Add git-mv
[git/jrn.git] / git-svnimport.perl
blob20a85724cb1660a3411fa34717cd5f1e98ebd731
1 #!/usr/bin/perl -w
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 v5.8.0; # for shell-safe open("-|",LIST)
14 use strict;
15 use warnings;
16 use Getopt::Std;
17 use File::Spec;
18 use File::Temp qw(tempfile);
19 use File::Path qw(mkpath);
20 use File::Basename qw(basename dirname);
21 use Time::Local;
22 use IO::Pipe;
23 use POSIX qw(strftime dup2);
24 use IPC::Open2;
25 use SVN::Core;
26 use SVN::Ra;
28 die "Need CVN:Core 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1";
30 $SIG{'PIPE'}="IGNORE";
31 $ENV{'TZ'}="UTC";
33 our($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);
35 sub usage() {
36 print STDERR <<END;
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]
41 END
42 exit(1);
45 getopts("b:C:dDhil:mM:o:s:t:T:uv") or usage();
46 usage if $opt_h;
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();
54 $opt_o ||= "origin";
55 $opt_s ||= 1;
56 $opt_l = 100 unless defined $opt_l;
57 my $git_tree = $opt_C;
58 $git_tree ||= ".";
60 my $svn_url = $ARGV[0];
61 my $svn_dir = $ARGV[1];
63 our @mergerx = ();
64 if ($opt_m) {
65 @mergerx = ( qr/\W(?:from|of|merge|merging|merged) (\w+)/i );
67 if ($opt_M) {
68 push (@mergerx, qr/$opt_M/);
71 select(STDERR); $|=1; select(STDOUT);
74 package SVNconn;
75 # Basic SVN connection.
76 # We're only interested in connecting and downloading, so ...
78 use File::Spec;
79 use File::Temp qw(tempfile);
80 use POSIX qw(strftime dup2);
82 sub new {
83 my($what,$repo) = @_;
84 $what=ref($what) if ref($what);
86 my $self = {};
87 $self->{'buffer'} = "";
88 bless($self,$what);
90 $repo =~ s#/+$##;
91 $self->{'fullrep'} = $repo;
92 $self->conn();
94 return $self;
97 sub conn {
98 my $self = shift;
99 my $repo = $self->{'fullrep'};
100 my $s = SVN::Ra->new($repo);
102 die "SVN connection to $repo: $!\n" unless defined $s;
103 $self->{'svn'} = $s;
104 $self->{'repo'} = $repo;
105 $self->{'maxrev'} = $s->get_latest_revnum();
108 sub file {
109 my($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 eval { $self->{'svn'}->get_file($path,$rev,$fh); };
116 if($@) {
117 return undef if $@ =~ /Attempted to get checksum/;
118 die $@;
120 close ($fh);
122 return $name;
125 package main;
126 use URI;
128 my $svn = $svn_url;
129 $svn .= "/$svn_dir" if defined $svn_dir;
130 $svn = SVNconn->new($svn);
132 my $lwp_ua;
133 if($opt_d or $opt_D) {
134 $svn_url = URI->new($svn_url)->canonical;
135 if($opt_D) {
136 $svn_dir =~ s#/*$#/#;
137 } else {
138 $svn_dir = "";
140 if ($svn_url->scheme eq "http") {
141 use LWP::UserAgent;
142 $lwp_ua = LWP::UserAgent->new(keep_alive => 1, requests_redirectable => []);
143 } else {
144 print STDERR "Warning: not HTTP; turning off direct file access\n";
145 $opt_d=0;
149 sub pdate($) {
150 my($d) = @_;
151 $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)#
152 or die "Unparseable date: $d\n";
153 my $y=$1; $y-=1900 if $y>1900;
154 return timegm($6||0,$5,$4,$3,$2-1,$y);
157 sub getwd() {
158 my $pwd = `pwd`;
159 chomp $pwd;
160 return $pwd;
164 sub get_headref($$) {
165 my $name = shift;
166 my $git_dir = shift;
167 my $sha;
169 if (open(C,"$git_dir/refs/heads/$name")) {
170 chomp($sha = <C>);
171 close(C);
172 length($sha) == 40
173 or die "Cannot get head id for $name ($sha): $!\n";
175 return $sha;
179 -d $git_tree
180 or mkdir($git_tree,0777)
181 or die "Could not create $git_tree: $!";
182 chdir($git_tree);
184 my $orig_branch = "";
185 my $forward_master = 0;
186 my %branches;
188 my $git_dir = $ENV{"GIT_DIR"} || ".git";
189 $git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#;
190 $ENV{"GIT_DIR"} = $git_dir;
191 my $orig_git_index;
192 $orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE};
193 my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx',
194 DIR => File::Spec->tmpdir());
195 close ($git_ih);
196 $ENV{GIT_INDEX_FILE} = $git_index;
197 my $maxnum = 0;
198 my $last_rev = "";
199 my $last_branch;
200 my $current_rev = $opt_s-1;
201 unless(-d $git_dir) {
202 system("git-init-db");
203 die "Cannot init the GIT db at $git_tree: $?\n" if $?;
204 system("git-read-tree");
205 die "Cannot init an empty tree: $?\n" if $?;
207 $last_branch = $opt_o;
208 $orig_branch = "";
209 } else {
210 -f "$git_dir/refs/heads/$opt_o"
211 or die "Branch '$opt_o' does not exist.\n".
