 while (1) {
1030 int made_progress = 0;
1031
1032 for (i = 0; i < diff_queued_diff.nr; i++) {
1033 struct diff_filepair *p = diff_queued_diff.queue[i];
1034 struct origin *norigin;
1035 mmfile_t file_p;
33494784
JH
1036 struct blame_entry this[3];
1037
1038 if (!DIFF_FILE_VALID(p->one))
1039 continue; /* does not exist in parent */
5209ac4d
AG
1040 if (S_ISGITLINK(p->one->mode))
1041 continue; /* ignore git links */
33494784
JH
1042 if (porigin && !strcmp(p->one->path, porigin->path))
1043 /* find_move already dealt with this path */
1044 continue;
1045
1046 norigin = get_origin(sb, parent, p->one->path);
1047 hashcpy(norigin->blob_sha1, p->one->sha1);
c2e525d9 1048 fill_origin_blob(norigin, &file_p);
33494784
JH
1049 if (!file_p.ptr)
1050 continue;
1051
1052 for (j = 0; j < num_ents; j++) {
1053 find_copy_in_blob(sb, blame_list[j].ent,
1054 norigin, this, &file_p);
1055 copy_split_if_better(sb, blame_list[j].split,
1056 this);
1057 decref_split(this);
1058 }
33494784
JH
1059 origin_decref(norigin);
1060 }
18abd745 1061
aec8fa1f 1062 for (j = 0; j < num_ents; j++) {
33494784
JH
1063 struct blame_entry *split = blame_list[j].split;
1064 if (split[1].suspect &&
1065 blame_copy_score < ent_score(sb, &split[1])) {
1066 split_blame(sb, split, blame_list[j].ent);
1067 made_progress = 1;
1068 }
eb46328e
AG
1069 else
1070 blame_list[j].ent->scanned = 1;
33494784 1071 decref_split(split);
18abd745 1072 }
33494784 1073 free(blame_list);
aec8fa1f 1074
33494784
JH
1075 if (!made_progress)
1076 break;
1617fe98 1077 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784
JH
1078 if (!blame_list) {
1079 retval = 1;
1080 break;
1081 }
18abd745 1082 }
eb46328e 1083 reset_scanned_flag(sb);
33494784 1084 diff_flush(&diff_opts);
03b69c76 1085 diff_tree_release_paths(&diff_opts);
33494784 1086 return retval;
18abd745
JH
1087}
1088
1732a1fd
JH
1089/*
1090 * The blobs of origin and porigin exactly match, so everything
c2e525d9
JH
1091 * origin is suspected for can be blamed on the parent.
1092 */
1093static void pass_whole_blame(struct scoreboard *sb,
1094 struct origin *origin, struct origin *porigin)
1095{
1096 struct blame_entry *e;
1097
1098 if (!porigin->file.ptr && origin->file.ptr) {
1099 /* Steal its file */
1100 porigin->file = origin->file;
1101 origin->file.ptr = NULL;
1102 }
1103 for (e = sb->ent; e; e = e->next) {
171dccd5 1104 if (!same_suspect(e->suspect, origin))
c2e525d9
JH
1105 continue;
1106 origin_incref(porigin);
1107 origin_decref(e->suspect);
1108 e->suspect = porigin;
1109 }
1110}
1111
69264f46
JH
1112/*
1113 * We pass blame from the current commit to its parents. We keep saying
1114 * "parent" (and "porigin"), but what we mean is to find scapegoat to
1115 * exonerate ourselves.
1116 */
85af7929 1117static struct commit_list *first_scapegoat(struct rev_info *revs, struct commit *commit)
69264f46 1118{
85af7929
JH
1119 if (!reverse)
1120 return commit->parents;
1121 return lookup_decoration(&revs->children, &commit->object);
69264f46
JH
1122}
1123
85af7929 1124static int num_scapegoats(struct rev_info *revs, struct commit *commit)
69264f46
JH
1125{
1126 int cnt;
85af7929 1127 struct commit_list *l = first_scapegoat(revs, commit);
69264f46
JH
1128 for (cnt = 0; l; l = l->next)
1129 cnt++;
1130 return cnt;
1131}
1132
1133#define MAXSG 16
cee7f245 1134
d24bba80 1135static void pass_blame(struct scoreboard *sb, struct origin *origin, int opt)
cee7f245 1136{
85af7929 1137 struct rev_info *revs = sb->revs;
69264f46 1138 int i, pass, num_sg;
cee7f245 1139 struct commit *commit = origin->commit;
69264f46
JH
1140 struct commit_list *sg;
1141 struct origin *sg_buf[MAXSG];
1142 struct origin *porigin, **sg_origin = sg_buf;
1143
85af7929 1144 num_sg = num_scapegoats(revs, commit);
69264f46
JH
1145 if (!num_sg)
1146 goto finish;
1147 else if (num_sg < ARRAY_SIZE(sg_buf))
1148 memset(sg_buf, 0, sizeof(sg_buf));
1149 else
1150 sg_origin = xcalloc(num_sg, sizeof(*sg_origin));
cee7f245 1151
69264f46
JH
1152 /*
1153 * The first pass looks for unrenamed path to optimize for
f69e743d
JH
1154 * common cases, then we look for renames in the second pass.
1155 */
1156 for (pass = 0; pass < 2; pass++) {
1157 struct origin *(*find)(struct scoreboard *,
1158 struct commit *, struct origin *);
1159 find = pass ? find_rename : find_origin;
1160
85af7929 1161 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1162 i < num_sg && sg;
1163 sg = sg->next, i++) {
1164 struct commit *p = sg->item;
0421d9f8 1165 int j, same;
f69e743d 1166
69264f46 1167 if (sg_origin[i])
f69e743d
JH
1168 continue;
1169 if (parse_commit(p))
1170 continue;
0d981c67 1171 porigin = find(sb, p, origin);
f69e743d
JH
1172 if (!porigin)
1173 continue;
1174 if (!hashcmp(porigin->blob_sha1, origin->blob_sha1)) {
c2e525d9 1175 pass_whole_blame(sb, origin, porigin);
f69e743d
JH
1176 origin_decref(porigin);
1177 goto finish;
1178 }
0421d9f8 1179 for (j = same = 0; j < i; j++)
69264f46
JH
1180 if (sg_origin[j] &&
1181 !hashcmp(sg_origin[j]->blob_sha1,
0421d9f8
JH
1182 porigin->blob_sha1)) {
1183 same = 1;
1184 break;
1185 }
1186 if (!same)
69264f46 1187 sg_origin[i] = porigin;
0421d9f8
JH
1188 else
1189 origin_decref(porigin);
cee7f245 1190 }
cee7f245
JH
1191 }
1192
c2e525d9 1193 num_commits++;
85af7929 1194 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1195 i < num_sg && sg;
1196 sg = sg->next, i++) {
1197 struct origin *porigin = sg_origin[i];
cee7f245
JH
1198 if (!porigin)
1199 continue;
1200 if (pass_blame_to_parent(sb, origin, porigin))
54a4c617 1201 goto finish;
cee7f245 1202 }
d24bba80
JH
1203
1204 /*
1732a1fd 1205 * Optionally find moves in parents' files.
d24bba80
JH
1206 */
1207 if (opt & PICKAXE_BLAME_MOVE)
85af7929 1208 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1209 i < num_sg && sg;
1210 sg = sg->next, i++) {
1211 struct origin *porigin = sg_origin[i];
d24bba80
JH
1212 if (!porigin)
1213 continue;
1214 if (find_move_in_parent(sb, origin, porigin))
54a4c617 1215 goto finish;
d24bba80
JH
1216 }
1217
18abd745 1218 /*
1732a1fd 1219 * Optionally find copies from parents' files.
