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 aggregate CVS check-ins into related changes.
7 # Fortunately, "cvsps" does that for us; all we have to do is to parse
10 # Checking out the files is done by a single long-running CVS connection
13 # The head revision is on branch "origin" by default.
14 # You can change that with the '-o' option.
20 use File::Temp qw(tempfile);
21 use File::Path qw(mkpath);
22 use File::Basename qw(basename dirname);
26 use POSIX qw(strftime dup2);
28 $SIG{'PIPE'}="IGNORE";
31 our($opt_h,$opt_o,$opt_v,$opt_k,$opt_d,$opt_p,$opt_C,$opt_z,$opt_i,$opt_s,$opt_m,$opt_M);
35 Usage: ${\basename $0} # fetch/update GIT from CVS
36 [ -o branch-for-HEAD ] [ -h ] [ -v ] [ -d CVSROOT ]
37 [ -p opts-for-cvsps ] [ -C GIT_repository ] [ -z fuzz ]
38 [ -i ] [ -k ] [-s subst] [ -m ] [ -M regex] [ CVS_module ]
43 getopts("hivmko:d:p:C:z:s:M:") or usage();
46 @ARGV <= 1 or usage();
49 $ENV{"CVSROOT"} = $opt_d;
50 } elsif(-f 'CVS/Root') {
51 open my $f, '<', 'CVS/Root' or die 'Failed to open CVS/Root';
55 $ENV{"CVSROOT"} = $opt_d;
56 } elsif($ENV{"CVSROOT"}) {
57 $opt_d = $ENV{"CVSROOT"};
59 die "CVSROOT needs to be set";
63 my $git_tree = $opt_C;
69 } elsif (-f 'CVS/Repository') {
70 open my $f, '<', 'CVS/Repository' or
71 die 'Failed to open CVS/Repository';
81 @mergerx = ( qr/\W(?:from|of|merge|merging|merged) (\w+)/i );
84 push (@mergerx, qr/$opt_M/);
87 select(STDERR); $|=1; select(STDOUT);
92 # We're only interested in connecting and downloading, so ...
95 use File::Temp qw(tempfile);
96 use POSIX qw(strftime dup2);
99 my($what,$repo,$subdir) = @_;
100 $what=ref($what) if ref($what);
103 $self->{'buffer'} = "";
107 $self->{'fullrep'} = $repo;
110 $self->{'subdir'} = $subdir;
111 $self->{'lines'} = undef;
118 my $repo = $self->{'fullrep'};
119 if($repo =~ s/^:pserver:(?:(.*?)(?::(.*?))?@)?([^:\/]*)(?::(\d*))?//) {
120 my($user,$pass,$serv,$port) = ($1,$2,$3,$4);
121 $user="anonymous" unless defined $user;
124 $rr2 = ":pserver:$user\@$serv:$repo";
127 my $rr = ":pserver:$user\@$serv:$port$repo";
130 open(H,$ENV{'HOME'}."/.cvspass") and do {
131 # :pserver:cvs@mea.tmt.tele.fi:/cvsroot/zmailer Ah<Z
135 my ($w,$p) = split(/\s/,$_,2);
136 if($w eq $rr or $w eq $rr2) {
143 $pass="A" unless $pass;
145 my $s = IO::Socket::INET->new(PeerHost => $serv, PeerPort => $port);
146 die "Socket to $serv: $!\n" unless defined $s;
147 $s->write("BEGIN AUTH REQUEST\n$repo\n$user\n$pass\nEND AUTH REQUEST\n")
148 or die "Write to $serv: $!\n";
153 if($rep ne "I LOVE YOU\n") {
154 $rep="<unknown>" unless $rep;
155 die "AuthReply: $rep\n";
157 $self->{'socketo'} = $s;
158 $self->{'socketi'} = $s;
159 } else { # local or ext: Fork off our own cvs server.
