git2svn

   1 #!/usr/bin/perl
   2 #
   3 # git2svn, converts a git branch to a svn ditto
   4 #
   5 # Copyright (c) 2008 Love Hörnquist Åstrand
   6 # All rights reserved.
   7 #
   8 # Redistribution and use in source and binary forms, with or without
   9 # modification, are permitted provided that the following conditions
  10 # are met:
  11 #
  12 # 1. Redistributions of source code must retain the above copyright
  13 #    notice, this list of conditions and the following disclaimer.
  14 #
  15 # 2. Redistributions in binary form must reproduce the above copyright
  16 #    notice, this list of conditions and the following disclaimer in the
  17 #    documentation and/or other materials provided with the distribution.
  18 #
  19 # THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  20 # ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  21 # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  22 # ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  23 # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  24 # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  25 # OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  26 # HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  27 # LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  28 # OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  29 # SUCH DAMAGE.
  30
  31 use strict;
  32 use POSIX qw(strftime);
  33 use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
  34 use Pod::Usage;
  35
  36 my $IN;
  37 my $OUT;
  38
  39 my ($help, $verbose, $keeplogs, $no_load);
  40
  41 # svn
  42 my $svntree = "repro";
  43 my $basedir = "trunk";
  44 my $revision = 1;
  45
  46 # git
  47 my $branch = "master";
  48 my $syncname;
  49 my $masterrev;
  50 my $fexport;
  51
  52 my %blob;
  53 my %paths;
  54
  55 sub read_data
  56 {
  57     my ($IN, $next, $length, $data, $l) = (shift, shift);
  58     unless($next =~ m/^data (\d+)/) { die "missing data: $next" ; }
  59     $length = $1;
  60
  61     $l = read(IN, $data, $length);
  62     unless ($l == $length) { die "failed to read data $l != $length"; }
  63     $data = "" if ($length == 0);
  64     return $data;
  65 }
  66
  67 sub prop
  68 {
  69     my ($key, $value) = (shift, shift);
  70     "K " . length($key) . "\n$key\nV " . length($value) . "\n$value\n";
  71 }
  72
  73 sub parse_svn_tree
  74 {
  75     my $url = shift;
  76     my ($SVN, $type, $name);
  77
  78     open(SVN, "svn ls -R $url|") or die "failed to open svn ls -R $url";
  79     while (<SVN>) {
  80         chomp;
  81         if (m@/?(.*)/$@) {
  82             $type = 1;
  83             $name = "$1";
  84         } else {
  85             $type = 2;
  86             $name = "$_";
  87         }
  88         $paths{$name} = $type;
  89     }
  90     close SVN;
  91
  92     open(SVN, "svn info $url|") or die "failed to open svn info $url";
  93     while (<SVN>) {
  94         chomp;
  95         if ($revision eq 1 && /^Revision: (\d+)/) {
  96             $revision = $1 + 1;
  97         }
  98     }
  99     close SVN;
 100 }
 101
 102 sub find_branch_id
 103 {
 104     my ($gittree, $name) = (shift, shift);
 105
 106     my $GIT;
 107
 108     open FOO, "cd $gittree ".
 109         "&& git rev-parse $name 2>/dev/null|" or
 110         die "git rev-parse $name failed";
 111
 112     my $id = <FOO>;
 113     close GIT;
 114     chomp($id);
 115     $id = "" if ($id eq $name);
 116     return $id;
 117 }
 118
 119 sub parse_git_tree
 120 {
 121     my ($gittree, $branch, $shortbranch) = (shift, shift, shift);
 122
 123     $syncname = "git2svn-syncpoint-${shortbranch}";
 124     print STDERR "syncname tag: $syncname\n" if ($verbose);
 125
 126     $masterrev = find_branch_id($gittree, $branch);
 127     die "No head found for ${branch}" if ($masterrev eq "");
 128
 129     my $oldmasterrev = find_branch_id($gittree, $syncname);
 130
 131     if ($oldmasterrev ne "") {
 132
 133         die "no $svntree, but incremental (have synctag) ?\n".
