Andrew's git / gitweb.git / git-svnimport.perl
?
git-svnimport.perlon commit simple euristic for further free packing improvements (4e8da19)
   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
  13use strict;
  14use warnings;
  15use Getopt::Std;
  16use File::Copy;
  17use File::Spec;
  18use File::Temp qw(tempfile);
  19use File::Path qw(mkpath);
  20use File::Basename qw(basename dirname);
  21use Time::Local;
  22use IO::Pipe;
  23use POSIX qw(strftime dup2);
  24use IPC::Open2;
  25use SVN::Core;
  26use SVN::Ra;
  27
  28die "Need SVN:Core 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
  29
  30$SIG{'PIPE'}="IGNORE";
  31$ENV{'TZ'}="UTC";
  32
  33our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,
  34    $opt_b,$opt_r,$opt_I,$opt_A,$opt_s,$opt_l,$opt_d,$opt_D);
  35
  36sub usage() {
  37        print STDERR <<END;
  38Usage: ${\basename $0}     # fetch/update GIT from SVN
  39       [-o branch-for-HEAD] [-h] [-v] [-l max_rev]
  40       [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
  41       [-d|-D] [-i] [-u] [-r] [-I ignorefilename] [-s start_chg]
  42       [-m] [-M regex] [-A author_file] [SVN_URL]
  43END
  44        exit(1);
  45}
  46
  47getopts("A:b:C:dDhiI:l:mM:o:rs:t:T:uv") or usage();
  48usage if $opt_h;
  49
  50my $tag_name = $opt_t || "tags";
  51my $trunk_name = $opt_T || "trunk";
  52my $branch_name = $opt_b || "branches";
  53
  54@ARGV == 1 or @ARGV == 2 or usage();
  55
  56$opt_o ||= "origin";
  57$opt_s ||= 1;
  58my $git_tree = $opt_C;
  59$git_tree ||= ".";
  60
  61my $svn_url = $ARGV[0];
  62my $svn_dir = $ARGV[1];
  63
  64our @mergerx = ();
  65if ($opt_m) {
  66        @mergerx = ( qr/\W(?:from|of|merge|merging|merged) (\w+)/i );
  67}
  68if ($opt_M) {
  69        push (@mergerx, qr/$opt_M/);
  70}
  71
  72# Absolutize filename now, since we will have chdir'ed by the time we
  73# get around to opening it.
  74$opt_A = File::Spec->rel2abs($opt_A) if $opt_A;
  75
  76our %users = ();
  77our $users_file = undef;
  78sub read_users($) {
  79        $users_file = File::Spec->rel2abs(@_);
  80        die "Cannot open $users_file\n" unless -f $users_file;
  81        open(my $authors,$users_file);
  82        while(<$authors>) {
  83                chomp;
  84                next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/;
  85                (my $user,my $name,my $email) = ($1,$2,$3);
  86                $users{$user} = [$name,$email];
  87        }
  88        close($authors);
  89}
  90
  91select(STDERR); $|=1; select(STDOUT);
  92
  93
  94package SVNconn;
  95# Basic SVN connection.
  96# We're only interested in connecting and downloading, so ...