160 my $pr = IO::Pipe->new();
161 my $pw = IO::Pipe->new();
163 die "Fork: $!\n" unless defined $pid;
165 $cvs = $ENV{CVS_SERVER} if exists $ENV{CVS_SERVER};
167 $rsh = $ENV{CVS_RSH} if exists $ENV{CVS_RSH};
169 my @cvs = ($cvs, 'server');
170 my ($local, $user, $host);
171 $local = $repo =~ s/:local://;
174 $local = !($repo =~ s/^(?:([^\@:]+)\@)?([^:]+)://);
175 ($user, $host) = ($1, $2);
179 unshift @cvs, $rsh, '-l', $user, $host;
181 unshift @cvs, $rsh, $host;
188 dup2($pw->fileno(),0);
189 dup2($pr->fileno(),1);
196 $self->{'socketo'} = $pw;
197 $self->{'socketi'} = $pr;
199 $self->{'socketo'}->write("Root $repo\n");
201 # Trial and error says that this probably is the minimum set
202 $self->{'socketo'}->write("Valid-responses ok error Valid-requests Mode M Mbinary E Checked-in Created Updated Merged Removed\n");
204 $self->{'socketo'}->write("valid-requests\n");
205 $self->{'socketo'}->flush();
207 chomp(my $rep=$self->readline());
208 if($rep !~ s/^Valid-requests\s*//) {
209 $rep="<unknown>" unless $rep;
210 die "Expected Valid-requests from server, but got: $rep\n";
212 chomp(my $res=$self->readline());
213 die "validReply: $res\n" if $res ne "ok";
215 $self->{'socketo'}->write("UseUnchanged\n") if $rep =~ /\bUseUnchanged\b/;
216 $self->{'repo'} = $repo;
221 return $self->{'socketi'}->getline();
225 # Request a file with a given revision.
226 # Trial and error says this is a good way to do it. :-/
227 my($self,$fn,$rev) = @_;
228 $self->{'socketo'}->write("Argument -N\n") or return undef;
229 $self->{'socketo'}->write("Argument -P\n") or return undef;
230 # -kk: Linus' version doesn't use it - defaults to off
232 $self->{'socketo'}->write("Argument -kk\n") or return undef;
234 $self->{'socketo'}->write("Argument -r\n") or return undef;
235 $self->{'socketo'}->write("Argument $rev\n") or return undef;
236 $self->{'socketo'}->write("Argument --\n") or return undef;
237 $self->{'socketo'}->write("Argument $self->{'subdir'}/$fn\n") or return undef;
238 $self->{'socketo'}->write("Directory .\n") or return undef;
239 $self->{'socketo'}->write("$self->{'repo'}\n") or return undef;
240 # $self->{'socketo'}->write("Sticky T1.0\n") or return undef;
241 $self->{'socketo'}->write("co\n") or return undef;
242 $self->{'socketo'}->flush() or return undef;
243 $self->{'lines'} = 0;
247 # Read a line from the server.
248 # ... except that 'line' may be an entire file. ;-)
250 die "Not in lines" unless defined $self->{'lines'};
254 while(defined($line = $self->readline())) {
255 # M U gnupg-cvs-rep/AUTHORS
256 # Updated gnupg-cvs-rep/
257 # /daten/src/rsync/gnupg-cvs-rep/AUTHORS
258 # /AUTHORS/1.1///T1.1
263 if($line =~ s/^(?:Created|Updated) //) {
264 $line = $self->readline(); # path
265 $line = $self->readline(); # Entries line
266 my $mode = $self->readline(); chomp $mode;
267 $self->{'mode'} = $mode;
268 defined (my $cnt = $self->readline())
269 or die "EOF from server after 'Changed'\n";
271 die "Duh: Filesize $cnt" if $cnt !~ /^\d+$/;
