git-svnimport.perl

   1 #!/usr/bin/perl -w
   2
   3 # This tool is copyright (c) 2005, Matthias Urlichs.
   4 # It is released under the Gnu Public License, version 2.
   5 #
   6 # The basic idea is to pull and analyze SVN changes.
   7 #
   8 # Checking out the files is done by a single long-running SVN connection.
   9 #
  10 # The head revision is on branch "origin" by default.
  11 # You can change that with the '-o' option.
  12
  13 use strict;
  14 use warnings;
  15 use Getopt::Std;
  16 use File::Spec;
  17 use File::Temp qw(tempfile);
  18 use File::Path qw(mkpath);
  19 use File::Basename qw(basename dirname);
  20 use Time::Local;
  21 use IO::Pipe;
  22 use POSIX qw(strftime dup2);
  23 use IPC::Open2;
  24 use SVN::Core;
  25 use SVN::Ra;
  26
  27 die "Need SVN:Core 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
  28
  29 $SIG{'PIPE'}="IGNORE";
  30 $ENV{'TZ'}="UTC";
  31
  32 our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,
  33     $opt_b,$opt_r,$opt_I,$opt_A,$opt_s,$opt_l,$opt_d,$opt_D);
  34
  35 sub usage() {
  36         print STDERR <<END;
  37 Usage: ${\basename $0}     # fetch/update GIT from SVN
  38        [-o branch-for-HEAD] [-h] [-v] [-l max_rev]
  39        [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
  40        [-d|-D] [-i] [-u] [-r] [-I ignorefilename] [-s start_chg]
  41        [-m] [-M regex] [-A author_file] [SVN_URL]
  42 END
  43         exit(1);
  44 }
  45
  46 getopts("A:b:C:dDhiI:l:mM:o:rs:t:T:uv") or usage();
  47 usage if $opt_h;
  48
  49 my $tag_name = $opt_t || "tags";
  50 my $trunk_name = $opt_T || "trunk";
  51 my $branch_name = $opt_b || "branches";
  52
  53 @ARGV == 1 or @ARGV == 2 or usage();
  54
  55 $opt_o ||= "origin";
  56 $opt_s ||= 1;
  57 my $git_tree = $opt_C;
  58 $git_tree ||= ".";
  59
  60 my $svn_url = $ARGV[0];
  61 my $svn_dir = $ARGV[1];
  62
  63 our @mergerx = ();
  64 if ($opt_m) {
  65         @mergerx = ( qr/\W(?:from|of|merge|merging|merged) (\w+)/i );
  66 }
  67 if ($opt_M) {
  68         push (@mergerx, qr/$opt_M/);
  69 }
  70
  71 our %users = ();
  72 if ($opt_A) {
  73         die "Cannot open $opt_A\n" unless -f $opt_A;
  74         open(my $authors,$opt_A);
  75         while(<$authors>) {
  76                 chomp;
  77                 next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/;
  78                 (my $user,my $name,my $email) = ($1,$2,$3);
  79                 $users{$user} = [$name,$email];
  80         }
  81         close($authors);
  82 }
  83
  84 select(STDERR); $|=1; select(STDOUT);
  85
  86
  87 package SVNconn;
  88 # Basic SVN connection.
  89 # We're only interested in connecting and downloading, so ...
