3 # This tool is copyright (c) 2005, Matthias Urlichs.
4 # It is released under the Gnu Public License, version 2.
6 # The basic idea is to pull and analyze SVN changes.
8 # Checking out the files is done by a single long-running SVN connection.
10 # The head revision is on branch "origin" by default.
11 # You can change that with the '-o' option.
13 require 5.008; # for shell-safe open("-|",LIST)
18 use File::Temp qw(tempfile);
19 use File::Path qw(mkpath);
20 use File::Basename qw(basename dirname);
23 use POSIX qw(strftime dup2);
28 die "Need CVN:Core 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
30 $SIG{'PIPE'}="IGNORE";
33 our($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);
37 Usage: ${\basename $0} # fetch/update GIT from CVS
38 [-o branch-for-HEAD] [-h] [-v] [-l max_num_changes]
39 [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
40 [-d|-D] [-i] [-u] [-s start_chg] [-m] [-M regex] [SVN_URL]
45 getopts("b:C:dDhil:mM:o:s:t:T:uv") or usage();
48 my $tag_name = $opt_t || "tags";
49 my $trunk_name = $opt_T || "trunk";
50 my $branch_name = $opt_b || "branches";
52 @ARGV == 1 or @ARGV == 2 or usage();
56 $opt_l = 100 unless defined $opt_l;
57 my $git_tree = $opt_C;
60 my $svn_url = $ARGV[0];
61 my $svn_dir = $ARGV[1];
65 @mergerx = ( qr/\W(?:from|of|merge|merging|merged) (\w+)/i );
68 push (@mergerx, qr/$opt_M/);
71 select(STDERR); $|=1; select(STDOUT);
75 # Basic SVN connection.
76 # We're only interested in connecting and downloading, so ...
79 use File::Temp qw(tempfile);
80 use POSIX qw(strftime dup2);
84 $what=ref($what) if ref($what);
87 $self->{'buffer'} = "";
91 $self->{'fullrep'} = $repo;
99 my $repo = $self->{'fullrep'};
100 my $s = SVN::Ra->new($repo);
102 die "SVN connection to $repo: $!\n" unless defined $s;
104 $self->{'repo'} = $repo;
105 $self->{'maxrev'} = $s->get_latest_revnum();
109 my($self,$path,$rev) = @_;
111 my ($fh, $name) = tempfile('gitsvn.XXXXXX',
112 DIR => File::Spec->tmpdir(), UNLINK => 1);
114 print "... $rev $path ...\n" if $opt_v;
115 my $pool = SVN::Pool->new();
116 eval { $self->{'svn'}->get_file($path,$rev,$fh,$pool); };
119 return undef if $@ =~ /Attempted to get checksum/;
131 $svn .= "/$svn_dir" if defined $svn_dir;
132 $svn = SVNconn->new($svn);
135 if($opt_d or $opt_D) {
136 $svn_url = URI->new($svn_url)->canonical;
138 $svn_dir =~ s#/*$#/#;
142 if ($svn_url->scheme eq "http") {
144 $lwp_ua = LWP::UserAgent->new(keep_alive => 1, requests_redirectable => []);
146 print STDERR "Warning: not HTTP; turning off direct file access\n";
153 $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
154 or die "Unparseable date: $d\n";
155 my $y=$1; $y-=1900 if $y>1900;
156 return timegm($6||0,$5,$4,$3,$2-1,$y);
166 sub get_headref($$) {
171 if (open(C,"$git_dir/refs/heads/$name")) {
175 or die "Cannot get head id for $name ($sha): $!\n";
182 or mkdir($git_tree,0777)
183 or die "Could not create $git_tree: $!";
186 my $orig_branch = "";
187 my $forward_master = 0;
190 my $git_dir = $ENV{"GIT_DIR"} || ".git";
191 $git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
192 $ENV{"GIT_DIR"} = $git_dir;
194 $orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
195 my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
196 DIR => File::Spec->tmpdir());
198 $ENV{GIT_INDEX_FILE} = $git_index;
202 my $current_rev = $opt_s-1;
203 unless(-d $git_dir) {
204 system("git-init-db");
205 die "Cannot init the GIT db at $git_tree: $?\n" if $?;
206 system("git-read-tree");
207 die "Cannot init an empty tree: $?\n" if $?;
209 $last_branch = $opt_o;
212 -f "$git_dir/refs/heads/$opt_o"
213 or die "Branch '$opt_o' does not exist.\n".