276 my $num = $self->{'socketi'}->read($buf,$cnt);
277 die "Server: Filesize $cnt: $num: $!\n" if not defined $num or $num<=0;
282 } elsif($line =~ s/^ //) {
284 $res += length($line);
285 } elsif($line =~ /^M\b/) {
287 } elsif($line =~ /^Mbinary\b/) {
289 die "EOF from server after 'Mbinary'" unless defined ($cnt = $self->readline());
291 die "Duh: Mbinary $cnt" if $cnt !~ /^\d+$/ or $cnt<1;
295 my $num = $self->{'socketi'}->read($buf,$cnt);
296 die "S: Mbinary $cnt: $num: $!\n" if not defined $num or $num<=0;
304 # print STDERR "S: ok (".length($res).")\n";
306 } elsif($line =~ s/^E //) {
307 # print STDERR "S: $line\n";
308 } elsif($line =~ /^Remove-entry /i) {
309 $line = $self->readline(); # filename
310 $line = $self->readline(); # OK
312 die "Unknown: $line" if $line ne "ok";
315 die "Unknown: $line\n";
321 my($self,$fn,$rev) = @_;
324 my ($fh, $name) = tempfile('gitcvs.XXXXXX',
325 DIR => File::Spec->tmpdir(), UNLINK => 1);
327 $self->_file($fn,$rev) and $res = $self->_line($fh);
332 $self->_file($fn,$rev)
333 or die "No file command send\n";
334 $res = $self->_line($fh);
335 die "No input: $fn $rev\n" unless defined $res;
339 return ($name, $res);
345 my $cvs = CVSconn->new($opt_d, $cvs_tree);
350 m#(\d{2,4})/(\d\d)/(\d\d)\s(\d\d):(\d\d)(?::(\d\d))?#
351 or die "Unparseable date: $d\n";
352 my $y=$1; $y-=1900 if $y>1900;
353 return timegm($6||0,$5,$4,$3,$2-1,$y);
361 for my $x(split(//,$mode)) {
366 } elsif($x eq "u") { $um |= 0700;
367 } elsif($x eq "g") { $um |= 0070;
368 } elsif($x eq "o") { $um |= 0007;
369 } elsif($x eq "r") { $mm |= 0444;
370 } elsif($x eq "w") { $mm |= 0222;
371 } elsif($x eq "x") { $mm |= 0111;
372 } elsif($x eq "=") { # do nothing
373 } else { die "Unknown mode: $mode\n";
387 sub get_headref($$) {
392 if (open(C,"$git_dir/refs/heads/$name")) {
396 or die "Cannot get head id for $name ($sha): $!\n";
403 or mkdir($git_tree,0777)
404 or die "Could not create $git_tree: $!";
407 my $last_branch = "";
408 my $orig_branch = "";
409 my $forward_master = 0;
412 my $git_dir = $ENV{"GIT_DIR"} || ".git";
413 $git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
414 $ENV{"GIT_DIR"} = $git_dir;
416 $orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
417 my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
418 DIR => File::Spec->tmpdir());
420 $ENV{GIT_INDEX_FILE} = $git_index;
421 unless(-d $git_dir) {
422 system("git-init-db");
423 die "Cannot init the GIT db at $git_tree: $?\n" if $?;
424 system("git-read-tree");
425 die "Cannot init an empty tree: $?\n" if $?;
427 $last_branch = $opt_o;
430 -f "$git_dir/refs/heads/$opt_o"
431 or die "Branch '$opt_o' does not exist.\n".
432 "Either use the correct '-o branch' option,\n".
433 "or import to a new repository.\n";
435 $last_branch = basename(readlink("$git_dir/HEAD"));
436 unless($last_branch) {
437 warn "Cannot read the last branch name: $! -- assuming 'master'\n";
438 $last_branch = "master";
440 $orig_branch = $last_branch;
