git-annotate.perl

   1 #!/usr/bin/perl
   2 # Copyright 2006, Ryan Anderson <ryan@michonline.com>
   3 #
   4 # GPL v2 (See COPYING)
   5 #
   6 # This file is licensed under the GPL v2, or a later version
   7 # at the discretion of Linus Torvalds.
   8
   9 use warnings;
  10 use strict;
  11 use Getopt::Std;
  12 use POSIX qw(strftime gmtime);
  13
  14 sub usage() {
  15         print STDERR 'Usage: ${\basename $0} [-s] [-S revs-file] file
  16
  17         -l              show long rev
  18         -r              follow renames
  19         -S commit       use revs from revs-file instead of calling git-rev-list
  20 ';
  21
  22         exit(1);
  23 }
  24
  25 our ($opt_h, $opt_l, $opt_r, $opt_S);
  26 getopts("hlrS:") or usage();
  27 $opt_h && usage();
  28
  29 my $filename = shift @ARGV;
  30
  31 my @stack = (
  32         {
  33                 'rev' => "HEAD",
  34                 'filename' => $filename,
  35         },
  36 );
  37
  38 our (@lineoffsets, @pendinglineoffsets);
  39 our @filelines = ();
  40 open(F,"<",$filename)
  41         or die "Failed to open filename: $!";
  42
  43 while(<F>) {
  44         chomp;
  45         push @filelines, $_;
  46 }
  47 close(F);
  48 our $leftover_lines = @filelines;
  49 our %revs;
  50 our @revqueue;
  51 our $head;
  52
  53 my $revsprocessed = 0;
  54 while (my $bound = pop @stack) {
  55         my @revisions = git_rev_list($bound->{'rev'}, $bound->{'filename'});
  56         foreach my $revinst (@revisions) {
  57                 my ($rev, @parents) = @$revinst;
  58                 $head ||= $rev;
  59
  60                 if (!defined($rev)) {
  61                         $rev = "";
  62                 }
  63                 $revs{$rev}{'filename'} = $bound->{'filename'};
  64                 if (scalar @parents > 0) {
  65                         $revs{$rev}{'parents'} = \@parents;
  66                         next;
  67                 }
  68
  69                 if (!$opt_r) {
  70                         next;
  71                 }
  72
  73                 my $newbound = find_parent_renames($rev, $bound->{'filename'});
  74                 if ( exists $newbound->{'filename'} && $newbound->{'filename'} ne $bound->{'filename'}) {
  75                         push @stack, $newbound;
  76                         $revs{$rev}{'parents'} = [$newbound->{'rev'}];
  77                 }
  78         }
  79 }
  80 push @revqueue, $head;
  81 init_claim($head);
  82 $revs{$head}{'lineoffsets'} = {};
  83 handle_rev();
  84
  85
  86 my $i = 0;
  87 foreach my $l (@filelines) {
  88         my ($output, $rev, $committer, $date);
  89         if (ref $l eq 'ARRAY') {
  90                 ($output, $rev, $committer, $date) = @$l;
  91                 if (!$opt_l && length($rev) > 8) {
  92                         $rev = substr($rev,0,8);
  93                 }
  94         } else {
  95                 $output = $l;
  96                 ($rev, $committer, $date) = ('unknown', 'unknown', 'unknown');
  97         }
  98
  99         printf("%s\t(%10s\t%10s\t%d)%s\n", $rev, $committer,
 100                 format_date($date), $i++, $output);
 101 }
 102
 103 sub init_claim {
 104         my ($rev) = @_;
 105         my %revinfo = git_commit_info($rev);
 106         for (my $i = 0; $i < @filelines; $i++) {
 107                 $filelines[$i] = [ $filelines[$i], '', '', '', 1];
 108                         # line,
 109                         # rev,
 110                         # author,
 111                         # date,
 112                         # 1 <-- belongs to the original file.
 113         }
 114         $revs{$rev}{'lines'} = \@filelines;
 115 }
 116
 117
 118 sub handle_rev {
 119         my $i = 0;
 120         while (my $rev = shift @revqueue) {
 121
 122                 my %revinfo = git_commit_info($rev);
 123
 124                 foreach my $p (@{$revs{$rev}{'parents'}}) {
 125
 126                         git_diff_parse($p, $rev, %revinfo);
 127                         push @revqueue, $p;
 128                 }
 129
 130
 131                 if (scalar @{$revs{$rev}{parents}} == 0) {
 132                         # We must be at the initial rev here, so claim everything that is left.
