handle https:// protocol in git-clone
[git.git] / git-annotate.perl
blob505b5ccb280e16e4279569810ed1f41cf247e570
1 #!/usr/bin/perl
2 # Copyright 2006, Ryan Anderson <ryan@michonline.com>
4 # GPL v2 (See COPYING)
6 # This file is licensed under the GPL v2, or a later version
7 # at the discretion of Linus Torvalds.
9 use warnings;
10 use strict;
11 use Getopt::Long;
12 use POSIX qw(strftime gmtime);
13 use File::Basename qw(basename dirname);
15 sub usage() {
16 print STDERR "Usage: ${\basename $0} [-s] [-S revs-file] file [ revision ]
17 -l, --long
18 Show long rev (Defaults off)
19 -t, --time
20 Show raw timestamp (Defaults off)
21 -r, --rename
22 Follow renames (Defaults on).
23 -S, --rev-file revs-file
24 Use revs from revs-file instead of calling git-rev-list
25 -h, --help
26 This message.
29 exit(1);
32 our ($help, $longrev, $rename, $rawtime, $starting_rev, $rev_file) = (0, 0, 1);
34 my $rc = GetOptions( "long|l" => \$longrev,
35 "time|t" => \$rawtime,
36 "help|h" => \$help,
37 "rename|r" => \$rename,
38 "rev-file|S=s" => \$rev_file);
39 if (!$rc or $help or !@ARGV) {
40 usage();
43 my $filename = shift @ARGV;
44 if (@ARGV) {
45 $starting_rev = shift @ARGV;
48 my @stack = (
50 'rev' => defined $starting_rev ? $starting_rev : "HEAD",
51 'filename' => $filename,
55 our @filelines = ();
57 if (defined $starting_rev) {
58 @filelines = git_cat_file($starting_rev, $filename);
59 } else {
60 open(F,"<",$filename)
61 or die "Failed to open filename: $!";
63 while(<F>) {
64 chomp;
65 push @filelines, $_;
67 close(F);
71 our %revs;
72 our @revqueue;
73 our $head;
75 my $revsprocessed = 0;
76 while (my $bound = pop @stack) {
77 my @revisions = git_rev_list($bound->{'rev'}, $bound->{'filename'});
78 foreach my $revinst (@revisions) {
79 my ($rev, @parents) = @$revinst;
80 $head ||= $rev;
82 if (!defined($rev)) {
83 $rev = "";
85 $revs{$rev}{'filename'} = $bound->{'filename'};
86 if (scalar @parents > 0) {
87 $revs{$rev}{'parents'} = \@parents;
88 next;
91 if (!$rename) {
92 next;
95 my $newbound = find_parent_renames($rev, $bound->{'filename'});
96 if ( exists $newbound->{'filename'} && $newbound->{'filename'} ne $bound->{'filename'}) {
97 push @stack, $newbound;
98 $revs{$rev}{'parents'} = [$newbound->{'rev'}];
102 push @revqueue, $head;
103 init_claim( defined $starting_rev ? $head : 'dirty');
104 unless (defined $starting_rev) {
105 my $diff = open_pipe("git","diff","HEAD", "--",$filename)
106 or die "Failed to call git diff to check for dirty state: $!";
108 _git_diff_parse($diff, [$head], "dirty", (
109 'author' => gitvar_name("GIT_AUTHOR_IDENT"),
110 'author_date' => sprintf("%s +0000",time()),
113 close($diff);
115 handle_rev();
118 my $i = 0;
119 foreach my $l (@filelines) {
120 my ($output, $rev, $committer, $date);
121 if (ref $l eq 'ARRAY') {
122 ($output, $rev, $committer, $date) = @$l;
123 if (!$longrev && length($rev) > 8) {
124 $rev = substr($rev,0,8);
126 } else {
127 $output = $l;
128 ($rev, $committer, $date) = ('unknown', 'unknown', 'unknown');
131 printf("%s\t(%10s\t%10s\t%d)%s\n", $rev, $committer,
132 format_date($date), ++$i, $output);
135 sub init_claim {
136 my ($rev) = @_;
137 for (my $i = 0; $i < @filelines; $i++) {
138 $filelines[$i] = [ $filelines[$i], '', '', '', 1];
139 # line,
140 # rev,
141 # author,
142 # date,
143 # 1 <-- belongs to the original file.
