1#!/usr/bin/env perl
2# Copyright (C) 2006, Eric Wong <normalperson@yhbt.net>
3# License: GPL v2 or later
4use warnings;
5use strict;
6use vars qw/ $AUTHOR $VERSION
7 $SVN_URL $SVN_INFO $SVN_WC $SVN_UUID
8 $GIT_SVN_INDEX $GIT_SVN
9 $GIT_DIR $REV_DIR/;
10$AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
11$VERSION = '0.11.0';
12use Cwd qw/abs_path/;
14$GIT_DIR = abs_path($ENV{GIT_DIR} || '.git');
15$ENV{GIT_DIR} = $GIT_DIR;
16# make sure the svn binary gives consistent output between locales and TZs:
18$ENV{TZ} = 'UTC';
19$ENV{LC_ALL} = 'C';
20# If SVN:: library support is added, please make the dependencies
22# optional and preserve the capability to use the command-line client.
23# use eval { require SVN::... } to make it lazy load
24# We don't use any modules not in the standard Perl distribution:
25use Carp qw/croak/;
26use IO::File qw//;
27use File::Basename qw/dirname basename/;
28use File::Path qw/mkpath/;
29use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
30use File::Spec qw//;
31use POSIX qw/strftime/;
32my $sha1 = qr/[a-f\d]{40}/;
33my $sha1_short = qr/[a-f\d]{4,40}/;
34my ($_revision,$_stdin,$_no_ignore_ext,$_no_stop_copy,$_help,$_rmdir,$_edit,
35 $_find_copies_harder, $_l, $_version, $_upgrade, $_authors);
36my (@_branch_from, %tree_map, %users);
37my $_svn_co_url_revs;
38my %fc_opts = ( 'no-ignore-externals' => \$_no_ignore_ext,
40 'branch|b=s' => \@_branch_from,
41 'authors-file|A=s' => \$_authors );
42my %cmd = (
43 fetch => [ \&fetch, "Download new revisions from SVN",
44 { 'revision|r=s' => \$_revision, %fc_opts } ],
45 init => [ \&init, "Initialize and fetch (import)", { } ],
46 commit => [ \&commit, "Commit git revisions to SVN",
47 { 'stdin|' => \$_stdin,
48 'edit|e' => \$_edit,
49 'rmdir' => \$_rmdir,
50 'find-copies-harder' => \$_find_copies_harder,
51 'l=i' => \$_l,
52 %fc_opts,
53 } ],
54 'show-ignore' => [ \&show_ignore, "Show svn:ignore listings", { } ],
55 rebuild => [ \&rebuild, "Rebuild git-svn metadata (after git clone)",
56 { 'no-ignore-externals' => \$_no_ignore_ext,
57 'upgrade' => \$_upgrade } ],
58);
59my $cmd;
60for (my $i = 0; $i < @ARGV; $i++) {
61 if (defined $cmd{$ARGV[$i]}) {
62 $cmd = $ARGV[$i];
63 splice @ARGV, $i, 1;
64 last;
65 }
66};
67my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
69GetOptions(%opts, 'help|H|h' => \$_help,
71 'version|V' => \$_version,
72 'id|i=s' => \$GIT_SVN) or exit 1;
73$GIT_SVN ||= $ENV{GIT_SVN_ID} || 'git-svn';
75$GIT_SVN_INDEX = "$GIT_DIR/$GIT_SVN/index";
76$SVN_URL = undef;
77$REV_DIR = "$GIT_DIR/$GIT_SVN/revs";
78$SVN_WC = "$GIT_DIR/$GIT_SVN/tree";
79usage(0) if $_help;
81version() if $_version;
82usage(1) unless defined $cmd;
83load_authors() if $_authors;
84svn_compat_check();
85$cmd{$cmd}->[0]->(@ARGV);
86exit 0;
87####################### primary functions ######################
89sub usage {
90 my $exit = shift || 0;
91 my $fd = $exit ? \*STDERR : \*STDOUT;
92 print $fd <<"";
93git-svn - bidirectional operations between a single Subversion tree and git
94Usage: $0 <command> [options] [arguments]\n
95 print $fd "Available commands:\n" unless $cmd;
97 foreach (sort keys %cmd) {
99 next if $cmd && $cmd ne $_;
100 print $fd ' ',pack('A13',$_),$cmd{$_}->[1],"\n";
101 foreach (keys %{$cmd{$_}->[2]}) {
102 # prints out arguments as they should be passed:
103 my $x = s#=s$## ? '<arg>' : s#=i$## ? '<num>' : '';
104 print $fd ' ' x 17, join(', ', map { length $_ > 1 ?
