Andrew's git / gitweb.git / git-svnimport.perl
?
git-svnimport.perlon commit Merge branch 'th/diff-extra' (f3c5b39)
   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        my $branch_esc = quotemeta ($branch_name);
  67        my $trunk_esc  = quotemeta ($trunk_name);
  68        @mergerx =
  69        (
  70                qr!\b(?:merg(?:ed?|ing))\b.*?\b((?:(?<=$branch_esc/)[\w\.\-]+)|(?:$trunk_esc))\b!i,
  71                qr!\b(?:from|of)\W+((?:(?<=$branch_esc/)[\w\.\-]+)|(?:$trunk_esc))\b!i,
  72                qr!\b(?:from|of)\W+(?:the )?([\w\.\-]+)[-\s]branch\b!i
  73        );
  74}
  75if ($opt_M) {
  76        unshift (@mergerx, qr/$opt_M/);
  77}
  78
  79# Absolutize filename now, since we will have chdir'ed by the time we
  80# get around to opening it.
  81$opt_A = File::Spec->rel2abs($opt_A) if $opt_A;
  82
  83our %users = ();
  84our $users_file = undef;
  85sub read_users($) {
  86        $users_file = File::Spec->rel2abs(@_);
  87        die "Cannot open $users_file\n" unless -f $users_file;
  88        open(my $authors,$users_file);
  89        while(<$authors>) {
  90                chomp;
  91                next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/;
  92                (my $user,my $name,my $email) = ($1,$2,$3);
  93                $users{$user} = [$name,$email];
  94        }
  95        close($authors);
  96}
  97
  98select(STDERR); $|=1; select(STDOUT);
  99
 100
 101package SVNconn;
 102# Basic SVN connection.
 103# We're only interested in connecting and downloading, so ...
 104
 105use File::Spec;
 106use File::Temp qw(tempfile);
 107use POSIX qw(strftime dup2);
 108use Fcntl qw(SEEK_SET);
 109
 110sub new {
 111        my($what,$repo) = @_;
 112        $what=ref($what) if ref($what);
 113
 114        my $self = {};
 115        $self->{'buffer'} = "";
 116        bless($self,$what);
 117
 118        $repo =~ s#/+$##;
 119        $self->{'fullrep'} = $repo;
 120        $self->conn();
 121
 122        return $self;
 123}
 124
 125sub conn {
 126        my $self = shift;
 127        my $repo = $self->{'fullrep'};
 128        my $auth = SVN::Core::auth_open ([SVN::Client::get_simple_provider,
 129                          SVN::Client::get_ssl_server_trust_file_provider,
 130                          SVN::Client::get_username_provider]);
 131        my $s = SVN::Ra->new(url => $repo, auth => $auth);
 132        die "SVN connection to $repo: $!\n" unless defined $s;
 133        $self->{'svn'} = $s;
 134        $self->{'repo'} = $repo;
 135        $self->{'maxrev'} = $s->get_latest_revnum();
 136}
 137
 138sub file {
 139        my($self,$path,$rev) = @_;
 140
 141        my ($fh, $name) = tempfile('gitsvn.XXXXXX',
 142                    DIR => File::Spec->tmpdir(), UNLINK => 1);
 143
 144        print "... $rev $path ...\n" if $opt_v;
 145        my (undef, $properties);
 146        my $pool = SVN::Pool->new();
 147        eval { (undef, $properties)
 148                   = $self->{'svn'}->get_file($path,$rev,$fh,$pool); };
 149        $pool->clear;
 150        if($@) {
 151                return undef if $@ =~ /Attempted to get checksum/;
 152                die $@;
 153        }
 154        my $mode;
 155        if (exists $properties->{'svn:executable'}) {
 156                $mode = '100755';
 157        } elsif (exists $properties->{'svn:special'}) {
 158                my ($special_content, $filesize);
 159                $filesize = tell $fh;
 160                seek $fh, 0, SEEK_SET;
 161                read $fh, $special_content, $filesize;
 162                if ($special_content =~ s/^link //) {
 163                        $mode = '120000';
 164                        seek $fh, 0, SEEK_SET;