  97
  98use File::Spec;
  99use File::Temp qw(tempfile);
 100use POSIX qw(strftime dup2);
 101use Fcntl qw(SEEK_SET);
 102
 103sub new {
 104        my($what,$repo) = @_;
 105        $what=ref($what) if ref($what);
 106
 107        my $self = {};
 108        $self->{'buffer'} = "";
 109        bless($self,$what);
 110
 111        $repo =~ s#/+$##;
 112        $self->{'fullrep'} = $repo;
 113        $self->conn();
 114
 115        return $self;
 116}
 117
 118sub conn {
 119        my $self = shift;
 120        my $repo = $self->{'fullrep'};
 121        my $auth = SVN::Core::auth_open ([SVN::Client::get_simple_provider,
 122                          SVN::Client::get_ssl_server_trust_file_provider,
 123                          SVN::Client::get_username_provider]);
 124        my $s = SVN::Ra->new(url => $repo, auth => $auth);
 125        die "SVN connection to $repo: $!\n" unless defined $s;
 126        $self->{'svn'} = $s;
 127        $self->{'repo'} = $repo;
 128        $self->{'maxrev'} = $s->get_latest_revnum();
 129}
 130
 131sub file {
 132        my($self,$path,$rev) = @_;
 133
 134        my ($fh, $name) = tempfile('gitsvn.XXXXXX',
 135                    DIR => File::Spec->tmpdir(), UNLINK => 1);
 136
 137        print "... $rev $path ...\n" if $opt_v;
 138        my (undef, $properties);
 139        my $pool = SVN::Pool->new();
 140        eval { (undef, $properties)
 141                   = $self->{'svn'}->get_file($path,$rev,$fh,$pool); };
 142        $pool->clear;
 143        if($@) {
 144                return undef if $@ =~ /Attempted to get checksum/;
 145                die $@;
 146        }
 147        my $mode;
 148        if (exists $properties->{'svn:executable'}) {
 149                $mode = '100755';
 150        } elsif (exists $properties->{'svn:special'}) {
 151                my ($special_content, $filesize);
 152                $filesize = tell $fh;
 153                seek $fh, 0, SEEK_SET;
 154                read $fh, $special_content, $filesize;
 155                if ($special_content =~ s/^link //) {
 156                        $mode = '120000';
 157                        seek $fh, 0, SEEK_SET;
 158                        truncate $fh, 0;
 159                        print $fh $special_content;
 160                } else {
 161                        die "unexpected svn:special file encountered";
 162                }
 163        } else {
 164                $mode = '100644';
 165        }
 166        close ($fh);
 167
 168        return ($name, $mode);
 169}
 170
 171sub ignore {
 172        my($self,$path,$rev) = @_;
 173
 174        print "... $rev $path ...\n" if $opt_v;
 175        my (undef,undef,$properties)
 176            = $self->{'svn'}->get_dir($path,$rev,undef);
 177        if (exists $properties->{'svn:ignore'}) {
 178                my ($fh, $name) = tempfile('gitsvn.XXXXXX',
 179                                           DIR => File::Spec->tmpdir(),
 180                                           UNLINK => 1);
 181                print $fh $properties->{'svn:ignore'};
 182                close($fh);
 183                return $name;
 184        } else {
 185                return undef;
 186        }
 187}
 188
 189package main;
 190use URI;
 191
 192our $svn = $svn_url;
 193$svn .= "/$svn_dir" if defined $svn_dir;
 194my $svn2 = SVNconn->new($svn);
 195$svn = SVNconn->new($svn);
 196
 197my $lwp_ua;
 198if($opt_d or $opt_D) {
 199        $svn_url = URI->new($svn_url)->canonical;
 200        if($opt_D) {
 201                $svn_dir =~ s#/*$#/#;
 202        } else {
 203                $svn_dir = "";
 204        }
 205        if ($svn_url->scheme eq "http") {
 206                use LWP::UserAgent;
 207                $lwp_ua = LWP::UserAgent->new(keep_alive => 1, requests_redirectable => []);
 208        } else {
 209                print STDERR "Warning: not HTTP; turning off direct file access\n";
 210                $opt_d=0;
 211        }
 212}
 213
 214sub pdate($) {
 215        my($d) = @_;
 216        $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
 217                or die "Unparseable date: $d\n";
 218        my $y=$1; $y-=1900 if $y>1900;
 219        return timegm($6||0,$5,$4,$3,$2-1,$y);
 220}
 221
 222sub getwd() {
 223        my $pwd = `pwd`;
 224        chomp $pwd;
 225        return $pwd;
 226}
 227
 228
 229sub get_headref($$) {
 230    my $name    = shift;
 231    my $git_dir = shift;
 232    my $sha;
 233
 234    if (open(C,"$git_dir/refs/heads/$name")) {
 235        chomp($sha = <C>);
 236        close(C);
 237        length($sha) == 40
 238            or die "Cannot get head id for $name ($sha): $!\n";
 239    }
 240    return $sha;
 241}
 242
 243
 244-d $git_tree
 245        or mkdir($git_tree,0777)
 246        or die "Could not create $git_tree: $!";
 247chdir($git_tree);
 248
 249my $orig_branch = "";
 250my $forward_master = 0;
 251my %branches;
 252
 253my $git_dir = $ENV{"GIT_DIR"} || ".git";
 254$git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
 255$ENV{"GIT_DIR"} = $git_dir;
 256my $orig_git_index;
 257$orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
 258my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
 259                                    DIR => File::Spec->tmpdir());
 260close ($git_ih);
 261$ENV{GIT_INDEX_FILE} = $git_index;
 262my $maxnum = 0;
 263my $last_rev = "";
 264my $last_branch;
 265my $current_rev = $opt_s || 1;
 266unless(-d $git_dir) {
 267        system("git-init-db");
 268        die "Cannot init the GIT db at $git_tree: $?\n" if $?;
 269        system("git-read-tree");
 270        die "Cannot init an empty tree: $?\n" if $?;
 271
 272        $last_branch = $opt_o;
 273        $orig_branch = "";
 274} else {
 275        -f "$git_dir/refs/heads/$opt_o"
 276                or die "Branch '$opt_o' does not exist.\n".
