blame: inline one-line function into its lone caller
[git/git.git] / builtin / blame.c
CommitLineData
cee7f245 1/*
31653c1a 2 * Blame
cee7f245
JH
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"
ffaf9cc0 22#include "utf8.h"
3b8a12e8 23#include "userdiff.h"
25ed3412 24#include "line-range.h"
cee7f245 25
efd2a8bd 26static char blame_usage[] = N_("git blame [options] [rev-opts] [rev] [--] file");
5817da01
PH
27
28static const char *blame_opt_usage[] = {
29 blame_usage,
30 "",
efd2a8bd 31 N_("[rev-opts] are documented in git-rev-list(1)"),
5817da01
PH
32 NULL
33};
cee7f245
JH
34
35static int longest_file;
36static int longest_author;
37static int max_orig_digits;
38static int max_digits;
5ff62c30 39static int max_score_digits;
4c10a5ca 40static int show_root;
85af7929 41static int reverse;
4c10a5ca 42static int blank_boundary;
717d1462 43static int incremental;
582aa00b 44static int xdl_opts;
84393bfd 45static int abbrev = -1;
3d1aa566 46static int no_whole_file_rename;
31653c1a
EL
47
48static enum date_mode blame_date_mode = DATE_ISO8601;
49static size_t blame_date_width;
50
c455c87c 51static struct string_list mailmap;
cee7f245 52
54a4c617
JH
53#ifndef DEBUG
54#define DEBUG 0
55#endif
56
c2e525d9
JH
57/* stats */
58static int num_read_blob;
59static int num_get_patch;
60static int num_commits;
61
d24bba80 62#define PICKAXE_BLAME_MOVE 01
18abd745
JH
63#define PICKAXE_BLAME_COPY 02
64#define PICKAXE_BLAME_COPY_HARDER 04
c63777c0 65#define PICKAXE_BLAME_COPY_HARDEST 010
d24bba80 66
4a0fc95f
JH
67/*
68 * blame for a blame_entry with score lower than these thresholds
69 * is not passed to the parent using move/copy logic.
70 */
71static unsigned blame_move_score;
72static unsigned blame_copy_score;
73#define BLAME_DEFAULT_MOVE_SCORE 20
74#define BLAME_DEFAULT_COPY_SCORE 40
75
cee7f245
JH
76/* bits #0..7 in revision.h, #8..11 used for merge_bases() in commit.c */
77#define METAINFO_SHOWN (1u<<12)
78#define MORE_THAN_ONE_PATH (1u<<13)
79
80/*
54a4c617 81 * One blob in a commit that is being suspected
cee7f245
JH
82 */
83struct origin {
54a4c617 84 int refcnt;
96e11709 85 struct origin *previous;
cee7f245 86 struct commit *commit;
c2e525d9 87 mmfile_t file;
cee7f245 88 unsigned char blob_sha1[20];
90064710 89 unsigned mode;
cee7f245
JH
90 char path[FLEX_ARRAY];
91};
92
4b4132fd
RS
93static int diff_hunks(mmfile_t *file_a, mmfile_t *file_b, long ctxlen,
94 xdl_emit_hunk_consume_func_t hunk_func, void *cb_data)
95{
96 xpparam_t xpp = {0};
97 xdemitconf_t xecfg = {0};
85c20c30 98 xdemitcb_t ecb = {NULL};
4b4132fd
RS
99
100 xpp.flags = xdl_opts;
101 xecfg.ctxlen = ctxlen;
102 xecfg.hunk_func = hunk_func;
103 ecb.priv = cb_data;
104 return xdi_diff(file_a, file_b, &xpp, &xecfg, &ecb);
105}
106
3b8a12e8
AB
107/*
108 * Prepare diff_filespec and convert it using diff textconv API
109 * if the textconv driver exists.
110 * Return 1 if the conversion succeeds, 0 otherwise.
111 */
e5fba602 112int textconv_object(const char *path,
90064710 113 unsigned mode,
e5fba602 114 const unsigned char *sha1,
e5450100 115 int sha1_valid,
e5fba602
CP
116 char **buf,
117 unsigned long *buf_size)
3b8a12e8
AB
118{
119 struct diff_filespec *df;
120 struct userdiff_driver *textconv;
121
122 df = alloc_filespec(path);
e5450100 123 fill_filespec(df, sha1, sha1_valid, mode);
3b8a12e8
AB
124 textconv = get_textconv(df);
125 if (!textconv) {
126 free_filespec(df);
127 return 0;
128 }
129
130 *buf_size = fill_textconv(textconv, df, buf);
131 free_filespec(df);
132 return 1;
133}
134
1732a1fd
JH
135/*
136 * Given an origin, prepare mmfile_t structure to be used by the
137 * diff machinery
138 */
3b8a12e8
AB
139static void fill_origin_blob(struct diff_options *opt,
140 struct origin *o, mmfile_t *file)
c2e525d9
JH
141{
142 if (!o->file.ptr) {
21666f1a 143 enum object_type type;
3b8a12e8
AB
144 unsigned long file_size;
145
c2e525d9 146 num_read_blob++;
3b8a12e8 147 if (DIFF_OPT_TST(opt, ALLOW_TEXTCONV) &&
e5450100 148 textconv_object(o->path, o->mode, o->blob_sha1, 1, &file->ptr, &file_size))
3b8a12e8
AB
149 ;
150 else
151 file->ptr = read_sha1_file(o->blob_sha1, &type, &file_size);
152 file->size = file_size;
153
ab43e495
JH
154 if (!file->ptr)
155 die("Cannot read blob %s for path %s",
156 sha1_to_hex(o->blob_sha1),
157 o->path);
c2e525d9
JH
158 o->file = *file;
159 }
160 else
161 *file = o->file;
c2e525d9
JH
162}
163
1732a1fd
JH
164/*
165 * Origin is refcounted and usually we keep the blob contents to be
166 * reused.
167 */
54a4c617
JH
168static inline struct origin *origin_incref(struct origin *o)
169{
170 if (o)
171 o->refcnt++;
172 return o;
173}
174
175static void origin_decref(struct origin *o)
176{
177 if (o && --o->refcnt <= 0) {
96e11709
JH
178 if (o->previous)
179 origin_decref(o->previous);
8e0f7003 180 free(o->file.ptr);
54a4c617
JH
181 free(o);
182 }
183}
184
7c3c7962
JH
185static void drop_origin_blob(struct origin *o)
186{
187 if (o->file.ptr) {
188 free(o->file.ptr);
189 o->file.ptr = NULL;
190 }
191}
192
1732a1fd
JH
193/*
194 * Each group of lines is described by a blame_entry; it can be split
195 * as we pass blame to the parents. They form a linked list in the
196 * scoreboard structure, sorted by the target line number.
197 */
cee7f245
JH
198struct blame_entry {
199 struct blame_entry *prev;
200 struct blame_entry *next;
201
202 /* the first line of this group in the final image;
203 * internally all line numbers are 0 based.
204 */
205 int lno;
206
207 /* how many lines this group has */
208 int num_lines;
209
210 /* the commit that introduced this group into the final image */
211 struct origin *suspect;
212
213 /* true if the suspect is truly guilty; false while we have not
214 * checked if the group came from one of its parents.
215 */
216 char guilty;
217
eb46328e
AG
218 /* true if the entry has been scanned for copies in the current parent
219 */
220 char scanned;
221
cee7f245
JH
222 /* the line number of the first line of this group in the
223 * suspect's file; internally all line numbers are 0 based.
224 */
225 int s_lno;
5ff62c30
JH
226
227 /* how significant this entry is -- cached to avoid
1732a1fd 228 * scanning the lines over and over.
5ff62c30
JH
229 */
230 unsigned score;
cee7f245
JH
231};
232
1732a1fd
JH
233/*
234 * The current state of the blame assignment.
235 */
cee7f245
JH
236struct scoreboard {
237 /* the final commit (i.e. where we started digging from) */
238 struct commit *final;
85af7929 239 struct rev_info *revs;
cee7f245
JH
240 const char *path;
241
1732a1fd
JH
242 /*
243 * The contents in the final image.
244 * Used by many functions to obtain contents of the nth line,
245 * indexed with scoreboard.lineno[blame_entry.lno].
cee7f245
JH
246 */
247 const char *final_buf;
248 unsigned long final_buf_size;
249
250 /* linked list of blames */
251 struct blame_entry *ent;
252
612702e8 253 /* look-up a line in the final buffer */
cee7f245
JH
254 int num_lines;
255 int *lineno;
256};
257
171dccd5 258static inline int same_suspect(struct origin *a, struct origin *b)
46014766 259{
171dccd5 260 if (a == b)
57584d9e 261 return 1;
171dccd5
JH
262 if (a->commit != b->commit)
263 return 0;
264 return !strcmp(a->path, b->path);
46014766
JH
265}
266
54a4c617
JH
267static void sanity_check_refcnt(struct scoreboard *);
268
1732a1fd
JH
269/*
270 * If two blame entries that are next to each other came from
271 * contiguous lines in the same origin (i.e. <commit, path> pair),
272 * merge them together.
273 */
cee7f245
JH
274static void coalesce(struct scoreboard *sb)
275{
276 struct blame_entry *ent, *next;
277
278 for (ent = sb->ent; ent && (next = ent->next); ent = next) {
171dccd5 279 if (same_suspect(ent->suspect, next->suspect) &&
cee7f245
JH
280 ent->guilty == next->guilty &&
281 ent->s_lno + ent->num_lines == next->s_lno) {
282 ent->num_lines += next->num_lines;
283 ent->next = next->next;
284 if (ent->next)
285 ent->next->prev = ent;
54a4c617 286 origin_decref(next->suspect);
cee7f245 287 free(next);
46014766 288 ent->score = 0;
cee7f245
JH
289 next = ent; /* again */
290 }
291 }
54a4c617
JH
292
293 if (DEBUG) /* sanity */
294 sanity_check_refcnt(sb);
cee7f245
JH
295}
296
1732a1fd
JH
297/*
298 * Given a commit and a path in it, create a new origin structure.
299 * The callers that add blame to the scoreboard should use
300 * get_origin() to obtain shared, refcounted copy instead of calling
301 * this function directly.
302 */
854b97f6
JH
303static struct origin *make_origin(struct commit *commit, const char *path)
304{
305 struct origin *o;
306 o = xcalloc(1, sizeof(*o) + strlen(path) + 1);
307 o->commit = commit;
308 o->refcnt = 1;
309 strcpy(o->path, path);
310 return o;
311}
312
1732a1fd
JH
313/*
314 * Locate an existing origin or create a new one.
315 */
f6c0e191
JH
316static struct origin *get_origin(struct scoreboard *sb,
317 struct commit *commit,
318 const char *path)
cee7f245 319{
f6c0e191 320 struct blame_entry *e;
cee7f245 321
f6c0e191
JH
322 for (e = sb->ent; e; e = e->next) {
323 if (e->suspect->commit == commit &&
324 !strcmp(e->suspect->path, path))
54a4c617 325 return origin_incref(e->suspect);
cee7f245 326 }
854b97f6 327 return make_origin(commit, path);
cee7f245
JH
328}
329
1732a1fd
JH
330/*
331 * Fill the blob_sha1 field of an origin if it hasn't, so that later
332 * call to fill_origin_blob() can use it to locate the data. blob_sha1
333 * for an origin is also used to pass the blame for the entire file to
334 * the parent to detect the case where a child's blob is identical to
335 * that of its parent's.
90064710
KS
336 *
337 * This also fills origin->mode for corresponding tree path.
1732a1fd 338 */
90064710 339static int fill_blob_sha1_and_mode(struct origin *origin)
f6c0e191 340{
f6c0e191
JH
341 if (!is_null_sha1(origin->blob_sha1))
342 return 0;
343 if (get_tree_entry(origin->commit->object.sha1,
344 origin->path,
90064710 345 origin->blob_sha1, &origin->mode))
f6c0e191 346 goto error_out;
21666f1a 347 if (sha1_object_info(origin->blob_sha1, NULL) != OBJ_BLOB)
f6c0e191
JH
348 goto error_out;
349 return 0;
350 error_out:
351 hashclr(origin->blob_sha1);
90064710 352 origin->mode = S_IFINVALID;
f6c0e191
JH
353 return -1;
354}
355
1732a1fd
JH
356/*
357 * We have an origin -- check if the same path exists in the
358 * parent and return an origin structure to represent it.
359 */
f6c0e191 360static struct origin *find_origin(struct scoreboard *sb,
cee7f245
JH
361 struct commit *parent,
362 struct origin *origin)
363{
364 struct origin *porigin = NULL;
365 struct diff_options diff_opts;
f6c0e191
JH
366 const char *paths[2];
367
0d981c67 368 if (parent->util) {
1732a1fd
JH
369 /*
370 * Each commit object can cache one origin in that
371 * commit. This is a freestanding copy of origin and
372 * not refcounted.
