From 89764f5d8b0ca8600e3a200e9f863c3c7a3ff5e8 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 11:15:09 +0200
Subject: cvsimport: report merge parents

Matching and reporting merge parents happens in a subprocess.
Re-open stdout before redirecting stdout to the pipe, so that printing
verbose messages doesn't go to the wrong place.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-cvsimport.perl b/git-cvsimport.perl
index f35c0d0..cc0eed2 100755
--- a/git-cvsimport.perl
+++ b/git-cvsimport.perl
@@ -567,6 +567,7 @@ my $commit = sub {
 	unless($pid) {
 		$pr->writer();
 		$pw->reader();
+		open(OUT,">&STDOUT");
 		dup2($pw->fileno(),0);
 		dup2($pr->fileno(),1);
 		$pr->close();
@@ -584,10 +585,9 @@ my $commit = sub {
 				if ( -e "$git_dir/refs/heads/$mparent") {
 					$mparent = get_headref($mparent, $git_dir);
 					push @par, '-p', $mparent;
-					# printing here breaks import # 
-					# # print "Merge parent branch: $mparent\n" if $opt_v;
+					print OUT "Merge parent branch: $mparent\n" if $opt_v;
 				}
-		    	} 
+			} 
 		}
 
 		exec("env",
-- 
cgit v0.10.2-6-g49f6


From eaf718f3ece277462de4e47391e5a965bbbaa297 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 11:40:43 +0200
Subject: New: git-svnimport.

As the name suggests, this script imports from SVN.

Only "normal" SVN repositories (with single trunk/, branches/, and tags/
subdrectories) are supported. Incremental imports require preserving
the file .git/svn2git.

Signed-Off-by: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/Documentation/git-svnimport.txt b/Documentation/git-svnimport.txt
new file mode 100644
index 0000000..be03a65
--- /dev/null
+++ b/Documentation/git-svnimport.txt
@@ -0,0 +1,99 @@
+git-svnimport(1)
+================
+v0.1, July 2005
+
+NAME
+----
+git-svnimport - Import a SVN repository into git
+
+
+SYNOPSIS
+--------
+'git-svnimport' [ -o <branch-for-HEAD> ] [ -h ] [ -v ]
+			[ -C <GIT_repository> ] [ -i ] [ -u ]
+			[ -b branch_subdir ] [ -t trunk_subdir ] [ -T tag_subdir ]
+			[ -m ] [ -M regex ] [ <SVN_repository_URL> ]
+
+
+DESCRIPTION
+-----------
+Imports a SVN repository into git. It will either create a new
+repository, or incrementally import into an existing one.
+
+SVN access is done by the SVN:: Perl module.
+
+git-svnimport assumes that SVN repositories are organized into one
+"trunk" directory where the main development happens, "branch/FOO"
+directories for branches, and "/tags/FOO" directories for tags.
+Other subdirectories are ignored.
+
+git-svnimport creates a file ".git/svn2git", which is required for
+incremental SVN imports. 
+
+OPTIONS
+-------
+-C <target-dir>::
+        The GIT repository to import to.  If the directory doesn't
+        exist, it will be created.  Default is the current directory.
+
+-i::
+	Import-only: don't perform a checkout after importing.  This option
+	ensures the working directory and cache remain untouched and will
+	not create them if they do not exist.
+
+-t <trunk_subdir>::
+	Name the SVN trunk. Default "trunk".
+	
+-T <tag_subdir>::
+	Name the SVN subdirectory for tags. Default "tags".
+	
+-b <branch_subdir>::
+	Name the SVN subdirectory for branches. Default "branches".
+	
+-o <branch-for-HEAD>::
+	The 'trunk' branch from SVN is imported to the 'origin' branch within
+	the git repository. Use this option if you want to import into a
+	different branch.
+
+-m::    
+	Attempt to detect merges based on the commit message. This option
+	will enable default regexes that try to capture the name source 
+	branch name from the commit message. 
+
+-M <regex>::
+	Attempt to detect merges based on the commit message with a custom
+	regex. It can be used with -m to also see the default regexes. 
+	You must escape forward slashes. 
+
+-v::
+	Verbosity: let 'svnimport' report what it is doing.
+
+<SVN_repository_URL>::
+	The URL of the SVN module you want to import. For local
+	repositories, use "file:///absolute/path".
+
+-h::
+	Print a short usage message and exit.
+
+OUTPUT
+------
+If '-v' is specified, the script reports what it is doing.
+
+Otherwise, success is indicated the Unix way, i.e. by simply exiting with
+a zero exit status.
+
+Author
+------
+Written by Matthias Urlichs <smurf@smurf.noris.de>, with help from
+various participants of the git-list <git@vger.kernel.org>.
+
+Based on a cvs2git script by the same author.
+
+Documentation
+--------------
+Documentation by Matthias Urlichs <smurf@smurf.noris.de>.
+
+GIT
+---
+Part of the gitlink:git[7] suite
+
diff --git a/Makefile b/Makefile
index fd4e163..30fda9e 100644
--- a/Makefile
+++ b/Makefile
@@ -91,7 +91,8 @@ SCRIPT_SH = \
 
 SCRIPT_PERL = \
 	git-archimport.perl git-cvsimport.perl git-relink.perl \
-	git-rename.perl git-shortlog.perl git-fmt-merge-msg.perl
+	git-rename.perl git-shortlog.perl git-fmt-merge-msg.perl \
+	git-svnimport-perl
 