105 "--$_" : "-$_" }
106 split /\|/,$_)," $x\n";
107 }
108 }
109 print $fd <<"";
110\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
111arbitrary identifier if you're tracking multiple SVN branches/repositories in
112one git repository and want to keep them separate. See git-svn(1) for more
113information.
114 exit $exit;
116}
117sub version {
119 print "git-svn version $VERSION\n";
120 exit 0;
121}
122sub rebuild {
124 $SVN_URL = shift or undef;
125 my $newest_rev = 0;
126 if ($_upgrade) {
127 sys('git-update-ref',"refs/remotes/$GIT_SVN","$GIT_SVN-HEAD");
128 } else {
129 check_upgrade_needed();
130 }
131 my $pid = open(my $rev_list,'-|');
133 defined $pid or croak $!;
134 if ($pid == 0) {
135 exec("git-rev-list","refs/remotes/$GIT_SVN") or croak $!;
136 }
137 my $latest;
138 while (<$rev_list>) {
139 chomp;
140 my $c = $_;
141 croak "Non-SHA1: $c\n" unless $c =~ /^$sha1$/o;
142 my @commit = grep(/^git-svn-id: /,`git-cat-file commit $c`);
143 next if (!@commit); # skip merges
144 my $id = $commit[$#commit];
145 my ($url, $rev, $uuid) = ($id =~ /^git-svn-id:\s(\S+?)\@(\d+)
146 \s([a-f\d\-]+)$/x);
147 if (!$rev || !$uuid || !$url) {
148 # some of the original repositories I made had
149 # indentifiers like this:
150 ($rev, $uuid) = ($id =~/^git-svn-id:\s(\d+)
151 \@([a-f\d\-]+)/x);
152 if (!$rev || !$uuid) {
153 croak "Unable to extract revision or UUID from ",
154 "$c, $id\n";
155 }
156 }
157 # if we merged or otherwise started elsewhere, this is
159 # how we break out of it
160 next if (defined $SVN_UUID && ($uuid ne $SVN_UUID));
161 next if (defined $SVN_URL && defined $url && ($url ne $SVN_URL));
162 print "r$rev = $c\n";
164 unless (defined $latest) {
165 if (!$SVN_URL && !$url) {
166 croak "SVN repository location required: $url\n";
167 }
168 $SVN_URL ||= $url;
169 $SVN_UUID ||= $uuid;
170 setup_git_svn();
171 $latest = $rev;
172 }
173 assert_revision_eq_or_unknown($rev, $c);
174 sys('git-update-ref',"$GIT_SVN/revs/$rev",$c);
175 $newest_rev = $rev if ($rev > $newest_rev);
176 }
177 close $rev_list or croak $?;
178 if (!chdir $SVN_WC) {
179 svn_cmd_checkout($SVN_URL, $latest, $SVN_WC);
180 chdir $SVN_WC or croak $!;
181 }
182 $pid = fork;
184 defined $pid or croak $!;
185 if ($pid == 0) {
186 my @svn_up = qw(svn up);
187 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
188 sys(@svn_up,"-r$newest_rev");
189 $ENV{GIT_INDEX_FILE} = $GIT_SVN_INDEX;
190 git_addremove();
191 exec('git-write-tree');
192 }
193 waitpid $pid, 0;
194 if ($_upgrade) {
196 print STDERR <<"";
197Keeping deprecated refs/head/$GIT_SVN-HEAD for now. Please remove it
198when you have upgraded your tools and habits to use refs/remotes/$GIT_SVN
199 }
201}
202sub init {
204 $SVN_URL = shift or croak "SVN repository location required\n";
205 unless (-d $GIT_DIR) {
206 sys('git-init-db');
207 }
208 setup_git_svn();
209}
210sub fetch {
212 my (@parents) = @_;
213 check_upgrade_needed();
214 $SVN_URL ||= file_to_s("$GIT_DIR/$GIT_SVN/info/url");
215 my @log_args = -d $SVN_WC ? ($SVN_WC) : ($SVN_URL);
216 unless ($_revision) {
217 $_revision = -d $SVN_WC ? 'BASE:HEAD' : '0:HEAD';