 165                        truncate $fh, 0;
 166                        print $fh $special_content;
 167                } else {
 168                        die "unexpected svn:special file encountered";
 169                }
 170        } else {
 171                $mode = '100644';
 172        }
 173        close ($fh);
 174
 175        return ($name, $mode);
 176}
 177
 178sub ignore {
 179        my($self,$path,$rev) = @_;
 180
 181        print "... $rev $path ...\n" if $opt_v;
 182        my (undef,undef,$properties)
 183            = $self->{'svn'}->get_dir($path,$rev,undef);
 184        if (exists $properties->{'svn:ignore'}) {
 185                my ($fh, $name) = tempfile('gitsvn.XXXXXX',
 186                                           DIR => File::Spec->tmpdir(),
 187                                           UNLINK => 1);
 188                print $fh $properties->{'svn:ignore'};
 189                close($fh);
 190                return $name;
 191        } else {
 192                return undef;
 193        }
 194}
 195
 196package main;
 197use URI;
 198
 199our $svn = $svn_url;
 200$svn .= "/$svn_dir" if defined $svn_dir;
 201my $svn2 = SVNconn->new($svn);
 202$svn = SVNconn->new($svn);
 203
 204my $lwp_ua;
 205if($opt_d or $opt_D) {
 206        $svn_url = URI->new($svn_url)->canonical;
 207        if($opt_D) {
 208                $svn_dir =~ s#/*$#/#;
 209        } else {
 210                $svn_dir = "";
 211        }
 212        if ($svn_url->scheme eq "http") {
 213                use LWP::UserAgent;
 214                $lwp_ua = LWP::UserAgent->new(keep_alive => 1, requests_redirectable => []);
 215        } else {
 216                print STDERR "Warning: not HTTP; turning off direct file access\n";
 217                $opt_d=0;
 218        }
 219}
 220
 221sub pdate($) {
 222        my($d) = @_;
 223        $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
 224                or die "Unparseable date: $d\n";
 225        my $y=$1; $y-=1900 if $y>1900;
 226        return timegm($6||0,$5,$4,$3,$2-1,$y);
 227}
 228
 229sub getwd() {
 230        my $pwd = `pwd`;
 231        chomp $pwd;
 232        return $pwd;
 233}
 234
 235
 236sub get_headref($$) {
 237    my $name    = shift;
 238    my $git_dir = shift;
 239    my $sha;
 240
 241    if (open(C,"$git_dir/refs/heads/$name")) {
 242        chomp($sha = <C>);
 243        close(C);
 244        length($sha) == 40
 245            or die "Cannot get head id for $name ($sha): $!\n";
 246    }
 247    return $sha;
 248}
 249
 250
 251-d $git_tree
 252        or mkdir($git_tree,0777)
 253        or die "Could not create $git_tree: $!";
 254chdir($git_tree);
 255
 256my $orig_branch = "";
 257my $forward_master = 0;
 258my %branches;
 259
 260my $git_dir = $ENV{"GIT_DIR"} || ".git";
 261$git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
 262$ENV{"GIT_DIR"} = $git_dir;
 263my $orig_git_index;
 264$orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
 265my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
 266                                    DIR => File::Spec->tmpdir());
 267close ($git_ih);
 268$ENV{GIT_INDEX_FILE} = $git_index;
 269my $maxnum = 0;
 270my $last_rev = "";
 271my $last_branch;
 272my $current_rev = $opt_s || 1;
 273unless(-d $git_dir) {
 274        system("git-init-db");
 275        die "Cannot init the GIT db at $git_tree: $?\n" if $?;
 276        system("git-read-tree");
 277        die "Cannot init an empty tree: $?\n" if $?;
 278
 279        $last_branch = $opt_o;
 280        $orig_branch = "";
 281} else {
 282        -f "$git_dir/refs/heads/$opt_o"
 283                or die "Branch '$opt_o' does not exist.\n".
