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, $ignore_author);
  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 $path = shift;
 191     ### given path, return prop("prop", "value")
 192     return "";
 193 }
 194
 195 sub isPrefix
 196 {
 197         my ($prefix, $string) = (shift, shift);
 198         return $prefix eq substr($string, 0, length($prefix));
 199 }
 200
 201 $|= 1;
 202
 203 my $result;
 204 $result = GetOptions ("git-branch=s" => \$branch,
 205                       "svn-prefix=s" => \$basedir,
 206                       "keep-logs" => \$keeplogs,
 207                       "no-load" => \$no_load,
 208                       "ignore-author" => \$ignore_author,
 209                       "verbose+" => \$verbose,
 210                       "help" => \$help) or pod2usage(2);
 211
 212 pod2usage(0) if ($help);
 213
 214 die "to few arguments" if ($#ARGV < 1);
 215
 216 mkdir ".data" unless (-d ".data");
 217
 218 die "cant find branch name" unless ($branch =~ m@/?([^/]+)$@);
 219 my $shortbranch = $1;
 220
 221 my $gittree = $ARGV[0];
 222 $svntree = $ARGV[1];
 223
 224 # create an identifier by replacing path separators
 225 # (i.e. "/", ":" and "\") with underscores
 226 my $svntree_id = $svntree;
 227 $svntree_id =~ s/[\/:\\]/_/g;
 228
 229 my $gitdump = ".data/git.dump-${svntree_id}-${shortbranch}";
 230 my $svndump = ".data/svn.dump-${svntree_id}-${shortbranch}";
 231 my $log = ".data/log-${svntree_id}-${shortbranch}";
 232
 233 parse_git_tree($gittree, $branch, $shortbranch);
 234
 235 my $cwd = `pwd`;
 236 chomp($cwd);
 237 parse_svn_tree("file://" . $cwd ."/". $svntree);
 238
 239 system(">$log");
 240
 241 print STDERR "git fast-export $branch ($fexport)\n" if ($verbose);
 242
 243 system("(cd $gittree && git fast-export $fexport) > $gitdump 2>$log") == 0 or
 244     die "git fast-export: $!";
 245
 246 open IN, "$gitdump" or
 247     die "failed to open $gitdump";
 248
 249 open OUT, ">$svndump" or
 250     die "failed to open $svndump";
 251
 252 print STDERR "creating svn dump from revision $revision...\n" if ($verbose);
 253
 254 print OUT "SVN-fs-dump-format-version: 3\n";
 255
 256 my $next = next_line();
 257 COMMAND: while (!eof(IN)) {
 258     if ($next eq "") {
 259         $next = next_line($IN);
 260         next COMMAND;
 261     } elsif ($next =~ /^commit (.*)/) {
 262
 263         my %commit;
 264
 265         $next = next_line($IN);
 266         if ($next =~ m/mark +(.*)/) {
 267             $commit{Mark} = $1;
 268             $next = next_line($IN);
 269         }
 270         if ($next =~ m/author +(.+) +<([^>]+)> +(\d+) +[+-](\d+)$/) {
 271             $commit{AuthorName} = $1;
 272             $commit{AuthorEmail} = $2;
 273             $commit{AuthorWhen} = $3;
 274             $commit{AuthorTZ} = $4;
 275             $next = next_line($IN);
 276         }
 277         unless ($next =~ m/committer +(.+) +<([^>]+)> +(\d+) +[+-](\d+)$/) {
 278             die "missing committer: $_";
 279         }
 280
 281         $commit{CommitterName} = $1;
 282         $commit{CommitterEmail} = $2;
 283         $commit{CommitterWhen} = $3;
 284         $commit{CommitterTZ} = $4;
 285
 286         $next = next_line($IN);
 287         my $log = read_data($IN, $next);
 288         $log =~ s/\s+$//;
 289
 290         $next = next_line($IN);
 291         if ($next =~ m/from (.*)/) {
 292             $commit{From} = $1;
 293             $next = next_line($IN);
 294         }
 295         if ($next =~ m/merge (.*)/) {
 296             $commit{Merge} = $1;
 297             $next = next_line($IN);
 298         }
 299
 300         my $date =
 301             strftime("%Y-%m-%dT%H:%M:%S.000000Z",
 302                      gmtime($commit{CommitterWhen}));
 303
 304         my $author = "git2svn-dump";
 305         if ($commit{CommitterEmail} =~ m/([^@]+)/) {
 306             $author = $1;
 307         }
 308         unless ($ignore_author) {
 309             if ($commit{AuthorEmail} =~ m/([^@]+)/) {
 310                 $author = $1;
 311             }
 312         }
 313
 314         my $props = "";
 315         $props .= prop("svn:author", $author);
 316         $props .= prop("svn:log", $log);
 317         $props .= prop("svn:date", $date);
 318         $props .= "PROPS-END";
 319
 320         # push out svn info
 321
 322         printf OUT "Revision-number: $revision\n"; $revision++;
 323         printf OUT "Prop-content-length: ". length($props) . "\n";
 324         printf OUT "Content-length: " . length($props) . "\n";
 325         printf OUT "\n";
 326         print OUT "$props\n";
 327
 328         while (1) {
 329             if ($next =~ m/M (\d+) (\S+) (.*)$/) {
 330                 my ($mode, $dataref, $path) = (oct $1, $2, $3);
 331                 my $content;
 332                 if ($dataref eq "inline") {
 333                     $next = next_line($IN);
 334                     $content = read_data($IN, $next);
 335                 } else {
 336                     if (defined $blob{$dataref}) {
 337                         $content = $blob{$dataref};
 338                     } else {
