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                     if (defined $blob{$dataref}) {
 323                         $content = $blob{$dataref};
 324                     } else {
 325                                 # Submodules cannot be converted
 326                                 print STDERR "Ignored line, please check if this is a submodule: $next\n" if ($verbose);
 327                                 $next = next_line($IN);
 328                                 next;
 329                     }
 330                     # here we really want to delete $blob{$dataref},
 331                     # but it might be referenced in the future. To
 332                     # avoid keepig everything in memory for larger
 333                     # repositories this must be written out to disk
 334                     # and removed when done.
 335                 }
 336
 337                 $path = "$basedir/$path";
 338                 checkdirs($path);
 339
 340                 my $action = "add";
 341
 342                 my $type = $mode & 0777000;
 343
 344                 if ($paths{$path}) {
 345                     if (($paths{$path} != 2) && ($type != 0120000)) {
 346                                 die "file was a dir";
 347                     } elsif (($paths{$path} != 2) && ($type == 0120000)) {
 348                                 print STDERR "Dir is now a symlink, deleting: $path\n" if ($verbose);
 349                             delete $paths{$path};
 350                             foreach ( keys( %paths ) ) {
 351                         delete $paths{$_} if ( /^$path\// );
 352                                 }
 353                                 # This is now a file and not a path anymore
 354                                 $paths{$path} = 2;
 355                                 printf OUT "Node-path: $path\nNode-action: delete\n\n";
 356                     } else {
 357                       $action = "change";
 358                     }
 359                 } else {
 360                     $paths{$path} = 2;
 361                 }
 362
 363
 364                 my $kind = "";
 365                 $kind = "file" if ($type == 0100000);
 366                 $kind = "symlink" if ($type == 0120000);
 367                 die "$type unknown" if ($kind eq "");
 368
 369                 $props = "";
 370                 $props .= prop("svn:executable", "on") if ($mode & 0111);
 371                 $props .= prop("svn:special", "*") if ($kind eq "symlink");
 372                 $props .= auto_props($path);
 373                 $props .= "PROPS-END\n" if ($props ne "");
 374
 375                 $content = "link $content\n" if ($kind eq "symlink");
 376
 377                 my $plen = length($props);
 378                 my $clen = length($content);
 379
 380                 printf OUT "Node-path: $path\n";
 381                 printf OUT "Node-kind: file\n";
 382                 printf OUT "Node-action: $action\n";
 383                 printf OUT "Text-content-length: $clen\n";
 384                 printf OUT "Content-length: " . ($clen + $plen) . "\n";
 385                 printf OUT "Prop-content-length: $plen\n" if ($plen);
 386                 printf OUT "\n";
 387
 388                 print OUT "$props" if ($plen);
 389
 390                 print OUT $content;
 391                 printf OUT "\n";
 392             } elsif ($next =~ m/D (.*)/) {
 393                 my $path = $basedir . "/". $1;
 394
 395                 if ($paths{$path}) {
 396                     delete $paths{$path};
 397                     foreach ( keys( %paths ) ) {
 398                 delete $paths{$_} if ( /^$path\// );
 399                         }
 400
 401                     printf OUT "Node-path: $path\n";
 402                     printf OUT "Node-action: delete\n";
 403                     printf OUT "\n";
 404                 } elsif ($verbose) {
 405                     print STDERR "deleting non existing object: $path\n";
 406                 }
 407
 408             } elsif ($next =~ m/^C (.*)/) {
 409                 die "file copy ?";
 410             } elsif ($next =~ m/^R (.*)/) {
 411                 die "file rename ?";
 412             } elsif ($next =~ m/^filedeleteall$/) {
 413                 die "file delete all ?";
 414             } else {
 415                 next COMMAND;
 416             }
 417             $next = next_line($IN);
 418         }
 419
 420     } elsif ($next =~ /^tag .*/) {
 421     } elsif ($next =~ /^reset .*/) {
 422     } elsif ($next =~ /^blob/) {
 423         my $mark = undef;
 424         $next = next_line($IN);
 425         if ($next =~ m/mark (.*)/) {
 426             $mark = $1;
 427             $next = next_line($IN);
 428         }
 429         my $data = read_data($IN, $next);
 430         $blob{$mark} = $data if (defined $mark);
 431     } elsif ($next =~ /^checkpoint .*/) {
 432     } elsif ($next =~ /^progress (.*)/) {
 433         print STDERR "progress: $1\n" if ($verbose);
 434     } else {
 435         die "unknown command $next";
 436     }
 437     $next = next_line($IN);
 438 }
 439
 440 close IN;
 441 close OUT;
 442
 443 print STDERR "...dumped to revision $revision\n" if ($verbose);
 444 print STDERR "(re-)setting sync-tag to new master\n" if ($verbose);
 445
 446 unless ($no_load) {
 447     system("cd $gittree && ".