 284                       "Either use the correct '-o branch' option,\n".
 285                       "or import to a new repository.\n";
 286
 287        -f "$git_dir/svn2git"
 288                or die "'$git_dir/svn2git' does not exist.\n".
 289                       "You need that file for incremental imports.\n";
 290        open(F, "git-symbolic-ref HEAD |") or
 291                die "Cannot run git-symbolic-ref: $!\n";
 292        chomp ($last_branch = <F>);
 293        $last_branch = basename($last_branch);
 294        close(F);
 295        unless($last_branch) {
 296                warn "Cannot read the last branch name: $! -- assuming 'master'\n";
 297                $last_branch = "master";
 298        }
 299        $orig_branch = $last_branch;
 300        $last_rev = get_headref($orig_branch, $git_dir);
 301        if (-f "$git_dir/SVN2GIT_HEAD") {
 302                die <<EOM;
 303SVN2GIT_HEAD exists.
 304Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
 305You may need to run
 306
 307    git-read-tree -m -u SVN2GIT_HEAD HEAD
 308EOM
 309        }
 310        system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
 311
 312        $forward_master =
 313            $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
 314            system('cmp', '-s', "$git_dir/refs/heads/master",
 315                                "$git_dir/refs/heads/$opt_o") == 0;
 316
 317        # populate index
 318        system('git-read-tree', $last_rev);
 319        die "read-tree failed: $?\n" if $?;
 320
 321        # Get the last import timestamps
 322        open my $B,"<", "$git_dir/svn2git";
 323        while(<$B>) {
 324                chomp;
 325                my($num,$branch,$ref) = split;
 326                $branches{$branch}{$num} = $ref;
 327                $branches{$branch}{"LAST"} = $ref;
 328                $current_rev = $num+1 if $current_rev <= $num;
 329        }
 330        close($B);
 331}
 332-d $git_dir
 333        or die "Could not create git subdir ($git_dir).\n";
 334
 335my $default_authors = "$git_dir/svn-authors";
 336if ($opt_A) {
 337        read_users($opt_A);
 338        copy($opt_A,$default_authors) or die "Copy failed: $!";
 339} else {
 340        read_users($default_authors) if -f $default_authors;
 341}
 342
 343open BRANCHES,">>", "$git_dir/svn2git";
 344
 345sub node_kind($$$) {
 346        my ($branch, $path, $revision) = @_;
 347        my $pool=SVN::Pool->new;
 348        my $kind = $svn->{'svn'}->check_path(revert_split_path($branch,$path),$revision,$pool);
 349        $pool->clear;
 350        return $kind;
 351}
 352
 353sub revert_split_path($$) {
 354        my($branch,$path) = @_;
 355
 356        my $svnpath;
 357        $path = "" if $path eq "/"; # this should not happen, but ...
 358        if($branch eq "/") {
 359                $svnpath = "$trunk_name/$path";
 360        } elsif($branch =~ m#^/#) {
 361                $svnpath = "$tag_name$branch/$path";
 362        } else {
 363                $svnpath = "$branch_name/$branch/$path";
 364        }
 365
 366        $svnpath =~ s#/+$##;
 367        return $svnpath;
 368}
 369
 370sub get_file($$$) {
 371        my($rev,$branch,$path) = @_;
 372
 373        my $svnpath = revert_split_path($branch,$path);
 374
 375        # now get it
 376        my ($name,$mode);
 377        if($opt_d) {
 378                my($req,$res);
 379
 380                # /svn/!svn/bc/2/django/trunk/django-docs/build.py
 381                my $url=$svn_url->clone();
 382                $url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath");
 383                print "... $path...\n" if $opt_v;
 384                $req = HTTP::Request->new(GET => $url);
 385                $res = $lwp_ua->request($req);
 386                if ($res->is_success) {
 387                        my $fh;
 388                        ($fh, $name) = tempfile('gitsvn.XXXXXX',
 389                        DIR => File::Spec->tmpdir(), UNLINK => 1);
 390                        print $fh $res->content;
 391                        close($fh) or die "Could not write $name: $!\n";
 392                } else {
 393                        return undef if $res->code == 301; # directory?