145 $revs{$rev}{'lines'} = \@filelines;
149 sub handle_rev {
150 my $i = 0;
151 my %seen;
152 while (my $rev = shift @revqueue) {
153 next if $seen{$rev}++;
155 my %revinfo = git_commit_info($rev);
157 if (exists $revs{$rev}{parents} &&
158 scalar @{$revs{$rev}{parents}} != 0) {
160 git_diff_parse($revs{$rev}{'parents'}, $rev, %revinfo);
161 push @revqueue, @{$revs{$rev}{'parents'}};
163 } else {
164 # We must be at the initial rev here, so claim everything that is left.
165 for (my $i = 0; $i < @{$revs{$rev}{lines}}; $i++) {
166 if (ref ${$revs{$rev}{lines}}[$i] eq '' || ${$revs{$rev}{lines}}[$i][1] eq '') {
167 claim_line($i, $rev, $revs{$rev}{lines}, %revinfo);
175 sub git_rev_list {
176 my ($rev, $file) = @_;
178 my $revlist;
179 if ($rev_file) {
180 open($revlist, '<' . $rev_file)
181 or die "Failed to open $rev_file : $!";
182 } else {
183 $revlist = open_pipe("git-rev-list","--parents","--remove-empty",$rev,"--",$file)
184 or die "Failed to exec git-rev-list: $!";
187 my @revs;
188 while(my $line = <$revlist>) {
189 chomp $line;
190 my ($rev, @parents) = split /\s+/, $line;
191 push @revs, [ $rev, @parents ];
193 close($revlist);
195 printf("0 revs found for rev %s (%s)\n", $rev, $file) if (@revs == 0);
196 return @revs;
199 sub find_parent_renames {
200 my ($rev, $file) = @_;
202 my $patch = open_pipe("git-diff-tree", "-M50", "-r","--name-status", "-z","$rev")
203 or die "Failed to exec git-diff: $!";
205 local $/ = "\0";
206 my %bound;
207 my $junk = <$patch>;
208 while (my $change = <$patch>) {
209 chomp $change;
210 my $filename = <$patch>;
211 if (!defined $filename) {
212 next;
214 chomp $filename;
216 if ($change =~ m/^[AMD]$/ ) {
217 next;
218 } elsif ($change =~ m/^R/ ) {
219 my $oldfilename = $filename;
220 $filename = <$patch>;
221 chomp $filename;
222 if ( $file eq $filename ) {
223 my $parent = git_find_parent($rev, $oldfilename);
224 @bound{'rev','filename'} = ($parent, $oldfilename);
225 last;
229 close($patch);
231 return \%bound;
235 sub git_find_parent {
236 my ($rev, $filename) = @_;
238 my $revparent = open_pipe("git-rev-list","--remove-empty", "--parents","--max-count=1","$rev","--",$filename)
239 or die "Failed to open git-rev-list to find a single parent: $!";
241 my $parentline = <$revparent>;
242 chomp $parentline;
243 my ($revfound,$parent) = split m/\s+/, $parentline;
245 close($revparent);
247 return $parent;
251 # Get a diff between the current revision and a parent.
252 # Record the commit information that results.