 448            "git tag -m \"sync $(date)\" -a -f ${syncname} ${masterrev}");
 449 }
 450
 451 print STDERR "loading dump into svn\n" if ($verbose);
 452
 453 unless ($no_load) {
 454     system("svnadmin load $svntree < $svndump >>$log 2>&1") == 0 or
 455         die "svnadmin load";
 456 }
 457
 458 unlink $svndump, $gitdump, $log unless ($keeplogs);
 459
 460 exit 0;
 461
 462
 463 __END__
 464
 465 =head1 NAME
 466
 467 B<git2svn> - converts a git branch to a svn ditto
 468
 469 =head1 SYNOPSIS
 470
 471 B<git2svn> [options] git-repro svn-repro
 472
 473 =head1 OPTIONS
 474
 475 =over 8
 476
 477 =item B<--git-branch>
 478
 479 The git branch to export. The default is branch is master.
 480
 481 =item B<--svn-prefix>
 482
 483 The svn prefix where the branch is import. The default is trunk to
 484 match the default GIT branch (master).
 485
 486 =item B<--no-load>
 487
 488 Don't load the svn repository or update the syncpoint tagname.
 489
 490 =item B<--keep-logs>
 491
 492 Don't delete the logs in $CWD/.data on success.
 493
 494 =item B<--verbose>
 495
 496 More verbose output, can be give more then once to increase the verbosity.
 497
 498 =item B<--help>
 499
 500 Print a brief help message and exits.
 501
 502 =back
 503
 504 =head1 DESCRIPTION
 505
 506 B<git2svn> will convert a git branch to a svn ditto, it also
 507 support incremantal updates.
 508
 509 B<git2svn> takes a git fast-export dump and converts it into a
 510 svn dump that is feed into svnadmin load.
 511
 512 B<git2svn> assumes its the only process that writes into the svn
 513 repository. This is because of the race between getting the to svn
 514 Revsion number from the svn, creating the dump with correct Revsions,
 515 and do the svnadmin load.
 516
 517 B<git2svn> also support incremental updates from a git branch to
 518 a svn reprositry. Its does this by setting a git tag
 519 (git2svn-syncpoint-<branchname>) where the last update was pulled
 520 from.
 521
 522 B<git2svn> was created as a hack over a weekend to support a
 523 smoother migration away from svn and allow users access to tools to
 524 browse and search code (fisheye) and use anonymouns svn servers.
 525
 526 =head1 EXAMPLES
 527
 528         B<git2svn> ~/src/heimdal svn-repro
 529
 530         B<git2svn> --git-branch heimdal-1-0-branch \
 531              --svn-prefix branches/heimdal-1-0-branch \
 532              ~/src/heimdal svn-repro
 533
 534 =head1 DOWNLOAD
 535
 536 B<git2svn> is avaible from repo.or.cz
 537
 538         git clone git://repo.or.cz/git2svn.git
 539
 540 =head1 AUTHORS
 541
 542         Love Hörnquist Åstrand <lha@kth.se>
 543
 544 =head1 BUGS
 545
 546         Send bug reports to lha@kth.se
 547
 548 =cut