textconv: make the API public
[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"
cee7f245 23
588c038a 24static char blame_usage[] = "git blame [options] [rev-opts] [rev] [--] file";
5817da01
PH
25
26static const char *blame_opt_usage[] = {
27 blame_usage,
28 "",
76447d23 29 "[rev-opts] are documented in git-rev-list(1)",
5817da01
PH
30 NULL
31};
cee7f245
JH
32
33static int longest_file;
34static int longest_author;
35static int max_orig_digits;
36static int max_digits;
5ff62c30 37static int max_score_digits;
4c10a5ca 38static int show_root;
85af7929 39static int reverse;
4c10a5ca 40static int blank_boundary;
717d1462 41static int incremental;
b82871b3 42static int xdl_opts = XDF_NEED_MINIMAL;
31653c1a
EL
43
44static enum date_mode blame_date_mode = DATE_ISO8601;
45static size_t blame_date_width;
46
c455c87c 47static struct string_list mailmap;
cee7f245 48
54a4c617
JH
49#ifndef DEBUG
50#define DEBUG 0
51#endif
52
c2e525d9
JH
53/* stats */
54static int num_read_blob;
55static int num_get_patch;
56static int num_commits;
57
d24bba80 58#define PICKAXE_BLAME_MOVE 01
18abd745
JH
59#define PICKAXE_BLAME_COPY 02
60#define PICKAXE_BLAME_COPY_HARDER 04
c63777c0 61#define PICKAXE_BLAME_COPY_HARDEST 010
d24bba80 62
4a0fc95f
JH
63/*
64 * blame for a blame_entry with score lower than these thresholds
65 * is not passed to the parent using move/copy logic.
66 */
67static unsigned blame_move_score;
68static unsigned blame_copy_score;
69#define BLAME_DEFAULT_MOVE_SCORE 20
70#define BLAME_DEFAULT_COPY_SCORE 40
71
cee7f245
JH
72/* bits #0..7 in revision.h, #8..11 used for merge_bases() in commit.c */
73#define METAINFO_SHOWN (1u<<12)
74#define MORE_THAN_ONE_PATH (1u<<13)
75
76/*
54a4c617 77 * One blob in a commit that is being suspected
cee7f245
JH
78 */
79struct origin {
54a4c617 80 int refcnt;
96e11709 81 struct origin *previous;
cee7f245 82 struct commit *commit;
c2e525d9 83 mmfile_t file;
cee7f245
JH
84 unsigned char blob_sha1[20];
85 char path[FLEX_ARRAY];
86};
87
1732a1fd
JH
88/*
89 * Given an origin, prepare mmfile_t structure to be used by the
90 * diff machinery
91 */
f6c07d7d 92static void fill_origin_blob(struct origin *o, mmfile_t *file)
c2e525d9
JH
93{
94 if (!o->file.ptr) {
21666f1a 95 enum object_type type;
c2e525d9 96 num_read_blob++;
21666f1a 97 file->ptr = read_sha1_file(o->blob_sha1, &type,
c2e525d9 98 (unsigned long *)(&(file->size)));
ab43e495
JH
99 if (!file->ptr)
100 die("Cannot read blob %s for path %s",
101 sha1_to_hex(o->blob_sha1),
102 o->path);
c2e525d9
JH
103 o->file = *file;
104 }
105 else
106 *file = o->file;
c2e525d9
JH
107}
108
1732a1fd
JH
109/*
110 * Origin is refcounted and usually we keep the blob contents to be
111 * reused.
112 */
54a4c617
JH
113static inline struct origin *origin_incref(struct origin *o)
114{
115 if (o)
116 o->refcnt++;
117 return o;
118}
119
120static void origin_decref(struct origin *o)
121{
122 if (o && --o->refcnt <= 0) {
96e11709
JH
123 if (o->previous)
124 origin_decref(o->previous);
8e0f7003 125 free(o->file.ptr);
54a4c617
JH
126 free(o);
127 }
128}
129
7c3c7962
JH
130static void drop_origin_blob(struct origin *o)
131{
132 if (o->file.ptr) {
133 free(o->file.ptr);
134 o->file.ptr = NULL;
135 }
136}
137
1732a1fd
JH
138/*
139 * Each group of lines is described by a blame_entry; it can be split
140 * as we pass blame to the parents. They form a linked list in the
141 * scoreboard structure, sorted by the target line number.
142 */
cee7f245
JH
143struct blame_entry {
144 struct blame_entry *prev;
145 struct blame_entry *next;
146
147 /* the first line of this group in the final image;
148 * internally all line numbers are 0 based.
149 */
150 int lno;
151
152 /* how many lines this group has */
153 int num_lines;
154
155 /* the commit that introduced this group into the final image */
156 struct origin *suspect;
157
158 /* true if the suspect is truly guilty; false while we have not
159 * checked if the group came from one of its parents.
160 */
161 char guilty;
162
eb46328e
AG
163 /* true if the entry has been scanned for copies in the current parent
164 */
165 char scanned;
166
cee7f245
JH
167 /* the line number of the first line of this group in the
168 * suspect's file; internally all line numbers are 0 based.
169 */
170 int s_lno;
5ff62c30
JH
171
172 /* how significant this entry is -- cached to avoid
1732a1fd 173 * scanning the lines over and over.
5ff62c30
JH
174 */
175 unsigned score;
cee7f245
JH
176};
177
1732a1fd
JH
178/*
179 * The current state of the blame assignment.
180 */
cee7f245
JH
181struct scoreboard {
182 /* the final commit (i.e. where we started digging from) */
183 struct commit *final;
85af7929 184 struct rev_info *revs;
cee7f245
JH
185 const char *path;
186
1732a1fd
JH
187 /*
188 * The contents in the final image.
189 * Used by many functions to obtain contents of the nth line,
190 * indexed with scoreboard.lineno[blame_entry.lno].
cee7f245
JH
191 */
192 const char *final_buf;
193 unsigned long final_buf_size;
194
195 /* linked list of blames */
196 struct blame_entry *ent;
197
612702e8 198 /* look-up a line in the final buffer */
cee7f245
JH
199 int num_lines;
200 int *lineno;
201};
202
171dccd5 203static inline int same_suspect(struct origin *a, struct origin *b)
46014766 204{
171dccd5 205 if (a == b)
57584d9e 206 return 1;
171dccd5
JH
207 if (a->commit != b->commit)
208 return 0;
209 return !strcmp(a->path, b->path);
46014766
JH
210}
211
54a4c617
JH
212static void sanity_check_refcnt(struct scoreboard *);
213
1732a1fd
JH
214/*
215 * If two blame entries that are next to each other came from
216 * contiguous lines in the same origin (i.e. <commit, path> pair),
217 * merge them together.
218 */
cee7f245
JH
219static void coalesce(struct scoreboard *sb)
220{
221 struct blame_entry *ent, *next;
222
223 for (ent = sb->ent; ent && (next = ent->next); ent = next) {
171dccd5 224 if (same_suspect(ent->suspect, next->suspect) &&
cee7f245
JH
225 ent->guilty == next->guilty &&
226 ent->s_lno + ent->num_lines == next->s_lno) {
227 ent->num_lines += next->num_lines;
228 ent->next = next->next;
229 if (ent->next)
230 ent->next->prev = ent;
54a4c617 231 origin_decref(next->suspect);
cee7f245 232 free(next);
46014766 233 ent->score = 0;
cee7f245
JH
234 next = ent; /* again */
235 }
236 }
54a4c617
JH
237
238 if (DEBUG) /* sanity */
239 sanity_check_refcnt(sb);
cee7f245
JH
240}
241
1732a1fd
JH
242/*
243 * Given a commit and a path in it, create a new origin structure.
244 * The callers that add blame to the scoreboard should use
245 * get_origin() to obtain shared, refcounted copy instead of calling
246 * this function directly.
247 */
854b97f6
JH
248static struct origin *make_origin(struct commit *commit, const char *path)
249{
250 struct origin *o;
251 o = xcalloc(1, sizeof(*o) + strlen(path) + 1);
252 o->commit = commit;
253 o->refcnt = 1;
254 strcpy(o->path, path);
255 return o;
256}
257
1732a1fd
JH
258/*
259 * Locate an existing origin or create a new one.
260 */
f6c0e191
JH
261static struct origin *get_origin(struct scoreboard *sb,
262 struct commit *commit,
263 const char *path)
cee7f245 264{
f6c0e191 265 struct blame_entry *e;
cee7f245 266
f6c0e191
JH
267 for (e = sb->ent; e; e = e->next) {
268 if (e->suspect->commit == commit &&
269 !strcmp(e->suspect->path, path))
54a4c617 270 return origin_incref(e->suspect);
cee7f245 271 }
854b97f6 272 return make_origin(commit, path);
cee7f245
JH
273}
274
1732a1fd
JH
275/*
276 * Fill the blob_sha1 field of an origin if it hasn't, so that later
277 * call to fill_origin_blob() can use it to locate the data. blob_sha1
278 * for an origin is also used to pass the blame for the entire file to
279 * the parent to detect the case where a child's blob is identical to
280 * that of its parent's.
281 */
f6c0e191
JH
282static int fill_blob_sha1(struct origin *origin)
283{
284 unsigned mode;
f6c0e191
JH
285
286 if (!is_null_sha1(origin->blob_sha1))
287 return 0;
288 if (get_tree_entry(origin->commit->object.sha1,
289 origin->path,
290 origin->blob_sha1, &mode))
291 goto error_out;
21666f1a 292 if (sha1_object_info(origin->blob_sha1, NULL) != OBJ_BLOB)
f6c0e191
JH
293 goto error_out;
294 return 0;
295 error_out:
296 hashclr(origin->blob_sha1);
297 return -1;
298}
299
1732a1fd
JH
300/*
301 * We have an origin -- check if the same path exists in the
302 * parent and return an origin structure to represent it.
303 */
f6c0e191 304static struct origin *find_origin(struct scoreboard *sb,
cee7f245
JH
305 struct commit *parent,
306 struct origin *origin)
307{
308 struct origin *porigin = NULL;
309 struct diff_options diff_opts;
f6c0e191
JH
310 const char *paths[2];
311
0d981c67 312 if (parent->util) {
1732a1fd
JH
313 /*
314 * Each commit object can cache one origin in that
315 * commit. This is a freestanding copy of origin and
316 * not refcounted.
854b97f6 317 */
0d981c67 318 struct origin *cached = parent->util;
854b97f6 319 if (!strcmp(cached->path, origin->path)) {
1732a1fd
JH
320 /*
321 * The same path between origin and its parent
322 * without renaming -- the most common case.
323 */
854b97f6 324 porigin = get_origin(sb, parent, cached->path);
1732a1fd
JH
325
326 /*
327 * If the origin was newly created (i.e. get_origin
328 * would call make_origin if none is found in the
329 * scoreboard), it does not know the blob_sha1,
330 * so copy it. Otherwise porigin was in the
331 * scoreboard and already knows blob_sha1.
332 */
854b97f6
JH
333 if (porigin->refcnt == 1)
334 hashcpy(porigin->blob_sha1, cached->blob_sha1);
335 return porigin;
336 }
337 /* otherwise it was not very useful; free it */
338 free(parent->util);
339 parent->util = NULL;
0d981c67
JH
340 }
341
f6c0e191
JH
342 /* See if the origin->path is different between parent
343 * and origin first. Most of the time they are the
344 * same and diff-tree is fairly efficient about this.
345 */
346 diff_setup(&diff_opts);
8f67f8ae 347 DIFF_OPT_SET(&diff_opts, RECURSIVE);
f6c0e191
JH
348 diff_opts.detect_rename = 0;
349 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
350 paths[0] = origin->path;
351 paths[1] = NULL;
352
353 diff_tree_setup_paths(paths, &diff_opts);
354 if (diff_setup_done(&diff_opts) < 0)
355 die("diff-setup");
1cfe7733
JH
356
357 if (is_null_sha1(origin->commit->object.sha1))
358 do_diff_cache(parent->tree->object.sha1, &diff_opts);
359 else
360 diff_tree_sha1(parent->tree->object.sha1,
361 origin->commit->tree->object.sha1,
362 "", &diff_opts);
f6c0e191
JH
363 diffcore_std(&diff_opts);
364
f6c0e191
JH
365 if (!diff_queued_diff.nr) {
366 /* The path is the same as parent */
367 porigin = get_origin(sb, parent, origin->path);
368 hashcpy(porigin->blob_sha1, origin->blob_sha1);
a9b2d424
JH
369 } else {
370 /*
371 * Since origin->path is a pathspec, if the parent
372 * commit had it as a directory, we will see a whole
373 * bunch of deletion of files in the directory that we
374 * do not care about.
