convert logmsg_reencode to get_commit_buffer
[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"
58dbfa2e 25#include "line-log.h"
cee7f245 26
efd2a8bd 27static char blame_usage[] = N_("git blame [options] [rev-opts] [rev] [--] file");
5817da01
PH
28
29static const char *blame_opt_usage[] = {
30 blame_usage,
31 "",
efd2a8bd 32 N_("[rev-opts] are documented in git-rev-list(1)"),
5817da01
PH
33 NULL
34};
cee7f245
JH
35
36static int longest_file;
37static int longest_author;
38static int max_orig_digits;
39static int max_digits;
5ff62c30 40static int max_score_digits;
4c10a5ca 41static int show_root;
85af7929 42static int reverse;
4c10a5ca 43static int blank_boundary;
717d1462 44static int incremental;
582aa00b 45static int xdl_opts;
84393bfd 46static int abbrev = -1;
3d1aa566 47static int no_whole_file_rename;
31653c1a
EL
48
49static enum date_mode blame_date_mode = DATE_ISO8601;
50static size_t blame_date_width;
51
c455c87c 52static struct string_list mailmap;
cee7f245 53
54a4c617
JH
54#ifndef DEBUG
55#define DEBUG 0
56#endif
57
c2e525d9
JH
58/* stats */
59static int num_read_blob;
60static int num_get_patch;
61static int num_commits;
62
d24bba80 63#define PICKAXE_BLAME_MOVE 01
18abd745
JH
64#define PICKAXE_BLAME_COPY 02
65#define PICKAXE_BLAME_COPY_HARDER 04
c63777c0 66#define PICKAXE_BLAME_COPY_HARDEST 010
d24bba80 67
4a0fc95f
JH
68/*
69 * blame for a blame_entry with score lower than these thresholds
70 * is not passed to the parent using move/copy logic.
71 */
72static unsigned blame_move_score;
73static unsigned blame_copy_score;
74#define BLAME_DEFAULT_MOVE_SCORE 20
75#define BLAME_DEFAULT_COPY_SCORE 40
76
208acbfb 77/* Remember to update object flag allocation in object.h */
cee7f245
JH
78#define METAINFO_SHOWN (1u<<12)
79#define MORE_THAN_ONE_PATH (1u<<13)
80
81/*
54a4c617 82 * One blob in a commit that is being suspected
cee7f245
JH
83 */
84struct origin {
54a4c617 85 int refcnt;
96e11709 86 struct origin *previous;
cee7f245 87 struct commit *commit;
c2e525d9 88 mmfile_t file;
cee7f245 89 unsigned char blob_sha1[20];
90064710 90 unsigned mode;
cee7f245
JH
91 char path[FLEX_ARRAY];
92};
93
4b4132fd
RS
94static int diff_hunks(mmfile_t *file_a, mmfile_t *file_b, long ctxlen,
95 xdl_emit_hunk_consume_func_t hunk_func, void *cb_data)
96{
97 xpparam_t xpp = {0};
98 xdemitconf_t xecfg = {0};
85c20c30 99 xdemitcb_t ecb = {NULL};
4b4132fd
RS
100
101 xpp.flags = xdl_opts;
102 xecfg.ctxlen = ctxlen;
103 xecfg.hunk_func = hunk_func;
104 ecb.priv = cb_data;
105 return xdi_diff(file_a, file_b, &xpp, &xecfg, &ecb);
106}
107
3b8a12e8
AB
108/*
109 * Prepare diff_filespec and convert it using diff textconv API
110 * if the textconv driver exists.
111 * Return 1 if the conversion succeeds, 0 otherwise.
112 */
e5fba602 113int textconv_object(const char *path,
90064710 114 unsigned mode,
e5fba602 115 const unsigned char *sha1,
e5450100 116 int sha1_valid,
e5fba602
CP
117 char **buf,
118 unsigned long *buf_size)
3b8a12e8
AB
119{
120 struct diff_filespec *df;
121 struct userdiff_driver *textconv;
122
123 df = alloc_filespec(path);
e5450100 124 fill_filespec(df, sha1, sha1_valid, mode);
3b8a12e8
AB
125 textconv = get_textconv(df);
126 if (!textconv) {
127 free_filespec(df);
128 return 0;
129 }
130
131 *buf_size = fill_textconv(textconv, df, buf);
132 free_filespec(df);
133 return 1;
134}
135
1732a1fd
JH
136/*
137 * Given an origin, prepare mmfile_t structure to be used by the
138 * diff machinery
139 */
3b8a12e8
AB
140static void fill_origin_blob(struct diff_options *opt,
141 struct origin *o, mmfile_t *file)
c2e525d9
JH
142{
143 if (!o->file.ptr) {
21666f1a 144 enum object_type type;
3b8a12e8
AB
145 unsigned long file_size;
146
c2e525d9 147 num_read_blob++;
3b8a12e8 148 if (DIFF_OPT_TST(opt, ALLOW_TEXTCONV) &&
e5450100 149 textconv_object(o->path, o->mode, o->blob_sha1, 1, &file->ptr, &file_size))
3b8a12e8
AB
150 ;
151 else
152 file->ptr = read_sha1_file(o->blob_sha1, &type, &file_size);
153 file->size = file_size;
154
ab43e495
JH
155 if (!file->ptr)
156 die("Cannot read blob %s for path %s",
157 sha1_to_hex(o->blob_sha1),
158 o->path);
c2e525d9
JH
159 o->file = *file;
160 }
161 else
162 *file = o->file;
c2e525d9
JH
163}
164
1732a1fd
JH
165/*
166 * Origin is refcounted and usually we keep the blob contents to be
167 * reused.
168 */
54a4c617
JH
169static inline struct origin *origin_incref(struct origin *o)
170{
171 if (o)
172 o->refcnt++;
173 return o;
174}
175
176static void origin_decref(struct origin *o)
177{
178 if (o && --o->refcnt <= 0) {
96e11709
JH
179 if (o->previous)
180 origin_decref(o->previous);
8e0f7003 181 free(o->file.ptr);
54a4c617
JH
182 free(o);
183 }
184}
185
7c3c7962
JH
186static void drop_origin_blob(struct origin *o)
187{
188 if (o->file.ptr) {
189 free(o->file.ptr);
190 o->file.ptr = NULL;
191 }
192}
193
1732a1fd
JH
194/*
195 * Each group of lines is described by a blame_entry; it can be split
196 * as we pass blame to the parents. They form a linked list in the
197 * scoreboard structure, sorted by the target line number.
198 */
cee7f245 199struct blame_entry {
cee7f245
JH
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
54a4c617
JH
258static void sanity_check_refcnt(struct scoreboard *);
259
1732a1fd
JH
260/*
261 * If two blame entries that are next to each other came from
262 * contiguous lines in the same origin (i.e. <commit, path> pair),
263 * merge them together.
264 */
cee7f245
JH
265static void coalesce(struct scoreboard *sb)
266{
267 struct blame_entry *ent, *next;
268
269 for (ent = sb->ent; ent && (next = ent->next); ent = next) {
0a88f08e 270 if (ent->suspect == next->suspect &&
cee7f245
JH
271 ent->guilty == next->guilty &&
272 ent->s_lno + ent->num_lines == next->s_lno) {
273 ent->num_lines += next->num_lines;
274 ent->next = next->next;
54a4c617 275 origin_decref(next->suspect);
cee7f245 276 free(next);
46014766 277 ent->score = 0;
cee7f245
JH
278 next = ent; /* again */
279 }
280 }
54a4c617
JH
281
282 if (DEBUG) /* sanity */
283 sanity_check_refcnt(sb);
cee7f245
JH
284}
285
1732a1fd
JH
286/*
287 * Given a commit and a path in it, create a new origin structure.
288 * The callers that add blame to the scoreboard should use
289 * get_origin() to obtain shared, refcounted copy instead of calling
290 * this function directly.
291 */
854b97f6
JH
292static struct origin *make_origin(struct commit *commit, const char *path)
293{
294 struct origin *o;
295 o = xcalloc(1, sizeof(*o) + strlen(path) + 1);
296 o->commit = commit;
297 o->refcnt = 1;
298 strcpy(o->path, path);
299 return o;
300}
301
1732a1fd
JH
302/*
303 * Locate an existing origin or create a new one.
304 */
f6c0e191
JH
305static struct origin *get_origin(struct scoreboard *sb,
306 struct commit *commit,
307 const char *path)
cee7f245 308{
f6c0e191 309 struct blame_entry *e;
cee7f245 310
f6c0e191
JH
311 for (e = sb->ent; e; e = e->next) {
312 if (e->suspect->commit == commit &&
313 !strcmp(e->suspect->path, path))
54a4c617 314 return origin_incref(e->suspect);
cee7f245 315 }
854b97f6 316 return make_origin(commit, path);
cee7f245
JH
317}
318
1732a1fd
JH
319/*
320 * Fill the blob_sha1 field of an origin if it hasn't, so that later
321 * call to fill_origin_blob() can use it to locate the data. blob_sha1
322 * for an origin is also used to pass the blame for the entire file to
323 * the parent to detect the case where a child's blob is identical to
324 * that of its parent's.
90064710
KS
325 *
326 * This also fills origin->mode for corresponding tree path.
1732a1fd 327 */
90064710 328static int fill_blob_sha1_and_mode(struct origin *origin)
f6c0e191 329{
f6c0e191
JH
330 if (!is_null_sha1(origin->blob_sha1))
331 return 0;
332 if (get_tree_entry(origin->commit->object.sha1,
333 origin->path,
90064710 334 origin->blob_sha1, &origin->mode))
f6c0e191 335 goto error_out;
21666f1a 336 if (sha1_object_info(origin->blob_sha1, NULL) != OBJ_BLOB)
f6c0e191
JH
337 goto error_out;
338 return 0;
339 error_out:
340 hashclr(origin->blob_sha1);
90064710 341 origin->mode = S_IFINVALID;
f6c0e191
JH
342 return -1;
343}
344
1732a1fd
JH
345/*
346 * We have an origin -- check if the same path exists in the
347 * parent and return an origin structure to represent it.
348 */
f6c0e191 349static struct origin *find_origin(struct scoreboard *sb,
cee7f245
JH
350 struct commit *parent,
351 struct origin *origin)
352{
353 struct origin *porigin = NULL;
354 struct diff_options diff_opts;
f6c0e191
JH
355 const char *paths[2];
356
0d981c67 357 if (parent->util) {
1732a1fd
JH
358 /*
359 * Each commit object can cache one origin in that
360 * commit. This is a freestanding copy of origin and
361 * not refcounted.
854b97f6 362 */
0d981c67 363 struct origin *cached = parent->util;
854b97f6 364 if (!strcmp(cached->path, origin->path)) {
1732a1fd
JH
365 /*
366 * The same path between origin and its parent
367 * without renaming -- the most common case.
368 */
854b97f6 369 porigin = get_origin(sb, parent, cached->path);
1732a1fd
JH
370
371 /*
372 * If the origin was newly created (i.e. get_origin
373 * would call make_origin if none is found in the
90064710 374 * scoreboard), it does not know the blob_sha1/mode,
1732a1fd 375 * so copy it. Otherwise porigin was in the
90064710 376 * scoreboard and already knows blob_sha1/mode.
1732a1fd 377 */
90064710 378 if (porigin->refcnt == 1) {
854b97f6 379 hashcpy(porigin->blob_sha1, cached->blob_sha1);
90064710
KS
380 porigin->mode = cached->mode;
381 }
854b97f6
JH
382 return porigin;
383 }
384 /* otherwise it was not very useful; free it */
385 free(parent->util);
386 parent->util = NULL;
0d981c67
JH
387 }
388
f6c0e191
JH
389 /* See if the origin->path is different between parent
390 * and origin first. Most of the time they are the
391 * same and diff-tree is fairly efficient about this.
392 */
393 diff_setup(&diff_opts);
8f67f8ae 394 DIFF_OPT_SET(&diff_opts, RECURSIVE);
f6c0e191
JH
395 diff_opts.detect_rename = 0;
396 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
397 paths[0] = origin->path;
398 paths[1] = NULL;
399
4a2d5ae2
NTND
400 parse_pathspec(&diff_opts.pathspec,
401 PATHSPEC_ALL_MAGIC & ~PATHSPEC_LITERAL,
402 PATHSPEC_LITERAL_PATH, "", paths);
28452655 403 diff_setup_done(&diff_opts);
1cfe7733
JH
404
405 if (is_null_sha1(origin->commit->object.sha1))
406 do_diff_cache(parent->tree->object.sha1, &diff_opts);
407 else
408 diff_tree_sha1(parent->tree->object.sha1,
409 origin->commit->tree->object.sha1,
410 "", &diff_opts);
f6c0e191
JH
411 diffcore_std(&diff_opts);
412
f6c0e191
JH
413 if (!diff_queued_diff.nr) {
414 /* The path is the same as parent */
415 porigin = get_origin(sb, parent, origin->path);
416 hashcpy(porigin->blob_sha1, origin->blob_sha1);
90064710 417 porigin->mode = origin->mode;
a9b2d424
JH
418 } else {
419 /*
420 * Since origin->path is a pathspec, if the parent
421 * commit had it as a directory, we will see a whole
422 * bunch of deletion of files in the directory that we
423 * do not care about.