18abd745
JH
1220 */
1221 if (opt & PICKAXE_BLAME_COPY)
85af7929 1222 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1223 i < num_sg && sg;
1224 sg = sg->next, i++) {
1225 struct origin *porigin = sg_origin[i];
1226 if (find_copy_in_parent(sb, origin, sg->item,
18abd745 1227 porigin, opt))
54a4c617 1228 goto finish;
18abd745 1229 }
54a4c617
JH
1230
1231 finish:
69264f46
JH
1232 for (i = 0; i < num_sg; i++) {
1233 if (sg_origin[i]) {
1234 drop_origin_blob(sg_origin[i]);
1235 origin_decref(sg_origin[i]);
7c3c7962
JH
1236 }
1237 }
1238 drop_origin_blob(origin);
69264f46
JH
1239 if (sg_buf != sg_origin)
1240 free(sg_origin);
cee7f245
JH
1241}
1242
1732a1fd
JH
1243/*
1244 * Information on commits, used for output.
1245 */
cee7f245
JH
1246struct commit_info
1247{
3a55602e
SP
1248 const char *author;
1249 const char *author_mail;
cee7f245 1250 unsigned long author_time;
3a55602e 1251 const char *author_tz;
cee7f245
JH
1252
1253 /* filled only when asked for details */
3a55602e
SP
1254 const char *committer;
1255 const char *committer_mail;
cee7f245 1256 unsigned long committer_time;
3a55602e 1257 const char *committer_tz;
cee7f245 1258
3a55602e 1259 const char *summary;
cee7f245
JH
1260};
1261
1732a1fd
JH
1262/*
1263 * Parse author/committer line in the commit object buffer
1264 */
cee7f245 1265static void get_ac_line(const char *inbuf, const char *what,
3a55602e
SP
1266 int bufsz, char *person, const char **mail,
1267 unsigned long *time, const char **tz)
cee7f245 1268{
f9567384
JH
1269 int len, tzlen, maillen;
1270 char *tmp, *endp, *timepos;
cee7f245
JH
1271
1272 tmp = strstr(inbuf, what);
1273 if (!tmp)
1274 goto error_out;
1275 tmp += strlen(what);
1276 endp = strchr(tmp, '\n');
1277 if (!endp)
1278 len = strlen(tmp);
1279 else
1280 len = endp - tmp;
1281 if (bufsz <= len) {
1282 error_out:
1283 /* Ugh */
3a55602e 1284 *mail = *tz = "(unknown)";
cee7f245
JH
1285 *time = 0;
1286 return;
1287 }
1288 memcpy(person, tmp, len);
1289
1290 tmp = person;
1291 tmp += len;
1292 *tmp = 0;
1293 while (*tmp != ' ')
1294 tmp--;
1295 *tz = tmp+1;
f9567384 1296 tzlen = (person+len)-(tmp+1);
cee7f245
JH
1297
1298 *tmp = 0;
1299 while (*tmp != ' ')
1300 tmp--;
1301 *time = strtoul(tmp, NULL, 10);
f9567384 1302 timepos = tmp;
cee7f245
JH
1303
1304 *tmp = 0;
1305 while (*tmp != ' ')
1306 tmp--;
1307 *mail = tmp + 1;
1308 *tmp = 0;
f9567384
JH
1309 maillen = timepos - tmp;
1310
1311 if (!mailmap.nr)
1312 return;
1313
1314 /*
1315 * mailmap expansion may make the name longer.
1316 * make room by pushing stuff down.
1317 */
1318 tmp = person + bufsz - (tzlen + 1);
1319 memmove(tmp, *tz, tzlen);
1320 tmp[tzlen] = 0;
1321 *tz = tmp;
1322
1323 tmp = tmp - (maillen + 1);
1324 memmove(tmp, *mail, maillen);
1325 tmp[maillen] = 0;
1326 *mail = tmp;
1327
1328 /*
1329 * Now, convert e-mail using mailmap
1330 */
1331 map_email(&mailmap, tmp + 1, person, tmp-person-1);
cee7f245
JH
1332}
1333
1334static void get_commit_info(struct commit *commit,
1335 struct commit_info *ret,
1336 int detailed)
1337{
1338 int len;
69cd8f63 1339 char *tmp, *endp, *reencoded, *message;
cee7f245
JH
1340 static char author_buf[1024];
1341 static char committer_buf[1024];
1342 static char summary_buf[1024];
1343
1732a1fd
JH
1344 /*
1345 * We've operated without save_commit_buffer, so
612702e8
JH
1346 * we now need to populate them for output.
1347 */
1348 if (!commit->buffer) {
21666f1a 1349 enum object_type type;
612702e8
JH
1350 unsigned long size;
1351 commit->buffer =
21666f1a 1352 read_sha1_file(commit->object.sha1, &type, &size);
ab43e495
JH
1353 if (!commit->buffer)
1354 die("Cannot read commit %s",
1355 sha1_to_hex(commit->object.sha1));
612702e8 1356 }
69cd8f63
AG
1357 reencoded = reencode_commit_message(commit, NULL);
1358 message = reencoded ? reencoded : commit->buffer;
cee7f245 1359 ret->author = author_buf;
69cd8f63 1360 get_ac_line(message, "\nauthor ",
cee7f245
JH
1361 sizeof(author_buf), author_buf, &ret->author_mail,
1362 &ret->author_time, &ret->author_tz);
1363
69cd8f63
AG
1364 if (!detailed) {
1365 free(reencoded);
cee7f245 1366 return;
69cd8f63 1367 }
cee7f245
JH
1368
1369 ret->committer = committer_buf;
69cd8f63 1370 get_ac_line(message, "\ncommitter ",
cee7f245
JH
1371 sizeof(committer_buf), committer_buf, &ret->committer_mail,
1372 &ret->committer_time, &ret->committer_tz);
1373
1374 ret->summary = summary_buf;
69cd8f63 1375 tmp = strstr(message, "\n\n");
cee7f245
JH
1376 if (!tmp) {
1377 error_out:
1378 sprintf(summary_buf, "(%s)", sha1_to_hex(commit->object.sha1));
69cd8f63 1379 free(reencoded);
cee7f245
JH
1380 return;
1381 }
1382 tmp += 2;
1383 endp = strchr(tmp, '\n');
1384 if (!endp)
1cfe7733 1385 endp = tmp + strlen(tmp);
cee7f245 1386 len = endp - tmp;
1cfe7733 1387 if (len >= sizeof(summary_buf) || len == 0)
cee7f245
JH
1388 goto error_out;
1389 memcpy(summary_buf, tmp, len);
1390 summary_buf[len] = 0;
69cd8f63 1391 free(reencoded);
cee7f245
JH
1392}
1393
1732a1fd
JH
1394/*
1395 * To allow LF and other nonportable characters in pathnames,
1396 * they are c-style quoted as needed.