 133                         for (my $i = 0; $i < @{$revs{$rev}{lines}}; $i++) {
 134                                 if (ref ${$revs{$rev}{lines}}[$i] eq '' || ${$revs{$rev}{lines}}[$i][1] eq '') {
 135                                         claim_line($i, $rev, $revs{$rev}{lines}, %revinfo);
 136                                 }
 137                         }
 138                 }
 139         }
 140 }
 141
 142
 143 sub git_rev_list {
 144         my ($rev, $file) = @_;
 145
 146         if ($opt_S) {
 147                 open(P, '<' . $opt_S);
 148         } else {
 149                 open(P,"-|","git-rev-list","--parents","--remove-empty",$rev,"--",$file)
 150                         or die "Failed to exec git-rev-list: $!";
 151         }
 152
 153         my @revs;
 154         while(my $line = <P>) {
 155                 chomp $line;
 156                 my ($rev, @parents) = split /\s+/, $line;
 157                 push @revs, [ $rev, @parents ];
 158         }
 159         close(P);
 160
 161         printf("0 revs found for rev %s (%s)\n", $rev, $file) if (@revs == 0);
 162         return @revs;
 163 }
 164
 165 sub find_parent_renames {
 166         my ($rev, $file) = @_;
 167
 168         open(P,"-|","git-diff-tree", "-M50", "-r","--name-status", "-z","$rev")
 169                 or die "Failed to exec git-diff: $!";
 170
 171         local $/ = "\0";
 172         my %bound;
 173         my $junk = <P>;
 174         while (my $change = <P>) {
 175                 chomp $change;
 176                 my $filename = <P>;
 177                 chomp $filename;
 178
 179                 if ($change =~ m/^[AMD]$/ ) {
 180                         next;
 181                 } elsif ($change =~ m/^R/ ) {
 182                         my $oldfilename = $filename;
 183                         $filename = <P>;
 184                         chomp $filename;
 185                         if ( $file eq $filename ) {
 186                                 my $parent = git_find_parent($rev, $oldfilename);
 187                                 @bound{'rev','filename'} = ($parent, $oldfilename);
 188                                 last;
 189                         }
 190                 }
 191         }
 192         close(P);
 193
 194         return \%bound;
 195 }
 196
 197
 198 sub git_find_parent {
 199         my ($rev, $filename) = @_;
 200
 201         open(REVPARENT,"-|","git-rev-list","--remove-empty", "--parents","--max-count=1","$rev","--",$filename)
 202                 or die "Failed to open git-rev-list to find a single parent: $!";
 203
 204         my $parentline = <REVPARENT>;
 205         chomp $parentline;
 206         my ($revfound,$parent) = split m/\s+/, $parentline;
 207
 208         close(REVPARENT);
 209
 210         return $parent;
 211 }
 212
 213
 214 # Get a diff between the current revision and a parent.
 215 # Record the commit information that results.
 216 sub git_diff_parse {
 217         my ($parent, $rev, %revinfo) = @_;
 218
 219         my ($ri, $pi) = (0,0);
 220         open(DIFF,"-|","git-diff-tree","-M","-p",$rev,$parent,"--",
 221                         $revs{$rev}{'filename'}, $revs{$parent}{'filename'})
 222                 or die "Failed to call git-diff for annotation: $!";
 223
 224         my $slines = $revs{$rev}{'lines'};
 225         my @plines;
 226
 227         my $gotheader = 0;
 228         my ($remstart, $remlength, $addstart, $addlength);
 229         my ($hunk_start, $hunk_index, $hunk_adds);
 230         while(<DIFF>) {
 231                 chomp;
 232                 if (m/^@@ -(\d+),(\d+) \+(\d+),(\d+)/) {
 233                         ($remstart, $remlength, $addstart, $addlength) = ($1, $2, $3, $4);
 234                         # Adjust for 0-based arrays
 235                         $remstart--;
 236                         $addstart--;
 237                         # Reinit hunk tracking.