 134             "(\"cd $gittree && git tag -d $syncname\" to remove)"
 135             unless ( -d $svntree);
 136
 137         if (${oldmasterrev} eq $masterrev) {
 138             print STDERR "nothing to sync, $syncname matches $branch\n"
 139                 if ($verbose);
 140             exit 0;
 141         }
 142
 143         $fexport = "$oldmasterrev..$masterrev";
 144     } else {
 145         $fexport="$masterrev";
 146
 147         system("svnadmin create ./$svntree") unless (-d $svntree);
 148     }
 149 }
 150
 151
 152 sub checkdirs
 153 {
 154     my $path = shift;
 155     my $base = "";
 156
 157     # pick first dir, create, take next dir, continue until we reached basename
 158     while ($path =~ m@^([^/]+)/(.*)$@) {
 159         my $first = $base . $1;
 160         $path = $2;
 161         $base = $first . "/";
 162         next if ($paths{$first});
 163
 164         $paths{$first} = 1;
 165
 166         printf OUT "Node-path: $first\n";
 167         printf OUT "Node-kind: dir\n";
 168         printf OUT "Node-action: add\n";
 169         printf OUT "Prop-content-length: 0\n";
 170         printf OUT "Content-length: 0\n";
 171         printf OUT "\n";
 172     }
 173 }
 174
 175 sub next_line
 176 {
 177     my $IN = shift;
 178     my $next = <IN>;
 179     chomp $next;
 180     return $next;
 181 }
 182
 183 #
 184 # This is to allow setting props for a path, XXX add configuration
 185 # file/options for this.
 186 #
 187
 188 sub auto_props
 189 {
 190     my $path = shift;
 191     ### given path, return prop("prop", "value")
 192     return "";
 193 }
 194
 195
 196 $|= 1;
 197
 198 my $result;
 199 $result = GetOptions ("git-branch=s" => \$branch,
 200                       "svn-prefix=s" => \$basedir,
 201                       "keep-logs" => \$keeplogs,
 202                       "no-load" => \$no_load,
 203                       "verbose+" => \$verbose,
 204                       "help" => \$help) or pod2usage(2);
 205
 206 pod2usage(0) if ($help);
 207
 208 die "to few arguments" if ($#ARGV < 1);
 209
 210 mkdir ".data" unless (-d ".data");
 211
 212 die "cant find branch name" unless ($branch =~ m@/?([^/]+)$@);
 213 my $shortbranch = $1;
 214
 215 my $gitdump = ".data/git.dump-${shortbranch}";
 216 my $svndump = ".data/svn.dump-${shortbranch}";
 217 my $log = ".data/log-${shortbranch}";
 218
 219 my $gittree = $ARGV[0];
 220 $svntree = $ARGV[1];
 221
 222 parse_git_tree($gittree, $branch, $shortbranch);
 223
 224 my $cwd = `pwd`;
 225 chomp($cwd);
 226 parse_svn_tree("file://" . $cwd ."/". $svntree);
 227
 228 system(">$log");
 229
 230 print STDERR "git fast-export $branch ($fexport)\n" if ($verbose);
 231
 232 system("(cd $gittree && git fast-export $fexport) > $gitdump 2>$log") == 0 or
 233     die "git fast-export: $!";
 234
 235 open IN, "$gitdump" or
 236     die "failed to open $gitdump";
 237
 238 open OUT, ">$svndump" or
 239     die "failed to open $svndump";
 240
 241 print STDERR "creating svn dump from revision $revision...\n" if ($verbose);
 242
 243 print OUT "SVN-fs-dump-format-version: 3\n";
 244
 245 my $next = next_line();
 246 COMMAND: while (!eof(IN)) {
 247     if ($next eq "") {
 248         $next = next_line($IN);
 249         next COMMAND;
 250     } elsif ($next =~ /^commit (.*)/) {
 251
 252         my %commit;
 253
 254         $next = next_line($IN);
 255         if ($next =~ m/mark +(.*)/) {
 256             $commit{Mark} = $1;
 257             $next = next_line($IN);
 258         }
 259         if ($next =~ m/author +(.*)/) {
 260             $commit{Author} = $1;
 261             $next = next_line($IN);
 262         }
 263         unless ($next =~ m/committer +(.+) +<([^>]+)> +(\d+) +[+-](\d+)$/) {
 264             die "missing comitter: $_";
 265         }
 266
 267         $commit{CommitterName} = $1;
 268         $commit{CommitterEmail} = $2;
 269         $commit{CommitterWhen} = $3;
 270         $commit{CommitterTZ} = $4;
 271
 272         $next = next_line($IN);
 273         my $log = read_data($IN, $next);
 274
 275         $next = next_line($IN);
 276         if ($next =~ m/from (.*)/) {
 277             $commit{From} = $1;
 278             $next = next_line($IN);
 279         }
 280         if ($next =~ m/merge (.*)/) {
 281             $commit{Merge} = $1;
 282             $next = next_line($IN);
 283         }
 284
 285         my $date =
 286             strftime("%Y-%m-%dT%H:%M:%S.000000Z",
 287                      gmtime($commit{CommitterWhen}));
 288
 289         my $author = "(no author)";
 290         if ($commit{CommitterEmail} =~ m/([^@]+)/) {
 291             $author = $1;
 292         }
 293         $author = "git2svn-dump" if ($author eq "(no author)");
 294
 295         my $props = "";
 296         $props .= prop("svn:author", $author);
 297         $props .= prop("svn:log", $log);
 298         $props .= prop("svn:date", $date);
