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);
  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 $url";
  69     while (<SVN>) {
  70         if (m@/(.*)/$@) {
  71             $type = 1;
  72             $name = "$basedir/$1";
  73         } else {
  74             $type = 2;
  75             $name = "$basedir/$_";
  76         }
  77         $paths{$name} = $type;
  78     }
  79     close SVN;
  80
  81     open(SVN, "svn info $url|") or die "failed to open svn info $url";
  82     while (<SVN>) {
  83         if (/^Revision: (\d+)/) {
  84             $revision = $1 + 1;
  85             last;
  86         }
  87     }
  88     close SVN;
  89 }
  90
  91 sub parse_git_tree
  92 {
  93     $masterrev= `cat ${gittree}/.git/refs/heads/${branch}`;
  94     chomp($masterrev);
  95
  96     die "No head found for ${branch}" if ($masterrev eq "");
  97
  98     my $syncpoint="${gittree}/.git/refs/tags/${syncname}";
  99
 100     if (-f ${syncpoint}) {
 101         my $oldmasterrev=`cat ${syncpoint}`;
 102         chomp($oldmasterrev);
 103
 104         if (${oldmasterrev} eq ${masterrev}) {
 105             print STDERR "nothing to sync\n" if ($verbose);
 106             exit 0;
 107         }
 108
 109         die "no $svntree, but incremental (have synctag) ?\n".
 110             "(\"cd $gittree && git tag -d $syncname\" to remove)"
 111             unless ( -d $svntree);
 112
 113         $fexport = "$oldmasterrev..$masterrev";
 114     } else {
 115         $fexport="${masterrev}";
 116
 117         system("svnadmin create ./$svntree") unless (-d $svntree);
 118     }
 119 }
 120
 121
 122 sub checkdirs
 123 {
 124     my $path = shift;
 125     my $base = "";
 126
 127     # pick first dir, create, take next dir, continue until we reached basename
 128     while ($path =~ m@^([^/]+)/(.*)$@) {
 129         my $first = $base . $1;
 130         $path = $2;
 131         $base = $first . "/";
 132         next if ($paths{$first});
 133
 134         $paths{$first} = 1;
 135
 136         printf OUT "Node-path: $first\n";
 137         printf OUT "Node-kind: dir\n";
 138         printf OUT "Node-action: add\n";
 139         printf OUT "Prop-content-length: 0\n";
 140         printf OUT "Content-length: 0\n";
 141         printf OUT "\n";
 142     }
 143 }
 144
 145 sub next_line
 146 {
 147     my $IN = shift;
 148     my $next = <IN>;
 149     chomp $next;
 150     return $next;
 151 }
 152
 153 $|= 1;
 154
 155 my $result;
 156 $result = GetOptions ("git-branch=s" => \$branch,
 157                       "svn-prefix=s" => \$basedir,
 158                       "verbose+" => \$verbose,
 159                       "help" => \$help) or pod2usage(2);
 160
 161 pod2usage(0) if ($help);
 162
 163 die "to few arguments" if ($#ARGV < 1);
 164
 165 mkdir ".data" unless (-d ".data");
 166
 167 $syncname = "git2svn-syncpoint-${branch}";
 168
 169 my $gitdump = ".data/git.dump-${branch}";
 170 my $svndump = ".data/svn.dump-${branch}";
 171 my $log = ".data/log-${branch}";
 172
 173 $gittree = $ARGV[0];
 174 $svntree = $ARGV[1];
 175
 176 parse_git_tree($gittree);
 177
 178 my $cwd = `pwd`;
 179 chomp($cwd);
 180 parse_svn_tree("file://" . $cwd ."/". $svntree);
 181
 182 system(">$log");
 183
 184 print STDERR "git fast-export $branch ($fexport)\n" if ($verbose);
 185
 186 system("(cd $gittree && git fast-export $fexport) > $gitdump 2>$log") == 0 or