253 sub git_diff_parse {
254 my ($parents, $rev, %revinfo) = @_;
256 my @filenames = ( $revs{$rev}{'filename'} );
257 foreach my $parent (@$parents) {
258 push @filenames, $revs{$parent}{'filename'};
261 my $diff = open_pipe("git-diff-tree","-M","-p","-c",$rev,"--",
262 @filenames )
263 or die "Failed to call git-diff for annotation: $!";
265 _git_diff_parse($diff, $parents, $rev, %revinfo);
267 close($diff);
270 sub _git_diff_parse {
271 my ($diff, $parents, $rev, %revinfo) = @_;
273 my $ri = 0;
275 my $slines = $revs{$rev}{'lines'};
276 my (%plines, %pi);
278 my $gotheader = 0;
279 my ($remstart);
280 my $parent_count = @$parents;
282 my $diff_header_regexp = "^@";
283 $diff_header_regexp .= "@" x @$parents;
284 $diff_header_regexp .= ' -\d+,\d+' x @$parents;
285 $diff_header_regexp .= ' \+(\d+),\d+';
287 my %claim_regexps;
288 my $allparentplus = '^' . '\\+' x @$parents . '(.*)$';
291 my $i = 0;
292 foreach my $parent (@$parents) {
294 $pi{$parent} = 0;
295 my $r = '^' . '.' x @$parents . '(.*)$';
296 my $p = $r;
297 substr($p,$i+1, 1) = '\\+';
299 my $m = $r;
300 substr($m,$i+1, 1) = '-';
302 $claim_regexps{$parent}{plus} = $p;
303 $claim_regexps{$parent}{minus} = $m;
305 $plines{$parent} = [];
307 $i++;
311 DIFF:
312 while(<$diff>) {
313 chomp;
314 if (m/$diff_header_regexp/) {
315 $remstart = $1 - 1;
316 # (0-based arrays)
318 $gotheader = 1;
320 foreach my $parent (@$parents) {
321 for (my $i = $ri; $i < $remstart; $i++) {
322 $plines{$parent}[$pi{$parent}++] = $slines->[$i];
325 $ri = $remstart;
327 next DIFF;
329 } elsif (!$gotheader) {
330 # Skip over the leadin.
331 next DIFF;
334 if (m/^\\/) {
336 # Skip \No newline at end of file.
337 # But this can be internationalized, so only look
338 # for an initial \
340 } else {
341 my %claims = ();
342 my $negclaim = 0;
343 my $allclaimed = 0;
344 my $line;
346 if (m/$allparentplus/) {
347 claim_line($ri, $rev, $slines, %revinfo);
348 $allclaimed = 1;
352 PARENT:
353 foreach my $parent (keys %claim_regexps) {
354 my $m = $claim_regexps{$parent}{minus};
355 my $p = $claim_regexps{$parent}{plus};
357 if (m/$m/) {
358 $line = $1;
359 $plines{$parent}[$pi{$parent}++] = [ $line, '', '', '', 0 ];
360 $negclaim++;
362 } elsif (m/$p/) {
363 $line = $1;
364 if (get_line($slines, $ri) eq $line) {
365 # Found a match, claim
366 $claims{$parent}++;
368 } else {
369 die sprintf("Sync error: %d\n|%s\n|%s\n%s => %s\n",
370 $ri, $line,
371 get_line($slines, $ri),
372 $rev, $parent);
377 if (%claims) {
378 foreach my $parent (@$parents) {
379 next if $claims{$parent} || $allclaimed;
380 $plines{$parent}[$pi{$parent}++] = $slines->[$ri];
381 #[ $line, '', '', '', 0 ];
383 $ri++;
385 } elsif ($negclaim) {
386 next DIFF;
388 } else {
389 if (substr($_,scalar @$parents) ne get_line($slines,$ri) ) {
390 foreach my $parent (@$parents) {
391 printf("parent %s is on line %d\n", $parent, $pi{$parent});
394 die sprintf("Line %d, does not match:\n|%s|\n|%s|\n%s\n",
395 $ri,
396 substr($_,scalar @$parents),
397 get_line($slines,$ri), $rev);
399 foreach my $parent (@$parents) {
400 $plines{$parent}[$pi{$parent}++] = $slines->[$ri];
402 $ri++;
407 for (my $i = $ri; $i < @{$slines} ; $i++) {
408 foreach my $parent (@$parents) {
409 push @{$plines{$parent}}, $slines->[$ri];
411 $ri++;
414 foreach my $parent (@$parents) {
415 $revs{$parent}{lines} = $plines{$parent};
418 return;
421 sub get_line {
422 my ($lines, $index) = @_;
424 return ref $lines->[$index] ne '' ? $lines->[$index][0] : $lines->[$index];
427 sub git_cat_file {
428 my ($rev, $filename) = @_;