 299         $props .= "PROPS-END";
 300
 301         # push out svn info
 302
 303         printf OUT "Revision-number: $revision\n"; $revision++;
 304         printf OUT "Prop-content-length: ". length($props) . "\n";
 305         printf OUT "Content-length: " . length($props) . "\n";
 306         printf OUT "\n";
 307         print OUT "$props\n";
 308
 309         while (1) {
 310             if ($next =~ m/M (\d+) (\S+) (.*)$/) {
 311                 my ($mode, $dataref, $path) = (oct $1, $2, $3);
 312                 my $content;
 313                 if ($dataref eq "inline") {
 314                     $next = next_line($IN);
 315                     $content = read_data($IN, $next);
 316                 } else {
 317                     die "Revision missing content ref $dataref"
 318                         unless(defined $blob{$dataref});
 319
 320                     $content = $blob{$dataref};
 321                     # here we really want to delete $blob{$dataref},
 322                     # but it might be referenced in the future. To
 323                     # avoid keepig everything in memory for larger
 324                     # repositories this must be written out to disk
 325                     # and removed when done.
 326                 }
 327
 328                 $path = "$basedir/$path";
 329                 checkdirs($path);
 330
 331                 my $action = "add";
 332
 333                 if ($paths{$path}) {
 334                     die "file was a dir" if ($paths{$path} != 2);
 335                     $action = "change";
 336                 } else {
 337                     $paths{$path} = 2;
 338                 }
 339
 340
 341                 my $type = $mode & 0777000;
 342                 my $kind = "";
 343                 $kind = "file" if ($type == 0100000);
 344                 $kind = "symlink" if ($type == 0120000);
 345                 die "$type unknown" if ($kind eq "");
 346
 347                 $props = "";
 348                 $props .= prop("svn:executable", "on") if ($mode & 0111);
 349                 $props .= prop("svn:special", "*") if ($kind eq "symlink");
 350                 $props .= auto_props($path);
 351                 $props .= "PROPS-END" if ($props ne "");
 352
 353                 $content = "link $content" if ($kind eq "symlink");
 354
 355                 my $plen = length($props);
 356                 my $clen = length($content);
 357
 358                 printf OUT "Node-path: $path\n";
 359                 printf OUT "Node-kind: file\n";
 360                 printf OUT "Node-action: $action\n";
 361                 printf OUT "Text-content-length: $clen\n";
 362                 printf OUT "Content-length: " . ($clen + $plen) . "\n";
 363                 printf OUT "Prop-content-length: $plen\n" if ($plen);
 364                 printf OUT "\n";
 365
 366                 print OUT "$props\n" if ($plen);
 367
 368                 print OUT $content;
 369                 printf OUT "\n";
 370             } elsif ($next =~ m/D (.*)/) {
 371                 my $path = $basedir . "/". $1;
 372
 373                 if ($paths{$path}) {
 374                     delete $paths{$path};
 375
 376                     printf OUT "Node-path: $path\n";
 377                     printf OUT "Node-action: delete\n";
 378                     printf OUT "\n";
 379                 } elsif ($verbose) {
 380                     print STDERR "deleting non existing object: $path\n";
 381                 }
 382
 383             } elsif ($next =~ m/^C (.*)/) {
 384                 die "file copy ?";
 385             } elsif ($next =~ m/^R (.*)/) {
 386                 die "file rename ?";
 387             } elsif ($next =~ m/^filedeleteall$/) {
 388                 die "file delete all ?";
 389             } else {
 390                 next COMMAND;
 391             }
 392             $next = next_line($IN);
 393         }
 394
 395     } elsif ($next =~ /^tag .*/) {
 396     } elsif ($next =~ /^reset .*/) {
 397     } elsif ($next =~ /^blob/) {
 398         my $mark = undef;
 399         $next = next_line($IN);
 400         if ($next =~ m/mark (.*)/) {
 401             $mark = $1;
 402             $next = next_line($IN);
 403         }
 404         my $data = read_data($IN, $next);
 405         $blob{$mark} = $data if (defined $mark);
 406     } elsif ($next =~ /^checkpoint .*/) {
 407     } elsif ($next =~ /^progress (.*)/) {
 408         print STDERR "progress: $1\n" if ($verbose);
 409     } else {
 410         die "unknown command $next";
 411     }
 412     $next = next_line($IN);
 413 }
 414
 415 close IN;
 416 close OUT;
 417
 418 print STDERR "...dumped to revision $revision\n" if ($verbose);
 419 print STDERR "(re-)setting sync-tag to new master\n" if ($verbose);
 420
 421 unless ($no_load) {
 422     system("cd $gittree && ".
