git2svn

   1 #!/usr/bin/perl
   2 #
   3 #    git2svn, converts a git branch to a svn ditto
   4 #    Copyright (C) 2008  Love Hörnquist Åstrand
   5 #
   6 #    This program is free software: you can redistribute it and/or modify
   7 #    it under the terms of the GNU General Public License as published by
   8 #    the Free Software Foundation, either version 3 of the License, or
   9 #    (at your option) any later version.
  10 #
  11 #    This program is distributed in the hope that it will be useful,
  12 #    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 #    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 #    GNU General Public License for more details.
  15 #
  16 #    You should have received a copy of the GNU General Public License
  17 #    along with this program.  If not, see <http://www.gnu.org/licenses/>.
  18
  19 use strict;
  20 use POSIX qw(strftime);
  21 use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
  22 use Pod::Usage;
  23
  24 my $IN;
  25 my $OUT;
  26
  27 my ($help, $verbose, $no_unlink, $no_load);
  28
  29 # svn
  30 my $svntree = "repro";
  31 my $basedir = "trunk";
  32 my $revision = 1;
  33
  34 # git
  35 my $branch = "master";
  36 my $gittree;
  37 my $syncname;
  38 my $masterrev;
  39 my $fexport;
  40
  41
  42 my %blob;
  43 my %paths;
  44
  45 sub read_data
  46 {
  47     my ($IN, $next, $length, $data, $l) = (shift, shift);
  48     unless($next =~ m/^data (\d+)/) { die "missing data: $next" ; }
  49     $length = $1;
  50
  51     $l = read(IN, $data, $length);
  52     unless ($l == $length) { die "failed to read data $l != $length"; }
  53     $data = "" if ($length == 0);
  54     return $data;
  55 }
  56
  57 sub prop
  58 {
  59     my ($key, $value) = (shift, shift);
  60     "K " . length($key) . "\n$key\nV " . length($value) . "\n$value\n";
  61 }
  62
  63 sub parse_svn_tree
  64 {
  65     my $url = shift;
  66     my ($SVN, $type, $name);
  67
  68     open(SVN, "svn ls -R $url|") or die "failed to open svn ls -R $url";
  69     while (<SVN>) {
  70         chomp;
  71         if (m@/?(.*)/$@) {
  72             $type = 1;
  73             $name = "$1";
  74         } else {
  75             $type = 2;
  76             $name = "$_";
  77         }
  78         $paths{$name} = $type;
  79     }
  80     close SVN;
  81
  82     open(SVN, "svn info $url|") or die "failed to open svn info $url";
  83     while (<SVN>) {
  84         chomp;
  85         if (/^Revision: (\d+)/) {
  86             $revision = $1 + 1;
  87             last;
  88         }
  89     }
  90     close SVN;
  91 }
  92
  93 sub find_branch_id
  94 {
  95     my $name = shift;
  96
  97     foreach my $m ("heads", "remotes") {
  98         my $n = "${gittree}/.git/refs/${m}/${name}";
  99         my $ID;
 100         next unless (-f $n);
 101         open FOO, "$n";
 102         my $id = <FOO>;
 103         chomp($id);
 104         close FOO;
 105         return $id;
 106     }
 107     return undef;
 108 }
 109
 110 sub parse_git_tree
 111 {
 112     $masterrev = find_branch_id($branch);
 113
 114     die "No head found for ${branch}" if ($masterrev eq "");
 115
 116     my $syncpoint="${gittree}/.git/refs/tags/${syncname}";
 117
 118     if (-f ${syncpoint}) {
 119         my $oldmasterrev=`cat ${syncpoint}`;
 120         chomp($oldmasterrev);
 121
 122         die "failed to get old parse name" if ($oldmasterrev eq "");
 123
 124         if (${oldmasterrev} eq ${masterrev}) {
 125             print STDERR "nothing to sync\n" if ($verbose);
 126             exit 0;
 127         }
 128
 129         die "no $svntree, but incremental (have synctag) ?\n".
