blame: refactor porcelain output
[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"
cee7f245 24
588c038a 25static char blame_usage[] = "git blame [options] [rev-opts] [rev] [--] file";
5817da01
PH
26
27static const char *blame_opt_usage[] = {
28 blame_usage,
29 "",
76447d23 30 "[rev-opts] are documented in git-rev-list(1)",
5817da01
PH
31 NULL
32};
cee7f245
JH
33
34static int longest_file;
35static int longest_author;
36static int max_orig_digits;
37static int max_digits;
5ff62c30 38static int max_score_digits;
4c10a5ca 39static int show_root;
85af7929 40static int reverse;
4c10a5ca 41static int blank_boundary;
717d1462 42static int incremental;
582aa00b 43static int xdl_opts;
84393bfd 44static int abbrev = -1;
31653c1a
EL
45
46static enum date_mode blame_date_mode = DATE_ISO8601;
47static size_t blame_date_width;
48
c455c87c 49static struct string_list mailmap;
cee7f245 50
54a4c617
JH
51#ifndef DEBUG
52#define DEBUG 0
53#endif
54
c2e525d9
JH
55/* stats */
56static int num_read_blob;
57static int num_get_patch;
58static int num_commits;
59
d24bba80 60#define PICKAXE_BLAME_MOVE 01
18abd745
JH
61#define PICKAXE_BLAME_COPY 02
62#define PICKAXE_BLAME_COPY_HARDER 04
c63777c0 63#define PICKAXE_BLAME_COPY_HARDEST 010
d24bba80 64
4a0fc95f
JH
65/*
66 * blame for a blame_entry with score lower than these thresholds
67 * is not passed to the parent using move/copy logic.
68 */
69static unsigned blame_move_score;
70static unsigned blame_copy_score;
71#define BLAME_DEFAULT_MOVE_SCORE 20
72#define BLAME_DEFAULT_COPY_SCORE 40
73
cee7f245
JH
74/* bits #0..7 in revision.h, #8..11 used for merge_bases() in commit.c */
75#define METAINFO_SHOWN (1u<<12)
76#define MORE_THAN_ONE_PATH (1u<<13)
77
78/*
54a4c617 79 * One blob in a commit that is being suspected
cee7f245
JH
80 */
81struct origin {
54a4c617 82 int refcnt;
96e11709 83 struct origin *previous;
cee7f245 84 struct commit *commit;
c2e525d9 85 mmfile_t file;
cee7f245 86 unsigned char blob_sha1[20];
90064710 87 unsigned mode;
cee7f245
JH
88 char path[FLEX_ARRAY];
89};
90
3b8a12e8
AB
91/*
92 * Prepare diff_filespec and convert it using diff textconv API
93 * if the textconv driver exists.
94 * Return 1 if the conversion succeeds, 0 otherwise.
95 */
e5fba602 96int textconv_object(const char *path,
90064710 97 unsigned mode,
e5fba602
CP
98 const unsigned char *sha1,
99 char **buf,
100 unsigned long *buf_size)
3b8a12e8
AB
101{
102 struct diff_filespec *df;
103 struct userdiff_driver *textconv;
104
105 df = alloc_filespec(path);
90064710 106 fill_filespec(df, sha1, mode);
3b8a12e8
AB
107 textconv = get_textconv(df);
108 if (!textconv) {
109 free_filespec(df);
110 return 0;
111 }
112
113 *buf_size = fill_textconv(textconv, df, buf);
114 free_filespec(df);
115 return 1;
116}
117
1732a1fd
JH
118/*
119 * Given an origin, prepare mmfile_t structure to be used by the
120 * diff machinery
121 */
3b8a12e8
AB
122static void fill_origin_blob(struct diff_options *opt,
123 struct origin *o, mmfile_t *file)
c2e525d9
JH
124{
125 if (!o->file.ptr) {
21666f1a 126 enum object_type type;
3b8a12e8
AB
127 unsigned long file_size;
128
c2e525d9 129 num_read_blob++;
3b8a12e8 130 if (DIFF_OPT_TST(opt, ALLOW_TEXTCONV) &&
90064710 131 textconv_object(o->path, o->mode, o->blob_sha1, &file->ptr, &file_size))
3b8a12e8
AB
132 ;
133 else
134 file->ptr = read_sha1_file(o->blob_sha1, &type, &file_size);
135 file->size = file_size;
136
ab43e495
JH
137 if (!file->ptr)
138 die("Cannot read blob %s for path %s",
139 sha1_to_hex(o->blob_sha1),
140 o->path);
c2e525d9
JH
141 o->file = *file;
142 }
143 else
144 *file = o->file;
c2e525d9
JH
145}
146
1732a1fd
JH
147/*
148 * Origin is refcounted and usually we keep the blob contents to be
149 * reused.
150 */
54a4c617
JH
151static inline struct origin *origin_incref(struct origin *o)
152{
153 if (o)
154 o->refcnt++;
155 return o;
156}
157
158static void origin_decref(struct origin *o)
159{
160 if (o && --o->refcnt <= 0) {
96e11709
JH
161 if (o->previous)
162 origin_decref(o->previous);
8e0f7003 163 free(o->file.ptr);
54a4c617
JH
164 free(o);
165 }
166}
167
7c3c7962
JH
168static void drop_origin_blob(struct origin *o)
169{
170 if (o->file.ptr) {
171 free(o->file.ptr);
172 o->file.ptr = NULL;
173 }
174}
175
1732a1fd
JH
176/*
177 * Each group of lines is described by a blame_entry; it can be split
178 * as we pass blame to the parents. They form a linked list in the
179 * scoreboard structure, sorted by the target line number.
180 */
cee7f245
JH
181struct blame_entry {
182 struct blame_entry *prev;
183 struct blame_entry *next;
184
185 /* the first line of this group in the final image;
186 * internally all line numbers are 0 based.
187 */
188 int lno;
189
190 /* how many lines this group has */
191 int num_lines;
192
193 /* the commit that introduced this group into the final image */
194 struct origin *suspect;
195
196 /* true if the suspect is truly guilty; false while we have not
197 * checked if the group came from one of its parents.
198 */
199 char guilty;
200
eb46328e
AG
201 /* true if the entry has been scanned for copies in the current parent
202 */
203 char scanned;
204
cee7f245
JH
205 /* the line number of the first line of this group in the
206 * suspect's file; internally all line numbers are 0 based.
207 */
208 int s_lno;
5ff62c30
JH
209
210 /* how significant this entry is -- cached to avoid
1732a1fd 211 * scanning the lines over and over.
5ff62c30
JH
212 */
213 unsigned score;
cee7f245
JH
214};
215
1732a1fd
JH
216/*
217 * The current state of the blame assignment.
218 */
cee7f245
JH
219struct scoreboard {
220 /* the final commit (i.e. where we started digging from) */
221 struct commit *final;
85af7929 222 struct rev_info *revs;
cee7f245
JH
223 const char *path;
224
1732a1fd
JH
225 /*
226 * The contents in the final image.
227 * Used by many functions to obtain contents of the nth line,
228 * indexed with scoreboard.lineno[blame_entry.lno].
cee7f245
JH
229 */
230 const char *final_buf;
231 unsigned long final_buf_size;
232
233 /* linked list of blames */
234 struct blame_entry *ent;
235
612702e8 236 /* look-up a line in the final buffer */
cee7f245
JH
237 int num_lines;
238 int *lineno;
239};
240
171dccd5 241static inline int same_suspect(struct origin *a, struct origin *b)
46014766 242{
171dccd5 243 if (a == b)
57584d9e 244 return 1;
171dccd5
JH
245 if (a->commit != b->commit)
246 return 0;
247 return !strcmp(a->path, b->path);
46014766
JH
248}
249
54a4c617
JH
250static void sanity_check_refcnt(struct scoreboard *);
251
1732a1fd
JH
252/*
253 * If two blame entries that are next to each other came from
254 * contiguous lines in the same origin (i.e. <commit, path> pair),
255 * merge them together.
256 */
cee7f245
JH
257static void coalesce(struct scoreboard *sb)
258{
259 struct blame_entry *ent, *next;
260
261 for (ent = sb->ent; ent && (next = ent->next); ent = next) {
171dccd5 262 if (same_suspect(ent->suspect, next->suspect) &&
cee7f245
JH
263 ent->guilty == next->guilty &&
264 ent->s_lno + ent->num_lines == next->s_lno) {
265 ent->num_lines += next->num_lines;
266 ent->next = next->next;
267 if (ent->next)
268 ent->next->prev = ent;
54a4c617 269 origin_decref(next->suspect);
cee7f245 270 free(next);
46014766 271 ent->score = 0;
cee7f245
JH
272 next = ent; /* again */
273 }
274 }
54a4c617
JH
275
276 if (DEBUG) /* sanity */
277 sanity_check_refcnt(sb);
cee7f245
JH
278}
279
1732a1fd
JH
280/*
281 * Given a commit and a path in it, create a new origin structure.
282 * The callers that add blame to the scoreboard should use
283 * get_origin() to obtain shared, refcounted copy instead of calling
284 * this function directly.
285 */
854b97f6
JH
286static struct origin *make_origin(struct commit *commit, const char *path)
287{
288 struct origin *o;
289 o = xcalloc(1, sizeof(*o) + strlen(path) + 1);
290 o->commit = commit;
291 o->refcnt = 1;
292 strcpy(o->path, path);
293 return o;
294}
295
1732a1fd
JH
296/*
297 * Locate an existing origin or create a new one.
298 */
f6c0e191
JH
299static struct origin *get_origin(struct scoreboard *sb,
300 struct commit *commit,
301 const char *path)
cee7f245 302{
f6c0e191 303 struct blame_entry *e;
cee7f245 304
f6c0e191
JH
305 for (e = sb->ent; e; e = e->next) {
306 if (e->suspect->commit == commit &&
307 !strcmp(e->suspect->path, path))
54a4c617 308 return origin_incref(e->suspect);
cee7f245 309 }
854b97f6 310 return make_origin(commit, path);
cee7f245
JH
311}
312
1732a1fd
JH
313/*
314 * Fill the blob_sha1 field of an origin if it hasn't, so that later
315 * call to fill_origin_blob() can use it to locate the data. blob_sha1
316 * for an origin is also used to pass the blame for the entire file to
317 * the parent to detect the case where a child's blob is identical to
318 * that of its parent's.
90064710
KS
319 *
320 * This also fills origin->mode for corresponding tree path.
1732a1fd 321 */
90064710 322static int fill_blob_sha1_and_mode(struct origin *origin)
f6c0e191 323{
f6c0e191
JH
324 if (!is_null_sha1(origin->blob_sha1))
325 return 0;
326 if (get_tree_entry(origin->commit->object.sha1,
327 origin->path,
90064710 328 origin->blob_sha1, &origin->mode))
f6c0e191 329 goto error_out;
21666f1a 330 if (sha1_object_info(origin->blob_sha1, NULL) != OBJ_BLOB)
f6c0e191
JH
331 goto error_out;
332 return 0;
333 error_out:
334 hashclr(origin->blob_sha1);
90064710 335 origin->mode = S_IFINVALID;
f6c0e191
JH
336 return -1;
337}
338
1732a1fd
JH
339/*
340 * We have an origin -- check if the same path exists in the
341 * parent and return an origin structure to represent it.
342 */
f6c0e191 343static struct origin *find_origin(struct scoreboard *sb,
cee7f245
JH
344 struct commit *parent,
345 struct origin *origin)
346{
347 struct origin *porigin = NULL;
348 struct diff_options diff_opts;
f6c0e191
JH
349 const char *paths[2];
350
0d981c67 351 if (parent->util) {
1732a1fd
JH
352 /*
353 * Each commit object can cache one origin in that
354 * commit. This is a freestanding copy of origin and
355 * not refcounted.
854b97f6 356 */
0d981c67 357 struct origin *cached = parent->util;
854b97f6 358 if (!strcmp(cached->path, origin->path)) {
1732a1fd
JH
359 /*
360 * The same path between origin and its parent
361 * without renaming -- the most common case.
362 */
854b97f6 363 porigin = get_origin(sb, parent, cached->path);
1732a1fd
JH
364
365 /*
366 * If the origin was newly created (i.e. get_origin
367 * would call make_origin if none is found in the
90064710 368 * scoreboard), it does not know the blob_sha1/mode,
1732a1fd 369 * so copy it. Otherwise porigin was in the
90064710 370 * scoreboard and already knows blob_sha1/mode.