  90
  91 use File::Spec;
  92 use File::Temp qw(tempfile);
  93 use POSIX qw(strftime dup2);
  94
  95 sub new {
  96         my($what,$repo) = @_;
  97         $what=ref($what) if ref($what);
  98
  99         my $self = {};
 100         $self->{'buffer'} = "";
 101         bless($self,$what);
 102
 103         $repo =~ s#/+$##;
 104         $self->{'fullrep'} = $repo;
 105         $self->conn();
 106
 107         return $self;
 108 }
 109
 110 sub conn {
 111         my $self = shift;
 112         my $repo = $self->{'fullrep'};
 113         my $auth = SVN::Core::auth_open ([SVN::Client::get_simple_provider,
 114                           SVN::Client::get_ssl_server_trust_file_provider,
 115                           SVN::Client::get_username_provider]);
 116         my $s = SVN::Ra->new(url => $repo, auth => $auth);
 117         die "SVN connection to $repo: $!\n" unless defined $s;
 118         $self->{'svn'} = $s;
 119         $self->{'repo'} = $repo;
 120         $self->{'maxrev'} = $s->get_latest_revnum();
 121 }
 122
 123 sub file {
 124         my($self,$path,$rev) = @_;
 125
 126         my ($fh, $name) = tempfile('gitsvn.XXXXXX',
 127                     DIR => File::Spec->tmpdir(), UNLINK => 1);
 128
 129         print "... $rev $path ...\n" if $opt_v;
 130         my (undef, $properties);
 131         eval { (undef, $properties)
 132                    = $self->{'svn'}->get_file($path,$rev,$fh); };
 133         if($@) {
 134                 return undef if $@ =~ /Attempted to get checksum/;
 135                 die $@;
 136         }
 137         my $mode;
 138         if (exists $properties->{'svn:executable'}) {
 139                 $mode = '0755';
 140         } else {
 141                 $mode = '0644';
 142         }
 143         close ($fh);
 144
 145         return ($name, $mode);
 146 }
 147
 148 sub ignore {
 149         my($self,$path,$rev) = @_;
 150
 151         print "... $rev $path ...\n" if $opt_v;
 152         my (undef,undef,$properties)
 153             = $self->{'svn'}->get_dir($path,$rev,undef);
 154         if (exists $properties->{'svn:ignore'}) {
 155                 my ($fh, $name) = tempfile('gitsvn.XXXXXX',
 156                                            DIR => File::Spec->tmpdir(),
 157                                            UNLINK => 1);
 158                 print $fh $properties->{'svn:ignore'};
 159                 close($fh);
 160                 return $name;
 161         } else {
 162                 return undef;
 163         }
 164 }
 165
 166 package main;
 167 use URI;
 168
 169 our $svn = $svn_url;
 170 $svn .= "/$svn_dir" if defined $svn_dir;
 171 my $svn2 = SVNconn->new($svn);
 172 $svn = SVNconn->new($svn);
 173
 174 my $lwp_ua;
 175 if($opt_d or $opt_D) {
 176         $svn_url = URI->new($svn_url)->canonical;
 177         if($opt_D) {
 178                 $svn_dir =~ s#/*$#/#;
 179         } else {
 180                 $svn_dir = "";
 181         }
 182         if ($svn_url->scheme eq "http") {
 183                 use LWP::UserAgent;
 184                 $lwp_ua = LWP::UserAgent->new(keep_alive => 1, requests_redirectable => []);
 185         } else {
 186                 print STDERR "Warning: not HTTP; turning off direct file access\n";
 187                 $opt_d=0;
 188         }
 189 }
 190
 191 sub pdate($) {
 192         my($d) = @_;
 193         $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
 194                 or die "Unparseable date: $d\n";
 195         my $y=$1; $y-=1900 if $y>1900;
 196         return timegm($6||0,$5,$4,$3,$2-1,$y);
 197 }
 198
 199 sub getwd() {
 200         my $pwd = `pwd`;
 201         chomp $pwd;
 202         return $pwd;
 203 }
 204
 205
 206 sub get_headref($$) {
 207     my $name    = shift;
 208     my $git_dir = shift;
 209     my $sha;
 210
 211     if (open(C,"$git_dir/refs/heads/$name")) {
 212         chomp($sha = <C>);
 213         close(C);
 214         length($sha) == 40
 215             or die "Cannot get head id for $name ($sha): $!\n";
 216     }
 217     return $sha;
 218 }
 219
 220
 221 -d $git_tree
 222         or mkdir($git_tree,0777)
 223         or die "Could not create $git_tree: $!";
 224 chdir($git_tree);
 225
 226 my $orig_branch = "";
 227 my $forward_master = 0;
 228 my %branches;
 229
 230 my $git_dir = $ENV{"GIT_DIR"} || ".git";
 231 $git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
 232 $ENV{"GIT_DIR"} = $git_dir;
 233 my $orig_git_index;
 234 $orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
 235 my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
 236                                     DIR => File::Spec->tmpdir());
 237 close ($git_ih);
 238 $ENV{GIT_INDEX_FILE} = $git_index;
 239 my $maxnum = 0;
 240 my $last_rev = "";
 241 my $last_branch;
 242 my $current_rev = $opt_s || 1;
 243 unless(-d $git_dir) {
 244         system("git-init-db");
 245         die "Cannot init the GIT db at $git_tree: $?\n" if $?;
 246         system("git-read-tree");
 247         die "Cannot init an empty tree: $?\n" if $?;
 248
 249         $last_branch = $opt_o;
 250         $orig_branch = "";
 251 } else {
 252         -f "$git_dir/refs/heads/$opt_o"
 253                 or die "Branch '$opt_o' does not exist.\n".