1397 */
46e5e69d
JH
1398static void write_filename_info(const char *path)
1399{
1400 printf("filename ");
663af342 1401 write_name_quoted(path, stdout, '\n');
46e5e69d
JH
1402}
1403
1732a1fd
JH
1404/*
1405 * The blame_entry is found to be guilty for the range. Mark it
1406 * as such, and show it in incremental output.
1407 */
717d1462
LT
1408static void found_guilty_entry(struct blame_entry *ent)
1409{
1410 if (ent->guilty)
1411 return;
1412 ent->guilty = 1;
1413 if (incremental) {
1414 struct origin *suspect = ent->suspect;
1415
1416 printf("%s %d %d %d\n",
1417 sha1_to_hex(suspect->commit->object.sha1),
1418 ent->s_lno + 1, ent->lno + 1, ent->num_lines);
1419 if (!(suspect->commit->object.flags & METAINFO_SHOWN)) {
1420 struct commit_info ci;
1421 suspect->commit->object.flags |= METAINFO_SHOWN;
1422 get_commit_info(suspect->commit, &ci, 1);
1423 printf("author %s\n", ci.author);
1424 printf("author-mail %s\n", ci.author_mail);
1425 printf("author-time %lu\n", ci.author_time);
1426 printf("author-tz %s\n", ci.author_tz);
1427 printf("committer %s\n", ci.committer);
1428 printf("committer-mail %s\n", ci.committer_mail);
1429 printf("committer-time %lu\n", ci.committer_time);
1430 printf("committer-tz %s\n", ci.committer_tz);
1431 printf("summary %s\n", ci.summary);
1432 if (suspect->commit->object.flags & UNINTERESTING)
1433 printf("boundary\n");
1434 }
46e5e69d 1435 write_filename_info(suspect->path);
06f59e9f 1436 maybe_flush_or_die(stdout, "stdout");
717d1462
LT
1437 }
1438}
1439
1732a1fd
JH
1440/*
1441 * The main loop -- while the scoreboard has lines whose true origin
3dff5379 1442 * is still unknown, pick one blame_entry, and allow its current
1732a1fd
JH
1443 * suspect to pass blames to its parents.
1444 */
85af7929 1445static void assign_blame(struct scoreboard *sb, int opt)
717d1462 1446{
85af7929
JH
1447 struct rev_info *revs = sb->revs;
1448
717d1462
LT
1449 while (1) {
1450 struct blame_entry *ent;
1451 struct commit *commit;
1452 struct origin *suspect = NULL;
1453
1454 /* find one suspect to break down */
1455 for (ent = sb->ent; !suspect && ent; ent = ent->next)
1456 if (!ent->guilty)
1457 suspect = ent->suspect;
1458 if (!suspect)
1459 return; /* all done */
1460
1732a1fd
JH
1461 /*
1462 * We will use this suspect later in the loop,
1463 * so hold onto it in the meantime.
1464 */
717d1462
LT
1465 origin_incref(suspect);
1466 commit = suspect->commit;
1467 if (!commit->object.parsed)
1468 parse_commit(commit);
85af7929
JH
1469 if (reverse ||
1470 (!(commit->object.flags & UNINTERESTING) &&
1471 !(revs->max_age != -1 && commit->date < revs->max_age)))
717d1462
LT
1472 pass_blame(sb, suspect, opt);
1473 else {
1474 commit->object.flags |= UNINTERESTING;
1475 if (commit->object.parsed)
1476 mark_parents_uninteresting(commit);
1477 }
1478 /* treat root commit as boundary */
1479 if (!commit->parents && !show_root)
1480 commit->object.flags |= UNINTERESTING;
1481
1482 /* Take responsibility for the remaining entries */
1483 for (ent = sb->ent; ent; ent = ent->next)
171dccd5 1484 if (same_suspect(ent->suspect, suspect))
717d1462
LT
1485 found_guilty_entry(ent);
1486 origin_decref(suspect);
1487
1488 if (DEBUG) /* sanity */
1489 sanity_check_refcnt(sb);
1490 }
1491}
1492
1493static const char *format_time(unsigned long time, const char *tz_str,
1494 int show_raw_time)
1495{
1496 static char time_buf[128];
1497 time_t t = time;
1498 int minutes, tz;
1499 struct tm *tm;
1500
1501 if (show_raw_time) {
1502 sprintf(time_buf, "%lu %s", time, tz_str);
1503 return time_buf;
1504 }
1505
1506 tz = atoi(tz_str);
1507 minutes = tz < 0 ? -tz : tz;
1508 minutes = (minutes / 100)*60 + (minutes % 100);
1509 minutes = tz < 0 ? -minutes : minutes;
1510 t = time + minutes * 60;
1511 tm = gmtime(&t);
1512
1513 strftime(time_buf, sizeof(time_buf), "%Y-%m-%d %H:%M:%S ", tm);
1514 strcat(time_buf, tz_str);
1515 return time_buf;
1516}
1517
cee7f245
JH
1518#define OUTPUT_ANNOTATE_COMPAT 001
1519#define OUTPUT_LONG_OBJECT_NAME 002
1520#define OUTPUT_RAW_TIMESTAMP 004
1521#define OUTPUT_PORCELAIN 010
1522#define OUTPUT_SHOW_NAME 020
1523#define OUTPUT_SHOW_NUMBER 040
5ff62c30 1524#define OUTPUT_SHOW_SCORE 0100
093dc5be 1525#define OUTPUT_NO_AUTHOR 0200
cee7f245
JH
1526
1527static void emit_porcelain(struct scoreboard *sb, struct blame_entry *ent)
1528{
1529 int cnt;
1530 const char *cp;
1531 struct origin *suspect = ent->suspect;
1532 char hex[41];
1533
1534 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1535 printf("%s%c%d %d %d\n",
1536 hex,
1537 ent->guilty ? ' ' : '*', // purely for debugging
1538 ent->s_lno + 1,
1539 ent->lno + 1,
1540 ent->num_lines);
1541 if (!(suspect->commit->object.flags & METAINFO_SHOWN)) {
1542 struct commit_info ci;
1543 suspect->commit->object.flags |= METAINFO_SHOWN;
1544 get_commit_info(suspect->commit, &ci, 1);
1545 printf("author %s\n", ci.author);
1546 printf("author-mail %s\n", ci.author_mail);
1547 printf("author-time %lu\n", ci.author_time);
1548 printf("author-tz %s\n", ci.author_tz);
1549 printf("committer %s\n", ci.committer);
1550 printf("committer-mail %s\n", ci.committer_mail);
1551 printf("committer-time %lu\n", ci.committer_time);
1552 printf("committer-tz %s\n", ci.committer_tz);
46e5e69d 1553 write_filename_info(suspect->path);
cee7f245 1554 printf("summary %s\n", ci.summary);
b11121d9
JH
