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,$opt_s);
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 start_chg] [-m] [-M regex] [SVN_URL]
46 getopts("b:C:hivmM:o:s: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 $self->{'repo'} = $repo;
114 $self->{'maxrev'} = $s->get_latest_revnum();
118 my($self,$path,$rev) = @_;
121 my ($fh, $name) = tempfile('gitsvn.XXXXXX',
122 DIR => File::Spec->tmpdir(), UNLINK => 1);
124 print "... $rev $path ...\n" if $opt_v;
125 eval { $self->{'svn'}->get_file($path,$rev,$fh); };
126 if ($@ and $@ !~ /Attempted to get checksum/) {
129 eval { $self->{'svn'}->get_file($path,$rev,$fh); };
131 return () if $@ and $@ !~ /Attempted to get checksum/;
135 return ($name, $res);
141 my $svn = SVNconn->new($cvs_tree);
146 $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
147 or die "Unparseable date: $d\n";
148 my $y=$1; $y-=1900 if $y>1900;
149 return timegm($6||0,$5,$4,$3,$2-1,$y);
159 sub get_headref($$) {
164 if (open(C,"$git_dir/refs/heads/$name")) {
168 or die "Cannot get head id for $name ($sha): $!\n";
175 or mkdir($git_tree,0777)
176 or die "Could not create $git_tree: $!";
179 my $orig_branch = "";
180 my $forward_master = 0;
183 my $git_dir = $ENV{"GIT_DIR"} || ".git";
184 $git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
185 $ENV{"GIT_DIR"} = $git_dir;
187 $orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
188 my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
189 DIR => File::Spec->tmpdir());
191 $ENV{GIT_INDEX_FILE} = $git_index;
195 my $current_rev = $opt_s ? ($opt_s-1) : 0;
196 unless(-d $git_dir) {
197 system("git-init-db");
198 die "Cannot init the GIT db at $git_tree: $?\n" if $?;
199 system("git-read-tree");
200 die "Cannot init an empty tree: $?\n" if $?;
202 $last_branch = $opt_o;
205 -f "$git_dir/refs/heads/$opt_o"
206 or die "Branch '$opt_o' does not exist.\n".
207 "Either use the correct '-o branch' option,\n".
208 "or import to a new repository.\n";
210 -f "$git_dir/svn2git"
211 or die "'$git_dir/svn2git' does not exist.\n".
212 "You need that file for incremental imports.\n";
213 $last_branch = basename(readlink("$git_dir/HEAD"));
214 unless($last_branch) {
215 warn "Cannot read the last branch name: $! -- assuming 'master'\n";
216 $last_branch = "master";
218 $orig_branch = $last_branch;
219 $last_rev = get_headref($orig_branch, $git_dir);
220 if (-f "$git_dir/SVN2GIT_HEAD") {
223 Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
226 git-read-tree -m -u SVN2GIT_HEAD HEAD
229 system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
232 $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
233 system('cmp', '-s', "$git_dir/refs/heads/master",
234 "$git_dir/refs/heads/$opt_o") == 0;
237 system('git-read-tree', $last_rev);
238 die "read-tree failed: $?\n" if $?;
240 # Get the last import timestamps
241 open my $B,"<", "$git_dir/svn2git";
244 my($num,$branch,$ref) = split;
245 $branches{$branch}{$num} = $ref;
246 $branches{$branch}{"LAST"} = $ref;
247 $current_rev = $num+1 if $current_rev < $num+1;
252 or die "Could not create git subdir ($git_dir).\n";
254 open BRANCHES,">>", "$git_dir/svn2git";
258 #---------------------
260 #Date: 1999/09/18 13:03:59
262 #Branch: STABLE-BRANCH-1-0
263 #Ancestor branch: HEAD
266 # See ChangeLog: Sat Sep 18 13:03:28 CEST 1999 Werner Koch
268 # README:1.57->1.57.2.1
269 # VERSION:1.96->1.96.2.1
271 #---------------------
276 my($rev,$branch,$path) = @_;
278 # revert split_path(), below
280 $path = "" if $path eq "/"; # this should not happen, but ...