 254                        "Either use the correct '-o branch' option,\n".
 255                        "or import to a new repository.\n";
 256
 257         -f "$git_dir/svn2git"
 258                 or die "'$git_dir/svn2git' does not exist.\n".
 259                        "You need that file for incremental imports.\n";
 260         open(F, "git-symbolic-ref HEAD |") or
 261                 die "Cannot run git-symbolic-ref: $!\n";
 262         chomp ($last_branch = <F>);
 263         $last_branch = basename($last_branch);
 264         close(F);
 265         unless($last_branch) {
 266                 warn "Cannot read the last branch name: $! -- assuming 'master'\n";
 267                 $last_branch = "master";
 268         }
 269         $orig_branch = $last_branch;
 270         $last_rev = get_headref($orig_branch, $git_dir);
 271         if (-f "$git_dir/SVN2GIT_HEAD") {
 272                 die <<EOM;
 273 SVN2GIT_HEAD exists.
 274 Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
 275 You may need to run
 276
 277     git-read-tree -m -u SVN2GIT_HEAD HEAD
 278 EOM
 279         }
 280         system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
 281
 282         $forward_master =
 283             $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
 284             system('cmp', '-s', "$git_dir/refs/heads/master",
 285                                 "$git_dir/refs/heads/$opt_o") == 0;
 286
 287         # populate index
 288         system('git-read-tree', $last_rev);
 289         die "read-tree failed: $?\n" if $?;
 290
 291         # Get the last import timestamps
 292         open my $B,"<", "$git_dir/svn2git";
 293         while(<$B>) {
 294                 chomp;
 295                 my($num,$branch,$ref) = split;
 296                 $branches{$branch}{$num} = $ref;
 297                 $branches{$branch}{"LAST"} = $ref;
 298                 $current_rev = $num+1 if $current_rev <= $num;
 299         }
 300         close($B);
 301 }
 302 -d $git_dir
 303         or die "Could not create git subdir ($git_dir).\n";
 304
 305 open BRANCHES,">>", "$git_dir/svn2git";
 306
 307 sub node_kind($$$) {
 308         my ($branch, $path, $revision) = @_;
 309         my $pool=SVN::Pool->new;
 310         my $kind = $svn->{'svn'}->check_path(revert_split_path($branch,$path),$revision,$pool);
 311         $pool->clear;
 312         return $kind;
 313 }
 314
 315 sub revert_split_path($$) {
 316         my($branch,$path) = @_;
 317
 318         my $svnpath;
 319         $path = "" if $path eq "/"; # this should not happen, but ...
 320         if($branch eq "/") {
 321                 $svnpath = "$trunk_name/$path";
 322         } elsif($branch =~ m#^/#) {
 323                 $svnpath = "$tag_name$branch/$path";
 324         } else {
 325                 $svnpath = "$branch_name/$branch/$path";
 326         }
 327
 328         $svnpath =~ s#/+$##;
 329         return $svnpath;
 330 }
 331
 332 sub get_file($$$) {
 333         my($rev,$branch,$path) = @_;
 334
 335         my $svnpath = revert_split_path($branch,$path);
 336
 337         # now get it
 338         my ($name,$mode);
 339         if($opt_d) {
 340                 my($req,$res);
 341
 342                 # /svn/!svn/bc/2/django/trunk/django-docs/build.py
 343                 my $url=$svn_url->clone();
 344                 $url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath");
 345                 print "... $path...\n" if $opt_v;
 346                 $req = HTTP::Request->new(GET => $url);
 347                 $res = $lwp_ua->request($req);
 348                 if ($res->is_success) {
 349                         my $fh;
 350                         ($fh, $name) = tempfile('gitsvn.XXXXXX',
 351                         DIR => File::Spec->tmpdir(), UNLINK => 1);
 352                         print $fh $res->content;
 353                         close($fh) or die "Could not write $name: $!\n";
 354                 } else {
 355                         return undef if $res->code == 301; # directory?