214 "Either use the correct '-o branch' option,\n".
215 "or import to a new repository.\n";
217 -f "$git_dir/svn2git"
218 or die "'$git_dir/svn2git' does not exist.\n".
219 "You need that file for incremental imports.\n";
220 $last_branch = basename(readlink("$git_dir/HEAD"));
221 unless($last_branch) {
222 warn "Cannot read the last branch name: $! -- assuming 'master'\n";
223 $last_branch = "master";
225 $orig_branch = $last_branch;
226 $last_rev = get_headref($orig_branch, $git_dir);
227 if (-f "$git_dir/SVN2GIT_HEAD") {
230 Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
233 git-read-tree -m -u SVN2GIT_HEAD HEAD
236 system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
239 $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
240 system('cmp', '-s', "$git_dir/refs/heads/master",
241 "$git_dir/refs/heads/$opt_o") == 0;
244 system('git-read-tree', $last_rev);
245 die "read-tree failed: $?\n" if $?;
247 # Get the last import timestamps
248 open my $B,"<", "$git_dir/svn2git";
251 my($num,$branch,$ref) = split;
252 $branches{$branch}{$num} = $ref;
253 $branches{$branch}{"LAST"} = $ref;
254 $current_rev = $num if $current_rev < $num;
259 or die "Could not create git subdir ($git_dir).\n";
261 open BRANCHES,">>", "$git_dir/svn2git";
264 my ($branch, $path, $revision) = @_;
265 my $pool=SVN::Pool->new;
266 my $kind = $svn->{'svn'}->check_path(revert_split_path($branch,$path),$revision,$pool);
271 sub revert_split_path($$) {
272 my($branch,$path) = @_;
275 $path = "" if $path eq "/"; # this should not happen, but ...
277 $svnpath = "$trunk_name/$path";
278 } elsif($branch =~ m#^/#) {
279 $svnpath = "$tag_name$branch/$path";
281 $svnpath = "$branch_name/$branch/$path";
288 my($rev,$branch,$path) = @_;
290 my $svnpath = revert_split_path($branch,$path);
297 # /svn/!svn/bc/2/django/trunk/django-docs/build.py
298 my $url=$svn_url->clone();
299 $url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath");
300 print "... $path...\n" if $opt_v;
301 $req = HTTP::Request->new(GET => $url);
302 $res = $lwp_ua->request($req);
303 if ($res->is_success) {
305 ($fh, $name) = tempfile('gitsvn.XXXXXX',
306 DIR => File::Spec->tmpdir(), UNLINK => 1);
307 print $fh $res->content;
308 close($fh) or die "Could not write $name: $!\n";
310 return undef if $res->code == 301; # directory?
311 die $res->status_line." at $url\n";
314 $name = $svn->file("/$svnpath",$rev);
315 return undef unless defined $name;
318 open my $F, '-|', "git-hash-object", "-w", $name
319 or die "Cannot create object: $!\n";
324 my $mode = "0644"; # SV does not seem to store any file modes
325 return [$mode, $sha, $path];
332 if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
334 } elsif($path =~ s#^/\Q$trunk_name\E/?##) {
336 } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
339 print STDERR "$rev: Unrecognized path: $path\n";
342 $path = "/" if $path eq "";
343 return ($branch,$path);
346 sub copy_subdir($$$$$$) {
347 # Somebody copied a whole subdirectory.
348 # We need to find the index entries from the old version which the
349 # SVN log entry points to, and add them to the new place.