 130             "(\"cd $gittree && git tag -d $syncname\" to remove)"
 131             unless ( -d $svntree);
 132
 133         $fexport = "$oldmasterrev..$masterrev";
 134     } else {
 135         $fexport="${masterrev}";
 136
 137         system("svnadmin create ./$svntree") unless (-d $svntree);
 138     }
 139 }
 140
 141
 142 sub checkdirs
 143 {
 144     my $path = shift;
 145     my $base = "";
 146
 147     # pick first dir, create, take next dir, continue until we reached basename
 148     while ($path =~ m@^([^/]+)/(.*)$@) {
 149         my $first = $base . $1;
 150         $path = $2;
 151         $base = $first . "/";
 152         next if ($paths{$first});
 153
 154         $paths{$first} = 1;
 155
 156         printf OUT "Node-path: $first\n";
 157         printf OUT "Node-kind: dir\n";
 158         printf OUT "Node-action: add\n";
 159         printf OUT "Prop-content-length: 0\n";
 160         printf OUT "Content-length: 0\n";
 161         printf OUT "\n";
 162     }
 163 }
 164
 165 sub next_line
 166 {
 167     my $IN = shift;
 168     my $next = <IN>;
 169     chomp $next;
 170     return $next;
 171 }
 172
 173 $|= 1;
 174
 175 my $result;
 176 $result = GetOptions ("git-branch=s" => \$branch,
 177                       "svn-prefix=s" => \$basedir,
 178                       "no-unlink" => \$no_unlink,
 179                       "no-load" => \$no_load,
 180                       "verbose+" => \$verbose,
 181                       "help" => \$help) or pod2usage(2);
 182
 183 pod2usage(0) if ($help);
 184
 185 die "to few arguments" if ($#ARGV < 1);
 186
 187 mkdir ".data" unless (-d ".data");
 188
 189 die "cant find branch name" unless ($branch =~ m@/?([^/]+)$@);
 190 my $shortbranch = $1;
 191
 192 $syncname = "git2svn-syncpoint-${shortbranch}";
 193
 194 print STDERR "syncname tag: $syncname\n" if ($verbose);
 195
 196 my $gitdump = ".data/git.dump-${shortbranch}";
 197 my $svndump = ".data/svn.dump-${shortbranch}";
 198 my $log = ".data/log-${shortbranch}";
 199
 200 $gittree = $ARGV[0];
 201 $svntree = $ARGV[1];
 202
 203 parse_git_tree($gittree);
 204
 205 my $cwd = `pwd`;
 206 chomp($cwd);
 207 parse_svn_tree("file://" . $cwd ."/". $svntree);
 208
 209 system(">$log");
 210
 211 print STDERR "git fast-export $branch ($fexport)\n" if ($verbose);
 212
 213 system("(cd $gittree && git fast-export $fexport) > $gitdump 2>$log") == 0 or
 214     die "git fast-export: $!";
 215
 216 open IN, "$gitdump" or
 217     die "failed to open $gitdump";
 218
 219 open OUT, ">$svndump" or
 220     die "failed to open $svndump";
 221
 222 print STDERR "creating svn dump from revision $revision...\n" if ($verbose);
 223
 224 print OUT "SVN-fs-dump-format-version: 3\n";
 225
 226 my $next = next_line();
 227 COMMAND: while (!eof(IN)) {
 228     my $mark = undef;
 229     if ($next eq "") {
 230         $next = next_line($IN);
 231         next COMMAND;
 232     } elsif ($next =~ /^commit (.*)/) {
 233
 234         my %commit;
 235
 236         $next = next_line($IN);
 237         if ($next =~ m/mark +(.*)/) {
 238             $mark = $1;
 239             $next = next_line($IN);
 240         }
 241         if ($next =~ m/author +(.*)/) {
 242             $commit{author} = $1;
 243             $next = next_line($IN);
 244         }
 245         unless ($next =~ m/committer +(.+) +<([^>]+)> +(\d+) +[+-](\d+)$/) {
 246             die "missing comitter: $_";
 247         }
 248
 249         $commit{CommitterName} = $1;
 250         $commit{CommitterEmail} = $2;
 251         $commit{CommitterWhen} = $3;
 252         $commit{CommitterTZ} = $4;
 253
 254         $next = next_line($IN);
 255         my $log = read_data($IN, $next);
 256
 257         $next = next_line($IN);
 258         if ($next =~ m/from (.*)/) {
 259             $next = next_line($IN);
 260         }