429 return () unless defined $rev && defined $filename;
431 my $blob = git_ls_tree($rev, $filename);
432 die "Failed to find a blob for $filename in rev $rev\n" if !defined $blob;
434 my $catfile = open_pipe("git","cat-file", "blob", $blob)
435 or die "Failed to git-cat-file blob $blob (rev $rev, file $filename): " . $!;
437 my @lines;
438 while(<$catfile>) {
439 chomp;
440 push @lines, $_;
442 close($catfile);
444 return @lines;
447 sub git_ls_tree {
448 my ($rev, $filename) = @_;
450 my $lstree = open_pipe("git","ls-tree",$rev,$filename)
451 or die "Failed to call git ls-tree: $!";
453 my ($mode, $type, $blob, $tfilename);
454 while(<$lstree>) {
455 chomp;
456 ($mode, $type, $blob, $tfilename) = split(/\s+/, $_, 4);
457 last if ($tfilename eq $filename);
459 close($lstree);
461 return $blob if ($tfilename eq $filename);
462 die "git-ls-tree failed to find blob for $filename";
468 sub claim_line {
469 my ($floffset, $rev, $lines, %revinfo) = @_;
470 my $oline = get_line($lines, $floffset);
471 @{$lines->[$floffset]} = ( $oline, $rev,
472 $revinfo{'author'}, $revinfo{'author_date'} );
473 #printf("Claiming line %d with rev %s: '%s'\n",
474 # $floffset, $rev, $oline) if 1;
477 sub git_commit_info {
478 my ($rev) = @_;
479 my $commit = open_pipe("git-cat-file", "commit", $rev)
480 or die "Failed to call git-cat-file: $!";
482 my %info;
483 while(<$commit>) {
484 chomp;
485 last if (length $_ == 0);
487 if (m/^author (.*) <(.*)> (.*)$/) {
488 $info{'author'} = $1;
489 $info{'author_email'} = $2;
490 $info{'author_date'} = $3;
491 } elsif (m/^committer (.*) <(.*)> (.*)$/) {
492 $info{'committer'} = $1;
493 $info{'committer_email'} = $2;
494 $info{'committer_date'} = $3;
497 close($commit);
499 return %info;
502 sub format_date {
503 if ($rawtime) {
504 return $_[0];
506 my ($timestamp, $timezone) = split(' ', $_[0]);
507 my $minutes = abs($timezone);
508 $minutes = int($minutes / 100) * 60 + ($minutes % 100);
509 if ($timezone < 0) {
510 $minutes = -$minutes;
512 my $t = $timestamp + $minutes * 60;
513 return strftime("%Y-%m-%d %H:%M:%S " . $timezone, gmtime($t));
516 # Copied from git-send-email.perl - We need a Git.pm module..
517 sub gitvar {
518 my ($var) = @_;
519 my $fh;
520 my $pid = open($fh, '-|');
521 die "$!" unless defined $pid;
522 if (!$pid) {
523 exec('git-var', $var) or die "$!";
525 my ($val) = <$fh>;
526 close $fh or die "$!";
527 chomp($val);
528 return $val;
531 sub gitvar_name {
532 my ($name) = @_;
533 my $val = gitvar($name);
534 my @field = split(/\s+/, $val);
535 return join(' ', @field[0...(@field-4)]);
538 sub open_pipe {
539 if ($^O eq '##INSERT_ACTIVESTATE_STRING_HERE##') {
540 return open_pipe_activestate(@_);
541 } else {
542 return open_pipe_normal(@_);
546 sub open_pipe_activestate {
547 tie *fh, "Git::ActiveStatePipe", @_;
548 return *fh;
551 sub open_pipe_normal {
552 my (@execlist) = @_;
554 my $pid = open my $kid, "-|";
555 defined $pid or die "Cannot fork: $!";
557 unless ($pid) {
558 exec @execlist;
559 die "Cannot exec @execlist: $!";
562 return $kid;
565 package Git::ActiveStatePipe;
566 use strict;
568 sub TIEHANDLE {
569 my ($class, @params) = @_;
570 my $cmdline = join " ", @params;
571 my @data = qx{$cmdline};
572 bless { i => 0, data => \@data }, $class;
575 sub READLINE {
576 my $self = shift;
577 if ($self->{i} >= scalar @{$self->{data}}) {
578 return undef;
580 return $self->{'data'}->[ $self->{i}++ ];
583 sub CLOSE {
584 my $self = shift;
585 delete $self->{data};
586 delete $self->{i};
589 sub EOF {
590 my $self = shift;
591 return ($self->{i} >= scalar @{$self->{data}});