375 */
376 int i;
377 struct diff_filepair *p = NULL;
378 for (i = 0; i < diff_queued_diff.nr; i++) {
379 const char *name;
380 p = diff_queued_diff.queue[i];
381 name = p->one->path ? p->one->path : p->two->path;
382 if (!strcmp(name, origin->path))
383 break;
384 }
385 if (!p)
386 die("internal error in blame::find_origin");
f6c0e191
JH
387 switch (p->status) {
388 default:
acca687f 389 die("internal error in blame::find_origin (%c)",
f6c0e191
JH
390 p->status);
391 case 'M':
392 porigin = get_origin(sb, parent, origin->path);
393 hashcpy(porigin->blob_sha1, p->one->sha1);
394 break;
395 case 'A':
396 case 'T':
397 /* Did not exist in parent, or type changed */
398 break;
399 }
400 }
401 diff_flush(&diff_opts);
03b69c76 402 diff_tree_release_paths(&diff_opts);
0d981c67 403 if (porigin) {
1732a1fd
JH
404 /*
405 * Create a freestanding copy that is not part of
406 * the refcounted origin found in the scoreboard, and
407 * cache it in the commit.
408 */
854b97f6 409 struct origin *cached;
1732a1fd 410
854b97f6
JH
411 cached = make_origin(porigin->commit, porigin->path);
412 hashcpy(cached->blob_sha1, porigin->blob_sha1);
413 parent->util = cached;
0d981c67 414 }
f69e743d
JH
415 return porigin;
416}
f6c0e191 417
1732a1fd
JH
418/*
419 * We have an origin -- find the path that corresponds to it in its
420 * parent and return an origin structure to represent it.
421 */
f69e743d
JH
422static struct origin *find_rename(struct scoreboard *sb,
423 struct commit *parent,
424 struct origin *origin)
425{
426 struct origin *porigin = NULL;
427 struct diff_options diff_opts;
428 int i;
429 const char *paths[2];
cee7f245
JH
430
431 diff_setup(&diff_opts);
8f67f8ae 432 DIFF_OPT_SET(&diff_opts, RECURSIVE);
cee7f245
JH
433 diff_opts.detect_rename = DIFF_DETECT_RENAME;
434 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
2f3f8b21 435 diff_opts.single_follow = origin->path;
cee7f245
JH
436 paths[0] = NULL;
437 diff_tree_setup_paths(paths, &diff_opts);
438 if (diff_setup_done(&diff_opts) < 0)
439 die("diff-setup");
1cfe7733
JH
440
441 if (is_null_sha1(origin->commit->object.sha1))
442 do_diff_cache(parent->tree->object.sha1, &diff_opts);
443 else
444 diff_tree_sha1(parent->tree->object.sha1,
445 origin->commit->tree->object.sha1,
446 "", &diff_opts);
cee7f245
JH
447 diffcore_std(&diff_opts);
448
449 for (i = 0; i < diff_queued_diff.nr; i++) {
450 struct diff_filepair *p = diff_queued_diff.queue[i];
612702e8 451 if ((p->status == 'R' || p->status == 'C') &&
f6c0e191
JH
452 !strcmp(p->two->path, origin->path)) {
453 porigin = get_origin(sb, parent, p->one->path);
454 hashcpy(porigin->blob_sha1, p->one->sha1);
cee7f245
JH
455 break;
456 }
457 }
458 diff_flush(&diff_opts);
03b69c76 459 diff_tree_release_paths(&diff_opts);
cee7f245
JH
460 return porigin;
461}
462
1732a1fd 463/*
3dff5379 464 * Link in a new blame entry to the scoreboard. Entries that cover the
1732a1fd
JH
465 * same line range have been removed from the scoreboard previously.
466 */
cee7f245
JH
467static void add_blame_entry(struct scoreboard *sb, struct blame_entry *e)
468{
469 struct blame_entry *ent, *prev = NULL;
470
54a4c617
JH
471 origin_incref(e->suspect);
472
cee7f245
JH
473 for (ent = sb->ent; ent && ent->lno < e->lno; ent = ent->next)
474 prev = ent;
475
476 /* prev, if not NULL, is the last one that is below e */
477 e->prev = prev;
478 if (prev) {
479 e->next = prev->next;
480 prev->next = e;
481 }
482 else {
483 e->next = sb->ent;
484 sb->ent = e;
485 }
486 if (e->next)
487 e->next->prev = e;
488}
489
1732a1fd
JH
490/*
491 * src typically is on-stack; we want to copy the information in it to
790296fd 492 * a malloced blame_entry that is already on the linked list of the
1732a1fd
JH
493 * scoreboard. The origin of dst loses a refcnt while the origin of src
494 * gains one.
495 */
cee7f245
JH
496static void dup_entry(struct blame_entry *dst, struct blame_entry *src)
497{
498 struct blame_entry *p, *n;
54a4c617 499
cee7f245
JH
500 p = dst->prev;
501 n = dst->next;
54a4c617
JH
502 origin_incref(src->suspect);
503 origin_decref(dst->suspect);
cee7f245
JH
504 memcpy(dst, src, sizeof(*src));
505 dst->prev = p;
506 dst->next = n;
5ff62c30 507 dst->score = 0;
cee7f245
JH
508}
509
510static const char *nth_line(struct scoreboard *sb, int lno)
511{
512 return sb->final_buf + sb->lineno[lno];
513}
514
1732a1fd
JH
515/*
516 * It is known that lines between tlno to same came from parent, and e
517 * has an overlap with that range. it also is known that parent's
518 * line plno corresponds to e's line tlno.
519 *
520 * <---- e ----->
521 * <------>
522 * <------------>
523 * <------------>
524 * <------------------>
525 *
526 * Split e into potentially three parts; before this chunk, the chunk
527 * to be blamed for the parent, and after that portion.
528 */
54a4c617 529static void split_overlap(struct blame_entry *split,
cee7f245
JH
530 struct blame_entry *e,
531 int tlno, int plno, int same,
532 struct origin *parent)
533{
cee7f245
JH
534 int chunk_end_lno;
535 memset(split, 0, sizeof(struct blame_entry [3]));
536
537 if (e->s_lno < tlno) {
538 /* there is a pre-chunk part not blamed on parent */
54a4c617 539 split[0].suspect = origin_incref(e->suspect);
cee7f245
JH
540 split[0].lno = e->lno;
541 split[0].s_lno = e->s_lno;
542 split[0].num_lines = tlno - e->s_lno;
543 split[1].lno = e->lno + tlno - e->s_lno;
544 split[1].s_lno = plno;
545 }
546 else {
547 split[1].lno = e->lno;
548 split[1].s_lno = plno + (e->s_lno - tlno);
549 }
550
551 if (same < e->s_lno + e->num_lines) {
552 /* there is a post-chunk part not blamed on parent */
54a4c617 553 split[2].suspect = origin_incref(e->suspect);
cee7f245
JH
554 split[2].lno = e->lno + (same - e->s_lno);
555 split[2].s_lno = e->s_lno + (same - e->s_lno);
556 split[2].num_lines = e->s_lno + e->num_lines - same;
557 chunk_end_lno = split[2].lno;
558 }
559 else
560 chunk_end_lno = e->lno + e->num_lines;
561 split[1].num_lines = chunk_end_lno - split[1].lno;
562
1732a1fd
JH
563 /*
564 * if it turns out there is nothing to blame the parent for,
565 * forget about the splitting. !split[1].suspect signals this.
566 */
cee7f245
JH
567 if (split[1].num_lines < 1)
568 return;
54a4c617 569 split[1].suspect = origin_incref(parent);
cee7f245
JH
570}
571
1732a1fd
JH
572/*
573 * split_overlap() divided an existing blame e into up to three parts
574 * in split. Adjust the linked list of blames in the scoreboard to
575 * reflect the split.
576 */
cee7f245 577static void split_blame(struct scoreboard *sb,
54a4c617 578 struct blame_entry *split,
cee7f245
JH
579 struct blame_entry *e)
580{
581 struct blame_entry *new_entry;
582
583 if (split[0].suspect && split[2].suspect) {
1732a1fd 584 /* The first part (reuse storage for the existing entry e) */
cee7f245
JH
585 dup_entry(e, &split[0]);
586
1732a1fd 587 /* The last part -- me */
cee7f245
JH
588 new_entry = xmalloc(sizeof(*new_entry));
589 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
590 add_blame_entry(sb, new_entry);
591
1732a1fd 592 /* ... and the middle part -- parent */
cee7f245
JH
593 new_entry = xmalloc(sizeof(*new_entry));
594 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
595 add_blame_entry(sb, new_entry);
596 }
597 else if (!split[0].suspect && !split[2].suspect)
1732a1fd
JH
598 /*
599 * The parent covers the entire area; reuse storage for
600 * e and replace it with the parent.
601 */
cee7f245
JH
602 dup_entry(e, &split[1]);
603 else if (split[0].suspect) {
1732a1fd 604 /* me and then parent */
cee7f245
JH
605 dup_entry(e, &split[0]);
606
607 new_entry = xmalloc(sizeof(*new_entry));
608 memcpy(new_entry, &(split[1]), sizeof(struct blame_entry));
609 add_blame_entry(sb, new_entry);
610 }
611 else {
1732a1fd 612 /* parent and then me */
cee7f245
JH
613 dup_entry(e, &split[1]);
614
615 new_entry = xmalloc(sizeof(*new_entry));
616 memcpy(new_entry, &(split[2]), sizeof(struct blame_entry));
617 add_blame_entry(sb, new_entry);
618 }
619
54a4c617 620 if (DEBUG) { /* sanity */
cee7f245 621 struct blame_entry *ent;
612702e8 622 int lno = sb->ent->lno, corrupt = 0;
cee7f245
JH
623
624 for (ent = sb->ent; ent; ent = ent->next) {
625 if (lno != ent->lno)
626 corrupt = 1;
627 if (ent->s_lno < 0)
628 corrupt = 1;
629 lno += ent->num_lines;
630 }
631 if (corrupt) {
612702e8 632 lno = sb->ent->lno;
cee7f245
JH
633 for (ent = sb->ent; ent; ent = ent->next) {
634 printf("L %8d l %8d n %8d\n",
635 lno, ent->lno, ent->num_lines);
636 lno = ent->lno + ent->num_lines;
637 }
638 die("oops");
639 }
640 }
641}
642
1732a1fd
JH
643/*
644 * After splitting the blame, the origins used by the
645 * on-stack blame_entry should lose one refcnt each.
646 */
54a4c617
JH
647static void decref_split(struct blame_entry *split)
648{
649 int i;
650
651 for (i = 0; i < 3; i++)
652 origin_decref(split[i].suspect);
653}
654
1732a1fd
JH
655/*
656 * Helper for blame_chunk(). blame_entry e is known to overlap with
657 * the patch hunk; split it and pass blame to the parent.
658 */
cee7f245
JH
659static void blame_overlap(struct scoreboard *sb, struct blame_entry *e,
660 int tlno, int plno, int same,
661 struct origin *parent)
662{
663 struct blame_entry split[3];
664
665 split_overlap(split, e, tlno, plno, same, parent);
54a4c617
JH
666 if (split[1].suspect)
667 split_blame(sb, split, e);
668 decref_split(split);
cee7f245
JH
669}
670
1732a1fd
JH
671/*
672 * Find the line number of the last line the target is suspected for.
673 */
cee7f245
JH
674static int find_last_in_target(struct scoreboard *sb, struct origin *target)
675{
676 struct blame_entry *e;
677 int last_in_target = -1;
678
679 for (e = sb->ent; e; e = e->next) {
171dccd5 680 if (e->guilty || !same_suspect(e->suspect, target))
cee7f245
JH
681 continue;
682 if (last_in_target < e->s_lno + e->num_lines)
683 last_in_target = e->s_lno + e->num_lines;
684 }
685 return last_in_target;
686}
687
1732a1fd
JH
688/*
689 * Process one hunk from the patch between the current suspect for
690 * blame_entry e and its parent. Find and split the overlap, and
691 * pass blame to the overlapping part to the parent.