 356                         die $res->status_line." at $url\n";
 357                 }
 358                 $mode = '0644'; # can't obtain mode via direct http request?
 359         } else {
 360                 ($name,$mode) = $svn->file("$svnpath",$rev);
 361                 return undef unless defined $name;
 362         }
 363
 364         my $pid = open(my $F, '-|');
 365         die $! unless defined $pid;
 366         if (!$pid) {
 367             exec("git-hash-object", "-w", $name)
 368                 or die "Cannot create object: $!\n";
 369         }
 370         my $sha = <$F>;
 371         chomp $sha;
 372         close $F;
 373         unlink $name;
 374         return [$mode, $sha, $path];
 375 }
 376
 377 sub get_ignore($$$$$) {
 378         my($new,$old,$rev,$branch,$path) = @_;
 379
 380         return unless $opt_I;
 381         my $svnpath = revert_split_path($branch,$path);
 382         my $name = $svn->ignore("$svnpath",$rev);
 383         if ($path eq '/') {
 384                 $path = $opt_I;
 385         } else {
 386                 $path = File::Spec->catfile($path,$opt_I);
 387         }
 388         if (defined $name) {
 389                 my $pid = open(my $F, '-|');
 390                 die $! unless defined $pid;
 391                 if (!$pid) {
 392                         exec("git-hash-object", "-w", $name)
 393                             or die "Cannot create object: $!\n";
 394                 }
 395                 my $sha = <$F>;
 396                 chomp $sha;
 397                 close $F;
 398                 unlink $name;
 399                 push(@$new,['0644',$sha,$path]);
 400         } else {
 401                 push(@$old,$path);
 402         }
 403 }
 404
 405 sub split_path($$) {
 406         my($rev,$path) = @_;
 407         my $branch;
 408
 409         if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
 410                 $branch = "/$1";
 411         } elsif($path =~ s#^/\Q$trunk_name\E/?##) {
 412                 $branch = "/";
 413         } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
 414                 $branch = $1;
 415         } else {
 416                 my %no_error = (
 417                         "/" => 1,
 418                         "/$tag_name" => 1,
 419                         "/$branch_name" => 1
 420                 );
 421                 print STDERR "$rev: Unrecognized path: $path\n" unless (defined $no_error{$path});
 422                 return ()
 423         }
 424         $path = "/" if $path eq "";
 425         return ($branch,$path);
 426 }
 427
 428 sub branch_rev($$) {
 429
 430         my ($srcbranch,$uptorev) = @_;
 431
 432         my $bbranches = $branches{$srcbranch};
 433         my @revs = reverse sort { ($a eq 'LAST' ? 0 : $a) <=> ($b eq 'LAST' ? 0 : $b) } keys %$bbranches;
 434         my $therev;
 435         foreach my $arev(@revs) {
 436                 next if  ($arev eq 'LAST');
 437                 if ($arev <= $uptorev) {
 438                         $therev = $arev;
 439                         last;
 440                 }
 441         }
 442         return $therev;
 443 }
 444
 445 sub copy_path($$$$$$$$) {
 446         # Somebody copied a whole subdirectory.
 447         # We need to find the index entries from the old version which the
 448         # SVN log entry points to, and add them to the new place.