218 }
219 push @log_args, "-r$_revision";
220 push @log_args, '--stop-on-copy' unless $_no_stop_copy;
221 my $svn_log = svn_log_raw(@log_args);
223 my $base = next_log_entry($svn_log) or croak "No base revision!\n";
225 my $last_commit = undef;
226 unless (-d $SVN_WC) {
227 svn_cmd_checkout($SVN_URL,$base->{revision},$SVN_WC);
228 chdir $SVN_WC or croak $!;
229 read_uuid();
230 $last_commit = git_commit($base, @parents);
231 assert_svn_wc_clean($base->{revision}, $last_commit);
232 } else {
233 chdir $SVN_WC or croak $!;
234 read_uuid();
235 $last_commit = file_to_s("$REV_DIR/$base->{revision}");
236 }
237 my @svn_up = qw(svn up);
238 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
239 my $last = $base;
240 while (my $log_msg = next_log_entry($svn_log)) {
241 assert_svn_wc_clean($last->{revision}, $last_commit);
242 if ($last->{revision} >= $log_msg->{revision}) {
243 croak "Out of order: last >= current: ",
244 "$last->{revision} >= $log_msg->{revision}\n";
245 }
246 sys(@svn_up,"-r$log_msg->{revision}");
247 $last_commit = git_commit($log_msg, $last_commit, @parents);
248 $last = $log_msg;
249 }
250 assert_svn_wc_clean($last->{revision}, $last_commit);
251 unless (-e "$GIT_DIR/refs/heads/master") {
252 sys(qw(git-update-ref refs/heads/master),$last_commit);
253 }
254 return $last;
255}
256sub commit {
258 my (@commits) = @_;
259 check_upgrade_needed();
260 if ($_stdin || !@commits) {
261 print "Reading from stdin...\n";
262 @commits = ();
263 while (<STDIN>) {
264 if (/\b($sha1_short)\b/o) {
265 unshift @commits, $1;
266 }
267 }
268 }
269 my @revs;
270 foreach my $c (@commits) {
271 chomp(my @tmp = safe_qx('git-rev-parse',$c));
272 if (scalar @tmp == 1) {
273 push @revs, $tmp[0];
274 } elsif (scalar @tmp > 1) {
275 push @revs, reverse (safe_qx('git-rev-list',@tmp));
276 } else {
277 die "Failed to rev-parse $c\n";
278 }
279 }
280 chomp @revs;
281 fetch();
283 chdir $SVN_WC or croak $!;
284 my $info = svn_info('.');
285 read_uuid($info);
286 my $svn_current_rev = $info->{'Last Changed Rev'};
287 foreach my $c (@revs) {
288 my $mods = svn_checkout_tree($svn_current_rev, $c);
289 if (scalar @$mods == 0) {
290 print "Skipping, no changes detected\n";
291 next;
292 }
293 $svn_current_rev = svn_commit_tree($svn_current_rev, $c);
294 }
295 print "Done committing ",scalar @revs," revisions to SVN\n";
296}
298sub show_ignore {
300 require File::Find or die $!;
301 my $exclude_file = "$GIT_DIR/info/exclude";
302 open my $fh, '<', $exclude_file or croak $!;
303 chomp(my @excludes = (<$fh>));
304 close $fh or croak $!;
305 $SVN_URL ||= file_to_s("$GIT_DIR/$GIT_SVN/info/url");
307 chdir $SVN_WC or croak $!;
308 my %ign;
309 File::Find::find({wanted=>sub{if(lstat $_ && -d _ && -d "$_/.svn"){
310 s#^\./##;
311 @{$ign{$_}} = safe_qx(qw(svn propget svn:ignore),$_);
312 }}, no_chdir=>1},'.');
313 print "\n# /\n";
315 foreach (@{$ign{'.'}}) { print '/',$_ if /\S/ }
316 delete $ign{'.'};
317 foreach my $i (sort keys %ign) {
318 print "\n# ",$i,"\n";
319 foreach (@{$ign{$i}}) { print '/',$i,'/',$_ if /\S/ }
320 }
321}
322########################### utility functions #########################