212 "Either use the correct '-o branch' option,\n".
213 "or import to a new repository.\n";
215 -f "$git_dir/svn2git"
216 or die "'$git_dir/svn2git' does not exist.\n".
217 "You need that file for incremental imports.\n";
218 $last_branch = basename(readlink("$git_dir/HEAD"));
219 unless($last_branch) {
220 warn "Cannot read the last branch name: $! -- assuming 'master'\n";
221 $last_branch = "master";
223 $orig_branch = $last_branch;
224 $last_rev = get_headref($orig_branch, $git_dir);
225 if (-f "$git_dir/SVN2GIT_HEAD") {
226 die <<EOM;
227 SVN2GIT_HEAD exists.
228 Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD.
229 You may need to run
231 git-read-tree -m -u SVN2GIT_HEAD HEAD
234 system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD");
236 $forward_master =
237 $opt_o ne 'master' && -f "$git_dir/refs/heads/master" &&
238 system('cmp', '-s', "$git_dir/refs/heads/master",
239 "$git_dir/refs/heads/$opt_o") == 0;
241 # populate index
242 system('git-read-tree', $last_rev);
243 die "read-tree failed: $?\n" if $?;
245 # Get the last import timestamps
246 open my $B,"<", "$git_dir/svn2git";
247 while(<$B>) {
248 chomp;
249 my($num,$branch,$ref) = split;
250 $branches{$branch}{$num} = $ref;
251 $branches{$branch}{"LAST"} = $ref;
252 $current_rev = $num if $current_rev < $num;
254 close($B);
256 -d $git_dir
257 or die "Could not create git subdir ($git_dir).\n";
259 open BRANCHES,">>", "$git_dir/svn2git";
261 sub get_file($$$) {
262 my($rev,$branch,$path) = @_;
264 # revert split_path(), below
265 my $svnpath;
266 $path = "" if $path eq "/"; # this should not happen, but ...
267 if($branch eq "/") {
268 $svnpath = "$trunk_name/$path";
269 } elsif($branch =~ m#^/#) {
270 $svnpath = "$tag_name$branch/$path";
271 } else {
272 $svnpath = "$branch_name/$branch/$path";
275 # now get it
276 my $name;
277 if($opt_d) {
278 my($req,$res);
280 # /svn/!svn/bc/2/django/trunk/django-docs/build.py
281 my $url=$svn_url->clone();
282 $url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath");
283 print "... $path...\n" if $opt_v;
284 $req = HTTP::Request->new(GET => $url);
285 $res = $lwp_ua->request($req);
286 if ($res->is_success) {
287 my $fh;
288 ($fh, $name) = tempfile('gitsvn.XXXXXX',
289 DIR => File::Spec->tmpdir(), UNLINK => 1);
290 print $fh $res->content;
291 close($fh) or die "Could not write $name: $!\n";
292 } else {
293 return undef if $res->code == 301; # directory?
294 die $res->status_line." at $url\n";
296 } else {
297 $name = $svn->file("/$svnpath",$rev);
298 return undef unless defined $name;
301 open my $F, '-|', "git-hash-object", "-w", $name
302 or die "Cannot create object: $!\n";
303 my $sha = <$F>;
304 chomp $sha;
305 close $F;
306 unlink $name;
307 my $mode = "0644"; # SV does not seem to store any file modes
308 return [$mode, $sha, $path];
311 sub split_path($$) {
312 my($rev,$path) = @_;
313 my $branch;
315 if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) {
316 $branch = "/$1";
317 } elsif($path =~ s#^/\Q$trunk_name\E/?##) {
318 $branch = "/";
319 } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) {
320 $branch = $1;
321 } else {
322 print STDERR "$rev: Unrecognized path: $path\n";
323 return ()
325 $path = "/" if $path eq "";
326 return ($branch,$path);
329 sub copy_subdir($$$$$$) {
330 # Somebody copied a whole subdirectory.