854b97f6 373 */
0d981c67 374 struct origin *cached = parent->util;
854b97f6 375 if (!strcmp(cached->path, origin->path)) {
1732a1fd
JH
376 /*
377 * The same path between origin and its parent
378 * without renaming -- the most common case.
379 */
854b97f6 380 porigin = get_origin(sb, parent, cached->path);
1732a1fd
JH
381
382 /*
383 * If the origin was newly created (i.e. get_origin
384 * would call make_origin if none is found in the
90064710 385 * scoreboard), it does not know the blob_sha1/mode,
1732a1fd 386 * so copy it. Otherwise porigin was in the
90064710 387 * scoreboard and already knows blob_sha1/mode.
1732a1fd 388 */
90064710 389 if (porigin->refcnt == 1) {
854b97f6 390 hashcpy(porigin->blob_sha1, cached->blob_sha1);
90064710
KS
391 porigin->mode = cached->mode;
392 }
854b97f6
JH
393 return porigin;
394 }
395 /* otherwise it was not very useful; free it */
396 free(parent->util);
397 parent->util = NULL;
0d981c67
JH
398 }
399
f6c0e191
JH
400 /* See if the origin->path is different between parent
401 * and origin first. Most of the time they are the
402 * same and diff-tree is fairly efficient about this.
403 */
404 diff_setup(&diff_opts);
8f67f8ae 405 DIFF_OPT_SET(&diff_opts, RECURSIVE);
f6c0e191
JH
406 diff_opts.detect_rename = 0;
407 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
408 paths[0] = origin->path;
409 paths[1] = NULL;
410
411 diff_tree_setup_paths(paths, &diff_opts);
28452655 412 diff_setup_done(&diff_opts);
1cfe7733
JH
413
414 if (is_null_sha1(origin->commit->object.sha1))
415 do_diff_cache(parent->tree->object.sha1, &diff_opts);
416 else
417 diff_tree_sha1(parent->tree->object.sha1,
418 origin->commit->tree->object.sha1,
419 "", &diff_opts);
f6c0e191
JH
420 diffcore_std(&diff_opts);
421
f6c0e191
JH
422 if (!diff_queued_diff.nr) {
423 /* The path is the same as parent */
424 porigin = get_origin(sb, parent, origin->path);
425 hashcpy(porigin->blob_sha1, origin->blob_sha1);
90064710 426 porigin->mode = origin->mode;
a9b2d424
JH
427 } else {
428 /*
429 * Since origin->path is a pathspec, if the parent
430 * commit had it as a directory, we will see a whole
431 * bunch of deletion of files in the directory that we
432 * do not care about.
433 */
434 int i;
435 struct diff_filepair *p = NULL;
436 for (i = 0; i < diff_queued_diff.nr; i++) {
437 const char *name;
438 p = diff_queued_diff.queue[i];
439 name = p->one->path ? p->one->path : p->two->path;
440 if (!strcmp(name, origin->path))
441 break;
442 }
443 if (!p)
444 die("internal error in blame::find_origin");
f6c0e191
JH
445 switch (p->status) {
446 default:
acca687f 447 die("internal error in blame::find_origin (%c)",
f6c0e191
JH
448 p->status);
449 case 'M':
450 porigin = get_origin(sb, parent, origin->path);
451 hashcpy(porigin->blob_sha1, p->one->sha1);
90064710 452 porigin->mode = p->one->mode;
f6c0e191
JH
453 break;
454 case 'A':
455 case 'T':
456 /* Did not exist in parent, or type changed */
457 break;
458 }
459 }
460 diff_flush(&diff_opts);
03b69c76 461 diff_tree_release_paths(&diff_opts);
0d981c67 462 if (porigin) {
1732a1fd
JH
463 /*
464 * Create a freestanding copy that is not part of
465 * the refcounted origin found in the scoreboard, and
466 * cache it in the commit.
467 */
854b97f6 468 struct origin *cached;
1732a1fd 469
854b97f6
JH
470 cached = make_origin(porigin->commit, porigin->path);
471 hashcpy(cached->blob_sha1, porigin->blob_sha1);
90064710 472 cached->mode = porigin->mode;
854b97f6 473 parent->util = cached;
0d981c67 474 }
f69e743d
JH
475 return porigin;
476}
f6c0e191 477
1732a1fd
JH
478/*
479 * We have an origin -- find the path that corresponds to it in its
480 * parent and return an origin structure to represent it.
481 */
f69e743d
JH
482static struct origin *find_rename(struct scoreboard *sb,
483 struct commit *parent,
484 struct origin *origin)
485{
486 struct origin *porigin = NULL;
487 struct diff_options diff_opts;
488 int i;
489 const char *paths[2];
cee7f245
JH
490
491 diff_setup(&diff_opts);
8f67f8ae 492 DIFF_OPT_SET(&diff_opts, RECURSIVE);
cee7f245
JH
493 diff_opts.detect_rename = DIFF_DETECT_RENAME;
494 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
2f3f8b21 495 diff_opts.single_follow = origin->path;
cee7f245
JH
496 paths[0] = NULL;
497 diff_tree_setup_paths(paths, &diff_opts);
28452655 498 diff_setup_done(&diff_opts);
1cfe7733
JH
499
500 if (is_null_sha1(origin->commit->object.sha1))
501 do_diff_cache(parent->tree->object.sha1, &diff_opts);
502 else
503 diff_tree_sha1(parent->tree->object.sha1,
504 origin->commit->tree->object.sha1,
505 "", &diff_opts);
cee7f245
JH
506 diffcore_std(&diff_opts);
507
508 for (i = 0; i < diff_queued_diff.nr; i++) {
509 struct diff_filepair *p = diff_queued_diff.queue[i];
612702e8 510 if ((p->status == 'R' || p->status == 'C') &&
f6c0e191
JH
511 !strcmp(p->two->path, origin->path)) {
512 porigin = get_origin(sb, parent, p->one->path);
513 hashcpy(porigin->blob_sha1, p->one->sha1);
90064710 514 porigin->mode = p->one->mode;
cee7f245
JH
515 break;
516 }
517 }
518 diff_flush(&diff_opts);
03b69c76 519 diff_tree_release_paths(&diff_opts);
cee7f245
JH
520 return porigin;
521}
522
1732a1fd 523/*
3dff5379 524 * Link in a new blame entry to the scoreboard. Entries that cover the
1732a1fd
JH
525 * same line range have been removed from the scoreboard previously.
526 */
cee7f245
JH
527static void add_blame_entry(struct scoreboard *sb, struct blame_entry *e)
528{
529 struct blame_entry *ent, *prev = NULL;
530
54a4c617
JH
531 origin_incref(e->suspect);
532
cee7f245
JH
533 for (ent = sb->ent; ent && ent->lno < e->lno; ent = ent->next)
534 prev = ent;
535
536 /* prev, if not NULL, is the last one that is below e */
537 e->prev = prev;
538 if (prev) {
539 e->next = prev->next;
540 prev->next = e;
541 }
542 else {
543 e->next = sb->ent;
544 sb->ent = e;
545 }
546 if (e->next)
547 e->next->prev = e;
548}
549
1732a1fd
JH
550/*
551 * src typically is on-stack; we want to copy the information in it to
790296fd 552 * a malloced blame_entry that is already on the linked list of the
1732a1fd
JH
553 * scoreboard. The origin of dst loses a refcnt while the origin of src
554 * gains one.
555 */
cee7f245
JH
556static void dup_entry(struct blame_entry *dst, struct blame_entry *src)
557{
558 struct blame_entry *p, *n;
54a4c617 559
cee7f245
JH
560 p = dst->prev;
561 n = dst->next;
54a4c617
JH
562 origin_incref(src->suspect);
563 origin_decref(dst->suspect);
cee7f245
JH
564 memcpy(dst, src, sizeof(*src));
565 dst->prev = p;
566 dst->next = n;
5ff62c30 567 dst->score = 0;
cee7f245
JH
568}
569
25ed3412 570static const char *nth_line(struct scoreboard *sb, long lno)
cee7f245
JH
571{
572 return sb->final_buf + sb->lineno[lno];
573}
574
25ed3412
BY
575static const char *nth_line_cb(void *data, long lno)
576{
577 return nth_line((struct scoreboard *)data, lno);
578}
579
1732a1fd
JH
580/*
581 * It is known that lines between tlno to same came from parent, and e
582 * has an overlap with that range. it also is known that parent's
583 * line plno corresponds to e's line tlno.
584 *
585 * <---- e ----->
586 * <------>
587 * <------------>
588 * <------------>
589 * <------------------>
590 *
591 * Split e into potentially three parts; before this chunk, the chunk
592 * to be blamed for the parent, and after that portion.
593 */
54a4c617 594static void split_overlap(struct blame_entry *split,
cee7f245
JH
595 struct blame_entry *e,
596 int tlno, int plno, int same,
597 struct origin *parent)
598{
cee7f245
JH
599 int chunk_end_lno;
600 memset(split, 0, sizeof(struct blame_entry [3]));
601
602 if (e->s_lno < tlno) {
603 /* there is a pre-chunk part not blamed on parent */
54a4c617 604 split[0].suspect = origin_incref(e->suspect);
cee7f245
JH
605 split[0].lno = e->lno;
606 split[0].s_lno = e->s_lno;
607 split[0].num_lines = tlno - e->s_lno;
608 split[1].lno = e->lno + tlno - e->s_lno;
609 split[1].s_lno = plno;
610 }
611 else {
612 split[1].lno = e->lno;
613 split[1].s_lno = plno + (e->s_lno - tlno);
614 }
615
616 if (same < e->s_lno + e->num_lines) {
617 /* there is a post-chunk part not blamed on parent */
54a4c617 618 split[2].suspect = origin_incref(e->suspect);
cee7f245
JH
619 split[2].lno = e->lno + (same - e->s_lno);
620 split[2].s_lno = e->s_lno + (same - e->s_lno);
621 split[2].num_lines = e->s_lno + e->num_lines - same;
622 chunk_end_lno = split[2].lno;
623 }
624 else
625 chunk_end_lno = e->lno + e->num_lines;
626 split[1].num_lines = chunk_end_lno - split[1].lno;
627
1732a1fd
JH
628 /*
629 * if it turns out there is nothing to blame the parent for,
630 * forget about the splitting. !split[1].suspect signals this.
631 */
cee7f245
JH
632 if (split[1].num_lines < 1)
633 return;
54a4c617 634 split[1].suspect = origin_incref(parent);
cee7f245
JH
635}
636
1732a1fd
JH
637/*
638 * split_overlap() divided an existing blame e into up to three parts
639 * in split. Adjust the linked list of blames in the scoreboard to
640 * reflect the split.
641 */
cee7f245 642static void split_blame(struct scoreboard *sb,
54a4c617 643 struct blame_entry *split,
cee7f245
JH
644 struct blame_entry *e)
645{
646 struct blame_entry *new_entry;
647
648 if (split[0].suspect && split[2].suspect) {
1732a1fd 649 /* The first part (reuse storage for the existing entry e) */
cee7f245
JH
650 dup_entry(e, &split[0]);
651
1732a1fd 652 /* The last part -- me */
cee7f245
JH
653 new_entry = xmalloc(sizeof(*new_entry));
654 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
655 add_blame_entry(sb, new_entry);
656
1732a1fd 657 /* ... and the middle part -- parent */
cee7f245
JH
658 new_entry = xmalloc(sizeof(*new_entry));
659 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
660 add_blame_entry(sb, new_entry);
661 }
662 else if (!split[0].suspect && !split[2].suspect)
1732a1fd
JH
663 /*
664 * The parent covers the entire area; reuse storage for
665 * e and replace it with the parent.
666 */
cee7f245
JH
667 dup_entry(e, &split[1]);
668 else if (split[0].suspect) {
1732a1fd 669 /* me and then parent */
cee7f245
JH
670 dup_entry(e, &split[0]);
671
672 new_entry = xmalloc(sizeof(*new_entry));
673 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
674 add_blame_entry(sb, new_entry);
675 }
676 else {
1732a1fd 677 /* parent and then me */
cee7f245
JH
678 dup_entry(e, &split[1]);
679
680 new_entry = xmalloc(sizeof(*new_entry));
681 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
682 add_blame_entry(sb, new_entry);
683 }
684
54a4c617 685 if (DEBUG) { /* sanity */
cee7f245 686 struct blame_entry *ent;
612702e8 687 int lno = sb->ent->lno, corrupt = 0;
cee7f245
JH
688
689 for (ent = sb->ent; ent; ent = ent->next) {
690 if (lno != ent->lno)
691 corrupt = 1;
692 if (ent->s_lno < 0)
693 corrupt = 1;
694 lno += ent->num_lines;
695 }
696 if (corrupt) {
612702e8 697 lno = sb->ent->lno;
cee7f245
JH
698 for (ent = sb->ent; ent; ent = ent->next) {
699 printf("L %8d l %8d n %8d\n",
700 lno, ent->lno, ent->num_lines);
701 lno = ent->lno + ent->num_lines;
702 }
703 die("oops");
704 }
705 }
706}
707
1732a1fd
JH
708/*
709 * After splitting the blame, the origins used by the
710 * on-stack blame_entry should lose one refcnt each.