 394                        die $res->status_line." at $url\n";
 395                }
 396                $mode = '0644'; # can't obtain mode via direct http request?
 397        } else {
 398                ($name,$mode) = $svn->file("$svnpath",$rev);
 399                return undef unless defined $name;
 400        }
 401
 402        my $pid = open(my $F, '-|');
 403        die $! unless defined $pid;
 404        if (!$pid) {
 405            exec("git-hash-object", "-w", $name)
 406                or die "Cannot create object: $!\n";
 407        }
 408        my $sha = <$F>;
 409        chomp $sha;
 410        close $F;
 411        unlink $name;
 412        return [$mode, $sha, $path];
 413}
 414
 415sub get_ignore($$$$$) {
 416        my($new,$old,$rev,$branch,$path) = @_;
 417
 418        return unless $opt_I;
 419        my $svnpath = revert_split_path($branch,$path);
 420        my $name = $svn->ignore("$svnpath",$rev);
 421        if ($path eq '/') {
 422                $path = $opt_I;
 423        } else {
 424                $path = File::Spec->catfile($path,$opt_I);
 425        }
 426        if (defined $name) {
 427                my $pid = open(my $F, '-|');
 428                die $! unless defined $pid;
 429                if (!$pid) {
 430                        exec("git-hash-object", "-w", $name)
 431                            or die "Cannot create object: $!\n";
 432                }
 433                my $sha = <$F>;
 434                chomp $sha;
 435                close $F;
 436                unlink $name;
 437                push(@$new,['0644',$sha,$path]);
 438        } else {
 439                push(@$old,$path);
 440        }
 441}
 442
 443sub split_path($$) {
 444        my($rev,$path) = @_;
 445        my $branch;
 446
 447        if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
 448                $branch = "/$1";
 449        } elsif($path =~ s#^/\Q$trunk_name\E/?##) {
 450                $branch = "/";
 451        } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
 452                $branch = $1;
 453        } else {
 454                my %no_error = (
 455                        "/" => 1,
 456                        "/$tag_name" => 1,
 457                        "/$branch_name" => 1
 458                );
 459                print STDERR "$rev: Unrecognized path: $path\n" unless (defined $no_error{$path});
 460                return ()
 461        }
 462        $path = "/" if $path eq "";
 463        return ($branch,$path);
 464}
 465
 466sub branch_rev($$) {
 467
 468        my ($srcbranch,$uptorev) = @_;
 469
 470        my $bbranches = $branches{$srcbranch};
 471        my @revs = reverse sort { ($a eq 'LAST' ? 0 : $a) <=> ($b eq 'LAST' ? 0 : $b) } keys %$bbranches;
 472        my $therev;
 473        foreach my $arev(@revs) {
 474                next if  ($arev eq 'LAST');
 475                if ($arev <= $uptorev) {
 476                        $therev = $arev;
 477                        last;
 478                }
 479        }
 480        return $therev;
 481}
 482
 483sub copy_path($$$$$$$$) {
 484        # Somebody copied a whole subdirectory.
 485        # We need to find the index entries from the old version which the
 486        # SVN log entry points to, and add them to the new place.
