1 /*
2  * Blame
3  *
4  * Copyright (c) 2006, 2014 by its authors
5  * See COPYING for licensing conditions
6  */
7 
8 #include "cache.h"
9 #include "config.h"
10 #include "color.h"
11 #include "builtin.h"
12 #include "repository.h"
13 #include "commit.h"
14 #include "diff.h"
15 #include "revision.h"
16 #include "quote.h"
17 #include "string-list.h"
18 #include "mailmap.h"
19 #include "parse-options.h"
20 #include "prio-queue.h"
21 #include "utf8.h"
22 #include "userdiff.h"
23 #include "line-range.h"
24 #include "line-log.h"
25 #include "dir.h"
26 #include "progress.h"
27 #include "object-store.h"
28 #include "blame.h"
29 #include "refs.h"
30 #include "tag.h"
31 
32 static char blame_usage[] = N_("git blame [<options>] [<rev-opts>] [<rev>] [--] <file>");
33 
34 static const char *blame_opt_usage[] = {
35 	blame_usage,
36 	"",
37 	N_("<rev-opts> are documented in git-rev-list(1)"),
38 	NULL
39 };
40 
41 static int longest_file;
42 static int longest_author;
43 static int max_orig_digits;
44 static int max_digits;
45 static int max_score_digits;
46 static int show_root;
47 static int reverse;
48 static int blank_boundary;
49 static int incremental;
50 static int xdl_opts;
51 static int abbrev = -1;
52 static int no_whole_file_rename;
53 static int show_progress;
54 static char repeated_meta_color[COLOR_MAXLEN];
55 static int coloring_mode;
56 static struct string_list ignore_revs_file_list = STRING_LIST_INIT_NODUP;
57 static int mark_unblamable_lines;
58 static int mark_ignored_lines;
59 
60 static struct date_mode blame_date_mode = { DATE_ISO8601 };
61 static size_t blame_date_width;
62 
63 static struct string_list mailmap = STRING_LIST_INIT_NODUP;
64 
65 #ifndef DEBUG_BLAME
66 #define DEBUG_BLAME 0
67 #endif
68 
69 static unsigned blame_move_score;
70 static unsigned blame_copy_score;
71 
72 /* Remember to update object flag allocation in object.h */
73 #define METAINFO_SHOWN		(1u<<12)
74 #define MORE_THAN_ONE_PATH	(1u<<13)
75 
76 struct progress_info {
77 	struct progress *progress;
78 	int blamed_lines;
79 };
80 
nth_line_cb(void * data,long lno)81 static const char *nth_line_cb(void *data, long lno)
82 {
83 	return blame_nth_line((struct blame_scoreboard *)data, lno);
84 }
85 
86 /*
87  * Information on commits, used for output.
88  */
89 struct commit_info {
90 	struct strbuf author;
91 	struct strbuf author_mail;
92 	timestamp_t author_time;
93 	struct strbuf author_tz;
94 
95 	/* filled only when asked for details */
96 	struct strbuf committer;
97 	struct strbuf committer_mail;
98 	timestamp_t committer_time;
99 	struct strbuf committer_tz;
100 
101 	struct strbuf summary;
102 };
103 
104 #define COMMIT_INFO_INIT { \
105 	.author = STRBUF_INIT, \
106 	.author_mail = STRBUF_INIT, \
107 	.author_tz = STRBUF_INIT, \
108 	.committer = STRBUF_INIT, \
109 	.committer_mail = STRBUF_INIT, \
110 	.committer_tz = STRBUF_INIT, \
111 	.summary = STRBUF_INIT, \
112 }
113 
114 /*
115  * Parse author/committer line in the commit object buffer
116  */
get_ac_line(const char * inbuf,const char * what,struct strbuf * name,struct strbuf * mail,timestamp_t * time,struct strbuf * tz)117 static void get_ac_line(const char *inbuf, const char *what,
118 	struct strbuf *name, struct strbuf *mail,
119 	timestamp_t *time, struct strbuf *tz)
120 {
121 	struct ident_split ident;
122 	size_t len, maillen, namelen;
123 	char *tmp, *endp;
124 	const char *namebuf, *mailbuf;
125 
126 	tmp = strstr(inbuf, what);
127 	if (!tmp)
128 		goto error_out;
129 	tmp += strlen(what);
130 	endp = strchr(tmp, '\n');
131 	if (!endp)
132 		len = strlen(tmp);
133 	else
134 		len = endp - tmp;
135 
136 	if (split_ident_line(&ident, tmp, len)) {
137 	error_out:
138 		/* Ugh */
139 		tmp = "(unknown)";
140 		strbuf_addstr(name, tmp);
141 		strbuf_addstr(mail, tmp);
142 		strbuf_addstr(tz, tmp);
143 		*time = 0;
144 		return;
145 	}
146 
147 	namelen = ident.name_end - ident.name_begin;
148 	namebuf = ident.name_begin;
149 
150 	maillen = ident.mail_end - ident.mail_begin;
151 	mailbuf = ident.mail_begin;
152 
153 	if (ident.date_begin && ident.date_end)
154 		*time = strtoul(ident.date_begin, NULL, 10);
155 	else
156 		*time = 0;
157 
158 	if (ident.tz_begin && ident.tz_end)
159 		strbuf_add(tz, ident.tz_begin, ident.tz_end - ident.tz_begin);
160 	else
161 		strbuf_addstr(tz, "(unknown)");
162 
163 	/*
164 	 * Now, convert both name and e-mail using mailmap
165 	 */
166 	map_user(&mailmap, &mailbuf, &maillen,
167 		 &namebuf, &namelen);
168 
169 	strbuf_addf(mail, "<%.*s>", (int)maillen, mailbuf);
170 	strbuf_add(name, namebuf, namelen);
171 }
172 
commit_info_destroy(struct commit_info * ci)173 static void commit_info_destroy(struct commit_info *ci)
174 {
175 
176 	strbuf_release(&ci->author);
177 	strbuf_release(&ci->author_mail);
178 	strbuf_release(&ci->author_tz);
179 	strbuf_release(&ci->committer);
180 	strbuf_release(&ci->committer_mail);
181 	strbuf_release(&ci->committer_tz);
182 	strbuf_release(&ci->summary);
183 }
184 
get_commit_info(struct commit * commit,struct commit_info * ret,int detailed)185 static void get_commit_info(struct commit *commit,
186 			    struct commit_info *ret,
187 			    int detailed)
188 {
189 	int len;
190 	const char *subject, *encoding;
191 	const char *message;
192 
193 	encoding = get_log_output_encoding();
194 	message = logmsg_reencode(commit, NULL, encoding);
195 	get_ac_line(message, "\nauthor ",
196 		    &ret->author, &ret->author_mail,
197 		    &ret->author_time, &ret->author_tz);
198 
199 	if (!detailed) {
200 		unuse_commit_buffer(commit, message);
201 		return;
202 	}
203 
204 	get_ac_line(message, "\ncommitter ",
205 		    &ret->committer, &ret->committer_mail,
206 		    &ret->committer_time, &ret->committer_tz);
207 
208 	len = find_commit_subject(message, &subject);
209 	if (len)
210 		strbuf_add(&ret->summary, subject, len);
211 	else
212 		strbuf_addf(&ret->summary, "(%s)", oid_to_hex(&commit->object.oid));
213 
214 	unuse_commit_buffer(commit, message);
215 }
216 
217 /*
218  * Write out any suspect information which depends on the path. This must be
219  * handled separately from emit_one_suspect_detail(), because a given commit
220  * may have changes in multiple paths. So this needs to appear each time
221  * we mention a new group.