424 */
425 int i;
426 struct diff_filepair *p = NULL;
427 for (i = 0; i < diff_queued_diff.nr; i++) {
428 const char *name;
429 p = diff_queued_diff.queue[i];
430 name = p->one->path ? p->one->path : p->two->path;
431 if (!strcmp(name, origin->path))
432 break;
433 }
434 if (!p)
435 die("internal error in blame::find_origin");
f6c0e191
JH
436 switch (p->status) {
437 default:
acca687f 438 die("internal error in blame::find_origin (%c)",
f6c0e191
JH
439 p->status);
440 case 'M':
441 porigin = get_origin(sb, parent, origin->path);
442 hashcpy(porigin->blob_sha1, p->one->sha1);
90064710 443 porigin->mode = p->one->mode;
f6c0e191
JH
444 break;
445 case 'A':
446 case 'T':
447 /* Did not exist in parent, or type changed */
448 break;
449 }
450 }
451 diff_flush(&diff_opts);
bd1928df 452 free_pathspec(&diff_opts.pathspec);
0d981c67 453 if (porigin) {
1732a1fd
JH
454 /*
455 * Create a freestanding copy that is not part of
456 * the refcounted origin found in the scoreboard, and
457 * cache it in the commit.
458 */
854b97f6 459 struct origin *cached;
1732a1fd 460
854b97f6
JH
461 cached = make_origin(porigin->commit, porigin->path);
462 hashcpy(cached->blob_sha1, porigin->blob_sha1);
90064710 463 cached->mode = porigin->mode;
854b97f6 464 parent->util = cached;
0d981c67 465 }
f69e743d
JH
466 return porigin;
467}
f6c0e191 468
1732a1fd
JH
469/*
470 * We have an origin -- find the path that corresponds to it in its
471 * parent and return an origin structure to represent it.
472 */
f69e743d
JH
473static struct origin *find_rename(struct scoreboard *sb,
474 struct commit *parent,
475 struct origin *origin)
476{
477 struct origin *porigin = NULL;
478 struct diff_options diff_opts;
479 int i;
cee7f245
JH
480
481 diff_setup(&diff_opts);
8f67f8ae 482 DIFF_OPT_SET(&diff_opts, RECURSIVE);
cee7f245
JH
483 diff_opts.detect_rename = DIFF_DETECT_RENAME;
484 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
2f3f8b21 485 diff_opts.single_follow = origin->path;
28452655 486 diff_setup_done(&diff_opts);
1cfe7733
JH
487
488 if (is_null_sha1(origin->commit->object.sha1))
489 do_diff_cache(parent->tree->object.sha1, &diff_opts);
490 else
491 diff_tree_sha1(parent->tree->object.sha1,
492 origin->commit->tree->object.sha1,
493 "", &diff_opts);
cee7f245
JH
494 diffcore_std(&diff_opts);
495
496 for (i = 0; i < diff_queued_diff.nr; i++) {
497 struct diff_filepair *p = diff_queued_diff.queue[i];
612702e8 498 if ((p->status == 'R' || p->status == 'C') &&
f6c0e191
JH
499 !strcmp(p->two->path, origin->path)) {
500 porigin = get_origin(sb, parent, p->one->path);
501 hashcpy(porigin->blob_sha1, p->one->sha1);
90064710 502 porigin->mode = p->one->mode;
cee7f245
JH
503 break;
504 }
505 }
506 diff_flush(&diff_opts);
bd1928df 507 free_pathspec(&diff_opts.pathspec);
cee7f245
JH
508 return porigin;
509}
510
1732a1fd 511/*
3dff5379 512 * Link in a new blame entry to the scoreboard. Entries that cover the
1732a1fd
JH
513 * same line range have been removed from the scoreboard previously.
514 */
cee7f245
JH
515static void add_blame_entry(struct scoreboard *sb, struct blame_entry *e)
516{
517 struct blame_entry *ent, *prev = NULL;
518
54a4c617
JH
519 origin_incref(e->suspect);
520
cee7f245
JH
521 for (ent = sb->ent; ent && ent->lno < e->lno; ent = ent->next)
522 prev = ent;
523
524 /* prev, if not NULL, is the last one that is below e */
a0f58c58 525
cee7f245
JH
526 if (prev) {
527 e->next = prev->next;
528 prev->next = e;
529 }
530 else {
531 e->next = sb->ent;
532 sb->ent = e;
533 }
cee7f245
JH
534}
535
1732a1fd
JH
536/*
537 * src typically is on-stack; we want to copy the information in it to
790296fd 538 * a malloced blame_entry that is already on the linked list of the
1732a1fd
JH
539 * scoreboard. The origin of dst loses a refcnt while the origin of src
540 * gains one.
541 */
cee7f245
JH
542static void dup_entry(struct blame_entry *dst, struct blame_entry *src)
543{
a0f58c58 544 struct blame_entry *n;
54a4c617 545
cee7f245 546 n = dst->next;
54a4c617
JH
547 origin_incref(src->suspect);
548 origin_decref(dst->suspect);
cee7f245 549 memcpy(dst, src, sizeof(*src));
cee7f245 550 dst->next = n;
5ff62c30 551 dst->score = 0;
cee7f245
JH
552}
553
25ed3412 554static const char *nth_line(struct scoreboard *sb, long lno)
cee7f245
JH
555{
556 return sb->final_buf + sb->lineno[lno];
557}
558
25ed3412
BY
559static const char *nth_line_cb(void *data, long lno)
560{
561 return nth_line((struct scoreboard *)data, lno);
562}
563
1732a1fd
JH
564/*
565 * It is known that lines between tlno to same came from parent, and e
566 * has an overlap with that range. it also is known that parent's
567 * line plno corresponds to e's line tlno.
568 *
569 * <---- e ----->
570 * <------>
571 * <------------>
572 * <------------>
573 * <------------------>
574 *
575 * Split e into potentially three parts; before this chunk, the chunk
576 * to be blamed for the parent, and after that portion.
577 */
54a4c617 578static void split_overlap(struct blame_entry *split,
cee7f245
JH
579 struct blame_entry *e,
580 int tlno, int plno, int same,
581 struct origin *parent)
582{
cee7f245
JH
583 int chunk_end_lno;
584 memset(split, 0, sizeof(struct blame_entry [3]));
585
586 if (e->s_lno < tlno) {
587 /* there is a pre-chunk part not blamed on parent */
54a4c617 588 split[0].suspect = origin_incref(e->suspect);
cee7f245
JH
589 split[0].lno = e->lno;
590 split[0].s_lno = e->s_lno;
591 split[0].num_lines = tlno - e->s_lno;
592 split[1].lno = e->lno + tlno - e->s_lno;
593 split[1].s_lno = plno;
594 }
595 else {
596 split[1].lno = e->lno;
597 split[1].s_lno = plno + (e->s_lno - tlno);
598 }
599
600 if (same < e->s_lno + e->num_lines) {
601 /* there is a post-chunk part not blamed on parent */
54a4c617 602 split[2].suspect = origin_incref(e->suspect);
cee7f245
JH
603 split[2].lno = e->lno + (same - e->s_lno);
604 split[2].s_lno = e->s_lno + (same - e->s_lno);
605 split[2].num_lines = e->s_lno + e->num_lines - same;
606 chunk_end_lno = split[2].lno;
607 }
608 else
609 chunk_end_lno = e->lno + e->num_lines;
610 split[1].num_lines = chunk_end_lno - split[1].lno;
611
1732a1fd
JH
612 /*
613 * if it turns out there is nothing to blame the parent for,
614 * forget about the splitting. !split[1].suspect signals this.
615 */
cee7f245
JH
616 if (split[1].num_lines < 1)
617 return;
54a4c617 618 split[1].suspect = origin_incref(parent);
cee7f245
JH
619}
620
1732a1fd
JH
621/*
622 * split_overlap() divided an existing blame e into up to three parts
623 * in split. Adjust the linked list of blames in the scoreboard to
624 * reflect the split.
625 */
cee7f245 626static void split_blame(struct scoreboard *sb,
54a4c617 627 struct blame_entry *split,
cee7f245
JH
628 struct blame_entry *e)
629{
630 struct blame_entry *new_entry;
631
632 if (split[0].suspect && split[2].suspect) {
1732a1fd 633 /* The first part (reuse storage for the existing entry e) */
cee7f245
JH
634 dup_entry(e, &split[0]);
635
1732a1fd 636 /* The last part -- me */
cee7f245
JH
637 new_entry = xmalloc(sizeof(*new_entry));
638 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
639 add_blame_entry(sb, new_entry);
640
1732a1fd 641 /* ... and the middle part -- parent */
cee7f245
JH
642 new_entry = xmalloc(sizeof(*new_entry));
643 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
644 add_blame_entry(sb, new_entry);
645 }
646 else if (!split[0].suspect && !split[2].suspect)
1732a1fd
JH
647 /*
648 * The parent covers the entire area; reuse storage for
649 * e and replace it with the parent.
650 */
cee7f245
JH
651 dup_entry(e, &split[1]);
652 else if (split[0].suspect) {
1732a1fd 653 /* me and then parent */
cee7f245
JH
654 dup_entry(e, &split[0]);
655
656 new_entry = xmalloc(sizeof(*new_entry));
657 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
658 add_blame_entry(sb, new_entry);
659 }
660 else {
1732a1fd 661 /* parent and then me */
cee7f245
JH
662 dup_entry(e, &split[1]);
663
664 new_entry = xmalloc(sizeof(*new_entry));
665 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
666 add_blame_entry(sb, new_entry);
667 }
668
54a4c617 669 if (DEBUG) { /* sanity */
cee7f245 670 struct blame_entry *ent;
612702e8 671 int lno = sb->ent->lno, corrupt = 0;
cee7f245
JH
672
673 for (ent = sb->ent; ent; ent = ent->next) {
674 if (lno != ent->lno)
675 corrupt = 1;
676 if (ent->s_lno < 0)
677 corrupt = 1;
678 lno += ent->num_lines;
679 }
680 if (corrupt) {
612702e8 681 lno = sb->ent->lno;
cee7f245
JH
682 for (ent = sb->ent; ent; ent = ent->next) {
683 printf("L %8d l %8d n %8d\n",
684 lno, ent->lno, ent->num_lines);
685 lno = ent->lno + ent->num_lines;
686 }
687 die("oops");
688 }
689 }
690}
691
1732a1fd
JH
692/*
693 * After splitting the blame, the origins used by the
694 * on-stack blame_entry should lose one refcnt each.
695 */
54a4c617
JH
696static void decref_split(struct blame_entry *split)
697{
698 int i;
699
700 for (i = 0; i < 3; i++)
701 origin_decref(split[i].suspect);
702}
703
1732a1fd
JH
704/*
705 * Helper for blame_chunk(). blame_entry e is known to overlap with
706 * the patch hunk; split it and pass blame to the parent.
707 */
cee7f245
JH
708static void blame_overlap(struct scoreboard *sb, struct blame_entry *e,
709 int tlno, int plno, int same,
710 struct origin *parent)
711{
712 struct blame_entry split[3];
713
714 split_overlap(split, e, tlno, plno, same, parent);
54a4c617
JH
715 if (split[1].suspect)
716 split_blame(sb, split, e);
717 decref_split(split);
cee7f245
JH
718}
719
1732a1fd
JH
720/*
721 * Find the line number of the last line the target is suspected for.
722 */
cee7f245
JH
723static int find_last_in_target(struct scoreboard *sb, struct origin *target)
724{
725 struct blame_entry *e;
726 int last_in_target = -1;
727
728 for (e = sb->ent; e; e = e->next) {
0a88f08e 729 if (e->guilty || e->suspect != target)
cee7f245
JH
730 continue;
731 if (last_in_target < e->s_lno + e->num_lines)
732 last_in_target = e->s_lno + e->num_lines;
733 }
734 return last_in_target;
735}
736
1732a1fd
JH
737/*
738 * Process one hunk from the patch between the current suspect for
739 * blame_entry e and its parent. Find and split the overlap, and
740 * pass blame to the overlapping part to the parent.