 SCRIPT_PYTHON = \
 	git-merge-recursive.py
diff --git a/git-svnimport.perl b/git-svnimport.perl
new file mode 100755
index 0000000..08645f7
--- /dev/null
+++ b/git-svnimport.perl
@@ -0,0 +1,671 @@
+#!/usr/bin/perl -w
+
+# This tool is copyright (c) 2005, Matthias Urlichs.
+# It is released under the Gnu Public License, version 2.
+#
+# The basic idea is to pull and analyze SVN changes.
+#
+# Checking out the files is done by a single long-running CVS connection
+# / server process.
+#
+# The head revision is on branch "origin" by default.
+# You can change that with the '-o' option.
+
+require v5.8.0; # for shell-safe open("-|",LIST)
+use strict;
+use warnings;
+use Getopt::Std;
+use File::Spec;
+use File::Temp qw(tempfile);
+use File::Path qw(mkpath);
+use File::Basename qw(basename dirname);
+use Time::Local;
+use IO::Pipe;
+use POSIX qw(strftime dup2);
+use IPC::Open2;
+use SVN::Core;
+use SVN::Ra;
+
+$SIG{'PIPE'}="IGNORE";
+$ENV{'TZ'}="UTC";
+
+our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b);
+
+sub usage() {
+	print STDERR <<END;
+Usage: ${\basename $0}     # fetch/update GIT from CVS
+       [-o branch-for-HEAD] [-h] [-v]
+       [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
+       [-i] [-u] [-s subst] [-m] [-M regex] [SVN_URL]
+END
+	exit(1);
+}
+
+getopts("b:C:hivmM:o:t:T:u") or usage();
+usage if $opt_h;
+
+my $tag_name = $opt_t || "tags";
+my $trunk_name = $opt_T || "trunk";
+my $branch_name = $opt_b || "branches";
+
+@ARGV <= 1 or usage();
+
+$opt_o ||= "origin";
+my $git_tree = $opt_C;
+$git_tree ||= ".";
+
+my $cvs_tree;
+if ($#ARGV == 0) {
+	$cvs_tree = $ARGV[0];
+} elsif (-f 'CVS/Repository') {
+	open my $f, '<', 'CVS/Repository' or 
+	    die 'Failed to open CVS/Repository';
+	$cvs_tree = <$f>;
+	chomp $cvs_tree;
+	close $f;
+} else {
+	usage();
+}
+
+our @mergerx = ();
+if ($opt_m) {
+	@mergerx = ( qr/\W(?:from|of|merge|merging|merged) (\w+)/i );
+}
+if ($opt_M) {
+	push (@mergerx, qr/$opt_M/);
+}
+
+select(STDERR); $|=1; select(STDOUT);
+
+
+package SVNconn;
+# Basic SVN connection.
+# We're only interested in connecting and downloading, so ...
+
+use File::Spec;
+use File::Temp qw(tempfile);
+use POSIX qw(strftime dup2);
+
+sub new {
+	my($what,$repo) = @_;
+	$what=ref($what) if ref($what);
+
+	my $self = {};
+	$self->{'buffer'} = "";
+	bless($self,$what);
+
+	$repo =~ s#/+$##;
+	$self->{'fullrep'} = $repo;
+	$self->conn();
+
+	$self->{'lines'} = undef;
+
+	return $self;
+}
+
+sub conn {
+	my $self = shift;
+	my $repo = $self->{'fullrep'};
+	my $s = SVN::Ra->new($repo);
+
+	die "SVN connection to $repo: $!\n" unless defined $s;
+	$self->{'svn'} = $s;
+	$self->{'repo'} = $repo;
+	$self->{'maxrev'} = $s->get_latest_revnum();
+}
+
+sub file {
+	my($self,$path,$rev) = @_;
+	my $res;
+
+	my ($fh, $name) = tempfile('gitsvn.XXXXXX', 
+		    DIR => File::Spec->tmpdir(), UNLINK => 1);
+
+	$self->{'svn'}->get_file($path,$rev,$fh) or do {
+	    # retry
+	    $self->conn();
+		$self->{'svn'}->get_file($path,$rev,$fh)
+		    or die "$rev: No file $path at $rev\n";
+	};
+	close ($fh);
+
+	return ($name, $res);
+}
+
+
+package main;
+
+my $svn = SVNconn->new($cvs_tree);
+
+
+sub pdate($) {
+	my($d) = @_;
+	$d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
+		or die "Unparseable date: $d\n";
+	my $y=$1; $y-=1900 if $y>1900;
+	return timegm($6||0,$5,$4,$3,$2-1,$y);
+}
+
+sub pmode($) {
+	my($mode) = @_;
+	my $m = 0;
+	my $mm = 0;
+	my $um = 0;
+	for my $x(split(//,$mode)) {
+		if($x eq ",") {
+			$m |= $mm&$um;
+			$mm = 0;
+			$um = 0;
+		} elsif($x eq "u") { $um |= 0700;
+		} elsif($x eq "g") { $um |= 0070;
+		} elsif($x eq "o") { $um |= 0007;
+		} elsif($x eq "r") { $mm |= 0444;
+		} elsif($x eq "w") { $mm |= 0222;
+		} elsif($x eq "x") { $mm |= 0111;
+		} elsif($x eq "=") { # do nothing
+		} else { die "Unknown mode: $mode\n";
+		}
+	}
+	$m |= $mm&$um;
+	return $m;
+}
+
+sub getwd() {
+	my $pwd = `pwd`;
+	chomp $pwd;
+	return $pwd;
+}
+
+
+sub get_headref($$) {
+    my $name    = shift;
+    my $git_dir = shift; 
+    my $sha;
+    
+    if (open(C,"$git_dir/refs/heads/$name")) {
+	chomp($sha = <C>);
+	close(C);
+	length($sha) == 40
+	    or die "Cannot get head id for $name ($sha): $!\n";
+    }
+    return $sha;
+}
+
+
+-d $git_tree
+	or mkdir($git_tree,0777)
+	or die "Could not create $git_tree: $!";
+chdir($git_tree);
+
+my $orig_branch = "";
+my $forward_master = 0;
+my %branches;
+
+my $git_dir = $ENV{"GIT_DIR"} || ".git";
+$git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
+$ENV{"GIT_DIR"} = $git_dir;
+my $orig_git_index;
+$orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
+my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
+				    DIR => File::Spec->tmpdir());
+close ($git_ih);
+$ENV{GIT_INDEX_FILE} = $git_index;
+my $maxnum = 0;
+my $last_rev = "";
+my $last_branch;
+my $current_rev = 0;
+unless(-d $git_dir) {
+	system("git-init-db");
+	die "Cannot init the GIT db at $git_tree: $?\n" if $?;
+	system("git-read-tree");
+	die "Cannot init an empty tree: $?\n" if $?;
+
+	$last_branch = $opt_o;
+	$orig_branch = "";
+} else {
+	-f "$git_dir/refs/heads/$opt_o"
+		or die "Branch '$opt_o' does not exist.\n".
+		       "Either use the correct '-o branch' option,\n".
+		       "or import to a new repository.\n";
+
+	-f "$git_dir/svn2git"
+		or die "'$git_dir/svn2git' does not exist.\n".
+		       "You need that file for incremental imports.\n";
+	$last_branch = basename(readlink("$git_dir/HEAD"));
+	unless($last_branch) {
+		warn "Cannot read the last branch name: $! -- assuming 'master'\n";
+		$last_branch = "master";
+	}
+	$orig_branch = $last_branch;
+	$last_rev = get_headref($orig_branch, $git_dir);
+	if (-f "$git_dir/SVN2GIT_HEAD") {
+		die <<EOM;
+SVN2GIT_HEAD exists.
+Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
+You may need to run
+
+    git-read-tree -m -u SVN2GIT_HEAD HEAD
+EOM
+	}
+	system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
+
+	$forward_master =
+	    $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
+	    system('cmp', '-s', "$git_dir/refs/heads/master", 
+				"$git_dir/refs/heads/$opt_o") == 0;
+
+	# populate index
+	system('git-read-tree', $last_rev);
+	die "read-tree failed: $?\n" if $?;
+
+	# Get the last import timestamps
+	open my $B,"<", "$git_dir/svn2git";
+	while(<$B>) {
+		chomp;
+		my($num,$branch,$ref) = split;
+		$branches{$branch}{$num} = $ref;
+		$branches{$branch}{"LAST"} = $ref;
+		$current_rev = $num+1 if $current_rev < $num+1;
+	}
+	close($B);
+}
+-d $git_dir
+	or die "Could not create git subdir ($git_dir).\n";
+
+open BRANCHES,">>", "$git_dir/svn2git";
+
+
+## cvsps output:
+#---------------------
+#PatchSet 314
+#Date: 1999/09/18 13:03:59
+#Author: wkoch
+#Branch: STABLE-BRANCH-1-0
+#Ancestor branch: HEAD
+#Tag: (none)
+#Log:
+#    See ChangeLog: Sat Sep 18 13:03:28 CEST 1999  Werner Koch
+#Members:
+#	README:1.57->1.57.2.1
+#	VERSION:1.96->1.96.2.1
+#
+#---------------------
+
+my $state = 0;
+
+sub get_file($$$) {
+	my($rev,$branch,$path) = @_;
+
+	# revert split_path(), below
+	my $svnpath;
+	$path = "" if $path eq "/"; # this should not happen, but ...
+	if($branch eq "/") {
+		$svnpath = "/$trunk_name/$path";
+	} elsif($branch =~ m#^/#) {
+		$svnpath = "/$tag_name$branch/$path";
+	} else {
+		$svnpath = "/$branch_name/$branch/$path";
+	}
+
+	# now get it
+	my ($name, $res) = $svn->file($svnpath,$rev);
+
+	open my $F, '-|', "git-hash-object -w $name"
+		or die "Cannot create object: $!\n";
+	my $sha = <$F>;
+	chomp $sha;
+	close $F;
+	# my $mode = pmode($cvs->{'mode'});
+	my $mode = "0644"; # SV does not seem to store any file modes
+	return [$mode, $sha, $path];
+}
+
+sub split_path($$) {
+	my($rev,$path) = @_;
+	my $branch;
+
+	if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
+		$branch = "/$1";
+	} elsif($path =~ s#^/\Q$trunk_name\E/?##) {
+		$branch = "/";
+	} elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
+		$branch = $1;
+	} else {
+		print STDERR "$rev: Unrecognized path: $path\n";
+		return ()
+	}
+	$path = "/" if $path eq "";
+	return ($branch,$path);
+}
+
+sub commit {
+	my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
+	my($author_name,$author_email,$dest);
+	my(@old,@new);
+
+	if ($author =~ /^(.*?)\s+<(.*)>$/) {
+		($author_name, $author_email) = ($1, $2);
+	} else {
+		$author =~ s/^<(.*)>$/$1/;
+		$author_name = $author_email = $author;
+	}
+	$date = pdate($date);
+
+	my $tag;
+	my $parent;
+	if($branch eq "/") { # trunk
+		$parent = $opt_o;
+	} elsif($branch =~ m#^/(.+)#) { # tag
+		$tag = 1;
+		$parent = $1;
+	} else { # "normal" branch
+		# nothing to do
+		$parent = $branch;
+	}
+	$dest = $parent;
+
+	my $prev = $changed_paths->{"/"};
+	if($prev and $prev->action eq "A") {
+		delete $changed_paths->{"/"};
+		my $oldpath = $prev->copyfrom_path;
+		my $rev;
+		if(defined $oldpath) {
+			my $p;
+			($parent,$p) = split_path($revision,$oldpath);
+			if($parent eq "/") {
+				$parent = $opt_o;
+			} else {
+				$parent =~ s#^/##; # if it's a tag
+			}
+		} else {
+			$parent = undef;
+		}
+	}
+
+	my $rev;
+	if(defined $parent) {
+		open(H,"git-rev-parse --verify $parent |");
+		$rev = <H>;
+		close(H) or do {
+			print STDERR "$revision: cannot find commit '$parent'!\n";
+			return;
+		};
+		chop $rev;
+		if(length($rev) != 40) {
+			print STDERR "$revision: cannot find commit '$parent'!\n";
+			return;
+		}
+		$rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
+		if($revision != 1 and not $rev) {
+			print STDERR "$revision: do not know ancestor for '$parent'!\n";
+			return;
+		}
+	} else {
+		$rev = undef;
+	}
+
+#	if($prev and $prev->action eq "A") {
+#		if(not $tag) {
+#			unless(open(H,"> $git_dir/refs/heads/$branch")) {
+#				print STDERR "$revision: Could not create branch $branch: $!\n";
+#				$state=11;
+#				next;
+#			}
+#			print H "$rev\n"
+#				or die "Could not write branch $branch: $!";
+#			close(H)
+#				or die "Could not write branch $branch: $!";
+#		}
+#	}
+	if(not defined $rev) {
+		unlink($git_index);
+	} elsif ($rev ne $last_rev) {
+		print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
+		system("git-read-tree", $rev);
+		die "read-tree failed for $rev: $?\n" if $?;
+		$last_rev = $rev;
+	}
+
+	while(my($path,$action) = each %$changed_paths) {
+		if ($action->action eq "A") {
+			my $f = get_file($revision,$branch,$path);
+			push(@new,$f) if $f;
+		} elsif ($action->action eq "D") {
+			push(@old,$path);
+		} elsif ($action->action eq "M") {
+			my $f = get_file($revision,$branch,$path);
+			push(@new,$f) if $f;
+		} elsif ($action->action eq "R") {
+			# refer to a file/tree in an earlier commit
+			push(@old,$path); # remove any old stuff
+
+			# ... and add any new stuff
+			my($b,$p) = split_path($revision,$action->oldpath);
+			open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->oldrev}, $p;
+			local $/ = '\0';
+			while(<$F>) {
+				chomp;
+				my($m,$p) = split(/\t/,$_,2);
+				my($mode,$type,$sha1) = split(/ /,$m);
+				next if $type ne "blob";
+				push(@new,[$mode,$sha1,$p]);
+			}
+		} else {
+			die "$revision: unknown action '".$action->action."' for $path\n";
+		}
+	}
+
+	if(@old) {
+		open F, "-│", "git-ls-files", "-z", @old or die $!;
+		@old = ();
+		local $/ = '\0';
+		while(<F>) {
+			chomp;
+			push(@old,$_);
+		}
+		close(F);
+
+		while(@old) {
+			my @o2;
+			if(@old > 55) {
+				@o2 = splice(@old,0,50);
+			} else {
+				@o2 = @old;
+				@old = ();
+			}
+			system("git-update-index","--force-remove","--",@o2);
+			die "Cannot remove files: $?\n" if $?;
+		}
+	}
+	while(@new) {
+		my @n2;
+		if(@new > 12) {
+			@n2 = splice(@new,0,10);
+		} else {
+			@n2 = @new;
+			@new = ();
+		}
+		system("git-update-index","--add",
+			(map { ('--cacheinfo', @$_) } @n2));
+		die "Cannot add files: $?\n" if $?;
+	}
+
+	my $pid = open(C,"-|");
+	die "Cannot fork: $!" unless defined $pid;
+	unless($pid) {
+		exec("git-write-tree");
+		die "Cannot exec git-write-tree: $!\n";
+	}
+	chomp(my $tree = <C>);
+	length($tree) == 40
+		or die "Cannot get tree id ($tree): $!\n";
+	close(C)
+		or die "Error running git-write-tree: $?\n";
+	print "Tree ID $tree\n" if $opt_v;
+
+	my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
+	my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
+	$pid = fork();
+	die "Fork: $!\n" unless defined $pid;
+	unless($pid) {
+		$pr->writer();
+		$pw->reader();
+		open(OUT,">&STDOUT");
+		dup2($pw->fileno(),0);
+		dup2($pr->fileno(),1);
+		$pr->close();
+		$pw->close();
+
+		my @par = ();
+		@par = ("-p",$rev) if defined $rev;
+
+		# loose detection of merges
+		# based on the commit msg
+		foreach my $rx (@mergerx) {
+			if ($message =~ $rx) {
+				my $mparent = $1;
+				if ($mparent eq 'HEAD') { $mparent = $opt_o };
+				if ( -e "$git_dir/refs/heads/$mparent") {
+					$mparent = get_headref($mparent, $git_dir);
+					push @par, '-p', $mparent;
+					print OUT "Merge parent branch: $mparent\n" if $opt_v;
+				}
+			} 
+		}
+
+		exec("env",
+			"GIT_AUTHOR_NAME=$author_name",
+			"GIT_AUTHOR_EMAIL=$author_email",
+			"GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
+			"GIT_COMMITTER_NAME=$author_name",
+			"GIT_COMMITTER_EMAIL=$author_email",
+			"GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
+			"git-commit-tree", $tree,@par);
+		die "Cannot exec git-commit-tree: $!\n";
+	}
+	$pw->writer();
+	$pr->reader();
+
+	$message =~ s/[\s\n]+\z//;
+
+	print $pw "$message\n"
+		or die "Error writing to git-commit-tree: $!\n";
+	$pw->close();
+
+	print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
+	chomp(my $cid = <$pr>);
+	length($cid) == 40
+		or die "Cannot get commit id ($cid): $!\n";
+	print "Commit ID $cid\n" if $opt_v;
+	$pr->close();
+
+	waitpid($pid,0);
+	die "Error running git-commit-tree: $?\n" if $?;
+
+	if(defined $dest) {
+		print "Writing to refs/heads/$dest\n" if $opt_v;
+		open(C,">$git_dir/refs/heads/$dest") and 
+		print C ("$cid\n") and
+		close(C)
+			or die "Cannot write branch $dest for update: $!\n";
+	} else {
+		print "... no known parent\n" if $opt_v;
+	}
+	$branches{$branch}{"LAST"} = $cid;
+	$branches{$branch}{$revision} = $cid;
+	$last_rev = $cid;
+	print BRANCHES "$revision $branch $cid\n";
+	print "DONE: $revision $dest $cid\n" if $opt_v;
+
+	if($tag) {
+		my($in, $out) = ('','');
+		$last_rev = "-" if %$changed_paths;
+		# the tag was 'complex', i.e. did not refer to a "real" revision
+		
+		$tag =~ tr/_/\./ if $opt_u;
+
+		my $pid = open2($in, $out, 'git-mktag');
+		print $out ("object $cid\n".
+		    "type commit\n".
+		    "tag $tag\n".
+		    "tagger $author_name <$author_email>\n") and
+		close($out)
+		    or die "Cannot create tag object $tag: $!\n";
+
+		my $tagobj = <$in>;
+		chomp $tagobj;
+
+		if ( !close($in) or waitpid($pid, 0) != $pid or
+				$? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
+			die "Cannot create tag object $tag: $!\n";
+		}
+		
+
+		open(C,">$git_dir/refs/tags/$tag")
+			or die "Cannot create tag $tag: $!\n";
+		print C "$tagobj\n"
+			or die "Cannot write tag $tag: $!\n";
+		close(C)
+			or die "Cannot write tag $tag: $!\n";
+
+		print "Created tag '$tag' on '$branch'\n" if $opt_v;
+	}
+}
+
+my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
+sub _commit_all {
+	($changed_paths, $revision, $author, $date, $message, $pool) = @_;
+}
+sub commit_all {
+	my %done;
+	my @col;
+	my $pref;
+	my $branch;
+
+	while(my($path,$action) = each %$changed_paths) {
+		($branch,$path) = split_path($revision,$path);
+		next if not defined $branch;
+		$done{$branch}{$path} = $action;
+	}
+	while(($branch,$changed_paths) = each %done) {
+		commit($branch, $changed_paths, $revision, $author, $date, $message);
+	}
+}
+
+while(++$current_rev < $svn->{'maxrev'}) {
+	$svn->{'svn'}->get_log("/",$current_rev,$current_rev,$current_rev,1,1,\&_commit_all,"");
+	commit_all();
+}
+
+
+unlink($git_index);
+
+if (defined $orig_git_index) {
+	$ENV{GIT_INDEX_FILE} = $orig_git_index;
+} else {
+	delete $ENV{GIT_INDEX_FILE};
+}
+
+# Now switch back to the branch we were in before all of this happened
+if($orig_branch) {
+	print "DONE\n" if $opt_v;
+	system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
+		if $forward_master;
+	unless ($opt_i) {
+		system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
+		die "read-tree failed: $?\n" if $?;
+	}
+} else {
+	$orig_branch = "master";
+	print "DONE; creating $orig_branch branch\n" if $opt_v;
+	system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
+		unless -f "$git_dir/refs/heads/master";
+	unlink("$git_dir/HEAD");
+	symlink("refs/heads/$orig_branch","$git_dir/HEAD");
+	unless ($opt_i) {
+		system('git checkout');
+		die "checkout failed: $?\n" if $?;
+	}
+}
+unlink("$git_dir/SVN2GIT_HEAD");
+close(BRANCHES);
-- 
cgit v0.10.2-6-g49f6