1732a1fd 371 */
90064710 372 if (porigin->refcnt == 1) {
854b97f6 373 hashcpy(porigin->blob_sha1, cached->blob_sha1);
90064710
KS
374 porigin->mode = cached->mode;
375 }
854b97f6
JH
376 return porigin;
377 }
378 /* otherwise it was not very useful; free it */
379 free(parent->util);
380 parent->util = NULL;
0d981c67
JH
381 }
382
f6c0e191
JH
383 /* See if the origin->path is different between parent
384 * and origin first. Most of the time they are the
385 * same and diff-tree is fairly efficient about this.
386 */
387 diff_setup(&diff_opts);
8f67f8ae 388 DIFF_OPT_SET(&diff_opts, RECURSIVE);
f6c0e191
JH
389 diff_opts.detect_rename = 0;
390 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
391 paths[0] = origin->path;
392 paths[1] = NULL;
393
394 diff_tree_setup_paths(paths, &diff_opts);
395 if (diff_setup_done(&diff_opts) < 0)
396 die("diff-setup");
1cfe7733
JH
397
398 if (is_null_sha1(origin->commit->object.sha1))
399 do_diff_cache(parent->tree->object.sha1, &diff_opts);
400 else
401 diff_tree_sha1(parent->tree->object.sha1,
402 origin->commit->tree->object.sha1,
403 "", &diff_opts);
f6c0e191
JH
404 diffcore_std(&diff_opts);
405
f6c0e191
JH
406 if (!diff_queued_diff.nr) {
407 /* The path is the same as parent */
408 porigin = get_origin(sb, parent, origin->path);
409 hashcpy(porigin->blob_sha1, origin->blob_sha1);
90064710 410 porigin->mode = origin->mode;
a9b2d424
JH
411 } else {
412 /*
413 * Since origin->path is a pathspec, if the parent
414 * commit had it as a directory, we will see a whole
415 * bunch of deletion of files in the directory that we
416 * do not care about.
417 */
418 int i;
419 struct diff_filepair *p = NULL;
420 for (i = 0; i < diff_queued_diff.nr; i++) {
421 const char *name;
422 p = diff_queued_diff.queue[i];
423 name = p->one->path ? p->one->path : p->two->path;
424 if (!strcmp(name, origin->path))
425 break;
426 }
427 if (!p)
428 die("internal error in blame::find_origin");
f6c0e191
JH
429 switch (p->status) {
430 default:
acca687f 431 die("internal error in blame::find_origin (%c)",
f6c0e191
JH
432 p->status);
433 case 'M':
434 porigin = get_origin(sb, parent, origin->path);
435 hashcpy(porigin->blob_sha1, p->one->sha1);
90064710 436 porigin->mode = p->one->mode;
f6c0e191
JH
437 break;
438 case 'A':
439 case 'T':
440 /* Did not exist in parent, or type changed */
441 break;
442 }
443 }
444 diff_flush(&diff_opts);
03b69c76 445 diff_tree_release_paths(&diff_opts);
0d981c67 446 if (porigin) {
1732a1fd
JH
447 /*
448 * Create a freestanding copy that is not part of
449 * the refcounted origin found in the scoreboard, and
450 * cache it in the commit.
451 */
854b97f6 452 struct origin *cached;
1732a1fd 453
854b97f6
JH
454 cached = make_origin(porigin->commit, porigin->path);
455 hashcpy(cached->blob_sha1, porigin->blob_sha1);
90064710 456 cached->mode = porigin->mode;
854b97f6 457 parent->util = cached;
0d981c67 458 }
f69e743d
JH
459 return porigin;
460}
f6c0e191 461
1732a1fd
JH
462/*
463 * We have an origin -- find the path that corresponds to it in its
464 * parent and return an origin structure to represent it.
465 */
f69e743d
JH
466static struct origin *find_rename(struct scoreboard *sb,
467 struct commit *parent,
468 struct origin *origin)
469{
470 struct origin *porigin = NULL;
471 struct diff_options diff_opts;
472 int i;
473 const char *paths[2];
cee7f245
JH
474
475 diff_setup(&diff_opts);
8f67f8ae 476 DIFF_OPT_SET(&diff_opts, RECURSIVE);
cee7f245
JH
477 diff_opts.detect_rename = DIFF_DETECT_RENAME;
478 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
2f3f8b21 479 diff_opts.single_follow = origin->path;
cee7f245
JH
480 paths[0] = NULL;
481 diff_tree_setup_paths(paths, &diff_opts);
482 if (diff_setup_done(&diff_opts) < 0)
483 die("diff-setup");
1cfe7733
JH
484
485 if (is_null_sha1(origin->commit->object.sha1))
486 do_diff_cache(parent->tree->object.sha1, &diff_opts);
487 else
488 diff_tree_sha1(parent->tree->object.sha1,
489 origin->commit->tree->object.sha1,
490 "", &diff_opts);
cee7f245
JH
491 diffcore_std(&diff_opts);
492
493 for (i = 0; i < diff_queued_diff.nr; i++) {
494 struct diff_filepair *p = diff_queued_diff.queue[i];
612702e8 495 if ((p->status == 'R' || p->status == 'C') &&
f6c0e191
JH
496 !strcmp(p->two->path, origin->path)) {
497 porigin = get_origin(sb, parent, p->one->path);
498 hashcpy(porigin->blob_sha1, p->one->sha1);
90064710 499 porigin->mode = p->one->mode;
cee7f245
JH
500 break;
501 }
502 }
503 diff_flush(&diff_opts);
03b69c76 504 diff_tree_release_paths(&diff_opts);
cee7f245
JH
505 return porigin;
506}
507
1732a1fd 508/*
3dff5379 509 * Link in a new blame entry to the scoreboard. Entries that cover the
1732a1fd
JH
510 * same line range have been removed from the scoreboard previously.
511 */
cee7f245
JH
512static void add_blame_entry(struct scoreboard *sb, struct blame_entry *e)
513{
514 struct blame_entry *ent, *prev = NULL;
515
54a4c617
JH
516 origin_incref(e->suspect);
517
cee7f245
JH
518 for (ent = sb->ent; ent && ent->lno < e->lno; ent = ent->next)
519 prev = ent;
520
521 /* prev, if not NULL, is the last one that is below e */
522 e->prev = prev;
523 if (prev) {
524 e->next = prev->next;
525 prev->next = e;
526 }
527 else {
528 e->next = sb->ent;
529 sb->ent = e;
530 }
531 if (e->next)
532 e->next->prev = e;
533}
534
1732a1fd
JH
535/*
536 * src typically is on-stack; we want to copy the information in it to
790296fd 537 * a malloced blame_entry that is already on the linked list of the
1732a1fd
JH
538 * scoreboard. The origin of dst loses a refcnt while the origin of src
539 * gains one.
540 */
cee7f245
JH
541static void dup_entry(struct blame_entry *dst, struct blame_entry *src)
542{
543 struct blame_entry *p, *n;
54a4c617 544
cee7f245
JH
545 p = dst->prev;
546 n = dst->next;
54a4c617
JH
547 origin_incref(src->suspect);
548 origin_decref(dst->suspect);
cee7f245
JH
549 memcpy(dst, src, sizeof(*src));
550 dst->prev = p;
551 dst->next = n;
5ff62c30 552 dst->score = 0;
cee7f245
JH
553}
554
555static const char *nth_line(struct scoreboard *sb, int lno)
556{
557 return sb->final_buf + sb->lineno[lno];
558}
559
1732a1fd
JH
560/*
561 * It is known that lines between tlno to same came from parent, and e
562 * has an overlap with that range. it also is known that parent's
563 * line plno corresponds to e's line tlno.
564 *
565 * <---- e ----->
566 * <------>
567 * <------------>
568 * <------------>
569 * <------------------>
570 *
571 * Split e into potentially three parts; before this chunk, the chunk
572 * to be blamed for the parent, and after that portion.
573 */
54a4c617 574static void split_overlap(struct blame_entry *split,
cee7f245
JH
575 struct blame_entry *e,
576 int tlno, int plno, int same,
577 struct origin *parent)
578{
cee7f245
JH
579 int chunk_end_lno;
580 memset(split, 0, sizeof(struct blame_entry [3]));
581
582 if (e->s_lno < tlno) {
583 /* there is a pre-chunk part not blamed on parent */
54a4c617 584 split[0].suspect = origin_incref(e->suspect);
cee7f245
JH
585 split[0].lno = e->lno;
586 split[0].s_lno = e->s_lno;
587 split[0].num_lines = tlno - e->s_lno;
588 split[1].lno = e->lno + tlno - e->s_lno;
589 split[1].s_lno = plno;
590 }
591 else {
592 split[1].lno = e->lno;
593 split[1].s_lno = plno + (e->s_lno - tlno);
594 }
595
596 if (same < e->s_lno + e->num_lines) {
597 /* there is a post-chunk part not blamed on parent */
54a4c617 598 split[2].suspect = origin_incref(e->suspect);
cee7f245
JH
599 split[2].lno = e->lno + (same - e->s_lno);
600 split[2].s_lno = e->s_lno + (same - e->s_lno);
601 split[2].num_lines = e->s_lno + e->num_lines - same;
602 chunk_end_lno = split[2].lno;
603 }
604 else
605 chunk_end_lno = e->lno + e->num_lines;
606 split[1].num_lines = chunk_end_lno - split[1].lno;
607
1732a1fd
JH
608 /*
609 * if it turns out there is nothing to blame the parent for,
610 * forget about the splitting. !split[1].suspect signals this.
611 */
cee7f245
JH
612 if (split[1].num_lines < 1)
613 return;
54a4c617 614 split[1].suspect = origin_incref(parent);
cee7f245
JH
615}
616
1732a1fd
JH
617/*
618 * split_overlap() divided an existing blame e into up to three parts
619 * in split. Adjust the linked list of blames in the scoreboard to
620 * reflect the split.
621 */
cee7f245 622static void split_blame(struct scoreboard *sb,
54a4c617 623 struct blame_entry *split,
cee7f245
JH
624 struct blame_entry *e)
625{
626 struct blame_entry *new_entry;
627
628 if (split[0].suspect && split[2].suspect) {
1732a1fd 629 /* The first part (reuse storage for the existing entry e) */
cee7f245
JH
630 dup_entry(e, &split[0]);
631
1732a1fd 632 /* The last part -- me */
cee7f245
JH
633 new_entry = xmalloc(sizeof(*new_entry));
634 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
635 add_blame_entry(sb, new_entry);
636
1732a1fd 637 /* ... and the middle part -- parent */
cee7f245
JH
638 new_entry = xmalloc(sizeof(*new_entry));
639 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
640 add_blame_entry(sb, new_entry);
641 }
642 else if (!split[0].suspect && !split[2].suspect)
1732a1fd
JH
643 /*
644 * The parent covers the entire area; reuse storage for
645 * e and replace it with the parent.
646 */
cee7f245
JH
647 dup_entry(e, &split[1]);
648 else if (split[0].suspect) {
1732a1fd 649 /* me and then parent */
cee7f245
JH
650 dup_entry(e, &split[0]);
651
652 new_entry = xmalloc(sizeof(*new_entry));
653 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
654 add_blame_entry(sb, new_entry);
655 }
656 else {
1732a1fd 657 /* parent and then me */
cee7f245
JH
658 dup_entry(e, &split[1]);
659
660 new_entry = xmalloc(sizeof(*new_entry));
661 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
662 add_blame_entry(sb, new_entry);
663 }
664
54a4c617 665 if (DEBUG) { /* sanity */
cee7f245 666 struct blame_entry *ent;
612702e8 667 int lno = sb->ent->lno, corrupt = 0;
cee7f245
JH
668
669 for (ent = sb->ent; ent; ent = ent->next) {
670 if (lno != ent->lno)
671 corrupt = 1;
672 if (ent->s_lno < 0)
673 corrupt = 1;
674 lno += ent->num_lines;
675 }
676 if (corrupt) {
612702e8 677 lno = sb->ent->lno;
cee7f245
JH
678 for (ent = sb->ent; ent; ent = ent->next) {
679 printf("L %8d l %8d n %8d\n",
680 lno, ent->lno, ent->num_lines);
681 lno = ent->lno + ent->num_lines;
682 }
683 die("oops");
684 }
685 }
686}
687
1732a1fd
JH
688/*
689 * After splitting the blame, the origins used by the
690 * on-stack blame_entry should lose one refcnt each.
691 */
54a4c617
JH
692static void decref_split(struct blame_entry *split)
693{
694 int i;
695
696 for (i = 0; i < 3; i++)
697 origin_decref(split[i].suspect);
698}
699
1732a1fd
JH
700/*
701 * Helper for blame_chunk(). blame_entry e is known to overlap with
702 * the patch hunk; split it and pass blame to the parent.