282 $svnpath = "/$trunk_name/$path";
283 } elsif($branch =~ m#^/#) {
284 $svnpath = "/$tag_name$branch/$path";
286 $svnpath = "/$branch_name/$branch/$path";
290 my ($name, $res) = eval { $svn->file($svnpath,$rev); };
291 return () unless defined $name;
293 open my $F, '-|', "git-hash-object", "-w", $name
294 or die "Cannot create object: $!\n";
298 my $mode = "0644"; # SV does not seem to store any file modes
299 return [$mode, $sha, $path];
306 if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
308 } elsif($path =~ s#^/\Q$trunk_name\E/?##) {
310 } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
313 print STDERR "$rev: Unrecognized path: $path\n";
316 $path = "/" if $path eq "";
317 return ($branch,$path);
321 my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
322 my($author_name,$author_email,$dest);
325 if (not defined $author) {
326 $author_name = $author_email = "unknown";
327 } elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
328 ($author_name, $author_email) = ($1, $2);
330 $author =~ s/^<(.*)>$/$1/;
331 $author_name = $author_email = $author;
333 $date = pdate($date);
337 if($branch eq "/") { # trunk
339 } elsif($branch =~ m#^/(.+)#) { # tag
342 } else { # "normal" branch
348 my $prev = $changed_paths->{"/"};
349 if($prev and $prev->[0] eq "A") {
350 delete $changed_paths->{"/"};
351 my $oldpath = $prev->[1];
353 if(defined $oldpath) {
355 ($parent,$p) = split_path($revision,$oldpath);
359 $parent =~ s#^/##; # if it's a tag
367 if($revision > $opt_s and defined $parent) {
368 open(H,"git-rev-parse --verify $parent |");
371 print STDERR "$revision: cannot find commit '$parent'!\n";
375 if(length($rev) != 40) {
376 print STDERR "$revision: cannot find commit '$parent'!\n";
379 $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
380 if($revision != $opt_s and not $rev) {
381 print STDERR "$revision: do not know ancestor for '$parent'!\n";
388 # if($prev and $prev->[0] eq "A") {
390 # unless(open(H,"> $git_dir/refs/heads/$branch")) {
391 # print STDERR "$revision: Could not create branch $branch: $!\n";
396 # or die "Could not write branch $branch: $!";
398 # or die "Could not write branch $branch: $!";
401 if(not defined $rev) {
403 } elsif ($rev ne $last_rev) {
404 print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
405 system("git-read-tree", $rev);
406 die "read-tree failed for $rev: $?\n" if $?;
411 if($tag and not %$changed_paths) {
414 while(my($path,$action) = each %$changed_paths) {
415 if ($action->[0] eq "A") {
416 my $f = get_file($revision,$branch,$path);
418 } elsif ($action->[0] eq "D") {
420 } elsif ($action->[0] eq "M") {
421 my $f = get_file($revision,$branch,$path);
423 } elsif ($action->[0] eq "R") {
424 # refer to a file/tree in an earlier commit
425 push(@old,$path); # remove any old stuff
427 # ... and add any new stuff
428 my($b,$p) = split_path($revision,$action->[1]);
429 open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $p;
433 my($m,$p) = split(/\t/,$_,2);
434 my($mode,$type,$sha1) = split(/ /,$m);
435 next if $type ne "blob";
436 push(@new,[$mode,$sha1,$p]);
439 die "$revision: unknown action '".$action->[0]."' for $path\n";
444 open my $F, "-|", "git-ls-files", "-z", @old or die $!;
456 @o2 = splice(@old,0,50);
461 system("git-update-index","--force-remove","--",@o2);
462 die "Cannot remove files: $?\n" if $?;
468 @n2 = splice(@new,0,10);
473 system("git-update-index","--add",
474 (map { ('--cacheinfo', @$_) } @n2));
475 die "Cannot add files: $?\n" if $?;
478 my $pid = open(C,"-|");
479 die "Cannot fork: $!" unless defined $pid;
481 exec("git-write-tree");
482 die "Cannot exec git-write-tree: $!\n";
484 chomp(my $tree = <C>);
486 or die "Cannot get tree id ($tree): $!\n";
488 or die "Error running git-write-tree: $?\n";
489 print "Tree ID $tree\n" if $opt_v;
491 my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