711 */
54a4c617
JH
712static void decref_split(struct blame_entry *split)
713{
714 int i;
715
716 for (i = 0; i < 3; i++)
717 origin_decref(split[i].suspect);
718}
719
1732a1fd
JH
720/*
721 * Helper for blame_chunk(). blame_entry e is known to overlap with
722 * the patch hunk; split it and pass blame to the parent.
723 */
cee7f245
JH
724static void blame_overlap(struct scoreboard *sb, struct blame_entry *e,
725 int tlno, int plno, int same,
726 struct origin *parent)
727{
728 struct blame_entry split[3];
729
730 split_overlap(split, e, tlno, plno, same, parent);
54a4c617
JH
731 if (split[1].suspect)
732 split_blame(sb, split, e);
733 decref_split(split);
cee7f245
JH
734}
735
1732a1fd
JH
736/*
737 * Find the line number of the last line the target is suspected for.
738 */
cee7f245
JH
739static int find_last_in_target(struct scoreboard *sb, struct origin *target)
740{
741 struct blame_entry *e;
742 int last_in_target = -1;
743
744 for (e = sb->ent; e; e = e->next) {
171dccd5 745 if (e->guilty || !same_suspect(e->suspect, target))
cee7f245
JH
746 continue;
747 if (last_in_target < e->s_lno + e->num_lines)
748 last_in_target = e->s_lno + e->num_lines;
749 }
750 return last_in_target;
751}
752
1732a1fd
JH
753/*
754 * Process one hunk from the patch between the current suspect for
755 * blame_entry e and its parent. Find and split the overlap, and
756 * pass blame to the overlapping part to the parent.
757 */
cee7f245
JH
758static void blame_chunk(struct scoreboard *sb,
759 int tlno, int plno, int same,
760 struct origin *target, struct origin *parent)
761{
612702e8 762 struct blame_entry *e;
cee7f245 763
612702e8 764 for (e = sb->ent; e; e = e->next) {
171dccd5 765 if (e->guilty || !same_suspect(e->suspect, target))
cee7f245
JH
766 continue;
767 if (same <= e->s_lno)
768 continue;
769 if (tlno < e->s_lno + e->num_lines)
770 blame_overlap(sb, e, tlno, plno, same, parent);
771 }
772}
773
cdcab133
RS
774struct blame_chunk_cb_data {
775 struct scoreboard *sb;
776 struct origin *target;
777 struct origin *parent;
778 long plno;
779 long tlno;
780};
781
0af596c6
RS
782static int blame_chunk_cb(long start_a, long count_a,
783 long start_b, long count_b, void *data)
cdcab133
RS
784{
785 struct blame_chunk_cb_data *d = data;
0af596c6
RS
786 blame_chunk(d->sb, d->tlno, d->plno, start_b, d->target, d->parent);
787 d->plno = start_a + count_a;
788 d->tlno = start_b + count_b;
789 return 0;
cdcab133
RS
790}
791
1732a1fd
JH
792/*
793 * We are looking at the origin 'target' and aiming to pass blame
794 * for the lines it is suspected to its parent. Run diff to find
795 * which lines came from parent and pass blame for them.
796 */
cee7f245
JH
797static int pass_blame_to_parent(struct scoreboard *sb,
798 struct origin *target,
799 struct origin *parent)
800{
cdcab133 801 int last_in_target;
7c4ed8c8 802 mmfile_t file_p, file_o;
66dbfd55 803 struct blame_chunk_cb_data d;
0af596c6 804
66dbfd55
GV
805 memset(&d, 0, sizeof(d));
806 d.sb = sb; d.target = target; d.parent = parent;
cee7f245
JH
807 last_in_target = find_last_in_target(sb, target);
808 if (last_in_target < 0)
809 return 1; /* nothing remains for this target */
810
3b8a12e8
AB
811 fill_origin_blob(&sb->revs->diffopt, parent, &file_p);
812 fill_origin_blob(&sb->revs->diffopt, target, &file_o);
7c4ed8c8 813 num_get_patch++;
cee7f245 814
4b4132fd 815 diff_hunks(&file_p, &file_o, 0, blame_chunk_cb, &d);
1732a1fd 816 /* The rest (i.e. anything after tlno) are the same as the parent */
cdcab133 817 blame_chunk(sb, d.tlno, d.plno, last_in_target, target, parent);
cee7f245 818
cee7f245
JH
819 return 0;
820}
821
1732a1fd
JH
822/*
823 * The lines in blame_entry after splitting blames many times can become
824 * very small and trivial, and at some point it becomes pointless to
825 * blame the parents. E.g. "\t\t}\n\t}\n\n" appears everywhere in any
826 * ordinary C program, and it is not worth to say it was copied from
827 * totally unrelated file in the parent.
828 *
829 * Compute how trivial the lines in the blame_entry are.
830 */
5ff62c30
JH
831static unsigned ent_score(struct scoreboard *sb, struct blame_entry *e)
832{
833 unsigned score;
834 const char *cp, *ep;
835
836 if (e->score)
837 return e->score;
838
612702e8 839 score = 1;
5ff62c30
JH
840 cp = nth_line(sb, e->lno);
841 ep = nth_line(sb, e->lno + e->num_lines);
842 while (cp < ep) {
843 unsigned ch = *((unsigned char *)cp);
844 if (isalnum(ch))
845 score++;
846 cp++;
847 }
848 e->score = score;
849 return score;
850}
851
1732a1fd
JH
852/*
853 * best_so_far[] and this[] are both a split of an existing blame_entry
854 * that passes blame to the parent. Maintain best_so_far the best split
855 * so far, by comparing this and best_so_far and copying this into
856 * bst_so_far as needed.
857 */
5ff62c30 858static void copy_split_if_better(struct scoreboard *sb,
54a4c617
JH
859 struct blame_entry *best_so_far,
860 struct blame_entry *this)
d24bba80 861{
54a4c617
JH
862 int i;
863
d24bba80
JH
864 if (!this[1].suspect)
865 return;
5ff62c30
JH
866 if (best_so_far[1].suspect) {
867 if (ent_score(sb, &this[1]) < ent_score(sb, &best_so_far[1]))
868 return;
869 }
54a4c617
JH
870
871 for (i = 0; i < 3; i++)
872 origin_incref(this[i].suspect);
873 decref_split(best_so_far);
d24bba80
JH
874 memcpy(best_so_far, this, sizeof(struct blame_entry [3]));
875}
876
dd166aa8
JH
877/*
878 * We are looking at a part of the final image represented by
879 * ent (tlno and same are offset by ent->s_lno).
880 * tlno is where we are looking at in the final image.
881 * up to (but not including) same match preimage.
882 * plno is where we are looking at in the preimage.
883 *
884 * <-------------- final image ---------------------->
885 * <------ent------>
886 * ^tlno ^same
887 * <---------preimage----->
888 * ^plno
889 *
890 * All line numbers are 0-based.
891 */
892static void handle_split(struct scoreboard *sb,
893 struct blame_entry *ent,
894 int tlno, int plno, int same,
895 struct origin *parent,
896 struct blame_entry *split)
897{
898 if (ent->num_lines <= tlno)
899 return;
900 if (tlno < same) {
901 struct blame_entry this[3];
902 tlno += ent->s_lno;
903 same += ent->s_lno;
904 split_overlap(this, ent, tlno, plno, same, parent);
905 copy_split_if_better(sb, split, this);
906 decref_split(this);
907 }
908}
909
cdcab133
RS
910struct handle_split_cb_data {
911 struct scoreboard *sb;
912 struct blame_entry *ent;
913 struct origin *parent;
914 struct blame_entry *split;
915 long plno;
916 long tlno;
917};
918
5d23ec76
RS
919static int handle_split_cb(long start_a, long count_a,
920 long start_b, long count_b, void *data)
cdcab133
RS
921{
922 struct handle_split_cb_data *d = data;
5d23ec76
RS
923 handle_split(d->sb, d->ent, d->tlno, d->plno, start_b, d->parent,
924 d->split);
925 d->plno = start_a + count_a;
926 d->tlno = start_b + count_b;
927 return 0;
cdcab133
RS
928}
929
1732a1fd
JH
930/*
931 * Find the lines from parent that are the same as ent so that
932 * we can pass blames to it. file_p has the blob contents for
933 * the parent.
934 */
d24bba80
JH
935static void find_copy_in_blob(struct scoreboard *sb,
936 struct blame_entry *ent,
937 struct origin *parent,
54a4c617 938 struct blame_entry *split,
d24bba80
JH
939 mmfile_t *file_p)
940{
941 const char *cp;
942 int cnt;
943 mmfile_t file_o;
66dbfd55 944 struct handle_split_cb_data d;
5d23ec76 945
66dbfd55
GV
946 memset(&d, 0, sizeof(d));
947 d.sb = sb; d.ent = ent; d.parent = parent; d.split = split;
1732a1fd
JH
948 /*
949 * Prepare mmfile that contains only the lines in ent.
950 */
d24bba80 951 cp = nth_line(sb, ent->lno);
4b25d091 952 file_o.ptr = (char *) cp;
d24bba80
JH
953 cnt = ent->num_lines;
954
955 while (cnt && cp < sb->final_buf + sb->final_buf_size) {
956 if (*cp++ == '\n')
957 cnt--;
958 }
959 file_o.size = cp - file_o.ptr;
960
dd166aa8
JH
961 /*
962 * file_o is a part of final image we are annotating.
963 * file_p partially may match that image.
964 */
d24bba80 965 memset(split, 0, sizeof(struct blame_entry [3]));
4b4132fd 966 diff_hunks(file_p, &file_o, 1, handle_split_cb, &d);
dd166aa8 967 /* remainder, if any, all match the preimage */
cdcab133 968 handle_split(sb, ent, d.tlno, d.plno, ent->num_lines, parent, split);
d24bba80
JH
969}
970
1732a1fd
JH
971/*
972 * See if lines currently target is suspected for can be attributed to
973 * parent.
974 */
d24bba80
JH
975static int find_move_in_parent(struct scoreboard *sb,
976 struct origin *target,
977 struct origin *parent)
978{
650e2f67 979 int last_in_target, made_progress;
46014766 980 struct blame_entry *e, split[3];
d24bba80 981 mmfile_t file_p;
d24bba80
JH
982
983 last_in_target = find_last_in_target(sb, target);
984 if (last_in_target < 0)
985 return 1; /* nothing remains for this target */
986
3b8a12e8 987 fill_origin_blob(&sb->revs->diffopt, parent, &file_p);
c2e525d9 988 if (!file_p.ptr)
d24bba80 989 return 0;
d24bba80 990
650e2f67
JH
991 made_progress = 1;
992 while (made_progress) {
993 made_progress = 0;
994 for (e = sb->ent; e; e = e->next) {
1617fe98
AG
995 if (e->guilty || !same_suspect(e->suspect, target) ||
996 ent_score(sb, e) < blame_move_score)
650e2f67
JH
997 continue;
998 find_copy_in_blob(sb, e, parent, split, &file_p);
999 if (split[1].suspect &&
1000 blame_move_score < ent_score(sb, &split[1])) {
1001 split_blame(sb, split, e);
1002 made_progress = 1;
1003 }
1004 decref_split(split);
1005 }
d24bba80 1006 }
d24bba80
JH
1007 return 0;
1008}
1009
33494784
JH
1010struct blame_list {
1011 struct blame_entry *ent;
1012 struct blame_entry split[3];
1013};
1014
1732a1fd
JH
1015/*
1016 * Count the number of entries the target is suspected for,
1017 * and prepare a list of entry and the best split.
1018 */
33494784
JH
1019static struct blame_list *setup_blame_list(struct scoreboard *sb,
1020 struct origin *target,
1617fe98 1021 int min_score,
33494784
JH
1022 int *num_ents_p)
1023{
1024 struct blame_entry *e;
1025 int num_ents, i;
1026 struct blame_list *blame_list = NULL;
1027
33494784 1028 for (e = sb->ent, num_ents = 0; e; e = e->next)
1617fe98
AG
1029 if (!e->scanned && !e->guilty &&
1030 same_suspect(e->suspect, target) &&
1031 min_score < ent_score(sb, e))
33494784
JH
1032 num_ents++;
1033 if (num_ents) {
1034 blame_list = xcalloc(num_ents, sizeof(struct blame_list));
1035 for (e = sb->ent, i = 0; e; e = e->next)
1617fe98
AG
1036 if (!e->scanned && !e->guilty &&
1037 same_suspect(e->suspect, target) &&
1038 min_score < ent_score(sb, e))
33494784
JH
1039 blame_list[i++].ent = e;
1040 }
1041 *num_ents_p = num_ents;
1042 return blame_list;
1043}
1044
eb46328e
AG
1045/*
1046 * Reset the scanned status on all entries.
