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 CVS connection
11 # The head revision is on branch "origin" by default.
12 # You can change that with the '-o' option.
14 require v5.8.0; # for shell-safe open("-|",LIST)
19 use File::Temp qw(tempfile);
20 use File::Path qw(mkpath);
21 use File::Basename qw(basename dirname);
24 use POSIX qw(strftime dup2);
29 die "Need CVN:COre 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
31 $SIG{'PIPE'}="IGNORE";
34 our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T,$opt_b);
38 Usage: ${\basename $0} # fetch/update GIT from CVS
39 [-o branch-for-HEAD] [-h] [-v]
40 [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname]
41 [-i] [-u] [-s subst] [-m] [-M regex] [SVN_URL]
46 getopts("b:C:hivmM:o:t:T:u") or usage();
49 my $tag_name = $opt_t || "tags";
50 my $trunk_name = $opt_T || "trunk";
51 my $branch_name = $opt_b || "branches";
53 @ARGV <= 1 or usage();
56 my $git_tree = $opt_C;
62 } elsif (-f 'CVS/Repository') {
63 open my $f, '<', 'CVS/Repository' or
64 die 'Failed to open CVS/Repository';
74 @mergerx = ( qr/\W(?:from|of|merge|merging|merged) (\w+)/i );
77 push (@mergerx, qr/$opt_M/);
80 select(STDERR); $|=1; select(STDOUT);
84 # Basic SVN connection.
85 # We're only interested in connecting and downloading, so ...
88 use File::Temp qw(tempfile);
89 use POSIX qw(strftime dup2);
93 $what=ref($what) if ref($what);
96 $self->{'buffer'} = "";
100 $self->{'fullrep'} = $repo;
108 my $repo = $self->{'fullrep'};
109 my $s = SVN::Ra->new($repo);
111 die "SVN connection to $repo: $!\n" unless defined $s;
113 print STDERR "*** SVN *** $s ***\n";
114 $self->{'repo'} = $repo;
115 $self->{'maxrev'} = $s->get_latest_revnum();
119 my($self,$path,$rev) = @_;
122 my ($fh, $name) = tempfile('gitsvn.XXXXXX',
123 DIR => File::Spec->tmpdir(), UNLINK => 1);
125 print "... $rev $path ...\n" if $opt_v;
126 my $s = $self->{'svn'};
127 print STDERR "*** GET *** $s ***\n";
128 eval { $s->get_file($path,$rev,$fh); };
129 if ($@ and $@ !~ /Attempted to get checksum/) {
132 eval { $self->{'svn'}->get_file($path,$rev,$fh); };
134 return () if $@ and $@ !~ /Attempted to get checksum/;
138 return ($name, $res);
144 my $svn = SVNconn->new($cvs_tree);
149 $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
150 or die "Unparseable date: $d\n";
151 my $y=$1; $y-=1900 if $y>1900;
152 return timegm($6||0,$5,$4,$3,$2-1,$y);
162 sub get_headref($$) {
167 if (open(C,"$git_dir/refs/heads/$name")) {
171 or die "Cannot get head id for $name ($sha): $!\n";
178 or mkdir($git_tree,0777)
179 or die "Could not create $git_tree: $!";
182 my $orig_branch = "";
183 my $forward_master = 0;
186 my $git_dir = $ENV{"GIT_DIR"} || ".git";
187 $git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
188 $ENV{"GIT_DIR"} = $git_dir;
190 $orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
191 my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
192 DIR => File::Spec->tmpdir());
194 $ENV{GIT_INDEX_FILE} = $git_index;
199 unless(-d $git_dir) {
200 system("git-init-db");
201 die "Cannot init the GIT db at $git_tree: $?\n" if $?;
202 system("git-read-tree");
203 die "Cannot init an empty tree: $?\n" if $?;
205 $last_branch = $opt_o;
208 -f "$git_dir/refs/heads/$opt_o"
209 or die "Branch '$opt_o' does not exist.\n".
210 "Either use the correct '-o branch' option,\n".
211 "or import to a new repository.\n";
213 -f "$git_dir/svn2git"
214 or die "'$git_dir/svn2git' does not exist.\n".
