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 $gittree = $ARGV[0];
 216 $svntree = $ARGV[1];
 217
 218 # create an identifier by replacing path separators
 219 # (i.e. "/", ":" and "\") with underscores
 220 my $svntree_id = $svntree;
 221 $svntree_id =~ s/[\/:\\]/_/g;
 222
 223 my $gitdump = ".data/git.dump-${svntree_id}-${shortbranch}";
 224 my $svndump = ".data/svn.dump-${svntree_id}-${shortbranch}";
 225 my $log = ".data/log-${svntree_id}-${shortbranch}";
 226
 227 parse_git_tree($gittree, $branch, $shortbranch);
 228
 229 my $cwd = `pwd`;
 230 chomp($cwd);
 231 parse_svn_tree("file://" . $cwd ."/". $svntree);
 232
 233 system(">$log");
 234
 235 print STDERR "git fast-export $branch ($fexport)\n" if ($verbose);
 236
 237 system("(cd $gittree && git fast-export $fexport) > $gitdump 2>$log") == 0 or
 238     die "git fast-export: $!";
 239
 240 open IN, "$gitdump" or
 241     die "failed to open $gitdump";
 242
 243 open OUT, ">$svndump" or
 244     die "failed to open $svndump";
 245
 246 print STDERR "creating svn dump from revision $revision...\n" if ($verbose);
 247
 248 print OUT "SVN-fs-dump-format-version: 3\n";
 249
 250 my $next = next_line();
 251 COMMAND: while (!eof(IN)) {
 252     if ($next eq "") {
 253         $next = next_line($IN);
 254         next COMMAND;
 255     } elsif ($next =~ /^commit (.*)/) {
 256
 257         my %commit;
 258
 259         $next = next_line($IN);
 260         if ($next =~ m/mark +(.*)/) {
 261             $commit{Mark} = $1;
 262             $next = next_line($IN);
 263         }
 264         if ($next =~ m/author +(.*)/) {
 265             $commit{Author} = $1;
 266             $next = next_line($IN);
 267         }
 268         unless ($next =~ m/committer +(.+) +<([^>]+)> +(\d+) +[+-](\d+)$/) {
 269             die "missing comitter: $_";
 270         }
 271
 272         $commit{CommitterName} = $1;
 273         $commit{CommitterEmail} = $2;
 274         $commit{CommitterWhen} = $3;
 275         $commit{CommitterTZ} = $4;
 276
 277         $next = next_line($IN);
 278         my $log = read_data($IN, $next);
 279
 280         $next = next_line($IN);
 281         if ($next =~ m/from (.*)/) {
 282             $commit{From} = $1;
 283             $next = next_line($IN);
 284         }
 285         if ($next =~ m/merge (.*)/) {
 286             $commit{Merge} = $1;
 287             $next = next_line($IN);
 288         }
 289
 290         my $date =
 291             strftime("%Y-%m-%dT%H:%M:%S.000000Z",
 292                      gmtime($commit{CommitterWhen}));
 293
 294         my $author = "(no author)";
 295         if ($commit{CommitterEmail} =~ m/([^@]+)/) {
 296             $author = $1;
 297         }
 298         $author = "git2svn-dump" if ($author eq "(no author)");
 299
 300         my $props = "";
 301         $props .= prop("svn:author", $author);
 302         $props .= prop("svn:log", $log);
 303         $props .= prop("svn:date", $date);
 304         $props .= "PROPS-END";
 305
 306         # push out svn info
 307
 308         printf OUT "Revision-number: $revision\n"; $revision++;
 309         printf OUT "Prop-content-length: ". length($props) . "\n";
 310         printf OUT "Content-length: " . length($props) . "\n";
 311         printf OUT "\n";
 312         print OUT "$props\n";
 313
 314         while (1) {
 315             if ($next =~ m/M (\d+) (\S+) (.*)$/) {
 316                 my ($mode, $dataref, $path) = (oct $1, $2, $3);
 317                 my $content;
 318                 if ($dataref eq "inline") {
 319                     $next = next_line($IN);
 320                     $content = read_data($IN, $next);
 321                 } else {
 322                     die "Revision missing content ref $dataref"
 323                         unless(defined $blob{$dataref});
 324
 325                     $content = $blob{$dataref};
 326                     # here we really want to delete $blob{$dataref},
 327                     # but it might be referenced in the future. To
 328                     # avoid keepig everything in memory for larger
 329                     # repositories this must be written out to disk
 330                     # and removed when done.