351 my($newrev,$newbranch,$path,$oldpath,$rev,$new) = @_;
352 my($branch,$srcpath) = split_path($rev,$oldpath);
354 my $gitrev = $branches{$branch}{$rev};
356 print STDERR "$newrev:$newbranch: could not find $oldpath \@ $rev\n";
359 print "$newrev:$newbranch:$path: copying from $branch:$srcpath @ $rev\n" if $opt_v;
360 $srcpath =~ s#/*$#/#;
361 open my $f,"-|","git-ls-tree","-r","-z",$gitrev,$srcpath;
365 my($m,$p) = split(/\t/,$_,2);
366 my($mode,$type,$sha1) = split(/ /,$m);
367 next if $type ne "blob";
368 $p = substr($p,length($srcpath)-1);
369 print "... found $path$p ...\n" if $opt_v;
370 push(@$new,[$mode,$sha1,$path.$p]);
373 print STDERR "$newrev:$newbranch: could not list files in $oldpath \@ $rev\n";
377 my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
378 my($author_name,$author_email,$dest);
381 if (not defined $author) {
382 $author_name = $author_email = "unknown";
383 } elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
384 ($author_name, $author_email) = ($1, $2);
386 $author =~ s/^<(.*)>$/$1/;
387 $author_name = $author_email = $author;
389 $date = pdate($date);
393 if($branch eq "/") { # trunk
395 } elsif($branch =~ m#^/(.+)#) { # tag
398 } else { # "normal" branch
404 my $prev = $changed_paths->{"/"};
405 if($prev and $prev->[0] eq "A") {
406 delete $changed_paths->{"/"};
407 my $oldpath = $prev->[1];
409 if(defined $oldpath) {
411 ($parent,$p) = split_path($revision,$oldpath);
415 $parent =~ s#^/##; # if it's a tag
423 if($revision > $opt_s and defined $parent) {
424 open(H,"git-rev-parse --verify $parent |");
427 print STDERR "$revision: cannot find commit '$parent'!\n";
431 if(length($rev) != 40) {
432 print STDERR "$revision: cannot find commit '$parent'!\n";
435 $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
436 if($revision != $opt_s and not $rev) {
437 print STDERR "$revision: do not know ancestor for '$parent'!\n";
444 # if($prev and $prev->[0] eq "A") {
446 # unless(open(H,"> $git_dir/refs/heads/$branch")) {
447 # print STDERR "$revision: Could not create branch $branch: $!\n";
452 # or die "Could not write branch $branch: $!";
454 # or die "Could not write branch $branch: $!";
457 if(not defined $rev) {
459 } elsif ($rev ne $last_rev) {
460 print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
461 system("git-read-tree", $rev);
462 die "read-tree failed for $rev: $?\n" if $?;
467 if($tag and not %$changed_paths) {
470 my @paths = sort keys %$changed_paths;
471 foreach my $path(@paths) {
472 my $action = $changed_paths->{$path};
474 if ($action->[0] eq "A") {
475 my $f = get_file($revision,$branch,$path);
478 } elsif($action->[1]) {
479 copy_subdir($revision,$branch,$path,$action->[1],$action->[2],\@new);
481 my $opath = $action->[3];
482 print STDERR "$revision: $branch: could not fetch '$opath'\n";
484 } elsif ($action->[0] eq "D") {
486 } elsif ($action->[0] eq "M") {
487 my $f = get_file($revision,$branch,$path);
489 } elsif ($action->[0] eq "R") {
490 # refer to a file/tree in an earlier commit
491 push(@old,$path); # remove any old stuff
493 # ... and add any new stuff
494 my($b,$srcpath) = split_path($revision,$action->[1]);
495 $srcpath =~ s#/*$#/#;
496 open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $srcpath;
500 my($m,$p) = split(/\t/,$_,2);
501 my($mode,$type,$sha1) = split(/ /,$m);
502 next if $type ne "blob";
503 $p = substr($p,length($srcpath)-1);
504 push(@new,[$mode,$sha1,$path.$p]);
508 die "$revision: unknown action '".$action->[0]."' for $path\n";
513 open my $F, "-|", "git-ls-files", "-z", @old or die $!;
525 @o2 = splice(@old,0,50);
530 system("git-update-index","--force-remove","--",@o2);
531 die "Cannot remove files: $?\n" if $?;
537 @n2 = splice(@new,0,10);
542 system("git-update-index","--add",
543 (map { ('--cacheinfo', @$_) } @n2));
544 die "Cannot add files: $?\n" if $?;
547 my $pid = open(C,"-|");
548 die "Cannot fork: $!" unless defined $pid;
550 exec("git-write-tree");
551 die "Cannot exec git-write-tree: $!\n";
553 chomp(my $tree = <C>);
555 or die "Cannot get tree id ($tree): $!\n";
557 or die "Error running git-write-tree: $?\n";
558 print "Tree ID $tree\n" if $opt_v;