 187     die "git fast-export: $!";
 188
 189 open IN, "$gitdump" or
 190     die "failed to open $gitdump";
 191
 192 open OUT, ">$svndump" or
 193     die "failed to open $svndump";
 194
 195 print STDERR "creating svn dump from revision $revision...\n" if ($verbose);
 196
 197 print OUT "SVN-fs-dump-format-version: 3\n";
 198
 199 my $next = next_line();
 200 COMMAND: while (!eof(IN)) {
 201     my $mark = undef;
 202     if ($next eq "") {
 203         $next = next_line($IN);
 204         next COMMAND;
 205     } elsif ($next =~ /^commit (.*)/) {
 206
 207         my %commit;
 208
 209         $next = next_line($IN);
 210         if ($next =~ m/mark (.*)/) {
 211             $mark = $1;
 212             $next = next_line($IN);
 213         }
 214         if ($next =~ m/author (.*)/) {
 215             $commit{author} = $1;
 216             $next = next_line($IN);
 217         }
 218         unless ($next =~ m/committer (.+) <([^>]+)> (\d+) \+(\d+)$/) {
 219             die "missing comitter"
 220         }
 221
 222         $commit{CommitterName} = $1;
 223         $commit{CommitterEmail} = $2;
 224         $commit{CommitterWhen} = $3;
 225         $commit{CommitterTZ} = $4;
 226
 227         $next = next_line($IN);
 228         my $log = read_data($IN, $next);
 229
 230         $next = next_line($IN);
 231         if ($next =~ m/from (.*)/) {
 232             $next = next_line($IN);
 233         }
 234         if ($next =~ m/merge (.*)/) {
 235             $next = next_line($IN);
 236         }
 237
 238         my $date =
 239             strftime("%Y-%m-%dT%H:%M:%S.000000Z",
 240                      gmtime($commit{CommitterWhen}));
 241
 242         my $author = "(no author)";
 243         if ($commit{CommitterEmail} =~ m/([^@]+)/) {
 244             $author = $1;
 245         }
 246         $author = "git2svn-dump" if ($author eq "(no author)");
 247
 248         my $props = "";
 249         $props .= prop("svn:author", $author);
 250         $props .= prop("svn:log", $log);
 251         $props .= prop("svn:date", $date);
 252         $props .= "PROPS-END";
 253
 254         # push out svn info
 255
 256         printf OUT "Revision-number: $revision\n"; $revision++;
 257         printf OUT "Prop-content-length: ". length($props) . "\n";
 258         printf OUT "Content-length: " . length($props) . "\n";
 259         printf OUT "\n";
 260         print OUT "$props\n";
 261
 262         while (1) {
 263             if ($next =~ m/M (\d+) (\S+) (.*)$/) {
 264                 my ($mode, $dataref, $path) = (oct $1, $2, $3);
 265                 my $content;
 266                 if ($dataref eq "inline") {
 267                     $next = next_line($IN);
 268                     $content = read_data($IN, $next);
 269                 } else {
 270                     die "Revision missing content ref $dataref"
 271                         unless(defined $blob{$dataref});
 272
 273                     $content = $blob{$dataref};
 274                     # here we really want to delete $blob{$dataref},
 275                     # but it might be referenced in the future. To
 276                     # avoid keepig everything in memory for larger
 277                     # repositories this must be written out to disk
 278                     # and removed when done.