1047 */
1048static void reset_scanned_flag(struct scoreboard *sb)
1049{
1050 struct blame_entry *e;
1051 for (e = sb->ent; e; e = e->next)
1052 e->scanned = 0;
1053}
1054
1732a1fd
JH
1055/*
1056 * For lines target is suspected for, see if we can find code movement
1057 * across file boundary from the parent commit. porigin is the path
1058 * in the parent we already tried.
1059 */
18abd745
JH
1060static int find_copy_in_parent(struct scoreboard *sb,
1061 struct origin *target,
1062 struct commit *parent,
1063 struct origin *porigin,
1064 int opt)
1065{
1066 struct diff_options diff_opts;
1067 const char *paths[1];
aec8fa1f 1068 int i, j;
33494784
JH
1069 int retval;
1070 struct blame_list *blame_list;
aec8fa1f 1071 int num_ents;
18abd745 1072
1617fe98 1073 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784 1074 if (!blame_list)
18abd745
JH
1075 return 1; /* nothing remains for this target */
1076
1077 diff_setup(&diff_opts);
8f67f8ae 1078 DIFF_OPT_SET(&diff_opts, RECURSIVE);
18abd745
JH
1079 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
1080
18abd745
JH
1081 paths[0] = NULL;
1082 diff_tree_setup_paths(paths, &diff_opts);
28452655 1083 diff_setup_done(&diff_opts);
33494784
JH
1084
1085 /* Try "find copies harder" on new path if requested;
1086 * we do not want to use diffcore_rename() actually to
1087 * match things up; find_copies_harder is set only to
1088 * force diff_tree_sha1() to feed all filepairs to diff_queue,
1089 * and this code needs to be after diff_setup_done(), which
1090 * usually makes find-copies-harder imply copy detection.
1091 */
c63777c0
JH
1092 if ((opt & PICKAXE_BLAME_COPY_HARDEST)
1093 || ((opt & PICKAXE_BLAME_COPY_HARDER)
1094 && (!porigin || strcmp(target->path, porigin->path))))
8f67f8ae 1095 DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
33494784 1096
1cfe7733
JH
1097 if (is_null_sha1(target->commit->object.sha1))
1098 do_diff_cache(parent->tree->object.sha1, &diff_opts);
1099 else
1100 diff_tree_sha1(parent->tree->object.sha1,
1101 target->commit->tree->object.sha1,
1102 "", &diff_opts);
18abd745 1103
8f67f8ae 1104 if (!DIFF_OPT_TST(&diff_opts, FIND_COPIES_HARDER))
33494784 1105 diffcore_std(&diff_opts);
18abd745 1106
33494784
JH
1107 retval = 0;
1108 while (1) {
1109 int made_progress = 0;
1110
1111 for (i = 0; i < diff_queued_diff.nr; i++) {
1112 struct diff_filepair *p = diff_queued_diff.queue[i];
1113 struct origin *norigin;
1114 mmfile_t file_p;
33494784
JH
1115 struct blame_entry this[3];
1116
1117 if (!DIFF_FILE_VALID(p->one))
1118 continue; /* does not exist in parent */
5209ac4d
AG
1119 if (S_ISGITLINK(p->one->mode))
1120 continue; /* ignore git links */
33494784
JH
1121 if (porigin && !strcmp(p->one->path, porigin->path))
1122 /* find_move already dealt with this path */
1123 continue;
1124
1125 norigin = get_origin(sb, parent, p->one->path);
1126 hashcpy(norigin->blob_sha1, p->one->sha1);
90064710 1127 norigin->mode = p->one->mode;
3b8a12e8 1128 fill_origin_blob(&sb->revs->diffopt, norigin, &file_p);
33494784
JH
1129 if (!file_p.ptr)
1130 continue;
1131
1132 for (j = 0; j < num_ents; j++) {
1133 find_copy_in_blob(sb, blame_list[j].ent,
1134 norigin, this, &file_p);
1135 copy_split_if_better(sb, blame_list[j].split,
1136 this);
1137 decref_split(this);
1138 }
33494784
JH
1139 origin_decref(norigin);
1140 }
18abd745 1141
aec8fa1f 1142 for (j = 0; j < num_ents; j++) {
33494784
JH
1143 struct blame_entry *split = blame_list[j].split;
1144 if (split[1].suspect &&
1145 blame_copy_score < ent_score(sb, &split[1])) {
1146 split_blame(sb, split, blame_list[j].ent);
1147 made_progress = 1;
1148 }
eb46328e
AG
1149 else
1150 blame_list[j].ent->scanned = 1;
33494784 1151 decref_split(split);
18abd745 1152 }
33494784 1153 free(blame_list);
aec8fa1f 1154
33494784
JH
1155 if (!made_progress)
1156 break;
1617fe98 1157 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784
JH
1158 if (!blame_list) {
1159 retval = 1;
1160 break;
1161 }
18abd745 1162 }
eb46328e 1163 reset_scanned_flag(sb);
33494784 1164 diff_flush(&diff_opts);
03b69c76 1165 diff_tree_release_paths(&diff_opts);
33494784 1166 return retval;
18abd745
JH
1167}
1168
1732a1fd
JH
1169/*
1170 * The blobs of origin and porigin exactly match, so everything
c2e525d9
JH
1171 * origin is suspected for can be blamed on the parent.
1172 */
1173static void pass_whole_blame(struct scoreboard *sb,
1174 struct origin *origin, struct origin *porigin)
1175{
1176 struct blame_entry *e;
1177
1178 if (!porigin->file.ptr && origin->file.ptr) {
1179 /* Steal its file */
1180 porigin->file = origin->file;
1181 origin->file.ptr = NULL;
1182 }
1183 for (e = sb->ent; e; e = e->next) {
171dccd5 1184 if (!same_suspect(e->suspect, origin))
c2e525d9
JH
1185 continue;
1186 origin_incref(porigin);
1187 origin_decref(e->suspect);
1188 e->suspect = porigin;
1189 }
1190}
1191
69264f46
JH
1192/*
1193 * We pass blame from the current commit to its parents. We keep saying
1194 * "parent" (and "porigin"), but what we mean is to find scapegoat to
1195 * exonerate ourselves.
1196 */
85af7929 1197static struct commit_list *first_scapegoat(struct rev_info *revs, struct commit *commit)
69264f46 1198{
85af7929
JH
1199 if (!reverse)
1200 return commit->parents;
1201 return lookup_decoration(&revs->children, &commit->object);
69264f46
JH
1202}
1203
85af7929 1204static int num_scapegoats(struct rev_info *revs, struct commit *commit)
69264f46
JH
1205{
1206 int cnt;
85af7929 1207 struct commit_list *l = first_scapegoat(revs, commit);
69264f46
JH
1208 for (cnt = 0; l; l = l->next)
1209 cnt++;
1210 return cnt;
1211}
1212
1213#define MAXSG 16
cee7f245 1214
d24bba80 1215static void pass_blame(struct scoreboard *sb, struct origin *origin, int opt)
cee7f245 1216{
85af7929 1217 struct rev_info *revs = sb->revs;
69264f46 1218 int i, pass, num_sg;
cee7f245 1219 struct commit *commit = origin->commit;
69264f46
JH
1220 struct commit_list *sg;
1221 struct origin *sg_buf[MAXSG];
1222 struct origin *porigin, **sg_origin = sg_buf;
1223
85af7929 1224 num_sg = num_scapegoats(revs, commit);
69264f46
JH
1225 if (!num_sg)
1226 goto finish;
1227 else if (num_sg < ARRAY_SIZE(sg_buf))
1228 memset(sg_buf, 0, sizeof(sg_buf));
1229 else
1230 sg_origin = xcalloc(num_sg, sizeof(*sg_origin));
cee7f245 1231
69264f46
JH
1232 /*
1233 * The first pass looks for unrenamed path to optimize for
f69e743d
JH
1234 * common cases, then we look for renames in the second pass.
1235 */
3d1aa566 1236 for (pass = 0; pass < 2 - no_whole_file_rename; pass++) {
f69e743d
JH
1237 struct origin *(*find)(struct scoreboard *,
1238 struct commit *, struct origin *);
1239 find = pass ? find_rename : find_origin;
1240
85af7929 1241 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1242 i < num_sg && sg;
1243 sg = sg->next, i++) {
1244 struct commit *p = sg->item;
0421d9f8 1245 int j, same;
f69e743d 1246
69264f46 1247 if (sg_origin[i])
f69e743d
JH
1248 continue;
1249 if (parse_commit(p))
1250 continue;
0d981c67 1251 porigin = find(sb, p, origin);
f69e743d
JH
1252 if (!porigin)
1253 continue;
1254 if (!hashcmp(porigin->blob_sha1, origin->blob_sha1)) {
c2e525d9 1255 pass_whole_blame(sb, origin, porigin);
f69e743d
JH
1256 origin_decref(porigin);
1257 goto finish;
1258 }
0421d9f8 1259 for (j = same = 0; j < i; j++)
69264f46
JH
1260 if (sg_origin[j] &&
1261 !hashcmp(sg_origin[j]->blob_sha1,
0421d9f8
JH
1262 porigin->blob_sha1)) {
1263 same = 1;
1264 break;
1265 }
1266 if (!same)
69264f46 1267 sg_origin[i] = porigin;
0421d9f8
JH
1268 else
1269 origin_decref(porigin);
cee7f245 1270 }
cee7f245
JH
1271 }
1272
c2e525d9 1273 num_commits++;
85af7929 1274 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1275 i < num_sg && sg;
1276 sg = sg->next, i++) {
1277 struct origin *porigin = sg_origin[i];
cee7f245
JH
1278 if (!porigin)
1279 continue;
96e11709
JH
1280 if (!origin->previous) {
1281 origin_incref(porigin);
1282 origin->previous = porigin;
1283 }
cee7f245 1284 if (pass_blame_to_parent(sb, origin, porigin))
54a4c617 1285 goto finish;
cee7f245 1286 }
d24bba80
JH
1287
1288 /*
1732a1fd 1289 * Optionally find moves in parents' files.
d24bba80
JH
1290 */
1291 if (opt & PICKAXE_BLAME_MOVE)
85af7929 1292 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1293 i < num_sg && sg;
1294 sg = sg->next, i++) {
1295 struct origin *porigin = sg_origin[i];
d24bba80
JH
1296 if (!porigin)
1297 continue;
1298 if (find_move_in_parent(sb, origin, porigin))
54a4c617 1299 goto finish;
d24bba80
JH
1300 }
1301
18abd745 1302 /*
1732a1fd 1303 * Optionally find copies from parents' files.
18abd745
JH
1304 */
1305 if (opt & PICKAXE_BLAME_COPY)
85af7929 1306 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1307 i < num_sg && sg;
1308 sg = sg->next, i++) {
1309 struct origin *porigin = sg_origin[i];
1310 if (find_copy_in_parent(sb, origin, sg->item,
18abd745 1311 porigin, opt))
54a4c617 1312 goto finish;
18abd745 1313 }
54a4c617
JH
1314
1315 finish:
69264f46
JH
1316 for (i = 0; i < num_sg; i++) {
1317 if (sg_origin[i]) {
1318 drop_origin_blob(sg_origin[i]);
1319 origin_decref(sg_origin[i]);
7c3c7962
JH
1320 }
1321 }
1322 drop_origin_blob(origin);
69264f46
JH
1323 if (sg_buf != sg_origin)
1324 free(sg_origin);
cee7f245
JH
1325}
1326
1732a1fd
JH
1327/*
1328 * Information on commits, used for output.