741 */
cee7f245
JH
742static void blame_chunk(struct scoreboard *sb,
743 int tlno, int plno, int same,
744 struct origin *target, struct origin *parent)
745{
612702e8 746 struct blame_entry *e;
cee7f245 747
612702e8 748 for (e = sb->ent; e; e = e->next) {
0a88f08e 749 if (e->guilty || e->suspect != target)
cee7f245
JH
750 continue;
751 if (same <= e->s_lno)
752 continue;
753 if (tlno < e->s_lno + e->num_lines)
754 blame_overlap(sb, e, tlno, plno, same, parent);
755 }
756}
757
cdcab133
RS
758struct blame_chunk_cb_data {
759 struct scoreboard *sb;
760 struct origin *target;
761 struct origin *parent;
762 long plno;
763 long tlno;
764};
765
0af596c6
RS
766static int blame_chunk_cb(long start_a, long count_a,
767 long start_b, long count_b, void *data)
cdcab133
RS
768{
769 struct blame_chunk_cb_data *d = data;
0af596c6
RS
770 blame_chunk(d->sb, d->tlno, d->plno, start_b, d->target, d->parent);
771 d->plno = start_a + count_a;
772 d->tlno = start_b + count_b;
773 return 0;
cdcab133
RS
774}
775
1732a1fd
JH
776/*
777 * We are looking at the origin 'target' and aiming to pass blame
778 * for the lines it is suspected to its parent. Run diff to find
779 * which lines came from parent and pass blame for them.
780 */
cee7f245
JH
781static int pass_blame_to_parent(struct scoreboard *sb,
782 struct origin *target,
783 struct origin *parent)
784{
cdcab133 785 int last_in_target;
7c4ed8c8 786 mmfile_t file_p, file_o;
66dbfd55 787 struct blame_chunk_cb_data d;
0af596c6 788
66dbfd55
GV
789 memset(&d, 0, sizeof(d));
790 d.sb = sb; d.target = target; d.parent = parent;
cee7f245
JH
791 last_in_target = find_last_in_target(sb, target);
792 if (last_in_target < 0)
793 return 1; /* nothing remains for this target */
794
3b8a12e8
AB
795 fill_origin_blob(&sb->revs->diffopt, parent, &file_p);
796 fill_origin_blob(&sb->revs->diffopt, target, &file_o);
7c4ed8c8 797 num_get_patch++;
cee7f245 798
4b4132fd 799 diff_hunks(&file_p, &file_o, 0, blame_chunk_cb, &d);
1732a1fd 800 /* The rest (i.e. anything after tlno) are the same as the parent */
cdcab133 801 blame_chunk(sb, d.tlno, d.plno, last_in_target, target, parent);
cee7f245 802
cee7f245
JH
803 return 0;
804}
805
1732a1fd
JH
806/*
807 * The lines in blame_entry after splitting blames many times can become
808 * very small and trivial, and at some point it becomes pointless to
809 * blame the parents. E.g. "\t\t}\n\t}\n\n" appears everywhere in any
810 * ordinary C program, and it is not worth to say it was copied from
811 * totally unrelated file in the parent.
812 *
813 * Compute how trivial the lines in the blame_entry are.
814 */
5ff62c30
JH
815static unsigned ent_score(struct scoreboard *sb, struct blame_entry *e)
816{
817 unsigned score;
818 const char *cp, *ep;
819
820 if (e->score)
821 return e->score;
822
612702e8 823 score = 1;
5ff62c30
JH
824 cp = nth_line(sb, e->lno);
825 ep = nth_line(sb, e->lno + e->num_lines);
826 while (cp < ep) {
827 unsigned ch = *((unsigned char *)cp);
828 if (isalnum(ch))
829 score++;
830 cp++;
831 }
832 e->score = score;
833 return score;
834}
835
1732a1fd
JH
836/*
837 * best_so_far[] and this[] are both a split of an existing blame_entry
838 * that passes blame to the parent. Maintain best_so_far the best split
839 * so far, by comparing this and best_so_far and copying this into
840 * bst_so_far as needed.
841 */
5ff62c30 842static void copy_split_if_better(struct scoreboard *sb,
54a4c617
JH
843 struct blame_entry *best_so_far,
844 struct blame_entry *this)
d24bba80 845{
54a4c617
JH
846 int i;
847
d24bba80
JH
848 if (!this[1].suspect)
849 return;
5ff62c30
JH
850 if (best_so_far[1].suspect) {
851 if (ent_score(sb, &this[1]) < ent_score(sb, &best_so_far[1]))
852 return;
853 }
54a4c617
JH
854
855 for (i = 0; i < 3; i++)
856 origin_incref(this[i].suspect);
857 decref_split(best_so_far);
d24bba80
JH
858 memcpy(best_so_far, this, sizeof(struct blame_entry [3]));
859}
860
dd166aa8
JH
861/*
862 * We are looking at a part of the final image represented by
863 * ent (tlno and same are offset by ent->s_lno).
864 * tlno is where we are looking at in the final image.
865 * up to (but not including) same match preimage.
866 * plno is where we are looking at in the preimage.
867 *
868 * <-------------- final image ---------------------->
869 * <------ent------>
870 * ^tlno ^same
871 * <---------preimage----->
872 * ^plno
873 *
874 * All line numbers are 0-based.
875 */
876static void handle_split(struct scoreboard *sb,
877 struct blame_entry *ent,
878 int tlno, int plno, int same,
879 struct origin *parent,
880 struct blame_entry *split)
881{
882 if (ent->num_lines <= tlno)
883 return;
884 if (tlno < same) {
885 struct blame_entry this[3];
886 tlno += ent->s_lno;
887 same += ent->s_lno;
888 split_overlap(this, ent, tlno, plno, same, parent);
889 copy_split_if_better(sb, split, this);
890 decref_split(this);
891 }
892}
893
cdcab133
RS
894struct handle_split_cb_data {
895 struct scoreboard *sb;
896 struct blame_entry *ent;
897 struct origin *parent;
898 struct blame_entry *split;
899 long plno;
900 long tlno;
901};
902
5d23ec76
RS
903static int handle_split_cb(long start_a, long count_a,
904 long start_b, long count_b, void *data)
cdcab133
RS
905{
906 struct handle_split_cb_data *d = data;
5d23ec76
RS
907 handle_split(d->sb, d->ent, d->tlno, d->plno, start_b, d->parent,
908 d->split);
909 d->plno = start_a + count_a;
910 d->tlno = start_b + count_b;
911 return 0;
cdcab133
RS
912}
913
1732a1fd
JH
914/*
915 * Find the lines from parent that are the same as ent so that
916 * we can pass blames to it. file_p has the blob contents for
917 * the parent.
918 */
d24bba80
JH
919static void find_copy_in_blob(struct scoreboard *sb,
920 struct blame_entry *ent,
921 struct origin *parent,
54a4c617 922 struct blame_entry *split,
d24bba80
JH
923 mmfile_t *file_p)
924{
925 const char *cp;
d24bba80 926 mmfile_t file_o;
66dbfd55 927 struct handle_split_cb_data d;
5d23ec76 928
66dbfd55
GV
929 memset(&d, 0, sizeof(d));
930 d.sb = sb; d.ent = ent; d.parent = parent; d.split = split;
1732a1fd
JH
931 /*
932 * Prepare mmfile that contains only the lines in ent.
933 */
d24bba80 934 cp = nth_line(sb, ent->lno);
4b25d091 935 file_o.ptr = (char *) cp;
3ee8944f 936 file_o.size = nth_line(sb, ent->lno + ent->num_lines) - cp;
d24bba80 937
dd166aa8
JH
938 /*
939 * file_o is a part of final image we are annotating.
940 * file_p partially may match that image.
941 */
d24bba80 942 memset(split, 0, sizeof(struct blame_entry [3]));
4b4132fd 943 diff_hunks(file_p, &file_o, 1, handle_split_cb, &d);
dd166aa8 944 /* remainder, if any, all match the preimage */
cdcab133 945 handle_split(sb, ent, d.tlno, d.plno, ent->num_lines, parent, split);
d24bba80
JH
946}
947
1732a1fd
JH
948/*
949 * See if lines currently target is suspected for can be attributed to
950 * parent.
951 */
d24bba80
JH
952static int find_move_in_parent(struct scoreboard *sb,
953 struct origin *target,
954 struct origin *parent)
955{
650e2f67 956 int last_in_target, made_progress;
46014766 957 struct blame_entry *e, split[3];
d24bba80 958 mmfile_t file_p;
d24bba80
JH
959
960 last_in_target = find_last_in_target(sb, target);
961 if (last_in_target < 0)
962 return 1; /* nothing remains for this target */
963
3b8a12e8 964 fill_origin_blob(&sb->revs->diffopt, parent, &file_p);
c2e525d9 965 if (!file_p.ptr)
d24bba80 966 return 0;
d24bba80 967
650e2f67
JH
968 made_progress = 1;
969 while (made_progress) {
970 made_progress = 0;
971 for (e = sb->ent; e; e = e->next) {
0a88f08e 972 if (e->guilty || e->suspect != target ||
1617fe98 973 ent_score(sb, e) < blame_move_score)
650e2f67
JH
974 continue;
975 find_copy_in_blob(sb, e, parent, split, &file_p);
976 if (split[1].suspect &&
977 blame_move_score < ent_score(sb, &split[1])) {
978 split_blame(sb, split, e);
979 made_progress = 1;
980 }
981 decref_split(split);
982 }
d24bba80 983 }
d24bba80
JH
984 return 0;
985}
986
33494784
JH
987struct blame_list {
988 struct blame_entry *ent;
989 struct blame_entry split[3];
990};
991
1732a1fd
JH
992/*
993 * Count the number of entries the target is suspected for,
994 * and prepare a list of entry and the best split.
995 */
33494784
JH
996static struct blame_list *setup_blame_list(struct scoreboard *sb,
997 struct origin *target,
1617fe98 998 int min_score,
33494784
JH
999 int *num_ents_p)
1000{
1001 struct blame_entry *e;
1002 int num_ents, i;
1003 struct blame_list *blame_list = NULL;
1004
33494784 1005 for (e = sb->ent, num_ents = 0; e; e = e->next)
1617fe98 1006 if (!e->scanned && !e->guilty &&
0a88f08e 1007 e->suspect == target &&
1617fe98 1008 min_score < ent_score(sb, e))
33494784
JH
1009 num_ents++;
1010 if (num_ents) {
1011 blame_list = xcalloc(num_ents, sizeof(struct blame_list));
1012 for (e = sb->ent, i = 0; e; e = e->next)
1617fe98 1013 if (!e->scanned && !e->guilty &&
0a88f08e 1014 e->suspect == target &&
1617fe98 1015 min_score < ent_score(sb, e))
33494784
JH
1016 blame_list[i++].ent = e;
1017 }
1018 *num_ents_p = num_ents;
1019 return blame_list;
1020}
1021
eb46328e
AG
1022/*
1023 * Reset the scanned status on all entries.
1024 */
1025static void reset_scanned_flag(struct scoreboard *sb)
1026{
1027 struct blame_entry *e;
1028 for (e = sb->ent; e; e = e->next)
1029 e->scanned = 0;
1030}
1031
1732a1fd
JH
1032/*
1033 * For lines target is suspected for, see if we can find code movement
1034 * across file boundary from the parent commit. porigin is the path
1035 * in the parent we already tried.
1036 */
18abd745
JH
1037static int find_copy_in_parent(struct scoreboard *sb,
1038 struct origin *target,
1039 struct commit *parent,
1040 struct origin *porigin,
1041 int opt)
1042{
1043 struct diff_options diff_opts;
aec8fa1f 1044 int i, j;
33494784
JH
1045 int retval;
1046 struct blame_list *blame_list;
aec8fa1f 1047 int num_ents;
18abd745 1048
1617fe98 1049 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784 1050 if (!blame_list)
18abd745
JH
1051 return 1; /* nothing remains for this target */
1052
1053 diff_setup(&diff_opts);
8f67f8ae 1054 DIFF_OPT_SET(&diff_opts, RECURSIVE);
18abd745
JH
1055 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
1056
28452655 1057 diff_setup_done(&diff_opts);
33494784
JH
1058
1059 /* Try "find copies harder" on new path if requested;
1060 * we do not want to use diffcore_rename() actually to
1061 * match things up; find_copies_harder is set only to
1062 * force diff_tree_sha1() to feed all filepairs to diff_queue,
1063 * and this code needs to be after diff_setup_done(), which
1064 * usually makes find-copies-harder imply copy detection.