 277                       "Either use the correct '-o branch' option,\n".
 278                       "or import to a new repository.\n";
 279
 280        -f "$git_dir/svn2git"
 281                or die "'$git_dir/svn2git' does not exist.\n".
 282                       "You need that file for incremental imports.\n";
 283        open(F, "git-symbolic-ref HEAD |") or
 284                die "Cannot run git-symbolic-ref: $!\n";
 285        chomp ($last_branch = <F>);
 286        $last_branch = basename($last_branch);
 287        close(F);
 288        unless($last_branch) {
 289                warn "Cannot read the last branch name: $! -- assuming 'master'\n";
 290                $last_branch = "master";
 291        }
 292        $orig_branch = $last_branch;
 293        $last_rev = get_headref($orig_branch, $git_dir);
 294        if (-f "$git_dir/SVN2GIT_HEAD") {
 295                die <<EOM;
 296SVN2GIT_HEAD exists.
 297Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
 298You may need to run
 299
 300    git-read-tree -m -u SVN2GIT_HEAD HEAD
 301EOM
 302        }
 303        system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
 304
 305        $forward_master =
 306            $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
 307            system('cmp', '-s', "$git_dir/refs/heads/master",
 308                                "$git_dir/refs/heads/$opt_o") == 0;
 309
 310        # populate index
 311        system('git-read-tree', $last_rev);
 312        die "read-tree failed: $?\n" if $?;
 313
 314        # Get the last import timestamps
 315        open my $B,"<", "$git_dir/svn2git";
 316        while(<$B>) {
 317                chomp;
 318                my($num,$branch,$ref) = split;
 319                $branches{$branch}{$num} = $ref;
 320                $branches{$branch}{"LAST"} = $ref;
 321                $current_rev = $num+1 if $current_rev <= $num;
 322        }
 323        close($B);
 324}
 325-d $git_dir
 326        or die "Could not create git subdir ($git_dir).\n";
 327
 328my $default_authors = "$git_dir/svn-authors";
 329if ($opt_A) {
 330        read_users($opt_A);
 331        copy($opt_A,$default_authors) or die "Copy failed: $!";
 332} else {
 333        read_users($default_authors) if -f $default_authors;
 334}
 335
 336open BRANCHES,">>", "$git_dir/svn2git";
 337
 338sub node_kind($$$) {
 339        my ($branch, $path, $revision) = @_;
 340        my $pool=SVN::Pool->new;
 341        my $kind = $svn->{'svn'}->check_path(revert_split_path($branch,$path),$revision,$pool);
 342        $pool->clear;
 343        return $kind;
 344}
 345
 346sub revert_split_path($$) {
 347        my($branch,$path) = @_;
 348
 349        my $svnpath;
 350        $path = "" if $path eq "/"; # this should not happen, but ...