1329 */
9cba13ca 1330struct commit_info {
ea02ffa3
AP
1331 struct strbuf author;
1332 struct strbuf author_mail;
cee7f245 1333 unsigned long author_time;
ea02ffa3 1334 struct strbuf author_tz;
cee7f245
JH
1335
1336 /* filled only when asked for details */
ea02ffa3
AP
1337 struct strbuf committer;
1338 struct strbuf committer_mail;
cee7f245 1339 unsigned long committer_time;
ea02ffa3 1340 struct strbuf committer_tz;
cee7f245 1341
ea02ffa3 1342 struct strbuf summary;
cee7f245
JH
1343};
1344
1732a1fd
JH
1345/*
1346 * Parse author/committer line in the commit object buffer
1347 */
cee7f245 1348static void get_ac_line(const char *inbuf, const char *what,
ea02ffa3
AP
1349 struct strbuf *name, struct strbuf *mail,
1350 unsigned long *time, struct strbuf *tz)
cee7f245 1351{
3c020bd5 1352 struct ident_split ident;
ea02ffa3
AP
1353 size_t len, maillen, namelen;
1354 char *tmp, *endp;
1355 const char *namebuf, *mailbuf;
cee7f245
JH
1356
1357 tmp = strstr(inbuf, what);
1358 if (!tmp)
1359 goto error_out;
1360 tmp += strlen(what);
1361 endp = strchr(tmp, '\n');
1362 if (!endp)
1363 len = strlen(tmp);
1364 else
1365 len = endp - tmp;
3c020bd5
AP
1366
1367 if (split_ident_line(&ident, tmp, len)) {
cee7f245
JH
1368 error_out:
1369 /* Ugh */
ea02ffa3
AP
1370 tmp = "(unknown)";
1371 strbuf_addstr(name, tmp);
1372 strbuf_addstr(mail, tmp);
1373 strbuf_addstr(tz, tmp);
cee7f245
JH
1374 *time = 0;
1375 return;
1376 }
cee7f245 1377
3c020bd5 1378 namelen = ident.name_end - ident.name_begin;
ea02ffa3 1379 namebuf = ident.name_begin;
cee7f245 1380
ea02ffa3
AP
1381 maillen = ident.mail_end - ident.mail_begin;
1382 mailbuf = ident.mail_begin;
f9567384 1383
de5abe9f
RS
1384 if (ident.date_begin && ident.date_end)
1385 *time = strtoul(ident.date_begin, NULL, 10);
1386 else
1387 *time = 0;
f9567384 1388
de5abe9f
RS
1389 if (ident.tz_begin && ident.tz_end)
1390 strbuf_add(tz, ident.tz_begin, ident.tz_end - ident.tz_begin);
1391 else
1392 strbuf_addstr(tz, "(unknown)");
f9567384 1393
f9567384 1394 /*
d20d654f 1395 * Now, convert both name and e-mail using mailmap
f9567384 1396 */
ea02ffa3
AP
1397 map_user(&mailmap, &mailbuf, &maillen,
1398 &namebuf, &namelen);
1399
1400 strbuf_addf(mail, "<%.*s>", (int)maillen, mailbuf);
1401 strbuf_add(name, namebuf, namelen);
1402}
1403
1404static void commit_info_init(struct commit_info *ci)
1405{
1406
1407 strbuf_init(&ci->author, 0);
1408 strbuf_init(&ci->author_mail, 0);
1409 strbuf_init(&ci->author_tz, 0);
1410 strbuf_init(&ci->committer, 0);
1411 strbuf_init(&ci->committer_mail, 0);
1412 strbuf_init(&ci->committer_tz, 0);
1413 strbuf_init(&ci->summary, 0);
1414}
1415
1416static void commit_info_destroy(struct commit_info *ci)
1417{
1418
1419 strbuf_release(&ci->author);
1420 strbuf_release(&ci->author_mail);
1421 strbuf_release(&ci->author_tz);
1422 strbuf_release(&ci->committer);
1423 strbuf_release(&ci->committer_mail);
1424 strbuf_release(&ci->committer_tz);
1425 strbuf_release(&ci->summary);
cee7f245
JH
1426}
1427
1428static void get_commit_info(struct commit *commit,
1429 struct commit_info *ret,
1430 int detailed)
1431{
1432 int len;
e297cf5a 1433 const char *subject, *encoding;
dd0d388c 1434 char *message;
ea02ffa3
AP
1435
1436 commit_info_init(ret);
cee7f245 1437
e297cf5a 1438 encoding = get_log_output_encoding();
5a10d236 1439 message = logmsg_reencode(commit, NULL, encoding);
69cd8f63 1440 get_ac_line(message, "\nauthor ",
ea02ffa3 1441 &ret->author, &ret->author_mail,
cee7f245
JH
1442 &ret->author_time, &ret->author_tz);
1443
69cd8f63 1444 if (!detailed) {
dd0d388c 1445 logmsg_free(message, commit);
cee7f245 1446 return;
69cd8f63 1447 }
cee7f245 1448
69cd8f63 1449 get_ac_line(message, "\ncommitter ",
ea02ffa3 1450 &ret->committer, &ret->committer_mail,
cee7f245
JH
1451 &ret->committer_time, &ret->committer_tz);
1452
ad98a58b 1453 len = find_commit_subject(message, &subject);
ea02ffa3
AP
1454 if (len)
1455 strbuf_add(&ret->summary, subject, len);
1456 else
1457 strbuf_addf(&ret->summary, "(%s)", sha1_to_hex(commit->object.sha1));
1458
dd0d388c 1459 logmsg_free(message, commit);
cee7f245
JH
1460}
1461
1732a1fd
JH
1462/*
1463 * To allow LF and other nonportable characters in pathnames,
1464 * they are c-style quoted as needed.
1465 */
46e5e69d
JH
1466static void write_filename_info(const char *path)
1467{
1468 printf("filename ");
663af342 1469 write_name_quoted(path, stdout, '\n');
46e5e69d
JH
1470}
1471
9991030c
JH
1472/*
1473 * Porcelain/Incremental format wants to show a lot of details per
1474 * commit. Instead of repeating this every line, emit it only once,
e86226e3
JK
1475 * the first time each commit appears in the output (unless the
1476 * user has specifically asked for us to repeat).
9991030c 1477 */
e86226e3 1478static int emit_one_suspect_detail(struct origin *suspect, int repeat)
9991030c
JH
1479{
1480 struct commit_info ci;
1481
e86226e3 1482 if (!repeat && (suspect->commit->object.flags & METAINFO_SHOWN))
9991030c
JH
1483 return 0;
1484
1485 suspect->commit->object.flags |= METAINFO_SHOWN;
1486 get_commit_info(suspect->commit, &ci, 1);
ea02ffa3
AP
1487 printf("author %s\n", ci.author.buf);
1488 printf("author-mail %s\n", ci.author_mail.buf);
9991030c 1489 printf("author-time %lu\n", ci.author_time);
ea02ffa3
AP
1490 printf("author-tz %s\n", ci.author_tz.buf);
1491 printf("committer %s\n", ci.committer.buf);
1492 printf("committer-mail %s\n", ci.committer_mail.buf);
9991030c 1493 printf("committer-time %lu\n", ci.committer_time);
ea02ffa3
AP
1494 printf("committer-tz %s\n", ci.committer_tz.buf);
1495 printf("summary %s\n", ci.summary.buf);
9991030c
JH
1496 if (suspect->commit->object.flags & UNINTERESTING)
1497 printf("boundary\n");
96e11709
JH
1498 if (suspect->previous) {
1499 struct origin *prev = suspect->previous;
1500 printf("previous %s ", sha1_to_hex(prev->commit->object.sha1));
1501 write_name_quoted(prev->path, stdout, '\n');
1502 }
ea02ffa3
AP
1503
1504 commit_info_destroy(&ci);
1505
9991030c
JH
1506 return 1;
1507}
1508
1732a1fd
JH
1509/*
1510 * The blame_entry is found to be guilty for the range. Mark it
1511 * as such, and show it in incremental output.
1512 */
717d1462
LT
1513static void found_guilty_entry(struct blame_entry *ent)
1514{
1515 if (ent->guilty)
1516 return;
1517 ent->guilty = 1;
1518 if (incremental) {
1519 struct origin *suspect = ent->suspect;
1520
1521 printf("%s %d %d %d\n",
1522 sha1_to_hex(suspect->commit->object.sha1),
1523 ent->s_lno + 1, ent->lno + 1, ent->num_lines);
e86226e3 1524 emit_one_suspect_detail(suspect, 0);
46e5e69d 1525 write_filename_info(suspect->path);
06f59e9f 1526 maybe_flush_or_die(stdout, "stdout");
717d1462
LT
1527 }
1528}
1529
1732a1fd
JH
1530/*
1531 * The main loop -- while the scoreboard has lines whose true origin
3dff5379 1532 * is still unknown, pick one blame_entry, and allow its current
1732a1fd
JH
1533 * suspect to pass blames to its parents.
1534 */
85af7929 1535static void assign_blame(struct scoreboard *sb, int opt)
717d1462 1536{
85af7929
JH
1537 struct rev_info *revs = sb->revs;
1538
717d1462
LT
1539 while (1) {
1540 struct blame_entry *ent;
1541 struct commit *commit;
1542 struct origin *suspect = NULL;
1543
1544 /* find one suspect to break down */
1545 for (ent = sb->ent; !suspect && ent; ent = ent->next)
1546 if (!ent->guilty)
1547 suspect = ent->suspect;
1548 if (!suspect)
1549 return; /* all done */
1550
1732a1fd
JH
1551 /*
1552 * We will use this suspect later in the loop,
1553 * so hold onto it in the meantime.
1554 */
717d1462
LT
1555 origin_incref(suspect);
1556 commit = suspect->commit;
1557 if (!commit->object.parsed)
1558 parse_commit(commit);
85af7929
JH
1559 if (reverse ||
1560 (!(commit->object.flags & UNINTERESTING) &&
1561 !(revs->max_age != -1 && commit->date < revs->max_age)))
717d1462
LT
1562 pass_blame(sb, suspect, opt);
1563 else {
1564 commit->object.flags |= UNINTERESTING;
1565 if (commit->object.parsed)
1566 mark_parents_uninteresting(commit);
1567 }
1568 /* treat root commit as boundary */
1569 if (!commit->parents && !show_root)
1570 commit->object.flags |= UNINTERESTING;
1571
1572 /* Take responsibility for the remaining entries */
1573 for (ent = sb->ent; ent; ent = ent->next)
171dccd5 1574 if (same_suspect(ent->suspect, suspect))
717d1462
LT
1575 found_guilty_entry(ent);
1576 origin_decref(suspect);
1577
1578 if (DEBUG) /* sanity */
1579 sanity_check_refcnt(sb);
1580 }
1581}
1582
1583static const char *format_time(unsigned long time, const char *tz_str,
1584 int show_raw_time)
1585{
1586 static char time_buf[128];
31653c1a
EL
1587 const char *time_str;
1588 int time_len;
1589 int tz;
717d1462
LT
1590
1591 if (show_raw_time) {
c3ea0515 1592 snprintf(time_buf, sizeof(time_buf), "%lu %s", time, tz_str);
717d1462 1593 }
31653c1a
EL
1594 else {
1595 tz = atoi(tz_str);
1596 time_str = show_date(time, tz, blame_date_mode);
1597 time_len = strlen(time_str);
1598 memcpy(time_buf, time_str, time_len);
1599 memset(time_buf + time_len, ' ', blame_date_width - time_len);
1600 }
717d1462
LT
1601 return time_buf;
1602}
1603
cee7f245
JH
1604#define OUTPUT_ANNOTATE_COMPAT 001
1605#define OUTPUT_LONG_OBJECT_NAME 002
1606#define OUTPUT_RAW_TIMESTAMP 004
1607#define OUTPUT_PORCELAIN 010
1608#define OUTPUT_SHOW_NAME 020
1609#define OUTPUT_SHOW_NUMBER 040
5ff62c30 1610#define OUTPUT_SHOW_SCORE 0100
093dc5be 1611#define OUTPUT_NO_AUTHOR 0200
1b8cdce9 1612#define OUTPUT_SHOW_EMAIL 0400
ed747dd5 1613#define OUTPUT_LINE_PORCELAIN 01000
cee7f245 1614
e86226e3
JK
1615static void emit_porcelain_details(struct origin *suspect, int repeat)
1616{
1617 if (emit_one_suspect_detail(suspect, repeat) ||
1618 (suspect->commit->object.flags & MORE_THAN_ONE_PATH))
1619 write_filename_info(suspect->path);
1620}
1621
1622static void emit_porcelain(struct scoreboard *sb, struct blame_entry *ent,
1623 int opt)
cee7f245 1624{
ed747dd5 1625 int repeat = opt & OUTPUT_LINE_PORCELAIN;
cee7f245
JH
1626 int cnt;
1627 const char *cp;
1628 struct origin *suspect = ent->suspect;
1629 char hex[41];
1630
1631 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1632 printf("%s%c%d %d %d\n",
1633 hex,
2543d9b6 1634 ent->guilty ? ' ' : '*', /* purely for debugging */
cee7f245
JH
1635 ent->s_lno + 1,
1636 ent->lno + 1,
1637 ent->num_lines);
ed747dd5 1638 emit_porcelain_details(suspect, repeat);
cee7f245
JH
1639
1640 cp = nth_line(sb, ent->lno);
1641 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1642 char ch;
ed747dd5 1643 if (cnt) {
cee7f245
JH
1644 printf("%s %d %d\n", hex,
1645 ent->s_lno + 1 + cnt,
1646 ent->lno + 1 + cnt);
ed747dd5
JK
1647 if (repeat)
1648 emit_porcelain_details(suspect, 1);
1649 }
cee7f245
JH
1650 putchar('\t');
1651 do {
1652 ch = *cp++;
1653 putchar(ch);
1654 } while (ch != '\n' &&
1655 cp < sb->final_buf + sb->final_buf_size);
1656 }
a5ca8367
JS
1657
1658 if (sb->final_buf_size && cp[-1] != '\n')
1659 putchar('\n');
cee7f245
JH
1660}
1661
1662static void emit_other(struct scoreboard *sb, struct blame_entry *ent, int opt)
1663{
1664 int cnt;
1665 const char *cp;
1666 struct origin *suspect = ent->suspect;