703 */
cee7f245
JH
704static void blame_overlap(struct scoreboard *sb, struct blame_entry *e,
705 int tlno, int plno, int same,
706 struct origin *parent)
707{
708 struct blame_entry split[3];
709
710 split_overlap(split, e, tlno, plno, same, parent);
54a4c617
JH
711 if (split[1].suspect)
712 split_blame(sb, split, e);
713 decref_split(split);
cee7f245
JH
714}
715
1732a1fd
JH
716/*
717 * Find the line number of the last line the target is suspected for.
718 */
cee7f245
JH
719static int find_last_in_target(struct scoreboard *sb, struct origin *target)
720{
721 struct blame_entry *e;
722 int last_in_target = -1;
723
724 for (e = sb->ent; e; e = e->next) {
171dccd5 725 if (e->guilty || !same_suspect(e->suspect, target))
cee7f245
JH
726 continue;
727 if (last_in_target < e->s_lno + e->num_lines)
728 last_in_target = e->s_lno + e->num_lines;
729 }
730 return last_in_target;
731}
732
1732a1fd
JH
733/*
734 * Process one hunk from the patch between the current suspect for
735 * blame_entry e and its parent. Find and split the overlap, and
736 * pass blame to the overlapping part to the parent.
737 */
cee7f245
JH
738static void blame_chunk(struct scoreboard *sb,
739 int tlno, int plno, int same,
740 struct origin *target, struct origin *parent)
741{
612702e8 742 struct blame_entry *e;
cee7f245 743
612702e8 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 (same <= e->s_lno)
748 continue;
749 if (tlno < e->s_lno + e->num_lines)
750 blame_overlap(sb, e, tlno, plno, same, parent);
751 }
752}
753
cdcab133
RS
754struct blame_chunk_cb_data {
755 struct scoreboard *sb;
756 struct origin *target;
757 struct origin *parent;
758 long plno;
759 long tlno;
760};
761
762static void blame_chunk_cb(void *data, long same, long p_next, long t_next)
763{
764 struct blame_chunk_cb_data *d = data;
765 blame_chunk(d->sb, d->tlno, d->plno, same, d->target, d->parent);
766 d->plno = p_next;
767 d->tlno = t_next;
768}
769
1732a1fd
JH
770/*
771 * We are looking at the origin 'target' and aiming to pass blame
772 * for the lines it is suspected to its parent. Run diff to find
773 * which lines came from parent and pass blame for them.
774 */
cee7f245
JH
775static int pass_blame_to_parent(struct scoreboard *sb,
776 struct origin *target,
777 struct origin *parent)
778{
cdcab133 779 int last_in_target;
7c4ed8c8 780 mmfile_t file_p, file_o;
66dbfd55 781 struct blame_chunk_cb_data d;
cdcab133
RS
782 xpparam_t xpp;
783 xdemitconf_t xecfg;
66dbfd55
GV
784 memset(&d, 0, sizeof(d));
785 d.sb = sb; d.target = target; d.parent = parent;
cee7f245
JH
786 last_in_target = find_last_in_target(sb, target);
787 if (last_in_target < 0)
788 return 1; /* nothing remains for this target */
789
3b8a12e8
AB
790 fill_origin_blob(&sb->revs->diffopt, parent, &file_p);
791 fill_origin_blob(&sb->revs->diffopt, target, &file_o);
7c4ed8c8 792 num_get_patch++;
cee7f245 793
cdcab133
RS
794 memset(&xpp, 0, sizeof(xpp));
795 xpp.flags = xdl_opts;
796 memset(&xecfg, 0, sizeof(xecfg));
797 xecfg.ctxlen = 0;
798 xdi_diff_hunks(&file_p, &file_o, blame_chunk_cb, &d, &xpp, &xecfg);
1732a1fd 799 /* The rest (i.e. anything after tlno) are the same as the parent */
cdcab133 800 blame_chunk(sb, d.tlno, d.plno, last_in_target, target, parent);
cee7f245 801
cee7f245
JH
802 return 0;
803}
804
1732a1fd
JH
805/*
806 * The lines in blame_entry after splitting blames many times can become
807 * very small and trivial, and at some point it becomes pointless to
808 * blame the parents. E.g. "\t\t}\n\t}\n\n" appears everywhere in any
809 * ordinary C program, and it is not worth to say it was copied from
810 * totally unrelated file in the parent.
811 *
812 * Compute how trivial the lines in the blame_entry are.
813 */
5ff62c30
JH
814static unsigned ent_score(struct scoreboard *sb, struct blame_entry *e)
815{
816 unsigned score;
817 const char *cp, *ep;
818
819 if (e->score)
820 return e->score;
821
612702e8 822 score = 1;
5ff62c30
JH
823 cp = nth_line(sb, e->lno);
824 ep = nth_line(sb, e->lno + e->num_lines);
825 while (cp < ep) {
826 unsigned ch = *((unsigned char *)cp);
827 if (isalnum(ch))
828 score++;
829 cp++;
830 }
831 e->score = score;
832 return score;
833}
834
1732a1fd
JH
835/*
836 * best_so_far[] and this[] are both a split of an existing blame_entry
837 * that passes blame to the parent. Maintain best_so_far the best split
838 * so far, by comparing this and best_so_far and copying this into
839 * bst_so_far as needed.
840 */
5ff62c30 841static void copy_split_if_better(struct scoreboard *sb,
54a4c617
JH
842 struct blame_entry *best_so_far,
843 struct blame_entry *this)
d24bba80 844{
54a4c617
JH
845 int i;
846
d24bba80
JH
847 if (!this[1].suspect)
848 return;
5ff62c30
JH
849 if (best_so_far[1].suspect) {
850 if (ent_score(sb, &this[1]) < ent_score(sb, &best_so_far[1]))
851 return;
852 }
54a4c617
JH
853
854 for (i = 0; i < 3; i++)
855 origin_incref(this[i].suspect);
856 decref_split(best_so_far);
d24bba80
JH
857 memcpy(best_so_far, this, sizeof(struct blame_entry [3]));
858}
859
dd166aa8
JH
860/*
861 * We are looking at a part of the final image represented by
862 * ent (tlno and same are offset by ent->s_lno).
863 * tlno is where we are looking at in the final image.
864 * up to (but not including) same match preimage.
865 * plno is where we are looking at in the preimage.
866 *
867 * <-------------- final image ---------------------->
868 * <------ent------>
869 * ^tlno ^same
870 * <---------preimage----->
871 * ^plno
872 *
873 * All line numbers are 0-based.
874 */
875static void handle_split(struct scoreboard *sb,
876 struct blame_entry *ent,
877 int tlno, int plno, int same,
878 struct origin *parent,
879 struct blame_entry *split)
880{
881 if (ent->num_lines <= tlno)
882 return;
883 if (tlno < same) {
884 struct blame_entry this[3];
885 tlno += ent->s_lno;
886 same += ent->s_lno;
887 split_overlap(this, ent, tlno, plno, same, parent);
888 copy_split_if_better(sb, split, this);
889 decref_split(this);
890 }
891}
892
cdcab133
RS
893struct handle_split_cb_data {
894 struct scoreboard *sb;
895 struct blame_entry *ent;
896 struct origin *parent;
897 struct blame_entry *split;
898 long plno;
899 long tlno;
900};
901
902static void handle_split_cb(void *data, long same, long p_next, long t_next)
903{
904 struct handle_split_cb_data *d = data;
905 handle_split(d->sb, d->ent, d->tlno, d->plno, same, d->parent, d->split);
906 d->plno = p_next;
907 d->tlno = t_next;
908}
909
1732a1fd
JH
910/*
911 * Find the lines from parent that are the same as ent so that
912 * we can pass blames to it. file_p has the blob contents for
913 * the parent.
914 */
d24bba80
JH
915static void find_copy_in_blob(struct scoreboard *sb,
916 struct blame_entry *ent,
917 struct origin *parent,
54a4c617 918 struct blame_entry *split,
d24bba80
JH
919 mmfile_t *file_p)
920{
921 const char *cp;
922 int cnt;
923 mmfile_t file_o;
66dbfd55 924 struct handle_split_cb_data d;
cdcab133
RS
925 xpparam_t xpp;
926 xdemitconf_t xecfg;
66dbfd55
GV
927 memset(&d, 0, sizeof(d));
928 d.sb = sb; d.ent = ent; d.parent = parent; d.split = split;
1732a1fd
JH
929 /*
930 * Prepare mmfile that contains only the lines in ent.
931 */
d24bba80 932 cp = nth_line(sb, ent->lno);
4b25d091 933 file_o.ptr = (char *) cp;
d24bba80
JH
934 cnt = ent->num_lines;
935
936 while (cnt && cp < sb->final_buf + sb->final_buf_size) {
937 if (*cp++ == '\n')
938 cnt--;
939 }
940 file_o.size = cp - file_o.ptr;
941
dd166aa8
JH
942 /*
943 * file_o is a part of final image we are annotating.
944 * file_p partially may match that image.
945 */
cdcab133
RS
946 memset(&xpp, 0, sizeof(xpp));
947 xpp.flags = xdl_opts;
948 memset(&xecfg, 0, sizeof(xecfg));
949 xecfg.ctxlen = 1;
d24bba80 950 memset(split, 0, sizeof(struct blame_entry [3]));
cdcab133 951 xdi_diff_hunks(file_p, &file_o, handle_split_cb, &d, &xpp, &xecfg);
dd166aa8 952 /* remainder, if any, all match the preimage */
cdcab133 953 handle_split(sb, ent, d.tlno, d.plno, ent->num_lines, parent, split);
d24bba80
JH
954}
955
1732a1fd
JH
956/*
957 * See if lines currently target is suspected for can be attributed to
958 * parent.
959 */
d24bba80
JH
960static int find_move_in_parent(struct scoreboard *sb,
961 struct origin *target,
962 struct origin *parent)
963{
650e2f67 964 int last_in_target, made_progress;
46014766 965 struct blame_entry *e, split[3];
d24bba80 966 mmfile_t file_p;
d24bba80
JH
967
968 last_in_target = find_last_in_target(sb, target);
969 if (last_in_target < 0)
970 return 1; /* nothing remains for this target */
971
3b8a12e8 972 fill_origin_blob(&sb->revs->diffopt, parent, &file_p);
c2e525d9 973 if (!file_p.ptr)
d24bba80 974 return 0;
d24bba80 975
650e2f67
JH
976 made_progress = 1;
977 while (made_progress) {
978 made_progress = 0;
979 for (e = sb->ent; e; e = e->next) {
1617fe98
AG
980 if (e->guilty || !same_suspect(e->suspect, target) ||
981 ent_score(sb, e) < blame_move_score)
650e2f67
JH
982 continue;
983 find_copy_in_blob(sb, e, parent, split, &file_p);
984 if (split[1].suspect &&
985 blame_move_score < ent_score(sb, &split[1])) {
986 split_blame(sb, split, e);
987 made_progress = 1;
988 }
989 decref_split(split);
990 }
d24bba80 991 }
d24bba80
JH
992 return 0;
993}
994
33494784
JH
995struct blame_list {
996 struct blame_entry *ent;
997 struct blame_entry split[3];
998};
999
1732a1fd
JH
1000/*
1001 * Count the number of entries the target is suspected for,
1002 * and prepare a list of entry and the best split.
1003 */
33494784
JH
1004static struct blame_list *setup_blame_list(struct scoreboard *sb,
1005 struct origin *target,
1617fe98 1006 int min_score,
33494784
JH
1007 int *num_ents_p)
1008{
1009 struct blame_entry *e;
1010 int num_ents, i;
1011 struct blame_list *blame_list = NULL;
1012
33494784 1013 for (e = sb->ent, num_ents = 0; e; e = e->next)
1617fe98
AG
1014 if (!e->scanned && !e->guilty &&
1015 same_suspect(e->suspect, target) &&
1016 min_score < ent_score(sb, e))
33494784
JH
1017 num_ents++;
1018 if (num_ents) {
1019 blame_list = xcalloc(num_ents, sizeof(struct blame_list));
1020 for (e = sb->ent, i = 0; e; e = e->next)
1617fe98
AG
1021 if (!e->scanned && !e->guilty &&
1022 same_suspect(e->suspect, target) &&
1023 min_score < ent_score(sb, e))
33494784
JH
1024 blame_list[i++].ent = e;
1025 }
1026 *num_ents_p = num_ents;
1027 return blame_list;
1028}
1029
eb46328e
AG
1030/*
1031 * Reset the scanned status on all entries.