1065 */
c63777c0
JH
1066 if ((opt & PICKAXE_BLAME_COPY_HARDEST)
1067 || ((opt & PICKAXE_BLAME_COPY_HARDER)
1068 && (!porigin || strcmp(target->path, porigin->path))))
8f67f8ae 1069 DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
33494784 1070
1cfe7733
JH
1071 if (is_null_sha1(target->commit->object.sha1))
1072 do_diff_cache(parent->tree->object.sha1, &diff_opts);
1073 else
1074 diff_tree_sha1(parent->tree->object.sha1,
1075 target->commit->tree->object.sha1,
1076 "", &diff_opts);
18abd745 1077
8f67f8ae 1078 if (!DIFF_OPT_TST(&diff_opts, FIND_COPIES_HARDER))
33494784 1079 diffcore_std(&diff_opts);
18abd745 1080
33494784
JH
1081 retval = 0;
1082 while (1) {
1083 int made_progress = 0;
1084
1085 for (i = 0; i < diff_queued_diff.nr; i++) {
1086 struct diff_filepair *p = diff_queued_diff.queue[i];
1087 struct origin *norigin;
1088 mmfile_t file_p;
33494784
JH
1089 struct blame_entry this[3];
1090
1091 if (!DIFF_FILE_VALID(p->one))
1092 continue; /* does not exist in parent */
5209ac4d
AG
1093 if (S_ISGITLINK(p->one->mode))
1094 continue; /* ignore git links */
33494784
JH
1095 if (porigin && !strcmp(p->one->path, porigin->path))
1096 /* find_move already dealt with this path */
1097 continue;
1098
1099 norigin = get_origin(sb, parent, p->one->path);
1100 hashcpy(norigin->blob_sha1, p->one->sha1);
90064710 1101 norigin->mode = p->one->mode;
3b8a12e8 1102 fill_origin_blob(&sb->revs->diffopt, norigin, &file_p);
33494784
JH
1103 if (!file_p.ptr)
1104 continue;
1105
1106 for (j = 0; j < num_ents; j++) {
1107 find_copy_in_blob(sb, blame_list[j].ent,
1108 norigin, this, &file_p);
1109 copy_split_if_better(sb, blame_list[j].split,
1110 this);
1111 decref_split(this);
1112 }
33494784
JH
1113 origin_decref(norigin);
1114 }
18abd745 1115
aec8fa1f 1116 for (j = 0; j < num_ents; j++) {
33494784
JH
1117 struct blame_entry *split = blame_list[j].split;
1118 if (split[1].suspect &&
1119 blame_copy_score < ent_score(sb, &split[1])) {
1120 split_blame(sb, split, blame_list[j].ent);
1121 made_progress = 1;
1122 }
eb46328e
AG
1123 else
1124 blame_list[j].ent->scanned = 1;
33494784 1125 decref_split(split);
18abd745 1126 }
33494784 1127 free(blame_list);
aec8fa1f 1128
33494784
JH
1129 if (!made_progress)
1130 break;
1617fe98 1131 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784
JH
1132 if (!blame_list) {
1133 retval = 1;
1134 break;
1135 }
18abd745 1136 }
eb46328e 1137 reset_scanned_flag(sb);
33494784 1138 diff_flush(&diff_opts);
bd1928df 1139 free_pathspec(&diff_opts.pathspec);
33494784 1140 return retval;
18abd745
JH
1141}
1142
1732a1fd
JH
1143/*
1144 * The blobs of origin and porigin exactly match, so everything
c2e525d9
JH
1145 * origin is suspected for can be blamed on the parent.
1146 */
1147static void pass_whole_blame(struct scoreboard *sb,
1148 struct origin *origin, struct origin *porigin)
1149{
1150 struct blame_entry *e;
1151
1152 if (!porigin->file.ptr && origin->file.ptr) {
1153 /* Steal its file */
1154 porigin->file = origin->file;
1155 origin->file.ptr = NULL;
1156 }
1157 for (e = sb->ent; e; e = e->next) {
0a88f08e 1158 if (e->suspect != origin)
c2e525d9
JH
1159 continue;
1160 origin_incref(porigin);
1161 origin_decref(e->suspect);
1162 e->suspect = porigin;
1163 }
1164}
1165
69264f46
JH
1166/*
1167 * We pass blame from the current commit to its parents. We keep saying
1168 * "parent" (and "porigin"), but what we mean is to find scapegoat to
1169 * exonerate ourselves.
1170 */
85af7929 1171static struct commit_list *first_scapegoat(struct rev_info *revs, struct commit *commit)
69264f46 1172{
85af7929
JH
1173 if (!reverse)
1174 return commit->parents;
1175 return lookup_decoration(&revs->children, &commit->object);
69264f46
JH
1176}
1177
85af7929 1178static int num_scapegoats(struct rev_info *revs, struct commit *commit)
69264f46
JH
1179{
1180 int cnt;
85af7929 1181 struct commit_list *l = first_scapegoat(revs, commit);
69264f46
JH
1182 for (cnt = 0; l; l = l->next)
1183 cnt++;
1184 return cnt;
1185}
1186
1187#define MAXSG 16
cee7f245 1188
d24bba80 1189static void pass_blame(struct scoreboard *sb, struct origin *origin, int opt)
cee7f245 1190{
85af7929 1191 struct rev_info *revs = sb->revs;
69264f46 1192 int i, pass, num_sg;
cee7f245 1193 struct commit *commit = origin->commit;
69264f46
JH
1194 struct commit_list *sg;
1195 struct origin *sg_buf[MAXSG];
1196 struct origin *porigin, **sg_origin = sg_buf;
1197
85af7929 1198 num_sg = num_scapegoats(revs, commit);
69264f46
JH
1199 if (!num_sg)
1200 goto finish;
1201 else if (num_sg < ARRAY_SIZE(sg_buf))
1202 memset(sg_buf, 0, sizeof(sg_buf));
1203 else
1204 sg_origin = xcalloc(num_sg, sizeof(*sg_origin));
cee7f245 1205
69264f46
JH
1206 /*
1207 * The first pass looks for unrenamed path to optimize for
f69e743d
JH
1208 * common cases, then we look for renames in the second pass.
1209 */
3d1aa566 1210 for (pass = 0; pass < 2 - no_whole_file_rename; pass++) {
f69e743d
JH
1211 struct origin *(*find)(struct scoreboard *,
1212 struct commit *, struct origin *);
1213 find = pass ? find_rename : find_origin;
1214
85af7929 1215 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1216 i < num_sg && sg;
1217 sg = sg->next, i++) {
1218 struct commit *p = sg->item;
0421d9f8 1219 int j, same;
f69e743d 1220
69264f46 1221 if (sg_origin[i])
f69e743d
JH
1222 continue;
1223 if (parse_commit(p))
1224 continue;
0d981c67 1225 porigin = find(sb, p, origin);
f69e743d
JH
1226 if (!porigin)
1227 continue;
1228 if (!hashcmp(porigin->blob_sha1, origin->blob_sha1)) {
c2e525d9 1229 pass_whole_blame(sb, origin, porigin);
f69e743d
JH
1230 origin_decref(porigin);
1231 goto finish;
1232 }
0421d9f8 1233 for (j = same = 0; j < i; j++)
69264f46
JH
1234 if (sg_origin[j] &&
1235 !hashcmp(sg_origin[j]->blob_sha1,
0421d9f8
JH
1236 porigin->blob_sha1)) {
1237 same = 1;
1238 break;
1239 }
1240 if (!same)
69264f46 1241 sg_origin[i] = porigin;
0421d9f8
JH
1242 else
1243 origin_decref(porigin);
cee7f245 1244 }
cee7f245
JH
1245 }
1246
c2e525d9 1247 num_commits++;
85af7929 1248 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1249 i < num_sg && sg;
1250 sg = sg->next, i++) {
1251 struct origin *porigin = sg_origin[i];
cee7f245
JH
1252 if (!porigin)
1253 continue;
96e11709
JH
1254 if (!origin->previous) {
1255 origin_incref(porigin);
1256 origin->previous = porigin;
1257 }
cee7f245 1258 if (pass_blame_to_parent(sb, origin, porigin))
54a4c617 1259 goto finish;
cee7f245 1260 }
d24bba80
JH
1261
1262 /*
1732a1fd 1263 * Optionally find moves in parents' files.
d24bba80
JH
1264 */
1265 if (opt & PICKAXE_BLAME_MOVE)
85af7929 1266 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1267 i < num_sg && sg;
1268 sg = sg->next, i++) {
1269 struct origin *porigin = sg_origin[i];
d24bba80
JH
1270 if (!porigin)
1271 continue;
1272 if (find_move_in_parent(sb, origin, porigin))
54a4c617 1273 goto finish;
d24bba80
JH
1274 }
1275
18abd745 1276 /*
1732a1fd 1277 * Optionally find copies from parents' files.
18abd745
JH
1278 */
1279 if (opt & PICKAXE_BLAME_COPY)
85af7929 1280 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1281 i < num_sg && sg;
1282 sg = sg->next, i++) {
1283 struct origin *porigin = sg_origin[i];
1284 if (find_copy_in_parent(sb, origin, sg->item,
18abd745 1285 porigin, opt))
54a4c617 1286 goto finish;
18abd745 1287 }
54a4c617
JH
1288
1289 finish:
69264f46
JH
1290 for (i = 0; i < num_sg; i++) {
1291 if (sg_origin[i]) {
1292 drop_origin_blob(sg_origin[i]);
1293 origin_decref(sg_origin[i]);
7c3c7962
JH
1294 }
1295 }
1296 drop_origin_blob(origin);
69264f46
JH
1297 if (sg_buf != sg_origin)
1298 free(sg_origin);
cee7f245
JH
1299}
1300
1732a1fd
JH
1301/*
1302 * Information on commits, used for output.
1303 */
9cba13ca 1304struct commit_info {
ea02ffa3
AP
1305 struct strbuf author;
1306 struct strbuf author_mail;
cee7f245 1307 unsigned long author_time;
ea02ffa3 1308 struct strbuf author_tz;
cee7f245
JH
1309
1310 /* filled only when asked for details */
ea02ffa3
AP
1311 struct strbuf committer;
1312 struct strbuf committer_mail;
cee7f245 1313 unsigned long committer_time;
ea02ffa3 1314 struct strbuf committer_tz;
cee7f245 1315
ea02ffa3 1316 struct strbuf summary;
cee7f245
JH
1317};
1318
1732a1fd
JH
1319/*
1320 * Parse author/committer line in the commit object buffer
1321 */
cee7f245 1322static void get_ac_line(const char *inbuf, const char *what,
ea02ffa3
AP
1323 struct strbuf *name, struct strbuf *mail,
1324 unsigned long *time, struct strbuf *tz)
cee7f245 1325{
3c020bd5 1326 struct ident_split ident;
ea02ffa3
AP
1327 size_t len, maillen, namelen;
1328 char *tmp, *endp;
1329 const char *namebuf, *mailbuf;
cee7f245
JH
1330
1331 tmp = strstr(inbuf, what);
1332 if (!tmp)
1333 goto error_out;
1334 tmp += strlen(what);
1335 endp = strchr(tmp, '\n');
1336 if (!endp)
1337 len = strlen(tmp);
1338 else
1339 len = endp - tmp;
3c020bd5
AP
1340
1341 if (split_ident_line(&ident, tmp, len)) {
cee7f245
JH
1342 error_out:
1343 /* Ugh */
ea02ffa3
AP
1344 tmp = "(unknown)";
1345 strbuf_addstr(name, tmp);
1346 strbuf_addstr(mail, tmp);
1347 strbuf_addstr(tz, tmp);
cee7f245
JH
1348 *time = 0;
1349 return;
1350 }
cee7f245 1351
3c020bd5 1352 namelen = ident.name_end - ident.name_begin;
ea02ffa3 1353 namebuf = ident.name_begin;
cee7f245 1354
ea02ffa3
AP
1355 maillen = ident.mail_end - ident.mail_begin;
1356 mailbuf = ident.mail_begin;
f9567384 1357
de5abe9f
RS
1358 if (ident.date_begin && ident.date_end)
1359 *time = strtoul(ident.date_begin, NULL, 10);
1360 else
1361 *time = 0;
f9567384 1362
de5abe9f
RS
1363 if (ident.tz_begin && ident.tz_end)
1364 strbuf_add(tz, ident.tz_begin, ident.tz_end - ident.tz_begin);
1365 else
1366 strbuf_addstr(tz, "(unknown)");
f9567384 1367
f9567384 1368 /*
d20d654f 1369 * Now, convert both name and e-mail using mailmap
f9567384 1370 */
ea02ffa3
AP
1371 map_user(&mailmap, &mailbuf, &maillen,
1372 &namebuf, &namelen);
1373
1374 strbuf_addf(mail, "<%.*s>", (int)maillen, mailbuf);
1375 strbuf_add(name, namebuf, namelen);
1376}
1377
1378static void commit_info_init(struct commit_info *ci)
1379{
1380
1381 strbuf_init(&ci->author, 0);
1382 strbuf_init(&ci->author_mail, 0);
1383 strbuf_init(&ci->author_tz, 0);
1384 strbuf_init(&ci->committer, 0);
1385 strbuf_init(&ci->committer_mail, 0);
1386 strbuf_init(&ci->committer_tz, 0);
1387 strbuf_init(&ci->summary, 0);
1388}
1389
1390static void commit_info_destroy(struct commit_info *ci)
1391{
1392
1393 strbuf_release(&ci->author);
1394 strbuf_release(&ci->author_mail);
1395 strbuf_release(&ci->author_tz);
1396 strbuf_release(&ci->committer);
1397 strbuf_release(&ci->committer_mail);
1398 strbuf_release(&ci->committer_tz);
1399 strbuf_release(&ci->summary);
cee7f245
JH
1400}
1401
1402static void get_commit_info(struct commit *commit,
1403 struct commit_info *ret,
1404 int detailed)
1405{
1406 int len;
e297cf5a 1407 const char *subject, *encoding;
b000c59b 1408 const char *message;
ea02ffa3
AP
1409
1410 commit_info_init(ret);
cee7f245 1411
e297cf5a 1412 encoding = get_log_output_encoding();
5a10d236 1413 message = logmsg_reencode(commit, NULL, encoding);
69cd8f63 1414 get_ac_line(message, "\nauthor ",
ea02ffa3 1415 &ret->author, &ret->author_mail,
cee7f245
JH
1416 &ret->author_time, &ret->author_tz);
1417
69cd8f63 1418 if (!detailed) {
b66103c3 1419 unuse_commit_buffer(commit, message);
cee7f245 1420 return;
69cd8f63 1421 }
cee7f245 1422
69cd8f63 1423 get_ac_line(message, "\ncommitter ",
ea02ffa3 1424 &ret->committer, &ret->committer_mail,
cee7f245
JH
1425 &ret->committer_time, &ret->committer_tz);
1426
ad98a58b 1427 len = find_commit_subject(message, &subject);
ea02ffa3
AP
1428 if (len)
1429 strbuf_add(&ret->summary, subject, len);
1430 else
1431 strbuf_addf(&ret->summary, "(%s)", sha1_to_hex(commit->object.sha1));
1432
b66103c3 1433 unuse_commit_buffer(commit, message);
cee7f245
JH
1434}
1435
1732a1fd
JH
1436/*
1437 * To allow LF and other nonportable characters in pathnames,
1438 * they are c-style quoted as needed.