 487
 488        my($newrev,$newbranch,$path,$oldpath,$rev,$node_kind,$new,$parents) = @_;
 489
 490        my($srcbranch,$srcpath) = split_path($rev,$oldpath);
 491        unless(defined $srcbranch) {
 492                print "Path not found when copying from $oldpath @ $rev\n";
 493                return;
 494        }
 495        my $therev = branch_rev($srcbranch, $rev);
 496        my $gitrev = $branches{$srcbranch}{$therev};
 497        unless($gitrev) {
 498                print STDERR "$newrev:$newbranch: could not find $oldpath \@ $rev\n";
 499                return;
 500        }
 501        if ($srcbranch ne $newbranch) {
 502                push(@$parents, $branches{$srcbranch}{'LAST'});
 503        }
 504        print "$newrev:$newbranch:$path: copying from $srcbranch:$srcpath @ $rev\n" if $opt_v;
 505        if ($node_kind eq $SVN::Node::dir) {
 506                        $srcpath =~ s#/*$#/#;
 507        }
 508        
 509        my $pid = open my $f,'-|';
 510        die $! unless defined $pid;
 511        if (!$pid) {
 512                exec("git-ls-tree","-r","-z",$gitrev,$srcpath)
 513                        or die $!;
 514        }
 515        local $/ = "\0";
 516        while(<$f>) {
 517                chomp;
 518                my($m,$p) = split(/\t/,$_,2);
 519                my($mode,$type,$sha1) = split(/ /,$m);
 520                next if $type ne "blob";
 521                if ($node_kind eq $SVN::Node::dir) {
 522                        $p = $path . substr($p,length($srcpath)-1);
 523                } else {
 524                        $p = $path;
 525                }
 526                push(@$new,[$mode,$sha1,$p]);   
 527        }
 528        close($f) or
 529                print STDERR "$newrev:$newbranch: could not list files in $oldpath \@ $rev\n";
 530}
 531
 532sub commit {
 533        my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
 534        my($author_name,$author_email,$dest);
 535        my(@old,@new,@parents);
 536
 537        if (not defined $author or $author eq "") {
 538                $author_name = $author_email = "unknown";
 539        } elsif (defined $users_file) {
 540                die "User $author is not listed in $users_file\n"
 541                    unless exists $users{$author};
 542                ($author_name,$author_email) = @{$users{$author}};
 543        } elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
 544                ($author_name, $author_email) = ($1, $2);
 545        } else {
 546                $author =~ s/^<(.*)>$/$1/;
 547                $author_name = $author_email = $author;
 548        }
 549        $date = pdate($date);
 550
 551        my $tag;
 552        my $parent;
 553        if($branch eq "/") { # trunk
 554                $parent = $opt_o;
 555        } elsif($branch =~ m#^/(.+)#) { # tag
 556                $tag = 1;
 557                $parent = $1;
 558        } else { # "normal" branch
 559                # nothing to do
 560                $parent = $branch;
 561        }
 562        $dest = $parent;
 563
 564        my $prev = $changed_paths->{"/"};
 565        if($prev and $prev->[0] eq "A") {
 566                delete $changed_paths->{"/"};
 567                my $oldpath = $prev->[1];
 568                my $rev;
 569                if(defined $oldpath) {
 570                        my $p;
 571                        ($parent,$p) = split_path($revision,$oldpath);
 572                        if($parent eq "/") {
 573                                $parent = $opt_o;
 574                        } else {
 575                                $parent =~ s#^/##; # if it's a tag
 576                        }
 577                } else {
 578                        $parent = undef;
 579                }
 580        }
 581
 582        my $rev;
 583        if($revision > $opt_s and defined $parent) {
 584                open(H,"git-rev-parse --verify $parent |");
 585                $rev = <H>;
 586                close(H) or do {
 587                        print STDERR "$revision: cannot find commit '$parent'!\n";
 588                        return;
 589                };
 590                chop $rev;
 591                if(length($rev) != 40) {
 592                        print STDERR "$revision: cannot find commit '$parent'!\n";
 593                        return;
 594                }
 595                $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
 596                if($revision != $opt_s and not $rev) {
 597                        print STDERR "$revision: do not know ancestor for '$parent'!\n";
 598                        return;
 599                }
 600        } else {
 601                $rev = undef;
 602        }
 603
 604#       if($prev and $prev->[0] eq "A") {
 605#               if(not $tag) {
 606#                       unless(open(H,"> $git_dir/refs/heads/$branch")) {
 607#                               print STDERR "$revision: Could not create branch $branch: $!\n";
 608#                               $state=11;
 609#                               next;
 610#                       }
 611#                       print H "$rev\n"
 612#                               or die "Could not write branch $branch: $!";
 613#                       close(H)