324sub read_uuid {
326 return if $SVN_UUID;
327 my $info = shift || svn_info('.');
328 $SVN_UUID = $info->{'Repository UUID'} or
329 croak "Repository UUID unreadable\n";
330 s_to_file($SVN_UUID,"$GIT_DIR/$GIT_SVN/info/uuid");
331}
332sub setup_git_svn {
334 defined $SVN_URL or croak "SVN repository location required\n";
335 unless (-d $GIT_DIR) {
336 croak "GIT_DIR=$GIT_DIR does not exist!\n";
337 }
338 mkpath(["$GIT_DIR/$GIT_SVN"]);
339 mkpath(["$GIT_DIR/$GIT_SVN/info"]);
340 mkpath([$REV_DIR]);
341 s_to_file($SVN_URL,"$GIT_DIR/$GIT_SVN/info/url");
342 open my $fd, '>>', "$GIT_DIR/$GIT_SVN/info/exclude" or croak $!;
344 print $fd '.svn',"\n";
345 close $fd or croak $!;
346}
347sub assert_svn_wc_clean {
349 my ($svn_rev, $treeish) = @_;
350 croak "$svn_rev is not an integer!\n" unless ($svn_rev =~ /^\d+$/);
351 croak "$treeish is not a sha1!\n" unless ($treeish =~ /^$sha1$/o);
352 my $lcr = svn_info('.')->{'Last Changed Rev'};
353 if ($svn_rev != $lcr) {
354 print STDERR "Checking for copy-tree ... ";
355 # use
356 my @diff = grep(/^Index: /,(safe_qx(qw(svn diff),
357 "-r$lcr:$svn_rev")));
358 if (@diff) {
359 croak "Nope! Expected r$svn_rev, got r$lcr\n";
360 } else {
361 print STDERR "OK!\n";
362 }
363 }
364 my @status = grep(!/^Performing status on external/,(`svn status`));
365 @status = grep(!/^\s*$/,@status);
366 if (scalar @status) {
367 print STDERR "Tree ($SVN_WC) is not clean:\n";
368 print STDERR $_ foreach @status;
369 croak;
370 }
371 assert_tree($treeish);
372}
373sub assert_tree {
375 my ($treeish) = @_;
376 croak "Not a sha1: $treeish\n" unless $treeish =~ /^$sha1$/o;
377 chomp(my $type = `git-cat-file -t $treeish`);
378 my $expected;
379 while ($type eq 'tag') {
380 chomp(($treeish, $type) = `git-cat-file tag $treeish`);
381 }
382 if ($type eq 'commit') {
383 $expected = (grep /^tree /,`git-cat-file commit $treeish`)[0];
384 ($expected) = ($expected =~ /^tree ($sha1)$/);
385 die "Unable to get tree from $treeish\n" unless $expected;
386 } elsif ($type eq 'tree') {
387 $expected = $treeish;
388 } else {
389 die "$treeish is a $type, expected tree, tag or commit\n";
390 }
391 my $old_index = $ENV{GIT_INDEX_FILE};
393 my $tmpindex = $GIT_SVN_INDEX.'.assert-tmp';
394 if (-e $tmpindex) {
395 unlink $tmpindex or croak $!;
396 }
397 $ENV{GIT_INDEX_FILE} = $tmpindex;
398 git_addremove();
399 chomp(my $tree = `git-write-tree`);
400 if ($old_index) {
401 $ENV{GIT_INDEX_FILE} = $old_index;
402 } else {
403 delete $ENV{GIT_INDEX_FILE};
404 }
405 if ($tree ne $expected) {
406 croak "Tree mismatch, Got: $tree, Expected: $expected\n";
407 }
408}
409sub parse_diff_tree {
411 my $diff_fh = shift;
412 local $/ = "\0";
413 my $state = 'meta';
414 my @mods;
415 while (<$diff_fh>) {
416 chomp $_; # this gets rid of the trailing "\0"
417 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
418 $sha1\s($sha1)\s([MTCRAD])\d*$/xo) {
419 push @mods, { mode_a => $1, mode_b => $2,
420 sha1_b => $3, chg => $4 };
421 if ($4 =~ /^(?:C|R)$/) {
422 $state = 'file_a';
423 } else {
424 $state = 'file_b';
425 }
426 } elsif ($state eq 'file_a') {