 351        if($branch eq "/") {
 352                $svnpath = "$trunk_name/$path";
 353        } elsif($branch =~ m#^/#) {
 354                $svnpath = "$tag_name$branch/$path";
 355        } else {
 356                $svnpath = "$branch_name/$branch/$path";
 357        }
 358
 359        $svnpath =~ s#/+$##;
 360        return $svnpath;
 361}
 362
 363sub get_file($$$) {
 364        my($rev,$branch,$path) = @_;
 365
 366        my $svnpath = revert_split_path($branch,$path);
 367
 368        # now get it
 369        my ($name,$mode);
 370        if($opt_d) {
 371                my($req,$res);
 372
 373                # /svn/!svn/bc/2/django/trunk/django-docs/build.py
 374                my $url=$svn_url->clone();
 375                $url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath");
 376                print "... $path...\n" if $opt_v;
 377                $req = HTTP::Request->new(GET => $url);
 378                $res = $lwp_ua->request($req);
 379                if ($res->is_success) {
 380                        my $fh;
 381                        ($fh, $name) = tempfile('gitsvn.XXXXXX',
 382                        DIR => File::Spec->tmpdir(), UNLINK => 1);
 383                        print $fh $res->content;
 384                        close($fh) or die "Could not write $name: $!\n";
 385                } else {
 386                        return undef if $res->code == 301; # directory?
 387                        die $res->status_line." at $url\n";
 388                }
 389                $mode = '0644'; # can't obtain mode via direct http request?
 390        } else {
 391                ($name,$mode) = $svn->file("$svnpath",$rev);
 392                return undef unless defined $name;
 393        }
 394
 395        my $pid = open(my $F, '-|');
 396        die $! unless defined $pid;
 397        if (!$pid) {
 398            exec("git-hash-object", "-w", $name)
 399                or die "Cannot create object: $!\n";
 400        }
 401        my $sha = <$F>;
 402        chomp $sha;
 403        close $F;
 404        unlink $name;
 405        return [$mode, $sha, $path];
 406}
 407
 408sub get_ignore($$$$$) {
 409        my($new,$old,$rev,$branch,$path) = @_;
 410
 411        return unless $opt_I;
 412        my $svnpath = revert_split_path($branch,$path);
 413        my $name = $svn->ignore("$svnpath",$rev);
 414        if ($path eq '/') {
 415                $path = $opt_I;
 416        } else {
 417                $path = File::Spec->catfile($path,$opt_I);
 418        }
 419        if (defined $name) {
 420                my $pid = open(my $F, '-|');
 421                die $! unless defined $pid;
 422                if (!$pid) {
 423                        exec("git-hash-object", "-w", $name)
 424                            or die "Cannot create object: $!\n";
 425                }
 426                my $sha = <$F>;
 427                chomp $sha;
 428                close $F;
 429                unlink $name;
 430                push(@$new,['0644',$sha,$path]);
 431        } else {
 432                push(@$old,$path);
 433        }
 434}
 435
 436sub split_path($$) {
 437        my($rev,$path) = @_;
 438        my $branch;
 439
 440        if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
 441                $branch = "/$1";
 442        } elsif($path =~ s#^/\Q$trunk_name\E/?##) {
 443                $branch = "/";
 444        } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
 445                $branch = $1;
 446        } else {
 447                my %no_error = (
 448                        "/" => 1,
 449                        "/$tag_name" => 1,
 450                        "/$branch_name" => 1
 451                );
 452                print STDERR "$rev: Unrecognized path: $path\n" unless (defined $no_error{$path});
 453                return ()
 454        }
 455        $path = "/" if $path eq "";
 456        return ($branch,$path);
 457}
 458
 459sub branch_rev($$) {
 460
 461        my ($srcbranch,$uptorev) = @_;
 462
 463        my $bbranches = $branches{$srcbranch};
 464        my @revs = reverse sort { ($a eq 'LAST' ? 0 : $a) <=> ($b eq 'LAST' ? 0 : $b) } keys %$bbranches;
 465        my $therev;
 466        foreach my $arev(@revs) {
 467                next if  ($arev eq 'LAST');
 468                if ($arev <= $uptorev) {
 469                        $therev = $arev;
 470                        last;
 471                }
 472        }
 473        return $therev;
 474}
 475
 476sub copy_path($$$$$$$$) {
 477        # Somebody copied a whole subdirectory.