215 "You need that file for incremental imports.\n";
216 $last_branch = basename(readlink("$git_dir/HEAD"));
217 unless($last_branch) {
218 warn "Cannot read the last branch name: $! -- assuming 'master'\n";
219 $last_branch = "master";
221 $orig_branch = $last_branch;
222 $last_rev = get_headref($orig_branch, $git_dir);
223 if (-f "$git_dir/SVN2GIT_HEAD") {
226 Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
229 git-read-tree -m -u SVN2GIT_HEAD HEAD
232 system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
235 $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
236 system('cmp', '-s', "$git_dir/refs/heads/master",
237 "$git_dir/refs/heads/$opt_o") == 0;
240 system('git-read-tree', $last_rev);
241 die "read-tree failed: $?\n" if $?;
243 # Get the last import timestamps
244 open my $B,"<", "$git_dir/svn2git";
247 my($num,$branch,$ref) = split;
248 $branches{$branch}{$num} = $ref;
249 $branches{$branch}{"LAST"} = $ref;
250 $current_rev = $num+1 if $current_rev < $num+1;
255 or die "Could not create git subdir ($git_dir).\n";
257 open BRANCHES,">>", "$git_dir/svn2git";
261 #---------------------
263 #Date: 1999/09/18 13:03:59
265 #Branch: STABLE-BRANCH-1-0
266 #Ancestor branch: HEAD
269 # See ChangeLog: Sat Sep 18 13:03:28 CEST 1999 Werner Koch
271 # README:1.57->1.57.2.1
272 # VERSION:1.96->1.96.2.1
274 #---------------------
279 my($rev,$branch,$path) = @_;
281 # revert split_path(), below
283 $path = "" if $path eq "/"; # this should not happen, but ...
285 $svnpath = "/$trunk_name/$path";
286 } elsif($branch =~ m#^/#) {
287 $svnpath = "/$tag_name$branch/$path";
289 $svnpath = "/$branch_name/$branch/$path";
293 my ($name, $res) = eval { $svn->file($svnpath,$rev); };
294 return () unless defined $name;
296 open my $F, '-|', "git-hash-object", "-w", $name
297 or die "Cannot create object: $!\n";
301 my $mode = "0644"; # SV does not seem to store any file modes
302 return [$mode, $sha, $path];
309 if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
311 } elsif($path =~ s#^/\Q$trunk_name\E/?##) {
313 } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
316 print STDERR "$rev: Unrecognized path: $path\n";
319 $path = "/" if $path eq "";
320 return ($branch,$path);
324 my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
325 my($author_name,$author_email,$dest);
328 if (not defined $author) {
329 $author_name = $author_email = "unknown";
330 } elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
331 ($author_name, $author_email) = ($1, $2);
333 $author =~ s/^<(.*)>$/$1/;
334 $author_name = $author_email = $author;
336 $date = pdate($date);
340 if($branch eq "/") { # trunk
342 } elsif($branch =~ m#^/(.+)#) { # tag
345 } else { # "normal" branch
351 my $prev = $changed_paths->{"/"};
352 if($prev and $prev->[0] eq "A") {
353 delete $changed_paths->{"/"};
354 my $oldpath = $prev->[1];
356 if(defined $oldpath) {
358 ($parent,$p) = split_path($revision,$oldpath);
362 $parent =~ s#^/##; # if it's a tag
370 if(defined $parent) {
371 open(H,"git-rev-parse --verify $parent |");
374 print STDERR "$revision: cannot find commit '$parent'!\n";
378 if(length($rev) != 40) {
379 print STDERR "$revision: cannot find commit '$parent'!\n";
382 $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
383 if($revision != 1 and not $rev) {
384 print STDERR "$revision: do not know ancestor for '$parent'!\n";
391 # if($prev and $prev->[0] eq "A") {
393 # unless(open(H,"> $git_dir/refs/heads/$branch")) {
394 # print STDERR "$revision: Could not create branch $branch: $!\n";
399 # or die "Could not write branch $branch: $!";
401 # or die "Could not write branch $branch: $!";
404 if(not defined $rev) {
406 } elsif ($rev ne $last_rev) {
407 print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
408 system("git-read-tree", $rev);
409 die "read-tree failed for $rev: $?\n" if $?;
413 while(my($path,$action) = each %$changed_paths) {
414 if ($action->[0] eq "A") {
415 my $f = get_file($revision,$branch,$path);
417 } elsif ($action->[0] eq "D") {
419 } elsif ($action->[0] eq "M") {
420 my $f = get_file($revision,$branch,$path);
422 } elsif ($action->[0] eq "R") {
423 # refer to a file/tree in an earlier commit
424 push(@old,$path); # remove any old stuff
426 # ... and add any new stuff
427 my($b,$p) = split_path($revision,$action->[1]);
428 open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $p;
432 my($m,$p) = split(/\t/,$_,2);
433 my($mode,$type,$sha1) = split(/ /,$m);
434 next if $type ne "blob";
435 push(@new,[$mode,$sha1,$p]);
438 die "$revision: unknown action '".$action->[0]."' for $path\n";