560 my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
561 my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
563 die "Fork: $!\n" unless defined $pid;
567 open(OUT,">&STDOUT");
568 dup2($pw->fileno(),0);
569 dup2($pr->fileno(),1);
574 @par = ("-p",$rev) if defined $rev;
576 # loose detection of merges
577 # based on the commit msg
578 foreach my $rx (@mergerx) {
579 if ($message =~ $rx) {
581 if ($mparent eq 'HEAD') { $mparent = $opt_o };
582 if ( -e "$git_dir/refs/heads/$mparent") {
583 $mparent = get_headref($mparent, $git_dir);
584 push @par, '-p', $mparent;
585 print OUT "Merge parent branch: $mparent\n" if $opt_v;
591 "GIT_AUTHOR_NAME=$author_name",
592 "GIT_AUTHOR_EMAIL=$author_email",
593 "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
594 "GIT_COMMITTER_NAME=$author_name",
595 "GIT_COMMITTER_EMAIL=$author_email",
596 "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
597 "git-commit-tree", $tree,@par);
598 die "Cannot exec git-commit-tree: $!\n";
603 $message =~ s/[\s\n]+\z//;
605 print $pw "$message\n"
606 or die "Error writing to git-commit-tree: $!\n";
609 print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
612 or die "Cannot get commit id ($cid): $!\n";
613 print "Commit ID $cid\n" if $opt_v;
617 die "Error running git-commit-tree: $?\n" if $?;
620 if(not defined $dest) {
621 print "... no known parent\n" if $opt_v;
623 print "Writing to refs/heads/$dest\n" if $opt_v;
624 open(C,">$git_dir/refs/heads/$dest") and
625 print C ("$cid\n") and
627 or die "Cannot write branch $dest for update: $!\n";
631 my($in, $out) = ('','');
632 $last_rev = "-" if %$changed_paths;
633 # the tag was 'complex', i.e. did not refer to a "real" revision
635 $dest =~ tr/_/\./ if $opt_u;
637 my $pid = open2($in, $out, 'git-mktag');
638 print $out ("object $cid\n".
641 "tagger $author_name <$author_email>\n") and
643 or die "Cannot create tag object $dest: $!\n";
648 if ( !close($in) or waitpid($pid, 0) != $pid or
649 $? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
650 die "Cannot create tag object $dest: $!\n";
653 open(C,">$git_dir/refs/tags/$dest") and
654 print C ("$tagobj\n") and
656 or die "Cannot create tag $branch: $!\n";
658 print "Created tag '$dest' on '$branch'\n" if $opt_v;
660 $branches{$branch}{"LAST"} = $cid;
661 $branches{$branch}{$revision} = $cid;
663 print BRANCHES "$revision $branch $cid\n";
664 print "DONE: $revision $dest $cid\n" if $opt_v;
667 my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
669 ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
671 while(my($path,$action) = each %$changed_paths) {
672 $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev, $path ];
674 $changed_paths = \%p;
683 while(my($path,$action) = each %$changed_paths) {
684 ($branch,$path) = split_path($revision,$path);
685 next if not defined $branch;
686 $done{$branch}{$path} = $action;
688 while(($branch,$changed_paths) = each %done) {
689 commit($branch, $changed_paths, $revision, $author, $date, $message);
693 while(++$current_rev <= $svn->{'maxrev'}) {
694 my $pool=SVN::Pool->new;
695 $svn->{'svn'}->get_log("/",$current_rev,$current_rev,1,1,1,\&_commit_all,$pool);
698 if($opt_l and not --$opt_l) {
699 print STDERR "Stopping, because there is a memory leak (in the SVN library).\n";
700 print STDERR "Please repeat this command; it will continue safely\n";
708 if (defined $orig_git_index) {
709 $ENV{GIT_INDEX_FILE} = $orig_git_index;
711 delete $ENV{GIT_INDEX_FILE};
714 # Now switch back to the branch we were in before all of this happened
716 print "DONE\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
717 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
720 system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
721 die "read-tree failed: $?\n" if $?;
724 $orig_branch = "master";
725 print "DONE; creating $orig_branch branch\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
726 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
727 unless -f "$git_dir/refs/heads/master";
728 unlink("$git_dir/HEAD");
729 symlink("refs/heads/$orig_branch","$git_dir/HEAD");
731 system('git checkout');
732 die "checkout failed: $?\n" if $?;
735 unlink("$git_dir/SVN2GIT_HEAD");