 449
 450         my($newrev,$newbranch,$path,$oldpath,$rev,$node_kind,$new,$parents) = @_;
 451
 452         my($srcbranch,$srcpath) = split_path($rev,$oldpath);
 453         unless(defined $srcbranch) {
 454                 print "Path not found when copying from $oldpath @ $rev\n";
 455                 return;
 456         }
 457         my $therev = branch_rev($srcbranch, $rev);
 458         my $gitrev = $branches{$srcbranch}{$therev};
 459         unless($gitrev) {
 460                 print STDERR "$newrev:$newbranch: could not find $oldpath \@ $rev\n";
 461                 return;
 462         }
 463         if ($srcbranch ne $newbranch) {
 464                 push(@$parents, $branches{$srcbranch}{'LAST'});
 465         }
 466         print "$newrev:$newbranch:$path: copying from $srcbranch:$srcpath @ $rev\n" if $opt_v;
 467         if ($node_kind eq $SVN::Node::dir) {
 468                         $srcpath =~ s#/*$#/#;
 469         }
 470
 471         my $pid = open my $f,'-|';
 472         die $! unless defined $pid;
 473         if (!$pid) {
 474                 exec("git-ls-tree","-r","-z",$gitrev,$srcpath)
 475                         or die $!;
 476         }
 477         local $/ = "\0";
 478         while(<$f>) {
 479                 chomp;
 480                 my($m,$p) = split(/\t/,$_,2);
 481                 my($mode,$type,$sha1) = split(/ /,$m);
 482                 next if $type ne "blob";
 483                 if ($node_kind eq $SVN::Node::dir) {
 484                         $p = $path . substr($p,length($srcpath)-1);
 485                 } else {
 486                         $p = $path;
 487                 }
 488                 push(@$new,[$mode,$sha1,$p]);
 489         }
 490         close($f) or
 491                 print STDERR "$newrev:$newbranch: could not list files in $oldpath \@ $rev\n";
 492 }
 493
 494 sub commit {
 495         my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
 496         my($author_name,$author_email,$dest);
 497         my(@old,@new,@parents);
 498
 499         if (not defined $author) {
 500                 $author_name = $author_email = "unknown";
 501         } elsif ($opt_A) {
 502                 die "User $author is not listed in $opt_A\n"
 503                     unless exists $users{$author};
 504                 ($author_name,$author_email) = @{$users{$author}};
 505         } elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
 506                 ($author_name, $author_email) = ($1, $2);
 507         } else {
 508                 $author =~ s/^<(.*)>$/$1/;
 509                 $author_name = $author_email = $author;
 510         }
 511         $date = pdate($date);
 512
 513         my $tag;
 514         my $parent;
 515         if($branch eq "/") { # trunk
 516                 $parent = $opt_o;
 517         } elsif($branch =~ m#^/(.+)#) { # tag
 518                 $tag = 1;
 519                 $parent = $1;
 520         } else { # "normal" branch
 521                 # nothing to do
 522                 $parent = $branch;
 523         }
 524         $dest = $parent;
 525
 526         my $prev = $changed_paths->{"/"};
 527         if($prev and $prev->[0] eq "A") {
 528                 delete $changed_paths->{"/"};
 529                 my $oldpath = $prev->[1];
 530                 my $rev;
 531                 if(defined $oldpath) {
 532                         my $p;
 533                         ($parent,$p) = split_path($revision,$oldpath);
 534                         if($parent eq "/") {
 535                                 $parent = $opt_o;
 536                         } else {
 537                                 $parent =~ s#^/##; # if it's a tag
 538                         }
 539                 } else {
 540                         $parent = undef;
 541                 }
 542         }
 543
 544         my $rev;
 545         if($revision > $opt_s and defined $parent) {
 546                 open(H,"git-rev-parse --verify $parent |");
 547                 $rev = <H>;
 548                 close(H) or do {
 549                         print STDERR "$revision: cannot find commit '$parent'!\n";
 550                         return;
 551                 };
 552                 chop $rev;
 553                 if(length($rev) != 40) {
 554                         print STDERR "$revision: cannot find commit '$parent'!\n";
 555                         return;
 556                 }
 557                 $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
 558                 if($revision != $opt_s and not $rev) {
 559                         print STDERR "$revision: do not know ancestor for '$parent'!\n";
 560                         return;
 561                 }
 562         } else {
 563                 $rev = undef;
 564         }
 565
 566 #       if($prev and $prev->[0] eq "A") {
 567 #               if(not $tag) {
 568 #                       unless(open(H,"> $git_dir/refs/heads/$branch")) {
 569 #                               print STDERR "$revision: Could not create branch $branch: $!\n";
 570 #                               $state=11;
 571 #                               next;
 572 #                       }
 573 #                       print H "$rev\n"