222  *
223  * To allow LF and other nonportable characters in pathnames,
224  * they are c-style quoted as needed.
225  */
write_filename_info(struct blame_origin * suspect)226 static void write_filename_info(struct blame_origin *suspect)
227 {
228 	if (suspect->previous) {
229 		struct blame_origin *prev = suspect->previous;
230 		printf("previous %s ", oid_to_hex(&prev->commit->object.oid));
231 		write_name_quoted(prev->path, stdout, '\n');
232 	}
233 	printf("filename ");
234 	write_name_quoted(suspect->path, stdout, '\n');
235 }
236 
237 /*
238  * Porcelain/Incremental format wants to show a lot of details per
239  * commit.  Instead of repeating this every line, emit it only once,
240  * the first time each commit appears in the output (unless the
241  * user has specifically asked for us to repeat).
242  */
emit_one_suspect_detail(struct blame_origin * suspect,int repeat)243 static int emit_one_suspect_detail(struct blame_origin *suspect, int repeat)
244 {
245 	struct commit_info ci = COMMIT_INFO_INIT;
246 
247 	if (!repeat && (suspect->commit->object.flags & METAINFO_SHOWN))
248 		return 0;
249 
250 	suspect->commit->object.flags |= METAINFO_SHOWN;
251 	get_commit_info(suspect->commit, &ci, 1);
252 	printf("author %s\n", ci.author.buf);
253 	printf("author-mail %s\n", ci.author_mail.buf);
254 	printf("author-time %"PRItime"\n", ci.author_time);
255 	printf("author-tz %s\n", ci.author_tz.buf);
256 	printf("committer %s\n", ci.committer.buf);
257 	printf("committer-mail %s\n", ci.committer_mail.buf);
258 	printf("committer-time %"PRItime"\n", ci.committer_time);
259 	printf("committer-tz %s\n", ci.committer_tz.buf);
260 	printf("summary %s\n", ci.summary.buf);
261 	if (suspect->commit->object.flags & UNINTERESTING)
262 		printf("boundary\n");
263 
264 	commit_info_destroy(&ci);
265 
266 	return 1;
267 }
268 
269 /*
270  * The blame_entry is found to be guilty for the range.
271  * Show it in incremental output.
272  */
found_guilty_entry(struct blame_entry * ent,void * data)273 static void found_guilty_entry(struct blame_entry *ent, void *data)
274 {
275 	struct progress_info *pi = (struct progress_info *)data;
276 
277 	if (incremental) {
278 		struct blame_origin *suspect = ent->suspect;
279 
280 		printf("%s %d %d %d\n",
281 		       oid_to_hex(&suspect->commit->object.oid),
282 		       ent->s_lno + 1, ent->lno + 1, ent->num_lines);
283 		emit_one_suspect_detail(suspect, 0);
284 		write_filename_info(suspect);
285 		maybe_flush_or_die(stdout, "stdout");
286 	}
287 	pi->blamed_lines += ent->num_lines;
288 	display_progress(pi->progress, pi->blamed_lines);
289 }
290 
format_time(timestamp_t time,const char * tz_str,int show_raw_time)291 static const char *format_time(timestamp_t time, const char *tz_str,
292 			       int show_raw_time)
293 {
294 	static struct strbuf time_buf = STRBUF_INIT;
295 
296 	strbuf_reset(&time_buf);
297 	if (show_raw_time) {
298 		strbuf_addf(&time_buf, "%"PRItime" %s", time, tz_str);
299 	}
300 	else {
301 		const char *time_str;
302 		size_t time_width;
303 		int tz;
304 		tz = atoi(tz_str);
305 		time_str = show_date(time, tz, &blame_date_mode);
306 		strbuf_addstr(&time_buf, time_str);
307 		/*
308 		 * Add space paddings to time_buf to display a fixed width
309 		 * string, and use time_width for display width calibration.