427 my $x = $mods[$#mods] or croak "Empty array\n";
428 if ($x->{chg} !~ /^(?:C|R)$/) {
429 croak "Error parsing $_, $x->{chg}\n";
430 }
431 $x->{file_a} = $_;
432 $state = 'file_b';
433 } elsif ($state eq 'file_b') {
434 my $x = $mods[$#mods] or croak "Empty array\n";
435 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
436 croak "Error parsing $_, $x->{chg}\n";
437 }
438 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
439 croak "Error parsing $_, $x->{chg}\n";
440 }
441 $x->{file_b} = $_;
442 $state = 'meta';
443 } else {
444 croak "Error parsing $_\n";
445 }
446 }
447 close $diff_fh or croak $!;
448 return \@mods;
450}
451sub svn_check_prop_executable {
453 my $m = shift;
454 return if -l $m->{file_b};
455 if ($m->{mode_b} =~ /755$/) {
456 chmod((0755 &~ umask),$m->{file_b}) or croak $!;
457 if ($m->{mode_a} !~ /755$/) {
458 sys(qw(svn propset svn:executable 1), $m->{file_b});
459 }
460 -x $m->{file_b} or croak "$m->{file_b} is not executable!\n";
461 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
462 sys(qw(svn propdel svn:executable), $m->{file_b});
463 chmod((0644 &~ umask),$m->{file_b}) or croak $!;
464 -x $m->{file_b} and croak "$m->{file_b} is executable!\n";
465 }
466}
467sub svn_ensure_parent_path {
469 my $dir_b = dirname(shift);
470 svn_ensure_parent_path($dir_b) if ($dir_b ne File::Spec->curdir);
471 mkpath([$dir_b]) unless (-d $dir_b);
472 sys(qw(svn add -N), $dir_b) unless (-d "$dir_b/.svn");
473}
474sub precommit_check {
476 my $mods = shift;
477 my (%rm_file, %rmdir_check, %added_check);
478 my %o = ( D => 0, R => 1, C => 2, A => 3, M => 3, T => 3 );
480 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
481 if ($m->{chg} eq 'R') {
482 if (-d $m->{file_b}) {
483 err_dir_to_file("$m->{file_a} => $m->{file_b}");
484 }
485 # dir/$file => dir/file/$file
486 my $dirname = dirname($m->{file_b});
487 while ($dirname ne File::Spec->curdir) {
488 if ($dirname ne $m->{file_a}) {
489 $dirname = dirname($dirname);
490 next;
491 }
492 err_file_to_dir("$m->{file_a} => $m->{file_b}");
493 }
494 # baz/zzz => baz (baz is a file)
495 $dirname = dirname($m->{file_a});
496 while ($dirname ne File::Spec->curdir) {
497 if ($dirname ne $m->{file_b}) {
498 $dirname = dirname($dirname);
499 next;
500 }
501 err_dir_to_file("$m->{file_a} => $m->{file_b}");
502 }
503 }
504 if ($m->{chg} =~ /^(D|R)$/) {
505 my $t = $1 eq 'D' ? 'file_b' : 'file_a';
506 $rm_file{ $m->{$t} } = 1;
507 my $dirname = dirname( $m->{$t} );
508 my $basename = basename( $m->{$t} );
509 $rmdir_check{$dirname}->{$basename} = 1;
510 } elsif ($m->{chg} =~ /^(?:A|C)$/) {
511 if (-d $m->{file_b}) {
512 err_dir_to_file($m->{file_b});
513 }
514 my $dirname = dirname( $m->{file_b} );
515 my $basename = basename( $m->{file_b} );
516 $added_check{$dirname}->{$basename} = 1;
517 while ($dirname ne File::Spec->curdir) {
518 if ($rm_file{$dirname}) {
519 err_file_to_dir($m->{file_b});
520 }
521 $dirname = dirname $dirname;
522 }
523 }
524 }
525 return (\%rmdir_check, \%added_check);
526 sub err_dir_to_file {
528 my $file = shift;
529 print STDERR "Node change from directory to file ",