331 # We need to find the index entries from the old version which the
332 # SVN log entry points to, and add them to the new place.
334 my($newrev,$newbranch,$path,$oldpath,$rev,$new) = @_;
335 my($branch,$srcpath) = split_path($rev,$oldpath);
337 my $gitrev = $branches{$branch}{$rev};
338 unless($gitrev) {
339 print STDERR "$newrev:$newbranch: could not find $oldpath \@ $rev\n";
340 return;
342 print "$newrev:$newbranch:$path: copying from $branch:$srcpath @ $rev\n" if $opt_v;
343 $srcpath =~ s#/*$#/#;
344 open my $f,"-|","git-ls-tree","-r","-z",$gitrev,$srcpath;
345 local $/ = "\0";
346 while(<$f>) {
347 chomp;
348 my($m,$p) = split(/\t/,$_,2);
349 my($mode,$type,$sha1) = split(/ /,$m);
350 next if $type ne "blob";
351 $p = substr($p,length($srcpath)-1);
352 print "... found $path$p ...\n" if $opt_v;
353 push(@$new,[$mode,$sha1,$path.$p]);
355 close($f) or
356 print STDERR "$newrev:$newbranch: could not list files in $oldpath \@ $rev\n";
359 sub commit {
360 my($branch, $changed_paths, $revision, $author, $date, $message) = @_;
361 my($author_name,$author_email,$dest);
362 my(@old,@new);
364 if (not defined $author) {
365 $author_name = $author_email = "unknown";
366 } elsif ($author =~ /^(.*?)\s+<(.*)>$/) {
367 ($author_name, $author_email) = ($1, $2);
368 } else {
369 $author =~ s/^<(.*)>$/$1/;
370 $author_name = $author_email = $author;
372 $date = pdate($date);
374 my $tag;
375 my $parent;
376 if($branch eq "/") { # trunk
377 $parent = $opt_o;
378 } elsif($branch =~ m#^/(.+)#) { # tag
379 $tag = 1;
380 $parent = $1;
381 } else { # "normal" branch
382 # nothing to do
383 $parent = $branch;
385 $dest = $parent;
387 my $prev = $changed_paths->{"/"};
388 if($prev and $prev->[0] eq "A") {
389 delete $changed_paths->{"/"};
390 my $oldpath = $prev->[1];
391 my $rev;
392 if(defined $oldpath) {
393 my $p;
394 ($parent,$p) = split_path($revision,$oldpath);
395 if($parent eq "/") {
396 $parent = $opt_o;
397 } else {
398 $parent =~ s#^/##; # if it's a tag
400 } else {
401 $parent = undef;
405 my $rev;
406 if($revision > $opt_s and defined $parent) {
407 open(H,"git-rev-parse --verify $parent |");
408 $rev = <H>;
409 close(H) or do {
410 print STDERR "$revision: cannot find commit '$parent'!\n";
411 return;
413 chop $rev;
414 if(length($rev) != 40) {
415 print STDERR "$revision: cannot find commit '$parent'!\n";
416 return;
418 $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"};
419 if($revision != $opt_s and not $rev) {
420 print STDERR "$revision: do not know ancestor for '$parent'!\n";
421 return;
423 } else {
424 $rev = undef;
427 # if($prev and $prev->[0] eq "A") {
428 # if(not $tag) {
429 # unless(open(H,"> $git_dir/refs/heads/$branch")) {
430 # print STDERR "$revision: Could not create branch $branch: $!\n";
431 # $state=11;
432 # next;
434 # print H "$rev\n"
435 # or die "Could not write branch $branch: $!";
436 # close(H)
437 # or die "Could not write branch $branch: $!";
440 if(not defined $rev) {
441 unlink($git_index);
442 } elsif ($rev ne $last_rev) {
443 print "Switching from $last_rev to $rev ($branch)\n" if $opt_v;
444 system("git-read-tree", $rev);
445 die "read-tree failed for $rev: $?\n" if $?;
446 $last_rev = $rev;
449 my $cid;
450 if($tag and not %$changed_paths) {
451 $cid = $rev;
452 } else {
453 my @paths = sort keys %$changed_paths;
454 foreach my $path(@paths) {
455 my $action = $changed_paths->{$path};
457 if ($action->[0] eq "A") {
458 my $f = get_file($revision,$branch,$path);
459 if($f) {
460 push(@new,$f) if $f;
461 } elsif($action->[1]) {
462 copy_subdir($revision,$branch,$path,$action->[1],$action->[2],\@new);
463 } else {
464 my $opath = $action->[3];
465 print STDERR "$revision: $branch: could not fetch '$opath'\n";
467 } elsif ($action->[0] eq "D") {
468 push(@old,$path);
469 } elsif ($action->[0] eq "M") {
470 my $f = get_file($revision,$branch,$path);