1032 */
1033static void reset_scanned_flag(struct scoreboard *sb)
1034{
1035 struct blame_entry *e;
1036 for (e = sb->ent; e; e = e->next)
1037 e->scanned = 0;
1038}
1039
1732a1fd
JH
1040/*
1041 * For lines target is suspected for, see if we can find code movement
1042 * across file boundary from the parent commit. porigin is the path
1043 * in the parent we already tried.
1044 */
18abd745
JH
1045static int find_copy_in_parent(struct scoreboard *sb,
1046 struct origin *target,
1047 struct commit *parent,
1048 struct origin *porigin,
1049 int opt)
1050{
1051 struct diff_options diff_opts;
1052 const char *paths[1];
aec8fa1f 1053 int i, j;
33494784
JH
1054 int retval;
1055 struct blame_list *blame_list;
aec8fa1f 1056 int num_ents;
18abd745 1057
1617fe98 1058 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784 1059 if (!blame_list)
18abd745
JH
1060 return 1; /* nothing remains for this target */
1061
1062 diff_setup(&diff_opts);
8f67f8ae 1063 DIFF_OPT_SET(&diff_opts, RECURSIVE);
18abd745
JH
1064 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
1065
18abd745
JH
1066 paths[0] = NULL;
1067 diff_tree_setup_paths(paths, &diff_opts);
1068 if (diff_setup_done(&diff_opts) < 0)
1069 die("diff-setup");
33494784
JH
1070
1071 /* Try "find copies harder" on new path if requested;
1072 * we do not want to use diffcore_rename() actually to
1073 * match things up; find_copies_harder is set only to
1074 * force diff_tree_sha1() to feed all filepairs to diff_queue,
1075 * and this code needs to be after diff_setup_done(), which
1076 * usually makes find-copies-harder imply copy detection.
1077 */
c63777c0
JH
1078 if ((opt & PICKAXE_BLAME_COPY_HARDEST)
1079 || ((opt & PICKAXE_BLAME_COPY_HARDER)
1080 && (!porigin || strcmp(target->path, porigin->path))))
8f67f8ae 1081 DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
33494784 1082
1cfe7733
JH
1083 if (is_null_sha1(target->commit->object.sha1))
1084 do_diff_cache(parent->tree->object.sha1, &diff_opts);
1085 else
1086 diff_tree_sha1(parent->tree->object.sha1,
1087 target->commit->tree->object.sha1,
1088 "", &diff_opts);
18abd745 1089
8f67f8ae 1090 if (!DIFF_OPT_TST(&diff_opts, FIND_COPIES_HARDER))
33494784 1091 diffcore_std(&diff_opts);
18abd745 1092
33494784
JH
1093 retval = 0;
1094 while (1) {
1095 int made_progress = 0;
1096
1097 for (i = 0; i < diff_queued_diff.nr; i++) {
1098 struct diff_filepair *p = diff_queued_diff.queue[i];
1099 struct origin *norigin;
1100 mmfile_t file_p;
33494784
JH
1101 struct blame_entry this[3];
1102
1103 if (!DIFF_FILE_VALID(p->one))
1104 continue; /* does not exist in parent */
5209ac4d
AG
1105 if (S_ISGITLINK(p->one->mode))
1106 continue; /* ignore git links */
33494784
JH
1107 if (porigin && !strcmp(p->one->path, porigin->path))
1108 /* find_move already dealt with this path */
1109 continue;
1110
1111 norigin = get_origin(sb, parent, p->one->path);
1112 hashcpy(norigin->blob_sha1, p->one->sha1);
90064710 1113 norigin->mode = p->one->mode;
3b8a12e8 1114 fill_origin_blob(&sb->revs->diffopt, norigin, &file_p);
33494784
JH
1115 if (!file_p.ptr)
1116 continue;
1117
1118 for (j = 0; j < num_ents; j++) {
1119 find_copy_in_blob(sb, blame_list[j].ent,
1120 norigin, this, &file_p);
1121 copy_split_if_better(sb, blame_list[j].split,
1122 this);
1123 decref_split(this);
1124 }
33494784
JH
1125 origin_decref(norigin);
1126 }
18abd745 1127
aec8fa1f 1128 for (j = 0; j < num_ents; j++) {
33494784
JH
1129 struct blame_entry *split = blame_list[j].split;
1130 if (split[1].suspect &&
1131 blame_copy_score < ent_score(sb, &split[1])) {
1132 split_blame(sb, split, blame_list[j].ent);
1133 made_progress = 1;
1134 }
eb46328e
AG
1135 else
1136 blame_list[j].ent->scanned = 1;
33494784 1137 decref_split(split);
18abd745 1138 }
33494784 1139 free(blame_list);
aec8fa1f 1140
33494784
JH
1141 if (!made_progress)
1142 break;
1617fe98 1143 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784
JH
1144 if (!blame_list) {
1145 retval = 1;
1146 break;
1147 }
18abd745 1148 }
eb46328e 1149 reset_scanned_flag(sb);
33494784 1150 diff_flush(&diff_opts);
03b69c76 1151 diff_tree_release_paths(&diff_opts);
33494784 1152 return retval;
18abd745
JH
1153}
1154
1732a1fd
JH
1155/*
1156 * The blobs of origin and porigin exactly match, so everything
c2e525d9
JH
1157 * origin is suspected for can be blamed on the parent.
1158 */
1159static void pass_whole_blame(struct scoreboard *sb,
1160 struct origin *origin, struct origin *porigin)
1161{
1162 struct blame_entry *e;
1163
1164 if (!porigin->file.ptr && origin->file.ptr) {
1165 /* Steal its file */
1166 porigin->file = origin->file;
1167 origin->file.ptr = NULL;
1168 }
1169 for (e = sb->ent; e; e = e->next) {
171dccd5 1170 if (!same_suspect(e->suspect, origin))
c2e525d9
JH
1171 continue;
1172 origin_incref(porigin);
1173 origin_decref(e->suspect);
1174 e->suspect = porigin;
1175 }
1176}
1177
69264f46
JH
1178/*
1179 * We pass blame from the current commit to its parents. We keep saying
1180 * "parent" (and "porigin"), but what we mean is to find scapegoat to
1181 * exonerate ourselves.
1182 */
85af7929 1183static struct commit_list *first_scapegoat(struct rev_info *revs, struct commit *commit)
69264f46 1184{
85af7929
JH
1185 if (!reverse)
1186 return commit->parents;
1187 return lookup_decoration(&revs->children, &commit->object);
69264f46
JH
1188}
1189
85af7929 1190static int num_scapegoats(struct rev_info *revs, struct commit *commit)
69264f46
JH
1191{
1192 int cnt;
85af7929 1193 struct commit_list *l = first_scapegoat(revs, commit);
69264f46
JH
1194 for (cnt = 0; l; l = l->next)
1195 cnt++;
1196 return cnt;
1197}
1198
1199#define MAXSG 16
cee7f245 1200
d24bba80 1201static void pass_blame(struct scoreboard *sb, struct origin *origin, int opt)
cee7f245 1202{
85af7929 1203 struct rev_info *revs = sb->revs;
69264f46 1204 int i, pass, num_sg;
cee7f245 1205 struct commit *commit = origin->commit;
69264f46
JH
1206 struct commit_list *sg;
1207 struct origin *sg_buf[MAXSG];
1208 struct origin *porigin, **sg_origin = sg_buf;
1209
85af7929 1210 num_sg = num_scapegoats(revs, commit);
69264f46
JH
1211 if (!num_sg)
1212 goto finish;
1213 else if (num_sg < ARRAY_SIZE(sg_buf))
1214 memset(sg_buf, 0, sizeof(sg_buf));
1215 else
1216 sg_origin = xcalloc(num_sg, sizeof(*sg_origin));
cee7f245 1217
69264f46
JH
1218 /*
1219 * The first pass looks for unrenamed path to optimize for
f69e743d
JH
1220 * common cases, then we look for renames in the second pass.
1221 */
1222 for (pass = 0; pass < 2; pass++) {
1223 struct origin *(*find)(struct scoreboard *,
1224 struct commit *, struct origin *);
1225 find = pass ? find_rename : find_origin;
1226
85af7929 1227 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1228 i < num_sg && sg;
1229 sg = sg->next, i++) {
1230 struct commit *p = sg->item;
0421d9f8 1231 int j, same;
f69e743d 1232
69264f46 1233 if (sg_origin[i])
f69e743d
JH
1234 continue;
1235 if (parse_commit(p))
1236 continue;
0d981c67 1237 porigin = find(sb, p, origin);
f69e743d
JH
1238 if (!porigin)
1239 continue;
1240 if (!hashcmp(porigin->blob_sha1, origin->blob_sha1)) {
c2e525d9 1241 pass_whole_blame(sb, origin, porigin);
f69e743d
JH
1242 origin_decref(porigin);
1243 goto finish;
1244 }
0421d9f8 1245 for (j = same = 0; j < i; j++)
69264f46
JH
1246 if (sg_origin[j] &&
1247 !hashcmp(sg_origin[j]->blob_sha1,
0421d9f8
JH
1248 porigin->blob_sha1)) {
1249 same = 1;
1250 break;
1251 }
1252 if (!same)
69264f46 1253 sg_origin[i] = porigin;
0421d9f8
JH
1254 else
1255 origin_decref(porigin);
cee7f245 1256 }
cee7f245
JH
1257 }
1258
c2e525d9 1259 num_commits++;
85af7929 1260 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1261 i < num_sg && sg;
1262 sg = sg->next, i++) {
1263 struct origin *porigin = sg_origin[i];
cee7f245
JH
1264 if (!porigin)
1265 continue;
96e11709
JH
1266 if (!origin->previous) {
1267 origin_incref(porigin);
1268 origin->previous = porigin;
1269 }
cee7f245 1270 if (pass_blame_to_parent(sb, origin, porigin))
54a4c617 1271 goto finish;
cee7f245 1272 }
d24bba80
JH
1273
1274 /*
1732a1fd 1275 * Optionally find moves in parents' files.
d24bba80
JH
1276 */
1277 if (opt & PICKAXE_BLAME_MOVE)
85af7929 1278 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1279 i < num_sg && sg;
1280 sg = sg->next, i++) {
1281 struct origin *porigin = sg_origin[i];
d24bba80
JH
1282 if (!porigin)
1283 continue;
1284 if (find_move_in_parent(sb, origin, porigin))
54a4c617 1285 goto finish;
d24bba80
JH
1286 }
1287
18abd745 1288 /*
1732a1fd 1289 * Optionally find copies from parents' files.
18abd745
JH
1290 */
1291 if (opt & PICKAXE_BLAME_COPY)
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];
1296 if (find_copy_in_parent(sb, origin, sg->item,
18abd745 1297 porigin, opt))
54a4c617 1298 goto finish;
18abd745 1299 }
54a4c617
JH
1300
1301 finish:
69264f46
JH
1302 for (i = 0; i < num_sg; i++) {
1303 if (sg_origin[i]) {
1304 drop_origin_blob(sg_origin[i]);
1305 origin_decref(sg_origin[i]);
7c3c7962
JH
1306 }
1307 }
1308 drop_origin_blob(origin);
69264f46
JH
1309 if (sg_buf != sg_origin)
1310 free(sg_origin);
cee7f245
JH
1311}
1312
1732a1fd
JH
1313/*
1314 * Information on commits, used for output.