692 */
cee7f245
JH
693static void blame_chunk(struct scoreboard *sb,
694 int tlno, int plno, int same,
695 struct origin *target, struct origin *parent)
696{
612702e8 697 struct blame_entry *e;
cee7f245 698
612702e8 699 for (e = sb->ent; e; e = e->next) {
171dccd5 700 if (e->guilty || !same_suspect(e->suspect, target))
cee7f245
JH
701 continue;
702 if (same <= e->s_lno)
703 continue;
704 if (tlno < e->s_lno + e->num_lines)
705 blame_overlap(sb, e, tlno, plno, same, parent);
706 }
707}
708
cdcab133
RS
709struct blame_chunk_cb_data {
710 struct scoreboard *sb;
711 struct origin *target;
712 struct origin *parent;
713 long plno;
714 long tlno;
715};
716
717static void blame_chunk_cb(void *data, long same, long p_next, long t_next)
718{
719 struct blame_chunk_cb_data *d = data;
720 blame_chunk(d->sb, d->tlno, d->plno, same, d->target, d->parent);
721 d->plno = p_next;
722 d->tlno = t_next;
723}
724
1732a1fd
JH
725/*
726 * We are looking at the origin 'target' and aiming to pass blame
727 * for the lines it is suspected to its parent. Run diff to find
728 * which lines came from parent and pass blame for them.
729 */
cee7f245
JH
730static int pass_blame_to_parent(struct scoreboard *sb,
731 struct origin *target,
732 struct origin *parent)
733{
cdcab133 734 int last_in_target;
7c4ed8c8 735 mmfile_t file_p, file_o;
cdcab133
RS
736 struct blame_chunk_cb_data d = { sb, target, parent, 0, 0 };
737 xpparam_t xpp;
738 xdemitconf_t xecfg;
cee7f245
JH
739
740 last_in_target = find_last_in_target(sb, target);
741 if (last_in_target < 0)
742 return 1; /* nothing remains for this target */
743
7c4ed8c8
RS
744 fill_origin_blob(parent, &file_p);
745 fill_origin_blob(target, &file_o);
7c4ed8c8 746 num_get_patch++;
cee7f245 747
cdcab133
RS
748 memset(&xpp, 0, sizeof(xpp));
749 xpp.flags = xdl_opts;
750 memset(&xecfg, 0, sizeof(xecfg));
751 xecfg.ctxlen = 0;
752 xdi_diff_hunks(&file_p, &file_o, blame_chunk_cb, &d, &xpp, &xecfg);
1732a1fd 753 /* The rest (i.e. anything after tlno) are the same as the parent */
cdcab133 754 blame_chunk(sb, d.tlno, d.plno, last_in_target, target, parent);
cee7f245 755
cee7f245
JH
756 return 0;
757}
758
1732a1fd
JH
759/*
760 * The lines in blame_entry after splitting blames many times can become
761 * very small and trivial, and at some point it becomes pointless to
762 * blame the parents. E.g. "\t\t}\n\t}\n\n" appears everywhere in any
763 * ordinary C program, and it is not worth to say it was copied from
764 * totally unrelated file in the parent.
765 *
766 * Compute how trivial the lines in the blame_entry are.
767 */
5ff62c30
JH
768static unsigned ent_score(struct scoreboard *sb, struct blame_entry *e)
769{
770 unsigned score;
771 const char *cp, *ep;
772
773 if (e->score)
774 return e->score;
775
612702e8 776 score = 1;
5ff62c30
JH
777 cp = nth_line(sb, e->lno);
778 ep = nth_line(sb, e->lno + e->num_lines);
779 while (cp < ep) {
780 unsigned ch = *((unsigned char *)cp);
781 if (isalnum(ch))
782 score++;
783 cp++;
784 }
785 e->score = score;
786 return score;
787}
788
1732a1fd
JH
789/*
790 * best_so_far[] and this[] are both a split of an existing blame_entry
791 * that passes blame to the parent. Maintain best_so_far the best split
792 * so far, by comparing this and best_so_far and copying this into
793 * bst_so_far as needed.
794 */
5ff62c30 795static void copy_split_if_better(struct scoreboard *sb,
54a4c617
JH
796 struct blame_entry *best_so_far,
797 struct blame_entry *this)
d24bba80 798{
54a4c617
JH
799 int i;
800
d24bba80
JH
801 if (!this[1].suspect)
802 return;
5ff62c30
JH
803 if (best_so_far[1].suspect) {
804 if (ent_score(sb, &this[1]) < ent_score(sb, &best_so_far[1]))
805 return;
806 }
54a4c617
JH
807
808 for (i = 0; i < 3; i++)
809 origin_incref(this[i].suspect);
810 decref_split(best_so_far);
d24bba80
JH
811 memcpy(best_so_far, this, sizeof(struct blame_entry [3]));
812}
813
dd166aa8
JH
814/*
815 * We are looking at a part of the final image represented by
816 * ent (tlno and same are offset by ent->s_lno).
817 * tlno is where we are looking at in the final image.
818 * up to (but not including) same match preimage.
819 * plno is where we are looking at in the preimage.
820 *
821 * <-------------- final image ---------------------->
822 * <------ent------>
823 * ^tlno ^same
824 * <---------preimage----->
825 * ^plno
826 *
827 * All line numbers are 0-based.
828 */
829static void handle_split(struct scoreboard *sb,
830 struct blame_entry *ent,
831 int tlno, int plno, int same,
832 struct origin *parent,
833 struct blame_entry *split)
834{
835 if (ent->num_lines <= tlno)
836 return;
837 if (tlno < same) {
838 struct blame_entry this[3];
839 tlno += ent->s_lno;
840 same += ent->s_lno;
841 split_overlap(this, ent, tlno, plno, same, parent);
842 copy_split_if_better(sb, split, this);
843 decref_split(this);
844 }
845}
846
cdcab133
RS
847struct handle_split_cb_data {
848 struct scoreboard *sb;
849 struct blame_entry *ent;
850 struct origin *parent;
851 struct blame_entry *split;
852 long plno;
853 long tlno;
854};
855
856static void handle_split_cb(void *data, long same, long p_next, long t_next)
857{
858 struct handle_split_cb_data *d = data;
859 handle_split(d->sb, d->ent, d->tlno, d->plno, same, d->parent, d->split);
860 d->plno = p_next;
861 d->tlno = t_next;
862}
863
1732a1fd
JH
864/*
865 * Find the lines from parent that are the same as ent so that
866 * we can pass blames to it. file_p has the blob contents for
867 * the parent.
868 */
d24bba80
JH
869static void find_copy_in_blob(struct scoreboard *sb,
870 struct blame_entry *ent,
871 struct origin *parent,
54a4c617 872 struct blame_entry *split,
d24bba80
JH
873 mmfile_t *file_p)
874{
875 const char *cp;
876 int cnt;
877 mmfile_t file_o;
cdcab133
RS
878 struct handle_split_cb_data d = { sb, ent, parent, split, 0, 0 };
879 xpparam_t xpp;
880 xdemitconf_t xecfg;
d24bba80 881
1732a1fd
JH
882 /*
883 * Prepare mmfile that contains only the lines in ent.
884 */
d24bba80 885 cp = nth_line(sb, ent->lno);
4b25d091 886 file_o.ptr = (char *) cp;
d24bba80
JH
887 cnt = ent->num_lines;
888
889 while (cnt && cp < sb->final_buf + sb->final_buf_size) {
890 if (*cp++ == '\n')
891 cnt--;
892 }
893 file_o.size = cp - file_o.ptr;
894
dd166aa8
JH
895 /*
896 * file_o is a part of final image we are annotating.
897 * file_p partially may match that image.
898 */
cdcab133
RS
899 memset(&xpp, 0, sizeof(xpp));
900 xpp.flags = xdl_opts;
901 memset(&xecfg, 0, sizeof(xecfg));
902 xecfg.ctxlen = 1;
d24bba80 903 memset(split, 0, sizeof(struct blame_entry [3]));
cdcab133 904 xdi_diff_hunks(file_p, &file_o, handle_split_cb, &d, &xpp, &xecfg);
dd166aa8 905 /* remainder, if any, all match the preimage */
cdcab133 906 handle_split(sb, ent, d.tlno, d.plno, ent->num_lines, parent, split);
d24bba80
JH
907}
908
1732a1fd
JH
909/*
910 * See if lines currently target is suspected for can be attributed to
911 * parent.
912 */
d24bba80
JH
913static int find_move_in_parent(struct scoreboard *sb,
914 struct origin *target,
915 struct origin *parent)
916{
650e2f67 917 int last_in_target, made_progress;
46014766 918 struct blame_entry *e, split[3];
d24bba80 919 mmfile_t file_p;
d24bba80
JH
920
921 last_in_target = find_last_in_target(sb, target);
922 if (last_in_target < 0)
923 return 1; /* nothing remains for this target */
924
c2e525d9
JH
925 fill_origin_blob(parent, &file_p);
926 if (!file_p.ptr)
d24bba80 927 return 0;
d24bba80 928
650e2f67
JH
929 made_progress = 1;
930 while (made_progress) {
931 made_progress = 0;
932 for (e = sb->ent; e; e = e->next) {
1617fe98
AG
933 if (e->guilty || !same_suspect(e->suspect, target) ||
934 ent_score(sb, e) < blame_move_score)
650e2f67
JH
935 continue;
936 find_copy_in_blob(sb, e, parent, split, &file_p);
937 if (split[1].suspect &&
938 blame_move_score < ent_score(sb, &split[1])) {
939 split_blame(sb, split, e);
940 made_progress = 1;
941 }
942 decref_split(split);
943 }
d24bba80 944 }
d24bba80
JH
945 return 0;
946}
947
33494784
JH
948struct blame_list {
949 struct blame_entry *ent;
950 struct blame_entry split[3];
951};
952
1732a1fd
JH
953/*
954 * Count the number of entries the target is suspected for,
955 * and prepare a list of entry and the best split.
956 */
33494784
JH
957static struct blame_list *setup_blame_list(struct scoreboard *sb,
958 struct origin *target,
1617fe98 959 int min_score,
33494784
JH
960 int *num_ents_p)
961{
962 struct blame_entry *e;
963 int num_ents, i;
964 struct blame_list *blame_list = NULL;
965
33494784 966 for (e = sb->ent, num_ents = 0; e; e = e->next)
1617fe98
AG
967 if (!e->scanned && !e->guilty &&
968 same_suspect(e->suspect, target) &&
969 min_score < ent_score(sb, e))
33494784
JH
970 num_ents++;
971 if (num_ents) {
972 blame_list = xcalloc(num_ents, sizeof(struct blame_list));
973 for (e = sb->ent, i = 0; e; e = e->next)
1617fe98
AG
974 if (!e->scanned && !e->guilty &&
975 same_suspect(e->suspect, target) &&
976 min_score < ent_score(sb, e))
33494784
JH
977 blame_list[i++].ent = e;
978 }
979 *num_ents_p = num_ents;
980 return blame_list;
981}
982
eb46328e
AG
983/*
984 * Reset the scanned status on all entries.
985 */
986static void reset_scanned_flag(struct scoreboard *sb)
987{
988 struct blame_entry *e;
989 for (e = sb->ent; e; e = e->next)
990 e->scanned = 0;
991}
992
1732a1fd
JH
993/*
994 * For lines target is suspected for, see if we can find code movement
995 * across file boundary from the parent commit. porigin is the path
996 * in the parent we already tried.