471 push(@new,$f) if $f;
472 } elsif ($action->[0] eq "R") {
473 # refer to a file/tree in an earlier commit
474 push(@old,$path); # remove any old stuff
476 # ... and add any new stuff
477 my($b,$srcpath) = split_path($revision,$action->[1]);
478 $srcpath =~ s#/*$#/#;
479 open my $F,"-|","git-ls-tree","-r","-z", $branches{$b}{$action->[2]}, $srcpath;
480 local $/ = "\0";
481 while(<$F>) {
482 chomp;
483 my($m,$p) = split(/\t/,$_,2);
484 my($mode,$type,$sha1) = split(/ /,$m);
485 next if $type ne "blob";
486 $p = substr($p,length($srcpath)-1);
487 push(@new,[$mode,$sha1,$path.$p]);
489 close($F);
490 } else {
491 die "$revision: unknown action '".$action->[0]."' for $path\n";
495 if(@old) {
496 open my $F, "-|", "git-ls-files", "-z", @old or die $!;
497 @old = ();
498 local $/ = "\0";
499 while(<$F>) {
500 chomp;
501 push(@old,$_);
503 close($F);
505 while(@old) {
506 my @o2;
507 if(@old > 55) {
508 @o2 = splice(@old,0,50);
509 } else {
510 @o2 = @old;
511 @old = ();
513 system("git-update-index","--force-remove","--",@o2);
514 die "Cannot remove files: $?\n" if $?;
517 while(@new) {
518 my @n2;
519 if(@new > 12) {
520 @n2 = splice(@new,0,10);
521 } else {
522 @n2 = @new;
523 @new = ();
525 system("git-update-index","--add",
526 (map { ('--cacheinfo', @$_) } @n2));
527 die "Cannot add files: $?\n" if $?;
530 my $pid = open(C,"-|");
531 die "Cannot fork: $!" unless defined $pid;
532 unless($pid) {
533 exec("git-write-tree");
534 die "Cannot exec git-write-tree: $!\n";
536 chomp(my $tree = <C>);
537 length($tree) == 40
538 or die "Cannot get tree id ($tree): $!\n";
539 close(C)
540 or die "Error running git-write-tree: $?\n";
541 print "Tree ID $tree\n" if $opt_v;
543 my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n";
544 my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n";
545 $pid = fork();
546 die "Fork: $!\n" unless defined $pid;
547 unless($pid) {
548 $pr->writer();
549 $pw->reader();
550 open(OUT,">&STDOUT");
551 dup2($pw->fileno(),0);
552 dup2($pr->fileno(),1);
553 $pr->close();
554 $pw->close();
556 my @par = ();
557 @par = ("-p",$rev) if defined $rev;
559 # loose detection of merges
560 # based on the commit msg
561 foreach my $rx (@mergerx) {
562 if ($message =~ $rx) {
563 my $mparent = $1;
564 if ($mparent eq 'HEAD') { $mparent = $opt_o };
565 if ( -e "$git_dir/refs/heads/$mparent") {
566 $mparent = get_headref($mparent, $git_dir);
567 push @par, '-p', $mparent;
568 print OUT "Merge parent branch: $mparent\n" if $opt_v;
573 exec("env",
574 "GIT_AUTHOR_NAME=$author_name",
575 "GIT_AUTHOR_EMAIL=$author_email",
576 "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
577 "GIT_COMMITTER_NAME=$author_name",
578 "GIT_COMMITTER_EMAIL=$author_email",
579 "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)),
580 "git-commit-tree", $tree,@par);
581 die "Cannot exec git-commit-tree: $!\n";
583 $pw->writer();
584 $pr->reader();
586 $message =~ s/[\s\n]+\z//;
588 print $pw "$message\n"
589 or die "Error writing to git-commit-tree: $!\n";
590 $pw->close();
592 print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v;
593 chomp($cid = <$pr>);
594 length($cid) == 40
595 or die "Cannot get commit id ($cid): $!\n";
596 print "Commit ID $cid\n" if $opt_v;
597 $pr->close();
599 waitpid($pid,0);
600 die "Error running git-commit-tree: $?\n" if $?;
603 if(not defined $dest) {
604 print "... no known parent\n" if $opt_v;
605 } elsif(not $tag) {
606 print "Writing to refs/heads/$dest\n" if $opt_v;
607 open(C,">$git_dir/refs/heads/$dest") and
608 print C ("$cid\n") and
609 close(C)
610 or die "Cannot write branch $dest for update: $!\n";
613 if($tag) {
614 my($in, $out) = ('','');
615 $last_rev = "-" if %$changed_paths;
616 # the tag was 'complex', i.e. did not refer to a "real" revision
618 $dest =~ tr/_/\./ if $opt_u;
620 my $pid = open2($in, $out, 'git-mktag');
621 print $out ("object $cid\n".