310 		 */
311 		for (time_width = utf8_strwidth(time_str);
312 		     time_width < blame_date_width;
313 		     time_width++)
314 			strbuf_addch(&time_buf, ' ');
315 	}
316 	return time_buf.buf;
317 }
318 
319 #define OUTPUT_ANNOTATE_COMPAT      (1U<<0)
320 #define OUTPUT_LONG_OBJECT_NAME     (1U<<1)
321 #define OUTPUT_RAW_TIMESTAMP        (1U<<2)
322 #define OUTPUT_PORCELAIN            (1U<<3)
323 #define OUTPUT_SHOW_NAME            (1U<<4)
324 #define OUTPUT_SHOW_NUMBER          (1U<<5)
325 #define OUTPUT_SHOW_SCORE           (1U<<6)
326 #define OUTPUT_NO_AUTHOR            (1U<<7)
327 #define OUTPUT_SHOW_EMAIL           (1U<<8)
328 #define OUTPUT_LINE_PORCELAIN       (1U<<9)
329 #define OUTPUT_COLOR_LINE           (1U<<10)
330 #define OUTPUT_SHOW_AGE_WITH_COLOR  (1U<<11)
331 
emit_porcelain_details(struct blame_origin * suspect,int repeat)332 static void emit_porcelain_details(struct blame_origin *suspect, int repeat)
333 {
334 	if (emit_one_suspect_detail(suspect, repeat) ||
335 	    (suspect->commit->object.flags & MORE_THAN_ONE_PATH))
336 		write_filename_info(suspect);
337 }
338 
emit_porcelain(struct blame_scoreboard * sb,struct blame_entry * ent,int opt)339 static void emit_porcelain(struct blame_scoreboard *sb, struct blame_entry *ent,
340 			   int opt)
341 {
342 	int repeat = opt & OUTPUT_LINE_PORCELAIN;
343 	int cnt;
344 	const char *cp;
345 	struct blame_origin *suspect = ent->suspect;
346 	char hex[GIT_MAX_HEXSZ + 1];
347 
348 	oid_to_hex_r(hex, &suspect->commit->object.oid);
349 	printf("%s %d %d %d\n",
350 	       hex,
351 	       ent->s_lno + 1,
352 	       ent->lno + 1,
353 	       ent->num_lines);
354 	emit_porcelain_details(suspect, repeat);
355 
356 	cp = blame_nth_line(sb, ent->lno);
357 	for (cnt = 0; cnt < ent->num_lines; cnt++) {
358 		char ch;
359 		if (cnt) {
360 			printf("%s %d %d\n", hex,
361 			       ent->s_lno + 1 + cnt,
362 			       ent->lno + 1 + cnt);
363 			if (repeat)
364 				emit_porcelain_details(suspect, 1);
365 		}
366 		putchar('\t');
367 		do {
368 			ch = *cp++;
369 			putchar(ch);
370 		} while (ch != '\n' &&
371 			 cp < sb->final_buf + sb->final_buf_size);
372 	}
373 
374 	if (sb->final_buf_size && cp[-1] != '\n')
375 		putchar('\n');
376 }
377 
378 static struct color_field {
379 	timestamp_t hop;
380 	char col[COLOR_MAXLEN];
381 } *colorfield;
382 static int colorfield_nr, colorfield_alloc;
383 
parse_color_fields(const char * s)384 static void parse_color_fields(const char *s)
385 {
386 	struct string_list l = STRING_LIST_INIT_DUP;
387 	struct string_list_item *item;
388 	enum { EXPECT_DATE, EXPECT_COLOR } next = EXPECT_COLOR;
389 
390 	colorfield_nr = 0;
391 
392 	/* Ideally this would be stripped and split at the same time? */
393 	string_list_split(&l, s, ',', -1);
394 	ALLOC_GROW(colorfield, colorfield_nr + 1, colorfield_alloc);
395 
396 	for_each_string_list_item(item, &l) {
397 		switch (next) {
398 		case EXPECT_DATE:
399 			colorfield[colorfield_nr].hop = approxidate(item->string);
400 			next = EXPECT_COLOR;
401 			colorfield_nr++;
402 			ALLOC_GROW(colorfield, colorfield_nr + 1, colorfield_alloc);
403 			break;
404 		case EXPECT_COLOR:
405 			if (color_parse(item->string, colorfield[colorfield_nr].col))
406 				die(_("expecting a color: %s"), item->string);
407 			next = EXPECT_DATE;
408 			break;
409 		}
410 	}
411 
412 	if (next == EXPECT_COLOR)
413 		die(_("must end with a color"));
414 
415 	colorfield[colorfield_nr].hop = TIME_MAX;
416 	string_list_clear(&l, 0);
417 }
418 
setup_default_color_by_age(void)419 static void setup_default_color_by_age(void)
420 {
421 	parse_color_fields("blue,12 month ago,white,1 month ago,red");
422 }
423 
determine_line_heat(struct commit_info * ci,const char ** dest_color)424 static void determine_line_heat(struct commit_info *ci, const char **dest_color)
425 {
426 	int i = 0;
427 
428 	while (i < colorfield_nr && ci->author_time > colorfield[i].hop)
429 		i++;
430 
431 	*dest_color = colorfield[i].col;
432 }
433 
emit_other(struct blame_scoreboard * sb,struct blame_entry * ent,int opt)434 static void emit_other(struct blame_scoreboard *sb, struct blame_entry *ent, int opt)
435 {
436 	int cnt;
437 	const char *cp;
438 	struct blame_origin *suspect = ent->suspect;
439 	struct commit_info ci = COMMIT_INFO_INIT;
440 	char hex[GIT_MAX_HEXSZ + 1];
441 	int show_raw_time = !!(opt & OUTPUT_RAW_TIMESTAMP);
442 	const char *default_color = NULL, *color = NULL, *reset = NULL;
443 
444 	get_commit_info(suspect->commit, &ci, 1);
445 	oid_to_hex_r(hex, &suspect->commit->object.oid);
446 
447 	cp = blame_nth_line(sb, ent->lno);
448 
449 	if (opt & OUTPUT_SHOW_AGE_WITH_COLOR) {
450 		determine_line_heat(&ci, &default_color);
451 		color = default_color;
452 		reset = GIT_COLOR_RESET;
453 	}
454 
455 	for (cnt = 0; cnt < ent->num_lines; cnt++) {
456 		char ch;
457 		int length = (opt & OUTPUT_LONG_OBJECT_NAME) ? the_hash_algo->hexsz : abbrev;
458 
459 		if (opt & OUTPUT_COLOR_LINE) {
460 			if (cnt > 0) {
461 				color = repeated_meta_color;
462 				reset = GIT_COLOR_RESET;
463 			} else  {