From 2b5e63d1b4a6f1b0ef58755a192ecb2d04e3569d Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 12:33:22 +0200
Subject: svn import: add eval()

Trying to downlaod a file that's really a subdirectory doesn't work too well.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 08645f7..920b23e 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -121,12 +121,15 @@ sub file {
 	my ($fh, $name) = tempfile('gitsvn.XXXXXX', 
 		    DIR => File::Spec->tmpdir(), UNLINK => 1);
 
-	$self->{'svn'}->get_file($path,$rev,$fh) or do {
+	print "... $rev $path ...\n" if $opt_v;
+	eval { $self->{'svn'}->get_file($path,$rev,$fh); };
+	if (defined $@ and $@ !~ /Attempted to get checksum/) {
 	    # retry
 	    $self->conn();
-		$self->{'svn'}->get_file($path,$rev,$fh)
-		    or die "$rev: No file $path at $rev\n";
+		eval { $self->{'svn'}->get_file($path,$rev,$fh); };
 	};
+	return () if defined $@ and $@ !~ /Attempted to get checksum/;
+	die $@ if $@;
 	close ($fh);
 
 	return ($name, $res);
@@ -308,7 +311,8 @@ sub get_file($$$) {
 	}
 
 	# now get it
-	my ($name, $res) = $svn->file($svnpath,$rev);
+	my ($name, $res) = eval { $svn->file($svnpath,$rev); };
+	return () unless defined $name;
 
 	open my $F, '-|', "git-hash-object -w $name"
 		or die "Cannot create object: $!\n";
@@ -343,7 +347,9 @@ sub commit {
 	my($author_name,$author_email,$dest);
 	my(@old,@new);
 
-	if ($author =~ /^(.*?)\s+<(.*)>$/) {
+	if (not defined $author) {
+		$author_name = $author_email = "unknown";
+	} elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
 		($author_name, $author_email) = ($1, $2);
 	} else {
 		$author =~ s/^<(.*)>$/$1/;
-- 
cgit v0.10.2-6-g49f6


From c015bf2bcbdc2021410753b4a024dd35dfef6a33 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 12:34:32 +0200
Subject: SVN import: No modes

svn doesn't seem to save file modes:
removed the code that analyzes them.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 920b23e..27d964c 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -149,30 +149,6 @@ sub pdate($) {
 	return timegm($6||0,$5,$4,$3,$2-1,$y);
 }
 
-sub pmode($) {
-	my($mode) = @_;
-	my $m = 0;
-	my $mm = 0;
-	my $um = 0;
-	for my $x(split(//,$mode)) {
-		if($x eq ",") {
-			$m |= $mm&$um;
-			$mm = 0;
-			$um = 0;
-		} elsif($x eq "u") { $um |= 0700;
-		} elsif($x eq "g") { $um |= 0070;
-		} elsif($x eq "o") { $um |= 0007;
-		} elsif($x eq "r") { $mm |= 0444;
-		} elsif($x eq "w") { $mm |= 0222;
-		} elsif($x eq "x") { $mm |= 0111;
-		} elsif($x eq "=") { # do nothing
-		} else { die "Unknown mode: $mode\n";
-		}
-	}
-	$m |= $mm&$um;
-	return $m;
-}
-
 sub getwd() {
 	my $pwd = `pwd`;
 	chomp $pwd;
@@ -319,7 +295,6 @@ sub get_file($$$) {
 	my $sha = <$F>;
 	chomp $sha;
 	close $F;
-	# my $mode = pmode($cvs->{'mode'});
 	my $mode = "0644"; # SV does not seem to store any file modes
 	return [$mode, $sha, $path];
 }
-- 
cgit v0.10.2-6-g49f6


From f0daa628a297819ed039e6b61d337fde02d8a6dd Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 12:41:15 +0200
Subject: svn import: copy path information

Due to a bug in the SVN library, path information is freed as soon as the
callback returns, even if it still refers to the data.