1555 if (suspect->commit->object.flags & UNINTERESTING)
1556 printf("boundary\n");
cee7f245
JH
1557 }
1558 else if (suspect->commit->object.flags & MORE_THAN_ONE_PATH)
46e5e69d 1559 write_filename_info(suspect->path);
cee7f245
JH
1560
1561 cp = nth_line(sb, ent->lno);
1562 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1563 char ch;
1564 if (cnt)
1565 printf("%s %d %d\n", hex,
1566 ent->s_lno + 1 + cnt,
1567 ent->lno + 1 + cnt);
1568 putchar('\t');
1569 do {
1570 ch = *cp++;
1571 putchar(ch);
1572 } while (ch != '\n' &&
1573 cp < sb->final_buf + sb->final_buf_size);
1574 }
1575}
1576
1577static void emit_other(struct scoreboard *sb, struct blame_entry *ent, int opt)
1578{
1579 int cnt;
1580 const char *cp;
1581 struct origin *suspect = ent->suspect;
1582 struct commit_info ci;
1583 char hex[41];
1584 int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP);
1585
1586 get_commit_info(suspect->commit, &ci, 1);
1587 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1588
1589 cp = nth_line(sb, ent->lno);
1590 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1591 char ch;
b11121d9
JH
1592 int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? 40 : 8;
1593
1594 if (suspect->commit->object.flags & UNINTERESTING) {
e68989a7
JH
1595 if (blank_boundary)
1596 memset(hex, ' ', length);
7ceacdff 1597 else if (!(opt & OUTPUT_ANNOTATE_COMPAT)) {
4c10a5ca
JH
1598 length--;
1599 putchar('^');
1600 }
b11121d9 1601 }
cee7f245 1602
b11121d9 1603 printf("%.*s", length, hex);
cee7f245
JH
1604 if (opt & OUTPUT_ANNOTATE_COMPAT)
1605 printf("\t(%10s\t%10s\t%d)", ci.author,
1606 format_time(ci.author_time, ci.author_tz,
1607 show_raw_time),
1608 ent->lno + 1 + cnt);
1609 else {
5ff62c30 1610 if (opt & OUTPUT_SHOW_SCORE)
54a4c617
JH
1611 printf(" %*d %02d",
1612 max_score_digits, ent->score,
1613 ent->suspect->refcnt);
cee7f245
JH
1614 if (opt & OUTPUT_SHOW_NAME)
1615 printf(" %-*.*s", longest_file, longest_file,
1616 suspect->path);
1617 if (opt & OUTPUT_SHOW_NUMBER)
1618 printf(" %*d", max_orig_digits,
1619 ent->s_lno + 1 + cnt);
093dc5be
JH
1620
1621 if (!(opt & OUTPUT_NO_AUTHOR))
1622 printf(" (%-*.*s %10s",
1623 longest_author, longest_author,
1624 ci.author,
1625 format_time(ci.author_time,
1626 ci.author_tz,
1627 show_raw_time));
1628 printf(" %*d) ",
cee7f245
JH
1629 max_digits, ent->lno + 1 + cnt);
1630 }
1631 do {
1632 ch = *cp++;
1633 putchar(ch);
1634 } while (ch != '\n' &&
1635 cp < sb->final_buf + sb->final_buf_size);
1636 }
1637}
1638
1639static void output(struct scoreboard *sb, int option)
1640{
1641 struct blame_entry *ent;
1642
1643 if (option & OUTPUT_PORCELAIN) {
1644 for (ent = sb->ent; ent; ent = ent->next) {
1645 struct blame_entry *oth;
1646 struct origin *suspect = ent->suspect;
1647 struct commit *commit = suspect->commit;
1648 if (commit->object.flags & MORE_THAN_ONE_PATH)
1649 continue;
1650 for (oth = ent->next; oth; oth = oth->next) {
1651 if ((oth->suspect->commit != commit) ||
1652 !strcmp(oth->suspect->path, suspect->path))
1653 continue;
1654 commit->object.flags |= MORE_THAN_ONE_PATH;
1655 break;
1656 }
1657 }
1658 }
1659
1660 for (ent = sb->ent; ent; ent = ent->next) {
1661 if (option & OUTPUT_PORCELAIN)
1662 emit_porcelain(sb, ent);
5ff62c30 1663 else {
cee7f245 1664 emit_other(sb, ent, option);
5ff62c30 1665 }
cee7f245
JH
1666 }
1667}
1668
1732a1fd
JH
1669/*
1670 * To allow quick access to the contents of nth line in the
1671 * final image, prepare an index in the scoreboard.
1672 */
cee7f245
JH
1673static int prepare_lines(struct scoreboard *sb)
1674{
1675 const char *buf = sb->final_buf;
1676 unsigned long len = sb->final_buf_size;
1677 int num = 0, incomplete = 0, bol = 1;
1678
1679 if (len && buf[len-1] != '\n')
1680 incomplete++; /* incomplete line at the end */
1681 while (len--) {
1682 if (bol) {
1683 sb->lineno = xrealloc(sb->lineno,
1684 sizeof(int* ) * (num + 1));
1685 sb->lineno[num] = buf - sb->final_buf;
1686 bol = 0;
1687 }
1688 if (*buf++ == '\n') {
1689 num++;
1690 bol = 1;
1691 }
1692 }
1ca6ca87
JH
1693 sb->lineno = xrealloc(sb->lineno,
1694 sizeof(int* ) * (num + incomplete + 1));
1695 sb->lineno[num + incomplete] = buf - sb->final_buf;
cee7f245
JH
1696 sb->num_lines = num + incomplete;
1697 return sb->num_lines;
1698}
1699
1732a1fd
JH
1700/*
1701 * Add phony grafts for use with -S; this is primarily to
34baebce 1702 * support git's cvsserver that wants to give a linear history
1732a1fd
JH
1703 * to its clients.
1704 */
cee7f245
JH
1705static int read_ancestry(const char *graft_file)
1706{
1707 FILE *fp = fopen(graft_file, "r");
1708 char buf[1024];
1709 if (!fp)
1710 return -1;
1711 while (fgets(buf, sizeof(buf), fp)) {
1712 /* The format is just "Commit Parent1 Parent2 ...\n" */
1713 int len = strlen(buf);
1714 struct commit_graft *graft = read_graft_line(buf, len);
8eaf7986
JH
1715 if (graft)
1716 register_commit_graft(graft, 0);
cee7f245
JH
1717 }
1718 fclose(fp);
1719 return 0;
1720}
1721
1732a1fd
JH
1722/*
1723 * How many columns do we need to show line numbers in decimal?
1724 */
cee7f245
JH
1725static int lineno_width(int lines)
1726{
f2915045 1727 int i, width;
cee7f245 1728
f2915045
JH
1729 for (width = 1, i = 10; i <= lines + 1; width++)
1730 i *= 10;
1731 return width;
cee7f245
JH
1732}
1733
1732a1fd
JH
1734/*
1735 * How many columns do we need to show line numbers, authors,
1736 * and filenames?