530 "is not supported by Subversion: ",$file,"\n";
531 exit 1;
532 }
533 sub err_file_to_dir {
534 my $file = shift;
535 print STDERR "Node change from file to directory ",
536 "is not supported by Subversion: ",$file,"\n";
537 exit 1;
538 }
539}
540sub svn_checkout_tree {
542 my ($svn_rev, $treeish) = @_;
543 my $from = file_to_s("$REV_DIR/$svn_rev");
544 assert_svn_wc_clean($svn_rev,$from);
545 print "diff-tree $from $treeish\n";
546 my $pid = open my $diff_fh, '-|';
547 defined $pid or croak $!;
548 if ($pid == 0) {
549 my @diff_tree = qw(git-diff-tree -z -r -C);
550 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
551 push @diff_tree, "-l$_l" if defined $_l;
552 exec(@diff_tree, $from, $treeish) or croak $!;
553 }
554 my $mods = parse_diff_tree($diff_fh);
555 unless (@$mods) {
556 # git can do empty commits, but SVN doesn't allow it...
557 return $mods;
558 }
559 my ($rm, $add) = precommit_check($mods);
560 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
562 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
563 if ($m->{chg} eq 'C') {
564 svn_ensure_parent_path( $m->{file_b} );
565 sys(qw(svn cp), $m->{file_a}, $m->{file_b});
566 apply_mod_line_blob($m);
567 svn_check_prop_executable($m);
568 } elsif ($m->{chg} eq 'D') {
569 sys(qw(svn rm --force), $m->{file_b});
570 } elsif ($m->{chg} eq 'R') {
571 svn_ensure_parent_path( $m->{file_b} );
572 sys(qw(svn mv --force), $m->{file_a}, $m->{file_b});
573 apply_mod_line_blob($m);
574 svn_check_prop_executable($m);
575 } elsif ($m->{chg} eq 'M') {
576 apply_mod_line_blob($m);
577 svn_check_prop_executable($m);
578 } elsif ($m->{chg} eq 'T') {
579 sys(qw(svn rm --force),$m->{file_b});
580 apply_mod_line_blob($m);
581 sys(qw(svn add --force), $m->{file_b});
582 svn_check_prop_executable($m);
583 } elsif ($m->{chg} eq 'A') {
584 svn_ensure_parent_path( $m->{file_b} );
585 apply_mod_line_blob($m);
586 sys(qw(svn add --force), $m->{file_b});
587 svn_check_prop_executable($m);
588 } else {
589 croak "Invalid chg: $m->{chg}\n";
590 }
591 }
592 assert_tree($treeish);
594 if ($_rmdir) { # remove empty directories
595 handle_rmdir($rm, $add);
596 }
597 assert_tree($treeish);
598 return $mods;
599}
600# svn ls doesn't work with respect to the current working tree, but what's
602# in the repository. There's not even an option for it... *sigh*
603# (added files don't show up and removed files remain in the ls listing)
604sub svn_ls_current {
605 my ($dir, $rm, $add) = @_;
606 chomp(my @ls = safe_qx('svn','ls',$dir));
607 my @ret = ();
608 foreach (@ls) {
609 s#/$##; # trailing slashes are evil
610 push @ret, $_ unless $rm->{$dir}->{$_};
611 }
612 if (exists $add->{$dir}) {
613 push @ret, keys %{$add->{$dir}};
614 }
615 return \@ret;
616}
617sub handle_rmdir {
619 my ($rm, $add) = @_;
620 foreach my $dir (sort {length $b <=> length $a} keys %$rm) {
622 my $ls = svn_ls_current($dir, $rm, $add);
623 next if (scalar @$ls);
624 sys(qw(svn rm --force),$dir);
625 my $dn = dirname $dir;
627 $rm->{ $dn }->{ basename $dir } = 1;
628 $ls = svn_ls_current($dn, $rm, $add);
629 while (scalar @$ls == 0 && $dn ne File::Spec->curdir) {