Workaround: Copy it. (Also fix a wrong-method-name bug while we're at it.)

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 27d964c..3844cc5 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -346,9 +346,9 @@ sub commit {
 	$dest = $parent;
 
 	my $prev = $changed_paths->{"/"};
-	if($prev and $prev->action eq "A") {
+	if($prev and $prev->[0] eq "A") {
 		delete $changed_paths->{"/"};
-		my $oldpath = $prev->copyfrom_path;
+		my $oldpath = $prev->[1];
 		my $rev;
 		if(defined $oldpath) {
 			my $p;
@@ -385,7 +385,7 @@ sub commit {
 		$rev = undef;
 	}
 
-#	if($prev and $prev->action eq "A") {
+#	if($prev and $prev->[0] eq "A") {
 #		if(not $tag) {
 #			unless(open(H,"> $git_dir/refs/heads/$branch")) {
 #				print STDERR "$revision: Could not create branch $branch: $!\n";
@@ -408,21 +408,21 @@ sub commit {
 	}
 
 	while(my($path,$action) = each %$changed_paths) {
-		if ($action->action eq "A") {
+		if ($action->[0] eq "A") {
 			my $f = get_file($revision,$branch,$path);
 			push(@new,$f) if $f;
-		} elsif ($action->action eq "D") {
+		} elsif ($action->[0] eq "D") {
 			push(@old,$path);
-		} elsif ($action->action eq "M") {
+		} elsif ($action->[0] eq "M") {
 			my $f = get_file($revision,$branch,$path);
 			push(@new,$f) if $f;
-		} elsif ($action->action eq "R") {
+		} elsif ($action->[0] eq "R") {
 			# refer to a file/tree in an earlier commit
 			push(@old,$path); # remove any old stuff
 
 			# ... and add any new stuff
-			my($b,$p) = split_path($revision,$action->oldpath);
-			open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->oldrev}, $p;
+			my($b,$p) = split_path($revision,$action->[1]);
+			open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $p;
 			local $/ = '\0';
 			while(<$F>) {
 				chomp;
@@ -432,7 +432,7 @@ sub commit {
 				push(@new,[$mode,$sha1,$p]);
 			}
 		} else {
-			die "$revision: unknown action '".$action->action."' for $path\n";
+			die "$revision: unknown action '".$action->[0]."' for $path\n";
 		}
 	}
 
@@ -596,7 +596,13 @@ sub commit {
 my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
 sub _commit_all {
 	($changed_paths, $revision, $author, $date, $message, $pool) = @_;
+	my %p;
+	while(my($path,$action) = each %$changed_paths) {
+		$p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev ];
+	}
+	$changed_paths = \%p;
 }
+
 sub commit_all {
 	my %done;
 	my @col;
-- 
cgit v0.10.2-6-g49f6


From 6d281217b930466717d0078285e5431c409b7aec Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 12:45:46 +0200
Subject: svn import: fixed two pipe open calls.

Perl's magic "different semantics of open() based on the number of
arguments" is really annoying at times...

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 3844cc5..415b50b 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -290,7 +290,7 @@ sub get_file($$$) {
 	my ($name, $res) = eval { $svn->file($svnpath,$rev); };
 	return () unless defined $name;
 
-	open my $F, '-|', "git-hash-object -w $name"
+	open my $F, '-|', "git-hash-object", "-w", $name
 		or die "Cannot create object: $!\n";
 	my $sha = <$F>;
 	chomp $sha;
@@ -437,14 +437,14 @@ sub commit {
 	}
 
 	if(@old) {
-		open F, "-│", "git-ls-files", "-z", @old or die $!;
+		open my $F, "-│", "git-ls-files", "-z", @old or die $!;
 		@old = ();
 		local $/ = '\0';
-		while(<F>) {
+		while(<$F>) {
 			chomp;
 			push(@old,$_);
 		}
-		close(F);
+		close($F);
 
 		while(@old) {
 			my @o2;
-- 
cgit v0.10.2-6-g49f6


From c6582aba5230301fe826c06d1891a4a8a9d32074 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 13:10:01 +0200
Subject: svn import: add libsvn-core-perl to Debian's control file

Added libsvn-core-perl to debian/control, "Recommends:" section.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/debian/control b/debian/control
index 5d75c32..0ec8933 100644
--- a/debian/control
+++ b/debian/control
@@ -8,7 +8,7 @@ Standards-Version: 3.6.1
 Package: git-core
 Architecture: any
 Depends: ${shlibs:Depends}, ${perl:Depends}, ${misc:Depends}, patch, rcs
-Recommends: rsync, curl, ssh, libmail-sendmail-perl, libemail-valid-perl, python (>= 2.4.0), less
+Recommends: rsync, curl, ssh, libmail-sendmail-perl, libemail-valid-perl, libsvn-core-perl (>= 1.2.1), python (>= 2.4.0), less
 Suggests: cogito
 Conflicts: git, cogito (<< 0.13)
 Description: The git content addressable filesystem
-- 
cgit v0.10.2-6-g49f6


From 37dcf6de60498dfedb5d97ad96789b0527493bcd Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 13:42:48 +0200
Subject: svn improt needs SVN::Core 1.2.1 or better

Die with a warning if Perl's svn module is too old.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 415b50b..ef73f36 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -26,6 +26,8 @@ use IPC::Open2;
 use SVN::Core;
 use SVN::Ra;
 
+die "Need CVN:COre 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
+
 $SIG{'PIPE'}="IGNORE";
 $ENV{'TZ'}="UTC";
 
-- 
cgit v0.10.2-6-g49f6


From 8cd4177d5ef651c8acb9c1f113a076e6fae2e1c2 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 14:14:44 +0200
Subject: svn import: avoid reconnecting

Perl's eval() sets $@ to empts, not undef, when it succeeds.
That caused excessive reconnect attempts.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index ef73f36..ba9f818 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -100,8 +100,6 @@ sub new {
 	$self->{'fullrep'} = $repo;
 	$self->conn();
 
-	$self->{'lines'} = undef;
-
 	return $self;
 }
 
@@ -112,6 +110,7 @@ sub conn {
 
 	die "SVN connection to $repo: $!\n" unless defined $s;
 	$self->{'svn'} = $s;
+	print STDERR "*** SVN *** $s ***\n";
 	$self->{'repo'} = $repo;
 	$self->{'maxrev'} = $s->get_latest_revnum();
 }
@@ -124,13 +123,15 @@ sub file {
 		    DIR => File::Spec->tmpdir(), UNLINK => 1);
 
 	print "... $rev $path ...\n" if $opt_v;
-	eval { $self->{'svn'}->get_file($path,$rev,$fh); };
-	if (defined $@ and $@ !~ /Attempted to get checksum/) {
+	my $s = $self->{'svn'};
+	print STDERR "*** GET *** $s ***\n";
+	eval { $s->get_file($path,$rev,$fh); };
+	if ($@ and $@ !~ /Attempted to get checksum/) {
 	    # retry
 	    $self->conn();
 		eval { $self->{'svn'}->get_file($path,$rev,$fh); };
 	};
-	return () if defined $@ and $@ !~ /Attempted to get checksum/;
+	return () if $@ and $@ !~ /Attempted to get checksum/;
 	die $@ if $@;
 	close ($fh);
 
-- 
cgit v0.10.2-6-g49f6


From c7ff5f1d7d6f103802fd6d6b1d4dcdf095e95d5f Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 14:18:38 +0200
Subject: svn import: wrong file open mode

There are multiple | characters in Unicode. Don't use the wrong one ...

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index ba9f818..2686f4a 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -440,7 +440,7 @@ sub commit {
 	}
 
 	if(@old) {
-		open my $F, "-│", "git-ls-files", "-z", @old or die $!;
+		open my $F, "-|", "git-ls-files", "-z", @old or die $!;
 		@old = ();
 		local $/ = '\0';
 		while(<$F>) {
-- 
cgit v0.10.2-6-g49f6


From 62d72e4b707d6931787e91f611ed49aaabb14391 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 14:19:15 +0200
Subject: svn import: remove debugging

Removed debugging output used to identify the too-many-connections problem.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 2686f4a..f7bf616 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -110,7 +110,6 @@ sub conn {
 
 	die "SVN connection to $repo: $!\n" unless defined $s;
 	$self->{'svn'} = $s;
-	print STDERR "*** SVN *** $s ***\n";
 	$self->{'repo'} = $repo;
 	$self->{'maxrev'} = $s->get_latest_revnum();
 }
@@ -124,7 +123,6 @@ sub file {
 
 	print "... $rev $path ...\n" if $opt_v;
 	my $s = $self->{'svn'};
-	print STDERR "*** GET *** $s ***\n";
 	eval { $s->get_file($path,$rev,$fh); };
 	if ($@ and $@ !~ /Attempted to get checksum/) {
 	    # retry
-- 
cgit v0.10.2-6-g49f6


From f02b3eba7fc6e3fca25e0063dde6ff149aed2b45 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 14:42:59 +0200
Subject: svn import: Fix tagging.