441 if (-f "$git_dir/CVS2GIT_HEAD") {
444 Make sure your working directory corresponds to HEAD and remove CVS2GIT_HEAD.
447 git-read-tree -m -u CVS2GIT_HEAD HEAD
450 system('cp', "$git_dir/HEAD", "$git_dir/CVS2GIT_HEAD");
453 $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
454 system('cmp', '-s', "$git_dir/refs/heads/master",
455 "$git_dir/refs/heads/$opt_o") == 0;
458 system('git-read-tree', $last_branch);
459 die "read-tree failed: $?\n" if $?;
461 # Get the last import timestamps
462 opendir(D,"$git_dir/refs/heads");
463 while(defined(my $head = readdir(D))) {
464 next if $head =~ /^\./;
465 open(F,"$git_dir/refs/heads/$head")
466 or die "Bad head branch: $head: $!\n";
467 chomp(my $ftag = <F>);
469 open(F,"git-cat-file commit $ftag |");
471 next unless /^author\s.*\s(\d+)\s[-+]\d{4}$/;
472 $branch_date{$head} = $1;
481 or die "Could not create git subdir ($git_dir).\n";
483 my $pid = open(CVS,"-|");
484 die "Cannot fork: $!\n" unless defined $pid;
487 @opt = split(/,/,$opt_p) if defined $opt_p;
488 unshift @opt, '-z', $opt_z if defined $opt_z;
489 exec("cvsps",@opt,"-u","-A","--cvs-direct",'--root',$opt_d,$cvs_tree);
490 die "Could not start cvsps: $!\n";
495 #---------------------
497 #Date: 1999/09/18 13:03:59
499 #Branch: STABLE-BRANCH-1-0
500 #Ancestor branch: HEAD
503 # See ChangeLog: Sat Sep 18 13:03:28 CEST 1999 Werner Koch
505 # README:1.57->1.57.2.1
506 # VERSION:1.96->1.96.2.1
508 #---------------------
512 my($patchset,$date,$author,$branch,$ancestor,$tag,$logmsg);
519 @o2 = splice(@old,0,50);
524 system("git-update-cache","--force-remove","--",@o2);
525 die "Cannot remove files: $?\n" if $?;
530 @n2 = splice(@new,0,10);
535 system("git-update-cache","--add",
536 (map { ('--cacheinfo', @$_) } @n2));
537 die "Cannot add files: $?\n" if $?;
541 die "Cannot fork: $!" unless defined $pid;
543 exec("git-write-tree");
544 die "Cannot exec git-write-tree: $!\n";
546 chomp(my $tree = <C>);
548 or die "Cannot get tree id ($tree): $!\n";
550 or die "Error running git-write-tree: $?\n";
551 print "Tree ID $tree\n" if $opt_v;
554 if(open(C,"$git_dir/refs/heads/$last_branch")) {
555 chomp($parent = <C>);
557 length($parent) == 40
558 or die "Cannot get parent id ($parent): $!\n";
559 print "Parent ID $parent\n" if $opt_v;
562 my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
563 my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
565 die "Fork: $!\n" unless defined $pid;
569 dup2($pw->fileno(),0);
570 dup2($pr->fileno(),1);
575 @par = ("-p",$parent) if $parent;
577 # loose detection of merges
578 # based on the commit msg
579 foreach my $rx (@mergerx) {
580 if ($logmsg =~ $rx) {
582 if ($mparent eq 'HEAD') { $mparent = $opt_o };
583 if ( -e "$git_dir/refs/heads/$mparent") {
584 $mparent = get_headref($mparent, $git_dir);
585 push @par, '-p', $mparent;
586 # printing here breaks import #
587 # # print "Merge parent branch: $mparent\n" if $opt_v;
593 "GIT_AUTHOR_NAME=$author",
594 "GIT_AUTHOR_EMAIL=$author",
595 "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
596 "GIT_COMMITTER_NAME=$author",
597 "GIT_COMMITTER_EMAIL=$author",
598 "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
599 "git-commit-tree", $tree,@par);
600 die "Cannot exec git-commit-tree: $!\n";
605 # compatibility with git2cvs
606 substr($logmsg,32767) = "" if length($logmsg) > 32767;
607 $logmsg =~ s/[\s\n]+\z//;
609 print $pw "$logmsg\n"
610 or die "Error writing to git-commit-tree: $!\n";
613 print "Committed patch $patchset ($branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
614 chomp(my $cid = <$pr>);
616 or die "Cannot get commit id ($cid): $!\n";
617 print "Commit ID $cid\n" if $opt_v;