1315 */
9cba13ca 1316struct commit_info {
3a55602e
SP
1317 const char *author;
1318 const char *author_mail;
cee7f245 1319 unsigned long author_time;
3a55602e 1320 const char *author_tz;
cee7f245
JH
1321
1322 /* filled only when asked for details */
3a55602e
SP
1323 const char *committer;
1324 const char *committer_mail;
cee7f245 1325 unsigned long committer_time;
3a55602e 1326 const char *committer_tz;
cee7f245 1327
3a55602e 1328 const char *summary;
cee7f245
JH
1329};
1330
1732a1fd
JH
1331/*
1332 * Parse author/committer line in the commit object buffer
1333 */
cee7f245 1334static void get_ac_line(const char *inbuf, const char *what,
d20d654f
MSO
1335 int person_len, char *person,
1336 int mail_len, char *mail,
3a55602e 1337 unsigned long *time, const char **tz)
cee7f245 1338{
f9567384 1339 int len, tzlen, maillen;
d20d654f 1340 char *tmp, *endp, *timepos, *mailpos;
cee7f245
JH
1341
1342 tmp = strstr(inbuf, what);
1343 if (!tmp)
1344 goto error_out;
1345 tmp += strlen(what);
1346 endp = strchr(tmp, '\n');
1347 if (!endp)
1348 len = strlen(tmp);
1349 else
1350 len = endp - tmp;
d20d654f 1351 if (person_len <= len) {
cee7f245
JH
1352 error_out:
1353 /* Ugh */
d20d654f 1354 *tz = "(unknown)";
c8cba791 1355 strcpy(person, *tz);
d20d654f 1356 strcpy(mail, *tz);
cee7f245
JH
1357 *time = 0;
1358 return;
1359 }
1360 memcpy(person, tmp, len);
1361
1362 tmp = person;
1363 tmp += len;
1364 *tmp = 0;
c8cba791 1365 while (person < tmp && *tmp != ' ')
cee7f245 1366 tmp--;
c8cba791
DR
1367 if (tmp <= person)
1368 goto error_out;
cee7f245 1369 *tz = tmp+1;
f9567384 1370 tzlen = (person+len)-(tmp+1);
cee7f245
JH
1371
1372 *tmp = 0;
c8cba791 1373 while (person < tmp && *tmp != ' ')
cee7f245 1374 tmp--;
c8cba791
DR
1375 if (tmp <= person)
1376 goto error_out;
cee7f245 1377 *time = strtoul(tmp, NULL, 10);
f9567384 1378 timepos = tmp;
cee7f245
JH
1379
1380 *tmp = 0;
9b01f003 1381 while (person < tmp && !(*tmp == ' ' && tmp[1] == '<'))
cee7f245 1382 tmp--;
c8cba791
DR
1383 if (tmp <= person)
1384 return;
d20d654f 1385 mailpos = tmp + 1;
cee7f245 1386 *tmp = 0;
f9567384 1387 maillen = timepos - tmp;
d20d654f 1388 memcpy(mail, mailpos, maillen);
f9567384
JH
1389
1390 if (!mailmap.nr)
1391 return;
1392
1393 /*
1394 * mailmap expansion may make the name longer.
1395 * make room by pushing stuff down.
1396 */
d20d654f 1397 tmp = person + person_len - (tzlen + 1);
f9567384
JH
1398 memmove(tmp, *tz, tzlen);
1399 tmp[tzlen] = 0;
1400 *tz = tmp;
1401
f9567384 1402 /*
d20d654f 1403 * Now, convert both name and e-mail using mailmap
f9567384 1404 */
eeefa7c9 1405 if (map_user(&mailmap, mail+1, mail_len-1, person, tmp-person-1)) {
d20d654f
MSO
1406 /* Add a trailing '>' to email, since map_user returns plain emails
1407 Note: It already has '<', since we replace from mail+1 */
1408 mailpos = memchr(mail, '\0', mail_len);
1409 if (mailpos && mailpos-mail < mail_len - 1) {
1410 *mailpos = '>';
1411 *(mailpos+1) = '\0';
1412 }
1413 }
cee7f245
JH
1414}
1415
1416static void get_commit_info(struct commit *commit,
1417 struct commit_info *ret,
1418 int detailed)
1419{
1420 int len;
ad98a58b
CC
1421 const char *subject;
1422 char *reencoded, *message;
d20d654f
MSO
1423 static char author_name[1024];
1424 static char author_mail[1024];
1425 static char committer_name[1024];
1426 static char committer_mail[1024];
cee7f245
JH
1427 static char summary_buf[1024];
1428
1732a1fd
JH
1429 /*
1430 * We've operated without save_commit_buffer, so
612702e8
JH
1431 * we now need to populate them for output.
1432 */
1433 if (!commit->buffer) {
21666f1a 1434 enum object_type type;
612702e8
JH
1435 unsigned long size;
1436 commit->buffer =
21666f1a 1437 read_sha1_file(commit->object.sha1, &type, &size);
ab43e495
JH
1438 if (!commit->buffer)
1439 die("Cannot read commit %s",
1440 sha1_to_hex(commit->object.sha1));
612702e8 1441 }
69cd8f63
AG
1442 reencoded = reencode_commit_message(commit, NULL);
1443 message = reencoded ? reencoded : commit->buffer;
d20d654f
MSO
1444 ret->author = author_name;
1445 ret->author_mail = author_mail;
69cd8f63 1446 get_ac_line(message, "\nauthor ",
d20d654f
MSO
1447 sizeof(author_name), author_name,
1448 sizeof(author_mail), author_mail,
cee7f245
JH
1449 &ret->author_time, &ret->author_tz);
1450
69cd8f63
AG
1451 if (!detailed) {
1452 free(reencoded);
cee7f245 1453 return;
69cd8f63 1454 }
cee7f245 1455
d20d654f
MSO
1456 ret->committer = committer_name;
1457 ret->committer_mail = committer_mail;
69cd8f63 1458 get_ac_line(message, "\ncommitter ",
d20d654f
MSO
1459 sizeof(committer_name), committer_name,
1460 sizeof(committer_mail), committer_mail,
cee7f245
JH
1461 &ret->committer_time, &ret->committer_tz);
1462
1463 ret->summary = summary_buf;
ad98a58b
CC
1464 len = find_commit_subject(message, &subject);
1465 if (len && len < sizeof(summary_buf)) {
1466 memcpy(summary_buf, subject, len);
1467 summary_buf[len] = 0;
1468 } else {
cee7f245 1469 sprintf(summary_buf, "(%s)", sha1_to_hex(commit->object.sha1));
cee7f245 1470 }
69cd8f63 1471 free(reencoded);
cee7f245
JH
1472}
1473
1732a1fd
JH
1474/*
1475 * To allow LF and other nonportable characters in pathnames,
1476 * they are c-style quoted as needed.
1477 */
46e5e69d
JH
1478static void write_filename_info(const char *path)
1479{
1480 printf("filename ");
663af342 1481 write_name_quoted(path, stdout, '\n');
46e5e69d
JH
1482}
1483
9991030c
JH
1484/*
1485 * Porcelain/Incremental format wants to show a lot of details per
1486 * commit. Instead of repeating this every line, emit it only once,
e86226e3
JK
1487 * the first time each commit appears in the output (unless the
1488 * user has specifically asked for us to repeat).
9991030c 1489 */
e86226e3 1490static int emit_one_suspect_detail(struct origin *suspect, int repeat)
9991030c
JH
1491{
1492 struct commit_info ci;
1493
e86226e3 1494 if (!repeat && (suspect->commit->object.flags & METAINFO_SHOWN))
9991030c
JH
1495 return 0;
1496
1497 suspect->commit->object.flags |= METAINFO_SHOWN;
1498 get_commit_info(suspect->commit, &ci, 1);
1499 printf("author %s\n", ci.author);
1500 printf("author-mail %s\n", ci.author_mail);
1501 printf("author-time %lu\n", ci.author_time);
1502 printf("author-tz %s\n", ci.author_tz);
1503 printf("committer %s\n", ci.committer);
1504 printf("committer-mail %s\n", ci.committer_mail);
1505 printf("committer-time %lu\n", ci.committer_time);
1506 printf("committer-tz %s\n", ci.committer_tz);
1507 printf("summary %s\n", ci.summary);
1508 if (suspect->commit->object.flags & UNINTERESTING)
1509 printf("boundary\n");
96e11709
JH
1510 if (suspect->previous) {
1511 struct origin *prev = suspect->previous;
1512 printf("previous %s ", sha1_to_hex(prev->commit->object.sha1));
1513 write_name_quoted(prev->path, stdout, '\n');
1514 }
9991030c
JH
1515 return 1;
1516}
1517
1732a1fd
JH
1518/*
1519 * The blame_entry is found to be guilty for the range. Mark it
1520 * as such, and show it in incremental output.
1521 */
717d1462
LT
1522static void found_guilty_entry(struct blame_entry *ent)
1523{
1524 if (ent->guilty)
1525 return;
1526 ent->guilty = 1;
1527 if (incremental) {
1528 struct origin *suspect = ent->suspect;
1529
1530 printf("%s %d %d %d\n",
1531 sha1_to_hex(suspect->commit->object.sha1),
1532 ent->s_lno + 1, ent->lno + 1, ent->num_lines);
e86226e3 1533 emit_one_suspect_detail(suspect, 0);
46e5e69d 1534 write_filename_info(suspect->path);
06f59e9f 1535 maybe_flush_or_die(stdout, "stdout");
717d1462
LT
1536 }
1537}
1538
1732a1fd
JH
1539/*
1540 * The main loop -- while the scoreboard has lines whose true origin
3dff5379 1541 * is still unknown, pick one blame_entry, and allow its current
1732a1fd
JH
1542 * suspect to pass blames to its parents.
1543 */
85af7929 1544static void assign_blame(struct scoreboard *sb, int opt)
717d1462 1545{
85af7929
JH
1546 struct rev_info *revs = sb->revs;
1547
717d1462
LT
1548 while (1) {
1549 struct blame_entry *ent;
1550 struct commit *commit;
1551 struct origin *suspect = NULL;
1552
1553 /* find one suspect to break down */
1554 for (ent = sb->ent; !suspect && ent; ent = ent->next)
1555 if (!ent->guilty)
1556 suspect = ent->suspect;
1557 if (!suspect)
1558 return; /* all done */
1559
1732a1fd
JH
1560 /*
1561 * We will use this suspect later in the loop,
1562 * so hold onto it in the meantime.