630 sys(qw(svn rm --force),$dn);
631 $dir = basename $dn;
632 $dn = dirname $dn;
633 $rm->{ $dn }->{ $dir } = 1;
634 $ls = svn_ls_current($dn, $rm, $add);
635 }
636 }
637}
638sub svn_commit_tree {
640 my ($svn_rev, $commit) = @_;
641 my $commit_msg = "$GIT_DIR/$GIT_SVN/.svn-commit.tmp.$$";
642 my %log_msg = ( msg => '' );
643 open my $msg, '>', $commit_msg or croak $!;
644 chomp(my $type = `git-cat-file -t $commit`);
646 if ($type eq 'commit') {
647 my $pid = open my $msg_fh, '-|';
648 defined $pid or croak $!;
649 if ($pid == 0) {
651 exec(qw(git-cat-file commit), $commit) or croak $!;
652 }
653 my $in_msg = 0;
654 while (<$msg_fh>) {
655 if (!$in_msg) {
656 $in_msg = 1 if (/^\s*$/);
657 } elsif (/^git-svn-id: /) {
658 # skip this, we regenerate the correct one
659 # on re-fetch anyways
660 } else {
661 print $msg $_ or croak $!;
662 }
663 }
664 close $msg_fh or croak $!;
665 }
666 close $msg or croak $!;
667 if ($_edit || ($type eq 'tree')) {
669 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
670 system($editor, $commit_msg);
671 }
672 # file_to_s removes all trailing newlines, so just use chomp() here:
674 open $msg, '<', $commit_msg or croak $!;
675 { local $/; chomp($log_msg{msg} = <$msg>); }
676 close $msg or croak $!;
677 my ($oneline) = ($log_msg{msg} =~ /([^\n\r]+)/);
679 print "Committing $commit: $oneline\n";
680 my @ci_output = safe_qx(qw(svn commit -F),$commit_msg);
682 my ($committed) = grep(/^Committed revision \d+\./,@ci_output);
683 unlink $commit_msg;
684 defined $committed or croak
685 "Commit output failed to parse committed revision!\n",
686 join("\n",@ci_output),"\n";
687 my ($rev_committed) = ($committed =~ /^Committed revision (\d+)\./);
688 my @svn_up = qw(svn up);
690 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
691 if ($rev_committed == ($svn_rev + 1)) {
692 push @svn_up, "-r$rev_committed";
693 sys(@svn_up);
694 my $info = svn_info('.');
695 my $date = $info->{'Last Changed Date'} or die "Missing date\n";
696 if ($info->{'Last Changed Rev'} != $rev_committed) {
697 croak "$info->{'Last Changed Rev'} != $rev_committed\n"
698 }
699 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
700 /(\d{4})\-(\d\d)\-(\d\d)\s
701 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
702 or croak "Failed to parse date: $date\n";
703 $log_msg{date} = "$tz $Y-$m-$d $H:$M:$S";
704 $log_msg{author} = $info->{'Last Changed Author'};
705 $log_msg{revision} = $rev_committed;
706 $log_msg{msg} .= "\n";
707 my $parent = file_to_s("$REV_DIR/$svn_rev");
708 git_commit(\%log_msg, $parent, $commit);
709 return $rev_committed;
710 }
711 # resync immediately
712 push @svn_up, "-r$svn_rev";
713 sys(@svn_up);
714 return fetch("$rev_committed=$commit")->{revision};
715}
716# read the entire log into a temporary file (which is removed ASAP)
718# and store the file handle + parser state
719sub svn_log_raw {
720 my (@log_args) = @_;
721 my $log_fh = IO::File->new_tmpfile or croak $!;
722 my $pid = fork;
723 defined $pid or croak $!;
724 if (!$pid) {
725 open STDOUT, '>&', $log_fh or croak $!;
726 exec (qw(svn log), @log_args) or croak $!