997 */
18abd745
JH
998static int find_copy_in_parent(struct scoreboard *sb,
999 struct origin *target,
1000 struct commit *parent,
1001 struct origin *porigin,
1002 int opt)
1003{
1004 struct diff_options diff_opts;
1005 const char *paths[1];
aec8fa1f 1006 int i, j;
33494784
JH
1007 int retval;
1008 struct blame_list *blame_list;
aec8fa1f 1009 int num_ents;
18abd745 1010
1617fe98 1011 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784 1012 if (!blame_list)
18abd745
JH
1013 return 1; /* nothing remains for this target */
1014
1015 diff_setup(&diff_opts);
8f67f8ae 1016 DIFF_OPT_SET(&diff_opts, RECURSIVE);
18abd745
JH
1017 diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
1018
18abd745
JH
1019 paths[0] = NULL;
1020 diff_tree_setup_paths(paths, &diff_opts);
1021 if (diff_setup_done(&diff_opts) < 0)
1022 die("diff-setup");
33494784
JH
1023
1024 /* Try "find copies harder" on new path if requested;
1025 * we do not want to use diffcore_rename() actually to
1026 * match things up; find_copies_harder is set only to
1027 * force diff_tree_sha1() to feed all filepairs to diff_queue,
1028 * and this code needs to be after diff_setup_done(), which
1029 * usually makes find-copies-harder imply copy detection.
1030 */
c63777c0
JH
1031 if ((opt & PICKAXE_BLAME_COPY_HARDEST)
1032 || ((opt & PICKAXE_BLAME_COPY_HARDER)
1033 && (!porigin || strcmp(target->path, porigin->path))))
8f67f8ae 1034 DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
33494784 1035
1cfe7733
JH
1036 if (is_null_sha1(target->commit->object.sha1))
1037 do_diff_cache(parent->tree->object.sha1, &diff_opts);
1038 else
1039 diff_tree_sha1(parent->tree->object.sha1,
1040 target->commit->tree->object.sha1,
1041 "", &diff_opts);
18abd745 1042
8f67f8ae 1043 if (!DIFF_OPT_TST(&diff_opts, FIND_COPIES_HARDER))
33494784 1044 diffcore_std(&diff_opts);
18abd745 1045
33494784
JH
1046 retval = 0;
1047 while (1) {
1048 int made_progress = 0;
1049
1050 for (i = 0; i < diff_queued_diff.nr; i++) {
1051 struct diff_filepair *p = diff_queued_diff.queue[i];
1052 struct origin *norigin;
1053 mmfile_t file_p;
33494784
JH
1054 struct blame_entry this[3];
1055
1056 if (!DIFF_FILE_VALID(p->one))
1057 continue; /* does not exist in parent */
5209ac4d
AG
1058 if (S_ISGITLINK(p->one->mode))
1059 continue; /* ignore git links */
33494784
JH
1060 if (porigin && !strcmp(p->one->path, porigin->path))
1061 /* find_move already dealt with this path */
1062 continue;
1063
1064 norigin = get_origin(sb, parent, p->one->path);
1065 hashcpy(norigin->blob_sha1, p->one->sha1);
c2e525d9 1066 fill_origin_blob(norigin, &file_p);
33494784
JH
1067 if (!file_p.ptr)
1068 continue;
1069
1070 for (j = 0; j < num_ents; j++) {
1071 find_copy_in_blob(sb, blame_list[j].ent,
1072 norigin, this, &file_p);
1073 copy_split_if_better(sb, blame_list[j].split,
1074 this);
1075 decref_split(this);
1076 }
33494784
JH
1077 origin_decref(norigin);
1078 }
18abd745 1079
aec8fa1f 1080 for (j = 0; j < num_ents; j++) {
33494784
JH
1081 struct blame_entry *split = blame_list[j].split;
1082 if (split[1].suspect &&
1083 blame_copy_score < ent_score(sb, &split[1])) {
1084 split_blame(sb, split, blame_list[j].ent);
1085 made_progress = 1;
1086 }
eb46328e
AG
1087 else
1088 blame_list[j].ent->scanned = 1;
33494784 1089 decref_split(split);
18abd745 1090 }
33494784 1091 free(blame_list);
aec8fa1f 1092
33494784
JH
1093 if (!made_progress)
1094 break;
1617fe98 1095 blame_list = setup_blame_list(sb, target, blame_copy_score, &num_ents);
33494784
JH
1096 if (!blame_list) {
1097 retval = 1;
1098 break;
1099 }
18abd745 1100 }
eb46328e 1101 reset_scanned_flag(sb);
33494784 1102 diff_flush(&diff_opts);
03b69c76 1103 diff_tree_release_paths(&diff_opts);
33494784 1104 return retval;
18abd745
JH
1105}
1106
1732a1fd
JH
1107/*
1108 * The blobs of origin and porigin exactly match, so everything
c2e525d9
JH
1109 * origin is suspected for can be blamed on the parent.
1110 */
1111static void pass_whole_blame(struct scoreboard *sb,
1112 struct origin *origin, struct origin *porigin)
1113{
1114 struct blame_entry *e;
1115
1116 if (!porigin->file.ptr && origin->file.ptr) {
1117 /* Steal its file */
1118 porigin->file = origin->file;
1119 origin->file.ptr = NULL;
1120 }
1121 for (e = sb->ent; e; e = e->next) {
171dccd5 1122 if (!same_suspect(e->suspect, origin))
c2e525d9
JH
1123 continue;
1124 origin_incref(porigin);
1125 origin_decref(e->suspect);
1126 e->suspect = porigin;
1127 }
1128}
1129
69264f46
JH
1130/*
1131 * We pass blame from the current commit to its parents. We keep saying
1132 * "parent" (and "porigin"), but what we mean is to find scapegoat to
1133 * exonerate ourselves.
1134 */
85af7929 1135static struct commit_list *first_scapegoat(struct rev_info *revs, struct commit *commit)
69264f46 1136{
85af7929
JH
1137 if (!reverse)
1138 return commit->parents;
1139 return lookup_decoration(&revs->children, &commit->object);
69264f46
JH
1140}
1141
85af7929 1142static int num_scapegoats(struct rev_info *revs, struct commit *commit)
69264f46
JH
1143{
1144 int cnt;
85af7929 1145 struct commit_list *l = first_scapegoat(revs, commit);
69264f46
JH
1146 for (cnt = 0; l; l = l->next)
1147 cnt++;
1148 return cnt;
1149}
1150
1151#define MAXSG 16
cee7f245 1152
d24bba80 1153static void pass_blame(struct scoreboard *sb, struct origin *origin, int opt)
cee7f245 1154{
85af7929 1155 struct rev_info *revs = sb->revs;
69264f46 1156 int i, pass, num_sg;
cee7f245 1157 struct commit *commit = origin->commit;
69264f46
JH
1158 struct commit_list *sg;
1159 struct origin *sg_buf[MAXSG];
1160 struct origin *porigin, **sg_origin = sg_buf;
1161
85af7929 1162 num_sg = num_scapegoats(revs, commit);
69264f46
JH
1163 if (!num_sg)
1164 goto finish;
1165 else if (num_sg < ARRAY_SIZE(sg_buf))
1166 memset(sg_buf, 0, sizeof(sg_buf));
1167 else
1168 sg_origin = xcalloc(num_sg, sizeof(*sg_origin));
cee7f245 1169
69264f46
JH
1170 /*
1171 * The first pass looks for unrenamed path to optimize for
f69e743d
JH
1172 * common cases, then we look for renames in the second pass.
1173 */
1174 for (pass = 0; pass < 2; pass++) {
1175 struct origin *(*find)(struct scoreboard *,
1176 struct commit *, struct origin *);
1177 find = pass ? find_rename : find_origin;
1178
85af7929 1179 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1180 i < num_sg && sg;
1181 sg = sg->next, i++) {
1182 struct commit *p = sg->item;
0421d9f8 1183 int j, same;
f69e743d 1184
69264f46 1185 if (sg_origin[i])
f69e743d
JH
1186 continue;
1187 if (parse_commit(p))
1188 continue;
0d981c67 1189 porigin = find(sb, p, origin);
f69e743d
JH
1190 if (!porigin)
1191 continue;
1192 if (!hashcmp(porigin->blob_sha1, origin->blob_sha1)) {
c2e525d9 1193 pass_whole_blame(sb, origin, porigin);
f69e743d
JH
1194 origin_decref(porigin);
1195 goto finish;
1196 }
0421d9f8 1197 for (j = same = 0; j < i; j++)
69264f46
JH
1198 if (sg_origin[j] &&
1199 !hashcmp(sg_origin[j]->blob_sha1,
0421d9f8
JH
1200 porigin->blob_sha1)) {
1201 same = 1;
1202 break;
1203 }
1204 if (!same)
69264f46 1205 sg_origin[i] = porigin;
0421d9f8
JH
1206 else
1207 origin_decref(porigin);
cee7f245 1208 }
cee7f245
JH
1209 }
1210
c2e525d9 1211 num_commits++;
85af7929 1212 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1213 i < num_sg && sg;
1214 sg = sg->next, i++) {
1215 struct origin *porigin = sg_origin[i];
cee7f245
JH
1216 if (!porigin)
1217 continue;
96e11709
JH
1218 if (!origin->previous) {
1219 origin_incref(porigin);
1220 origin->previous = porigin;
1221 }
cee7f245 1222 if (pass_blame_to_parent(sb, origin, porigin))
54a4c617 1223 goto finish;
cee7f245 1224 }
d24bba80
JH
1225
1226 /*
1732a1fd 1227 * Optionally find moves in parents' files.
d24bba80
JH
1228 */
1229 if (opt & PICKAXE_BLAME_MOVE)
85af7929 1230 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1231 i < num_sg && sg;
1232 sg = sg->next, i++) {
1233 struct origin *porigin = sg_origin[i];
d24bba80
JH
1234 if (!porigin)
1235 continue;
1236 if (find_move_in_parent(sb, origin, porigin))
54a4c617 1237 goto finish;
d24bba80
JH
1238 }
1239
18abd745 1240 /*
1732a1fd 1241 * Optionally find copies from parents' files.
18abd745
JH
1242 */
1243 if (opt & PICKAXE_BLAME_COPY)
85af7929 1244 for (i = 0, sg = first_scapegoat(revs, commit);
69264f46
JH
1245 i < num_sg && sg;
1246 sg = sg->next, i++) {
1247 struct origin *porigin = sg_origin[i];
1248 if (find_copy_in_parent(sb, origin, sg->item,
18abd745 1249 porigin, opt))
54a4c617 1250 goto finish;
18abd745 1251 }
54a4c617
JH
1252
1253 finish:
69264f46
JH
1254 for (i = 0; i < num_sg; i++) {
1255 if (sg_origin[i]) {
1256 drop_origin_blob(sg_origin[i]);
1257 origin_decref(sg_origin[i]);
7c3c7962
JH
1258 }
1259 }
1260 drop_origin_blob(origin);
69264f46
JH
1261 if (sg_buf != sg_origin)
1262 free(sg_origin);
cee7f245
JH
1263}
1264
1732a1fd
JH
1265/*
1266 * Information on commits, used for output.
1267 */
cee7f245
JH
1268struct commit_info
1269{
3a55602e
SP
1270 const char *author;
1271 const char *author_mail;
cee7f245 1272 unsigned long author_time;
3a55602e 1273 const char *author_tz;
cee7f245
JH
1274
1275 /* filled only when asked for details */
3a55602e
SP
1276 const char *committer;
1277 const char *committer_mail;
cee7f245 1278 unsigned long committer_time;
3a55602e 1279 const char *committer_tz;
cee7f245 1280
3a55602e 1281 const char *summary;
cee7f245
JH
1282};
1283
1732a1fd
JH
1284/*
1285 * Parse author/committer line in the commit object buffer
1286 */
cee7f245 1287static void get_ac_line(const char *inbuf, const char *what,
d20d654f
MSO
1288 int person_len, char *person,
1289 int mail_len, char *mail,
3a55602e 1290 unsigned long *time, const char **tz)
cee7f245 1291{
f9567384 1292 int len, tzlen, maillen;
d20d654f 1293 char *tmp, *endp, *timepos, *mailpos;
cee7f245
JH
1294
1295 tmp = strstr(inbuf, what);
1296 if (!tmp)
1297 goto error_out;
1298 tmp += strlen(what);
1299 endp = strchr(tmp, '\n');
1300 if (!endp)
1301 len = strlen(tmp);
1302 else
1303 len = endp - tmp;
d20d654f 1304 if (person_len <= len) {
cee7f245
JH
1305 error_out:
1306 /* Ugh */
d20d654f 1307 *tz = "(unknown)";
c8cba791 1308 strcpy(person, *tz);
d20d654f 1309 strcpy(mail, *tz);
cee7f245
JH
1310 *time = 0;
1311 return;
1312 }
1313 memcpy(person, tmp, len);
1314
1315 tmp = person;
1316 tmp += len;
1317 *tmp = 0;
c8cba791 1318 while (person < tmp && *tmp != ' ')
cee7f245 1319 tmp--;
c8cba791
DR
1320 if (tmp <= person)
1321 goto error_out;
cee7f245 1322 *tz = tmp+1;
f9567384 1323 tzlen = (person+len)-(tmp+1);
cee7f245
JH
1324
1325 *tmp = 0;
c8cba791 1326 while (person < tmp && *tmp != ' ')
cee7f245 1327 tmp--;
c8cba791
DR
1328 if (tmp <= person)
1329 goto error_out;
cee7f245 1330 *time = strtoul(tmp, NULL, 10);
f9567384 1331 timepos = tmp;
cee7f245
JH
1332
1333 *tmp = 0;
c8cba791 1334 while (person < tmp && *tmp != ' ')
cee7f245 1335 tmp--;
c8cba791
DR
1336 if (tmp <= person)
1337 return;
d20d654f 1338 mailpos = tmp + 1;
cee7f245 1339 *tmp = 0;
f9567384 1340 maillen = timepos - tmp;
d20d654f 1341 memcpy(mail, mailpos, maillen);
f9567384
JH
1342
1343 if (!mailmap.nr)
1344 return;
1345
1346 /*
1347 * mailmap expansion may make the name longer.