 614#                               or die "Could not write branch $branch: $!";
 615#               }
 616#       }
 617        if(not defined $rev) {
 618                unlink($git_index);
 619        } elsif ($rev ne $last_rev) {
 620                print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
 621                system("git-read-tree", $rev);
 622                die "read-tree failed for $rev: $?\n" if $?;
 623                $last_rev = $rev;
 624        }
 625
 626        push (@parents, $rev) if defined $rev;
 627
 628        my $cid;
 629        if($tag and not %$changed_paths) {
 630                $cid = $rev;
 631        } else {
 632                my @paths = sort keys %$changed_paths;
 633                foreach my $path(@paths) {
 634                        my $action = $changed_paths->{$path};
 635
 636                        if ($action->[0] eq "R") {
 637                                # refer to a file/tree in an earlier commit
 638                                push(@old,$path); # remove any old stuff
 639                        }
 640                        if(($action->[0] eq "A") || ($action->[0] eq "R")) {
 641                                my $node_kind = node_kind($branch,$path,$revision);
 642                                if ($node_kind eq $SVN::Node::file) {
 643                                        my $f = get_file($revision,$branch,$path);
 644                                        if ($f) {
 645                                                push(@new,$f) if $f;
 646                                        } else {
 647                                                my $opath = $action->[3];
 648                                                print STDERR "$revision: $branch: could not fetch '$opath'\n";
 649                                        }
 650                                } elsif ($node_kind eq $SVN::Node::dir) {
 651                                        if($action->[1]) {
 652                                                copy_path($revision, $branch,
 653                                                          $path, $action->[1],
 654                                                          $action->[2], $node_kind,
 655                                                          \@new, \@parents);
 656                                        } else {
 657                                                get_ignore(\@new, \@old, $revision,
 658                                                           $branch, $path);
 659                                        }
 660                                }
 661                        } elsif ($action->[0] eq "D") {
 662                                push(@old,$path);
 663                        } elsif ($action->[0] eq "M") {
 664                                my $node_kind = node_kind($branch,$path,$revision);
 665                                if ($node_kind eq $SVN::Node::file) {
 666                                        my $f = get_file($revision,$branch,$path);
 667                                        push(@new,$f) if $f;
 668                                } elsif ($node_kind eq $SVN::Node::dir) {
 669                                        get_ignore(\@new, \@old, $revision,
 670                                                   $branch,$path);
 671                                }
 672                        } else {
 673                                die "$revision: unknown action '".$action->[0]."' for $path\n";
 674                        }
 675                }
 676
 677                while(@old) {
 678                        my @o1;
 679                        if(@old > 55) {
 680                                @o1 = splice(@old,0,50);
 681                        } else {
 682                                @o1 = @old;
 683                                @old = ();
 684                        }
 685                        my $pid = open my $F, "-|";
 686                        die "$!" unless defined $pid;
 687                        if (!$pid) {
 688                                exec("git-ls-files", "-z", @o1) or die $!;
 689                        }
 690                        @o1 = ();
 691                        local $/ = "\0";
 692                        while(<$F>) {
 693                                chomp;
 694                                push(@o1,$_);
 695                        }
 696                        close($F);
 697
 698                        while(@o1) {
 699                                my @o2;
 700                                if(@o1 > 55) {
 701                                        @o2 = splice(@o1,0,50);
 702                                } else {
 703                                        @o2 = @o1;
 704                                        @o1 = ();
 705                                }
 706                                system("git-update-index","--force-remove","--",@o2);
 707                                die "Cannot remove files: $?\n" if $?;
 708                        }
 709                }
 710                while(@new) {
 711                        my @n2;
 712                        if(@new > 12) {
 713                                @n2 = splice(@new,0,10);
 714                        } else {