 238                         $hunk_start = $remstart;
 239                         $hunk_index = 0;
 240                         $gotheader = 1;
 241
 242                         for (my $i = $ri; $i < $remstart; $i++) {
 243                                 $plines[$pi++] = $slines->[$i];
 244                                 $ri++;
 245                         }
 246                         next;
 247                 } elsif (!$gotheader) {
 248                         next;
 249                 }
 250
 251                 if (m/^\+(.*)$/) {
 252                         my $line = $1;
 253                         $plines[$pi++] = [ $line, '', '', '', 0 ];
 254                         next;
 255
 256                 } elsif (m/^-(.*)$/) {
 257                         my $line = $1;
 258                         if (get_line($slines, $ri) eq $line) {
 259                                 # Found a match, claim
 260                                 claim_line($ri, $rev, $slines, %revinfo);
 261                         } else {
 262                                 die sprintf("Sync error: %d/%d\n|%s\n|%s\n%s => %s\n",
 263                                                 $ri, $hunk_start + $hunk_index,
 264                                                 $line,
 265                                                 get_line($slines, $ri),
 266                                                 $rev, $parent);
 267                         }
 268                         $ri++;
 269
 270                 } else {
 271                         if (substr($_,1) ne get_line($slines,$ri) ) {
 272                                 die sprintf("Line %d (%d) does not match:\n|%s\n|%s\n%s => %s\n",
 273                                                 $hunk_start + $hunk_index, $ri,
 274                                                 substr($_,1),
 275                                                 get_line($slines,$ri),
 276                                                 $rev, $parent);
 277                         }
 278                         $plines[$pi++] = $slines->[$ri++];
 279                 }
 280                 $hunk_index++;
 281         }
 282         close(DIFF);
 283         for (my $i = $ri; $i < @{$slines} ; $i++) {
 284                 push @plines, $slines->[$ri++];
 285         }
 286
 287         $revs{$parent}{lines} = \@plines;
 288         return;
 289 }
 290
 291 sub get_line {
 292         my ($lines, $index) = @_;
 293
 294         return ref $lines->[$index] ne '' ? $lines->[$index][0] : $lines->[$index];
 295 }
 296
 297 sub git_cat_file {
 298         my ($parent, $filename) = @_;
 299         return () unless defined $parent && defined $filename;
 300         my $blobline = `git-ls-tree $parent $filename`;
 301         my ($mode, $type, $blob, $tfilename) = split(/\s+/, $blobline, 4);
 302
 303         open(C,"-|","git-cat-file", "blob", $blob)
 304                 or die "Failed to git-cat-file blob $blob (rev $parent, file $filename): " . $!;
 305
 306         my @lines;
 307         while(<C>) {
 308                 chomp;
 309                 push @lines, $_;
 310         }
 311         close(C);
 312
 313         return @lines;
 314 }
 315
 316
 317 sub claim_line {
 318         my ($floffset, $rev, $lines, %revinfo) = @_;
 319         my $oline = get_line($lines, $floffset);
 320         @{$lines->[$floffset]} = ( $oline, $rev,
 321                 $revinfo{'author'}, $revinfo{'author_date'} );
 322         #printf("Claiming line %d with rev %s: '%s'\n",
 323         #               $floffset, $rev, $oline) if 1;
 324 }
 325
 326 sub git_commit_info {
 327         my ($rev) = @_;
 328         open(COMMIT, "-|","git-cat-file", "commit", $rev)
 329                 or die "Failed to call git-cat-file: $!";
 330
 331         my %info;
 332         while(<COMMIT>) {
 333                 chomp;
 334                 last if (length $_ == 0);
 335
 336                 if (m/^author (.*) <(.*)> (.*)$/) {
 337                         $info{'author'} = $1;
 338                         $info{'author_email'} = $2;
 339                         $info{'author_date'} = $3;
 340                 } elsif (m/^committer (.*) <(.*)> (.*)$/) {
 341                         $info{'committer'} = $1;
 342                         $info{'committer_email'} = $2;
 343                         $info{'committer_date'} = $3;
 344                 }
 345         }
 346         close(COMMIT);
 347
 348         return %info;
 349 }
 350
 351 sub format_date {
 352         my ($timestamp, $timezone) = split(' ', $_[0]);
 353
 354         return strftime("%Y-%m-%d %H:%M:%S " . $timezone, gmtime($timestamp));
 355 }
 356