1348 * make room by pushing stuff down.
1349 */
d20d654f 1350 tmp = person + person_len - (tzlen + 1);
f9567384
JH
1351 memmove(tmp, *tz, tzlen);
1352 tmp[tzlen] = 0;
1353 *tz = tmp;
1354
f9567384 1355 /*
d20d654f 1356 * Now, convert both name and e-mail using mailmap
f9567384 1357 */
eeefa7c9 1358 if (map_user(&mailmap, mail+1, mail_len-1, person, tmp-person-1)) {
d20d654f
MSO
1359 /* Add a trailing '>' to email, since map_user returns plain emails
1360 Note: It already has '<', since we replace from mail+1 */
1361 mailpos = memchr(mail, '\0', mail_len);
1362 if (mailpos && mailpos-mail < mail_len - 1) {
1363 *mailpos = '>';
1364 *(mailpos+1) = '\0';
1365 }
1366 }
cee7f245
JH
1367}
1368
1369static void get_commit_info(struct commit *commit,
1370 struct commit_info *ret,
1371 int detailed)
1372{
1373 int len;
69cd8f63 1374 char *tmp, *endp, *reencoded, *message;
d20d654f
MSO
1375 static char author_name[1024];
1376 static char author_mail[1024];
1377 static char committer_name[1024];
1378 static char committer_mail[1024];
cee7f245
JH
1379 static char summary_buf[1024];
1380
1732a1fd
JH
1381 /*
1382 * We've operated without save_commit_buffer, so
612702e8
JH
1383 * we now need to populate them for output.
1384 */
1385 if (!commit->buffer) {
21666f1a 1386 enum object_type type;
612702e8
JH
1387 unsigned long size;
1388 commit->buffer =
21666f1a 1389 read_sha1_file(commit->object.sha1, &type, &size);
ab43e495
JH
1390 if (!commit->buffer)
1391 die("Cannot read commit %s",
1392 sha1_to_hex(commit->object.sha1));
612702e8 1393 }
69cd8f63
AG
1394 reencoded = reencode_commit_message(commit, NULL);
1395 message = reencoded ? reencoded : commit->buffer;
d20d654f
MSO
1396 ret->author = author_name;
1397 ret->author_mail = author_mail;
69cd8f63 1398 get_ac_line(message, "\nauthor ",
d20d654f
MSO
1399 sizeof(author_name), author_name,
1400 sizeof(author_mail), author_mail,
cee7f245
JH
1401 &ret->author_time, &ret->author_tz);
1402
69cd8f63
AG
1403 if (!detailed) {
1404 free(reencoded);
cee7f245 1405 return;
69cd8f63 1406 }
cee7f245 1407
d20d654f
MSO
1408 ret->committer = committer_name;
1409 ret->committer_mail = committer_mail;
69cd8f63 1410 get_ac_line(message, "\ncommitter ",
d20d654f
MSO
1411 sizeof(committer_name), committer_name,
1412 sizeof(committer_mail), committer_mail,
cee7f245
JH
1413 &ret->committer_time, &ret->committer_tz);
1414
1415 ret->summary = summary_buf;
69cd8f63 1416 tmp = strstr(message, "\n\n");
cee7f245
JH
1417 if (!tmp) {
1418 error_out:
1419 sprintf(summary_buf, "(%s)", sha1_to_hex(commit->object.sha1));
69cd8f63 1420 free(reencoded);
cee7f245
JH
1421 return;
1422 }
1423 tmp += 2;
1424 endp = strchr(tmp, '\n');
1425 if (!endp)
1cfe7733 1426 endp = tmp + strlen(tmp);
cee7f245 1427 len = endp - tmp;
1cfe7733 1428 if (len >= sizeof(summary_buf) || len == 0)
cee7f245
JH
1429 goto error_out;
1430 memcpy(summary_buf, tmp, len);
1431 summary_buf[len] = 0;
69cd8f63 1432 free(reencoded);
cee7f245
JH
1433}
1434
1732a1fd
JH
1435/*
1436 * To allow LF and other nonportable characters in pathnames,
1437 * they are c-style quoted as needed.
1438 */
46e5e69d
JH
1439static void write_filename_info(const char *path)
1440{
1441 printf("filename ");
663af342 1442 write_name_quoted(path, stdout, '\n');
46e5e69d
JH
1443}
1444
9991030c
JH
1445/*
1446 * Porcelain/Incremental format wants to show a lot of details per
1447 * commit. Instead of repeating this every line, emit it only once,
1448 * the first time each commit appears in the output.
1449 */
1450static int emit_one_suspect_detail(struct origin *suspect)
1451{
1452 struct commit_info ci;
1453
1454 if (suspect->commit->object.flags & METAINFO_SHOWN)
1455 return 0;
1456
1457 suspect->commit->object.flags |= METAINFO_SHOWN;
1458 get_commit_info(suspect->commit, &ci, 1);
1459 printf("author %s\n", ci.author);
1460 printf("author-mail %s\n", ci.author_mail);
1461 printf("author-time %lu\n", ci.author_time);
1462 printf("author-tz %s\n", ci.author_tz);
1463 printf("committer %s\n", ci.committer);
1464 printf("committer-mail %s\n", ci.committer_mail);
1465 printf("committer-time %lu\n", ci.committer_time);
1466 printf("committer-tz %s\n", ci.committer_tz);
1467 printf("summary %s\n", ci.summary);
1468 if (suspect->commit->object.flags & UNINTERESTING)
1469 printf("boundary\n");
96e11709
JH
1470 if (suspect->previous) {
1471 struct origin *prev = suspect->previous;
1472 printf("previous %s ", sha1_to_hex(prev->commit->object.sha1));
1473 write_name_quoted(prev->path, stdout, '\n');
1474 }
9991030c
JH
1475 return 1;
1476}
1477
1732a1fd
JH
1478/*
1479 * The blame_entry is found to be guilty for the range. Mark it
1480 * as such, and show it in incremental output.
1481 */
717d1462
LT
1482static void found_guilty_entry(struct blame_entry *ent)
1483{
1484 if (ent->guilty)
1485 return;
1486 ent->guilty = 1;
1487 if (incremental) {
1488 struct origin *suspect = ent->suspect;
1489
1490 printf("%s %d %d %d\n",
1491 sha1_to_hex(suspect->commit->object.sha1),
1492 ent->s_lno + 1, ent->lno + 1, ent->num_lines);
9991030c 1493 emit_one_suspect_detail(suspect);
46e5e69d 1494 write_filename_info(suspect->path);
06f59e9f 1495 maybe_flush_or_die(stdout, "stdout");
717d1462
LT
1496 }
1497}
1498
1732a1fd
JH
1499/*
1500 * The main loop -- while the scoreboard has lines whose true origin
3dff5379 1501 * is still unknown, pick one blame_entry, and allow its current
1732a1fd
JH
1502 * suspect to pass blames to its parents.
1503 */
85af7929 1504static void assign_blame(struct scoreboard *sb, int opt)
717d1462 1505{
85af7929
JH
1506 struct rev_info *revs = sb->revs;
1507
717d1462
LT
1508 while (1) {
1509 struct blame_entry *ent;
1510 struct commit *commit;
1511 struct origin *suspect = NULL;
1512
1513 /* find one suspect to break down */
1514 for (ent = sb->ent; !suspect && ent; ent = ent->next)
1515 if (!ent->guilty)
1516 suspect = ent->suspect;
1517 if (!suspect)
1518 return; /* all done */
1519
1732a1fd
JH
1520 /*
1521 * We will use this suspect later in the loop,
1522 * so hold onto it in the meantime.