Tagging was 100% broken. :-/

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index f7bf616..f977ed2 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -122,8 +122,7 @@ sub file {
 		    DIR => File::Spec->tmpdir(), UNLINK => 1);
 
 	print "... $rev $path ...\n" if $opt_v;
-	my $s = $self->{'svn'};
-	eval { $s->get_file($path,$rev,$fh); };
+	eval { $self->{'svn'}->get_file($path,$rev,$fh); };
 	if ($@ and $@ !~ /Attempted to get checksum/) {
 	    # retry
 	    $self->conn();
@@ -544,14 +543,14 @@ sub commit {
 	waitpid($pid,0);
 	die "Error running git-commit-tree: $?\n" if $?;
 
-	if(defined $dest) {
+	if(not defined $dest) {
+		print "... no known parent\n" if $opt_v;
+	} elsif(not $tag) {
 		print "Writing to refs/heads/$dest\n" if $opt_v;
 		open(C,">$git_dir/refs/heads/$dest") and 
 		print C ("$cid\n") and
 		close(C)
 			or die "Cannot write branch $dest for update: $!\n";
-	} else {
-		print "... no known parent\n" if $opt_v;
 	}
 	$branches{$branch}{"LAST"} = $cid;
 	$branches{$branch}{$revision} = $cid;
@@ -564,33 +563,30 @@ sub commit {
 		$last_rev = "-" if %$changed_paths;
 		# the tag was 'complex', i.e. did not refer to a "real" revision
 		
-		$tag =~ tr/_/\./ if $opt_u;
+		$dest =~ tr/_/\./ if $opt_u;
 
 		my $pid = open2($in, $out, 'git-mktag');
 		print $out ("object $cid\n".
 		    "type commit\n".
-		    "tag $tag\n".
+		    "tag $dest\n".
 		    "tagger $author_name <$author_email>\n") and
 		close($out)
-		    or die "Cannot create tag object $tag: $!\n";
+		    or die "Cannot create tag object $dest: $!\n";
 
 		my $tagobj = <$in>;
 		chomp $tagobj;
 
 		if ( !close($in) or waitpid($pid, 0) != $pid or
 				$? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
-			die "Cannot create tag object $tag: $!\n";
+			die "Cannot create tag object $dest: $!\n";
 		}
-		
 
-		open(C,">$git_dir/refs/tags/$tag")
-			or die "Cannot create tag $tag: $!\n";
-		print C "$tagobj\n"
-			or die "Cannot write tag $tag: $!\n";
+		open(C,">$git_dir/refs/tags/$dest") and
+		print C ("$tagobj\n") and
 		close(C)
-			or die "Cannot write tag $tag: $!\n";
+			or die "Cannot create tag $branch: $!\n";
 
-		print "Created tag '$tag' on '$branch'\n" if $opt_v;
+		print "Created tag '$dest' on '$branch'\n" if $opt_v;
 	}
 }
 
-- 
cgit v0.10.2-6-g49f6


From bf267d99e88410a31452d4d1ccae0b9b4432b11e Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 14:51:13 +0200
Subject: svn import: Do not create empty tags

If a tag is "clean", do not create a commit for it.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index f977ed2..10ffb54 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -407,141 +407,146 @@ sub commit {
 		$last_rev = $rev;
 	}
 
-	while(my($path,$action) = each %$changed_paths) {
-		if ($action->[0] eq "A") {
-			my $f = get_file($revision,$branch,$path);
-			push(@new,$f) if $f;
-		} elsif ($action->[0] eq "D") {
-			push(@old,$path);
-		} elsif ($action->[0] eq "M") {
-			my $f = get_file($revision,$branch,$path);
-			push(@new,$f) if $f;
-		} elsif ($action->[0] eq "R") {
-			# refer to a file/tree in an earlier commit
-			push(@old,$path); # remove any old stuff
-
-			# ... and add any new stuff
-			my($b,$p) = split_path($revision,$action->[1]);
-			open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $p;
+	my $cid;
+	if($tag and not %$changed_paths) {
+		$cid = $rev;
+	} else {
+		while(my($path,$action) = each %$changed_paths) {
+			if ($action->[0] eq "A") {
+				my $f = get_file($revision,$branch,$path);
+				push(@new,$f) if $f;
+			} elsif ($action->[0] eq "D") {
+				push(@old,$path);
+			} elsif ($action->[0] eq "M") {
+				my $f = get_file($revision,$branch,$path);
+				push(@new,$f) if $f;
+			} elsif ($action->[0] eq "R") {
+				# refer to a file/tree in an earlier commit
+				push(@old,$path); # remove any old stuff
+
+				# ... and add any new stuff
+				my($b,$p) = split_path($revision,$action->[1]);
+				open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $p;
+				local $/ = '\0';
+				while(<$F>) {
+					chomp;
+					my($m,$p) = split(/\t/,$_,2);
+					my($mode,$type,$sha1) = split(/ /,$m);
+					next if $type ne "blob";
+					push(@new,[$mode,$sha1,$p]);
+				}
+			} else {
+				die "$revision: unknown action '".$action->[0]."' for $path\n";
+			}
+		}
+
+		if(@old) {
+			open my $F, "-|", "git-ls-files", "-z", @old or die $!;
+			@old = ();
 			local $/ = '\0';
 			while(<$F>) {
 				chomp;
-				my($m,$p) = split(/\t/,$_,2);
-				my($mode,$type,$sha1) = split(/ /,$m);
-				next if $type ne "blob";
-				push(@new,[$mode,$sha1,$p]);
+				push(@old,$_);
+			}
+			close($F);
+
+			while(@old) {
+				my @o2;
+				if(@old > 55) {
+					@o2 = splice(@old,0,50);
+				} else {
+					@o2 = @old;
+					@old = ();
+				}
+				system("git-update-index","--force-remove","--",@o2);
+				die "Cannot remove files: $?\n" if $?;
 			}
-		} else {
-			die "$revision: unknown action '".$action->[0]."' for $path\n";
-		}
-	}
-
-	if(@old) {
-		open my $F, "-|", "git-ls-files", "-z", @old or die $!;
-		@old = ();
-		local $/ = '\0';
-		while(<$F>) {
-			chomp;
-			push(@old,$_);
 		}
-		close($F);
-
-		while(@old) {
-			my @o2;
-			if(@old > 55) {
-				@o2 = splice(@old,0,50);
+		while(@new) {
+			my @n2;
+			if(@new > 12) {
+				@n2 = splice(@new,0,10);
 			} else {
-				@o2 = @old;
-				@old = ();
+				@n2 = @new;
+				@new = ();
 			}
-			system("git-update-index","--force-remove","--",@o2);
-			die "Cannot remove files: $?\n" if $?;
-		}
-	}
-	while(@new) {
-		my @n2;
-		if(@new > 12) {
-			@n2 = splice(@new,0,10);
-		} else {
-			@n2 = @new;
-			@new = ();
+			system("git-update-index","--add",
+				(map { ('--cacheinfo', @$_) } @n2));
+			die "Cannot add files: $?\n" if $?;
 		}
-		system("git-update-index","--add",
-			(map { ('--cacheinfo', @$_) } @n2));
-		die "Cannot add files: $?\n" if $?;
-	}
-
-	my $pid = open(C,"-|");
-	die "Cannot fork: $!" unless defined $pid;
-	unless($pid) {
-		exec("git-write-tree");
-		die "Cannot exec git-write-tree: $!\n";
-	}
-	chomp(my $tree = <C>);
-	length($tree) == 40
-		or die "Cannot get tree id ($tree): $!\n";
-	close(C)
-		or die "Error running git-write-tree: $?\n";
-	print "Tree ID $tree\n" if $opt_v;
-
-	my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
-	my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
-	$pid = fork();
-	die "Fork: $!\n" unless defined $pid;
-	unless($pid) {
-		$pr->writer();
-		$pw->reader();
-		open(OUT,">&STDOUT");
-		dup2($pw->fileno(),0);
-		dup2($pr->fileno(),1);
-		$pr->close();
-		$pw->close();
 
-		my @par = ();
-		@par = ("-p",$rev) if defined $rev;
-
-		# loose detection of merges
-		# based on the commit msg
-		foreach my $rx (@mergerx) {
-			if ($message =~ $rx) {
-				my $mparent = $1;
-				if ($mparent eq 'HEAD') { $mparent = $opt_o };
-				if ( -e "$git_dir/refs/heads/$mparent") {
-					$mparent = get_headref($mparent, $git_dir);
-					push @par, '-p', $mparent;
-					print OUT "Merge parent branch: $mparent\n" if $opt_v;
-				}
-			} 
+		my $pid = open(C,"-|");
+		die "Cannot fork: $!" unless defined $pid;
+		unless($pid) {
+			exec("git-write-tree");
+			die "Cannot exec git-write-tree: $!\n";
 		}
+		chomp(my $tree = <C>);
+		length($tree) == 40
+			or die "Cannot get tree id ($tree): $!\n";
+		close(C)
+			or die "Error running git-write-tree: $?\n";
+		print "Tree ID $tree\n" if $opt_v;
+
+		my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
+		my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
+		$pid = fork();
+		die "Fork: $!\n" unless defined $pid;
+		unless($pid) {
+			$pr->writer();
+			$pw->reader();
+			open(OUT,">&STDOUT");
+			dup2($pw->fileno(),0);
+			dup2($pr->fileno(),1);
+			$pr->close();
+			$pw->close();
+
+			my @par = ();
+			@par = ("-p",$rev) if defined $rev;
+
+			# loose detection of merges
+			# based on the commit msg
+			foreach my $rx (@mergerx) {
+				if ($message =~ $rx) {
+					my $mparent = $1;
+					if ($mparent eq 'HEAD') { $mparent = $opt_o };
+					if ( -e "$git_dir/refs/heads/$mparent") {
+						$mparent = get_headref($mparent, $git_dir);
+						push @par, '-p', $mparent;
+						print OUT "Merge parent branch: $mparent\n" if $opt_v;
+					}
+				} 
+			}
 
-		exec("env",
-			"GIT_AUTHOR_NAME=$author_name",
-			"GIT_AUTHOR_EMAIL=$author_email",
-			"GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
-			"GIT_COMMITTER_NAME=$author_name",
-			"GIT_COMMITTER_EMAIL=$author_email",
-			"GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
-			"git-commit-tree", $tree,@par);
-		die "Cannot exec git-commit-tree: $!\n";
-	}
-	$pw->writer();
-	$pr->reader();
+			exec("env",
+				"GIT_AUTHOR_NAME=$author_name",
+				"GIT_AUTHOR_EMAIL=$author_email",
+				"GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
+				"GIT_COMMITTER_NAME=$author_name",
+				"GIT_COMMITTER_EMAIL=$author_email",
+				"GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
+				"git-commit-tree", $tree,@par);
+			die "Cannot exec git-commit-tree: $!\n";
+		}
+		$pw->writer();
+		$pr->reader();
 
-	$message =~ s/[\s\n]+\z//;
+		$message =~ s/[\s\n]+\z//;
 