622 "type commit\n".
623 "tag $dest\n".
624 "tagger $author_name <$author_email>\n") and
625 close($out)
626 or die "Cannot create tag object $dest: $!\n";
628 my $tagobj = <$in>;
629 chomp $tagobj;
631 if ( !close($in) or waitpid($pid, 0) != $pid or
632 $? != 0 or $tagobj !~ /^[0123456789abcdef]{40}$/ ) {
633 die "Cannot create tag object $dest: $!\n";
636 open(C,">$git_dir/refs/tags/$dest") and
637 print C ("$tagobj\n") and
638 close(C)
639 or die "Cannot create tag $branch: $!\n";
641 print "Created tag '$dest' on '$branch'\n" if $opt_v;
643 $branches{$branch}{"LAST"} = $cid;
644 $branches{$branch}{$revision} = $cid;
645 $last_rev = $cid;
646 print BRANCHES "$revision $branch $cid\n";
647 print "DONE: $revision $dest $cid\n" if $opt_v;
650 my ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
651 sub _commit_all {
652 ($changed_paths, $revision, $author, $date, $message, $pool) = @_;
653 my %p;
654 while(my($path,$action) = each %$changed_paths) {
655 $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev, $path ];
657 $changed_paths = \%p;
660 sub commit_all {
661 my %done;
662 my @col;
663 my $pref;
664 my $branch;
666 while(my($path,$action) = each %$changed_paths) {
667 ($branch,$path) = split_path($revision,$path);
668 next if not defined $branch;
669 $done{$branch}{$path} = $action;
671 while(($branch,$changed_paths) = each %done) {
672 commit($branch, $changed_paths, $revision, $author, $date, $message);
676 while(++$current_rev <= $svn->{'maxrev'}) {
677 $svn->{'svn'}->get_log("/",$current_rev,$current_rev,$current_rev,1,1,\&_commit_all,"");
678 commit_all();
679 if($opt_l and not --$opt_l) {
680 print STDERR "Stopping, because there is a memory leak (in the SVN library).\n";
681 print STDERR "Please repeat this command; it will continue safely\n";
682 last;
687 unlink($git_index);
689 if (defined $orig_git_index) {
690 $ENV{GIT_INDEX_FILE} = $orig_git_index;
691 } else {
692 delete $ENV{GIT_INDEX_FILE};
695 # Now switch back to the branch we were in before all of this happened
696 if($orig_branch) {
697 print "DONE\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
698 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
699 if $forward_master;
700 unless ($opt_i) {
701 system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD');
702 die "read-tree failed: $?\n" if $?;
704 } else {
705 $orig_branch = "master";
706 print "DONE; creating $orig_branch branch\n" if $opt_v and (not defined $opt_l or $opt_l > 0);
707 system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master")
708 unless -f "$git_dir/refs/heads/master";
709 unlink("$git_dir/HEAD");
710 symlink("refs/heads/$orig_branch","$git_dir/HEAD");
711 unless ($opt_i) {
712 system('git checkout');
713 die "checkout failed: $?\n" if $?;
716 unlink("$git_dir/SVN2GIT_HEAD");
717 close(BRANCHES);