464 				color = default_color ? default_color : NULL;
465 				reset = default_color ? GIT_COLOR_RESET : NULL;
466 			}
467 		}
468 		if (color)
469 			fputs(color, stdout);
470 
471 		if (suspect->commit->object.flags & UNINTERESTING) {
472 			if (blank_boundary)
473 				memset(hex, ' ', length);
474 			else if (!(opt & OUTPUT_ANNOTATE_COMPAT)) {
475 				length--;
476 				putchar('^');
477 			}
478 		}
479 
480 		if (mark_unblamable_lines && ent->unblamable) {
481 			length--;
482 			putchar('*');
483 		}
484 		if (mark_ignored_lines && ent->ignored) {
485 			length--;
486 			putchar('?');
487 		}
488 		printf("%.*s", length, hex);
489 		if (opt & OUTPUT_ANNOTATE_COMPAT) {
490 			const char *name;
491 			if (opt & OUTPUT_SHOW_EMAIL)
492 				name = ci.author_mail.buf;
493 			else
494 				name = ci.author.buf;
495 			printf("\t(%10s\t%10s\t%d)", name,
496 			       format_time(ci.author_time, ci.author_tz.buf,
497 					   show_raw_time),
498 			       ent->lno + 1 + cnt);
499 		} else {
500 			if (opt & OUTPUT_SHOW_SCORE)
501 				printf(" %*d %02d",
502 				       max_score_digits, ent->score,
503 				       ent->suspect->refcnt);
504 			if (opt & OUTPUT_SHOW_NAME)
505 				printf(" %-*.*s", longest_file, longest_file,
506 				       suspect->path);
507 			if (opt & OUTPUT_SHOW_NUMBER)
508 				printf(" %*d", max_orig_digits,
509 				       ent->s_lno + 1 + cnt);
510 
511 			if (!(opt & OUTPUT_NO_AUTHOR)) {
512 				const char *name;
513 				int pad;
514 				if (opt & OUTPUT_SHOW_EMAIL)
515 					name = ci.author_mail.buf;
516 				else
517 					name = ci.author.buf;
518 				pad = longest_author - utf8_strwidth(name);
519 				printf(" (%s%*s %10s",
520 				       name, pad, "",
521 				       format_time(ci.author_time,
522 						   ci.author_tz.buf,
523 						   show_raw_time));
524 			}
525 			printf(" %*d) ",
526 			       max_digits, ent->lno + 1 + cnt);
527 		}
528 		if (reset)
529 			fputs(reset, stdout);
530 		do {
531 			ch = *cp++;
532 			putchar(ch);
533 		} while (ch != '\n' &&
534 			 cp < sb->final_buf + sb->final_buf_size);
535 	}
536 
537 	if (sb->final_buf_size && cp[-1] != '\n')
538 		putchar('\n');
539 
540 	commit_info_destroy(&ci);
541 }
542 
output(struct blame_scoreboard * sb,int option)543 static void output(struct blame_scoreboard *sb, int option)
544 {
545 	struct blame_entry *ent;
546 
547 	if (option & OUTPUT_PORCELAIN) {
548 		for (ent = sb->ent; ent; ent = ent->next) {
549 			int count = 0;
550 			struct blame_origin *suspect;
551 			struct commit *commit = ent->suspect->commit;
552 			if (commit->object.flags & MORE_THAN_ONE_PATH)
553 				continue;
554 			for (suspect = get_blame_suspects(commit); suspect; suspect = suspect->next) {
555 				if (suspect->guilty && count++) {
556 					commit->object.flags |= MORE_THAN_ONE_PATH;
557 					break;
558 				}
559 			}
560 		}
561 	}
562 
563 	for (ent = sb->ent; ent; ent = ent->next) {
564 		if (option & OUTPUT_PORCELAIN)
565 			emit_porcelain(sb, ent, option);
566 		else {
567 			emit_other(sb, ent, option);
568 		}
569 	}
570 }
571 
572 /*
573  * Add phony grafts for use with -S; this is primarily to
574  * support git's cvsserver that wants to give a linear history
575  * to its clients.
576  */
read_ancestry(const char * graft_file)577 static int read_ancestry(const char *graft_file)
578 {
579 	FILE *fp = fopen_or_warn(graft_file, "r");
580 	struct strbuf buf = STRBUF_INIT;
581 	if (!fp)
582 		return -1;
583 	while (!strbuf_getwholeline(&buf, fp, '\n')) {
584 		/* The format is just "Commit Parent1 Parent2 ...\n" */
585 		struct commit_graft *graft = read_graft_line(&buf);
586 		if (graft)
587 			register_commit_graft(the_repository, graft, 0);
588 	}
589 	fclose(fp);
590 	strbuf_release(&buf);
591 	return 0;
592 }
593 
update_auto_abbrev(int auto_abbrev,struct blame_origin * suspect)594 static int update_auto_abbrev(int auto_abbrev, struct blame_origin *suspect)
595 {
596 	const char *uniq = find_unique_abbrev(&suspect->commit->object.oid,
597 					      auto_abbrev);
598 	int len = strlen(uniq);
599 	if (auto_abbrev < len)
600 		return len;
601 	return auto_abbrev;
602 }
603 
604 /*
605  * How many columns do we need to show line numbers, authors,
606  * and filenames?
607  */
find_alignment(struct blame_scoreboard * sb,int * option)608 static void find_alignment(struct blame_scoreboard *sb, int *option)
609 {
610 	int longest_src_lines = 0;
611 	int longest_dst_lines = 0;
612 	unsigned largest_score = 0;
613 	struct blame_entry *e;
614 	int compute_auto_abbrev = (abbrev < 0);
615 	int auto_abbrev = DEFAULT_ABBREV;
616 
617 	for (e = sb->ent; e; e = e->next) {
618 		struct blame_origin *suspect = e->suspect;
619 		int num;
620 
621 		if (compute_auto_abbrev)
622 			auto_abbrev = update_auto_abbrev(auto_abbrev, suspect);
623 		if (strcmp(suspect->path, sb->path))
624 			*option |= OUTPUT_SHOW_NAME;
625 		num = strlen(suspect->path);
626 		if (longest_file < num)
627 			longest_file = num;
628 		if (!(suspect->commit->object.flags & METAINFO_SHOWN)) {
629 			struct commit_info ci = COMMIT_INFO_INIT;
630 			suspect->commit->object.flags |= METAINFO_SHOWN;