1737 */
cee7f245
JH
1738static void find_alignment(struct scoreboard *sb, int *option)
1739{
1740 int longest_src_lines = 0;
1741 int longest_dst_lines = 0;
5ff62c30 1742 unsigned largest_score = 0;
cee7f245
JH
1743 struct blame_entry *e;
1744
1745 for (e = sb->ent; e; e = e->next) {
1746 struct origin *suspect = e->suspect;
1747 struct commit_info ci;
1748 int num;
1749
ab3bb800
JH
1750 if (strcmp(suspect->path, sb->path))
1751 *option |= OUTPUT_SHOW_NAME;
1752 num = strlen(suspect->path);
1753 if (longest_file < num)
1754 longest_file = num;
cee7f245
JH
1755 if (!(suspect->commit->object.flags & METAINFO_SHOWN)) {
1756 suspect->commit->object.flags |= METAINFO_SHOWN;
1757 get_commit_info(suspect->commit, &ci, 1);
cee7f245
JH
1758 num = strlen(ci.author);
1759 if (longest_author < num)
1760 longest_author = num;
1761 }
1762 num = e->s_lno + e->num_lines;
1763 if (longest_src_lines < num)
1764 longest_src_lines = num;
1765 num = e->lno + e->num_lines;
1766 if (longest_dst_lines < num)
1767 longest_dst_lines = num;
5ff62c30
JH
1768 if (largest_score < ent_score(sb, e))
1769 largest_score = ent_score(sb, e);
cee7f245
JH
1770 }
1771 max_orig_digits = lineno_width(longest_src_lines);
1772 max_digits = lineno_width(longest_dst_lines);
5ff62c30 1773 max_score_digits = lineno_width(largest_score);
cee7f245
JH
1774}
1775
1732a1fd
JH
1776/*
1777 * For debugging -- origin is refcounted, and this asserts that
1778 * we do not underflow.
1779 */
54a4c617
JH
1780static void sanity_check_refcnt(struct scoreboard *sb)
1781{
1782 int baa = 0;
1783 struct blame_entry *ent;
1784
1785 for (ent = sb->ent; ent; ent = ent->next) {
ae86ad65 1786 /* Nobody should have zero or negative refcnt */
854b97f6
JH
1787 if (ent->suspect->refcnt <= 0) {
1788 fprintf(stderr, "%s in %s has negative refcnt %d\n",
1789 ent->suspect->path,
1790 sha1_to_hex(ent->suspect->commit->object.sha1),
1791 ent->suspect->refcnt);
ae86ad65 1792 baa = 1;
854b97f6 1793 }
ae86ad65
JH
1794 }
1795 for (ent = sb->ent; ent; ent = ent->next) {
1796 /* Mark the ones that haven't been checked */
54a4c617
JH
1797 if (0 < ent->suspect->refcnt)
1798 ent->suspect->refcnt = -ent->suspect->refcnt;
54a4c617
JH
1799 }
1800 for (ent = sb->ent; ent; ent = ent->next) {
1732a1fd
JH
1801 /*
1802 * ... then pick each and see if they have the the
1803 * correct refcnt.
54a4c617
JH
1804 */
1805 int found;
1806 struct blame_entry *e;
1807 struct origin *suspect = ent->suspect;
1808
1809 if (0 < suspect->refcnt)
1810 continue;
ae86ad65 1811 suspect->refcnt = -suspect->refcnt; /* Unmark */
54a4c617
JH
1812 for (found = 0, e = sb->ent; e; e = e->next) {
1813 if (e->suspect != suspect)
1814 continue;
1815 found++;
1816 }
854b97f6
JH
1817 if (suspect->refcnt != found) {
1818 fprintf(stderr, "%s in %s has refcnt %d, not %d\n",
1819 ent->suspect->path,
1820 sha1_to_hex(ent->suspect->commit->object.sha1),
1821 ent->suspect->refcnt, found);
1822 baa = 2;
1823 }
54a4c617
JH
1824 }
1825 if (baa) {
1826 int opt = 0160;
1827 find_alignment(sb, &opt);
1828 output(sb, opt);
854b97f6 1829 die("Baa %d!", baa);
54a4c617
JH
1830 }
1831}
1832
1732a1fd
JH
1833/*
1834 * Used for the command line parsing; check if the path exists
1835 * in the working tree.
1836 */
c455c87c 1837static int has_string_in_work_tree(const char *path)
cee7f245
JH
1838{
1839 struct stat st;
1840 return !lstat(path, &st);
1841}
1842
4a0fc95f
JH
1843static unsigned parse_score(const char *arg)
1844{
1845 char *end;
1846 unsigned long score = strtoul(arg, &end, 10);
1847 if (*end)
1848 return 0;
1849 return score;
1850}
1851
20239bae
JK
1852static const char *add_prefix(const char *prefix, const char *path)
1853{
097971f5 1854 return prefix_path(prefix, prefix ? strlen(prefix) : 0, path);
20239bae
JK
1855}
1856
1732a1fd
JH
1857/*
1858 * Parsing of (comma separated) one item in the -L option
1859 */
931233bc
JH
1860static const char *parse_loc(const char *spec,
1861 struct scoreboard *sb, long lno,
1862 long begin, long *ret)
1863{
1864 char *term;
1865 const char *line;
1866 long num;
1867 int reg_error;
1868 regex_t regexp;
1869 regmatch_t match[1];
1870
7bd9641d
JH
1871 /* Allow "-L <something>,+20" to mean starting at <something>
1872 * for 20 lines, or "-L <something>,-5" for 5 lines ending at
1873 * <something>.
1874 */
1875 if (1 < begin && (spec[0] == '+' || spec[0] == '-')) {
1876 num = strtol(spec + 1, &term, 10);
1877 if (term != spec + 1) {
1878 if (spec[0] == '-')
1879 num = 0 - num;
1880 if (0 < num)
1881 *ret = begin + num - 2;
1882 else if (!num)
1883 *ret = begin;
1884 else
1885 *ret = begin + num;
1886 return term;
1887 }
1888 return spec;
1889 }
931233bc
JH
1890 num = strtol(spec, &term, 10);
1891 if (term != spec) {
1892 *ret = num;
1893 return term;
1894 }
1895 if (spec[0] != '/')
1896 return spec;
1897
1898 /* it could be a regexp of form /.../ */
1899 for (term = (char*) spec + 1; *term && *term != '/'; term++) {
1900 if (*term == '\\')
1901 term++;
1902 }
1903 if (*term != '/')
1904 return spec;
1905
1906 /* try [spec+1 .. term-1] as regexp */
1907 *term = 0;
1908 begin--; /* input is in human terms */
1909 line = nth_line(sb, begin);
1910
1911 if (!(reg_error = regcomp(&regexp, spec + 1, REG_NEWLINE)) &&
1912 !(reg_error = regexec(&regexp, line, 1, match, 0))) {
1913 const char *cp = line + match[0].rm_so;
1914 const char *nline;
1915
1916 while (begin++ < lno) {
1917 nline = nth_line(sb, begin);
1918 if (line <= cp && cp < nline)
1919 break;
1920 line = nline;
1921 }
1922 *ret = begin;
1923 regfree(&regexp);
1924 *term++ = '/';
1925 return term;
1926 }
1927 else {
1928 char errbuf[1024];
1929 regerror(reg_error, &regexp, errbuf, 1024);
1930 die("-L parameter '%s': %s", spec + 1, errbuf);
1931 }
1932}
1933
1732a1fd
JH
1934/*
1935 * Parsing of -L option
1936 */
931233bc
JH
1937static void prepare_blame_range(struct scoreboard *sb,
1938 const char *bottomtop,
1939 long lno,
1940 long *bottom, long *top)
1941{
1942 const char *term;
1943
1944 term = parse_loc(bottomtop, sb, lno, 1, bottom);
1945 if (*term == ',') {
1946 term = parse_loc(term + 1, sb, lno, *bottom + 1, top);
1947 if (*term)
acca687f 1948 usage(blame_usage);
931233bc
JH
1949 }
1950 if (*term)
acca687f 1951 usage(blame_usage);
931233bc
JH
1952}
1953
ef90d6d4 1954static int git_blame_config(const char *var, const char *value, void *cb)
4c10a5ca
JH
1955{
1956 if (!strcmp(var, "blame.showroot")) {
1957 show_root = git_config_bool(var, value);
1958 return 0;
1959 }
1960 if (!strcmp(var, "blame.blankboundary")) {
1961 blank_boundary = git_config_bool(var, value);
1962 return 0;
1963 }
ef90d6d4 1964 return git_default_config(var, value, cb);
4c10a5ca
JH
1965}
1966
f6c07d7d
JH
1967/*
1968 * Prepare a dummy commit that represents the work tree (or staged) item.