 574 #                               or die "Could not write branch $branch: $!";
 575 #                       close(H)
 576 #                               or die "Could not write branch $branch: $!";
 577 #               }
 578 #       }
 579         if(not defined $rev) {
 580                 unlink($git_index);
 581         } elsif ($rev ne $last_rev) {
 582                 print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
 583                 system("git-read-tree", $rev);
 584                 die "read-tree failed for $rev: $?\n" if $?;
 585                 $last_rev = $rev;
 586         }
 587
 588         push (@parents, $rev) if defined $rev;
 589
 590         my $cid;
 591         if($tag and not %$changed_paths) {
 592                 $cid = $rev;
 593         } else {
 594                 my @paths = sort keys %$changed_paths;
 595                 foreach my $path(@paths) {
 596                         my $action = $changed_paths->{$path};
 597
 598                         if ($action->[0] eq "R") {
 599                                 # refer to a file/tree in an earlier commit
 600                                 push(@old,$path); # remove any old stuff
 601                         }
 602                         if(($action->[0] eq "A") || ($action->[0] eq "R")) {
 603                                 my $node_kind = node_kind($branch,$path,$revision);
 604                                 if($action->[1]) {
 605                                         copy_path($revision,$branch,$path,$action->[1],$action->[2],$node_kind,\@new,\@parents);
 606                                 } elsif ($node_kind eq $SVN::Node::file) {
 607                                         my $f = get_file($revision,$branch,$path);
 608                                         if ($f) {
 609                                                 push(@new,$f) if $f;
 610                                         } else {
 611                                                 my $opath = $action->[3];
 612                                                 print STDERR "$revision: $branch: could not fetch '$opath'\n";
 613                                         }
 614                                 } elsif ($node_kind eq $SVN::Node::dir) {
 615                                         get_ignore(\@new, \@old, $revision,
 616                                                    $branch,$path);
 617                                 }
 618                         } elsif ($action->[0] eq "D") {
 619                                 push(@old,$path);
 620                         } elsif ($action->[0] eq "M") {
 621                                 my $node_kind = node_kind($branch,$path,$revision);
 622                                 if ($node_kind eq $SVN::Node::file) {
 623                                         my $f = get_file($revision,$branch,$path);
 624                                         push(@new,$f) if $f;
 625                                 } elsif ($node_kind eq $SVN::Node::dir) {
 626                                         get_ignore(\@new, \@old, $revision,
 627                                                    $branch,$path);
 628                                 }
 629                         } else {
 630                                 die "$revision: unknown action '".$action->[0]."' for $path\n";
 631                         }
 632                 }
 633
 634                 while(@old) {
 635                         my @o1;
 636                         if(@old > 55) {
 637                                 @o1 = splice(@old,0,50);
 638                         } else {
 639                                 @o1 = @old;
 640                                 @old = ();
 641                         }
 642                         my $pid = open my $F, "-|";
 643                         die "$!" unless defined $pid;
 644                         if (!$pid) {
 645                                 exec("git-ls-files", "-z", @o1) or die $!;
 646                         }
 647                         @o1 = ();
 648                         local $/ = "\0";
 649                         while(<$F>) {
 650                                 chomp;
 651                                 push(@o1,$_);
 652                         }
 653                         close($F);
 654
 655                         while(@o1) {
 656                                 my @o2;
 657                                 if(@o1 > 55) {
 658                                         @o2 = splice(@o1,0,50);
 659                                 } else {
 660                                         @o2 = @o1;
 661                                         @o1 = ();
 662                                 }
 663                                 system("git-update-index","--force-remove","--",@o2);
 664                                 die "Cannot remove files: $?\n" if $?;
 665                         }
 666                 }
 667                 while(@new) {
 668                         my @n2;
 669                         if(@new > 12) {
 670                                 @n2 = splice(@new,0,10);
 671                         } else {
 672                                 @n2 = @new;
 673                                 @new = ();
 674                         }