1523 */
717d1462
LT
1524 origin_incref(suspect);
1525 commit = suspect->commit;
1526 if (!commit->object.parsed)
1527 parse_commit(commit);
85af7929
JH
1528 if (reverse ||
1529 (!(commit->object.flags & UNINTERESTING) &&
1530 !(revs->max_age != -1 && commit->date < revs->max_age)))
717d1462
LT
1531 pass_blame(sb, suspect, opt);
1532 else {
1533 commit->object.flags |= UNINTERESTING;
1534 if (commit->object.parsed)
1535 mark_parents_uninteresting(commit);
1536 }
1537 /* treat root commit as boundary */
1538 if (!commit->parents && !show_root)
1539 commit->object.flags |= UNINTERESTING;
1540
1541 /* Take responsibility for the remaining entries */
1542 for (ent = sb->ent; ent; ent = ent->next)
171dccd5 1543 if (same_suspect(ent->suspect, suspect))
717d1462
LT
1544 found_guilty_entry(ent);
1545 origin_decref(suspect);
1546
1547 if (DEBUG) /* sanity */
1548 sanity_check_refcnt(sb);
1549 }
1550}
1551
1552static const char *format_time(unsigned long time, const char *tz_str,
1553 int show_raw_time)
1554{
1555 static char time_buf[128];
31653c1a
EL
1556 const char *time_str;
1557 int time_len;
1558 int tz;
717d1462
LT
1559
1560 if (show_raw_time) {
1561 sprintf(time_buf, "%lu %s", time, tz_str);
717d1462 1562 }
31653c1a
EL
1563 else {
1564 tz = atoi(tz_str);
1565 time_str = show_date(time, tz, blame_date_mode);
1566 time_len = strlen(time_str);
1567 memcpy(time_buf, time_str, time_len);
1568 memset(time_buf + time_len, ' ', blame_date_width - time_len);
1569 }
717d1462
LT
1570 return time_buf;
1571}
1572
cee7f245
JH
1573#define OUTPUT_ANNOTATE_COMPAT 001
1574#define OUTPUT_LONG_OBJECT_NAME 002
1575#define OUTPUT_RAW_TIMESTAMP 004
1576#define OUTPUT_PORCELAIN 010
1577#define OUTPUT_SHOW_NAME 020
1578#define OUTPUT_SHOW_NUMBER 040
5ff62c30 1579#define OUTPUT_SHOW_SCORE 0100
093dc5be 1580#define OUTPUT_NO_AUTHOR 0200
cee7f245
JH
1581
1582static void emit_porcelain(struct scoreboard *sb, struct blame_entry *ent)
1583{
1584 int cnt;
1585 const char *cp;
1586 struct origin *suspect = ent->suspect;
1587 char hex[41];
1588
1589 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1590 printf("%s%c%d %d %d\n",
1591 hex,
2543d9b6 1592 ent->guilty ? ' ' : '*', /* purely for debugging */
cee7f245
JH
1593 ent->s_lno + 1,
1594 ent->lno + 1,
1595 ent->num_lines);
9991030c
JH
1596 if (emit_one_suspect_detail(suspect) ||
1597 (suspect->commit->object.flags & MORE_THAN_ONE_PATH))
46e5e69d 1598 write_filename_info(suspect->path);
cee7f245
JH
1599
1600 cp = nth_line(sb, ent->lno);
1601 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1602 char ch;
1603 if (cnt)
1604 printf("%s %d %d\n", hex,
1605 ent->s_lno + 1 + cnt,
1606 ent->lno + 1 + cnt);
1607 putchar('\t');
1608 do {
1609 ch = *cp++;
1610 putchar(ch);
1611 } while (ch != '\n' &&
1612 cp < sb->final_buf + sb->final_buf_size);
1613 }
a5ca8367
JS
1614
1615 if (sb->final_buf_size && cp[-1] != '\n')
1616 putchar('\n');
cee7f245
JH
1617}
1618
1619static void emit_other(struct scoreboard *sb, struct blame_entry *ent, int opt)
1620{
1621 int cnt;
1622 const char *cp;
1623 struct origin *suspect = ent->suspect;
1624 struct commit_info ci;
1625 char hex[41];
1626 int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP);
1627
1628 get_commit_info(suspect->commit, &ci, 1);
1629 strcpy(hex, sha1_to_hex(suspect->commit->object.sha1));
1630
1631 cp = nth_line(sb, ent->lno);
1632 for (cnt = 0; cnt < ent->num_lines; cnt++) {
1633 char ch;
b11121d9
JH
1634 int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? 40 : 8;
1635
1636 if (suspect->commit->object.flags & UNINTERESTING) {
e68989a7
JH
1637 if (blank_boundary)
1638 memset(hex, ' ', length);
7ceacdff 1639 else if (!(opt & OUTPUT_ANNOTATE_COMPAT)) {
4c10a5ca
JH
1640 length--;
1641 putchar('^');
1642 }
b11121d9 1643 }
cee7f245 1644
b11121d9 1645 printf("%.*s", length, hex);
cee7f245
JH
1646 if (opt & OUTPUT_ANNOTATE_COMPAT)
1647 printf("\t(%10s\t%10s\t%d)", ci.author,
1648 format_time(ci.author_time, ci.author_tz,
1649 show_raw_time),
1650 ent->lno + 1 + cnt);
1651 else {
5ff62c30 1652 if (opt & OUTPUT_SHOW_SCORE)
54a4c617
JH
1653 printf(" %*d %02d",
1654 max_score_digits, ent->score,
1655 ent->suspect->refcnt);
cee7f245
JH
1656 if (opt & OUTPUT_SHOW_NAME)
1657 printf(" %-*.*s", longest_file, longest_file,
1658 suspect->path);
1659 if (opt & OUTPUT_SHOW_NUMBER)
1660 printf(" %*d", max_orig_digits,
1661 ent->s_lno + 1 + cnt);
093dc5be 1662
ffaf9cc0
GT
1663 if (!(opt & OUTPUT_NO_AUTHOR)) {
1664 int pad = longest_author - utf8_strwidth(ci.author);
1665 printf(" (%s%*s %10s",
1666 ci.author, pad, "",
093dc5be
JH
1667 format_time(ci.author_time,
1668 ci.author_tz,
1669 show_raw_time));
ffaf9cc0 1670 }
093dc5be 1671 printf(" %*d) ",
cee7f245
JH
1672 max_digits, ent->lno + 1 + cnt);
1673 }
1674 do {
1675 ch = *cp++;
1676 putchar(ch);
1677 } while (ch != '\n' &&
1678 cp < sb->final_buf + sb->final_buf_size);
1679 }
a5ca8367
JS
1680
1681 if (sb->final_buf_size && cp[-1] != '\n')
1682 putchar('\n');
cee7f245
JH
1683}
1684
1685static void output(struct scoreboard *sb, int option)
1686{
1687 struct blame_entry *ent;
1688
1689 if (option & OUTPUT_PORCELAIN) {
1690 for (ent = sb->ent; ent; ent = ent->next) {
1691 struct blame_entry *oth;
1692 struct origin *suspect = ent->suspect;
1693 struct commit *commit = suspect->commit;
1694 if (commit->object.flags & MORE_THAN_ONE_PATH)
1695 continue;
1696 for (oth = ent->next; oth; oth = oth->next) {
1697 if ((oth->suspect->commit != commit) ||
1698 !strcmp(oth->suspect->path, suspect->path))
1699 continue;
1700 commit->object.flags |= MORE_THAN_ONE_PATH;
1701 break;
1702 }
1703 }
1704 }
1705
1706 for (ent = sb->ent; ent; ent = ent->next) {
1707 if (option & OUTPUT_PORCELAIN)
1708 emit_porcelain(sb, ent);
5ff62c30 1709 else {
cee7f245 1710 emit_other(sb, ent, option);
5ff62c30 1711 }
cee7f245
JH
1712 }
1713}
1714
1732a1fd
JH
1715/*
1716 * To allow quick access to the contents of nth line in the
1717 * final image, prepare an index in the scoreboard.
1718 */
cee7f245
JH
1719static int prepare_lines(struct scoreboard *sb)
1720{
1721 const char *buf = sb->final_buf;
1722 unsigned long len = sb->final_buf_size;
1723 int num = 0, incomplete = 0, bol = 1;
1724
1725 if (len && buf[len-1] != '\n')
1726 incomplete++; /* incomplete line at the end */
1727 while (len--) {
1728 if (bol) {
1729 sb->lineno = xrealloc(sb->lineno,
4b25d091 1730 sizeof(int *) * (num + 1));
cee7f245
JH
1731 sb->lineno[num] = buf - sb->final_buf;
1732 bol = 0;
1733 }
1734 if (*buf++ == '\n') {
1735 num++;
1736 bol = 1;
1737 }
1738 }
1ca6ca87 1739 sb->lineno = xrealloc(sb->lineno,
4b25d091 1740 sizeof(int *) * (num + incomplete + 1));
1ca6ca87 1741 sb->lineno[num + incomplete] = buf - sb->final_buf;
cee7f245
JH
1742 sb->num_lines = num + incomplete;
1743 return sb->num_lines;
1744}
1745
1732a1fd
JH
1746/*
1747 * Add phony grafts for use with -S; this is primarily to
34baebce 1748 * support git's cvsserver that wants to give a linear history
1732a1fd
JH
1749 * to its clients.
1750 */
cee7f245
JH
1751static int read_ancestry(const char *graft_file)
1752{
1753 FILE *fp = fopen(graft_file, "r");
1754 char buf[1024];
1755 if (!fp)
1756 return -1;
1757 while (fgets(buf, sizeof(buf), fp)) {
1758 /* The format is just "Commit Parent1 Parent2 ...\n" */
1759 int len = strlen(buf);
1760 struct commit_graft *graft = read_graft_line(buf, len);
8eaf7986
JH
1761 if (graft)
1762 register_commit_graft(graft, 0);
cee7f245
JH
1763 }
1764 fclose(fp);
1765 return 0;
1766}
1767
1732a1fd
JH
1768/*
1769 * How many columns do we need to show line numbers in decimal?
1770 */
cee7f245
JH
1771static int lineno_width(int lines)
1772{
f2915045 1773 int i, width;
cee7f245 1774
00fb3d21 1775 for (width = 1, i = 10; i <= lines; width++)
f2915045
JH
1776 i *= 10;
1777 return width;
cee7f245
JH
1778}
1779
1732a1fd
JH
1780/*
1781 * How many columns do we need to show line numbers, authors,
1782 * and filenames?
1783 */
cee7f245
JH
1784static void find_alignment(struct scoreboard *sb, int *option)
1785{
1786 int longest_src_lines = 0;
1787 int longest_dst_lines = 0;
5ff62c30 1788 unsigned largest_score = 0;
cee7f245
JH
1789 struct blame_entry *e;
1790
1791 for (e = sb->ent; e; e = e->next) {
1792 struct origin *suspect = e->suspect;
1793 struct commit_info ci;
1794 int num;
1795
ab3bb800
JH
1796 if (strcmp(suspect->path, sb->path))
1797 *option |= OUTPUT_SHOW_NAME;
1798 num = strlen(suspect->path);
1799 if (longest_file < num)
1800 longest_file = num;
cee7f245
JH
1801 if (!(suspect->commit->object.flags & METAINFO_SHOWN)) {
1802 suspect->commit->object.flags |= METAINFO_SHOWN;
1803 get_commit_info(suspect->commit, &ci, 1);
ffaf9cc0 1804 num = utf8_strwidth(ci.author);
cee7f245
JH
1805 if (longest_author < num)
1806 longest_author = num;
1807 }
1808 num = e->s_lno + e->num_lines;
1809 if (longest_src_lines < num)
1810 longest_src_lines = num;
1811 num = e->lno + e->num_lines;
1812 if (longest_dst_lines < num)
1813 longest_dst_lines = num;
5ff62c30
JH
1814 if (largest_score < ent_score(sb, e))
1815 largest_score = ent_score(sb, e);
cee7f245
JH
1816 }
1817 max_orig_digits = lineno_width(longest_src_lines);
1818 max_digits = lineno_width(longest_dst_lines);
5ff62c30 1819 max_score_digits = lineno_width(largest_score);
cee7f245
JH
1820}
1821
1732a1fd
JH
1822/*
1823 * For debugging -- origin is refcounted, and this asserts that
1824 * we do not underflow.
1825 */
54a4c617
JH
1826static void sanity_check_refcnt(struct scoreboard *sb)
1827{
1828 int baa = 0;
1829 struct blame_entry *ent;
1830
1831 for (ent = sb->ent; ent; ent = ent->next) {
ae86ad65 1832 /* Nobody should have zero or negative refcnt */
854b97f6
JH
1833 if (ent->suspect->refcnt <= 0) {
1834 fprintf(stderr, "%s in %s has negative refcnt %d\n",
1835 ent->suspect->path,
1836 sha1_to_hex(ent->suspect->commit->object.sha1),
1837 ent->suspect->refcnt);
ae86ad65 1838 baa = 1;
854b97f6 1839 }
ae86ad65 1840 }
54a4c617
JH
1841 if (baa) {
1842 int opt = 0160;
1843 find_alignment(sb, &opt);
1844 output(sb, opt);
854b97f6 1845 die("Baa %d!", baa);
54a4c617
JH
1846 }
1847}
1848
1732a1fd
JH
1849/*
1850 * Used for the command line parsing; check if the path exists
1851 * in the working tree.
1852 */
c455c87c 1853static int has_string_in_work_tree(const char *path)
cee7f245
JH
1854{
1855 struct stat st;
1856 return !lstat(path, &st);
1857}
1858
4a0fc95f
JH
1859static unsigned parse_score(const char *arg)
1860{
1861 char *end;
1862 unsigned long score = strtoul(arg, &end, 10);
1863 if (*end)
1864 return 0;
1865 return score;
1866}
1867
20239bae
JK
1868static const char *add_prefix(const char *prefix, const char *path)
1869{
097971f5 1870 return prefix_path(prefix, prefix ? strlen(prefix) : 0, path);
20239bae
JK
1871}
1872
1732a1fd
JH
1873/*
1874 * Parsing of (comma separated) one item in the -L option
1875 */
931233bc
JH
1876static const char *parse_loc(const char *spec,
1877 struct scoreboard *sb, long lno,
1878 long begin, long *ret)
1879{
1880 char *term;
1881 const char *line;
1882 long num;
1883 int reg_error;
1884 regex_t regexp;
1885 regmatch_t match[1];
1886
7bd9641d
JH
1887 /* Allow "-L <something>,+20" to mean starting at <something>
1888 * for 20 lines, or "-L <something>,-5" for 5 lines ending at
1889 * <something>.