1969 * Note that annotating work tree item never works in the reverse.
1970 */
1cfe7733
JH
1971static struct commit *fake_working_tree_commit(const char *path, const char *contents_from)
1972{
1973 struct commit *commit;
1974 struct origin *origin;
1975 unsigned char head_sha1[20];
f285a2d7 1976 struct strbuf buf = STRBUF_INIT;
1cfe7733 1977 const char *ident;
1cfe7733 1978 time_t now;
1cfe7733
JH
1979 int size, len;
1980 struct cache_entry *ce;
1981 unsigned mode;
1982
1983 if (get_sha1("HEAD", head_sha1))
1984 die("No such ref: HEAD");
1985
1986 time(&now);
1987 commit = xcalloc(1, sizeof(*commit));
1988 commit->parents = xcalloc(1, sizeof(*commit->parents));
1989 commit->parents->item = lookup_commit_reference(head_sha1);
1990 commit->object.parsed = 1;
1991 commit->date = now;
1992 commit->object.type = OBJ_COMMIT;
1993
1994 origin = make_origin(commit, path);
1995
1996 if (!contents_from || strcmp("-", contents_from)) {
1997 struct stat st;
1998 const char *read_from;
1999
2000 if (contents_from) {
2001 if (stat(contents_from, &st) < 0)
2002 die("Cannot stat %s", contents_from);
2003 read_from = contents_from;
2004 }
2005 else {
2006 if (lstat(path, &st) < 0)
2007 die("Cannot lstat %s", path);
2008 read_from = path;
2009 }
1cfe7733
JH
2010 mode = canon_mode(st.st_mode);
2011 switch (st.st_mode & S_IFMT) {
2012 case S_IFREG:
387e7e19
PH
2013 if (strbuf_read_file(&buf, read_from, st.st_size) != st.st_size)
2014 die("cannot open or read %s", read_from);
1cfe7733
JH
2015 break;
2016 case S_IFLNK:
edfd45d2 2017 if (strbuf_readlink(&buf, read_from, st.st_size) < 0)
1cfe7733
JH
2018 die("cannot readlink %s", read_from);
2019 break;
2020 default:
2021 die("unsupported file type %s", read_from);
2022 }
2023 }
2024 else {
2025 /* Reading from stdin */
2026 contents_from = "standard input";
1cfe7733 2027 mode = 0;
f1696ee3 2028 if (strbuf_read(&buf, 0, 0) < 0)
af6eb822 2029 die("read error %s from stdin", strerror(errno));
1cfe7733 2030 }
21e5ad50 2031 convert_to_git(path, buf.buf, buf.len, &buf, 0);
af6eb822
PH
2032 origin->file.ptr = buf.buf;
2033 origin->file.size = buf.len;
2034 pretend_sha1_file(buf.buf, buf.len, OBJ_BLOB, origin->blob_sha1);
1cfe7733
JH
2035 commit->util = origin;
2036
2037 /*
2038 * Read the current index, replace the path entry with
2039 * origin->blob_sha1 without mucking with its mode or type
2040 * bits; we are not going to write this index out -- we just
2041 * want to run "diff-index --cached".
2042 */
2043 discard_cache();
2044 read_cache();
2045
2046 len = strlen(path);
2047 if (!mode) {
2048 int pos = cache_name_pos(path, len);
2049 if (0 <= pos)
7a51ed66 2050 mode = active_cache[pos]->ce_mode;
1cfe7733
JH
2051 else
2052 /* Let's not bother reading from HEAD tree */
2053 mode = S_IFREG | 0644;
2054 }
2055 size = cache_entry_size(len);
2056 ce = xcalloc(1, size);
2057 hashcpy(ce->sha1, origin->blob_sha1);
2058 memcpy(ce->name, path, len);
2059 ce->ce_flags = create_ce_flags(len, 0);
2060 ce->ce_mode = create_ce_mode(mode);
2061 add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
2062
2063 /*
2064 * We are not going to write this out, so this does not matter
2065 * right now, but someday we might optimize diff-index --cached
2066 * with cache-tree information.
2067 */
2068 cache_tree_invalidate_path(active_cache_tree, path);
2069
2070 commit->buffer = xmalloc(400);
2071 ident = fmt_ident("Not Committed Yet", "not.committed.yet", NULL, 0);
1bb88be9 2072 snprintf(commit->buffer, 400,
1cfe7733
JH
2073 "tree 0000000000000000000000000000000000000000\n"
2074 "parent %s\n"
2075 "author %s\n"
2076 "committer %s\n\n"
2077 "Version of %s from %s\n",
2078 sha1_to_hex(head_sha1),
2079 ident, ident, path, contents_from ? contents_from : path);
2080 return commit;
2081}
2082
85af7929 2083static const char *prepare_final(struct scoreboard *sb)
f6c07d7d
JH
2084{
2085 int i;
2086 const char *final_commit_name = NULL;
85af7929 2087 struct rev_info *revs = sb->revs;
f6c07d7d
JH
2088
2089 /*
2090 * There must be one and only one positive commit in the
2091 * revs->pending array.
2092 */
2093 for (i = 0; i < revs->pending.nr; i++) {
2094 struct object *obj = revs->pending.objects[i].item;
2095 if (obj->flags & UNINTERESTING)
2096 continue;
2097 while (obj->type == OBJ_TAG)
2098 obj = deref_tag(obj, NULL, 0);
2099 if (obj->type != OBJ_COMMIT)
2100 die("Non commit %s?", revs->pending.objects[i].name);
2101 if (sb->final)
2102 die("More than one commit to dig from %s and %s?",
2103 revs->pending.objects[i].name,
2104 final_commit_name);
2105 sb->final = (struct commit *) obj;
2106 final_commit_name = revs->pending.objects[i].name;
2107 }
2108 return final_commit_name;
2109}
2110
85af7929
JH
2111static const char *prepare_initial(struct scoreboard *sb)
2112{
2113 int i;
2114 const char *final_commit_name = NULL;
2115 struct rev_info *revs = sb->revs;
2116
2117 /*
2118 * There must be one and only one negative commit, and it must be
2119 * the boundary.