 331                 }
 332
 333                 $path = "$basedir/$path";
 334                 checkdirs($path);
 335
 336                 my $action = "add";
 337
 338                 if ($paths{$path}) {
 339                     die "file was a dir" if ($paths{$path} != 2);
 340                     $action = "change";
 341                 } else {
 342                     $paths{$path} = 2;
 343                 }
 344
 345
 346                 my $type = $mode & 0777000;
 347                 my $kind = "";
 348                 $kind = "file" if ($type == 0100000);
 349                 $kind = "symlink" if ($type == 0120000);
 350                 die "$type unknown" if ($kind eq "");
 351
 352                 $props = "";
 353                 $props .= prop("svn:executable", "on") if ($mode & 0111);
 354                 $props .= prop("svn:special", "*") if ($kind eq "symlink");
 355                 $props .= auto_props($path);
 356                 $props .= "PROPS-END" if ($props ne "");
 357
 358                 $content = "link $content" if ($kind eq "symlink");
 359
 360                 my $plen = length($props);
 361                 my $clen = length($content);
 362
 363                 printf OUT "Node-path: $path\n";
 364                 printf OUT "Node-kind: file\n";
 365                 printf OUT "Node-action: $action\n";
 366                 printf OUT "Text-content-length: $clen\n";
 367                 printf OUT "Content-length: " . ($clen + $plen) . "\n";
 368                 printf OUT "Prop-content-length: $plen\n" if ($plen);
 369                 printf OUT "\n";
 370
 371                 print OUT "$props\n" if ($plen);
 372
 373                 print OUT $content;
 374                 printf OUT "\n";
 375             } elsif ($next =~ m/D (.*)/) {
 376                 my $path = $basedir . "/". $1;
 377
 378                 if ($paths{$path}) {
 379                     delete $paths{$path};
 380
 381                     printf OUT "Node-path: $path\n";
 382                     printf OUT "Node-action: delete\n";
 383                     printf OUT "\n";
 384                 } elsif ($verbose) {
 385                     print STDERR "deleting non existing object: $path\n";
 386                 }
 387
 388             } elsif ($next =~ m/^C (.*)/) {
 389                 die "file copy ?";
 390             } elsif ($next =~ m/^R (.*)/) {
 391                 die "file rename ?";
 392             } elsif ($next =~ m/^filedeleteall$/) {
 393                 die "file delete all ?";
 394             } else {
 395                 next COMMAND;
 396             }
 397             $next = next_line($IN);
 398         }
 399
 400     } elsif ($next =~ /^tag .*/) {
 401     } elsif ($next =~ /^reset .*/) {
 402     } elsif ($next =~ /^blob/) {
 403         my $mark = undef;
 404         $next = next_line($IN);
 405         if ($next =~ m/mark (.*)/) {
 406             $mark = $1;
 407             $next = next_line($IN);
 408         }
 409         my $data = read_data($IN, $next);
 410         $blob{$mark} = $data if (defined $mark);
 411     } elsif ($next =~ /^checkpoint .*/) {
 412     } elsif ($next =~ /^progress (.*)/) {
 413         print STDERR "progress: $1\n" if ($verbose);
 414     } else {
 415         die "unknown command $next";
 416     }
 417     $next = next_line($IN);
 418 }
 419
 420 close IN;
 421 close OUT;
 422
 423 print STDERR "...dumped to revision $revision\n" if ($verbose);
 424 print STDERR "(re-)setting sync-tag to new master\n" if ($verbose);
 425
 426 unless ($no_load) {
 427     system("cd $gittree && ".
 428            "git tag -m \"sync $(date)\" -a -f ${syncname} ${masterrev}");
 429 }
 430
 431 print STDERR "loading dump into svn\n" if ($verbose);
 432
 433 unless ($no_load) {
 434     system("svnadmin load $svntree < $svndump >>$log 2>&1") == 0 or
 435         die "svnadmin load";
 436 }
 437
 438 unlink $svndump, $gitdump, $log unless ($keeplogs);
 439
 440 exit 0;
 441
 442
 443 __END__
 444
 445 =head1 NAME
 446
 447 B<git2svn> - converts a git branch to a svn ditto
 448
 449 =head1 SYNOPSIS
 450
 451 B<git2svn> [options] git-repro svn-repro
 452
 453 =head1 OPTIONS
 454
 455 =over 8
 456
 457 =item B<--git-branch>
 458
 459 The git branch to export. The default is branch is master.
 460
 461 =item B<--svn-prefix>
 462
 463 The svn prefix where the branch is import. The default is trunk to
 464 match the default GIT branch (master).
 465
 466 =item B<--no-load>
 467
 468 Don't load the svn repository or update the syncpoint tagname.
 469
 470 =item B<--keep-logs>
 471
 472 Don't delete the logs in $CWD/.data on success.
 473
 474 =item B<--verbose>
 475
 476 More verbose output, can be give more then once to increase the verbosity.
 477
 478 =item B<--help>
 479
 480 Print a brief help message and exits.
 481
 482 =back
 483
 484 =head1 DESCRIPTION
 485
 486 B<git2svn> will convert a git branch to a svn ditto, it also
 487 support incremantal updates.
 488
 489 B<git2svn> takes a git fast-export dump and converts it into a
 490 svn dump that is feed into svnadmin load.
 491
 492 B<git2svn> assumes its the only process that writes into the svn
 493 repository. This is because of the race between getting the to svn
 494 Revsion number from the svn, creating the dump with correct Revsions,
 495 and do the svnadmin load.
 496
 497 B<git2svn> also support incremental updates from a git branch to
 498 a svn reprositry. Its does this by setting a git tag
 499 (git2svn-syncpoint-<branchname>) where the last update was pulled
 500 from.
 501
 502 B<git2svn> was created as a hack over a weekend to support a
 503 smoother migration away from svn and allow users access to tools to
 504 browse and search code (fisheye) and use anonymouns svn servers.
 505
 506 =head1 EXAMPLES
 507
 508         B<git2svn> ~/src/heimdal svn-repro
 509
 510         B<git2svn> --git-branch heimdal-1-0-branch \
 511              --svn-prefix branches/heimdal-1-0-branch \
 512              ~/src/heimdal svn-repro
 513
 514 =head1 DOWNLOAD
 515
 516 B<git2svn> is avaible from repo.or.cz
 517
 518         git clone git://repo.or.cz/git2svn.git
 519
 520 =head1 AUTHORS
 521
 522         Love Hörnquist Åstrand <lha@kth.se>
 523
 524 =head1 BUGS
 525
 526         Send bug reports to lha@kth.se
 527
 528 =cut