 478        # We need to find the index entries from the old version which the
 479        # SVN log entry points to, and add them to the new place.
 480
 481        my($newrev,$newbranch,$path,$oldpath,$rev,$node_kind,$new,$parents) = @_;
 482
 483        my($srcbranch,$srcpath) = split_path($rev,$oldpath);
 484        unless(defined $srcbranch) {
 485                print "Path not found when copying from $oldpath @ $rev\n";
 486                return;
 487        }
 488        my $therev = branch_rev($srcbranch, $rev);
 489        my $gitrev = $branches{$srcbranch}{$therev};
 490        unless($gitrev) {
 491                print STDERR "$newrev:$newbranch: could not find $oldpath \@ $rev\n";
 492                return;
 493        }
 494        if ($srcbranch ne $newbranch) {
 495                push(@$parents, $branches{$srcbranch}{'LAST'});
 496        }
 497        print "$newrev:$newbranch:$path: copying from $srcbranch:$srcpath @ $rev\n" if $opt_v;
 498        if ($node_kind eq $SVN::Node::dir) {
 499                        $srcpath =~ s#/*$#/#;
 500        }
 501        
 502        my $pid = open my $f,'-|';
 503        die $! unless defined $pid;
 504        if (!$pid) {
 505                exec("git-ls-tree","-r","-z",$gitrev,$srcpath)
 506                        or die $!;
 507        }
 508        local $/ = "\0";
 509        while(<$f>) {
 510                chomp;
 511                my($m,$p) = split(/\t/,$_,2);
 512                my($mode,$type,$sha1) = split(/ /,$m);
 513                next if $type ne "blob";
 514                if ($node_kind eq $SVN::Node::dir) {
 515                        $p = $path . substr($p,length($srcpath)-1);
 516                } else {
 517                        $p = $path;
 518                }
 519                push(@$new,[$mode,$sha1,$p]);   
 520        }
 521        close($f) or
 522                print STDERR "$newrev:$newbranch: could not list files in $oldpath \@ $rev\n";
 523}
 524
 525sub commit {
 526        my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
 527        my($author_name,$author_email,$dest);
 528        my(@old,@new,@parents);
 529
 530        if (not defined $author) {
 531                $author_name = $author_email = "unknown";
 532        } elsif (defined $users_file) {
 533                die "User $author is not listed in $users_file\n"
 534                    unless exists $users{$author};
 535                ($author_name,$author_email) = @{$users{$author}};
 536        } elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
 537                ($author_name, $author_email) = ($1, $2);
 538        } else {
 539                $author =~ s/^<(.*)>$/$1/;
 540                $author_name = $author_email = $author;
 541        }
 542        $date = pdate($date);
 543
 544        my $tag;
 545        my $parent;
 546        if($branch eq "/") { # trunk
 547                $parent = $opt_o;
 548        } elsif($branch =~ m#^/(.+)#) { # tag
 549                $tag = 1;
 550                $parent = $1;
 551        } else { # "normal" branch
 552                # nothing to do
 553                $parent = $branch;
 554        }
 555        $dest = $parent;
 556
 557        my $prev = $changed_paths->{"/"};
 558        if($prev and $prev->[0] eq "A") {
 559                delete $changed_paths->{"/"};
 560                my $oldpath = $prev->[1];
 561                my $rev;
 562                if(defined $oldpath) {
 563                        my $p;
 564                        ($parent,$p) = split_path($revision,$oldpath);
 565                        if($parent eq "/") {
 566                                $parent = $opt_o;
 567                        } else {
 568                                $parent =~ s#^/##; # if it's a tag
 569                        }
 570                } else {
 571                        $parent = undef;
 572                }
 573        }
 574
 575        my $rev;
 576        if($revision > $opt_s and defined $parent) {
 577                open(H,"git-rev-parse --verify $parent |");
 578                $rev = <H>;
 579                close(H) or do {
 580                        print STDERR "$revision: cannot find commit '$parent'!\n";
 581                        return;
 582                };
 583                chop $rev;
 584                if(length($rev) != 40) {
 585                        print STDERR "$revision: cannot find commit '$parent'!\n";
 586                        return;
 587                }
 588                $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
 589                if($revision != $opt_s and not $rev) {
 590                        print STDERR "$revision: do not know ancestor for '$parent'!\n";
 591                        return;
 592                }
 593        } else {
 594                $rev = undef;
 595        }
 596
 597#       if($prev and $prev->[0] eq "A") {
 598#               if(not $tag) {
 599#                       unless(open(H,"> $git_dir/refs/heads/$branch")) {
 600#                               print STDERR "$revision: Could not create branch $branch: $!\n";
 601#                               $state=11;
 602#                               next;
 603#                       }
 604#                       print H "$rev\n"