631 			get_commit_info(suspect->commit, &ci, 1);
632 			if (*option & OUTPUT_SHOW_EMAIL)
633 				num = utf8_strwidth(ci.author_mail.buf);
634 			else
635 				num = utf8_strwidth(ci.author.buf);
636 			if (longest_author < num)
637 				longest_author = num;
638 			commit_info_destroy(&ci);
639 		}
640 		num = e->s_lno + e->num_lines;
641 		if (longest_src_lines < num)
642 			longest_src_lines = num;
643 		num = e->lno + e->num_lines;
644 		if (longest_dst_lines < num)
645 			longest_dst_lines = num;
646 		if (largest_score < blame_entry_score(sb, e))
647 			largest_score = blame_entry_score(sb, e);
648 	}
649 	max_orig_digits = decimal_width(longest_src_lines);
650 	max_digits = decimal_width(longest_dst_lines);
651 	max_score_digits = decimal_width(largest_score);
652 
653 	if (compute_auto_abbrev)
654 		/* one more abbrev length is needed for the boundary commit */
655 		abbrev = auto_abbrev + 1;
656 }
657 
sanity_check_on_fail(struct blame_scoreboard * sb,int baa)658 static void sanity_check_on_fail(struct blame_scoreboard *sb, int baa)
659 {
660 	int opt = OUTPUT_SHOW_SCORE | OUTPUT_SHOW_NUMBER | OUTPUT_SHOW_NAME;
661 	find_alignment(sb, &opt);
662 	output(sb, opt);
663 	die("Baa %d!", baa);
664 }
665 
parse_score(const char * arg)666 static unsigned parse_score(const char *arg)
667 {
668 	char *end;
669 	unsigned long score = strtoul(arg, &end, 10);
670 	if (*end)
671 		return 0;
672 	return score;
673 }
674 
add_prefix(const char * prefix,const char * path)675 static const char *add_prefix(const char *prefix, const char *path)
676 {
677 	return prefix_path(prefix, prefix ? strlen(prefix) : 0, path);
678 }
679 
git_blame_config(const char * var,const char * value,void * cb)680 static int git_blame_config(const char *var, const char *value, void *cb)
681 {
682 	if (!strcmp(var, "blame.showroot")) {
683 		show_root = git_config_bool(var, value);
684 		return 0;
685 	}
686 	if (!strcmp(var, "blame.blankboundary")) {
687 		blank_boundary = git_config_bool(var, value);
688 		return 0;
689 	}
690 	if (!strcmp(var, "blame.showemail")) {
691 		int *output_option = cb;
692 		if (git_config_bool(var, value))
693 			*output_option |= OUTPUT_SHOW_EMAIL;
694 		else
695 			*output_option &= ~OUTPUT_SHOW_EMAIL;
696 		return 0;
697 	}
698 	if (!strcmp(var, "blame.date")) {
699 		if (!value)
700 			return config_error_nonbool(var);
701 		parse_date_format(value, &blame_date_mode);
702 		return 0;
703 	}
704 	if (!strcmp(var, "blame.ignorerevsfile")) {
705 		const char *str;
706 		int ret;
707 
708 		ret = git_config_pathname(&str, var, value);
709 		if (ret)
710 			return ret;
711 		string_list_insert(&ignore_revs_file_list, str);
712 		return 0;
713 	}
714 	if (!strcmp(var, "blame.markunblamablelines")) {
715 		mark_unblamable_lines = git_config_bool(var, value);
716 		return 0;
717 	}
718 	if (!strcmp(var, "blame.markignoredlines")) {
719 		mark_ignored_lines = git_config_bool(var, value);
720 		return 0;
721 	}
722 	if (!strcmp(var, "color.blame.repeatedlines")) {
723 		if (color_parse_mem(value, strlen(value), repeated_meta_color))
724 			warning(_("invalid color '%s' in color.blame.repeatedLines"),
725 				value);
726 		return 0;
727 	}
728 	if (!strcmp(var, "color.blame.highlightrecent")) {
729 		parse_color_fields(value);
730 		return 0;
731 	}
732 
733 	if (!strcmp(var, "blame.coloring")) {
734 		if (!strcmp(value, "repeatedLines")) {
735 			coloring_mode |= OUTPUT_COLOR_LINE;
736 		} else if (!strcmp(value, "highlightRecent")) {
737 			coloring_mode |= OUTPUT_SHOW_AGE_WITH_COLOR;
738 		} else if (!strcmp(value, "none")) {
739 			coloring_mode &= ~(OUTPUT_COLOR_LINE |
740 					    OUTPUT_SHOW_AGE_WITH_COLOR);
741 		} else {
742 			warning(_("invalid value for blame.coloring"));
743 			return 0;
744 		}
745 	}
746 
747 	if (git_diff_heuristic_config(var, value, cb) < 0)
748 		return -1;
749 	if (userdiff_config(var, value) < 0)
750 		return -1;
751 
752 	return git_default_config(var, value, cb);
753 }
754 
blame_copy_callback(const struct option * option,const char * arg,int unset)755 static int blame_copy_callback(const struct option *option, const char *arg, int unset)
756 {
757 	int *opt = option->value;
758 
759 	BUG_ON_OPT_NEG(unset);
760 
761 	/*
762 	 * -C enables copy from removed files;
763 	 * -C -C enables copy from existing files, but only
764 	 *       when blaming a new file;
765 	 * -C -C -C enables copy from existing files for
766 	 *          everybody
767 	 */
768 	if (*opt & PICKAXE_BLAME_COPY_HARDER)
769 		*opt |= PICKAXE_BLAME_COPY_HARDEST;
770 	if (*opt & PICKAXE_BLAME_COPY)
771 		*opt |= PICKAXE_BLAME_COPY_HARDER;
772 	*opt |= PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE;
773 
774 	if (arg)
775 		blame_copy_score = parse_score(arg);
776 	return 0;
777 }
778 
blame_move_callback(const struct option * option,const char * arg,int unset)779 static int blame_move_callback(const struct option *option, const char *arg, int unset)
780 {
781 	int *opt = option->value;
782 
783 	BUG_ON_OPT_NEG(unset);
784 
785 	*opt |= PICKAXE_BLAME_MOVE;
786 
787 	if (arg)