-	print $pw "$message\n"
-		or die "Error writing to git-commit-tree: $!\n";
-	$pw->close();
+		print $pw "$message\n"
+			or die "Error writing to git-commit-tree: $!\n";
+		$pw->close();
 
-	print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
-	chomp(my $cid = <$pr>);
-	length($cid) == 40
-		or die "Cannot get commit id ($cid): $!\n";
-	print "Commit ID $cid\n" if $opt_v;
-	$pr->close();
+		print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
+		chomp($cid = <$pr>);
+		length($cid) == 40
+			or die "Cannot get commit id ($cid): $!\n";
+		print "Commit ID $cid\n" if $opt_v;
+		$pr->close();
 
-	waitpid($pid,0);
-	die "Error running git-commit-tree: $?\n" if $?;
+		waitpid($pid,0);
+		die "Error running git-commit-tree: $?\n" if $?;
+	}
 
 	if(not defined $dest) {
 		print "... no known parent\n" if $opt_v;
-- 
cgit v0.10.2-6-g49f6


From 7ee74a99b2b710b5f5adc22db2fe0aca8a74c809 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 15:14:21 +0200
Subject: svn import: skip initial revisions

Add a flag to skip initial revisions: some SVN repositories have
initial setup cruft in their logs which we might want to ignore.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/Documentation/git-svnimport.txt b/Documentation/git-svnimport.txt
index be03a65..a144c6c 100644
--- a/Documentation/git-svnimport.txt
+++ b/Documentation/git-svnimport.txt
@@ -12,7 +12,7 @@ SYNOPSIS
 'git-svnimport' [ -o <branch-for-HEAD> ] [ -h ] [ -v ]
 			[ -C <GIT_repository> ] [ -i ] [ -u ]
 			[ -b branch_subdir ] [ -t trunk_subdir ] [ -T tag_subdir ]
-			[ -m ] [ -M regex ] [ <SVN_repository_URL> ]
+			[ -s start_chg ] [ -m ] [ -M regex ] [ <SVN_repository_URL> ]
 
 
 DESCRIPTION
@@ -36,6 +36,11 @@ OPTIONS
         The GIT repository to import to.  If the directory doesn't
         exist, it will be created.  Default is the current directory.
 
+-s <start_rev>::
+        Start importing at this SVN change number. The  default is 1.
++
+When importing incementally, you might need to edit the .git/svn2git file.
+
 -i::
 	Import-only: don't perform a checkout after importing.  This option
 	ensures the working directory and cache remain untouched and will
diff --git a/git-svnimport.perl b/git-svnimport.perl
index 10ffb54..896222b 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -31,19 +31,19 @@ die "Need CVN:COre 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
 $SIG{'PIPE'}="IGNORE";
 $ENV{'TZ'}="UTC";
 
-our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b);
+our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b,$opt_s);
 
 sub usage() {
 	print STDERR <<END;
 Usage: ${\basename $0}     # fetch/update GIT from CVS
        [-o branch-for-HEAD] [-h] [-v]
        [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
-       [-i] [-u] [-s subst] [-m] [-M regex] [SVN_URL]
+       [-i] [-u] [-s start_chg] [-m] [-M regex] [SVN_URL]
 END
 	exit(1);
 }
 
-getopts("b:C:hivmM:o:t:T:u") or usage();
+getopts("b:C:hivmM:o:s:t:T:u") or usage();
 usage if $opt_h;
 
 my $tag_name = $opt_t || "tags";
@@ -192,7 +192,7 @@ $ENV{GIT_INDEX_FILE} = $git_index;
 my $maxnum = 0;
 my $last_rev = "";
 my $last_branch;
-my $current_rev = 0;
+my $current_rev = $opt_s ? ($opt_s-1) : 0;
 unless(-d $git_dir) {
 	system("git-init-db");
 	die "Cannot init the GIT db at $git_tree: $?\n" if $?;
@@ -364,7 +364,7 @@ sub commit {
 	}
 
 	my $rev;
-	if(defined $parent) {
+	if($revision > $opt_s and defined $parent) {
 		open(H,"git-rev-parse --verify $parent |");
 		$rev = <H>;
 		close(H) or do {
@@ -377,7 +377,7 @@ sub commit {
 			return;
 		}
 		$rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
-		if($revision != 1 and not $rev) {
+		if($revision != $opt_s and not $rev) {
 			print STDERR "$revision: do not know ancestor for '$parent'!\n";
 			return;
 		}
-- 
cgit v0.10.2-6-g49f6


From e7e477dfacecaf362f08a65644d806da5231a9f0 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 15:28:00 +0200
Subject: svn import: incremental imports

Incremental imports skipped a revision.

Also improve interrupt safety -- ^C while writing a tag caused the tag
to be skipped.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 896222b..00f563d 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -244,7 +244,7 @@ EOM
 		my($num,$branch,$ref) = split;
 		$branches{$branch}{$num} = $ref;
 		$branches{$branch}{"LAST"} = $ref;
-		$current_rev = $num+1 if $current_rev < $num+1;
+		$current_rev = $num if $current_rev < $num;
 	}
 	close($B);
 }
@@ -557,11 +557,6 @@ sub commit {
 		close(C)
 			or die "Cannot write branch $dest for update: $!\n";
 	}
-	$branches{$branch}{"LAST"} = $cid;
-	$branches{$branch}{$revision} = $cid;
-	$last_rev = $cid;
-	print BRANCHES "$revision $branch $cid\n";
-	print "DONE: $revision $dest $cid\n" if $opt_v;
 
 	if($tag) {
 		my($in, $out) = ('','');
@@ -593,6 +588,11 @@ sub commit {
 
 		print "Created tag '$dest' on '$branch'\n" if $opt_v;
 	}
+	$branches{$branch}{"LAST"} = $cid;
+	$branches{$branch}{$revision} = $cid;
+	$last_rev = $cid;
+	print BRANCHES "$revision $branch $cid\n";
+	print "DONE: $revision $dest $cid\n" if $opt_v;
 }
 
 my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
-- 
cgit v0.10.2-6-g49f6


From 3ef378a67b5b39441837700ab4dfb800eb443c5a Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 18:45:00 +0200
Subject: svn import: Add a loop limit option

The svn library has a serious memory leak.
Added a new option (-l NUM) which causes git-svnimport to exit cleanly
after fetching that many changes, in order to .

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/Documentation/git-svnimport.txt b/Documentation/git-svnimport.txt
index a144c6c..047f8f3 100644
--- a/Documentation/git-svnimport.txt
+++ b/Documentation/git-svnimport.txt
@@ -10,7 +10,7 @@ git-svnimport - Import a SVN repository into git
 SYNOPSIS
 --------
 'git-svnimport' [ -o <branch-for-HEAD> ] [ -h ] [ -v ]
-			[ -C <GIT_repository> ] [ -i ] [ -u ]
+			[ -C <GIT_repository> ] [ -i ] [ -u ] [-l limit_nr_changes]
 			[ -b branch_subdir ] [ -t trunk_subdir ] [ -T tag_subdir ]
 			[ -s start_chg ] [ -m ] [ -M regex ] [ <SVN_repository_URL> ]
 
@@ -70,6 +70,15 @@ When importing incementally, you might need to edit the .git/svn2git file.
 	regex. It can be used with -m to also see the default regexes. 
 	You must escape forward slashes. 
 
+-l <max_num_changes>::
+	Limit the number of SVN changesets we pull before quitting.
+	This option is necessary because the SVN library has serious memory
+	leaks; the recommended value for nontrivial imports is 100.
+
+	git-svnimport will still exit with a zero exit code. You can check
+	the size of the file ".git/svn2git" to determine whether to call
+	the importer again.
+
 -v::
 	Verbosity: let 'svnimport' report what it is doing.
 
diff --git a/git-svnimport.perl b/git-svnimport.perl
index 00f563d..b880297 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -31,19 +31,19 @@ die "Need CVN:COre 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
 $SIG{'PIPE'}="IGNORE";
 $ENV{'TZ'}="UTC";
 
-our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b,$opt_s);
+our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b,$opt_s,$opt_l);
 
 sub usage() {
 	print STDERR <<END;
 Usage: ${\basename $0}     # fetch/update GIT from CVS
-       [-o branch-for-HEAD] [-h] [-v]
+       [-o branch-for-HEAD] [-h] [-v] [-l max_num_changes]
        [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
        [-i] [-u] [-s start_chg] [-m] [-M regex] [SVN_URL]
 END
 	exit(1);
 }
 
-getopts("b:C:hivmM:o:s:t:T:u") or usage();
+getopts("b:C:hil:mM:o:s:t:T:uv") or usage();
 usage if $opt_h;
 
 my $tag_name = $opt_t || "tags";
@@ -53,6 +53,7 @@ my $branch_name = $opt_b || "branches";
 @ARGV <= 1 or usage();
 
 $opt_o ||= "origin";
+$opt_l = 100 unless defined $opt_l;
 my $git_tree = $opt_C;
 $git_tree ||= ".";
 
@@ -624,6 +625,10 @@ sub commit_all {
 while(++$current_rev < $svn->{'maxrev'}) {
 	$svn->{'svn'}->get_log("/",$current_rev,$current_rev,$current_rev,1,1,\&_commit_all,"");
 	commit_all();
+	if($opt_l and not --$opt_l) {
+		print STDERR "Exiting due to a memory leak. Repeat, please.\n";
+		last;
+	}
 }
 
 
@@ -637,7 +642,7 @@ if (defined $orig_git_index) {
 
 # Now switch back to the branch we were in before all of this happened
 if($orig_branch) {
-	print "DONE\n" if $opt_v;
+	print "DONE\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 	system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 		if $forward_master;
 	unless ($opt_i) {
@@ -646,7 +651,7 @@ if($orig_branch) {
 	}
 } else {
 	$orig_branch = "master";
-	print "DONE; creating $orig_branch branch\n" if $opt_v;
+	print "DONE; creating $orig_branch branch\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
 	system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
 		unless -f "$git_dir/refs/heads/master";
 	unlink("$git_dir/HEAD");
-- 
cgit v0.10.2-6-g49f6


From 22dcbb75129b3124a9fd71ed449030b79093b634 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 18:54:53 +0200
Subject: svn import: unlink downlaoded files

Actually removing the files that have been checked out of SVN,
after checking them into git of course, is a good idea...