621 die "Error running git-commit-tree: $?\n" if $?;
623 open(C,">$git_dir/refs/heads/$branch")
624 or die "Cannot open branch $branch for update: $!\n";
626 or die "Cannot write branch $branch for update: $!\n";
628 or die "Cannot write branch $branch for update: $!\n";
631 open(C,">$git_dir/refs/tags/$tag")
632 or die "Cannot create tag $tag: $!\n";
634 or die "Cannot write tag $branch: $!\n";
636 or die "Cannot write tag $branch: $!\n";
637 print "Created tag '$tag' on '$branch'\n" if $opt_v;
643 if($state == 0 and /^-+$/) {
645 } elsif($state == 0) {
648 } elsif(($state==0 or $state==1) and s/^PatchSet\s+//) {
651 } elsif($state == 2 and s/^Date:\s+//) {
654 print STDERR "Could not parse date: $_\n";
659 } elsif($state == 3 and s/^Author:\s+//) {
663 } elsif($state == 4 and s/^Branch:\s+//) {
668 } elsif($state == 5 and s/^Ancestor branch:\s+//) {
671 $ancestor = $opt_o if $ancestor eq "HEAD";
673 } elsif($state == 5) {
677 } elsif($state == 6 and s/^Tag:\s+//) {
685 } elsif($state == 7 and /^Log:/) {
688 } elsif($state == 8 and /^Members:/) {
689 $branch = $opt_o if $branch eq "HEAD";
690 if(defined $branch_date{$branch} and $branch_date{$branch} >= $date) {
692 print "skip patchset $patchset: $date before $branch_date{$branch}\n" if $opt_v;
697 if(-f "$git_dir/refs/heads/$branch") {
698 print STDERR "Branch $branch already exists!\n";
702 unless(open(H,"$git_dir/refs/heads/$ancestor")) {
703 print STDERR "Branch $ancestor does not exist!\n";
709 unless(open(H,"> $git_dir/refs/heads/$branch")) {
710 print STDERR "Could not create branch $branch: $!\n";
715 or die "Could not write branch $branch: $!";
717 or die "Could not write branch $branch: $!";
719 if(($ancestor || $branch) ne $last_branch) {
720 print "Switching from $last_branch to $branch\n" if $opt_v;
721 system("git-read-tree", $branch);
722 die "read-tree failed: $?\n" if $?;
724 $last_branch = $branch if $branch ne $last_branch;
726 } elsif($state == 8) {
728 } elsif($state == 9 and /^\s+(.+?):(INITIAL|\d+(?:\.\d+)+)->(\d+(?:\.\d+)+)\s*$/) {
729 # VERSION:1.96->1.96.2.1
730 my $init = ($2 eq "INITIAL");
734 my ($tmpname, $size) = $cvs->file($fn,$rev);
737 print "Drop $fn\n" if $opt_v;
739 print "".($init ? "New" : "Update")." $fn: $size bytes\n" if $opt_v;
740 open my $F, '-|', "git-hash-object -w $tmpname"
741 or die "Cannot create object: $!\n";
745 my $mode = pmode($cvs->{'mode'});
746 push(@new,[$mode, $sha, $fn]); # may be resurrected!
749 } elsif($state == 9 and /^\s+(.+?):\d+(?:\.\d+)+->(\d+(?:\.\d+)+)\(DEAD\)\s*$/) {
753 print "Delete $fn\n" if $opt_v;
754 } elsif($state == 9 and /^\s*$/) {
756 } elsif(($state == 9 or $state == 10) and /^-+$/) {
759 } elsif($state == 11 and /^-+$/) {
761 } elsif(/^-+$/) { # end of unknown-line processing
763 } elsif($state != 11) { # ignore stuff when skipping
764 print "* UNKNOWN LINE * $_\n";
767 &$commit() if $branch and $state != 11;
771 if (defined $orig_git_index) {
772 $ENV{GIT_INDEX_FILE} = $orig_git_index;
774 delete $ENV{GIT_INDEX_FILE};
777 # Now switch back to the branch we were in before all of this happened
779 print "DONE\n" if $opt_v;
780 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
783 system('git-read-tree', '-m', '-u', 'CVS2GIT_HEAD', 'HEAD');
784 die "read-tree failed: $?\n" if $?;
787 $orig_branch = "master";
788 print "DONE; creating $orig_branch branch\n" if $opt_v;
789 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
790 unless -f "$git_dir/refs/heads/master";
791 unlink("$git_dir/HEAD");
792 symlink("refs/heads/$orig_branch","$git_dir/HEAD");
794 system('git checkout');
795 die "checkout failed: $?\n" if $?;
798 unlink("$git_dir/CVS2GIT_HEAD");