 605#                               or die "Could not write branch $branch: $!";
 606#                       close(H)
 607#                               or die "Could not write branch $branch: $!";
 608#               }
 609#       }
 610        if(not defined $rev) {
 611                unlink($git_index);
 612        } elsif ($rev ne $last_rev) {
 613                print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
 614                system("git-read-tree", $rev);
 615                die "read-tree failed for $rev: $?\n" if $?;
 616                $last_rev = $rev;
 617        }
 618
 619        push (@parents, $rev) if defined $rev;
 620
 621        my $cid;
 622        if($tag and not %$changed_paths) {
 623                $cid = $rev;
 624        } else {
 625                my @paths = sort keys %$changed_paths;
 626                foreach my $path(@paths) {
 627                        my $action = $changed_paths->{$path};
 628
 629                        if ($action->[0] eq "R") {
 630                                # refer to a file/tree in an earlier commit
 631                                push(@old,$path); # remove any old stuff
 632                        }
 633                        if(($action->[0] eq "A") || ($action->[0] eq "R")) {
 634                                my $node_kind = node_kind($branch,$path,$revision);
 635                                if ($node_kind eq $SVN::Node::file) {
 636                                        my $f = get_file($revision,$branch,$path);
 637                                        if ($f) {
 638                                                push(@new,$f) if $f;
 639                                        } else {
 640                                                my $opath = $action->[3];
 641                                                print STDERR "$revision: $branch: could not fetch '$opath'\n";
 642                                        }
 643                                } elsif ($node_kind eq $SVN::Node::dir) {
 644                                        if($action->[1]) {
 645                                                copy_path($revision, $branch,
 646                                                          $path, $action->[1],
 647                                                          $action->[2], $node_kind,
 648                                                          \@new, \@parents);
 649                                        } else {
 650                                                get_ignore(\@new, \@old, $revision,
 651                                                           $branch, $path);
 652                                        }
 653                                }
 654                        } elsif ($action->[0] eq "D") {
 655                                push(@old,$path);
 656                        } elsif ($action->[0] eq "M") {
 657                                my $node_kind = node_kind($branch,$path,$revision);
 658                                if ($node_kind eq $SVN::Node::file) {
 659                                        my $f = get_file($revision,$branch,$path);
 660                                        push(@new,$f) if $f;
 661                                } elsif ($node_kind eq $SVN::Node::dir) {
 662                                        get_ignore(\@new, \@old, $revision,
 663                                                   $branch,$path);
 664                                }
 665                        } else {
 666                                die "$revision: unknown action '".$action->[0]."' for $path\n";
 667                        }
 668                }
 669
 670                while(@old) {
 671                        my @o1;
 672                        if(@old > 55) {
 673                                @o1 = splice(@old,0,50);
 674                        } else {
 675                                @o1 = @old;
 676                                @old = ();
 677                        }
 678                        my $pid = open my $F, "-|";
 679                        die "$!" unless defined $pid;
 680                        if (!$pid) {
 681                                exec("git-ls-files", "-z", @o1) or die $!;
 682                        }
 683                        @o1 = ();
 684                        local $/ = "\0";
 685                        while(<$F>) {
 686                                chomp;
 687                                push(@o1,$_);
 688                        }
 689                        close($F);
 690
 691                        while(@o1) {
 692                                my @o2;
 693                                if(@o1 > 55) {
 694                                        @o2 = splice(@o1,0,50);
 695                                } else {
 696                                        @o2 = @o1;
 697                                        @o1 = ();
 698                                }
 699                                system("git-update-index","--force-remove","--",@o2);
 700                                die "Cannot remove files: $?\n" if $?;
 701                        }
 702                }
 703                while(@new) {
 704                        my @n2;
 705                        if(@new > 12) {
 706                                @n2 = splice(@new,0,10);