1439 */
46e5e69d
JH
1440static void write_filename_info(const char *path)
1441{
1442 printf("filename ");
663af342 1443 write_name_quoted(path, stdout, '\n');
46e5e69d
JH
1444}
1445
9991030c
JH
1446/*
1447 * Porcelain/Incremental format wants to show a lot of details per
1448 * commit. Instead of repeating this every line, emit it only once,
e86226e3
JK
1449 * the first time each commit appears in the output (unless the
1450 * user has specifically asked for us to repeat).
9991030c 1451 */
e86226e3 1452static int emit_one_suspect_detail(struct origin *suspect, int repeat)
9991030c
JH
1453{
1454 struct commit_info ci;
1455
e86226e3 1456 if (!repeat && (suspect->commit->object.flags & METAINFO_SHOWN))
9991030c
JH
1457 return 0;
1458
1459 suspect->commit->object.flags |= METAINFO_SHOWN;
1460 get_commit_info(suspect->commit, &ci, 1);
ea02ffa3
AP
1461 printf("author %s\n", ci.author.buf);
1462 printf("author-mail %s\n", ci.author_mail.buf);
9991030c 1463 printf("author-time %lu\n", ci.author_time);
ea02ffa3
AP
1464 printf("author-tz %s\n", ci.author_tz.buf);
1465 printf("committer %s\n", ci.committer.buf);
1466 printf("committer-mail %s\n", ci.committer_mail.buf);
9991030c 1467 printf("committer-time %lu\n", ci.committer_time);
ea02ffa3
AP
1468 printf("committer-tz %s\n", ci.committer_tz.buf);
1469 printf("summary %s\n", ci.summary.buf);
9991030c
JH
1470 if (suspect->commit->object.flags & UNINTERESTING)
1471 printf("boundary\n");
96e11709
JH
1472 if (suspect->previous) {
1473 struct origin *prev = suspect->previous;
1474 printf("previous %s ", sha1_to_hex(prev->commit->object.sha1));
1475 write_name_quoted(prev->path, stdout, '\n');
1476 }
ea02ffa3
AP
1477
1478 commit_info_destroy(&ci);
1479
9991030c
JH
1480 return 1;
1481}
1482
1732a1fd
JH
1483/*
1484 * The blame_entry is found to be guilty for the range. Mark it
1485 * as such, and show it in incremental output.
1486 */
717d1462
LT
1487static void found_guilty_entry(struct blame_entry *ent)
1488{
1489 if (ent->guilty)
1490 return;
1491 ent->guilty = 1;
1492 if (incremental) {
1493 struct origin *suspect = ent->suspect;
1494
1495 printf("%s %d %d %d\n",
1496 sha1_to_hex(suspect->commit->object.sha1),
1497 ent->s_lno + 1, ent->lno + 1, ent->num_lines);
e86226e3 1498 emit_one_suspect_detail(suspect, 0);
46e5e69d 1499 write_filename_info(suspect->path);
06f59e9f 1500 maybe_flush_or_die(stdout, "stdout");
717d1462
LT
1501 }
1502}
1503
1732a1fd
JH
1504/*
1505 * The main loop -- while the scoreboard has lines whose true origin
3dff5379 1506 * is still unknown, pick one blame_entry, and allow its current
1732a1fd
JH
1507 * suspect to pass blames to its parents.
1508 */
85af7929 1509static void assign_blame(struct scoreboard *sb, int opt)
717d1462 1510{
85af7929
JH
1511 struct rev_info *revs = sb->revs;
1512
717d1462
LT
1513 while (1) {
1514 struct blame_entry *ent;
1515 struct commit *commit;
1516 struct origin *suspect = NULL;
1517
1518 /* find one suspect to break down */
1519 for (ent = sb->ent; !suspect && ent; ent = ent->next)
1520 if (!ent->guilty)
1521 suspect = ent->suspect;
1522 if (!suspect)
1523 return; /* all done */
1524
1732a1fd
JH
1525 /*
1526 * We will use this suspect later in the loop,
1527 * so hold onto it in the meantime.
1528 */
717d1462
LT
1529 origin_incref(suspect);
1530 commit = suspect->commit;
0064053b 1531 parse_commit(commit);
85af7929
JH
1532 if (reverse ||
1533 (!(commit->object.flags & UNINTERESTING) &&
1534 !(revs->max_age != -1 && commit->date < revs->max_age)))
717d1462
LT
1535 pass_blame(sb, suspect, opt);
1536 else {
1537 commit->object.flags |= UNINTERESTING;
1538 if (commit->object.parsed)
1539 mark_parents_uninteresting(commit);
1540 }
1541 /* treat root commit as boundary */
1542 if (!commit->parents && !show_root)
1543 commit->object.flags |= UNINTERESTING;
1544
1545 /* Take responsibility for the remaining entries */
1546 for (ent = sb->ent; ent; ent = ent->next)
0a88f08e 1547 if (ent->suspect == suspect)
717d1462
LT
1548 found_guilty_entry(ent);
1549 origin_decref(suspect);
1550
1551 if (DEBUG) /* sanity */
1552 sanity_check_refcnt(sb);
1553 }
1554}
1555
1556static const char *format_time(unsigned long time, const char *tz_str,
1557 int show_raw_time)
1558{
1559 static char time_buf[128];
717d1462
LT
1560
1561 if (show_raw_time) {
c3ea0515 1562 snprintf(time_buf, sizeof(time_buf), "%lu %s", time, tz_str);
717d1462 1563 }
31653c1a 1564 else {
ac39b277
EP
1565 const char *time_str;
1566 int time_len;
1567 int tz;
31653c1a
EL
1568 tz = atoi(tz_str);
1569 time_str = show_date(time, tz, blame_date_mode);
1570 time_len = strlen(time_str);
1571 memcpy(time_buf, time_str, time_len);
1572 memset(time_buf + time_len, ' ', blame_date_width - time_len);
1573 }
717d1462
LT
1574 return time_buf;
1575}
1576
cee7f245
JH
1577#define OUTPUT_ANNOTATE_COMPAT 001
1578#define OUTPUT_LONG_OBJECT_NAME 002
1579#define OUTPUT_RAW_TIMESTAMP 004
1580#define OUTPUT_PORCELAIN 010
1581#define OUTPUT_SHOW_NAME 020
1582#define OUTPUT_SHOW_NUMBER 040
5ff62c30 1583#define OUTPUT_SHOW_SCORE 0100
093dc5be 1584#define OUTPUT_NO_AUTHOR 0200
1b8cdce9 1585#define OUTPUT_SHOW_EMAIL 0400
ed747dd5 1586#define OUTPUT_LINE_PORCELAIN 01000
cee7f245 1587
e86226e3
JK
1588static void emit_porcelain_details(struct origin *suspect, int repeat)
1589{
1590 if (emit_one_suspect_detail(suspect, repeat) ||
1591 (suspect->commit->object.flags & MORE_THAN_ONE_PATH))
1592 write_filename_info(suspect->path);
1593}
1594
1595static void emit_porcelain(struct scoreboard *sb, struct blame_entry *ent,
1596 int opt)
cee7f245 1597{
ed747dd5 1598 int repeat = opt & OUTPUT_LINE_PORCELAIN;
cee7f245
JH
1599 int cnt;
1600 const char *cp;
1601 struct origin *suspect = ent->suspect;
1602 char hex[41];
1603
1604 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1605 printf("%s%c%d %d %d\n",
1606 hex,
2543d9b6 1607 ent->guilty ? ' ' : '*', /* purely for debugging */
cee7f245
JH
1608 ent->s_lno + 1,
1609 ent->lno + 1,
1610 ent->num_lines);
ed747dd5 1611 emit_porcelain_details(suspect, repeat);
cee7f245
JH
1612
1613 cp = nth_line(sb, ent->lno);
1614 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1615 char ch;
ed747dd5 1616 if (cnt) {
cee7f245
JH
1617 printf("%s %d %d\n", hex,
1618 ent->s_lno + 1 + cnt,
1619 ent->lno + 1 + cnt);
ed747dd5
JK
1620 if (repeat)
1621 emit_porcelain_details(suspect, 1);
1622 }
cee7f245
JH
1623 putchar('\t');
1624 do {
1625 ch = *cp++;
1626 putchar(ch);
1627 } while (ch != '\n' &&
1628 cp < sb->final_buf + sb->final_buf_size);
1629 }
a5ca8367
JS
1630
1631 if (sb->final_buf_size && cp[-1] != '\n')
1632 putchar('\n');
cee7f245
JH
1633}
1634
1635static void emit_other(struct scoreboard *sb, struct blame_entry *ent, int opt)
1636{
1637 int cnt;
1638 const char *cp;
1639 struct origin *suspect = ent->suspect;
1640 struct commit_info ci;
1641 char hex[41];
1642 int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP);
1643
1644 get_commit_info(suspect->commit, &ci, 1);
1645 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1646
1647 cp = nth_line(sb, ent->lno);
1648 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1649 char ch;
84393bfd 1650 int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? 40 : abbrev;
b11121d9
JH
1651
1652 if (suspect->commit->object.flags & UNINTERESTING) {
e68989a7
JH
1653 if (blank_boundary)
1654 memset(hex, ' ', length);
7ceacdff 1655 else if (!(opt & OUTPUT_ANNOTATE_COMPAT)) {
4c10a5ca
JH
1656 length--;
1657 putchar('^');
1658 }
b11121d9 1659 }
cee7f245 1660
b11121d9 1661 printf("%.*s", length, hex);
1b8cdce9
KB
1662 if (opt & OUTPUT_ANNOTATE_COMPAT) {
1663 const char *name;
1664 if (opt & OUTPUT_SHOW_EMAIL)
ea02ffa3 1665 name = ci.author_mail.buf;
1b8cdce9 1666 else
ea02ffa3 1667 name = ci.author.buf;
1b8cdce9 1668 printf("\t(%10s\t%10s\t%d)", name,
ea02ffa3 1669 format_time(ci.author_time, ci.author_tz.buf,
cee7f245
JH
1670 show_raw_time),
1671 ent->lno + 1 + cnt);
1b8cdce9 1672 } else {
5ff62c30 1673 if (opt & OUTPUT_SHOW_SCORE)
54a4c617
JH
1674 printf(" %*d %02d",
1675 max_score_digits, ent->score,
1676 ent->suspect->refcnt);
cee7f245
JH
1677 if (opt & OUTPUT_SHOW_NAME)
1678 printf(" %-*.*s", longest_file, longest_file,
1679 suspect->path);
1680 if (opt & OUTPUT_SHOW_NUMBER)
1681 printf(" %*d", max_orig_digits,
1682 ent->s_lno + 1 + cnt);
093dc5be 1683
ffaf9cc0 1684 if (!(opt & OUTPUT_NO_AUTHOR)) {
1b8cdce9
KB
1685 const char *name;
1686 int pad;
1687 if (opt & OUTPUT_SHOW_EMAIL)
ea02ffa3 1688 name = ci.author_mail.buf;
1b8cdce9 1689 else
ea02ffa3 1690 name = ci.author.buf;
1b8cdce9 1691 pad = longest_author - utf8_strwidth(name);
ffaf9cc0 1692 printf(" (%s%*s %10s",
1b8cdce9 1693 name, pad, "",
093dc5be 1694 format_time(ci.author_time,
ea02ffa3 1695 ci.author_tz.buf,
093dc5be 1696 show_raw_time));
ffaf9cc0 1697 }
093dc5be 1698 printf(" %*d) ",
cee7f245
JH
1699 max_digits, ent->lno + 1 + cnt);
1700 }
1701 do {
1702 ch = *cp++;
1703 putchar(ch);
1704 } while (ch != '\n' &&
1705 cp < sb->final_buf + sb->final_buf_size);
1706 }
a5ca8367
JS
1707
1708 if (sb->final_buf_size && cp[-1] != '\n')
1709 putchar('\n');
ea02ffa3
AP
1710
1711 commit_info_destroy(&ci);
cee7f245
JH
1712}
1713
1714static void output(struct scoreboard *sb, int option)
1715{
1716 struct blame_entry *ent;
1717
1718 if (option & OUTPUT_PORCELAIN) {
1719 for (ent = sb->ent; ent; ent = ent->next) {
1720 struct blame_entry *oth;
1721 struct origin *suspect = ent->suspect;
1722 struct commit *commit = suspect->commit;
1723 if (commit->object.flags & MORE_THAN_ONE_PATH)
1724 continue;
1725 for (oth = ent->next; oth; oth = oth->next) {
1726 if ((oth->suspect->commit != commit) ||
1727 !strcmp(oth->suspect->path, suspect->path))
1728 continue;
1729 commit->object.flags |= MORE_THAN_ONE_PATH;
1730 break;
1731 }
1732 }
1733 }
1734
1735 for (ent = sb->ent; ent; ent = ent->next) {
1736 if (option & OUTPUT_PORCELAIN)
e86226e3 1737 emit_porcelain(sb, ent, option);
5ff62c30 1738 else {
cee7f245 1739 emit_other(sb, ent, option);
5ff62c30 1740 }
cee7f245
JH
1741 }
1742}
1743
1732a1fd
JH
1744/*
1745 * To allow quick access to the contents of nth line in the
1746 * final image, prepare an index in the scoreboard.