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index b880297..102fa6e 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -296,6 +296,7 @@ sub get_file($$$) {
 	my $sha = <$F>;
 	chomp $sha;
 	close $F;
+	unlink $name;
 	my $mode = "0644"; # SV does not seem to store any file modes
 	return [$mode, $sha, $path];
 }
-- 
cgit v0.10.2-6-g49f6


From 8470b7f3a3cd5a70ff0b05486e335d351843890f Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Mon, 10 Oct 2005 20:10:48 +0200
Subject: svn import: get all revisions

Not skipping the last revision is generally seen as Good Thing. ;-)

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 102fa6e..f9318c8 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -623,7 +623,7 @@ sub commit_all {
 	}
 }
 
-while(++$current_rev < $svn->{'maxrev'}) {
+while(++$current_rev <= $svn->{'maxrev'}) {
 	$svn->{'svn'}->get_log("/",$current_rev,$current_rev,$current_rev,1,1,\&_commit_all,"");
 	commit_all();
 	if($opt_l and not --$opt_l) {
-- 
cgit v0.10.2-6-g49f6


From 16e685967d638fd50860fa24bd5f2d06ab4f8e96 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Tue, 11 Oct 2005 14:18:01 +0200
Subject: svn import: typo fix

Fixed a minor typo

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index f9318c8..821f51f 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -26,7 +26,7 @@ use IPC::Open2;
 use SVN::Core;
 use SVN::Ra;
 
-die "Need CVN:COre 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
+die "Need CVN:Core 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
 
 $SIG{'PIPE'}="IGNORE";
 $ENV{'TZ'}="UTC";
-- 
cgit v0.10.2-6-g49f6


From 2fa92046a8ae8548a8dff15abbf730506d49009e Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Tue, 11 Oct 2005 16:22:03 +0200
Subject: svn import: make -s option actually optional

The -s option was accidentally not optional.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 821f51f..0462c31 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -53,6 +53,7 @@ my $branch_name = $opt_b || "branches";
 @ARGV <= 1 or usage();
 
 $opt_o ||= "origin";
+$opt_s ||= 1;
 $opt_l = 100 unless defined $opt_l;
 my $git_tree = $opt_C;
 $git_tree ||= ".";
@@ -193,7 +194,7 @@ $ENV{GIT_INDEX_FILE} = $git_index;
 my $maxnum = 0;
 my $last_rev = "";
 my $last_branch;
-my $current_rev = $opt_s ? ($opt_s-1) : 0;
+my $current_rev = $opt_s-1;
 unless(-d $git_dir) {
 	system("git-init-db");
 	die "Cannot init the GIT db at $git_tree: $?\n" if $?;
-- 
cgit v0.10.2-6-g49f6


From 4a91b796e3f48ad01d06cce46df28321b7a1500b Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Tue, 11 Oct 2005 17:02:45 +0200
Subject: svn import: remove some CVS cruft

Some remains of CVS still lingered. Removed.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 0462c31..b976841 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -5,8 +5,7 @@
 #
 # The basic idea is to pull and analyze SVN changes.
 #
-# Checking out the files is done by a single long-running CVS connection
-# / server process.
+# Checking out the files is done by a single long-running SVN connection.
 #
 # The head revision is on branch "origin" by default.
 # You can change that with the '-o' option.
@@ -50,7 +49,7 @@ my $tag_name = $opt_t || "tags";
 my $trunk_name = $opt_T || "trunk";
 my $branch_name = $opt_b || "branches";
 
-@ARGV <= 1 or usage();
+@ARGV == 1 or usage();
 
 $opt_o ||= "origin";
 $opt_s ||= 1;
@@ -58,18 +57,7 @@ $opt_l = 100 unless defined $opt_l;
 my $git_tree = $opt_C;
 $git_tree ||= ".";
 
-my $cvs_tree;
-if ($#ARGV == 0) {
-	$cvs_tree = $ARGV[0];
-} elsif (-f 'CVS/Repository') {
-	open my $f, '<', 'CVS/Repository' or 
-	    die 'Failed to open CVS/Repository';
-	$cvs_tree = <$f>;
-	chomp $cvs_tree;
-	close $f;
-} else {
-	usage();
-}
+my $svn_url = $ARGV[0];
 
 our @mergerx = ();
 if ($opt_m) {
@@ -140,7 +128,7 @@ sub file {
 
 package main;
 
-my $svn = SVNconn->new($cvs_tree);
+my $svn = SVNconn->new($svn_url);
 
 
 sub pdate($) {
@@ -255,25 +243,6 @@ EOM
 
 open BRANCHES,">>", "$git_dir/svn2git";
 
-
-## cvsps output:
-#---------------------
-#PatchSet 314
-#Date: 1999/09/18 13:03:59
-#Author: wkoch
-#Branch: STABLE-BRANCH-1-0
-#Ancestor branch: HEAD
-#Tag: (none)
-#Log:
-#    See ChangeLog: Sat Sep 18 13:03:28 CEST 1999  Werner Koch
-#Members:
-#	README:1.57->1.57.2.1
-#	VERSION:1.96->1.96.2.1
-#
-#---------------------
-
-my $state = 0;
-
 sub get_file($$$) {
 	my($rev,$branch,$path) = @_;
 
-- 
cgit v0.10.2-6-g49f6


From 25f6f325d7a8f7cb686a9ffd9fa2c00b3aa85a60 Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Tue, 11 Oct 2005 18:13:30 +0200
Subject: svn import: Add direct HTTP access

Some SVN repositories that are accessible through HTTP don't like when I
retrieve files using SVN methods ("internal server error").

Therefore, I added an option to get the contents using (persistent) HTTP
directly. This also reduces round-trip time, from two or three requests
down to one.

Also corrected error handling a bit.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/Documentation/git-svnimport.txt b/Documentation/git-svnimport.txt
index 047f8f3..a27a835 100644
--- a/Documentation/git-svnimport.txt
+++ b/Documentation/git-svnimport.txt
@@ -9,10 +9,11 @@ git-svnimport - Import a SVN repository into git
 
 SYNOPSIS
 --------
-'git-svnimport' [ -o <branch-for-HEAD> ] [ -h ] [ -v ]
+'git-svnimport' [ -o <branch-for-HEAD> ] [ -h ] [ -v ] [ -d | -D ]
 			[ -C <GIT_repository> ] [ -i ] [ -u ] [-l limit_nr_changes]
 			[ -b branch_subdir ] [ -t trunk_subdir ] [ -T tag_subdir ]
-			[ -s start_chg ] [ -m ] [ -M regex ] [ <SVN_repository_URL> ]
+			[ -s start_chg ] [ -m ] [ -M regex ]
+			<SVN_repository_URL> [ <path> ]
 
 
 DESCRIPTION
@@ -82,9 +83,32 @@ When importing incementally, you might need to edit the .git/svn2git file.
 -v::
 	Verbosity: let 'svnimport' report what it is doing.
 
+-d::
+	Use direct HTTP requests if possible. The "<path>" argument is used
+	only for retrieving the SVN logs; the path to the contents is
+	included in the SVN log.
+
+-D::
+	Use direct HTTP requests if possible. The "<path>" argument is used
+	for retrieving the logs, as well as for the contents.
++
+There's no safe way to automatically find out which of these options to
+use, so you need to try both. Usually, the one that's wrong will die
+with a 40x error pretty quickly.
+
 <SVN_repository_URL>::
 	The URL of the SVN module you want to import. For local
 	repositories, use "file:///absolute/path".
++
+If you're using the "-d" or "-D" option, this is the URL of the SVN
+repository itself; it usually ends in "/svn".
+
+<SVN_repository_URL>::
+	The URL of the SVN module you want to import. For local
+	repositories, use "file:///absolute/path".
+
+<path>
+	The path to the module you want to check out.
 
 -h::
 	Print a short usage message and exit.
diff --git a/git-svnimport.perl b/git-svnimport.perl
index b976841..2f89c31 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -30,26 +30,26 @@ die "Need CVN:Core 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
 $SIG{'PIPE'}="IGNORE";
 $ENV{'TZ'}="UTC";
 
-our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b,$opt_s,$opt_l);
+our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b,$opt_s,$opt_l,$opt_d,$opt_D);
 
 sub usage() {
 	print STDERR <<END;
 Usage: ${\basename $0}     # fetch/update GIT from CVS
        [-o branch-for-HEAD] [-h] [-v] [-l max_num_changes]
        [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
-       [-i] [-u] [-s start_chg] [-m] [-M regex] [SVN_URL]
+       [-d|-D] [-i] [-u] [-s start_chg] [-m] [-M regex] [SVN_URL]
 END
 	exit(1);
 }
 
-getopts("b:C:hil:mM:o:s:t:T:uv") or usage();
+getopts("b:C:dDhil:mM:o:s:t:T:uv") or usage();
 usage if $opt_h;
 
 my $tag_name = $opt_t || "tags";
 my $trunk_name = $opt_T || "trunk";
 my $branch_name = $opt_b || "branches";
 
-@ARGV == 1 or usage();
+@ARGV == 1 or @ARGV == 2 or usage();
 
 $opt_o ||= "origin";
 $opt_s ||= 1;
@@ -58,6 +58,7 @@ my $git_tree = $opt_C;
 $git_tree ||= ".";
 
 my $svn_url = $ARGV[0];
+my $svn_dir = $ARGV[1];
 
 our @mergerx = ();
 if ($opt_m) {
@@ -106,30 +107,44 @@ sub conn {
 
 sub file {
 	my($self,$path,$rev) = @_;
-	my $res;
 
 	my ($fh, $name) = tempfile('gitsvn.XXXXXX', 
 		    DIR => File::Spec->tmpdir(), UNLINK => 1);
 
 	print "... $rev $path ...\n" if $opt_v;
 	eval { $self->{'svn'}->get_file($path,$rev,$fh); };
-	if ($@ and $@ !~ /Attempted to get checksum/) {
-	    # retry
-	    $self->conn();
-		eval { $self->{'svn'}->get_file($path,$rev,$fh); };
-	};
-	return () if $@ and $@ !~ /Attempted to get checksum/;
-	die $@ if $@;
+	if($@) {
+		return undef if $@ =~ /Attempted to get checksum/;
+		die $@;
+	}
 	close ($fh);
 