 707                        } else {
 708                                @n2 = @new;
 709                                @new = ();
 710                        }
 711                        system("git-update-index","--add",
 712                                (map { ('--cacheinfo', @$_) } @n2));
 713                        die "Cannot add files: $?\n" if $?;
 714                }
 715
 716                my $pid = open(C,"-|");
 717                die "Cannot fork: $!" unless defined $pid;
 718                unless($pid) {
 719                        exec("git-write-tree");
 720                        die "Cannot exec git-write-tree: $!\n";
 721                }
 722                chomp(my $tree = <C>);
 723                length($tree) == 40
 724                        or die "Cannot get tree id ($tree): $!\n";
 725                close(C)
 726                        or die "Error running git-write-tree: $?\n";
 727                print "Tree ID $tree\n" if $opt_v;
 728
 729                my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
 730                my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
 731                $pid = fork();
 732                die "Fork: $!\n" unless defined $pid;
 733                unless($pid) {
 734                        $pr->writer();
 735                        $pw->reader();
 736                        open(OUT,">&STDOUT");
 737                        dup2($pw->fileno(),0);
 738                        dup2($pr->fileno(),1);
 739                        $pr->close();
 740                        $pw->close();
 741
 742                        my @par = ();
 743
 744                        # loose detection of merges
 745                        # based on the commit msg
 746                        foreach my $rx (@mergerx) {
 747                                if ($message =~ $rx) {
 748                                        my $mparent = $1;
 749                                        if ($mparent eq 'HEAD') { $mparent = $opt_o };
 750                                        if ( -e "$git_dir/refs/heads/$mparent") {
 751                                                $mparent = get_headref($mparent, $git_dir);
 752                                                push (@parents, $mparent);
 753                                                print OUT "Merge parent branch: $mparent\n" if $opt_v;
 754                                        }
 755                                }
 756                        }
 757                        my %seen_parents = ();
 758                        my @unique_parents = grep { ! $seen_parents{$_} ++ } @parents;
 759                        foreach my $bparent (@unique_parents) {
 760                                push @par, '-p', $bparent;
 761                                print OUT "Merge parent branch: $bparent\n" if $opt_v;
 762                        }
 763
 764                        exec("env",
 765                                "GIT_AUTHOR_NAME=$author_name",
 766                                "GIT_AUTHOR_EMAIL=$author_email",
 767                                "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
 768                                "GIT_COMMITTER_NAME=$author_name",
 769                                "GIT_COMMITTER_EMAIL=$author_email",
 770                                "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
 771                                "git-commit-tree", $tree,@par);
 772                        die "Cannot exec git-commit-tree: $!\n";
 773                }
 774                $pw->writer();
 775                $pr->reader();
 776
 777                $message =~ s/[\s\n]+\z//;
 778                $message = "r$revision: $message" if $opt_r;
 779
 780                print $pw "$message\n"
 781                        or die "Error writing to git-commit-tree: $!\n";
 782                $pw->close();
 783
 784                print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
 785                chomp($cid = <$pr>);
 786                length($cid) == 40
 787                        or die "Cannot get commit id ($cid): $!\n";
 788                print "Commit ID $cid\n" if $opt_v;
 789                $pr->close();
 790
 791                waitpid($pid,0);
 792                die "Error running git-commit-tree: $?\n" if $?;
 793        }
 794
 795        if (not defined $cid) {
 796                $cid = $branches{"/"}{"LAST"};
 797        }
 798
 799        if(not defined $dest) {
 800                print "... no known parent\n" if $opt_v;
 801        } elsif(not $tag) {
 802                print "Writing to refs/heads/$dest\n" if $opt_v;
 803                open(C,">$git_dir/refs/heads/$dest") and
 804                print C ("$cid\n") and
 805                close(C)
 806                        or die "Cannot write branch $dest for update: $!\n";
 807        }
 808
 809        if($tag) {
 810                my($in, $out) = ('','');
 811                $last_rev = "-" if %$changed_paths;
 812                # the tag was 'complex', i.e. did not refer to a "real" revision
 813
 814                $dest =~ tr/_/\./ if $opt_u;
 815                $branch = $dest;
 816
 817                my $pid = open2($in, $out, 'git-mktag');
 818                print $out ("object $cid\n".