1747 */
cee7f245
JH
1748static int prepare_lines(struct scoreboard *sb)
1749{
1750 const char *buf = sb->final_buf;
1751 unsigned long len = sb->final_buf_size;
352bbbd9
DK
1752 const char *end = buf + len;
1753 const char *p;
1754 int *lineno;
1755 int num = 0, incomplete = 0;
cee7f245 1756
352bbbd9
DK
1757 for (p = buf;;) {
1758 p = memchr(p, '\n', end - p);
1759 if (p) {
1760 p++;
cee7f245 1761 num++;
352bbbd9 1762 continue;
cee7f245 1763 }
352bbbd9 1764 break;
cee7f245 1765 }
352bbbd9
DK
1766
1767 if (len && end[-1] != '\n')
1768 incomplete++; /* incomplete line at the end */
1769
1770 sb->lineno = xmalloc(sizeof(*sb->lineno) * (num + incomplete + 1));
1771 lineno = sb->lineno;
1772
1773 *lineno++ = 0;
1774 for (p = buf;;) {
1775 p = memchr(p, '\n', end - p);
1776 if (p) {
1777 p++;
1778 *lineno++ = p - buf;
1779 continue;
1780 }
1781 break;
1782 }
1783
1784 if (incomplete)
1785 *lineno++ = len;
1786
cee7f245
JH
1787 sb->num_lines = num + incomplete;
1788 return sb->num_lines;
1789}
1790
1732a1fd
JH
1791/*
1792 * Add phony grafts for use with -S; this is primarily to
34baebce 1793 * support git's cvsserver that wants to give a linear history
1732a1fd
JH
1794 * to its clients.
1795 */
cee7f245
JH
1796static int read_ancestry(const char *graft_file)
1797{
1798 FILE *fp = fopen(graft_file, "r");
e228c173 1799 struct strbuf buf = STRBUF_INIT;
cee7f245
JH
1800 if (!fp)
1801 return -1;
e228c173 1802 while (!strbuf_getwholeline(&buf, fp, '\n')) {
cee7f245 1803 /* The format is just "Commit Parent1 Parent2 ...\n" */
e228c173 1804 struct commit_graft *graft = read_graft_line(buf.buf, buf.len);
8eaf7986
JH
1805 if (graft)
1806 register_commit_graft(graft, 0);
cee7f245
JH
1807 }
1808 fclose(fp);
e228c173 1809 strbuf_release(&buf);
cee7f245
JH
1810 return 0;
1811}
1812
b31272f7
JH
1813static int update_auto_abbrev(int auto_abbrev, struct origin *suspect)
1814{
1815 const char *uniq = find_unique_abbrev(suspect->commit->object.sha1,
1816 auto_abbrev);
1817 int len = strlen(uniq);
1818 if (auto_abbrev < len)
1819 return len;
1820 return auto_abbrev;
1821}
1822
1732a1fd
JH
1823/*
1824 * How many columns do we need to show line numbers, authors,
1825 * and filenames?
1826 */
cee7f245
JH
1827static void find_alignment(struct scoreboard *sb, int *option)
1828{
1829 int longest_src_lines = 0;
1830 int longest_dst_lines = 0;
5ff62c30 1831 unsigned largest_score = 0;
cee7f245 1832 struct blame_entry *e;
b31272f7
JH
1833 int compute_auto_abbrev = (abbrev < 0);
1834 int auto_abbrev = default_abbrev;
cee7f245
JH
1835
1836 for (e = sb->ent; e; e = e->next) {
1837 struct origin *suspect = e->suspect;
1838 struct commit_info ci;
1839 int num;
1840
b31272f7
JH
1841 if (compute_auto_abbrev)
1842 auto_abbrev = update_auto_abbrev(auto_abbrev, suspect);
ab3bb800
JH
1843 if (strcmp(suspect->path, sb->path))
1844 *option |= OUTPUT_SHOW_NAME;
1845 num = strlen(suspect->path);
1846 if (longest_file < num)
1847 longest_file = num;
cee7f245
JH
1848 if (!(suspect->commit->object.flags & METAINFO_SHOWN)) {
1849 suspect->commit->object.flags |= METAINFO_SHOWN;
1850 get_commit_info(suspect->commit, &ci, 1);
1b8cdce9 1851 if (*option & OUTPUT_SHOW_EMAIL)
ea02ffa3 1852 num = utf8_strwidth(ci.author_mail.buf);
1b8cdce9 1853 else
ea02ffa3 1854 num = utf8_strwidth(ci.author.buf);
cee7f245
JH
1855 if (longest_author < num)
1856 longest_author = num;
1857 }
1858 num = e->s_lno + e->num_lines;
1859 if (longest_src_lines < num)
1860 longest_src_lines = num;
1861 num = e->lno + e->num_lines;
1862 if (longest_dst_lines < num)
1863 longest_dst_lines = num;
5ff62c30
JH
1864 if (largest_score < ent_score(sb, e))
1865 largest_score = ent_score(sb, e);
ea02ffa3
AP
1866
1867 commit_info_destroy(&ci);
cee7f245 1868 }
ec7ff5ba
ZJS
1869 max_orig_digits = decimal_width(longest_src_lines);
1870 max_digits = decimal_width(longest_dst_lines);
1871 max_score_digits = decimal_width(largest_score);
b31272f7
JH
1872
1873 if (compute_auto_abbrev)
1874 /* one more abbrev length is needed for the boundary commit */
1875 abbrev = auto_abbrev + 1;
cee7f245
JH
1876}
1877
1732a1fd
JH
1878/*
1879 * For debugging -- origin is refcounted, and this asserts that
1880 * we do not underflow.
1881 */
54a4c617
JH
1882static void sanity_check_refcnt(struct scoreboard *sb)
1883{
1884 int baa = 0;
1885 struct blame_entry *ent;
1886
1887 for (ent = sb->ent; ent; ent = ent->next) {
ae86ad65 1888 /* Nobody should have zero or negative refcnt */
854b97f6
JH
1889 if (ent->suspect->refcnt <= 0) {
1890 fprintf(stderr, "%s in %s has negative refcnt %d\n",
1891 ent->suspect->path,
1892 sha1_to_hex(ent->suspect->commit->object.sha1),
1893 ent->suspect->refcnt);
ae86ad65 1894 baa = 1;
854b97f6 1895 }
ae86ad65 1896 }
54a4c617
JH
1897 if (baa) {
1898 int opt = 0160;
1899 find_alignment(sb, &opt);
1900 output(sb, opt);
854b97f6 1901 die("Baa %d!", baa);
54a4c617
JH
1902 }
1903}
1904
1732a1fd
JH
1905/*
1906 * Used for the command line parsing; check if the path exists
1907 * in the working tree.
1908 */
c455c87c 1909static int has_string_in_work_tree(const char *path)
cee7f245
JH
1910{
1911 struct stat st;
1912 return !lstat(path, &st);
1913}
1914
4a0fc95f
JH
1915static unsigned parse_score(const char *arg)
1916{
1917 char *end;
1918 unsigned long score = strtoul(arg, &end, 10);
1919 if (*end)
1920 return 0;
1921 return score;
1922}
1923
20239bae
JK
1924static const char *add_prefix(const char *prefix, const char *path)
1925{
097971f5 1926 return prefix_path(prefix, prefix ? strlen(prefix) : 0, path);
20239bae
JK
1927}
1928
ef90d6d4 1929static int git_blame_config(const char *var, const char *value, void *cb)
4c10a5ca
JH
1930{
1931 if (!strcmp(var, "blame.showroot")) {
1932 show_root = git_config_bool(var, value);
1933 return 0;
1934 }
1935 if (!strcmp(var, "blame.blankboundary")) {
1936 blank_boundary = git_config_bool(var, value);
1937 return 0;
1938 }
31653c1a
EL
1939 if (!strcmp(var, "blame.date")) {
1940 if (!value)
1941 return config_error_nonbool(var);
1942 blame_date_mode = parse_date_format(value);
1943 return 0;
1944 }
3b8a12e8 1945
6680a087 1946 if (userdiff_config(var, value) < 0)
3b8a12e8 1947 return -1;
3b8a12e8 1948
ef90d6d4 1949 return git_default_config(var, value, cb);
4c10a5ca
JH
1950}
1951
9aeaab68 1952static void verify_working_tree_path(struct commit *work_tree, const char *path)
ffcabccf 1953{
9aeaab68 1954 struct commit_list *parents;
ffcabccf 1955
9aeaab68
JH
1956 for (parents = work_tree->parents; parents; parents = parents->next) {
1957 const unsigned char *commit_sha1 = parents->item->object.sha1;
1958 unsigned char blob_sha1[20];
1959 unsigned mode;
1960
1961 if (!get_tree_entry(commit_sha1, path, blob_sha1, &mode) &&
1962 sha1_object_info(blob_sha1, NULL) == OBJ_BLOB)
1963 return;
1964 }
1965 die("no such path '%s' in HEAD", path);
1966}
1967
1968static struct commit_list **append_parent(struct commit_list **tail, const unsigned char *sha1)
1969{
1970 struct commit *parent;
1971
1972 parent = lookup_commit_reference(sha1);
1973 if (!parent)
1974 die("no such commit %s", sha1_to_hex(sha1));
1975 return &commit_list_insert(parent, tail)->next;
1976}
1977
1978static void append_merge_parents(struct commit_list **tail)
1979{
1980 int merge_head;
1981 const char *merge_head_file = git_path("MERGE_HEAD");
1982 struct strbuf line = STRBUF_INIT;
1983
1984 merge_head = open(merge_head_file, O_RDONLY);
1985 if (merge_head < 0) {
1986 if (errno == ENOENT)
1987 return;
1988 die("cannot open '%s' for reading", merge_head_file);
1989 }
1990
1991 while (!strbuf_getwholeline_fd(&line, merge_head, '\n')) {
1992 unsigned char sha1[20];
1993 if (line.len < 40 || get_sha1_hex(line.buf, sha1))
1994 die("unknown line in '%s': %s", merge_head_file, line.buf);
1995 tail = append_parent(tail, sha1);
1996 }
1997 close(merge_head);
1998 strbuf_release(&line);
ffcabccf
JH
1999}
2000
f6c07d7d
JH
2001/*
2002 * Prepare a dummy commit that represents the work tree (or staged) item.
2003 * Note that annotating work tree item never works in the reverse.