 715                                @n2 = @new;
 716                                @new = ();
 717                        }
 718                        system("git-update-index","--add",
 719                                (map { ('--cacheinfo', @$_) } @n2));
 720                        die "Cannot add files: $?\n" if $?;
 721                }
 722
 723                my $pid = open(C,"-|");
 724                die "Cannot fork: $!" unless defined $pid;
 725                unless($pid) {
 726                        exec("git-write-tree");
 727                        die "Cannot exec git-write-tree: $!\n";
 728                }
 729                chomp(my $tree = <C>);
 730                length($tree) == 40
 731                        or die "Cannot get tree id ($tree): $!\n";
 732                close(C)
 733                        or die "Error running git-write-tree: $?\n";
 734                print "Tree ID $tree\n" if $opt_v;
 735
 736                my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
 737                my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
 738                $pid = fork();
 739                die "Fork: $!\n" unless defined $pid;
 740                unless($pid) {
 741                        $pr->writer();
 742                        $pw->reader();
 743                        open(OUT,">&STDOUT");
 744                        dup2($pw->fileno(),0);
 745                        dup2($pr->fileno(),1);
 746                        $pr->close();
 747                        $pw->close();
 748
 749                        my @par = ();
 750
 751                        # loose detection of merges
 752                        # based on the commit msg
 753                        foreach my $rx (@mergerx) {
 754                                if ($message =~ $rx) {
 755                                        my $mparent = $1;
 756                                        if ($mparent eq 'HEAD') { $mparent = $opt_o };
 757                                        if ( -e "$git_dir/refs/heads/$mparent") {
 758                                                $mparent = get_headref($mparent, $git_dir);
 759                                                push (@parents, $mparent);
 760                                                print OUT "Merge parent branch: $mparent\n" if $opt_v;
 761                                        }
 762                                }
 763                        }
 764                        my %seen_parents = ();
 765                        my @unique_parents = grep { ! $seen_parents{$_} ++ } @parents;
 766                        foreach my $bparent (@unique_parents) {
 767                                push @par, '-p', $bparent;
 768                                print OUT "Merge parent branch: $bparent\n" if $opt_v;
 769                        }
 770
 771                        exec("env",
 772                                "GIT_AUTHOR_NAME=$author_name",
 773                                "GIT_AUTHOR_EMAIL=$author_email",
 774                                "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
 775                                "GIT_COMMITTER_NAME=$author_name",
 776                                "GIT_COMMITTER_EMAIL=$author_email",
 777                                "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
 778                                "git-commit-tree", $tree,@par);
 779                        die "Cannot exec git-commit-tree: $!\n";
 780                }
 781                $pw->writer();
 782                $pr->reader();
 783
 784                $message =~ s/[\s\n]+\z//;
 785                $message = "r$revision: $message" if $opt_r;
 786
 787                print $pw "$message\n"
 788                        or die "Error writing to git-commit-tree: $!\n";
 789                $pw->close();
 790
 791                print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
 792                chomp($cid = <$pr>);
 793                length($cid) == 40
 794                        or die "Cannot get commit id ($cid): $!\n";
 795                print "Commit ID $cid\n" if $opt_v;
 796                $pr->close();
 797
 798                waitpid($pid,0);
 799                die "Error running git-commit-tree: $?\n" if $?;
 800        }
 801
 802        if (not defined $cid) {
 803                $cid = $branches{"/"}{"LAST"};
 804        }
 805
 806        if(not defined $dest) {
 807                print "... no known parent\n" if $opt_v;
 808        } elsif(not $tag) {
 809                print "Writing to refs/heads/$dest\n" if $opt_v;
 810                open(C,">$git_dir/refs/heads/$dest") and
 811                print C ("$cid\n") and
 812                close(C)
 813                        or die "Cannot write branch $dest for update: $!\n";
 814        }
 815
 816        if($tag) {
 817                my($in, $out) = ('','');
 818                $last_rev = "-" if %$changed_paths;
 819                # the tag was 'complex', i.e. did not refer to a "real" revision
 820
 821                $dest =~ tr/_/\./ if $opt_u;
 822                $branch = $dest;
 823
 824                my $pid = open2($in, $out, 'git-mktag');
 825                print $out ("object $cid\n".