 261         if ($next =~ m/merge (.*)/) {
 262             $next = next_line($IN);
 263         }
 264
 265         my $date =
 266             strftime("%Y-%m-%dT%H:%M:%S.000000Z",
 267                      gmtime($commit{CommitterWhen}));
 268
 269         my $author = "(no author)";
 270         if ($commit{CommitterEmail} =~ m/([^@]+)/) {
 271             $author = $1;
 272         }
 273         $author = "git2svn-dump" if ($author eq "(no author)");
 274
 275         my $props = "";
 276         $props .= prop("svn:author", $author);
 277         $props .= prop("svn:log", $log);
 278         $props .= prop("svn:date", $date);
 279         $props .= "PROPS-END";
 280
 281         # push out svn info
 282
 283         printf OUT "Revision-number: $revision\n"; $revision++;
 284         printf OUT "Prop-content-length: ". length($props) . "\n";
 285         printf OUT "Content-length: " . length($props) . "\n";
 286         printf OUT "\n";
 287         print OUT "$props\n";
 288
 289         while (1) {
 290             if ($next =~ m/M (\d+) (\S+) (.*)$/) {
 291                 my ($mode, $dataref, $path) = (oct $1, $2, $3);
 292                 my $content;
 293                 if ($dataref eq "inline") {
 294                     $next = next_line($IN);
 295                     $content = read_data($IN, $next);
 296                 } else {
 297                     die "Revision missing content ref $dataref"
 298                         unless(defined $blob{$dataref});
 299
 300                     $content = $blob{$dataref};
 301                     # here we really want to delete $blob{$dataref},
 302                     # but it might be referenced in the future. To
 303                     # avoid keepig everything in memory for larger
 304                     # repositories this must be written out to disk
 305                     # and removed when done.
 306                 }
 307
 308                 $path = "$basedir/$path";
 309                 checkdirs($path);
 310
 311                 my $action = "add";
 312
 313                 if ($paths{$path}) {
 314                     die "file was a dir" if ($paths{$path} != 2);
 315                     $action = "change";
 316                 } else {
 317                     $paths{$path} = 2;
 318                 }
 319
 320
 321                 my $type = $mode & 0777000;
 322                 my $kind = "";
 323                 $kind = "file" if ($type == 0100000);
 324                 $kind = "symlink" if ($type == 0120000);
 325                 die "$type unknown" if ($kind eq "");
 326
 327                 $props = "";
 328                 $props .= prop("svn:executable", "on") if ($mode & 0111);
 329                 $props .= prop("svn:special", "*") if ($kind eq "symlink");
 330                 $props .= "PROPS-END" if ($props ne "");
 331
 332                 $content = "link $content" if ($kind eq "symlink");
 333
 334                 my $plen = length($props);
 335                 my $clen = length($content);
 336
 337                 printf OUT "Node-path: $path\n";
 338                 printf OUT "Node-kind: file\n";
 339                 printf OUT "Node-action: $action\n";
 340                 printf OUT "Text-content-length: $clen\n";
 341                 printf OUT "Content-length: " . ($clen + $plen) . "\n";
 342                 printf OUT "Prop-content-length: $plen\n" if ($plen);
 343                 printf OUT "\n";
 344
 345                 print OUT "$props\n" if ($plen);
 346
 347                 print OUT $content;
 348                 printf OUT "\n";
 349             } elsif ($next =~ m/D (.*)/) {
 350                 my $path = $basedir . "/". $1;
 351
 352                 if ($paths{$path}) {
 353                     delete $paths{$path};
 354