2004 */
3b8a12e8
AB
2005static struct commit *fake_working_tree_commit(struct diff_options *opt,
2006 const char *path,
2007 const char *contents_from)
1cfe7733
JH
2008{
2009 struct commit *commit;
2010 struct origin *origin;
9aeaab68 2011 struct commit_list **parent_tail, *parent;
1cfe7733 2012 unsigned char head_sha1[20];
f285a2d7 2013 struct strbuf buf = STRBUF_INIT;
1cfe7733 2014 const char *ident;
1cfe7733 2015 time_t now;
1cfe7733
JH
2016 int size, len;
2017 struct cache_entry *ce;
2018 unsigned mode;
9aeaab68 2019 struct strbuf msg = STRBUF_INIT;
1cfe7733
JH
2020
2021 time(&now);
10322a0a 2022 commit = alloc_commit_node();
1cfe7733
JH
2023 commit->object.parsed = 1;
2024 commit->date = now;
2025 commit->object.type = OBJ_COMMIT;
9aeaab68
JH
2026 parent_tail = &commit->parents;
2027
2028 if (!resolve_ref_unsafe("HEAD", head_sha1, 1, NULL))
2029 die("no such ref: HEAD");
2030
2031 parent_tail = append_parent(parent_tail, head_sha1);
2032 append_merge_parents(parent_tail);
2033 verify_working_tree_path(commit, path);
1cfe7733
JH
2034
2035 origin = make_origin(commit, path);
2036
9aeaab68
JH
2037 ident = fmt_ident("Not Committed Yet", "not.committed.yet", NULL, 0);
2038 strbuf_addstr(&msg, "tree 0000000000000000000000000000000000000000\n");
2039 for (parent = commit->parents; parent; parent = parent->next)
2040 strbuf_addf(&msg, "parent %s\n",
2041 sha1_to_hex(parent->item->object.sha1));
2042 strbuf_addf(&msg,
2043 "author %s\n"
2044 "committer %s\n\n"
2045 "Version of %s from %s\n",
2046 ident, ident, path,
2047 (!contents_from ? path :
2048 (!strcmp(contents_from, "-") ? "standard input" : contents_from)));
66c2827e 2049 set_commit_buffer(commit, strbuf_detach(&msg, NULL));
9aeaab68 2050
1cfe7733
JH
2051 if (!contents_from || strcmp("-", contents_from)) {
2052 struct stat st;
2053 const char *read_from;
8518088f 2054 char *buf_ptr;
3b8a12e8 2055 unsigned long buf_len;
1cfe7733
JH
2056
2057 if (contents_from) {
2058 if (stat(contents_from, &st) < 0)
0721c314 2059 die_errno("Cannot stat '%s'", contents_from);
1cfe7733
JH
2060 read_from = contents_from;
2061 }
2062 else {
2063 if (lstat(path, &st) < 0)
0721c314 2064 die_errno("Cannot lstat '%s'", path);
1cfe7733
JH
2065 read_from = path;
2066 }
1cfe7733 2067 mode = canon_mode(st.st_mode);
3b8a12e8 2068
1cfe7733
JH
2069 switch (st.st_mode & S_IFMT) {
2070 case S_IFREG:
3b8a12e8 2071 if (DIFF_OPT_TST(opt, ALLOW_TEXTCONV) &&
e5450100 2072 textconv_object(read_from, mode, null_sha1, 0, &buf_ptr, &buf_len))
8518088f 2073 strbuf_attach(&buf, buf_ptr, buf_len, buf_len + 1);
3b8a12e8 2074 else if (strbuf_read_file(&buf, read_from, st.st_size) != st.st_size)
0721c314 2075 die_errno("cannot open or read '%s'", read_from);
1cfe7733
JH
2076 break;
2077 case S_IFLNK:
edfd45d2 2078 if (strbuf_readlink(&buf, read_from, st.st_size) < 0)
0721c314 2079 die_errno("cannot readlink '%s'", read_from);
1cfe7733
JH
2080 break;
2081 default:
2082 die("unsupported file type %s", read_from);
2083 }
2084 }
2085 else {
2086 /* Reading from stdin */
1cfe7733 2087 mode = 0;
f1696ee3 2088 if (strbuf_read(&buf, 0, 0) < 0)
d824cbba 2089 die_errno("failed to read from stdin");
1cfe7733 2090 }
21e5ad50 2091 convert_to_git(path, buf.buf, buf.len, &buf, 0);
af6eb822
PH
2092 origin->file.ptr = buf.buf;
2093 origin->file.size = buf.len;
2094 pretend_sha1_file(buf.buf, buf.len, OBJ_BLOB, origin->blob_sha1);
1cfe7733
JH
2095 commit->util = origin;
2096
2097 /*
2098 * Read the current index, replace the path entry with
2099 * origin->blob_sha1 without mucking with its mode or type
2100 * bits; we are not going to write this index out -- we just
2101 * want to run "diff-index --cached".
2102 */
2103 discard_cache();
2104 read_cache();
2105
2106 len = strlen(path);
2107 if (!mode) {
2108 int pos = cache_name_pos(path, len);
2109 if (0 <= pos)
7a51ed66 2110 mode = active_cache[pos]->ce_mode;
1cfe7733
JH
2111 else
2112 /* Let's not bother reading from HEAD tree */
2113 mode = S_IFREG | 0644;
2114 }
2115 size = cache_entry_size(len);
2116 ce = xcalloc(1, size);
2117 hashcpy(ce->sha1, origin->blob_sha1);
2118 memcpy(ce->name, path, len);
b60e188c
TG
2119 ce->ce_flags = create_ce_flags(0);
2120 ce->ce_namelen = len;
1cfe7733
JH
2121 ce->ce_mode = create_ce_mode(mode);
2122 add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
2123
2124 /*
2125 * We are not going to write this out, so this does not matter
2126 * right now, but someday we might optimize diff-index --cached
2127 * with cache-tree information.
2128 */
2129 cache_tree_invalidate_path(active_cache_tree, path);
2130
1cfe7733
JH
2131 return commit;
2132}
2133
85af7929 2134static const char *prepare_final(struct scoreboard *sb)
f6c07d7d
JH
2135{
2136 int i;
2137 const char *final_commit_name = NULL;
85af7929 2138 struct rev_info *revs = sb->revs;
f6c07d7d
JH
2139
2140 /*
2141 * There must be one and only one positive commit in the
2142 * revs->pending array.
2143 */
2144 for (i = 0; i < revs->pending.nr; i++) {
2145 struct object *obj = revs->pending.objects[i].item;
2146 if (obj->flags & UNINTERESTING)
2147 continue;
2148 while (obj->type == OBJ_TAG)
2149 obj = deref_tag(obj, NULL, 0);
2150 if (obj->type != OBJ_COMMIT)
2151 die("Non commit %s?", revs->pending.objects[i].name);
2152 if (sb->final)
2153 die("More than one commit to dig from %s and %s?",
2154 revs->pending.objects[i].name,
2155 final_commit_name);
2156 sb->final = (struct commit *) obj;
2157 final_commit_name = revs->pending.objects[i].name;
2158 }
2159 return final_commit_name;
2160}
2161
85af7929
JH
2162static const char *prepare_initial(struct scoreboard *sb)
2163{
2164 int i;
2165 const char *final_commit_name = NULL;
2166 struct rev_info *revs = sb->revs;
2167
2168 /*
2169 * There must be one and only one negative commit, and it must be
2170 * the boundary.
2171 */
2172 for (i = 0; i < revs->pending.nr; i++) {
2173 struct object *obj = revs->pending.objects[i].item;
2174 if (!(obj->flags & UNINTERESTING))
2175 continue;
2176 while (obj->type == OBJ_TAG)
2177 obj = deref_tag(obj, NULL, 0);
2178 if (obj->type != OBJ_COMMIT)
2179 die("Non commit %s?", revs->pending.objects[i].name);
2180 if (sb->final)
2181 die("More than one commit to dig down to %s and %s?",
2182 revs->pending.objects[i].name,
2183 final_commit_name);
2184 sb->final = (struct commit *) obj;
2185 final_commit_name = revs->pending.objects[i].name;
2186 }
2187 if (!final_commit_name)
2188 die("No commit to dig down to?");
2189 return final_commit_name;
2190}
2191
5817da01
PH
2192static int blame_copy_callback(const struct option *option, const char *arg, int unset)
2193{
2194 int *opt = option->value;
2195
2196 /*
2197 * -C enables copy from removed files;
2198 * -C -C enables copy from existing files, but only
2199 * when blaming a new file;
2200 * -C -C -C enables copy from existing files for
2201 * everybody
2202 */
2203 if (*opt & PICKAXE_BLAME_COPY_HARDER)
2204 *opt |= PICKAXE_BLAME_COPY_HARDEST;
2205 if (*opt & PICKAXE_BLAME_COPY)
2206 *opt |= PICKAXE_BLAME_COPY_HARDER;
2207 *opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE;
2208
2209 if (arg)
2210 blame_copy_score = parse_score(arg);
2211 return 0;
2212}
2213
2214static int blame_move_callback(const struct option *option, const char *arg, int unset)
2215{
2216 int *opt = option->value;
2217
2218 *opt |= PICKAXE_BLAME_MOVE;
2219
2220 if (arg)
2221 blame_move_score = parse_score(arg);
2222 return 0;
2223}
2224
acca687f 2225int cmd_blame(int argc, const char **argv, const char *prefix)
cee7f245
JH
2226{
2227 struct rev_info revs;
2228 const char *path;
2229 struct scoreboard sb;
2230 struct origin *o;
58dbfa2e
ES
2231 struct blame_entry *ent = NULL;
2232 long dashdash_pos, lno;
cee7f245 2233 const char *final_commit_name = NULL;
21666f1a 2234 enum object_type type;
5817da01 2235
58dbfa2e 2236 static struct string_list range_list;
5817da01
PH
2237 static int output_option = 0, opt = 0;
2238 static int show_stats = 0;
2239 static const char *revs_file = NULL;
2240 static const char *contents_from = NULL;
2241 static const struct option options[] = {
d5d09d47
SB
2242 OPT_BOOL(0, "incremental", &incremental, N_("Show blame entries as we find them, incrementally")),
2243 OPT_BOOL('b', NULL, &blank_boundary, N_("Show blank SHA-1 for boundary commits (Default: off)")),
2244 OPT_BOOL(0, "root", &show_root, N_("Do not treat root commits as boundaries (Default: off)")),
2245 OPT_BOOL(0, "show-stats", &show_stats, N_("Show work cost statistics")),
efd2a8bd
NTND
2246 OPT_BIT(0, "score-debug", &output_option, N_("Show output score for blame entries"), OUTPUT_SHOW_SCORE),
2247 OPT_BIT('f', "show-name", &output_option, N_("Show original filename (Default: auto)"), OUTPUT_SHOW_NAME),
2248 OPT_BIT('n', "show-number", &output_option, N_("Show original linenumber (Default: off)"), OUTPUT_SHOW_NUMBER),
2249 OPT_BIT('p', "porcelain", &output_option, N_("Show in a format designed for machine consumption"), OUTPUT_PORCELAIN),
2250 OPT_BIT(0, "line-porcelain", &output_option, N_("Show porcelain format with per-line commit information"), OUTPUT_PORCELAIN|OUTPUT_LINE_PORCELAIN),
2251 OPT_BIT('c', NULL, &output_option, N_("Use the same output mode as git-annotate (Default: off)"), OUTPUT_ANNOTATE_COMPAT),
2252 OPT_BIT('t', NULL, &output_option, N_("Show raw timestamp (Default: off)"), OUTPUT_RAW_TIMESTAMP),
2253 OPT_BIT('l', NULL, &output_option, N_("Show long commit SHA1 (Default: off)"), OUTPUT_LONG_OBJECT_NAME),
2254 OPT_BIT('s', NULL, &output_option, N_("Suppress author name and timestamp (Default: off)"), OUTPUT_NO_AUTHOR),
2255 OPT_BIT('e', "show-email", &output_option, N_("Show author email instead of name (Default: off)"), OUTPUT_SHOW_EMAIL),
2256 OPT_BIT('w', NULL, &xdl_opts, N_("Ignore whitespace differences"), XDF_IGNORE_WHITESPACE),