788 		blame_move_score = parse_score(arg);
789 	return 0;
790 }
791 
is_a_rev(const char * name)792 static int is_a_rev(const char *name)
793 {
794 	struct object_id oid;
795 
796 	if (get_oid(name, &oid))
797 		return 0;
798 	return OBJ_NONE < oid_object_info(the_repository, &oid, NULL);
799 }
800 
peel_to_commit_oid(struct object_id * oid_ret,void * cbdata)801 static int peel_to_commit_oid(struct object_id *oid_ret, void *cbdata)
802 {
803 	struct repository *r = ((struct blame_scoreboard *)cbdata)->repo;
804 	struct object_id oid;
805 
806 	oidcpy(&oid, oid_ret);
807 	while (1) {
808 		struct object *obj;
809 		int kind = oid_object_info(r, &oid, NULL);
810 		if (kind == OBJ_COMMIT) {
811 			oidcpy(oid_ret, &oid);
812 			return 0;
813 		}
814 		if (kind != OBJ_TAG)
815 			return -1;
816 		obj = deref_tag(r, parse_object(r, &oid), NULL, 0);
817 		if (!obj)
818 			return -1;
819 		oidcpy(&oid, &obj->oid);
820 	}
821 }
822 
build_ignorelist(struct blame_scoreboard * sb,struct string_list * ignore_revs_file_list,struct string_list * ignore_rev_list)823 static void build_ignorelist(struct blame_scoreboard *sb,
824 			     struct string_list *ignore_revs_file_list,
825 			     struct string_list *ignore_rev_list)
826 {
827 	struct string_list_item *i;
828 	struct object_id oid;
829 
830 	oidset_init(&sb->ignore_list, 0);
831 	for_each_string_list_item(i, ignore_revs_file_list) {
832 		if (!strcmp(i->string, ""))
833 			oidset_clear(&sb->ignore_list);
834 		else
835 			oidset_parse_file_carefully(&sb->ignore_list, i->string,
836 						    peel_to_commit_oid, sb);
837 	}
838 	for_each_string_list_item(i, ignore_rev_list) {
839 		if (get_oid_committish(i->string, &oid) ||
840 		    peel_to_commit_oid(&oid, sb))
841 			die(_("cannot find revision %s to ignore"), i->string);
842 		oidset_insert(&sb->ignore_list, &oid);
843 	}
844 }
845 
cmd_blame(int argc,const char ** argv,const char * prefix)846 int cmd_blame(int argc, const char **argv, const char *prefix)
847 {
848 	struct rev_info revs;
849 	const char *path;
850 	struct blame_scoreboard sb;
851 	struct blame_origin *o;
852 	struct blame_entry *ent = NULL;
853 	long dashdash_pos, lno;
854 	struct progress_info pi = { NULL, 0 };
855 
856 	struct string_list range_list = STRING_LIST_INIT_NODUP;
857 	struct string_list ignore_rev_list = STRING_LIST_INIT_NODUP;
858 	int output_option = 0, opt = 0;
859 	int show_stats = 0;
860 	const char *revs_file = NULL;
861 	const char *contents_from = NULL;
862 	const struct option options[] = {
863 		OPT_BOOL(0, "incremental", &incremental, N_("show blame entries as we find them, incrementally")),
864 		OPT_BOOL('b', NULL, &blank_boundary, N_("do not show object names of boundary commits (Default: off)")),
865 		OPT_BOOL(0, "root", &show_root, N_("do not treat root commits as boundaries (Default: off)")),
866 		OPT_BOOL(0, "show-stats", &show_stats, N_("show work cost statistics")),
867 		OPT_BOOL(0, "progress", &show_progress, N_("force progress reporting")),
868 		OPT_BIT(0, "score-debug", &output_option, N_("show output score for blame entries"), OUTPUT_SHOW_SCORE),
869 		OPT_BIT('f', "show-name", &output_option, N_("show original filename (Default: auto)"), OUTPUT_SHOW_NAME),
870 		OPT_BIT('n', "show-number", &output_option, N_("show original linenumber (Default: off)"), OUTPUT_SHOW_NUMBER),
871 		OPT_BIT('p', "porcelain", &output_option, N_("show in a format designed for machine consumption"), OUTPUT_PORCELAIN),
872 		OPT_BIT(0, "line-porcelain", &output_option, N_("show porcelain format with per-line commit information"), OUTPUT_PORCELAIN|OUTPUT_LINE_PORCELAIN),
873 		OPT_BIT('c', NULL, &output_option, N_("use the same output mode as git-annotate (Default: off)"), OUTPUT_ANNOTATE_COMPAT),
874 		OPT_BIT('t', NULL, &output_option, N_("show raw timestamp (Default: off)"), OUTPUT_RAW_TIMESTAMP),
875 		OPT_BIT('l', NULL, &output_option, N_("show long commit SHA1 (Default: off)"), OUTPUT_LONG_OBJECT_NAME),
876 		OPT_BIT('s', NULL, &output_option, N_("suppress author name and timestamp (Default: off)"), OUTPUT_NO_AUTHOR),
877 		OPT_BIT('e', "show-email", &output_option, N_("show author email instead of name (Default: off)"), OUTPUT_SHOW_EMAIL),
878 		OPT_BIT('w', NULL, &xdl_opts, N_("ignore whitespace differences"), XDF_IGNORE_WHITESPACE),
879 		OPT_STRING_LIST(0, "ignore-rev", &ignore_rev_list, N_("rev"), N_("ignore <rev> when blaming")),
880 		OPT_STRING_LIST(0, "ignore-revs-file", &ignore_revs_file_list, N_("file"), N_("ignore revisions from <file>")),
881 		OPT_BIT(0, "color-lines", &output_option, N_("color redundant metadata from previous line differently"), OUTPUT_COLOR_LINE),
882 		OPT_BIT(0, "color-by-age", &output_option, N_("color lines by age"), OUTPUT_SHOW_AGE_WITH_COLOR),
883 		OPT_BIT(0, "minimal", &xdl_opts, N_("spend extra cycles to find better match"), XDF_NEED_MINIMAL),
884 		OPT_STRING('S', NULL, &revs_file, N_("file"), N_("use revisions from <file> instead of calling git-rev-list")),
885 		OPT_STRING(0, "contents", &contents_from, N_("file"), N_("use <file>'s contents as the final image")),