1563 */
717d1462
LT
1564 origin_incref(suspect);
1565 commit = suspect->commit;
1566 if (!commit->object.parsed)
1567 parse_commit(commit);
85af7929
JH
1568 if (reverse ||
1569 (!(commit->object.flags & UNINTERESTING) &&
1570 !(revs->max_age != -1 && commit->date < revs->max_age)))
717d1462
LT
1571 pass_blame(sb, suspect, opt);
1572 else {
1573 commit->object.flags |= UNINTERESTING;
1574 if (commit->object.parsed)
1575 mark_parents_uninteresting(commit);
1576 }
1577 /* treat root commit as boundary */
1578 if (!commit->parents && !show_root)
1579 commit->object.flags |= UNINTERESTING;
1580
1581 /* Take responsibility for the remaining entries */
1582 for (ent = sb->ent; ent; ent = ent->next)
171dccd5 1583 if (same_suspect(ent->suspect, suspect))
717d1462
LT
1584 found_guilty_entry(ent);
1585 origin_decref(suspect);
1586
1587 if (DEBUG) /* sanity */
1588 sanity_check_refcnt(sb);
1589 }
1590}
1591
1592static const char *format_time(unsigned long time, const char *tz_str,
1593 int show_raw_time)
1594{
1595 static char time_buf[128];
31653c1a
EL
1596 const char *time_str;
1597 int time_len;
1598 int tz;
717d1462
LT
1599
1600 if (show_raw_time) {
1601 sprintf(time_buf, "%lu %s", time, tz_str);
717d1462 1602 }
31653c1a
EL
1603 else {
1604 tz = atoi(tz_str);
1605 time_str = show_date(time, tz, blame_date_mode);
1606 time_len = strlen(time_str);
1607 memcpy(time_buf, time_str, time_len);
1608 memset(time_buf + time_len, ' ', blame_date_width - time_len);
1609 }
717d1462
LT
1610 return time_buf;
1611}
1612
cee7f245
JH
1613#define OUTPUT_ANNOTATE_COMPAT 001
1614#define OUTPUT_LONG_OBJECT_NAME 002
1615#define OUTPUT_RAW_TIMESTAMP 004
1616#define OUTPUT_PORCELAIN 010
1617#define OUTPUT_SHOW_NAME 020
1618#define OUTPUT_SHOW_NUMBER 040
5ff62c30 1619#define OUTPUT_SHOW_SCORE 0100
093dc5be 1620#define OUTPUT_NO_AUTHOR 0200
1b8cdce9 1621#define OUTPUT_SHOW_EMAIL 0400
cee7f245 1622
e86226e3
JK
1623static void emit_porcelain_details(struct origin *suspect, int repeat)
1624{
1625 if (emit_one_suspect_detail(suspect, repeat) ||
1626 (suspect->commit->object.flags & MORE_THAN_ONE_PATH))
1627 write_filename_info(suspect->path);
1628}
1629
1630static void emit_porcelain(struct scoreboard *sb, struct blame_entry *ent,
1631 int opt)
cee7f245
JH
1632{
1633 int cnt;
1634 const char *cp;
1635 struct origin *suspect = ent->suspect;
1636 char hex[41];
1637
1638 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1639 printf("%s%c%d %d %d\n",
1640 hex,
2543d9b6 1641 ent->guilty ? ' ' : '*', /* purely for debugging */
cee7f245
JH
1642 ent->s_lno + 1,
1643 ent->lno + 1,
1644 ent->num_lines);
e86226e3 1645 emit_porcelain_details(suspect, 0);
cee7f245
JH
1646
1647 cp = nth_line(sb, ent->lno);
1648 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1649 char ch;
1650 if (cnt)
1651 printf("%s %d %d\n", hex,
1652 ent->s_lno + 1 + cnt,
1653 ent->lno + 1 + cnt);
1654 putchar('\t');
1655 do {
1656 ch = *cp++;
1657 putchar(ch);
1658 } while (ch != '\n' &&
1659 cp < sb->final_buf + sb->final_buf_size);
1660 }
a5ca8367
JS
1661
1662 if (sb->final_buf_size && cp[-1] != '\n')
1663 putchar('\n');
cee7f245
JH
1664}
1665
1666static void emit_other(struct scoreboard *sb, struct blame_entry *ent, int opt)
1667{
1668 int cnt;
1669 const char *cp;
1670 struct origin *suspect = ent->suspect;
1671 struct commit_info ci;
1672 char hex[41];
1673 int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP);
1674
1675 get_commit_info(suspect->commit, &ci, 1);
1676 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1677
1678 cp = nth_line(sb, ent->lno);
1679 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1680 char ch;
84393bfd 1681 int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? 40 : abbrev;
b11121d9
JH
1682
1683 if (suspect->commit->object.flags & UNINTERESTING) {
e68989a7
JH
1684 if (blank_boundary)
1685 memset(hex, ' ', length);
7ceacdff 1686 else if (!(opt & OUTPUT_ANNOTATE_COMPAT)) {
4c10a5ca
JH
1687 length--;
1688 putchar('^');
1689 }
b11121d9 1690 }
cee7f245 1691
b11121d9 1692 printf("%.*s", length, hex);
1b8cdce9
KB
1693 if (opt & OUTPUT_ANNOTATE_COMPAT) {
1694 const char *name;
1695 if (opt & OUTPUT_SHOW_EMAIL)
1696 name = ci.author_mail;
1697 else
1698 name = ci.author;
1699 printf("\t(%10s\t%10s\t%d)", name,
cee7f245
JH
1700 format_time(ci.author_time, ci.author_tz,
1701 show_raw_time),
1702 ent->lno + 1 + cnt);
1b8cdce9 1703 } else {
5ff62c30 1704 if (opt & OUTPUT_SHOW_SCORE)
54a4c617
JH
1705 printf(" %*d %02d",
1706 max_score_digits, ent->score,
1707 ent->suspect->refcnt);
cee7f245
JH
1708 if (opt & OUTPUT_SHOW_NAME)
1709 printf(" %-*.*s", longest_file, longest_file,
1710 suspect->path);
1711 if (opt & OUTPUT_SHOW_NUMBER)
1712 printf(" %*d", max_orig_digits,
1713 ent->s_lno + 1 + cnt);
093dc5be 1714
ffaf9cc0 1715 if (!(opt & OUTPUT_NO_AUTHOR)) {
1b8cdce9
KB
1716 const char *name;
1717 int pad;
1718 if (opt & OUTPUT_SHOW_EMAIL)
1719 name = ci.author_mail;
1720 else
1721 name = ci.author;
1722 pad = longest_author - utf8_strwidth(name);
ffaf9cc0 1723 printf(" (%s%*s %10s",
1b8cdce9 1724 name, pad, "",
093dc5be
JH
1725 format_time(ci.author_time,
1726 ci.author_tz,
1727 show_raw_time));
ffaf9cc0 1728 }
093dc5be 1729 printf(" %*d) ",
cee7f245
JH
1730 max_digits, ent->lno + 1 + cnt);
1731 }
1732 do {
1733 ch = *cp++;
1734 putchar(ch);
1735 } while (ch != '\n' &&
1736 cp < sb->final_buf + sb->final_buf_size);
1737 }
a5ca8367
JS
1738
1739 if (sb->final_buf_size && cp[-1] != '\n')
1740 putchar('\n');
cee7f245
JH
1741}
1742
1743static void output(struct scoreboard *sb, int option)
1744{
1745 struct blame_entry *ent;
1746
1747 if (option & OUTPUT_PORCELAIN) {
1748 for (ent = sb->ent; ent; ent = ent->next) {
1749 struct blame_entry *oth;
1750 struct origin *suspect = ent->suspect;
1751 struct commit *commit = suspect->commit;
1752 if (commit->object.flags & MORE_THAN_ONE_PATH)
1753 continue;
1754 for (oth = ent->next; oth; oth = oth->next) {
1755 if ((oth->suspect->commit != commit) ||
1756 !strcmp(oth->suspect->path, suspect->path))
1757 continue;
1758 commit->object.flags |= MORE_THAN_ONE_PATH;
1759 break;
1760 }
1761 }
1762 }
1763
1764 for (ent = sb->ent; ent; ent = ent->next) {
1765 if (option & OUTPUT_PORCELAIN)
e86226e3 1766 emit_porcelain(sb, ent, option);
5ff62c30 1767 else {
cee7f245 1768 emit_other(sb, ent, option);
5ff62c30 1769 }
cee7f245
JH
1770 }
1771}
1772
1732a1fd
JH
1773/*
1774 * To allow quick access to the contents of nth line in the
1775 * final image, prepare an index in the scoreboard.
1776 */
cee7f245
JH
1777static int prepare_lines(struct scoreboard *sb)
1778{
1779 const char *buf = sb->final_buf;
1780 unsigned long len = sb->final_buf_size;
1781 int num = 0, incomplete = 0, bol = 1;
1782
1783 if (len && buf[len-1] != '\n')
1784 incomplete++; /* incomplete line at the end */
1785 while (len--) {
1786 if (bol) {
1787 sb->lineno = xrealloc(sb->lineno,
4b25d091 1788 sizeof(int *) * (num + 1));
cee7f245
JH
1789 sb->lineno[num] = buf - sb->final_buf;
1790 bol = 0;
1791 }
1792 if (*buf++ == '\n') {
1793 num++;
1794 bol = 1;
1795 }
1796 }
1ca6ca87 1797 sb->lineno = xrealloc(sb->lineno,
4b25d091 1798 sizeof(int *) * (num + incomplete + 1));
1ca6ca87 1799 sb->lineno[num + incomplete] = buf - sb->final_buf;
cee7f245
JH
1800 sb->num_lines = num + incomplete;
1801 return sb->num_lines;
1802}
1803
1732a1fd
JH
1804/*
1805 * Add phony grafts for use with -S; this is primarily to
34baebce 1806 * support git's cvsserver that wants to give a linear history
1732a1fd
JH
1807 * to its clients.
1808 */
cee7f245
JH
1809static int read_ancestry(const char *graft_file)
1810{
1811 FILE *fp = fopen(graft_file, "r");
1812 char buf[1024];
1813 if (!fp)
1814 return -1;
1815 while (fgets(buf, sizeof(buf), fp)) {
1816 /* The format is just "Commit Parent1 Parent2 ...\n" */
1817 int len = strlen(buf);
1818 struct commit_graft *graft = read_graft_line(buf, len);
8eaf7986
JH
1819 if (graft)
1820 register_commit_graft(graft, 0);
cee7f245
JH
1821 }
1822 fclose(fp);
1823 return 0;
1824}
1825
1732a1fd
JH
1826/*
1827 * How many columns do we need to show line numbers in decimal?
1828 */
cee7f245
JH
1829static int lineno_width(int lines)
1830{
f2915045 1831 int i, width;
cee7f245 1832
00fb3d21 1833 for (width = 1, i = 10; i <= lines; width++)
f2915045
JH
1834 i *= 10;
1835 return width;
cee7f245
JH
1836}
1837
1732a1fd
JH
1838/*
1839 * How many columns do we need to show line numbers, authors,
1840 * and filenames?
1841 */
cee7f245
JH
1842static void find_alignment(struct scoreboard *sb, int *option)
1843{
1844 int longest_src_lines = 0;
1845 int longest_dst_lines = 0;
5ff62c30 1846 unsigned largest_score = 0;
cee7f245
JH
1847 struct blame_entry *e;
1848
1849 for (e = sb->ent; e; e = e->next) {
1850 struct origin *suspect = e->suspect;
1851 struct commit_info ci;
1852 int num;
1853
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
KB
1862 if (*option & OUTPUT_SHOW_EMAIL)
1863 num = utf8_strwidth(ci.author_mail);
1864 else
1865 num = utf8_strwidth(ci.author);
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);
cee7f245
JH
1877 }
1878 max_orig_digits = lineno_width(longest_src_lines);
1879 max_digits = lineno_width(longest_dst_lines);
5ff62c30 1880 max_score_digits = lineno_width(largest_score);
cee7f245
JH
1881}
1882
1732a1fd
JH
1883/*
1884 * For debugging -- origin is refcounted, and this asserts that
1885 * we do not underflow.
1886 */
54a4c617
JH
1887static void sanity_check_refcnt(struct scoreboard *sb)
1888{
1889 int baa = 0;
1890 struct blame_entry *ent;
1891
1892 for (ent = sb->ent; ent; ent = ent->next) {
ae86ad65 1893 /* Nobody should have zero or negative refcnt */
854b97f6
JH
1894 if (ent->suspect->refcnt <= 0) {
1895 fprintf(stderr, "%s in %s has negative refcnt %d\n",
1896 ent->suspect->path,
1897 sha1_to_hex(ent->suspect->commit->object.sha1),
1898 ent->suspect->refcnt);
ae86ad65 1899 baa = 1;
854b97f6 1900 }
ae86ad65 1901 }
54a4c617
JH
1902 if (baa) {
1903 int opt = 0160;
1904 find_alignment(sb, &opt);
1905 output(sb, opt);
854b97f6 1906 die("Baa %d!", baa);
54a4c617
JH
1907 }
1908}
1909
1732a1fd
JH
1910/*
1911 * Used for the command line parsing; check if the path exists
1912 * in the working tree.
1913 */
c455c87c 1914static int has_string_in_work_tree(const char *path)
cee7f245
JH
1915{
1916 struct stat st;
1917 return !lstat(path, &st);
1918}
1919
4a0fc95f
JH
1920static unsigned parse_score(const char *arg)
1921{
1922 char *end;
1923 unsigned long score = strtoul(arg, &end, 10);
1924 if (*end)
1925 return 0;
1926 return score;
1927}
1928
20239bae
JK
1929static const char *add_prefix(const char *prefix, const char *path)
1930{
097971f5 1931 return prefix_path(prefix, prefix ? strlen(prefix) : 0, path);
20239bae
JK
1932}
1933
1732a1fd
JH
1934/*
1935 * Parsing of (comma separated) one item in the -L option
1936 */
931233bc
JH
1937static const char *parse_loc(const char *spec,
1938 struct scoreboard *sb, long lno,
1939 long begin, long *ret)
1940{
1941 char *term;
1942 const char *line;
1943 long num;
1944 int reg_error;
1945 regex_t regexp;
1946 regmatch_t match[1];
1947
7bd9641d
JH
1948 /* Allow "-L <something>,+20" to mean starting at <something>
1949 * for 20 lines, or "-L <something>,-5" for 5 lines ending at
1950 * <something>.