2257 OPT_BIT(0, "minimal", &xdl_opts, N_("Spend extra cycles to find better match"), XDF_NEED_MINIMAL),
2258 OPT_STRING('S', NULL, &revs_file, N_("file"), N_("Use revisions from <file> instead of calling git-rev-list")),
2259 OPT_STRING(0, "contents", &contents_from, N_("file"), N_("Use <file>'s contents as the final image")),
2260 { OPTION_CALLBACK, 'C', NULL, &opt, N_("score"), N_("Find line copies within and across files"), PARSE_OPT_OPTARG, blame_copy_callback },
2261 { OPTION_CALLBACK, 'M', NULL, &opt, N_("score"), N_("Find line movements within and across files"), PARSE_OPT_OPTARG, blame_move_callback },
58dbfa2e 2262 OPT_STRING_LIST('L', NULL, &range_list, N_("n,m"), N_("Process only line range n,m, counting from 1")),
84393bfd 2263 OPT__ABBREV(&abbrev),
5817da01
PH
2264 OPT_END()
2265 };
2266
2267 struct parse_opt_ctx_t ctx;
7ceacdff 2268 int cmd_is_annotate = !strcmp(argv[0], "annotate");
58dbfa2e
ES
2269 struct range_set ranges;
2270 unsigned int range_i;
52f4d126 2271 long anchor;
e68989a7 2272
ef90d6d4 2273 git_config(git_blame_config, NULL);
5817da01 2274 init_revisions(&revs, NULL);
31653c1a 2275 revs.date_mode = blame_date_mode;
3b8a12e8 2276 DIFF_OPT_SET(&revs.diffopt, ALLOW_TEXTCONV);
3d1aa566 2277 DIFF_OPT_SET(&revs.diffopt, FOLLOW_RENAMES);
31653c1a 2278
612702e8 2279 save_commit_buffer = 0;
3f8d5204 2280 dashdash_pos = 0;
612702e8 2281
9ca1169f
SB
2282 parse_options_start(&ctx, argc, argv, prefix, options,
2283 PARSE_OPT_KEEP_DASHDASH | PARSE_OPT_KEEP_ARGV0);
5817da01 2284 for (;;) {
5817da01
PH
2285 switch (parse_options_step(&ctx, options, blame_opt_usage)) {
2286 case PARSE_OPT_HELP:
2287 exit(129);
2288 case PARSE_OPT_DONE:
3f8d5204
PH
2289 if (ctx.argv[0])
2290 dashdash_pos = ctx.cpidx;
5817da01
PH
2291 goto parse_done;
2292 }
2293
2294 if (!strcmp(ctx.argv[0], "--reverse")) {
2295 ctx.argv[0] = "--children";
2296 reverse = 1;
2297 }
6b61ec05 2298 parse_revision_opt(&revs, &ctx, options, blame_opt_usage);
5817da01
PH
2299 }
2300parse_done:
3d1aa566
JH
2301 no_whole_file_rename = !DIFF_OPT_TST(&revs.diffopt, FOLLOW_RENAMES);
2302 DIFF_OPT_CLR(&revs.diffopt, FOLLOW_RENAMES);
5817da01
PH
2303 argc = parse_options_end(&ctx);
2304
b31272f7
JH
2305 if (0 < abbrev)
2306 /* one more abbrev length is needed for the boundary commit */
2307 abbrev++;
84393bfd 2308
aa9ea77d 2309 if (revs_file && read_ancestry(revs_file))
d824cbba 2310 die_errno("reading graft file '%s' failed", revs_file);
aa9ea77d 2311
31653c1a 2312 if (cmd_is_annotate) {
7ceacdff 2313 output_option |= OUTPUT_ANNOTATE_COMPAT;
31653c1a
EL
2314 blame_date_mode = DATE_ISO8601;
2315 } else {
2316 blame_date_mode = revs.date_mode;
2317 }
2318
2319 /* The maximum width used to show the dates */
2320 switch (blame_date_mode) {
2321 case DATE_RFC2822:
2322 blame_date_width = sizeof("Thu, 19 Oct 2006 16:00:04 -0700");
2323 break;
2324 case DATE_ISO8601:
2325 blame_date_width = sizeof("2006-10-19 16:00:04 -0700");
2326 break;
2327 case DATE_RAW:
2328 blame_date_width = sizeof("1161298804 -0700");
2329 break;
2330 case DATE_SHORT:
2331 blame_date_width = sizeof("2006-10-19");
2332 break;
2333 case DATE_RELATIVE:
2334 /* "normal" is used as the fallback for "relative" */
2335 case DATE_LOCAL:
2336 case DATE_NORMAL:
2337 blame_date_width = sizeof("Thu Oct 19 16:00:04 2006 -0700");
2338 break;
2339 }
2340 blame_date_width -= 1; /* strip the null */
7ceacdff 2341
b3123f98
JH
2342 if (DIFF_OPT_TST(&revs.diffopt, FIND_COPIES_HARDER))
2343 opt |= (PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE |
2344 PICKAXE_BLAME_COPY_HARDER);
2345
4a0fc95f
JH
2346 if (!blame_move_score)
2347 blame_move_score = BLAME_DEFAULT_MOVE_SCORE;
2348 if (!blame_copy_score)
2349 blame_copy_score = BLAME_DEFAULT_COPY_SCORE;
2350
1732a1fd
JH
2351 /*
2352 * We have collected options unknown to us in argv[1..unk]
cee7f245 2353 * which are to be passed to revision machinery if we are
3dff5379 2354 * going to do the "bottom" processing.
cee7f245
JH
2355 *
2356 * The remaining are:
2357 *
22e5e58a 2358 * (1) if dashdash_pos != 0, it is either
3f8d5204
PH
2359 * "blame [revisions] -- <path>" or
2360 * "blame -- <path> <rev>"
cee7f245 2361 *
22e5e58a 2362 * (2) otherwise, it is one of the two:
3f8d5204
PH
2363 * "blame [revisions] <path>"
2364 * "blame <path> <rev>"
cee7f245 2365 *
3f8d5204
PH
2366 * Note that we must strip out <path> from the arguments: we do not
2367 * want the path pruning but we may want "bottom" processing.
cee7f245 2368 */
3f8d5204
PH
2369 if (dashdash_pos) {
2370 switch (argc - dashdash_pos - 1) {
2371 case 2: /* (1b) */
2372 if (argc != 4)
5817da01 2373 usage_with_options(blame_opt_usage, options);
3f8d5204
PH
2374 /* reorder for the new way: <rev> -- <path> */
2375 argv[1] = argv[3];
2376 argv[3] = argv[2];
2377 argv[2] = "--";
2378 /* FALLTHROUGH */
2379 case 1: /* (1a) */
2380 path = add_prefix(prefix, argv[--argc]);
2381 argv[argc] = NULL;
2382 break;
2383 default:
2384 usage_with_options(blame_opt_usage, options);
cee7f245 2385 }
3f8d5204
PH
2386 } else {
2387 if (argc < 2)
5817da01 2388 usage_with_options(blame_opt_usage, options);
3f8d5204 2389 path = add_prefix(prefix, argv[argc - 1]);
c455c87c 2390 if (argc == 3 && !has_string_in_work_tree(path)) { /* (2b) */
3f8d5204
PH
2391 path = add_prefix(prefix, argv[1]);
2392 argv[1] = argv[2];
cee7f245 2393 }
3f8d5204 2394 argv[argc - 1] = "--";
cee7f245 2395
3f8d5204 2396 setup_work_tree();
c455c87c 2397 if (!has_string_in_work_tree(path))
d824cbba 2398 die_errno("cannot stat path '%s'", path);
cee7f245
JH
2399 }
2400
8b3dce56 2401 revs.disable_stdin = 1;
3f8d5204 2402 setup_revisions(argc, argv, &revs, NULL);
cee7f245
JH
2403 memset(&sb, 0, sizeof(sb));
2404
85af7929
JH
2405 sb.revs = &revs;
2406 if (!reverse)
2407 final_commit_name = prepare_final(&sb);
2408 else if (contents_from)
2409 die("--contents and --children do not blend well.");
2410 else
2411 final_commit_name = prepare_initial(&sb);
cee7f245
JH
2412
2413 if (!sb.final) {
1732a1fd
JH
2414 /*
2415 * "--not A B -- path" without anything positive;
1cfe7733
JH
2416 * do not default to HEAD, but use the working tree
2417 * or "--contents".
1732a1fd 2418 */
1981820b 2419 setup_work_tree();
3b8a12e8
AB
2420 sb.final = fake_working_tree_commit(&sb.revs->diffopt,
2421 path, contents_from);
1cfe7733 2422 add_pending_object(&revs, &(sb.final->object), ":");
cee7f245 2423 }
1cfe7733
JH
2424 else if (contents_from)
2425 die("Cannot use --contents with final commit object name");
cee7f245 2426
1732a1fd
JH
2427 /*
2428 * If we have bottom, this will mark the ancestors of the
cee7f245
JH
2429 * bottom commits we would reach while traversing as
2430 * uninteresting.
2431 */
3d51e1b5
MK
2432 if (prepare_revision_walk(&revs))
2433 die("revision walk setup failed");
cee7f245 2434
1cfe7733
JH
2435 if (is_null_sha1(sb.final->object.sha1)) {
2436 char *buf;
2437 o = sb.final->util;
2438 buf = xmalloc(o->file.size + 1);
2439 memcpy(buf, o->file.ptr, o->file.size + 1);
2440 sb.final_buf = buf;
2441 sb.final_buf_size = o->file.size;
2442 }
2443 else {
2444 o = get_origin(&sb, sb.final, path);
90064710 2445 if (fill_blob_sha1_and_mode(o))
1cfe7733 2446 die("no such path %s in %s", path, final_commit_name);
cee7f245 2447
3b8a12e8 2448 if (DIFF_OPT_TST(&sb.revs->diffopt, ALLOW_TEXTCONV) &&
e5450100 2449 textconv_object(path, o->mode, o->blob_sha1, 1, (char **) &sb.final_buf,
3b8a12e8
AB
2450 &sb.final_buf_size))
2451 ;
2452 else
2453 sb.final_buf = read_sha1_file(o->blob_sha1, &type,
2454 &sb.final_buf_size);
2455
ab43e495
JH
2456 if (!sb.final_buf)
2457 die("Cannot read blob %s for path %s",
2458 sha1_to_hex(o->blob_sha1),
2459 path);
1cfe7733 2460 }
c2e525d9 2461 num_read_blob++;
cee7f245
JH
2462 lno = prepare_lines(&sb);
2463
58dbfa2e
ES
2464 if (lno && !range_list.nr)
2465 string_list_append(&range_list, xstrdup("1"));
2466
52f4d126 2467 anchor = 1;
58dbfa2e
ES
2468 range_set_init(&ranges, range_list.nr);
2469 for (range_i = 0; range_i < range_list.nr; ++range_i) {
2470 long bottom, top;
2471 if (parse_range_arg(range_list.items[range_i].string,
52f4d126 2472 nth_line_cb, &sb, lno, anchor,
58dbfa2e
ES
2473 &bottom, &top, sb.path))
2474 usage(blame_usage);
2475 if (lno < top || ((lno || bottom) && lno < bottom))
2476 die("file %s has only %lu lines", path, lno);
2477 if (bottom < 1)
2478 bottom = 1;
2479 if (top < 1)
2480 top = lno;
2481 bottom--;
2482 range_set_append_unsafe(&ranges, bottom, top);
52f4d126 2483 anchor = top + 1;
58dbfa2e
ES
2484 }
2485 sort_and_merge_range_set(&ranges);
2486
2487 for (range_i = ranges.nr; range_i > 0; --range_i) {
2488 const struct range *r = &ranges.ranges[range_i - 1];
2489 long bottom = r->start;
2490 long top = r->end;
2491 struct blame_entry *next = ent;
2492 ent = xcalloc(1, sizeof(*ent));
2493 ent->lno = bottom;
2494 ent->num_lines = top - bottom;
2495 ent->suspect = o;
2496 ent->s_lno = bottom;
2497 ent->next = next;
58dbfa2e
ES
2498 origin_incref(o);
2499 }
2500 origin_decref(o);
2501
2502 range_set_release(&ranges);
2503 string_list_clear(&range_list, 0);
cee7f245
JH
2504
2505 sb.ent = ent;
2506 sb.path = path;
2507
d551a488 2508 read_mailmap(&mailmap, NULL);
f9567384 2509
b92565dc
MH
2510 if (!incremental)
2511 setup_pager();
2512
85af7929 2513 assign_blame(&sb, opt);
cee7f245 2514
717d1462
LT
2515 if (incremental)
2516 return 0;
2517
cee7f245
JH
2518 coalesce(&sb);
2519
2520 if (!(output_option & OUTPUT_PORCELAIN))
2521 find_alignment(&sb, &output_option);
2522
2523 output(&sb, output_option);
2524 free((void *)sb.final_buf);
2525 for (ent = sb.ent; ent; ) {
2526 struct blame_entry *e = ent->next;
2527 free(ent);
2528 ent = e;
2529 }
c2e525d9 2530
870b39c1 2531 if (show_stats) {
c2e525d9
JH
2532 printf("num read blob: %d\n", num_read_blob);
2533 printf("num get patch: %d\n", num_get_patch);
2534 printf("num commits: %d\n", num_commits);
2535 }
cee7f245
JH
2536 return 0;
2537}