2120 */
2121 for (i = 0; i < revs->pending.nr; i++) {
2122 struct object *obj = revs->pending.objects[i].item;
2123 if (!(obj->flags & UNINTERESTING))
2124 continue;
2125 while (obj->type == OBJ_TAG)
2126 obj = deref_tag(obj, NULL, 0);
2127 if (obj->type != OBJ_COMMIT)
2128 die("Non commit %s?", revs->pending.objects[i].name);
2129 if (sb->final)
2130 die("More than one commit to dig down to %s and %s?",
2131 revs->pending.objects[i].name,
2132 final_commit_name);
2133 sb->final = (struct commit *) obj;
2134 final_commit_name = revs->pending.objects[i].name;
2135 }
2136 if (!final_commit_name)
2137 die("No commit to dig down to?");
2138 return final_commit_name;
2139}
2140
5817da01
PH
2141static int blame_copy_callback(const struct option *option, const char *arg, int unset)
2142{
2143 int *opt = option->value;
2144
2145 /*
2146 * -C enables copy from removed files;
2147 * -C -C enables copy from existing files, but only
2148 * when blaming a new file;
2149 * -C -C -C enables copy from existing files for
2150 * everybody
2151 */
2152 if (*opt & PICKAXE_BLAME_COPY_HARDER)
2153 *opt |= PICKAXE_BLAME_COPY_HARDEST;
2154 if (*opt & PICKAXE_BLAME_COPY)
2155 *opt |= PICKAXE_BLAME_COPY_HARDER;
2156 *opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE;
2157
2158 if (arg)
2159 blame_copy_score = parse_score(arg);
2160 return 0;
2161}
2162
2163static int blame_move_callback(const struct option *option, const char *arg, int unset)
2164{
2165 int *opt = option->value;
2166
2167 *opt |= PICKAXE_BLAME_MOVE;
2168
2169 if (arg)
2170 blame_move_score = parse_score(arg);
2171 return 0;
2172}
2173
2174static int blame_bottomtop_callback(const struct option *option, const char *arg, int unset)
2175{
2176 const char **bottomtop = option->value;
2177 if (!arg)
2178 return -1;
2179 if (*bottomtop)
2180 die("More than one '-L n,m' option given");
2181 *bottomtop = arg;
2182 return 0;
2183}
2184
acca687f 2185int cmd_blame(int argc, const char **argv, const char *prefix)
cee7f245
JH
2186{
2187 struct rev_info revs;
2188 const char *path;
2189 struct scoreboard sb;
2190 struct origin *o;
2191 struct blame_entry *ent;
3f8d5204 2192 long dashdash_pos, bottom, top, lno;
cee7f245 2193 const char *final_commit_name = NULL;
21666f1a 2194 enum object_type type;
5817da01
PH
2195
2196 static const char *bottomtop = NULL;
2197 static int output_option = 0, opt = 0;
2198 static int show_stats = 0;
2199 static const char *revs_file = NULL;
2200 static const char *contents_from = NULL;
2201 static const struct option options[] = {
2202 OPT_BOOLEAN(0, "incremental", &incremental, "Show blame entries as we find them, incrementally"),
2203 OPT_BOOLEAN('b', NULL, &blank_boundary, "Show blank SHA-1 for boundary commits (Default: off)"),
2204 OPT_BOOLEAN(0, "root", &show_root, "Do not treat root commits as boundaries (Default: off)"),
2205 OPT_BOOLEAN(0, "show-stats", &show_stats, "Show work cost statistics"),
2206 OPT_BIT(0, "score-debug", &output_option, "Show output score for blame entries", OUTPUT_SHOW_SCORE),
2207 OPT_BIT('f', "show-name", &output_option, "Show original filename (Default: auto)", OUTPUT_SHOW_NAME),
2208 OPT_BIT('n', "show-number", &output_option, "Show original linenumber (Default: off)", OUTPUT_SHOW_NUMBER),
2209 OPT_BIT('p', "porcelain", &output_option, "Show in a format designed for machine consumption", OUTPUT_PORCELAIN),
2210 OPT_BIT('c', NULL, &output_option, "Use the same output mode as git-annotate (Default: off)", OUTPUT_ANNOTATE_COMPAT),
2211 OPT_BIT('t', NULL, &output_option, "Show raw timestamp (Default: off)", OUTPUT_RAW_TIMESTAMP),
2212 OPT_BIT('l', NULL, &output_option, "Show long commit SHA1 (Default: off)", OUTPUT_LONG_OBJECT_NAME),
2213 OPT_BIT('s', NULL, &output_option, "Suppress author name and timestamp (Default: off)", OUTPUT_NO_AUTHOR),
2214 OPT_BIT('w', NULL, &xdl_opts, "Ignore whitespace differences", XDF_IGNORE_WHITESPACE),
2215 OPT_STRING('S', NULL, &revs_file, "file", "Use revisions from <file> instead of calling git-rev-list"),
2216 OPT_STRING(0, "contents", &contents_from, "file", "Use <file>'s contents as the final image"),
2217 { OPTION_CALLBACK, 'C', NULL, &opt, "score", "Find line copies within and across files", PARSE_OPT_OPTARG, blame_copy_callback },
2218 { OPTION_CALLBACK, 'M', NULL, &opt, "score", "Find line movements within and across files", PARSE_OPT_OPTARG, blame_move_callback },
2219 OPT_CALLBACK('L', NULL, &bottomtop, "n,m", "Process only line range n,m, counting from 1", blame_bottomtop_callback),
2220 OPT_END()
2221 };
2222
2223 struct parse_opt_ctx_t ctx;
7ceacdff 2224 int cmd_is_annotate = !strcmp(argv[0], "annotate");
e68989a7 2225
ef90d6d4 2226 git_config(git_blame_config, NULL);
5817da01 2227 init_revisions(&revs, NULL);
612702e8 2228 save_commit_buffer = 0;
3f8d5204 2229 dashdash_pos = 0;
612702e8 2230
5817da01
PH
2231 parse_options_start(&ctx, argc, argv, PARSE_OPT_KEEP_DASHDASH |
2232 PARSE_OPT_KEEP_ARGV0);
2233 for (;;) {
5817da01
PH
2234 switch (parse_options_step(&ctx, options, blame_opt_usage)) {
2235 case PARSE_OPT_HELP:
2236 exit(129);
2237 case PARSE_OPT_DONE:
3f8d5204
PH
2238 if (ctx.argv[0])
2239 dashdash_pos = ctx.cpidx;
5817da01
PH
2240 goto parse_done;
2241 }
2242
2243 if (!strcmp(ctx.argv[0], "--reverse")) {
2244 ctx.argv[0] = "--children";
2245 reverse = 1;
2246 }
6b61ec05 2247 parse_revision_opt(&revs, &ctx, options, blame_opt_usage);
5817da01
PH
2248 }
2249parse_done:
2250 argc = parse_options_end(&ctx);
2251
7ceacdff
JH
2252 if (cmd_is_annotate)
2253 output_option |= OUTPUT_ANNOTATE_COMPAT;
2254
b3123f98
JH
2255 if (DIFF_OPT_TST(&revs.diffopt, FIND_COPIES_HARDER))
2256 opt |= (PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE |
2257 PICKAXE_BLAME_COPY_HARDER);
2258
4a0fc95f
JH
2259 if (!blame_move_score)
2260 blame_move_score = BLAME_DEFAULT_MOVE_SCORE;
2261 if (!blame_copy_score)
2262 blame_copy_score = BLAME_DEFAULT_COPY_SCORE;
2263
1732a1fd
JH
2264 /*
2265 * We have collected options unknown to us in argv[1..unk]
cee7f245 2266 * which are to be passed to revision machinery if we are
3dff5379 2267 * going to do the "bottom" processing.