1890 */
1891 if (1 < begin && (spec[0] == '+' || spec[0] == '-')) {
1892 num = strtol(spec + 1, &term, 10);
1893 if (term != spec + 1) {
1894 if (spec[0] == '-')
1895 num = 0 - num;
1896 if (0 < num)
1897 *ret = begin + num - 2;
1898 else if (!num)
1899 *ret = begin;
1900 else
1901 *ret = begin + num;
1902 return term;
1903 }
1904 return spec;
1905 }
931233bc
JH
1906 num = strtol(spec, &term, 10);
1907 if (term != spec) {
1908 *ret = num;
1909 return term;
1910 }
1911 if (spec[0] != '/')
1912 return spec;
1913
1914 /* it could be a regexp of form /.../ */
4b25d091 1915 for (term = (char *) spec + 1; *term && *term != '/'; term++) {
931233bc
JH
1916 if (*term == '\\')
1917 term++;
1918 }
1919 if (*term != '/')
1920 return spec;
1921
1922 /* try [spec+1 .. term-1] as regexp */
1923 *term = 0;
1924 begin--; /* input is in human terms */
1925 line = nth_line(sb, begin);
1926
1927 if (!(reg_error = regcomp(&regexp, spec + 1, REG_NEWLINE)) &&
1928 !(reg_error = regexec(&regexp, line, 1, match, 0))) {
1929 const char *cp = line + match[0].rm_so;
1930 const char *nline;
1931
1932 while (begin++ < lno) {
1933 nline = nth_line(sb, begin);
1934 if (line <= cp && cp < nline)
1935 break;
1936 line = nline;
1937 }
1938 *ret = begin;
1939 regfree(&regexp);
1940 *term++ = '/';
1941 return term;
1942 }
1943 else {
1944 char errbuf[1024];
1945 regerror(reg_error, &regexp, errbuf, 1024);
1946 die("-L parameter '%s': %s", spec + 1, errbuf);
1947 }
1948}
1949
1732a1fd
JH
1950/*
1951 * Parsing of -L option
1952 */
931233bc
JH
1953static void prepare_blame_range(struct scoreboard *sb,
1954 const char *bottomtop,
1955 long lno,
1956 long *bottom, long *top)
1957{
1958 const char *term;
1959
1960 term = parse_loc(bottomtop, sb, lno, 1, bottom);
1961 if (*term == ',') {
1962 term = parse_loc(term + 1, sb, lno, *bottom + 1, top);
1963 if (*term)
acca687f 1964 usage(blame_usage);
931233bc
JH
1965 }
1966 if (*term)
acca687f 1967 usage(blame_usage);
931233bc
JH
1968}
1969
ef90d6d4 1970static int git_blame_config(const char *var, const char *value, void *cb)
4c10a5ca
JH
1971{
1972 if (!strcmp(var, "blame.showroot")) {
1973 show_root = git_config_bool(var, value);
1974 return 0;
1975 }
1976 if (!strcmp(var, "blame.blankboundary")) {
1977 blank_boundary = git_config_bool(var, value);
1978 return 0;
1979 }
31653c1a
EL
1980 if (!strcmp(var, "blame.date")) {
1981 if (!value)
1982 return config_error_nonbool(var);
1983 blame_date_mode = parse_date_format(value);
1984 return 0;
1985 }
ef90d6d4 1986 return git_default_config(var, value, cb);
4c10a5ca
JH
1987}
1988
f6c07d7d
JH
1989/*
1990 * Prepare a dummy commit that represents the work tree (or staged) item.
1991 * Note that annotating work tree item never works in the reverse.
1992 */
1cfe7733
JH
1993static struct commit *fake_working_tree_commit(const char *path, const char *contents_from)
1994{
1995 struct commit *commit;
1996 struct origin *origin;
1997 unsigned char head_sha1[20];
f285a2d7 1998 struct strbuf buf = STRBUF_INIT;
1cfe7733 1999 const char *ident;
1cfe7733 2000 time_t now;
1cfe7733
JH
2001 int size, len;
2002 struct cache_entry *ce;
2003 unsigned mode;
2004
2005 if (get_sha1("HEAD", head_sha1))
2006 die("No such ref: HEAD");
2007
2008 time(&now);
2009 commit = xcalloc(1, sizeof(*commit));
2010 commit->parents = xcalloc(1, sizeof(*commit->parents));
2011 commit->parents->item = lookup_commit_reference(head_sha1);
2012 commit->object.parsed = 1;
2013 commit->date = now;
2014 commit->object.type = OBJ_COMMIT;
2015
2016 origin = make_origin(commit, path);
2017
2018 if (!contents_from || strcmp("-", contents_from)) {
2019 struct stat st;
2020 const char *read_from;
2021
2022 if (contents_from) {
2023 if (stat(contents_from, &st) < 0)
0721c314 2024 die_errno("Cannot stat '%s'", contents_from);
1cfe7733
JH
2025 read_from = contents_from;
2026 }
2027 else {
2028 if (lstat(path, &st) < 0)
0721c314 2029 die_errno("Cannot lstat '%s'", path);
1cfe7733
JH
2030 read_from = path;
2031 }
1cfe7733
JH
2032 mode = canon_mode(st.st_mode);
2033 switch (st.st_mode & S_IFMT) {
2034 case S_IFREG:
387e7e19 2035 if (strbuf_read_file(&buf, read_from, st.st_size) != st.st_size)
0721c314 2036 die_errno("cannot open or read '%s'", read_from);
1cfe7733
JH
2037 break;
2038 case S_IFLNK:
edfd45d2 2039 if (strbuf_readlink(&buf, read_from, st.st_size) < 0)
0721c314 2040 die_errno("cannot readlink '%s'", read_from);
1cfe7733
JH
2041 break;
2042 default:
2043 die("unsupported file type %s", read_from);
2044 }
2045 }
2046 else {
2047 /* Reading from stdin */
2048 contents_from = "standard input";
1cfe7733 2049 mode = 0;
f1696ee3 2050 if (strbuf_read(&buf, 0, 0) < 0)
d824cbba 2051 die_errno("failed to read from stdin");
1cfe7733 2052 }
21e5ad50 2053 convert_to_git(path, buf.buf, buf.len, &buf, 0);
af6eb822
PH
2054 origin->file.ptr = buf.buf;
2055 origin->file.size = buf.len;
2056 pretend_sha1_file(buf.buf, buf.len, OBJ_BLOB, origin->blob_sha1);
1cfe7733
JH
2057 commit->util = origin;
2058
2059 /*
2060 * Read the current index, replace the path entry with
2061 * origin->blob_sha1 without mucking with its mode or type
2062 * bits; we are not going to write this index out -- we just
2063 * want to run "diff-index --cached".
2064 */
2065 discard_cache();
2066 read_cache();
2067
2068 len = strlen(path);
2069 if (!mode) {
2070 int pos = cache_name_pos(path, len);
2071 if (0 <= pos)
7a51ed66 2072 mode = active_cache[pos]->ce_mode;
1cfe7733
JH
2073 else
2074 /* Let's not bother reading from HEAD tree */
2075 mode = S_IFREG | 0644;
2076 }
2077 size = cache_entry_size(len);
2078 ce = xcalloc(1, size);
2079 hashcpy(ce->sha1, origin->blob_sha1);
2080 memcpy(ce->name, path, len);
2081 ce->ce_flags = create_ce_flags(len, 0);
2082 ce->ce_mode = create_ce_mode(mode);
2083 add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
2084
2085 /*
2086 * We are not going to write this out, so this does not matter
2087 * right now, but someday we might optimize diff-index --cached
2088 * with cache-tree information.
2089 */
2090 cache_tree_invalidate_path(active_cache_tree, path);
2091
2092 commit->buffer = xmalloc(400);
2093 ident = fmt_ident("Not Committed Yet", "not.committed.yet", NULL, 0);
1bb88be9 2094 snprintf(commit->buffer, 400,
1cfe7733
JH
2095 "tree 0000000000000000000000000000000000000000\n"
2096 "parent %s\n"
2097 "author %s\n"
2098 "committer %s\n\n"
2099 "Version of %s from %s\n",
2100 sha1_to_hex(head_sha1),
2101 ident, ident, path, contents_from ? contents_from : path);
2102 return commit;
2103}
2104
85af7929 2105static const char *prepare_final(struct scoreboard *sb)
f6c07d7d
JH
2106{
2107 int i;
2108 const char *final_commit_name = NULL;
85af7929 2109 struct rev_info *revs = sb->revs;
f6c07d7d
JH
2110
2111 /*
2112 * There must be one and only one positive commit in the
2113 * revs->pending array.
2114 */
2115 for (i = 0; i < revs->pending.nr; i++) {
2116 struct object *obj = revs->pending.objects[i].item;
2117 if (obj->flags & UNINTERESTING)
2118 continue;
2119 while (obj->type == OBJ_TAG)
2120 obj = deref_tag(obj, NULL, 0);
2121 if (obj->type != OBJ_COMMIT)
2122 die("Non commit %s?", revs->pending.objects[i].name);
2123 if (sb->final)
2124 die("More than one commit to dig from %s and %s?",
2125 revs->pending.objects[i].name,
2126 final_commit_name);
2127 sb->final = (struct commit *) obj;
2128 final_commit_name = revs->pending.objects[i].name;
2129 }
2130 return final_commit_name;
2131}
2132
85af7929
JH
2133static const char *prepare_initial(struct scoreboard *sb)
2134{
2135 int i;
2136 const char *final_commit_name = NULL;
2137 struct rev_info *revs = sb->revs;
2138
2139 /*
2140 * There must be one and only one negative commit, and it must be
2141 * the boundary.