 675                         system("git-update-index","--add",
 676                                 (map { ('--cacheinfo', @$_) } @n2));
 677                         die "Cannot add files: $?\n" if $?;
 678                 }
 679
 680                 my $pid = open(C,"-|");
 681                 die "Cannot fork: $!" unless defined $pid;
 682                 unless($pid) {
 683                         exec("git-write-tree");
 684                         die "Cannot exec git-write-tree: $!\n";
 685                 }
 686                 chomp(my $tree = <C>);
 687                 length($tree) == 40
 688                         or die "Cannot get tree id ($tree): $!\n";
 689                 close(C)
 690                         or die "Error running git-write-tree: $?\n";
 691                 print "Tree ID $tree\n" if $opt_v;
 692
 693                 my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
 694                 my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
 695                 $pid = fork();
 696                 die "Fork: $!\n" unless defined $pid;
 697                 unless($pid) {
 698                         $pr->writer();
 699                         $pw->reader();
 700                         open(OUT,">&STDOUT");
 701                         dup2($pw->fileno(),0);
 702                         dup2($pr->fileno(),1);
 703                         $pr->close();
 704                         $pw->close();
 705
 706                         my @par = ();
 707
 708                         # loose detection of merges
 709                         # based on the commit msg
 710                         foreach my $rx (@mergerx) {
 711                                 if ($message =~ $rx) {
 712                                         my $mparent = $1;
 713                                         if ($mparent eq 'HEAD') { $mparent = $opt_o };
 714                                         if ( -e "$git_dir/refs/heads/$mparent") {
 715                                                 $mparent = get_headref($mparent, $git_dir);
 716                                                 push (@parents, $mparent);
 717                                                 print OUT "Merge parent branch: $mparent\n" if $opt_v;
 718                                         }
 719                                 }
 720                         }
 721                         my %seen_parents = ();
 722                         my @unique_parents = grep { ! $seen_parents{$_} ++ } @parents;
 723                         foreach my $bparent (@unique_parents) {
 724                                 push @par, '-p', $bparent;
 725                                 print OUT "Merge parent branch: $bparent\n" if $opt_v;
 726                         }
 727
 728                         exec("env",
 729                                 "GIT_AUTHOR_NAME=$author_name",
 730                                 "GIT_AUTHOR_EMAIL=$author_email",
 731                                 "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
 732                                 "GIT_COMMITTER_NAME=$author_name",
 733                                 "GIT_COMMITTER_EMAIL=$author_email",
 734                                 "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
 735                                 "git-commit-tree", $tree,@par);
 736                         die "Cannot exec git-commit-tree: $!\n";
 737                 }
 738                 $pw->writer();
 739                 $pr->reader();
 740
 741                 $message =~ s/[\s\n]+\z//;
 742                 $message = "r$revision: $message" if $opt_r;
 743
 744                 print $pw "$message\n"
 745                         or die "Error writing to git-commit-tree: $!\n";
 746                 $pw->close();
 747
 748                 print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
 749                 chomp($cid = <$pr>);
 750                 length($cid) == 40
 751                         or die "Cannot get commit id ($cid): $!\n";
 752                 print "Commit ID $cid\n" if $opt_v;
 753                 $pr->close();
 754
 755                 waitpid($pid,0);
 756                 die "Error running git-commit-tree: $?\n" if $?;
 757         }
 758
 759         if (not defined $cid) {
 760                 $cid = $branches{"/"}{"LAST"};
 761         }
 762
 763         if(not defined $dest) {
 764                 print "... no known parent\n" if $opt_v;
 765         } elsif(not $tag) {
 766                 print "Writing to refs/heads/$dest\n" if $opt_v;
 767                 open(C,">$git_dir/refs/heads/$dest") and
 768                 print C ("$cid\n") and
 769                 close(C)
 770                         or die "Cannot write branch $dest for update: $!\n";
 771         }
 772
 773         if($tag) {
 774                 my($in, $out) = ('','');
 775                 $last_rev = "-" if %$changed_paths;
 776                 # the tag was 'complex', i.e. did not refer to a "real" revision
 777
 778                 $dest =~ tr/_/\./ if $opt_u;
 779                 $branch = $dest;
 780
 781                 my $pid = open2($in, $out, 'git-mktag');
 782                 print $out ("object $cid\n".