1951 */
1952 if (1 < begin && (spec[0] == '+' || spec[0] == '-')) {
1953 num = strtol(spec + 1, &term, 10);
1954 if (term != spec + 1) {
1955 if (spec[0] == '-')
1956 num = 0 - num;
1957 if (0 < num)
1958 *ret = begin + num - 2;
1959 else if (!num)
1960 *ret = begin;
1961 else
1962 *ret = begin + num;
1963 return term;
1964 }
1965 return spec;
1966 }
931233bc
JH
1967 num = strtol(spec, &term, 10);
1968 if (term != spec) {
1969 *ret = num;
1970 return term;
1971 }
1972 if (spec[0] != '/')
1973 return spec;
1974
1975 /* it could be a regexp of form /.../ */
4b25d091 1976 for (term = (char *) spec + 1; *term && *term != '/'; term++) {
931233bc
JH
1977 if (*term == '\\')
1978 term++;
1979 }
1980 if (*term != '/')
1981 return spec;
1982
1983 /* try [spec+1 .. term-1] as regexp */
1984 *term = 0;
1985 begin--; /* input is in human terms */
1986 line = nth_line(sb, begin);
1987
1988 if (!(reg_error = regcomp(&regexp, spec + 1, REG_NEWLINE)) &&
1989 !(reg_error = regexec(&regexp, line, 1, match, 0))) {
1990 const char *cp = line + match[0].rm_so;
1991 const char *nline;
1992
1993 while (begin++ < lno) {
1994 nline = nth_line(sb, begin);
1995 if (line <= cp && cp < nline)
1996 break;
1997 line = nline;
1998 }
1999 *ret = begin;
2000 regfree(&regexp);
2001 *term++ = '/';
2002 return term;
2003 }
2004 else {
2005 char errbuf[1024];
2006 regerror(reg_error, &regexp, errbuf, 1024);
2007 die("-L parameter '%s': %s", spec + 1, errbuf);
2008 }
2009}
2010
1732a1fd
JH
2011/*
2012 * Parsing of -L option
2013 */
931233bc
JH
2014static void prepare_blame_range(struct scoreboard *sb,
2015 const char *bottomtop,
2016 long lno,
2017 long *bottom, long *top)
2018{
2019 const char *term;
2020
2021 term = parse_loc(bottomtop, sb, lno, 1, bottom);
2022 if (*term == ',') {
2023 term = parse_loc(term + 1, sb, lno, *bottom + 1, top);
2024 if (*term)
acca687f 2025 usage(blame_usage);
931233bc
JH
2026 }
2027 if (*term)
acca687f 2028 usage(blame_usage);
931233bc
JH
2029}
2030
ef90d6d4 2031static int git_blame_config(const char *var, const char *value, void *cb)
4c10a5ca
JH
2032{
2033 if (!strcmp(var, "blame.showroot")) {
2034 show_root = git_config_bool(var, value);
2035 return 0;
2036 }
2037 if (!strcmp(var, "blame.blankboundary")) {
2038 blank_boundary = git_config_bool(var, value);
2039 return 0;
2040 }
31653c1a
EL
2041 if (!strcmp(var, "blame.date")) {
2042 if (!value)
2043 return config_error_nonbool(var);
2044 blame_date_mode = parse_date_format(value);
2045 return 0;
2046 }
3b8a12e8
AB
2047
2048 switch (userdiff_config(var, value)) {
2049 case 0:
2050 break;
2051 case -1:
2052 return -1;
2053 default:
2054 return 0;
2055 }
2056
ef90d6d4 2057 return git_default_config(var, value, cb);
4c10a5ca
JH
2058}
2059
f6c07d7d
JH
2060/*
2061 * Prepare a dummy commit that represents the work tree (or staged) item.
2062 * Note that annotating work tree item never works in the reverse.
2063 */
3b8a12e8
AB
2064static struct commit *fake_working_tree_commit(struct diff_options *opt,
2065 const char *path,
2066 const char *contents_from)
1cfe7733
JH
2067{
2068 struct commit *commit;
2069 struct origin *origin;
2070 unsigned char head_sha1[20];
f285a2d7 2071 struct strbuf buf = STRBUF_INIT;
1cfe7733 2072 const char *ident;
1cfe7733 2073 time_t now;
1cfe7733
JH
2074 int size, len;
2075 struct cache_entry *ce;
2076 unsigned mode;
2077
2078 if (get_sha1("HEAD", head_sha1))
2079 die("No such ref: HEAD");
2080
2081 time(&now);
2082 commit = xcalloc(1, sizeof(*commit));
2083 commit->parents = xcalloc(1, sizeof(*commit->parents));
2084 commit->parents->item = lookup_commit_reference(head_sha1);
2085 commit->object.parsed = 1;
2086 commit->date = now;
2087 commit->object.type = OBJ_COMMIT;
2088
2089 origin = make_origin(commit, path);
2090
2091 if (!contents_from || strcmp("-", contents_from)) {
2092 struct stat st;
2093 const char *read_from;
3b8a12e8 2094 unsigned long buf_len;
1cfe7733
JH
2095
2096 if (contents_from) {
2097 if (stat(contents_from, &st) < 0)
0721c314 2098 die_errno("Cannot stat '%s'", contents_from);
1cfe7733
JH
2099 read_from = contents_from;
2100 }
2101 else {
2102 if (lstat(path, &st) < 0)
0721c314 2103 die_errno("Cannot lstat '%s'", path);
1cfe7733
JH
2104 read_from = path;
2105 }
1cfe7733 2106 mode = canon_mode(st.st_mode);
3b8a12e8 2107
1cfe7733
JH
2108 switch (st.st_mode & S_IFMT) {
2109 case S_IFREG:
3b8a12e8 2110 if (DIFF_OPT_TST(opt, ALLOW_TEXTCONV) &&
90064710 2111 textconv_object(read_from, mode, null_sha1, &buf.buf, &buf_len))
3b8a12e8
AB
2112 buf.len = buf_len;
2113 else if (strbuf_read_file(&buf, read_from, st.st_size) != st.st_size)
0721c314 2114 die_errno("cannot open or read '%s'", read_from);
1cfe7733
JH
2115 break;
2116 case S_IFLNK:
edfd45d2 2117 if (strbuf_readlink(&buf, read_from, st.st_size) < 0)
0721c314 2118 die_errno("cannot readlink '%s'", read_from);
1cfe7733
JH
2119 break;
2120 default:
2121 die("unsupported file type %s", read_from);
2122 }
2123 }
2124 else {
2125 /* Reading from stdin */
2126 contents_from = "standard input";
1cfe7733 2127 mode = 0;
f1696ee3 2128 if (strbuf_read(&buf, 0, 0) < 0)
d824cbba 2129 die_errno("failed to read from stdin");
1cfe7733 2130 }
21e5ad50 2131 convert_to_git(path, buf.buf, buf.len, &buf, 0);
af6eb822
PH
2132 origin->file.ptr = buf.buf;
2133 origin->file.size = buf.len;
2134 pretend_sha1_file(buf.buf, buf.len, OBJ_BLOB, origin->blob_sha1);
1cfe7733
JH
2135 commit->util = origin;
2136
2137 /*
2138 * Read the current index, replace the path entry with
2139 * origin->blob_sha1 without mucking with its mode or type
2140 * bits; we are not going to write this index out -- we just
2141 * want to run "diff-index --cached".
2142 */
2143 discard_cache();
2144 read_cache();
2145
2146 len = strlen(path);
2147 if (!mode) {
2148 int pos = cache_name_pos(path, len);
2149 if (0 <= pos)
7a51ed66 2150 mode = active_cache[pos]->ce_mode;
1cfe7733
JH
2151 else
2152 /* Let's not bother reading from HEAD tree */
2153 mode = S_IFREG | 0644;
2154 }
2155 size = cache_entry_size(len);
2156 ce = xcalloc(1, size);
2157 hashcpy(ce->sha1, origin->blob_sha1);
2158 memcpy(ce->name, path, len);
2159 ce->ce_flags = create_ce_flags(len, 0);
2160 ce->ce_mode = create_ce_mode(mode);
2161 add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
2162
2163 /*
2164 * We are not going to write this out, so this does not matter
2165 * right now, but someday we might optimize diff-index --cached
2166 * with cache-tree information.
2167 */
2168 cache_tree_invalidate_path(active_cache_tree, path);
2169
2170 commit->buffer = xmalloc(400);
2171 ident = fmt_ident("Not Committed Yet", "not.committed.yet", NULL, 0);
1bb88be9 2172 snprintf(commit->buffer, 400,
1cfe7733
JH
2173 "tree 0000000000000000000000000000000000000000\n"
2174 "parent %s\n"
2175 "author %s\n"
2176 "committer %s\n\n"
2177 "Version of %s from %s\n",
2178 sha1_to_hex(head_sha1),
2179 ident, ident, path, contents_from ? contents_from : path);
2180 return commit;
2181}
2182
85af7929 2183static const char *prepare_final(struct scoreboard *sb)
f6c07d7d
JH
2184{
2185 int i;
2186 const char *final_commit_name = NULL;
85af7929 2187 struct rev_info *revs = sb->revs;
f6c07d7d
JH
2188
2189 /*
2190 * There must be one and only one positive commit in the
2191 * revs->pending array.
2192 */
2193 for (i = 0; i < revs->pending.nr; i++) {
2194 struct object *obj = revs->pending.objects[i].item;
2195 if (obj->flags & UNINTERESTING)
2196 continue;
2197 while (obj->type == OBJ_TAG)
2198 obj = deref_tag(obj, NULL, 0);
2199 if (obj->type != OBJ_COMMIT)
2200 die("Non commit %s?", revs->pending.objects[i].name);
2201 if (sb->final)
2202 die("More than one commit to dig from %s and %s?",
2203 revs->pending.objects[i].name,
2204 final_commit_name);
2205 sb->final = (struct commit *) obj;
2206 final_commit_name = revs->pending.objects[i].name;
2207 }
2208 return final_commit_name;
2209}
2210
85af7929
JH
2211static const char *prepare_initial(struct scoreboard *sb)
2212{
2213 int i;
2214 const char *final_commit_name = NULL;
2215 struct rev_info *revs = sb->revs;
2216
2217 /*
2218 * There must be one and only one negative commit, and it must be
2219 * the boundary.
2220 */
2221 for (i = 0; i < revs->pending.nr; i++) {
2222 struct object *obj = revs->pending.objects[i].item;
2223 if (!(obj->flags & UNINTERESTING))
2224 continue;
2225 while (obj->type == OBJ_TAG)
2226 obj = deref_tag(obj, NULL, 0);
2227 if (obj->type != OBJ_COMMIT)
2228 die("Non commit %s?", revs->pending.objects[i].name);
2229 if (sb->final)
2230 die("More than one commit to dig down to %s and %s?",
2231 revs->pending.objects[i].name,
2232 final_commit_name);
2233 sb->final = (struct commit *) obj;
2234 final_commit_name = revs->pending.objects[i].name;
2235 }
2236 if (!final_commit_name)
2237 die("No commit to dig down to?");
2238 return final_commit_name;
2239}
2240
5817da01
PH
2241static int blame_copy_callback(const struct option *option, const char *arg, int unset)
2242{
2243 int *opt = option->value;
2244
2245 /*
2246 * -C enables copy from removed files;
2247 * -C -C enables copy from existing files, but only
2248 * when blaming a new file;
2249 * -C -C -C enables copy from existing files for
2250 * everybody
2251 */
2252 if (*opt & PICKAXE_BLAME_COPY_HARDER)
2253 *opt |= PICKAXE_BLAME_COPY_HARDEST;
2254 if (*opt & PICKAXE_BLAME_COPY)
2255 *opt |= PICKAXE_BLAME_COPY_HARDER;
2256 *opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE;
2257
2258 if (arg)
2259 blame_copy_score = parse_score(arg);
2260 return 0;
2261}
2262
2263static int blame_move_callback(const struct option *option, const char *arg, int unset)
2264{
2265 int *opt = option->value;
2266
2267 *opt |= PICKAXE_BLAME_MOVE;
2268
2269 if (arg)
2270 blame_move_score = parse_score(arg);
2271 return 0;
2272}
2273
2274static int blame_bottomtop_callback(const struct option *option, const char *arg, int unset)
2275{
2276 const char **bottomtop = option->value;
2277 if (!arg)
2278 return -1;
2279 if (*bottomtop)
2280 die("More than one '-L n,m' option given");
2281 *bottomtop = arg;
2282 return 0;
2283}
2284
acca687f 2285int cmd_blame(int argc, const char **argv, const char *prefix)
cee7f245
JH
2286{
2287 struct rev_info revs;
2288 const char *path;
2289 struct scoreboard sb;
2290 struct origin *o;
2291 struct blame_entry *ent;
3f8d5204 2292 long dashdash_pos, bottom, top, lno;
cee7f245 2293 const char *final_commit_name = NULL;
21666f1a 2294 enum object_type type;
5817da01
PH
2295