 819                    "type commit\n".
 820                    "tag $dest\n".
 821                    "tagger $author_name <$author_email>\n") and
 822                close($out)
 823                    or die "Cannot create tag object $dest: $!\n";
 824
 825                my $tagobj = <$in>;
 826                chomp $tagobj;
 827
 828                if ( !close($in) or waitpid($pid, 0) != $pid or
 829                                $? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
 830                        die "Cannot create tag object $dest: $!\n";
 831                }
 832
 833                open(C,">$git_dir/refs/tags/$dest") and
 834                print C ("$tagobj\n") and
 835                close(C)
 836                        or die "Cannot create tag $branch: $!\n";
 837
 838                print "Created tag '$dest' on '$branch'\n" if $opt_v;
 839        }
 840        $branches{$branch}{"LAST"} = $cid;
 841        $branches{$branch}{$revision} = $cid;
 842        $last_rev = $cid;
 843        print BRANCHES "$revision $branch $cid\n";
 844        print "DONE: $revision $dest $cid\n" if $opt_v;
 845}
 846
 847sub commit_all {
 848        # Recursive use of the SVN connection does not work
 849        local $svn = $svn2;
 850
 851        my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
 852        my %p;
 853        while(my($path,$action) = each %$changed_paths) {
 854                $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev, $path ];
 855        }
 856        $changed_paths = \%p;
 857
 858        my %done;
 859        my @col;
 860        my $pref;
 861        my $branch;
 862
 863        while(my($path,$action) = each %$changed_paths) {
 864                ($branch,$path) = split_path($revision,$path);
 865                next if not defined $branch;
 866                $done{$branch}{$path} = $action;
 867        }
 868        while(($branch,$changed_paths) = each %done) {
 869                commit($branch, $changed_paths, $revision, $author, $date, $message);
 870        }
 871}
 872
 873$opt_l = $svn->{'maxrev'} if not defined $opt_l or $opt_l > $svn->{'maxrev'};
 874
 875if ($opt_l < $current_rev) {
 876    print "Up to date: no new revisions to fetch!\n" if $opt_v;
 877    unlink("$git_dir/SVN2GIT_HEAD");
 878    exit;
 879}
 880
 881print "Fetching from $current_rev to $opt_l ...\n" if $opt_v;
 882
 883my $pool=SVN::Pool->new;
 884$svn->{'svn'}->get_log("/",$current_rev,$opt_l,0,1,1,\&commit_all,$pool);
 885$pool->clear;
 886
 887
 888unlink($git_index);
 889
 890if (defined $orig_git_index) {
 891        $ENV{GIT_INDEX_FILE} = $orig_git_index;
 892} else {
 893        delete $ENV{GIT_INDEX_FILE};
 894}
 895
 896# Now switch back to the branch we were in before all of this happened
 897if($orig_branch) {
 898        print "DONE\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 899        system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 900                if $forward_master;
 901        unless ($opt_i) {
 902                system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
 903                die "read-tree failed: $?\n" if $?;
 904        }
 905} else {
 906        $orig_branch = "master";
 907        print "DONE; creating $orig_branch branch\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 908        system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 909                unless -f "$git_dir/refs/heads/master";
 910        system('git-update-ref', 'HEAD', "$orig_branch");
 911        unless ($opt_i) {
 912                system('git checkout');
 913                die "checkout failed: $?\n" if $?;
 914        }
 915}
 916unlink("$git_dir/SVN2GIT_HEAD");
 917close(BRANCHES);