cee7f245
JH
2268 *
2269 * The remaining are:
2270 *
3f8d5204
PH
2271 * (1) if dashdash_pos != 0, its either
2272 * "blame [revisions] -- <path>" or
2273 * "blame -- <path> <rev>"
cee7f245 2274 *
3f8d5204
PH
2275 * (2) otherwise, its one of the two:
2276 * "blame [revisions] <path>"
2277 * "blame <path> <rev>"
cee7f245 2278 *
3f8d5204
PH
2279 * Note that we must strip out <path> from the arguments: we do not
2280 * want the path pruning but we may want "bottom" processing.
cee7f245 2281 */
3f8d5204
PH
2282 if (dashdash_pos) {
2283 switch (argc - dashdash_pos - 1) {
2284 case 2: /* (1b) */
2285 if (argc != 4)
5817da01 2286 usage_with_options(blame_opt_usage, options);
3f8d5204
PH
2287 /* reorder for the new way: <rev> -- <path> */
2288 argv[1] = argv[3];
2289 argv[3] = argv[2];
2290 argv[2] = "--";
2291 /* FALLTHROUGH */
2292 case 1: /* (1a) */
2293 path = add_prefix(prefix, argv[--argc]);
2294 argv[argc] = NULL;
2295 break;
2296 default:
2297 usage_with_options(blame_opt_usage, options);
cee7f245 2298 }
3f8d5204
PH
2299 } else {
2300 if (argc < 2)
5817da01 2301 usage_with_options(blame_opt_usage, options);
3f8d5204 2302 path = add_prefix(prefix, argv[argc - 1]);
c455c87c 2303 if (argc == 3 && !has_string_in_work_tree(path)) { /* (2b) */
3f8d5204
PH
2304 path = add_prefix(prefix, argv[1]);
2305 argv[1] = argv[2];
cee7f245 2306 }
3f8d5204 2307 argv[argc - 1] = "--";
cee7f245 2308
3f8d5204 2309 setup_work_tree();
c455c87c 2310 if (!has_string_in_work_tree(path))
3f8d5204 2311 die("cannot stat path %s: %s", path, strerror(errno));
cee7f245
JH
2312 }
2313
3f8d5204 2314 setup_revisions(argc, argv, &revs, NULL);
cee7f245
JH
2315 memset(&sb, 0, sizeof(sb));
2316
85af7929
JH
2317 sb.revs = &revs;
2318 if (!reverse)
2319 final_commit_name = prepare_final(&sb);
2320 else if (contents_from)
2321 die("--contents and --children do not blend well.");
2322 else
2323 final_commit_name = prepare_initial(&sb);
cee7f245
JH
2324
2325 if (!sb.final) {
1732a1fd
JH
2326 /*
2327 * "--not A B -- path" without anything positive;
1cfe7733
JH
2328 * do not default to HEAD, but use the working tree
2329 * or "--contents".
1732a1fd 2330 */
1981820b 2331 setup_work_tree();
1cfe7733
JH
2332 sb.final = fake_working_tree_commit(path, contents_from);
2333 add_pending_object(&revs, &(sb.final->object), ":");
cee7f245 2334 }
1cfe7733
JH
2335 else if (contents_from)
2336 die("Cannot use --contents with final commit object name");
cee7f245 2337
1732a1fd
JH
2338 /*
2339 * If we have bottom, this will mark the ancestors of the
cee7f245
JH
2340 * bottom commits we would reach while traversing as
2341 * uninteresting.
2342 */
3d51e1b5
MK
2343 if (prepare_revision_walk(&revs))
2344 die("revision walk setup failed");
cee7f245 2345
1cfe7733
JH
2346 if (is_null_sha1(sb.final->object.sha1)) {
2347 char *buf;
2348 o = sb.final->util;
2349 buf = xmalloc(o->file.size + 1);
2350 memcpy(buf, o->file.ptr, o->file.size + 1);
2351 sb.final_buf = buf;
2352 sb.final_buf_size = o->file.size;
2353 }
2354 else {
2355 o = get_origin(&sb, sb.final, path);
2356 if (fill_blob_sha1(o))
2357 die("no such path %s in %s", path, final_commit_name);
cee7f245 2358
21666f1a 2359 sb.final_buf = read_sha1_file(o->blob_sha1, &type,
1cfe7733 2360 &sb.final_buf_size);
ab43e495
JH
2361 if (!sb.final_buf)
2362 die("Cannot read blob %s for path %s",
2363 sha1_to_hex(o->blob_sha1),
2364 path);
1cfe7733 2365 }
c2e525d9 2366 num_read_blob++;
cee7f245
JH
2367 lno = prepare_lines(&sb);
2368
931233bc
JH
2369 bottom = top = 0;
2370 if (bottomtop)
2371 prepare_blame_range(&sb, bottomtop, lno, &bottom, &top);
2372 if (bottom && top && top < bottom) {
2373 long tmp;
2374 tmp = top; top = bottom; bottom = tmp;
2375 }
cee7f245
JH
2376 if (bottom < 1)
2377 bottom = 1;
2378 if (top < 1)
2379 top = lno;
2380 bottom--;
2381 if (lno < top)
2382 die("file %s has only %lu lines", path, lno);
2383
2384 ent = xcalloc(1, sizeof(*ent));
2385 ent->lno = bottom;
2386 ent->num_lines = top - bottom;
2387 ent->suspect = o;
2388 ent->s_lno = bottom;
2389
2390 sb.ent = ent;
2391 sb.path = path;
2392
2393 if (revs_file && read_ancestry(revs_file))
2394 die("reading graft file %s failed: %s",
2395 revs_file, strerror(errno));
2396
d551a488 2397 read_mailmap(&mailmap, NULL);
f9567384 2398
b92565dc
MH
2399 if (!incremental)
2400 setup_pager();
2401
85af7929 2402 assign_blame(&sb, opt);
cee7f245 2403
717d1462
LT
2404 if (incremental)
2405 return 0;
2406
cee7f245
JH
2407 coalesce(&sb);
2408
2409 if (!(output_option & OUTPUT_PORCELAIN))
2410 find_alignment(&sb, &output_option);
2411
2412 output(&sb, output_option);
2413 free((void *)sb.final_buf);
2414 for (ent = sb.ent; ent; ) {
2415 struct blame_entry *e = ent->next;
2416 free(ent);
2417 ent = e;
2418 }
c2e525d9 2419
870b39c1 2420 if (show_stats) {
c2e525d9
JH
2421 printf("num read blob: %d\n", num_read_blob);
2422 printf("num get patch: %d\n", num_get_patch);
2423 printf("num commits: %d\n", num_commits);
2424 }
cee7f245
JH
2425 return 0;
2426}