443 open my $F, "-|", "git-ls-files", "-z", @old or die $!;
455 @o2 = splice(@old,0,50);
460 system("git-update-index","--force-remove","--",@o2);
461 die "Cannot remove files: $?\n" if $?;
467 @n2 = splice(@new,0,10);
472 system("git-update-index","--add",
473 (map { ('--cacheinfo', @$_) } @n2));
474 die "Cannot add files: $?\n" if $?;
477 my $pid = open(C,"-|");
478 die "Cannot fork: $!" unless defined $pid;
480 exec("git-write-tree");
481 die "Cannot exec git-write-tree: $!\n";
483 chomp(my $tree = <C>);
485 or die "Cannot get tree id ($tree): $!\n";
487 or die "Error running git-write-tree: $?\n";
488 print "Tree ID $tree\n" if $opt_v;
490 my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
491 my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
493 die "Fork: $!\n" unless defined $pid;
497 open(OUT,">&STDOUT");
498 dup2($pw->fileno(),0);
499 dup2($pr->fileno(),1);
504 @par = ("-p",$rev) if defined $rev;
506 # loose detection of merges
507 # based on the commit msg
508 foreach my $rx (@mergerx) {
509 if ($message =~ $rx) {
511 if ($mparent eq 'HEAD') { $mparent = $opt_o };
512 if ( -e "$git_dir/refs/heads/$mparent") {
513 $mparent = get_headref($mparent, $git_dir);
514 push @par, '-p', $mparent;
515 print OUT "Merge parent branch: $mparent\n" if $opt_v;
521 "GIT_AUTHOR_NAME=$author_name",
522 "GIT_AUTHOR_EMAIL=$author_email",
523 "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
524 "GIT_COMMITTER_NAME=$author_name",
525 "GIT_COMMITTER_EMAIL=$author_email",
526 "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
527 "git-commit-tree", $tree,@par);
528 die "Cannot exec git-commit-tree: $!\n";
533 $message =~ s/[\s\n]+\z//;
535 print $pw "$message\n"
536 or die "Error writing to git-commit-tree: $!\n";
539 print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
540 chomp(my $cid = <$pr>);
542 or die "Cannot get commit id ($cid): $!\n";
543 print "Commit ID $cid\n" if $opt_v;
547 die "Error running git-commit-tree: $?\n" if $?;
550 print "Writing to refs/heads/$dest\n" if $opt_v;
551 open(C,">$git_dir/refs/heads/$dest") and
552 print C ("$cid\n") and
554 or die "Cannot write branch $dest for update: $!\n";
556 print "... no known parent\n" if $opt_v;
558 $branches{$branch}{"LAST"} = $cid;
559 $branches{$branch}{$revision} = $cid;
561 print BRANCHES "$revision $branch $cid\n";
562 print "DONE: $revision $dest $cid\n" if $opt_v;
565 my($in, $out) = ('','');
566 $last_rev = "-" if %$changed_paths;
567 # the tag was 'complex', i.e. did not refer to a "real" revision
569 $tag =~ tr/_/\./ if $opt_u;
571 my $pid = open2($in, $out, 'git-mktag');
572 print $out ("object $cid\n".
575 "tagger $author_name <$author_email>\n") and
577 or die "Cannot create tag object $tag: $!\n";
582 if ( !close($in) or waitpid($pid, 0) != $pid or
583 $? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
584 die "Cannot create tag object $tag: $!\n";
588 open(C,">$git_dir/refs/tags/$tag")
589 or die "Cannot create tag $tag: $!\n";
591 or die "Cannot write tag $tag: $!\n";
593 or die "Cannot write tag $tag: $!\n";
595 print "Created tag '$tag' on '$branch'\n" if $opt_v;
599 my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
601 ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
603 while(my($path,$action) = each %$changed_paths) {
604 $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev ];
606 $changed_paths = \%p;
615 while(my($path,$action) = each %$changed_paths) {
616 ($branch,$path) = split_path($revision,$path);
617 next if not defined $branch;
618 $done{$branch}{$path} = $action;
620 while(($branch,$changed_paths) = each %done) {
621 commit($branch, $changed_paths, $revision, $author, $date, $message);
625 while(++$current_rev < $svn->{'maxrev'}) {
626 $svn->{'svn'}->get_log("/",$current_rev,$current_rev,$current_rev,1,1,\&_commit_all,"");
633 if (defined $orig_git_index) {
634 $ENV{GIT_INDEX_FILE} = $orig_git_index;
636 delete $ENV{GIT_INDEX_FILE};
639 # Now switch back to the branch we were in before all of this happened
641 print "DONE\n" if $opt_v;
642 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
645 system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
646 die "read-tree failed: $?\n" if $?;
649 $orig_branch = "master";
650 print "DONE; creating $orig_branch branch\n" if $opt_v;
651 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
652 unless -f "$git_dir/refs/heads/master";
653 unlink("$git_dir/HEAD");
654 symlink("refs/heads/$orig_branch","$git_dir/HEAD");
656 system('git checkout');
657 die "checkout failed: $?\n" if $?;
660 unlink("$git_dir/SVN2GIT_HEAD");