1667 struct commit_info ci;
1668 char hex[41];
1669 int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP);
1670
1671 get_commit_info(suspect->commit, &ci, 1);
1672 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1673
1674 cp = nth_line(sb, ent->lno);
1675 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1676 char ch;
84393bfd 1677 int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? 40 : abbrev;
b11121d9
JH
1678
1679 if (suspect->commit->object.flags & UNINTERESTING) {
e68989a7
JH
1680 if (blank_boundary)
1681 memset(hex, ' ', length);
7ceacdff 1682 else if (!(opt & OUTPUT_ANNOTATE_COMPAT)) {
4c10a5ca
JH
1683 length--;
1684 putchar('^');
1685 }
b11121d9 1686 }
cee7f245 1687
b11121d9 1688 printf("%.*s", length, hex);
1b8cdce9
KB
1689 if (opt & OUTPUT_ANNOTATE_COMPAT) {
1690 const char *name;
1691 if (opt & OUTPUT_SHOW_EMAIL)
ea02ffa3 1692 name = ci.author_mail.buf;
1b8cdce9 1693 else
ea02ffa3 1694 name = ci.author.buf;
1b8cdce9 1695 printf("\t(%10s\t%10s\t%d)", name,
ea02ffa3 1696 format_time(ci.author_time, ci.author_tz.buf,
cee7f245
JH
1697 show_raw_time),
1698 ent->lno + 1 + cnt);
1b8cdce9 1699 } else {
5ff62c30 1700 if (opt & OUTPUT_SHOW_SCORE)
54a4c617
JH
1701 printf(" %*d %02d",
1702 max_score_digits, ent->score,
1703 ent->suspect->refcnt);
cee7f245
JH
1704 if (opt & OUTPUT_SHOW_NAME)
1705 printf(" %-*.*s", longest_file, longest_file,
1706 suspect->path);
1707 if (opt & OUTPUT_SHOW_NUMBER)
1708 printf(" %*d", max_orig_digits,
1709 ent->s_lno + 1 + cnt);
093dc5be 1710
ffaf9cc0 1711 if (!(opt & OUTPUT_NO_AUTHOR)) {
1b8cdce9
KB
1712 const char *name;
1713 int pad;
1714 if (opt & OUTPUT_SHOW_EMAIL)
ea02ffa3 1715 name = ci.author_mail.buf;
1b8cdce9 1716 else
ea02ffa3 1717 name = ci.author.buf;
1b8cdce9 1718 pad = longest_author - utf8_strwidth(name);
ffaf9cc0 1719 printf(" (%s%*s %10s",
1b8cdce9 1720 name, pad, "",
093dc5be 1721 format_time(ci.author_time,
ea02ffa3 1722 ci.author_tz.buf,
093dc5be 1723 show_raw_time));
ffaf9cc0 1724 }
093dc5be 1725 printf(" %*d) ",
cee7f245
JH
1726 max_digits, ent->lno + 1 + cnt);
1727 }
1728 do {
1729 ch = *cp++;
1730 putchar(ch);
1731 } while (ch != '\n' &&
1732 cp < sb->final_buf + sb->final_buf_size);
1733 }
a5ca8367
JS
1734
1735 if (sb->final_buf_size && cp[-1] != '\n')
1736 putchar('\n');
ea02ffa3
AP
1737
1738 commit_info_destroy(&ci);
cee7f245
JH
1739}
1740
1741static void output(struct scoreboard *sb, int option)
1742{
1743 struct blame_entry *ent;
1744
1745 if (option & OUTPUT_PORCELAIN) {
1746 for (ent = sb->ent; ent; ent = ent->next) {
1747 struct blame_entry *oth;
1748 struct origin *suspect = ent->suspect;
1749 struct commit *commit = suspect->commit;
1750 if (commit->object.flags & MORE_THAN_ONE_PATH)
1751 continue;
1752 for (oth = ent->next; oth; oth = oth->next) {
1753 if ((oth->suspect->commit != commit) ||
1754 !strcmp(oth->suspect->path, suspect->path))
1755 continue;
1756 commit->object.flags |= MORE_THAN_ONE_PATH;
1757 break;
1758 }
1759 }
1760 }
1761
1762 for (ent = sb->ent; ent; ent = ent->next) {
1763 if (option & OUTPUT_PORCELAIN)
e86226e3 1764 emit_porcelain(sb, ent, option);
5ff62c30 1765 else {
cee7f245 1766 emit_other(sb, ent, option);
5ff62c30 1767 }
cee7f245
JH
1768 }
1769}
1770
1732a1fd
JH
1771/*
1772 * To allow quick access to the contents of nth line in the
1773 * final image, prepare an index in the scoreboard.
1774 */
cee7f245
JH
1775static int prepare_lines(struct scoreboard *sb)
1776{
1777 const char *buf = sb->final_buf;
1778 unsigned long len = sb->final_buf_size;
1779 int num = 0, incomplete = 0, bol = 1;
1780
1781 if (len && buf[len-1] != '\n')
1782 incomplete++; /* incomplete line at the end */
1783 while (len--) {
1784 if (bol) {
1785 sb->lineno = xrealloc(sb->lineno,
4b25d091 1786 sizeof(int *) * (num + 1));
cee7f245
JH
1787 sb->lineno[num] = buf - sb->final_buf;
1788 bol = 0;
1789 }
1790 if (*buf++ == '\n') {
1791 num++;
1792 bol = 1;
1793 }
1794 }
1ca6ca87 1795 sb->lineno = xrealloc(sb->lineno,
4b25d091 1796 sizeof(int *) * (num + incomplete + 1));
1ca6ca87 1797 sb->lineno[num + incomplete] = buf - sb->final_buf;
cee7f245
JH
1798 sb->num_lines = num + incomplete;
1799 return sb->num_lines;
1800}
1801
1732a1fd
JH
1802/*
1803 * Add phony grafts for use with -S; this is primarily to
34baebce 1804 * support git's cvsserver that wants to give a linear history
1732a1fd
JH
1805 * to its clients.
1806 */
cee7f245
JH
1807static int read_ancestry(const char *graft_file)
1808{
1809 FILE *fp = fopen(graft_file, "r");
1810 char buf[1024];
1811 if (!fp)
1812 return -1;
1813 while (fgets(buf, sizeof(buf), fp)) {
1814 /* The format is just "Commit Parent1 Parent2 ...\n" */
1815 int len = strlen(buf);
1816 struct commit_graft *graft = read_graft_line(buf, len);
8eaf7986
JH
1817 if (graft)
1818 register_commit_graft(graft, 0);
cee7f245
JH
1819 }
1820 fclose(fp);
1821 return 0;
1822}
1823
b31272f7
JH
1824static int update_auto_abbrev(int auto_abbrev, struct origin *suspect)
1825{
1826 const char *uniq = find_unique_abbrev(suspect->commit->object.sha1,
1827 auto_abbrev);
1828 int len = strlen(uniq);
1829 if (auto_abbrev < len)
1830 return len;
1831 return auto_abbrev;
1832}
1833
1732a1fd
JH
1834/*
1835 * How many columns do we need to show line numbers, authors,
1836 * and filenames?
1837 */
cee7f245
JH
1838static void find_alignment(struct scoreboard *sb, int *option)
1839{
1840 int longest_src_lines = 0;
1841 int longest_dst_lines = 0;
5ff62c30 1842 unsigned largest_score = 0;
cee7f245 1843 struct blame_entry *e;
b31272f7
JH
1844 int compute_auto_abbrev = (abbrev < 0);
1845 int auto_abbrev = default_abbrev;
cee7f245
JH
1846
1847 for (e = sb->ent; e; e = e->next) {
1848 struct origin *suspect = e->suspect;
1849 struct commit_info ci;
1850 int num;
1851
b31272f7
JH
1852 if (compute_auto_abbrev)
1853 auto_abbrev = update_auto_abbrev(auto_abbrev, suspect);
ab3bb800
JH
1854 if (strcmp(suspect->path, sb->path))
1855 *option |= OUTPUT_SHOW_NAME;
1856 num = strlen(suspect->path);
1857 if (longest_file < num)
1858 longest_file = num;
cee7f245
JH
1859 if (!(suspect->commit->object.flags & METAINFO_SHOWN)) {
1860 suspect->commit->object.flags |= METAINFO_SHOWN;
1861 get_commit_info(suspect->commit, &ci, 1);
1b8cdce9 1862 if (*option & OUTPUT_SHOW_EMAIL)
ea02ffa3 1863 num = utf8_strwidth(ci.author_mail.buf);
1b8cdce9 1864 else
ea02ffa3 1865 num = utf8_strwidth(ci.author.buf);
cee7f245
JH
1866 if (longest_author < num)
1867 longest_author = num;
1868 }
1869 num = e->s_lno + e->num_lines;
1870 if (longest_src_lines < num)
1871 longest_src_lines = num;
1872 num = e->lno + e->num_lines;
1873 if (longest_dst_lines < num)
1874 longest_dst_lines = num;
5ff62c30
JH
1875 if (largest_score < ent_score(sb, e))
1876 largest_score = ent_score(sb, e);
ea02ffa3
AP
1877
1878 commit_info_destroy(&ci);
cee7f245 1879 }
ec7ff5ba
ZJS
1880 max_orig_digits = decimal_width(longest_src_lines);
1881 max_digits = decimal_width(longest_dst_lines);
1882 max_score_digits = decimal_width(largest_score);
b31272f7
JH
1883
1884 if (compute_auto_abbrev)
1885 /* one more abbrev length is needed for the boundary commit */
1886 abbrev = auto_abbrev + 1;
cee7f245
JH
1887}
1888
1732a1fd
JH
1889/*
1890 * For debugging -- origin is refcounted, and this asserts that
1891 * we do not underflow.
1892 */
54a4c617
JH
1893static void sanity_check_refcnt(struct scoreboard *sb)
1894{
1895 int baa = 0;
1896 struct blame_entry *ent;
1897
1898 for (ent = sb->ent; ent; ent = ent->next) {
ae86ad65 1899 /* Nobody should have zero or negative refcnt */
854b97f6
JH
1900 if (ent->suspect->refcnt <= 0) {
1901 fprintf(stderr, "%s in %s has negative refcnt %d\n",
1902 ent->suspect->path,
1903 sha1_to_hex(ent->suspect->commit->object.sha1),
1904 ent->suspect->refcnt);
ae86ad65 1905 baa = 1;
854b97f6 1906 }
ae86ad65 1907 }
54a4c617
JH
1908 if (baa) {
1909 int opt = 0160;
1910 find_alignment(sb, &opt);
1911 output(sb, opt);
854b97f6 1912 die("Baa %d!", baa);
54a4c617
JH
1913 }
1914}
1915
1732a1fd
JH
1916/*
1917 * Used for the command line parsing; check if the path exists
1918 * in the working tree.
1919 */
c455c87c 1920static int has_string_in_work_tree(const char *path)
cee7f245
JH
1921{
1922 struct stat st;
1923 return !lstat(path, &st);
1924}
1925
4a0fc95f
JH
1926static unsigned parse_score(const char *arg)
1927{
1928 char *end;
1929 unsigned long score = strtoul(arg, &end, 10);
1930 if (*end)
1931 return 0;
1932 return score;
1933}
1934
20239bae
JK
1935static const char *add_prefix(const char *prefix, const char *path)
1936{
097971f5 1937 return prefix_path(prefix, prefix ? strlen(prefix) : 0, path);
20239bae
JK
1938}
1939
ef90d6d4 1940static int git_blame_config(const char *var, const char *value, void *cb)
4c10a5ca
JH
1941{
1942 if (!strcmp(var, "blame.showroot")) {
1943 show_root = git_config_bool(var, value);
1944 return 0;
1945 }
1946 if (!strcmp(var, "blame.blankboundary")) {
1947 blank_boundary = git_config_bool(var, value);
1948 return 0;
1949 }
31653c1a
EL
1950 if (!strcmp(var, "blame.date")) {
1951 if (!value)
1952 return config_error_nonbool(var);
1953 blame_date_mode = parse_date_format(value);
1954 return 0;
1955 }
3b8a12e8 1956
6680a087 1957 if (userdiff_config(var, value) < 0)
3b8a12e8 1958 return -1;
3b8a12e8 1959
ef90d6d4 1960 return git_default_config(var, value, cb);
4c10a5ca
JH
1961}
1962
9aeaab68 1963static void verify_working_tree_path(struct commit *work_tree, const char *path)
ffcabccf 1964{
9aeaab68 1965 struct commit_list *parents;
ffcabccf 1966
9aeaab68
JH
1967 for (parents = work_tree->parents; parents; parents = parents->next) {
1968 const unsigned char *commit_sha1 = parents->item->object.sha1;
1969 unsigned char blob_sha1[20];
1970 unsigned mode;
1971
1972 if (!get_tree_entry(commit_sha1, path, blob_sha1, &mode) &&
1973 sha1_object_info(blob_sha1, NULL) == OBJ_BLOB)
1974 return;
1975 }
1976 die("no such path '%s' in HEAD", path);
1977}
1978
1979static struct commit_list **append_parent(struct commit_list **tail, const unsigned char *sha1)
1980{
1981 struct commit *parent;
1982
1983 parent = lookup_commit_reference(sha1);
1984 if (!parent)
1985 die("no such commit %s", sha1_to_hex(sha1));
1986 return &commit_list_insert(parent, tail)->next;
1987}
1988
1989static void append_merge_parents(struct commit_list **tail)
1990{
1991 int merge_head;
1992 const char *merge_head_file = git_path("MERGE_HEAD");
1993 struct strbuf line = STRBUF_INIT;
1994
1995 merge_head = open(merge_head_file, O_RDONLY);
1996 if (merge_head < 0) {
1997 if (errno == ENOENT)
1998 return;
1999 die("cannot open '%s' for reading", merge_head_file);
2000 }
2001
2002 while (!strbuf_getwholeline_fd(&line, merge_head, '\n')) {
2003 unsigned char sha1[20];
2004 if (line.len < 40 || get_sha1_hex(line.buf, sha1))
2005 die("unknown line in '%s': %s", merge_head_file, line.buf);
2006 tail = append_parent(tail, sha1);
2007 }
2008 close(merge_head);
2009 strbuf_release(&line);
ffcabccf
JH
2010}
2011
f6c07d7d
JH
2012/*
2013 * Prepare a dummy commit that represents the work tree (or staged) item.