2296 static const char *bottomtop = NULL;
2297 static int output_option = 0, opt = 0;
2298 static int show_stats = 0;
2299 static const char *revs_file = NULL;
2300 static const char *contents_from = NULL;
2301 static const struct option options[] = {
2302 OPT_BOOLEAN(0, "incremental", &incremental, "Show blame entries as we find them, incrementally"),
2303 OPT_BOOLEAN('b', NULL, &blank_boundary, "Show blank SHA-1 for boundary commits (Default: off)"),
2304 OPT_BOOLEAN(0, "root", &show_root, "Do not treat root commits as boundaries (Default: off)"),
2305 OPT_BOOLEAN(0, "show-stats", &show_stats, "Show work cost statistics"),
2306 OPT_BIT(0, "score-debug", &output_option, "Show output score for blame entries", OUTPUT_SHOW_SCORE),
2307 OPT_BIT('f', "show-name", &output_option, "Show original filename (Default: auto)", OUTPUT_SHOW_NAME),
2308 OPT_BIT('n', "show-number", &output_option, "Show original linenumber (Default: off)", OUTPUT_SHOW_NUMBER),
2309 OPT_BIT('p', "porcelain", &output_option, "Show in a format designed for machine consumption", OUTPUT_PORCELAIN),
2310 OPT_BIT('c', NULL, &output_option, "Use the same output mode as git-annotate (Default: off)", OUTPUT_ANNOTATE_COMPAT),
2311 OPT_BIT('t', NULL, &output_option, "Show raw timestamp (Default: off)", OUTPUT_RAW_TIMESTAMP),
2312 OPT_BIT('l', NULL, &output_option, "Show long commit SHA1 (Default: off)", OUTPUT_LONG_OBJECT_NAME),
2313 OPT_BIT('s', NULL, &output_option, "Suppress author name and timestamp (Default: off)", OUTPUT_NO_AUTHOR),
1b8cdce9 2314 OPT_BIT('e', "show-email", &output_option, "Show author email instead of name (Default: off)", OUTPUT_SHOW_EMAIL),
5817da01
PH
2315 OPT_BIT('w', NULL, &xdl_opts, "Ignore whitespace differences", XDF_IGNORE_WHITESPACE),
2316 OPT_STRING('S', NULL, &revs_file, "file", "Use revisions from <file> instead of calling git-rev-list"),
2317 OPT_STRING(0, "contents", &contents_from, "file", "Use <file>'s contents as the final image"),
2318 { OPTION_CALLBACK, 'C', NULL, &opt, "score", "Find line copies within and across files", PARSE_OPT_OPTARG, blame_copy_callback },
2319 { OPTION_CALLBACK, 'M', NULL, &opt, "score", "Find line movements within and across files", PARSE_OPT_OPTARG, blame_move_callback },
2320 OPT_CALLBACK('L', NULL, &bottomtop, "n,m", "Process only line range n,m, counting from 1", blame_bottomtop_callback),
84393bfd 2321 OPT__ABBREV(&abbrev),
5817da01
PH
2322 OPT_END()
2323 };
2324
2325 struct parse_opt_ctx_t ctx;
7ceacdff 2326 int cmd_is_annotate = !strcmp(argv[0], "annotate");
e68989a7 2327
ef90d6d4 2328 git_config(git_blame_config, NULL);
5817da01 2329 init_revisions(&revs, NULL);
31653c1a 2330 revs.date_mode = blame_date_mode;
3b8a12e8 2331 DIFF_OPT_SET(&revs.diffopt, ALLOW_TEXTCONV);
31653c1a 2332
612702e8 2333 save_commit_buffer = 0;
3f8d5204 2334 dashdash_pos = 0;
612702e8 2335
9ca1169f
SB
2336 parse_options_start(&ctx, argc, argv, prefix, options,
2337 PARSE_OPT_KEEP_DASHDASH | PARSE_OPT_KEEP_ARGV0);
5817da01 2338 for (;;) {
5817da01
PH
2339 switch (parse_options_step(&ctx, options, blame_opt_usage)) {
2340 case PARSE_OPT_HELP:
2341 exit(129);
2342 case PARSE_OPT_DONE:
3f8d5204
PH
2343 if (ctx.argv[0])
2344 dashdash_pos = ctx.cpidx;
5817da01
PH
2345 goto parse_done;
2346 }
2347
2348 if (!strcmp(ctx.argv[0], "--reverse")) {
2349 ctx.argv[0] = "--children";
2350 reverse = 1;
2351 }
6b61ec05 2352 parse_revision_opt(&revs, &ctx, options, blame_opt_usage);
5817da01
PH
2353 }
2354parse_done:
2355 argc = parse_options_end(&ctx);
2356
84393bfd
NK
2357 if (abbrev == -1)
2358 abbrev = default_abbrev;
2359 /* one more abbrev length is needed for the boundary commit */
2360 abbrev++;
2361
aa9ea77d 2362 if (revs_file && read_ancestry(revs_file))
d824cbba 2363 die_errno("reading graft file '%s' failed", revs_file);
aa9ea77d 2364
31653c1a 2365 if (cmd_is_annotate) {
7ceacdff 2366 output_option |= OUTPUT_ANNOTATE_COMPAT;
31653c1a
EL
2367 blame_date_mode = DATE_ISO8601;
2368 } else {
2369 blame_date_mode = revs.date_mode;
2370 }
2371
2372 /* The maximum width used to show the dates */
2373 switch (blame_date_mode) {
2374 case DATE_RFC2822:
2375 blame_date_width = sizeof("Thu, 19 Oct 2006 16:00:04 -0700");
2376 break;
2377 case DATE_ISO8601:
2378 blame_date_width = sizeof("2006-10-19 16:00:04 -0700");
2379 break;
2380 case DATE_RAW:
2381 blame_date_width = sizeof("1161298804 -0700");
2382 break;
2383 case DATE_SHORT:
2384 blame_date_width = sizeof("2006-10-19");
2385 break;
2386 case DATE_RELATIVE:
2387 /* "normal" is used as the fallback for "relative" */
2388 case DATE_LOCAL:
2389 case DATE_NORMAL:
2390 blame_date_width = sizeof("Thu Oct 19 16:00:04 2006 -0700");
2391 break;
2392 }
2393 blame_date_width -= 1; /* strip the null */
7ceacdff 2394
b3123f98
JH
2395 if (DIFF_OPT_TST(&revs.diffopt, FIND_COPIES_HARDER))
2396 opt |= (PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE |
2397 PICKAXE_BLAME_COPY_HARDER);
2398
4a0fc95f
JH
2399 if (!blame_move_score)
2400 blame_move_score = BLAME_DEFAULT_MOVE_SCORE;
2401 if (!blame_copy_score)
2402 blame_copy_score = BLAME_DEFAULT_COPY_SCORE;
2403
1732a1fd
JH
2404 /*
2405 * We have collected options unknown to us in argv[1..unk]
cee7f245 2406 * which are to be passed to revision machinery if we are
3dff5379 2407 * going to do the "bottom" processing.
cee7f245
JH
2408 *
2409 * The remaining are:
2410 *
22e5e58a 2411 * (1) if dashdash_pos != 0, it is either
3f8d5204
PH
2412 * "blame [revisions] -- <path>" or
2413 * "blame -- <path> <rev>"
cee7f245 2414 *
22e5e58a 2415 * (2) otherwise, it is one of the two:
3f8d5204
PH
2416 * "blame [revisions] <path>"
2417 * "blame <path> <rev>"
cee7f245 2418 *
3f8d5204
PH
2419 * Note that we must strip out <path> from the arguments: we do not
2420 * want the path pruning but we may want "bottom" processing.
cee7f245 2421 */
3f8d5204
PH
2422 if (dashdash_pos) {
2423 switch (argc - dashdash_pos - 1) {
2424 case 2: /* (1b) */
2425 if (argc != 4)
5817da01 2426 usage_with_options(blame_opt_usage, options);
3f8d5204
PH
2427 /* reorder for the new way: <rev> -- <path> */
2428 argv[1] = argv[3];
2429 argv[3] = argv[2];
2430 argv[2] = "--";
2431 /* FALLTHROUGH */
2432 case 1: /* (1a) */
2433 path = add_prefix(prefix, argv[--argc]);
2434 argv[argc] = NULL;
2435 break;
2436 default:
2437 usage_with_options(blame_opt_usage, options);
cee7f245 2438 }
3f8d5204
PH
2439 } else {
2440 if (argc < 2)
5817da01 2441 usage_with_options(blame_opt_usage, options);
3f8d5204 2442 path = add_prefix(prefix, argv[argc - 1]);
c455c87c 2443 if (argc == 3 && !has_string_in_work_tree(path)) { /* (2b) */
3f8d5204
PH
2444 path = add_prefix(prefix, argv[1]);
2445 argv[1] = argv[2];
cee7f245 2446 }
3f8d5204 2447 argv[argc - 1] = "--";
cee7f245 2448
3f8d5204 2449 setup_work_tree();
c455c87c 2450 if (!has_string_in_work_tree(path))
d824cbba 2451 die_errno("cannot stat path '%s'", path);
cee7f245
JH
2452 }
2453
8b3dce56 2454 revs.disable_stdin = 1;
3f8d5204 2455 setup_revisions(argc, argv, &revs, NULL);
cee7f245
JH
2456 memset(&sb, 0, sizeof(sb));
2457
85af7929
JH
2458 sb.revs = &revs;
2459 if (!reverse)
2460 final_commit_name = prepare_final(&sb);
2461 else if (contents_from)
2462 die("--contents and --children do not blend well.");
2463 else
2464 final_commit_name = prepare_initial(&sb);
cee7f245
JH
2465
2466 if (!sb.final) {
1732a1fd
JH
2467 /*
2468 * "--not A B -- path" without anything positive;
1cfe7733
JH
2469 * do not default to HEAD, but use the working tree
2470 * or "--contents".
1732a1fd 2471 */
1981820b 2472 setup_work_tree();
3b8a12e8
AB
2473 sb.final = fake_working_tree_commit(&sb.revs->diffopt,
2474 path, contents_from);
1cfe7733 2475 add_pending_object(&revs, &(sb.final->object), ":");
cee7f245 2476 }
1cfe7733
JH
2477 else if (contents_from)
2478 die("Cannot use --contents with final commit object name");
cee7f245 2479
1732a1fd
JH
2480 /*
2481 * If we have bottom, this will mark the ancestors of the
cee7f245
JH
2482 * bottom commits we would reach while traversing as
2483 * uninteresting.
2484 */
3d51e1b5
MK
2485 if (prepare_revision_walk(&revs))
2486 die("revision walk setup failed");
cee7f245 2487
1cfe7733
JH
2488 if (is_null_sha1(sb.final->object.sha1)) {
2489 char *buf;
2490 o = sb.final->util;
2491 buf = xmalloc(o->file.size + 1);
2492 memcpy(buf, o->file.ptr, o->file.size + 1);
2493 sb.final_buf = buf;
2494 sb.final_buf_size = o->file.size;
2495 }
2496 else {
2497 o = get_origin(&sb, sb.final, path);
90064710 2498 if (fill_blob_sha1_and_mode(o))
1cfe7733 2499 die("no such path %s in %s", path, final_commit_name);
cee7f245 2500
3b8a12e8 2501 if (DIFF_OPT_TST(&sb.revs->diffopt, ALLOW_TEXTCONV) &&
90064710 2502 textconv_object(path, o->mode, o->blob_sha1, (char **) &sb.final_buf,
3b8a12e8
AB
2503 &sb.final_buf_size))
2504 ;
2505 else
2506 sb.final_buf = read_sha1_file(o->blob_sha1, &type,
2507 &sb.final_buf_size);
2508
ab43e495
JH
2509 if (!sb.final_buf)
2510 die("Cannot read blob %s for path %s",
2511 sha1_to_hex(o->blob_sha1),
2512 path);
1cfe7733 2513 }
c2e525d9 2514 num_read_blob++;
cee7f245
JH
2515 lno = prepare_lines(&sb);
2516
931233bc
JH
2517 bottom = top = 0;
2518 if (bottomtop)
2519 prepare_blame_range(&sb, bottomtop, lno, &bottom, &top);
2520 if (bottom && top && top < bottom) {
2521 long tmp;
2522 tmp = top; top = bottom; bottom = tmp;
2523 }
cee7f245
JH
2524 if (bottom < 1)
2525 bottom = 1;
2526 if (top < 1)
2527 top = lno;
2528 bottom--;
92f9e273 2529 if (lno < top || lno < bottom)
cee7f245
JH
2530 die("file %s has only %lu lines", path, lno);
2531
2532 ent = xcalloc(1, sizeof(*ent));
2533 ent->lno = bottom;
2534 ent->num_lines = top - bottom;
2535 ent->suspect = o;
2536 ent->s_lno = bottom;
2537
2538 sb.ent = ent;
2539 sb.path = path;
2540
d551a488 2541 read_mailmap(&mailmap, NULL);
f9567384 2542
b92565dc
MH
2543 if (!incremental)
2544 setup_pager();
2545
85af7929 2546 assign_blame(&sb, opt);
cee7f245 2547
717d1462
LT
2548 if (incremental)
2549 return 0;
2550
cee7f245
JH
2551 coalesce(&sb);
2552
2553 if (!(output_option & OUTPUT_PORCELAIN))
2554 find_alignment(&sb, &output_option);
2555
2556 output(&sb, output_option);
2557 free((void *)sb.final_buf);
2558 for (ent = sb.ent; ent; ) {
2559 struct blame_entry *e = ent->next;
2560 free(ent);
2561 ent = e;
2562 }
c2e525d9 2563
870b39c1 2564 if (show_stats) {
c2e525d9
JH
2565 printf("num read blob: %d\n", num_read_blob);
2566 printf("num get patch: %d\n", num_get_patch);
2567 printf("num commits: %d\n", num_commits);
2568 }
cee7f245
JH
2569 return 0;
2570}