2142 */
2143 for (i = 0; i < revs->pending.nr; i++) {
2144 struct object *obj = revs->pending.objects[i].item;
2145 if (!(obj->flags & UNINTERESTING))
2146 continue;
2147 while (obj->type == OBJ_TAG)
2148 obj = deref_tag(obj, NULL, 0);
2149 if (obj->type != OBJ_COMMIT)
2150 die("Non commit %s?", revs->pending.objects[i].name);
2151 if (sb->final)
2152 die("More than one commit to dig down to %s and %s?",
2153 revs->pending.objects[i].name,
2154 final_commit_name);
2155 sb->final = (struct commit *) obj;
2156 final_commit_name = revs->pending.objects[i].name;
2157 }
2158 if (!final_commit_name)
2159 die("No commit to dig down to?");
2160 return final_commit_name;
2161}
2162
5817da01
PH
2163static int blame_copy_callback(const struct option *option, const char *arg, int unset)
2164{
2165 int *opt = option->value;
2166
2167 /*
2168 * -C enables copy from removed files;
2169 * -C -C enables copy from existing files, but only
2170 * when blaming a new file;
2171 * -C -C -C enables copy from existing files for
2172 * everybody
2173 */
2174 if (*opt & PICKAXE_BLAME_COPY_HARDER)
2175 *opt |= PICKAXE_BLAME_COPY_HARDEST;
2176 if (*opt & PICKAXE_BLAME_COPY)
2177 *opt |= PICKAXE_BLAME_COPY_HARDER;
2178 *opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE;
2179
2180 if (arg)
2181 blame_copy_score = parse_score(arg);
2182 return 0;
2183}
2184
2185static int blame_move_callback(const struct option *option, const char *arg, int unset)
2186{
2187 int *opt = option->value;
2188
2189 *opt |= PICKAXE_BLAME_MOVE;
2190
2191 if (arg)
2192 blame_move_score = parse_score(arg);
2193 return 0;
2194}
2195
2196static int blame_bottomtop_callback(const struct option *option, const char *arg, int unset)
2197{
2198 const char **bottomtop = option->value;
2199 if (!arg)
2200 return -1;
2201 if (*bottomtop)
2202 die("More than one '-L n,m' option given");
2203 *bottomtop = arg;
2204 return 0;
2205}
2206
acca687f 2207int cmd_blame(int argc, const char **argv, const char *prefix)
cee7f245
JH
2208{
2209 struct rev_info revs;
2210 const char *path;
2211 struct scoreboard sb;
2212 struct origin *o;
2213 struct blame_entry *ent;
3f8d5204 2214 long dashdash_pos, bottom, top, lno;
cee7f245 2215 const char *final_commit_name = NULL;
21666f1a 2216 enum object_type type;
5817da01
PH
2217
2218 static const char *bottomtop = NULL;
2219 static int output_option = 0, opt = 0;
2220 static int show_stats = 0;
2221 static const char *revs_file = NULL;
2222 static const char *contents_from = NULL;
2223 static const struct option options[] = {
2224 OPT_BOOLEAN(0, "incremental", &incremental, "Show blame entries as we find them, incrementally"),
2225 OPT_BOOLEAN('b', NULL, &blank_boundary, "Show blank SHA-1 for boundary commits (Default: off)"),
2226 OPT_BOOLEAN(0, "root", &show_root, "Do not treat root commits as boundaries (Default: off)"),
2227 OPT_BOOLEAN(0, "show-stats", &show_stats, "Show work cost statistics"),
2228 OPT_BIT(0, "score-debug", &output_option, "Show output score for blame entries", OUTPUT_SHOW_SCORE),
2229 OPT_BIT('f', "show-name", &output_option, "Show original filename (Default: auto)", OUTPUT_SHOW_NAME),
2230 OPT_BIT('n', "show-number", &output_option, "Show original linenumber (Default: off)", OUTPUT_SHOW_NUMBER),
2231 OPT_BIT('p', "porcelain", &output_option, "Show in a format designed for machine consumption", OUTPUT_PORCELAIN),
2232 OPT_BIT('c', NULL, &output_option, "Use the same output mode as git-annotate (Default: off)", OUTPUT_ANNOTATE_COMPAT),
2233 OPT_BIT('t', NULL, &output_option, "Show raw timestamp (Default: off)", OUTPUT_RAW_TIMESTAMP),
2234 OPT_BIT('l', NULL, &output_option, "Show long commit SHA1 (Default: off)", OUTPUT_LONG_OBJECT_NAME),
2235 OPT_BIT('s', NULL, &output_option, "Suppress author name and timestamp (Default: off)", OUTPUT_NO_AUTHOR),
2236 OPT_BIT('w', NULL, &xdl_opts, "Ignore whitespace differences", XDF_IGNORE_WHITESPACE),
2237 OPT_STRING('S', NULL, &revs_file, "file", "Use revisions from <file> instead of calling git-rev-list"),
2238 OPT_STRING(0, "contents", &contents_from, "file", "Use <file>'s contents as the final image"),
2239 { OPTION_CALLBACK, 'C', NULL, &opt, "score", "Find line copies within and across files", PARSE_OPT_OPTARG, blame_copy_callback },
2240 { OPTION_CALLBACK, 'M', NULL, &opt, "score", "Find line movements within and across files", PARSE_OPT_OPTARG, blame_move_callback },
2241 OPT_CALLBACK('L', NULL, &bottomtop, "n,m", "Process only line range n,m, counting from 1", blame_bottomtop_callback),
2242 OPT_END()
2243 };
2244
2245 struct parse_opt_ctx_t ctx;
7ceacdff 2246 int cmd_is_annotate = !strcmp(argv[0], "annotate");
e68989a7 2247
ef90d6d4 2248 git_config(git_blame_config, NULL);
5817da01 2249 init_revisions(&revs, NULL);
31653c1a
EL
2250 revs.date_mode = blame_date_mode;
2251
612702e8 2252 save_commit_buffer = 0;
3f8d5204 2253 dashdash_pos = 0;
612702e8 2254
37782920 2255 parse_options_start(&ctx, argc, argv, prefix, PARSE_OPT_KEEP_DASHDASH |
5817da01
PH
2256 PARSE_OPT_KEEP_ARGV0);
2257 for (;;) {
5817da01
PH
2258 switch (parse_options_step(&ctx, options, blame_opt_usage)) {
2259 case PARSE_OPT_HELP:
2260 exit(129);
2261 case PARSE_OPT_DONE:
3f8d5204
PH
2262 if (ctx.argv[0])
2263 dashdash_pos = ctx.cpidx;
5817da01
PH
2264 goto parse_done;
2265 }
2266
2267 if (!strcmp(ctx.argv[0], "--reverse")) {
2268 ctx.argv[0] = "--children";
2269 reverse = 1;
2270 }
6b61ec05 2271 parse_revision_opt(&revs, &ctx, options, blame_opt_usage);
5817da01
PH
2272 }
2273parse_done:
2274 argc = parse_options_end(&ctx);
2275
aa9ea77d 2276 if (revs_file && read_ancestry(revs_file))
d824cbba 2277 die_errno("reading graft file '%s' failed", revs_file);
aa9ea77d 2278
31653c1a 2279 if (cmd_is_annotate) {
7ceacdff 2280 output_option |= OUTPUT_ANNOTATE_COMPAT;
31653c1a
EL
2281 blame_date_mode = DATE_ISO8601;
2282 } else {
2283 blame_date_mode = revs.date_mode;
2284 }
2285
2286 /* The maximum width used to show the dates */
2287 switch (blame_date_mode) {
2288 case DATE_RFC2822:
2289 blame_date_width = sizeof("Thu, 19 Oct 2006 16:00:04 -0700");
2290 break;
2291 case DATE_ISO8601:
2292 blame_date_width = sizeof("2006-10-19 16:00:04 -0700");
2293 break;
2294 case DATE_RAW:
2295 blame_date_width = sizeof("1161298804 -0700");
2296 break;
2297 case DATE_SHORT:
2298 blame_date_width = sizeof("2006-10-19");
2299 break;
2300 case DATE_RELATIVE:
2301 /* "normal" is used as the fallback for "relative" */
2302 case DATE_LOCAL:
2303 case DATE_NORMAL:
2304 blame_date_width = sizeof("Thu Oct 19 16:00:04 2006 -0700");
2305 break;
2306 }
2307 blame_date_width -= 1; /* strip the null */
7ceacdff 2308
b3123f98
JH
2309 if (DIFF_OPT_TST(&revs.diffopt, FIND_COPIES_HARDER))
2310 opt |= (PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE |
2311 PICKAXE_BLAME_COPY_HARDER);
2312
4a0fc95f
JH
2313 if (!blame_move_score)
2314 blame_move_score = BLAME_DEFAULT_MOVE_SCORE;
2315 if (!blame_copy_score)
2316 blame_copy_score = BLAME_DEFAULT_COPY_SCORE;
2317
1732a1fd
JH
2318 /*
2319 * We have collected options unknown to us in argv[1..unk]
cee7f245 2320 * which are to be passed to revision machinery if we are
3dff5379 2321 * going to do the "bottom" processing.
cee7f245
JH
2322 *
2323 * The remaining are:
2324 *
3f8d5204
PH
2325 * (1) if dashdash_pos != 0, its either
2326 * "blame [revisions] -- <path>" or
2327 * "blame -- <path> <rev>"
cee7f245 2328 *
3f8d5204
PH
2329 * (2) otherwise, its one of the two:
2330 * "blame [revisions] <path>"
2331 * "blame <path> <rev>"
cee7f245 2332 *
3f8d5204
PH
2333 * Note that we must strip out <path> from the arguments: we do not
2334 * want the path pruning but we may want "bottom" processing.
cee7f245 2335 */
3f8d5204
PH
2336 if (dashdash_pos) {
2337 switch (argc - dashdash_pos - 1) {
2338 case 2: /* (1b) */
2339 if (argc != 4)
5817da01 2340 usage_with_options(blame_opt_usage, options);
3f8d5204
PH
2341 /* reorder for the new way: <rev> -- <path> */
2342 argv[1] = argv[3];
2343 argv[3] = argv[2];
2344 argv[2] = "--";
2345 /* FALLTHROUGH */
2346 case 1: /* (1a) */
2347 path = add_prefix(prefix, argv[--argc]);
2348 argv[argc] = NULL;
2349 break;
2350 default:
2351 usage_with_options(blame_opt_usage, options);
cee7f245 2352 }
3f8d5204
PH
2353 } else {
2354 if (argc < 2)
5817da01 2355 usage_with_options(blame_opt_usage, options);
3f8d5204 2356 path = add_prefix(prefix, argv[argc - 1]);
c455c87c 2357 if (argc == 3 && !has_string_in_work_tree(path)) { /* (2b) */
3f8d5204
PH
2358 path = add_prefix(prefix, argv[1]);
2359 argv[1] = argv[2];
cee7f245 2360 }
3f8d5204 2361 argv[argc - 1] = "--";
cee7f245 2362
3f8d5204 2363 setup_work_tree();
c455c87c 2364 if (!has_string_in_work_tree(path))
d824cbba 2365 die_errno("cannot stat path '%s'", path);
cee7f245
JH
2366 }
2367
8b3dce56 2368 revs.disable_stdin = 1;
3f8d5204 2369 setup_revisions(argc, argv, &revs, NULL);
cee7f245
JH
2370 memset(&sb, 0, sizeof(sb));
2371
85af7929
JH
2372 sb.revs = &revs;
2373 if (!reverse)
2374 final_commit_name = prepare_final(&sb);
2375 else if (contents_from)
2376 die("--contents and --children do not blend well.");
2377 else
2378 final_commit_name = prepare_initial(&sb);
cee7f245
JH
2379
2380 if (!sb.final) {
1732a1fd
JH
2381 /*
2382 * "--not A B -- path" without anything positive;
1cfe7733
JH
2383 * do not default to HEAD, but use the working tree
2384 * or "--contents".
1732a1fd 2385 */
1981820b 2386 setup_work_tree();
1cfe7733
JH
2387 sb.final = fake_working_tree_commit(path, contents_from);
2388 add_pending_object(&revs, &(sb.final->object), ":");
cee7f245 2389 }
1cfe7733
JH
2390 else if (contents_from)
2391 die("Cannot use --contents with final commit object name");
cee7f245 2392
1732a1fd
JH
2393 /*
2394 * If we have bottom, this will mark the ancestors of the
cee7f245
JH
2395 * bottom commits we would reach while traversing as
2396 * uninteresting.
2397 */
3d51e1b5
MK
2398 if (prepare_revision_walk(&revs))
2399 die("revision walk setup failed");
cee7f245 2400
1cfe7733
JH
2401 if (is_null_sha1(sb.final->object.sha1)) {
2402 char *buf;
2403 o = sb.final->util;
2404 buf = xmalloc(o->file.size + 1);
2405 memcpy(buf, o->file.ptr, o->file.size + 1);
2406 sb.final_buf = buf;
2407 sb.final_buf_size = o->file.size;
2408 }
2409 else {
2410 o = get_origin(&sb, sb.final, path);
2411 if (fill_blob_sha1(o))
2412 die("no such path %s in %s", path, final_commit_name);
cee7f245 2413
21666f1a 2414 sb.final_buf = read_sha1_file(o->blob_sha1, &type,
1cfe7733 2415 &sb.final_buf_size);
ab43e495
JH
2416 if (!sb.final_buf)
2417 die("Cannot read blob %s for path %s",
2418 sha1_to_hex(o->blob_sha1),
2419 path);
1cfe7733 2420 }
c2e525d9 2421 num_read_blob++;
cee7f245
JH
2422 lno = prepare_lines(&sb);
2423
931233bc
JH
2424 bottom = top = 0;
2425 if (bottomtop)
2426 prepare_blame_range(&sb, bottomtop, lno, &bottom, &top);
2427 if (bottom && top && top < bottom) {
2428 long tmp;
2429 tmp = top; top = bottom; bottom = tmp;
2430 }
cee7f245
JH
2431 if (bottom < 1)
2432 bottom = 1;
2433 if (top < 1)
2434 top = lno;
2435 bottom--;
92f9e273 2436 if (lno < top || lno < bottom)
cee7f245
JH
2437 die("file %s has only %lu lines", path, lno);
2438
2439 ent = xcalloc(1, sizeof(*ent));
2440 ent->lno = bottom;
2441 ent->num_lines = top - bottom;
2442 ent->suspect = o;
2443 ent->s_lno = bottom;
2444
2445 sb.ent = ent;
2446 sb.path = path;
2447
d551a488 2448 read_mailmap(&mailmap, NULL);
f9567384 2449
b92565dc
MH
2450 if (!incremental)
2451 setup_pager();
2452
85af7929 2453 assign_blame(&sb, opt);
cee7f245 2454
717d1462
LT
2455 if (incremental)
2456 return 0;
2457
cee7f245
JH
2458 coalesce(&sb);
2459
2460 if (!(output_option & OUTPUT_PORCELAIN))
2461 find_alignment(&sb, &output_option);
2462
2463 output(&sb, output_option);
2464 free((void *)sb.final_buf);
2465 for (ent = sb.ent; ent; ) {
2466 struct blame_entry *e = ent->next;
2467 free(ent);
2468 ent = e;
2469 }
c2e525d9 2470
870b39c1 2471 if (show_stats) {
c2e525d9
JH
2472 printf("num read blob: %d\n", num_read_blob);
2473 printf("num get patch: %d\n", num_get_patch);
2474 printf("num commits: %d\n", num_commits);
2475 }
cee7f245
JH
2476 return 0;
2477}