 423            "git tag -m \"sync $(date)\" -a -f ${syncname} ${masterrev}");
 424 }
 425
 426 print STDERR "loading dump into svn\n" if ($verbose);
 427
 428 unless ($no_load) {
 429     system("svnadmin load $svntree < $svndump >>$log 2>&1") == 0 or
 430         die "svnadmin load";
 431 }
 432
 433 unlink $svndump, $gitdump, $log unless ($keeplogs);
 434
 435 exit 0;
 436
 437
 438 __END__
 439
 440 =head1 NAME
 441
 442 B<git2svn> - converts a git branch to a svn ditto
 443
 444 =head1 SYNOPSIS
 445
 446 B<git2svn> [options] git-repro svn-repro
 447
 448 =head1 OPTIONS
 449
 450 =over 8
 451
 452 =item B<--git-branch>
 453
 454 The git branch to export. The default is branch is master.
 455
 456 =item B<--svn-prefix>
 457
 458 The svn prefix where the branch is import. The default is trunk to
 459 match the default GIT branch (master).
 460
 461 =item B<--no-load>
 462
 463 Don't load the svn repository or update the syncpoint tagname.
 464
 465 =item B<--keep-logs>
 466
 467 Don't delete the logs in $CWD/.data on success.
 468
 469 =item B<--verbose>
 470
 471 More verbose output, can be give more then once to increase the verbosity.
 472
 473 =item B<--help>
 474
 475 Print a brief help message and exits.
 476
 477 =back
 478
 479 =head1 DESCRIPTION
 480
 481 B<git2svn> will convert a git branch to a svn ditto, it also
 482 support incremantal updates.
 483
 484 B<git2svn> takes a git fast-export dump and converts it into a
 485 svn dump that is feed into svnadmin load.
 486
 487 B<git2svn> assumes its the only process that writes into the svn
 488 repository. This is because of the race between getting the to svn
 489 Revsion number from the svn, creating the dump with correct Revsions,
 490 and do the svnadmin load.
 491
 492 B<git2svn> also support incremental updates from a git branch to
 493 a svn reprositry. Its does this by setting a git tag
 494 (git2svn-syncpoint-<branchname>) where the last update was pulled
 495 from.
 496
 497 B<git2svn> was created as a hack over a weekend to support a
 498 smoother migration away from svn and allow users access to tools to
 499 browse and search code (fisheye) and use anonymouns svn servers.
 500
 501 =head1 EXAMPLES
 502
 503         B<git2svn> ~/src/heimdal svn-repro
 504
 505         B<git2svn> --git-branch heimdal-1-0-branch \
 506              --svn-prefix branches/heimdal-1-0-branch \
 507              ~/src/heimdal svn-repro
 508
 509 =head1 DOWNLOAD
 510
 511 B<git2svn> is avaible from repo.or.cz
 512
 513         git clone git://repo.or.cz/git2svn.git
 514
 515 =head1 AUTHORS
 516
 517         Love Hörnquist Åstrand <lha@kth.se>
 518
 519 =head1 BUGS
 520
 521         Send bug reports to lha@kth.se
 522
 523 =cut