 339                                 # Submodules cannot be converted
 340                                 print STDERR "Ignored line, please check if this is a submodule: $next\n" if ($verbose);
 341                                 $next = next_line($IN);
 342                                 next;
 343                     }
 344                     # here we really want to delete $blob{$dataref},
 345                     # but it might be referenced in the future. To
 346                     # avoid keepig everything in memory for larger
 347                     # repositories this must be written out to disk
 348                     # and removed when done.
 349                 }
 350
 351                 $path = "$basedir/$path";
 352                 checkdirs($path);
 353
 354                 my $action = "add";
 355
 356                 my $type = $mode & 0777000;
 357
 358                 if ($paths{$path}) {
 359                     if (($paths{$path} != 2) && ($type != 0120000)) {
 360                                 die "file was a dir";
 361                     } elsif (($paths{$path} != 2) && ($type == 0120000)) {
 362                                 print STDERR "Dir is now a symlink, deleting: $path\n" if ($verbose);
 363                             foreach ( keys( %paths ) ) {
 364                         delete $paths{$_} if ( /^$path\// );
 365                                 }
 366                                 # This is now a file and not a path anymore
 367                                 $paths{$path} = 2;
 368                                 printf OUT "Node-path: $path\nNode-action: delete\n\n";
 369                     } else {
 370                       $action = "change";
 371                     }
 372                 } else {
 373                     $paths{$path} = 2;
 374                 }
 375
 376
 377                 my $kind = "";
 378                 $kind = "file" if ($type == 0100000);
 379                 $kind = "symlink" if ($type == 0120000);
 380                 die "$type unknown" if ($kind eq "");
 381
 382                 $props = "";
 383                 $props .= prop("svn:executable", "on") if ($mode & 0111);
 384                 $props .= prop("svn:special", "*") if ($kind eq "symlink");
 385                 $props .= auto_props($path);
 386                 $props .= "PROPS-END\n" if ($props ne "");
 387
 388                 $content = "link $content\n" if ($kind eq "symlink");
 389
 390                 my $plen = length($props);
 391                 my $clen = length($content);
 392
 393                 printf OUT "Node-path: $path\n";
 394                 printf OUT "Node-kind: file\n";
 395                 printf OUT "Node-action: $action\n";
 396                 printf OUT "Text-content-length: $clen\n";
 397                 printf OUT "Content-length: " . ($clen + $plen) . "\n";
 398                 printf OUT "Prop-content-length: $plen\n" if ($plen);
 399                 printf OUT "\n";
 400
 401                 print OUT "$props" if ($plen);
 402
 403                 print OUT $content;
 404                 printf OUT "\n";
 405             } elsif ($next =~ m/D (.*)/) {
 406                 my $path = $basedir . "/". $1;
 407
 408                 if ($paths{$path}) {
 409                     delete $paths{$path};
 410                     $path .= "/";
 411                     foreach ( keys( %paths ) ) {
 412                         delete $paths{$_} if ( isPrefix($path, $_) );
 413                         }
 414
 415                     printf OUT "Node-path: $path\n";
 416                     printf OUT "Node-action: delete\n";
 417                     printf OUT "\n";
 418                 } elsif ($verbose) {
 419                     print STDERR "deleting non existing object: $path\n";
 420                 }
 421
 422             } elsif ($next =~ m/^C (.*)/) {
 423                 die "file copy ?";
 424             } elsif ($next =~ m/^R (.*)/) {
 425                 die "file rename ?";
 426             } elsif ($next =~ m/^filedeleteall$/) {
 427                 die "file delete all ?";
 428             } else {
 429                 next COMMAND;
 430             }
 431             $next = next_line($IN);
 432         }
 433
 434     } elsif ($next =~ /^tag .*/) {
 435     } elsif ($next =~ /^reset .*/) {
 436     } elsif ($next =~ /^blob/) {
 437         my $mark = undef;
 438         $next = next_line($IN);
 439         if ($next =~ m/mark (.*)/) {
 440             $mark = $1;
 441             $next = next_line($IN);
 442         }
 443         my $data = read_data($IN, $next);
 444         $blob{$mark} = $data if (defined $mark);
 445     } elsif ($next =~ /^checkpoint .*/) {
 446     } elsif ($next =~ /^progress (.*)/) {
 447         print STDERR "progress: $1\n" if ($verbose);
 448     } else {
 449         die "unknown command $next";
 450     }
 451     $next = next_line($IN);
 452 }
 453
 454 close IN;
 455 close OUT;
 456
 457 print STDERR "...dumped to revision $revision\n" if ($verbose);
 458 print STDERR "(re-)setting sync-tag to new master\n" if ($verbose);
 459
 460 unless ($no_load) {
 461     system("cd $gittree && ".