 355                     printf OUT "Node-path: $path\n";
 356                     printf OUT "Node-action: delete\n";
 357                     printf OUT "\n";
 358                 } elsif ($verbose) {
 359                     print STDERR "deleting non existing object: $path\n";
 360                 }
 361
 362             } elsif ($next =~ m/^C (.*)/) {
 363                 die "file copy ?";
 364             } elsif ($next =~ m/^R (.*)/) {
 365                 die "file rename ?";
 366             } elsif ($next =~ m/^filedeleteall$/) {
 367                 die "file delete all ?";
 368             } else {
 369                 next COMMAND;
 370             }
 371             $next = next_line($IN);
 372         }
 373
 374     } elsif ($next =~ /^tag .*/) {
 375     } elsif ($next =~ /^reset .*/) {
 376     } elsif ($next =~ /^blob/) {
 377         $next = next_line($IN);
 378         if ($next =~ m/mark (.*)/) {
 379             $mark = $1;
 380             $next = next_line($IN);
 381         }
 382         my $data = read_data($IN, $next);
 383         $blob{$mark} = $data if (defined $mark);
 384     } elsif ($next =~ /^checkpoint .*/) {
 385     } elsif ($next =~ /^progress (.*)/) {
 386         print STDERR "progress: $1\n" if ($verbose);
 387     } else {
 388         die "unknown command $next";
 389     }
 390     $next = next_line($IN);
 391 }
 392
 393 close IN;
 394 close OUT;
 395
 396 print STDERR "(re-)setting sync-tag to new master\n" if ($verbose);
 397
 398 unless ($no_load) {
 399     system("cd $gittree && ".
 400            "git tag -m \"sync $(date)\" -a -f ${syncname} ${masterrev}");
 401 }
 402
 403 print STDERR "loading dump into svn\n" if ($verbose);
 404
 405 unless ($no_load) {
 406     system("svnadmin load $svntree < $svndump >>$log 2>&1") == 0 or
 407         die "svnadmin load";
 408 }
 409
 410 unlink $svndump, $gitdump, $log unless ($no_unlink);
 411
 412 exit 0;
 413
 414
 415 __END__
 416
 417 =head1 NAME
 418
 419 git2svn - converts a git branch to a svn ditto
 420
 421 =head1 SYNOPSIS
 422
 423 git2svn [options] git-repro svn-repro
 424
 425 =head1 OPTIONS
 426
 427 =over 8
 428
 429 =item B<-git-branch>
 430
 431 The git branch to export. The default is branch is master.
 432
 433 =item B<-svn-prefix>
 434
 435 The svn prefix where the branch is import. The default is trunk to
 436 match the default GIT branch (master).
 437
 438 =item B<-verbose>
 439
 440 More verbose output, can be give more then once to increase the verbosity.
 441
 442 =item B<-help>
 443
 444 Print a brief help message and exits.
 445
 446 =back
 447
 448 =head1 DESCRIPTION
 449
 450 B<This program> will convert a git branch to a svn ditto, it also
 451 support incremantal updates.
 452
 453 git2svn takes a git fast-export dump and converts it into a svn dump
 454 that is feed into svnadmin load.
 455
 456 git2svn assumes its the only process that writes into the svn
 457 repository. This is because of the race between getting the to svn
 458 Revsion number from the svn, creating the dump with correct Revsions,
 459 and do the svnadmin load.
 460
 461 git2svn also support incremental updates from a git branch to a svn
 462 reprositry. Its does this by setting a git tag
 463 (git2svn-syncpoint-<branchname>) where the last update was pulled from.
 464
 465 =head1 EXAMPLES
 466
 467         ./git2svn ~/src/heimdal svn-repro
 468         ./git2svn --git-branch heimdal-1-0-branch \
 469                   --svn-prefix branches/heimdal-1-0-branch \
 470                 ~/src/heimdal svn-repro
 471
 472 =head1 DOWNLOAD
 473
 474 git2svn is avaible from repo.or.cz
 475
 476         git clone git://repo.or.cz/git2svn.git
 477
 478 =head1 AUTHORS
 479
 480         Love Hörnquist Åstrand <lha@kth.se>
 481
 482 =head1 BUGS
 483
 484         Send bug reports to lha@kth.se
 485
 486 =cut