2014 * Note that annotating work tree item never works in the reverse.
2015 */
3b8a12e8
AB
2016static struct commit *fake_working_tree_commit(struct diff_options *opt,
2017 const char *path,
2018 const char *contents_from)
1cfe7733
JH
2019{
2020 struct commit *commit;
2021 struct origin *origin;
9aeaab68 2022 struct commit_list **parent_tail, *parent;
1cfe7733 2023 unsigned char head_sha1[20];
f285a2d7 2024 struct strbuf buf = STRBUF_INIT;
1cfe7733 2025 const char *ident;
1cfe7733 2026 time_t now;
1cfe7733
JH
2027 int size, len;
2028 struct cache_entry *ce;
2029 unsigned mode;
9aeaab68 2030 struct strbuf msg = STRBUF_INIT;
1cfe7733
JH
2031
2032 time(&now);
2033 commit = xcalloc(1, sizeof(*commit));
1cfe7733
JH
2034 commit->object.parsed = 1;
2035 commit->date = now;
2036 commit->object.type = OBJ_COMMIT;
9aeaab68
JH
2037 parent_tail = &commit->parents;
2038
2039 if (!resolve_ref_unsafe("HEAD", head_sha1, 1, NULL))
2040 die("no such ref: HEAD");
2041
2042 parent_tail = append_parent(parent_tail, head_sha1);
2043 append_merge_parents(parent_tail);
2044 verify_working_tree_path(commit, path);
1cfe7733
JH
2045
2046 origin = make_origin(commit, path);
2047
9aeaab68
JH
2048 ident = fmt_ident("Not Committed Yet", "not.committed.yet", NULL, 0);
2049 strbuf_addstr(&msg, "tree 0000000000000000000000000000000000000000\n");
2050 for (parent = commit->parents; parent; parent = parent->next)
2051 strbuf_addf(&msg, "parent %s\n",
2052 sha1_to_hex(parent->item->object.sha1));
2053 strbuf_addf(&msg,
2054 "author %s\n"
2055 "committer %s\n\n"
2056 "Version of %s from %s\n",
2057 ident, ident, path,
2058 (!contents_from ? path :
2059 (!strcmp(contents_from, "-") ? "standard input" : contents_from)));
2060 commit->buffer = strbuf_detach(&msg, NULL);
2061
1cfe7733
JH
2062 if (!contents_from || strcmp("-", contents_from)) {
2063 struct stat st;
2064 const char *read_from;
8518088f 2065 char *buf_ptr;
3b8a12e8 2066 unsigned long buf_len;
1cfe7733
JH
2067
2068 if (contents_from) {
2069 if (stat(contents_from, &st) < 0)
0721c314 2070 die_errno("Cannot stat '%s'", contents_from);
1cfe7733
JH
2071 read_from = contents_from;
2072 }
2073 else {
2074 if (lstat(path, &st) < 0)
0721c314 2075 die_errno("Cannot lstat '%s'", path);
1cfe7733
JH
2076 read_from = path;
2077 }
1cfe7733 2078 mode = canon_mode(st.st_mode);
3b8a12e8 2079
1cfe7733
JH
2080 switch (st.st_mode & S_IFMT) {
2081 case S_IFREG:
3b8a12e8 2082 if (DIFF_OPT_TST(opt, ALLOW_TEXTCONV) &&
e5450100 2083 textconv_object(read_from, mode, null_sha1, 0, &buf_ptr, &buf_len))
8518088f 2084 strbuf_attach(&buf, buf_ptr, buf_len, buf_len + 1);
3b8a12e8 2085 else if (strbuf_read_file(&buf, read_from, st.st_size) != st.st_size)
0721c314 2086 die_errno("cannot open or read '%s'", read_from);
1cfe7733
JH
2087 break;
2088 case S_IFLNK:
edfd45d2 2089 if (strbuf_readlink(&buf, read_from, st.st_size) < 0)
0721c314 2090 die_errno("cannot readlink '%s'", read_from);
1cfe7733
JH
2091 break;
2092 default:
2093 die("unsupported file type %s", read_from);
2094 }
2095 }
2096 else {
2097 /* Reading from stdin */
1cfe7733 2098 mode = 0;
f1696ee3 2099 if (strbuf_read(&buf, 0, 0) < 0)
d824cbba 2100 die_errno("failed to read from stdin");
1cfe7733 2101 }
21e5ad50 2102 convert_to_git(path, buf.buf, buf.len, &buf, 0);
af6eb822
PH
2103 origin->file.ptr = buf.buf;
2104 origin->file.size = buf.len;
2105 pretend_sha1_file(buf.buf, buf.len, OBJ_BLOB, origin->blob_sha1);
1cfe7733
JH
2106 commit->util = origin;
2107
2108 /*
2109 * Read the current index, replace the path entry with
2110 * origin->blob_sha1 without mucking with its mode or type
2111 * bits; we are not going to write this index out -- we just
2112 * want to run "diff-index --cached".
2113 */
2114 discard_cache();
2115 read_cache();
2116
2117 len = strlen(path);
2118 if (!mode) {
2119 int pos = cache_name_pos(path, len);
2120 if (0 <= pos)
7a51ed66 2121 mode = active_cache[pos]->ce_mode;
1cfe7733
JH
2122 else
2123 /* Let's not bother reading from HEAD tree */
2124 mode = S_IFREG | 0644;
2125 }
2126 size = cache_entry_size(len);
2127 ce = xcalloc(1, size);
2128 hashcpy(ce->sha1, origin->blob_sha1);
2129 memcpy(ce->name, path, len);
b60e188c
TG
2130 ce->ce_flags = create_ce_flags(0);
2131 ce->ce_namelen = len;
1cfe7733
JH
2132 ce->ce_mode = create_ce_mode(mode);
2133 add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
2134
2135 /*
2136 * We are not going to write this out, so this does not matter
2137 * right now, but someday we might optimize diff-index --cached
2138 * with cache-tree information.
2139 */
2140 cache_tree_invalidate_path(active_cache_tree, path);
2141
1cfe7733
JH
2142 return commit;
2143}
2144
85af7929 2145static const char *prepare_final(struct scoreboard *sb)
f6c07d7d
JH
2146{
2147 int i;
2148 const char *final_commit_name = NULL;
85af7929 2149 struct rev_info *revs = sb->revs;
f6c07d7d
JH
2150
2151 /*
2152 * There must be one and only one positive commit in the
2153 * revs->pending array.
2154 */
2155 for (i = 0; i < revs->pending.nr; i++) {
2156 struct object *obj = revs->pending.objects[i].item;
2157 if (obj->flags & UNINTERESTING)
2158 continue;
2159 while (obj->type == OBJ_TAG)
2160 obj = deref_tag(obj, NULL, 0);
2161 if (obj->type != OBJ_COMMIT)
2162 die("Non commit %s?", revs->pending.objects[i].name);
2163 if (sb->final)
2164 die("More than one commit to dig from %s and %s?",
2165 revs->pending.objects[i].name,
2166 final_commit_name);
2167 sb->final = (struct commit *) obj;
2168 final_commit_name = revs->pending.objects[i].name;
2169 }
2170 return final_commit_name;
2171}
2172
85af7929
JH
2173static const char *prepare_initial(struct scoreboard *sb)
2174{
2175 int i;
2176 const char *final_commit_name = NULL;
2177 struct rev_info *revs = sb->revs;
2178
2179 /*
2180 * There must be one and only one negative commit, and it must be
2181 * the boundary.
2182 */
2183 for (i = 0; i < revs->pending.nr; i++) {
2184 struct object *obj = revs->pending.objects[i].item;
2185 if (!(obj->flags & UNINTERESTING))
2186 continue;
2187 while (obj->type == OBJ_TAG)
2188 obj = deref_tag(obj, NULL, 0);
2189 if (obj->type != OBJ_COMMIT)
2190 die("Non commit %s?", revs->pending.objects[i].name);
2191 if (sb->final)
2192 die("More than one commit to dig down to %s and %s?",
2193 revs->pending.objects[i].name,
2194 final_commit_name);
2195 sb->final = (struct commit *) obj;
2196 final_commit_name = revs->pending.objects[i].name;
2197 }
2198 if (!final_commit_name)
2199 die("No commit to dig down to?");
2200 return final_commit_name;
2201}
2202
5817da01
PH
2203static int blame_copy_callback(const struct option *option, const char *arg, int unset)
2204{
2205 int *opt = option->value;
2206
2207 /*
2208 * -C enables copy from removed files;
2209 * -C -C enables copy from existing files, but only
2210 * when blaming a new file;
2211 * -C -C -C enables copy from existing files for
2212 * everybody
2213 */
2214 if (*opt & PICKAXE_BLAME_COPY_HARDER)
2215 *opt |= PICKAXE_BLAME_COPY_HARDEST;
2216 if (*opt & PICKAXE_BLAME_COPY)
2217 *opt |= PICKAXE_BLAME_COPY_HARDER;
2218 *opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE;
2219
2220 if (arg)
2221 blame_copy_score = parse_score(arg);
2222 return 0;
2223}
2224
2225static int blame_move_callback(const struct option *option, const char *arg, int unset)
2226{
2227 int *opt = option->value;
2228
2229 *opt |= PICKAXE_BLAME_MOVE;
2230
2231 if (arg)
2232 blame_move_score = parse_score(arg);
2233 return 0;
2234}
2235
2236static int blame_bottomtop_callback(const struct option *option, const char *arg, int unset)
2237{
2238 const char **bottomtop = option->value;
2239 if (!arg)
2240 return -1;
2241 if (*bottomtop)
2242 die("More than one '-L n,m' option given");
2243 *bottomtop = arg;
2244 return 0;
2245}
2246
acca687f 2247int cmd_blame(int argc, const char **argv, const char *prefix)
cee7f245
JH
2248{
2249 struct rev_info revs;
2250 const char *path;
2251 struct scoreboard sb;
2252 struct origin *o;
2253 struct blame_entry *ent;
3f8d5204 2254 long dashdash_pos, bottom, top, lno;
cee7f245 2255 const char *final_commit_name = NULL;
21666f1a 2256 enum object_type type;
5817da01
PH
2257
2258 static const char *bottomtop = NULL;
2259 static int output_option = 0, opt = 0;
2260 static int show_stats = 0;
2261 static const char *revs_file = NULL;
2262 static const char *contents_from = NULL;
2263 static const struct option options[] = {
efd2a8bd
NTND
2264 OPT_BOOLEAN(0, "incremental", &incremental, N_("Show blame entries as we find them, incrementally")),
2265 OPT_BOOLEAN('b', NULL, &blank_boundary, N_("Show blank SHA-1 for boundary commits (Default: off)")),
2266 OPT_BOOLEAN(0, "root", &show_root, N_("Do not treat root commits as boundaries (Default: off)")),
2267 OPT_BOOLEAN(0, "show-stats", &show_stats, N_("Show work cost statistics")),
2268 OPT_BIT(0, "score-debug", &output_option, N_("Show output score for blame entries"), OUTPUT_SHOW_SCORE),
2269 OPT_BIT('f', "show-name", &output_option, N_("Show original filename (Default: auto)"), OUTPUT_SHOW_NAME),
2270 OPT_BIT('n', "show-number", &output_option, N_("Show original linenumber (Default: off)"), OUTPUT_SHOW_NUMBER),
2271 OPT_BIT('p', "porcelain", &output_option, N_("Show in a format designed for machine consumption"), OUTPUT_PORCELAIN),
2272 OPT_BIT(0, "line-porcelain", &output_option, N_("Show porcelain format with per-line commit information"), OUTPUT_PORCELAIN|OUTPUT_LINE_PORCELAIN),
2273 OPT_BIT('c', NULL, &output_option, N_("Use the same output mode as git-annotate (Default: off)"), OUTPUT_ANNOTATE_COMPAT),