 462            "git tag -m \"sync $(date)\" -a -f ${syncname} ${masterrev}");
 463 }
 464
 465 print STDERR "loading dump into svn\n" if ($verbose);
 466
 467 unless ($no_load) {
 468     system("svnadmin load $svntree < $svndump >>$log 2>&1") == 0 or
 469         die "svnadmin load";
 470 }
 471
 472 unlink $svndump, $gitdump, $log unless ($keeplogs);
 473
 474 exit 0;
 475
 476
 477 __END__
 478
 479 =head1 NAME
 480
 481 B<git2svn> - converts a git branch to a svn ditto
 482
 483 =head1 SYNOPSIS
 484
 485 B<git2svn> [options] git-repro svn-repro
 486
 487 =head1 OPTIONS
 488
 489 =over 8
 490
 491 =item B<--git-branch>
 492
 493 The git branch to export. The default is branch is master.
 494
 495 =item B<--svn-prefix>
 496
 497 The svn prefix where the branch is import. The default is trunk to
 498 match the default GIT branch (master).
 499
 500 =item B<--no-load>
 501
 502 Don't load the svn repository or update the syncpoint tagname.
 503
 504 =item B<--ignore-author>
 505
 506 Ignore "author" lines in the fast-import stream. Use "committer"
 507 information instead.
 508
 509 =item B<--keep-logs>
 510
 511 Don't delete the logs in $CWD/.data on success.
 512
 513 =item B<--verbose>
 514
 515 More verbose output, can be give more then once to increase the verbosity.
 516
 517 =item B<--help>
 518
 519 Print a brief help message and exits.
 520
 521 =back
 522
 523 =head1 DESCRIPTION
 524
 525 B<git2svn> will convert a git branch to a svn ditto, it also
 526 support incremantal updates.
 527
 528 B<git2svn> takes a git fast-export dump and converts it into a
 529 svn dump that is feed into svnadmin load.
 530
 531 B<git2svn> assumes its the only process that writes into the svn
 532 repository. This is because of the race between getting the to svn
 533 Revsion number from the svn, creating the dump with correct Revsions,
 534 and do the svnadmin load.
 535
 536 B<git2svn> also support incremental updates from a git branch to
 537 a svn reprositry. Its does this by setting a git tag
 538 (git2svn-syncpoint-<branchname>) where the last update was pulled
 539 from.
 540
 541 B<git2svn> was created as a hack over a weekend to support a
 542 smoother migration away from svn and allow users access to tools to
 543 browse and search code (fisheye) and use anonymouns svn servers.
 544
 545 =head1 EXAMPLES
 546
 547         B<git2svn> ~/src/heimdal svn-repro
 548
 549         B<git2svn> --git-branch heimdal-1-0-branch \
 550              --svn-prefix branches/heimdal-1-0-branch \
 551              ~/src/heimdal svn-repro
 552
 553 =head1 DOWNLOAD
 554
 555 B<git2svn> is avaible from repo.or.cz
 556
 557         git clone git://repo.or.cz/git2svn.git
 558
 559 =head1 AUTHORS
 560
 561         Love Hörnquist Åstrand <lha@kth.se>
 562
 563 =head1 BUGS
 564
 565         Send bug reports to lha@kth.se
 566
 567 =cut