492 my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
494 die "Fork: $!\n" unless defined $pid;
498 open(OUT,">&STDOUT");
499 dup2($pw->fileno(),0);
500 dup2($pr->fileno(),1);
505 @par = ("-p",$rev) if defined $rev;
507 # loose detection of merges
508 # based on the commit msg
509 foreach my $rx (@mergerx) {
510 if ($message =~ $rx) {
512 if ($mparent eq 'HEAD') { $mparent = $opt_o };
513 if ( -e "$git_dir/refs/heads/$mparent") {
514 $mparent = get_headref($mparent, $git_dir);
515 push @par, '-p', $mparent;
516 print OUT "Merge parent branch: $mparent\n" if $opt_v;
522 "GIT_AUTHOR_NAME=$author_name",
523 "GIT_AUTHOR_EMAIL=$author_email",
524 "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
525 "GIT_COMMITTER_NAME=$author_name",
526 "GIT_COMMITTER_EMAIL=$author_email",
527 "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
528 "git-commit-tree", $tree,@par);
529 die "Cannot exec git-commit-tree: $!\n";
534 $message =~ s/[\s\n]+\z//;
536 print $pw "$message\n"
537 or die "Error writing to git-commit-tree: $!\n";
540 print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
543 or die "Cannot get commit id ($cid): $!\n";
544 print "Commit ID $cid\n" if $opt_v;
548 die "Error running git-commit-tree: $?\n" if $?;
551 if(not defined $dest) {
552 print "... no known parent\n" if $opt_v;
554 print "Writing to refs/heads/$dest\n" if $opt_v;
555 open(C,">$git_dir/refs/heads/$dest") and
556 print C ("$cid\n") and
558 or die "Cannot write branch $dest for update: $!\n";
560 $branches{$branch}{"LAST"} = $cid;
561 $branches{$branch}{$revision} = $cid;
563 print BRANCHES "$revision $branch $cid\n";
564 print "DONE: $revision $dest $cid\n" if $opt_v;
567 my($in, $out) = ('','');
568 $last_rev = "-" if %$changed_paths;
569 # the tag was 'complex', i.e. did not refer to a "real" revision
571 $dest =~ tr/_/\./ if $opt_u;
573 my $pid = open2($in, $out, 'git-mktag');
574 print $out ("object $cid\n".
577 "tagger $author_name <$author_email>\n") and
579 or die "Cannot create tag object $dest: $!\n";
584 if ( !close($in) or waitpid($pid, 0) != $pid or
585 $? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
586 die "Cannot create tag object $dest: $!\n";
589 open(C,">$git_dir/refs/tags/$dest") and
590 print C ("$tagobj\n") and
592 or die "Cannot create tag $branch: $!\n";
594 print "Created tag '$dest' on '$branch'\n" if $opt_v;
598 my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
600 ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
602 while(my($path,$action) = each %$changed_paths) {
603 $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev ];
605 $changed_paths = \%p;
614 while(my($path,$action) = each %$changed_paths) {
615 ($branch,$path) = split_path($revision,$path);
616 next if not defined $branch;
617 $done{$branch}{$path} = $action;
619 while(($branch,$changed_paths) = each %done) {
620 commit($branch, $changed_paths, $revision, $author, $date, $message);
624 while(++$current_rev < $svn->{'maxrev'}) {
625 $svn->{'svn'}->get_log("/",$current_rev,$current_rev,$current_rev,1,1,\&_commit_all,"");
632 if (defined $orig_git_index) {
633 $ENV{GIT_INDEX_FILE} = $orig_git_index;
635 delete $ENV{GIT_INDEX_FILE};
638 # Now switch back to the branch we were in before all of this happened
640 print "DONE\n" if $opt_v;
641 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
644 system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
645 die "read-tree failed: $?\n" if $?;
648 $orig_branch = "master";
649 print "DONE; creating $orig_branch branch\n" if $opt_v;
650 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
651 unless -f "$git_dir/refs/heads/master";
652 unlink("$git_dir/HEAD");
653 symlink("refs/heads/$orig_branch","$git_dir/HEAD");
655 system('git checkout');
656 die "checkout failed: $?\n" if $?;
659 unlink("$git_dir/SVN2GIT_HEAD");