2274 OPT_BIT('t', NULL, &output_option, N_("Show raw timestamp (Default: off)"), OUTPUT_RAW_TIMESTAMP),
2275 OPT_BIT('l', NULL, &output_option, N_("Show long commit SHA1 (Default: off)"), OUTPUT_LONG_OBJECT_NAME),
2276 OPT_BIT('s', NULL, &output_option, N_("Suppress author name and timestamp (Default: off)"), OUTPUT_NO_AUTHOR),
2277 OPT_BIT('e', "show-email", &output_option, N_("Show author email instead of name (Default: off)"), OUTPUT_SHOW_EMAIL),
2278 OPT_BIT('w', NULL, &xdl_opts, N_("Ignore whitespace differences"), XDF_IGNORE_WHITESPACE),
2279 OPT_BIT(0, "minimal", &xdl_opts, N_("Spend extra cycles to find better match"), XDF_NEED_MINIMAL),
2280 OPT_STRING('S', NULL, &revs_file, N_("file"), N_("Use revisions from <file> instead of calling git-rev-list")),
2281 OPT_STRING(0, "contents", &contents_from, N_("file"), N_("Use <file>'s contents as the final image")),
2282 { OPTION_CALLBACK, 'C', NULL, &opt, N_("score"), N_("Find line copies within and across files"), PARSE_OPT_OPTARG, blame_copy_callback },
2283 { OPTION_CALLBACK, 'M', NULL, &opt, N_("score"), N_("Find line movements within and across files"), PARSE_OPT_OPTARG, blame_move_callback },
2284 OPT_CALLBACK('L', NULL, &bottomtop, N_("n,m"), N_("Process only line range n,m, counting from 1"), blame_bottomtop_callback),
84393bfd 2285 OPT__ABBREV(&abbrev),
5817da01
PH
2286 OPT_END()
2287 };
2288
2289 struct parse_opt_ctx_t ctx;
7ceacdff 2290 int cmd_is_annotate = !strcmp(argv[0], "annotate");
e68989a7 2291
ef90d6d4 2292 git_config(git_blame_config, NULL);
5817da01 2293 init_revisions(&revs, NULL);
31653c1a 2294 revs.date_mode = blame_date_mode;
3b8a12e8 2295 DIFF_OPT_SET(&revs.diffopt, ALLOW_TEXTCONV);
3d1aa566 2296 DIFF_OPT_SET(&revs.diffopt, FOLLOW_RENAMES);
31653c1a 2297
612702e8 2298 save_commit_buffer = 0;
3f8d5204 2299 dashdash_pos = 0;
612702e8 2300
9ca1169f
SB
2301 parse_options_start(&ctx, argc, argv, prefix, options,
2302 PARSE_OPT_KEEP_DASHDASH | PARSE_OPT_KEEP_ARGV0);
5817da01 2303 for (;;) {
5817da01
PH
2304 switch (parse_options_step(&ctx, options, blame_opt_usage)) {
2305 case PARSE_OPT_HELP:
2306 exit(129);
2307 case PARSE_OPT_DONE:
3f8d5204
PH
2308 if (ctx.argv[0])
2309 dashdash_pos = ctx.cpidx;
5817da01
PH
2310 goto parse_done;
2311 }
2312
2313 if (!strcmp(ctx.argv[0], "--reverse")) {
2314 ctx.argv[0] = "--children";
2315 reverse = 1;
2316 }
6b61ec05 2317 parse_revision_opt(&revs, &ctx, options, blame_opt_usage);
5817da01
PH
2318 }
2319parse_done:
3d1aa566
JH
2320 no_whole_file_rename = !DIFF_OPT_TST(&revs.diffopt, FOLLOW_RENAMES);
2321 DIFF_OPT_CLR(&revs.diffopt, FOLLOW_RENAMES);
5817da01
PH
2322 argc = parse_options_end(&ctx);
2323
b31272f7
JH
2324 if (0 < abbrev)
2325 /* one more abbrev length is needed for the boundary commit */
2326 abbrev++;
84393bfd 2327
aa9ea77d 2328 if (revs_file && read_ancestry(revs_file))
d824cbba 2329 die_errno("reading graft file '%s' failed", revs_file);
aa9ea77d 2330
31653c1a 2331 if (cmd_is_annotate) {
7ceacdff 2332 output_option |= OUTPUT_ANNOTATE_COMPAT;
31653c1a
EL
2333 blame_date_mode = DATE_ISO8601;
2334 } else {
2335 blame_date_mode = revs.date_mode;
2336 }
2337
2338 /* The maximum width used to show the dates */
2339 switch (blame_date_mode) {
2340 case DATE_RFC2822:
2341 blame_date_width = sizeof("Thu, 19 Oct 2006 16:00:04 -0700");
2342 break;
2343 case DATE_ISO8601:
2344 blame_date_width = sizeof("2006-10-19 16:00:04 -0700");
2345 break;
2346 case DATE_RAW:
2347 blame_date_width = sizeof("1161298804 -0700");
2348 break;
2349 case DATE_SHORT:
2350 blame_date_width = sizeof("2006-10-19");
2351 break;
2352 case DATE_RELATIVE:
2353 /* "normal" is used as the fallback for "relative" */
2354 case DATE_LOCAL:
2355 case DATE_NORMAL:
2356 blame_date_width = sizeof("Thu Oct 19 16:00:04 2006 -0700");
2357 break;
2358 }
2359 blame_date_width -= 1; /* strip the null */
7ceacdff 2360
b3123f98
JH
2361 if (DIFF_OPT_TST(&revs.diffopt, FIND_COPIES_HARDER))
2362 opt |= (PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE |
2363 PICKAXE_BLAME_COPY_HARDER);
2364
4a0fc95f
JH
2365 if (!blame_move_score)
2366 blame_move_score = BLAME_DEFAULT_MOVE_SCORE;
2367 if (!blame_copy_score)
2368 blame_copy_score = BLAME_DEFAULT_COPY_SCORE;
2369
1732a1fd
JH
2370 /*
2371 * We have collected options unknown to us in argv[1..unk]
cee7f245 2372 * which are to be passed to revision machinery if we are
3dff5379 2373 * going to do the "bottom" processing.
cee7f245
JH
2374 *
2375 * The remaining are:
2376 *
22e5e58a 2377 * (1) if dashdash_pos != 0, it is either
3f8d5204
PH
2378 * "blame [revisions] -- <path>" or
2379 * "blame -- <path> <rev>"
cee7f245 2380 *
22e5e58a 2381 * (2) otherwise, it is one of the two:
3f8d5204
PH
2382 * "blame [revisions] <path>"
2383 * "blame <path> <rev>"
cee7f245 2384 *
3f8d5204
PH
2385 * Note that we must strip out <path> from the arguments: we do not
2386 * want the path pruning but we may want "bottom" processing.
cee7f245 2387 */
3f8d5204
PH
2388 if (dashdash_pos) {
2389 switch (argc - dashdash_pos - 1) {
2390 case 2: /* (1b) */
2391 if (argc != 4)
5817da01 2392 usage_with_options(blame_opt_usage, options);
3f8d5204
PH
2393 /* reorder for the new way: <rev> -- <path> */
2394 argv[1] = argv[3];
2395 argv[3] = argv[2];
2396 argv[2] = "--";
2397 /* FALLTHROUGH */
2398 case 1: /* (1a) */
2399 path = add_prefix(prefix, argv[--argc]);
2400 argv[argc] = NULL;
2401 break;
2402 default:
2403 usage_with_options(blame_opt_usage, options);
cee7f245 2404 }
3f8d5204
PH
2405 } else {
2406 if (argc < 2)
5817da01 2407 usage_with_options(blame_opt_usage, options);
3f8d5204 2408 path = add_prefix(prefix, argv[argc - 1]);
c455c87c 2409 if (argc == 3 && !has_string_in_work_tree(path)) { /* (2b) */
3f8d5204
PH
2410 path = add_prefix(prefix, argv[1]);
2411 argv[1] = argv[2];
cee7f245 2412 }
3f8d5204 2413 argv[argc - 1] = "--";
cee7f245 2414
3f8d5204 2415 setup_work_tree();
c455c87c 2416 if (!has_string_in_work_tree(path))
d824cbba 2417 die_errno("cannot stat path '%s'", path);
cee7f245
JH
2418 }
2419
8b3dce56 2420 revs.disable_stdin = 1;
3f8d5204 2421 setup_revisions(argc, argv, &revs, NULL);
cee7f245
JH
2422 memset(&sb, 0, sizeof(sb));
2423
85af7929
JH
2424 sb.revs = &revs;
2425 if (!reverse)
2426 final_commit_name = prepare_final(&sb);
2427 else if (contents_from)
2428 die("--contents and --children do not blend well.");
2429 else
2430 final_commit_name = prepare_initial(&sb);
cee7f245
JH
2431
2432 if (!sb.final) {
1732a1fd
JH
2433 /*
2434 * "--not A B -- path" without anything positive;
1cfe7733
JH
2435 * do not default to HEAD, but use the working tree
2436 * or "--contents".
1732a1fd 2437 */
1981820b 2438 setup_work_tree();
3b8a12e8
AB
2439 sb.final = fake_working_tree_commit(&sb.revs->diffopt,
2440 path, contents_from);
1cfe7733 2441 add_pending_object(&revs, &(sb.final->object), ":");
cee7f245 2442 }
1cfe7733
JH
2443 else if (contents_from)
2444 die("Cannot use --contents with final commit object name");
cee7f245 2445
1732a1fd
JH
2446 /*
2447 * If we have bottom, this will mark the ancestors of the
cee7f245
JH
2448 * bottom commits we would reach while traversing as
2449 * uninteresting.
2450 */
3d51e1b5
MK
2451 if (prepare_revision_walk(&revs))
2452 die("revision walk setup failed");
cee7f245 2453
1cfe7733
JH
2454 if (is_null_sha1(sb.final->object.sha1)) {
2455 char *buf;
2456 o = sb.final->util;
2457 buf = xmalloc(o->file.size + 1);
2458 memcpy(buf, o->file.ptr, o->file.size + 1);
2459 sb.final_buf = buf;
2460 sb.final_buf_size = o->file.size;
2461 }
2462 else {
2463 o = get_origin(&sb, sb.final, path);
90064710 2464 if (fill_blob_sha1_and_mode(o))
1cfe7733 2465 die("no such path %s in %s", path, final_commit_name);
cee7f245 2466
3b8a12e8 2467 if (DIFF_OPT_TST(&sb.revs->diffopt, ALLOW_TEXTCONV) &&
e5450100 2468 textconv_object(path, o->mode, o->blob_sha1, 1, (char **) &sb.final_buf,
3b8a12e8
AB
2469 &sb.final_buf_size))
2470 ;
2471 else
2472 sb.final_buf = read_sha1_file(o->blob_sha1, &type,
2473 &sb.final_buf_size);
2474
ab43e495
JH
2475 if (!sb.final_buf)
2476 die("Cannot read blob %s for path %s",
2477 sha1_to_hex(o->blob_sha1),
2478 path);
1cfe7733 2479 }
c2e525d9 2480 num_read_blob++;
cee7f245
JH
2481 lno = prepare_lines(&sb);
2482
931233bc 2483 bottom = top = 0;
75393574
ES
2484 if (bottomtop && parse_range_arg(bottomtop, nth_line_cb, &sb, lno,
2485 &bottom, &top, sb.path))
2486 usage(blame_usage);
164a9cf4
ES
2487 if (lno < top || ((lno || bottom) && lno < bottom))
2488 die("file %s has only %lu lines", path, lno);
cee7f245
JH
2489 if (bottom < 1)
2490 bottom = 1;
2491 if (top < 1)
2492 top = lno;
2493 bottom--;
cee7f245
JH
2494
2495 ent = xcalloc(1, sizeof(*ent));
2496 ent->lno = bottom;
2497 ent->num_lines = top - bottom;
2498 ent->suspect = o;
2499 ent->s_lno = bottom;
2500
2501 sb.ent = ent;
2502 sb.path = path;
2503
d551a488 2504 read_mailmap(&mailmap, NULL);
f9567384 2505
b92565dc
MH
2506 if (!incremental)
2507 setup_pager();
2508
85af7929 2509 assign_blame(&sb, opt);
cee7f245 2510
717d1462
LT
2511 if (incremental)
2512 return 0;
2513
cee7f245
JH
2514 coalesce(&sb);
2515
2516 if (!(output_option & OUTPUT_PORCELAIN))
2517 find_alignment(&sb, &output_option);
2518
2519 output(&sb, output_option);
2520 free((void *)sb.final_buf);
2521 for (ent = sb.ent; ent; ) {
2522 struct blame_entry *e = ent->next;
2523 free(ent);
2524 ent = e;
2525 }
c2e525d9 2526
870b39c1 2527 if (show_stats) {
c2e525d9
JH
2528 printf("num read blob: %d\n", num_read_blob);
2529 printf("num get patch: %d\n", num_get_patch);
2530 printf("num commits: %d\n", num_commits);
2531 }
cee7f245
JH
2532 return 0;
2533}