 783                     "type commit\n".
 784                     "tag $dest\n".
 785                     "tagger $author_name <$author_email>\n") and
 786                 close($out)
 787                     or die "Cannot create tag object $dest: $!\n";
 788
 789                 my $tagobj = <$in>;
 790                 chomp $tagobj;
 791
 792                 if ( !close($in) or waitpid($pid, 0) != $pid or
 793                                 $? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
 794                         die "Cannot create tag object $dest: $!\n";
 795                 }
 796
 797                 open(C,">$git_dir/refs/tags/$dest") and
 798                 print C ("$tagobj\n") and
 799                 close(C)
 800                         or die "Cannot create tag $branch: $!\n";
 801
 802                 print "Created tag '$dest' on '$branch'\n" if $opt_v;
 803         }
 804         $branches{$branch}{"LAST"} = $cid;
 805         $branches{$branch}{$revision} = $cid;
 806         $last_rev = $cid;
 807         print BRANCHES "$revision $branch $cid\n";
 808         print "DONE: $revision $dest $cid\n" if $opt_v;
 809 }
 810
 811 sub commit_all {
 812         # Recursive use of the SVN connection does not work
 813         local $svn = $svn2;
 814
 815         my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
 816         my %p;
 817         while(my($path,$action) = each %$changed_paths) {
 818                 $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev, $path ];
 819         }
 820         $changed_paths = \%p;
 821
 822         my %done;
 823         my @col;
 824         my $pref;
 825         my $branch;
 826
 827         while(my($path,$action) = each %$changed_paths) {
 828                 ($branch,$path) = split_path($revision,$path);
 829                 next if not defined $branch;
 830                 $done{$branch}{$path} = $action;
 831         }
 832         while(($branch,$changed_paths) = each %done) {
 833                 commit($branch, $changed_paths, $revision, $author, $date, $message);
 834         }
 835 }
 836
 837 $opt_l = $svn->{'maxrev'} if not defined $opt_l or $opt_l > $svn->{'maxrev'};
 838
 839 if ($svn->{'maxrev'} < $current_rev) {
 840     print "Up to date: no new revisions to fetch!\n" if $opt_v;
 841     unlink("$git_dir/SVN2GIT_HEAD");
 842     exit;
 843 }
 844
 845 print "Fetching from $current_rev to $opt_l ...\n" if $opt_v;
 846
 847 my $pool=SVN::Pool->new;
 848 $svn->{'svn'}->get_log("/",$current_rev,$opt_l,0,1,1,\&commit_all,$pool);
 849 $pool->clear;
 850
 851
 852 unlink($git_index);
 853
 854 if (defined $orig_git_index) {
 855         $ENV{GIT_INDEX_FILE} = $orig_git_index;
 856 } else {
 857         delete $ENV{GIT_INDEX_FILE};
 858 }
 859
 860 # Now switch back to the branch we were in before all of this happened
 861 if($orig_branch) {
 862         print "DONE\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 863         system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 864                 if $forward_master;
 865         unless ($opt_i) {
 866                 system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
 867                 die "read-tree failed: $?\n" if $?;
 868         }
 869 } else {
 870         $orig_branch = "master";
 871         print "DONE; creating $orig_branch branch\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 872         system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 873                 unless -f "$git_dir/refs/heads/master";
 874         system('git-update-ref', 'HEAD', "$orig_branch");
 875         unless ($opt_i) {
 876                 system('git checkout');
 877                 die "checkout failed: $?\n" if $?;
 878         }
 879 }
 880 unlink("$git_dir/SVN2GIT_HEAD");
 881 close(BRANCHES);