-	return ($name, $res);
+	return $name;
 }
 
-
 package main;
+use URI;
 
-my $svn = SVNconn->new($svn_url);
+my $svn = $svn_url;
+$svn .= "/$svn_dir" if defined $svn_dir;
+$svn = SVNconn->new($svn);
 
+my $lwp_ua;
+if($opt_d or $opt_D) {
+	$svn_url = URI->new($svn_url)->canonical;
+	if($opt_D) {
+		$svn_dir =~ s#/*$#/#;
+	} else {
+		$svn_dir = "";
+	}
+	if ($svn_url->scheme eq "http") {
+		use LWP::UserAgent;
+		$lwp_ua = LWP::UserAgent->new(keep_alive => 1, requests_redirectable => []);
+	} else {
+		print STDERR "Warning: not HTTP; turning off direct file access\n";
+		$opt_d=0;
+	}
+}
 
 sub pdate($) {
 	my($d) = @_;
@@ -258,8 +273,30 @@ sub get_file($$$) {
 	}
 
 	# now get it
-	my ($name, $res) = eval { $svn->file($svnpath,$rev); };
-	return () unless defined $name;
+	my $name;
+	if($opt_d) {
+		my($req,$res);
+
+		# /svn/!svn/bc/2/django/trunk/django-docs/build.py
+		my $url=$svn_url->clone();
+		$url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath");
+		print "Fetching $url...\n" if $opt_v;
+		$req = HTTP::Request->new(GET => $url);
+		$res = $lwp_ua->request($req);
+		if ($res->is_success) {
+			my $fh;
+			($fh, $name) = tempfile('gitsvn.XXXXXX', 
+		    	DIR => File::Spec->tmpdir(), UNLINK => 1);
+			print $fh $res->content;
+			close($fh) or die "Could not write $name: $!\n";
+		} else {
+			return undef if $res->code == 301; # directory?
+			die $res->status_line." at $url\n";
+		}
+	} else {
+		$name = $svn->file($svnpath,$rev);
+		return undef unless defined $name;
+	}
 
 	open my $F, '-|', "git-hash-object", "-w", $name
 		or die "Cannot create object: $!\n";
-- 
cgit v0.10.2-6-g49f6


From 0090a6185af18a199d5bc12781adf411177f1e1a Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Tue, 11 Oct 2005 19:42:27 +0200
Subject: svn import: copy directories

Import SVN-copied and -referenced directories correctly.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index 2f89c31..aa8fa40 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -265,11 +265,11 @@ sub get_file($$$) {
 	my $svnpath;
 	$path = "" if $path eq "/"; # this should not happen, but ...
 	if($branch eq "/") {
-		$svnpath = "/$trunk_name/$path";
+		$svnpath = "$trunk_name/$path";
 	} elsif($branch =~ m#^/#) {
-		$svnpath = "/$tag_name$branch/$path";
+		$svnpath = "$tag_name$branch/$path";
 	} else {
-		$svnpath = "/$branch_name/$branch/$path";
+		$svnpath = "$branch_name/$branch/$path";
 	}
 
 	# now get it
@@ -280,7 +280,7 @@ sub get_file($$$) {
 		# /svn/!svn/bc/2/django/trunk/django-docs/build.py
 		my $url=$svn_url->clone();
 		$url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath");
-		print "Fetching $url...\n" if $opt_v;
+		print "... $path...\n" if $opt_v;
 		$req = HTTP::Request->new(GET => $url);
 		$res = $lwp_ua->request($req);
 		if ($res->is_success) {
@@ -294,7 +294,7 @@ sub get_file($$$) {
 			die $res->status_line." at $url\n";
 		}
 	} else {
-		$name = $svn->file($svnpath,$rev);
+		$name = $svn->file("/$svnpath",$rev);
 		return undef unless defined $name;
 	}
 
@@ -326,6 +326,36 @@ sub split_path($$) {
 	return ($branch,$path);
 }
 
+sub copy_subdir($$$$$$) {
+	# Somebody copied a whole subdirectory.
+	# We need to find the index entries from the old version which the
+	# SVN log entry points to, and add them to the new place.
+
+	my($newrev,$newbranch,$path,$oldpath,$rev,$new) = @_;
+	my($branch,$srcpath) = split_path($rev,$oldpath);
+
+	my $gitrev = $branches{$branch}{$rev};
+	unless($gitrev) {
+		print STDERR "$newrev:$newbranch: could not find $oldpath \@ $rev\n";
+		return;
+	}
+	print "$newrev:$newbranch:$path: copying from $branch:$srcpath @ $rev\n" if $opt_v;
+	$srcpath =~ s#/*$#/#;
+	open my $f,"-|","git-ls-tree","-r","-z",$gitrev,$srcpath;
+	local $/ = "\0";
+	while(<$f>) {
+		chomp;
+		my($m,$p) = split(/\t/,$_,2);
+		my($mode,$type,$sha1) = split(/ /,$m);
+		next if $type ne "blob";
+		$p = substr($p,length($srcpath)-1);
+		print "... found $path$p ...\n" if $opt_v;
+		push(@$new,[$mode,$sha1,$path.$p]);
+	}
+	close($f) or 
+		print STDERR "$newrev:$newbranch: could not list files in $oldpath \@ $rev\n";
+}
+
 sub commit {
 	my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
 	my($author_name,$author_email,$dest);
@@ -420,10 +450,20 @@ sub commit {
 	if($tag and not %$changed_paths) {
 		$cid = $rev;
 	} else {
-		while(my($path,$action) = each %$changed_paths) {
+		my @paths = sort keys %$changed_paths;
+		foreach my $path(@paths) {
+			my $action = $changed_paths->{$path};
+
 			if ($action->[0] eq "A") {
 				my $f = get_file($revision,$branch,$path);
-				push(@new,$f) if $f;
+				if($f) {
+					push(@new,$f) if $f;
+				} elsif($action->[1]) {
+					copy_subdir($revision,$branch,$path,$action->[1],$action->[2],\@new);
+				} else {
+					my $opath = $action->[3];
+					print STDERR "$revision: $branch: could not fetch '$opath'\n";
+				}
 			} elsif ($action->[0] eq "D") {
 				push(@old,$path);
 			} elsif ($action->[0] eq "M") {
@@ -434,16 +474,19 @@ sub commit {
 				push(@old,$path); # remove any old stuff
 
 				# ... and add any new stuff
-				my($b,$p) = split_path($revision,$action->[1]);
-				open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $p;
-				local $/ = '\0';
+				my($b,$srcpath) = split_path($revision,$action->[1]);
+				$srcpath =~ s#/*$#/#;
+				open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $srcpath;
+				local $/ = "\0";
 				while(<$F>) {
 					chomp;
 					my($m,$p) = split(/\t/,$_,2);
 					my($mode,$type,$sha1) = split(/ /,$m);
 					next if $type ne "blob";
-					push(@new,[$mode,$sha1,$p]);
+					$p = substr($p,length($srcpath)-1);
+					push(@new,[$mode,$sha1,$path.$p]);
 				}
+				close($F);
 			} else {
 				die "$revision: unknown action '".$action->[0]."' for $path\n";
 			}
@@ -452,7 +495,7 @@ sub commit {
 		if(@old) {
 			open my $F, "-|", "git-ls-files", "-z", @old or die $!;
 			@old = ();
-			local $/ = '\0';
+			local $/ = "\0";
 			while(<$F>) {
 				chomp;
 				push(@old,$_);
@@ -609,7 +652,7 @@ sub _commit_all {
 	($changed_paths, $revision, $author, $date, $message, $pool) = @_;
 	my %p;
 	while(my($path,$action) = each %$changed_paths) {
-		$p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev ];
+		$p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev, $path ];
 	}
 	$changed_paths = \%p;
 }
-- 
cgit v0.10.2-6-g49f6


From f005dba7c15af5de74e05f8667da4e5d27cf6d1b Mon Sep 17 00:00:00 2001
From: Kalle Valo <Kalle.Valo@iki.fi>
Date: Sun, 16 Oct 2005 19:37:25 +0200
Subject: Makefile entry for git-svnimport contained a small typo.

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/Makefile b/Makefile
index 30fda9e..13b949b 100644
--- a/Makefile
+++ b/Makefile
@@ -92,7 +92,7 @@ SCRIPT_SH = \
 SCRIPT_PERL = \
 	git-archimport.perl git-cvsimport.perl git-relink.perl \
 	git-rename.perl git-shortlog.perl git-fmt-merge-msg.perl \
-	git-svnimport-perl
+	git-svnimport.perl
 
 SCRIPT_PYTHON = \
 	git-merge-recursive.py
-- 
cgit v0.10.2-6-g49f6


From 40dad96e41fcb96e31bdf11deec3c7bf6261adbe Mon Sep 17 00:00:00 2001
From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Sun, 16 Oct 2005 19:57:38 +0200
Subject: svn commit: re-word the exit-due-to-memory-leak message

Reworded the exit message, as per Kalle Valo's suggestion (but shorter).

Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>

diff --git a/git-svnimport.perl b/git-svnimport.perl
index aa8fa40..45486a8 100755
--- a/git-svnimport.perl
+++ b/git-svnimport.perl
@@ -677,7 +677,8 @@ while(++$current_rev <= $svn->{'maxrev'}) {
 	$svn->{'svn'}->get_log("/",$current_rev,$current_rev,$current_rev,1,1,\&_commit_all,"");
 	commit_all();
 	if($opt_l and not --$opt_l) {
-		print STDERR "Exiting due to a memory leak. Repeat, please.\n";
+		print STDERR "Stopping, because there is a memory leak (in the SVN library).\n";
+		print STDERR "Please repeat this command; it will continue safely\n";
 		last;
 	}
 }
-- 
cgit v0.10.2-6-g49f6