886 		OPT_CALLBACK_F('C', NULL, &opt, N_("score"), N_("find line copies within and across files"), PARSE_OPT_OPTARG, blame_copy_callback),
887 		OPT_CALLBACK_F('M', NULL, &opt, N_("score"), N_("find line movements within and across files"), PARSE_OPT_OPTARG, blame_move_callback),
888 		OPT_STRING_LIST('L', NULL, &range_list, N_("range"),
889 				N_("process only line range <start>,<end> or function :<funcname>")),
890 		OPT__ABBREV(&abbrev),
891 		OPT_END()
892 	};
893 
894 	struct parse_opt_ctx_t ctx;
895 	int cmd_is_annotate = !strcmp(argv[0], "annotate");
896 	struct range_set ranges;
897 	unsigned int range_i;
898 	long anchor;
899 	const int hexsz = the_hash_algo->hexsz;
900 
901 	setup_default_color_by_age();
902 	git_config(git_blame_config, &output_option);
903 	repo_init_revisions(the_repository, &revs, NULL);
904 	revs.date_mode = blame_date_mode;
905 	revs.diffopt.flags.allow_textconv = 1;
906 	revs.diffopt.flags.follow_renames = 1;
907 
908 	save_commit_buffer = 0;
909 	dashdash_pos = 0;
910 	show_progress = -1;
911 
912 	parse_options_start(&ctx, argc, argv, prefix, options,
913 			    PARSE_OPT_KEEP_DASHDASH | PARSE_OPT_KEEP_ARGV0);
914 	for (;;) {
915 		switch (parse_options_step(&ctx, options, blame_opt_usage)) {
916 		case PARSE_OPT_NON_OPTION:
917 		case PARSE_OPT_UNKNOWN:
918 			break;
919 		case PARSE_OPT_HELP:
920 		case PARSE_OPT_ERROR:
921 			exit(129);
922 		case PARSE_OPT_COMPLETE:
923 			exit(0);
924 		case PARSE_OPT_DONE:
925 			if (ctx.argv[0])
926 				dashdash_pos = ctx.cpidx;
927 			goto parse_done;
928 		}
929 
930 		if (!strcmp(ctx.argv[0], "--reverse")) {
931 			ctx.argv[0] = "--children";
932 			reverse = 1;
933 		}
934 		parse_revision_opt(&revs, &ctx, options, blame_opt_usage);
935 	}
936 parse_done:
937 	no_whole_file_rename = !revs.diffopt.flags.follow_renames;
938 	xdl_opts |= revs.diffopt.xdl_opts & XDF_INDENT_HEURISTIC;
939 	revs.diffopt.flags.follow_renames = 0;
940 	argc = parse_options_end(&ctx);
941 
942 	if (incremental || (output_option & OUTPUT_PORCELAIN)) {
943 		if (show_progress > 0)
944 			die(_("--progress can't be used with --incremental or porcelain formats"));
945 		show_progress = 0;
946 	} else if (show_progress < 0)
947 		show_progress = isatty(2);
948 
949 	if (0 < abbrev && abbrev < hexsz)
950 		/* one more abbrev length is needed for the boundary commit */
951 		abbrev++;
952 	else if (!abbrev)
953 		abbrev = hexsz;
954 
955 	if (revs_file && read_ancestry(revs_file))
956 		die_errno("reading graft file '%s' failed", revs_file);
957 
958 	if (cmd_is_annotate) {
959 		output_option |= OUTPUT_ANNOTATE_COMPAT;
960 		blame_date_mode.type = DATE_ISO8601;
961 	} else {
962 		blame_date_mode = revs.date_mode;
963 	}
964 
965 	/* The maximum width used to show the dates */
966 	switch (blame_date_mode.type) {
967 	case DATE_RFC2822:
968 		blame_date_width = sizeof("Thu, 19 Oct 2006 16:00:04 -0700");
969 		break;
970 	case DATE_ISO8601_STRICT:
971 		blame_date_width = sizeof("2006-10-19T16:00:04-07:00");
972 		break;
973 	case DATE_ISO8601:
974 		blame_date_width = sizeof("2006-10-19 16:00:04 -0700");
975 		break;
976 	case DATE_RAW:
977 		blame_date_width = sizeof("1161298804 -0700");
978 		break;
979 	case DATE_UNIX:
980 		blame_date_width = sizeof("1161298804");
981 		break;
982 	case DATE_SHORT:
983 		blame_date_width = sizeof("2006-10-19");
984 		break;
985 	case DATE_RELATIVE:
986 		/*
987 		 * TRANSLATORS: This string is used to tell us the
988 		 * maximum display width for a relative timestamp in
989 		 * "git blame" output.  For C locale, "4 years, 11
990 		 * months ago", which takes 22 places, is the longest
991 		 * among various forms of relative timestamps, but
992 		 * your language may need more or fewer display
993 		 * columns.
994 		 */
995 		blame_date_width = utf8_strwidth(_("4 years, 11 months ago")) + 1; /* add the null */
996 		break;
997 	case DATE_HUMAN:
998 		/* If the year is shown, no time is shown */
999 		blame_date_width = sizeof("Thu Oct 19 16:00");
1000 		break;
1001 	case DATE_NORMAL:
1002 		blame_date_width = sizeof("Thu Oct 19 16:00:04 2006 -0700");
1003 		break;
1004 	case DATE_STRFTIME:
1005 		blame_date_width = strlen(show_date(0, 0, &blame_date_mode)) + 1; /* add the null */
1006 		break;
1007 	}
1008 	blame_date_width -= 1; /* strip the null */
1009 
1010 	if (revs.diffopt.flags.find_copies_harder)
1011 		opt |= (PICKAXE_BLAME_COPY | PICKAXE_BLAME_MOVE |
1012 			PICKAXE_BLAME_COPY_HARDER);
1013 
1014 	/*
1015 	 * We have collected options unknown to us in argv[1..unk]
1016 	 * which are to be passed to revision machinery if we are
1017 	 * going to do the "bottom" processing.
1018 	 *
1019 	 * The remaining are:
1020 	 *
1021 	 * (1) if dashdash_pos != 0, it is either
1022 	 *     "blame [revisions] -- <path>" or
1023 	 *     "blame -- <path> <rev>"
1024 	 *
1025 	 * (2) otherwise, it is one of the two:
1026 	 *     "blame [revisions] <path>"
1027 	 *     "blame <path> <rev>"
1028 	 *
1029 	 * Note that we must strip out <path> from the arguments: we do not
1030 	 * want the path pruning but we may want "bottom" processing.