 279                 }
 280
 281                 checkdirs("$basedir/$path");
 282
 283                 my $action = "add";
 284
 285                 if ($paths{$path}) {
 286                     die "file was a dir" if ($paths{$path} != 2);
 287                     $action = "change";
 288                 } else {
 289                     $paths{$path} = 2;
 290                 }
 291
 292
 293                 my $type = $mode & 0777000;
 294                 my $kind = "";
 295                 $kind = "file" if ($type == 0100000);
 296                 $kind = "symlink" if ($type == 0120000);
 297                 die "$type unknown" if ($kind eq "");
 298
 299                 $props = "";
 300                 $props .= prop("svn:executable", "on") if ($mode & 0111);
 301                 $props .= "PROPS-END" if ($props ne "");
 302
 303                 my $plen = length($props);
 304                 my $clen = length($content);
 305
 306                 printf OUT "Node-path: $basedir/$path\n";
 307                 printf OUT "Node-kind: $kind\n";
 308                 printf OUT "Node-action: $action\n";
 309                 printf OUT "Text-content-length: $clen\n";
 310                 printf OUT "Content-length: " . ($clen + $plen) . "\n";
 311                 printf OUT "Prop-content-length: $plen\n" if ($plen);
 312                 printf OUT "\n";
 313
 314                 print OUT "$props\n" if ($plen);
 315
 316                 print OUT $content;
 317                 printf OUT "\n";
 318             } elsif ($next =~ m/D (.*)/) {
 319                 my $path = $1;
 320
 321                 die "deleting non existing object" unless ($paths{$path});
 322
 323                 delete $paths{$path};
 324
 325                 printf OUT "Node-path: $basedir/$path\n";
 326                 printf OUT "Node-action: delete\n";
 327                 printf OUT "\n";
 328
 329             } elsif ($next =~ m/^C (.*)/) {
 330                 die "file copy ?";
 331             } elsif ($next =~ m/^R (.*)/) {
 332                 die "file rename ?";
 333             } elsif ($next =~ m/^filedeleteall$/) {
 334                 die "file delete all ?";
 335             } else {
 336                 next COMMAND;
 337             }
 338             $next = next_line($IN);
 339         }
 340
 341     } elsif ($next =~ /^tag .*/) {
 342     } elsif ($next =~ /^reset .*/) {
 343     } elsif ($next =~ /^blob/) {
 344         $next = next_line($IN);
 345         if ($next =~ m/mark (.*)/) {
 346             $mark = $1;
 347             $next = next_line($IN);
 348         }
 349         my $data = read_data($IN, $next);
 350         $blob{$mark} = $data if (defined $mark);
 351     } elsif ($next =~ /^checkpoint .*/) {
 352     } elsif ($next =~ /^progress (.*)/) {
 353         print STDERR "progress: $1\n" if ($verbose);
 354     } else {
 355         die "unknown command $next";
 356     }
 357     $next = next_line($IN);
 358 }
 359
 360 close IN;
 361 close OUT;
 362
 363 print STDERR "(re-)setting sync-tag to new master\n" if ($verbose);
 364
 365 system("cd $gittree && ".
 366        "git tag -m \"sync $(date)\" -a -f ${syncname} ${masterrev}");
 367
 368 print STDERR "loading dump into svn\n" if ($verbose);
 369
 370 system("svnadmin load $svntree < $svndump >>$log 2>&1") == 0 or
 371     die "svnadmin load";
 372
 373 unlink $svndump, $gitdump, $log;
 374
 375 exit 0;
 376
 377
 378 __END__
 379
 380 =head1 NAME
 381
 382 git2svn - converts a git branch to a svn ditto
 383
 384 =head1 SYNOPSIS
 385
 386 git2svn [options] git-repro svn-repro
 387
 388 =head1 OPTIONS
 389
 390 =over 8
 391
 392 =item B<-git-branch>
 393
 394 The git branch to export. The default is branch is master.
 395
 396 =item B<-svn-prefix>
 397
 398 The svn prefix where the branch is import. The default is trunk to
 399 match the default GIT branch (master).
 400
 401 =item B<-verbose>
 402
 403 More verbose output, can be give more then once to increase the verbosity.
 404
 405 =item B<-help>
 406
 407 Print a brief help message and exits.
 408
 409 =back
 410
 411 =head1 DESCRIPTION
 412
 413 B<This program> will convert a git branch to a svn ditto, it also
 414 support incremantal updates.
 415
 416 git2svn takes a git fast-export dump and converts it into a svn dump
 417 that is feed into svnadmin load.
 418
 419 git2svn assumes its the only process that writes into the svn
 420 repository. This is because of the race between getting the to svn
 421 Revsion number from the svn, creating the dump with correct Revsions,
 422 and do the svnadmin load.
 423
 424 git2svn also support incremental updates from a git branch to a svn
 425 reprositry. Its does this by setting a git tag
 426 (git2svn-syncpoint-<branchname>) where the last update was pulled from.
 427
 428 =head1 EXAMPLES
 429
 430         ./git2svn ~/src/heimdal svn-repro
 431         ./git2svn --git-branch heimdal-1-0-branch \
 432                   --svn-prefix branches/heimdal-1-0-branch \
 433                 ~/src/heimdal svn-repro
 434
 435 =head1 DOWNLOAD
 436
 437 git2svn is avaible from repo.or.cz
 438
 439         git clone git://repo.or.cz/git2svn.git
 440
 441 =cut