 826                    "type commit\n".
 827                    "tag $dest\n".
 828                    "tagger $author_name <$author_email>\n") and
 829                close($out)
 830                    or die "Cannot create tag object $dest: $!\n";
 831
 832                my $tagobj = <$in>;
 833                chomp $tagobj;
 834
 835                if ( !close($in) or waitpid($pid, 0) != $pid or
 836                                $? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
 837                        die "Cannot create tag object $dest: $!\n";
 838                }
 839
 840                open(C,">$git_dir/refs/tags/$dest") and
 841                print C ("$tagobj\n") and
 842                close(C)
 843                        or die "Cannot create tag $branch: $!\n";
 844
 845                print "Created tag '$dest' on '$branch'\n" if $opt_v;
 846        }
 847        $branches{$branch}{"LAST"} = $cid;
 848        $branches{$branch}{$revision} = $cid;
 849        $last_rev = $cid;
 850        print BRANCHES "$revision $branch $cid\n";
 851        print "DONE: $revision $dest $cid\n" if $opt_v;
 852}
 853
 854sub commit_all {
 855        # Recursive use of the SVN connection does not work
 856        local $svn = $svn2;
 857
 858        my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
 859        my %p;
 860        while(my($path,$action) = each %$changed_paths) {
 861                $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev, $path ];
 862        }
 863        $changed_paths = \%p;
 864
 865        my %done;
 866        my @col;
 867        my $pref;
 868        my $branch;
 869
 870        while(my($path,$action) = each %$changed_paths) {
 871                ($branch,$path) = split_path($revision,$path);
 872                next if not defined $branch;
 873                $done{$branch}{$path} = $action;
 874        }
 875        while(($branch,$changed_paths) = each %done) {
 876                commit($branch, $changed_paths, $revision, $author, $date, $message);
 877        }
 878}
 879
 880$opt_l = $svn->{'maxrev'} if not defined $opt_l or $opt_l > $svn->{'maxrev'};
 881
 882if ($opt_l < $current_rev) {
 883    print "Up to date: no new revisions to fetch!\n" if $opt_v;
 884    unlink("$git_dir/SVN2GIT_HEAD");
 885    exit;
 886}
 887
 888print "Fetching from $current_rev to $opt_l ...\n" if $opt_v;
 889
 890my $pool=SVN::Pool->new;
 891$svn->{'svn'}->get_log("/",$current_rev,$opt_l,0,1,1,\&commit_all,$pool);
 892$pool->clear;
 893
 894
 895unlink($git_index);
 896
 897if (defined $orig_git_index) {
 898        $ENV{GIT_INDEX_FILE} = $orig_git_index;
 899} else {
 900        delete $ENV{GIT_INDEX_FILE};
 901}
 902
 903# Now switch back to the branch we were in before all of this happened
 904if($orig_branch) {
 905        print "DONE\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 906        system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 907                if $forward_master;
 908        unless ($opt_i) {
 909                system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
 910                die "read-tree failed: $?\n" if $?;
 911        }
 912} else {
 913        $orig_branch = "master";
 914        print "DONE; creating $orig_branch branch\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 915        system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 916                unless -f "$git_dir/refs/heads/master";
 917        system('git-update-ref', 'HEAD', "$orig_branch");
 918        unless ($opt_i) {
 919                system('git checkout');
 920                die "checkout failed: $?\n" if $?;
 921        }
 922}
 923unlink("$git_dir/SVN2GIT_HEAD");
 924close(BRANCHES);