1031 	 */
1032 	if (dashdash_pos) {
1033 		switch (argc - dashdash_pos - 1) {
1034 		case 2: /* (1b) */
1035 			if (argc != 4)
1036 				usage_with_options(blame_opt_usage, options);
1037 			/* reorder for the new way: <rev> -- <path> */
1038 			argv[1] = argv[3];
1039 			argv[3] = argv[2];
1040 			argv[2] = "--";
1041 			/* FALLTHROUGH */
1042 		case 1: /* (1a) */
1043 			path = add_prefix(prefix, argv[--argc]);
1044 			argv[argc] = NULL;
1045 			break;
1046 		default:
1047 			usage_with_options(blame_opt_usage, options);
1048 		}
1049 	} else {
1050 		if (argc < 2)
1051 			usage_with_options(blame_opt_usage, options);
1052 		if (argc == 3 && is_a_rev(argv[argc - 1])) { /* (2b) */
1053 			path = add_prefix(prefix, argv[1]);
1054 			argv[1] = argv[2];
1055 		} else {	/* (2a) */
1056 			if (argc == 2 && is_a_rev(argv[1]) && !get_git_work_tree())
1057 				die("missing <path> to blame");
1058 			path = add_prefix(prefix, argv[argc - 1]);
1059 		}
1060 		argv[argc - 1] = "--";
1061 	}
1062 
1063 	revs.disable_stdin = 1;
1064 	setup_revisions(argc, argv, &revs, NULL);
1065 	if (!revs.pending.nr && is_bare_repository()) {
1066 		struct commit *head_commit;
1067 		struct object_id head_oid;
1068 
1069 		if (!resolve_ref_unsafe("HEAD", RESOLVE_REF_READING,
1070 					&head_oid, NULL) ||
1071 		    !(head_commit = lookup_commit_reference_gently(revs.repo,
1072 							     &head_oid, 1)))
1073 			die("no such ref: HEAD");
1074 
1075 		add_pending_object(&revs, &head_commit->object, "HEAD");
1076 	}
1077 
1078 	init_scoreboard(&sb);
1079 	sb.revs = &revs;
1080 	sb.contents_from = contents_from;
1081 	sb.reverse = reverse;
1082 	sb.repo = the_repository;
1083 	sb.path = path;
1084 	build_ignorelist(&sb, &ignore_revs_file_list, &ignore_rev_list);
1085 	string_list_clear(&ignore_revs_file_list, 0);
1086 	string_list_clear(&ignore_rev_list, 0);
1087 	setup_scoreboard(&sb, &o);
1088 
1089 	/*
1090 	 * Changed-path Bloom filters are disabled when looking
1091 	 * for copies.
1092 	 */
1093 	if (!(opt & PICKAXE_BLAME_COPY))
1094 		setup_blame_bloom_data(&sb);
1095 
1096 	lno = sb.num_lines;
1097 
1098 	if (lno && !range_list.nr)
1099 		string_list_append(&range_list, "1");
1100 
1101 	anchor = 1;
1102 	range_set_init(&ranges, range_list.nr);
1103 	for (range_i = 0; range_i < range_list.nr; ++range_i) {
1104 		long bottom, top;
1105 		if (parse_range_arg(range_list.items[range_i].string,
1106 				    nth_line_cb, &sb, lno, anchor,
1107 				    &bottom, &top, sb.path,
1108 				    the_repository->index))
1109 			usage(blame_usage);
1110 		if ((!lno && (top || bottom)) || lno < bottom)
1111 			die(Q_("file %s has only %lu line",
1112 			       "file %s has only %lu lines",
1113 			       lno), sb.path, lno);
1114 		if (bottom < 1)
1115 			bottom = 1;
1116 		if (top < 1 || lno < top)
1117 			top = lno;
1118 		bottom--;
1119 		range_set_append_unsafe(&ranges, bottom, top);
1120 		anchor = top + 1;
1121 	}
1122 	sort_and_merge_range_set(&ranges);
1123 
1124 	for (range_i = ranges.nr; range_i > 0; --range_i) {
1125 		const struct range *r = &ranges.ranges[range_i - 1];
1126 		ent = blame_entry_prepend(ent, r->start, r->end, o);
1127 	}
1128 
1129 	o->suspects = ent;
1130 	prio_queue_put(&sb.commits, o->commit);
1131 
1132 	blame_origin_decref(o);
1133 
1134 	range_set_release(&ranges);
1135 	string_list_clear(&range_list, 0);
1136 
1137 	sb.ent = NULL;
1138 
1139 	if (blame_move_score)
1140 		sb.move_score = blame_move_score;
1141 	if (blame_copy_score)
1142 		sb.copy_score = blame_copy_score;
1143 
1144 	sb.debug = DEBUG_BLAME;
1145 	sb.on_sanity_fail = &sanity_check_on_fail;
1146 
1147 	sb.show_root = show_root;
1148 	sb.xdl_opts = xdl_opts;
1149 	sb.no_whole_file_rename = no_whole_file_rename;
1150 
1151 	read_mailmap(&mailmap);
1152 
1153 	sb.found_guilty_entry = &found_guilty_entry;
1154 	sb.found_guilty_entry_data = &pi;
1155 	if (show_progress)
1156 		pi.progress = start_delayed_progress(_("Blaming lines"), sb.num_lines);
1157 
1158 	assign_blame(&sb, opt);
1159 
1160 	stop_progress(&pi.progress);
1161 
1162 	if (!incremental)
1163 		setup_pager();
1164 	else
1165 		return 0;
1166 
1167 	blame_sort_final(&sb);
1168 
1169 	blame_coalesce(&sb);
1170 
1171 	if (!(output_option & (OUTPUT_COLOR_LINE | OUTPUT_SHOW_AGE_WITH_COLOR)))
1172 		output_option |= coloring_mode;
1173 
1174 	if (!(output_option & OUTPUT_PORCELAIN)) {
1175 		find_alignment(&sb, &output_option);
1176 		if (!*repeated_meta_color &&
1177 		    (output_option & OUTPUT_COLOR_LINE))
1178 			xsnprintf(repeated_meta_color,
1179 				  sizeof(repeated_meta_color),
1180 				  "%s", GIT_COLOR_CYAN);
1181 	}
1182 	if (output_option & OUTPUT_ANNOTATE_COMPAT)
1183 		output_option &= ~(OUTPUT_COLOR_LINE | OUTPUT_SHOW_AGE_WITH_COLOR);
1184 
1185 	output(&sb, output_option);
1186 	free((void *)sb.final_buf);
1187 	for (ent = sb.ent; ent; ) {
1188 		struct blame_entry *e = ent->next;
1189 		free(ent);
1190 		ent = e;
1191 	}
1192 
1193 	if (show_stats) {
1194 		printf("num read blob: %d\n", sb.num_read_blob);
1195 		printf("num get patch: %d\n", sb.num_get_patch);
1196 		printf("num commits: %d\n", sb.num_commits);
1197 	}
1198 
1199 	cleanup_scoreboard(&sb);
1200 	return 0;
1201 }
1202