727 }
728 waitpid $pid, 0;
729 croak if $?;
730 seek $log_fh, 0, 0 or croak $!;
731 return { state => 'sep', fh => $log_fh };
732}
733sub next_log_entry {
735 my $log = shift; # retval of svn_log_raw()
736 my $ret = undef;
737 my $fh = $log->{fh};
738 while (<$fh>) {
740 chomp;
741 if (/^\-{72}$/) {
742 if ($log->{state} eq 'msg') {
743 if ($ret->{lines}) {
744 $ret->{msg} .= $_."\n";
745 unless(--$ret->{lines}) {
746 $log->{state} = 'sep';
747 }
748 } else {
749 croak "Log parse error at: $_\n",
750 $ret->{revision},
751 "\n";
752 }
753 next;
754 }
755 if ($log->{state} ne 'sep') {
756 croak "Log parse error at: $_\n",
757 "state: $log->{state}\n",
758 $ret->{revision},
759 "\n";
760 }
761 $log->{state} = 'rev';
762 # if we have an empty log message, put something there:
764 if ($ret) {
765 $ret->{msg} ||= "\n";
766 delete $ret->{lines};
767 return $ret;
768 }
769 next;
770 }
771 if ($log->{state} eq 'rev' && s/^r(\d+)\s*\|\s*//) {
772 my $rev = $1;
773 my ($author, $date, $lines) = split(/\s*\|\s*/, $_, 3);
774 ($lines) = ($lines =~ /(\d+)/);
775 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
776 /(\d{4})\-(\d\d)\-(\d\d)\s
777 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
778 or croak "Failed to parse date: $date\n";
779 $ret = { revision => $rev,
780 date => "$tz $Y-$m-$d $H:$M:$S",
781 author => $author,
782 lines => $lines,
783 msg => '' };
784 if (defined $_authors && ! defined $users{$author}) {
785 die "Author: $author not defined in ",
786 "$_authors file\n";
787 }
788 $log->{state} = 'msg_start';
789 next;
790 }
791 # skip the first blank line of the message:
792 if ($log->{state} eq 'msg_start' && /^$/) {
793 $log->{state} = 'msg';
794 } elsif ($log->{state} eq 'msg') {
795 if ($ret->{lines}) {
796 $ret->{msg} .= $_."\n";
797 unless (--$ret->{lines}) {
798 $log->{state} = 'sep';
799 }
800 } else {
801 croak "Log parse error at: $_\n",
802 $ret->{revision},"\n";
803 }
804 }
805 }
806 return $ret;
807}
808sub svn_info {
810 my $url = shift || $SVN_URL;
811 my $pid = open my $info_fh, '-|';
813 defined $pid or croak $!;
814 if ($pid == 0) {
816 exec(qw(svn info),$url) or croak $!;
817 }
818 my $ret = {};
820 # only single-lines seem to exist in svn info output
821 while (<$info_fh>) {
822 chomp $_;
823 if (m#^([^:]+)\s*:\s*(\S.*)$#) {
824 $ret->{$1} = $2;
825 push @{$ret->{-order}}, $1;
826 }
827 }
828 close $info_fh or croak $!;
829 return $ret;
830}
831sub sys { system(@_) == 0 or croak $? }
833sub git_addremove {
835 system( "git-diff-files --name-only -z ".
836 " | git-update-index --remove -z --stdin && ".
837 "git-ls-files -z --others ".
838 "'--exclude-from=$GIT_DIR/$GIT_SVN/info/exclude'".
839 " | git-update-index --add -z --stdin"
840 ) == 0 or croak $?
841}
842sub s_to_file {
844 my ($str, $file, $mode) = @_;
845 open my $fd,'>',$file or croak $!;
846 print $fd $str,"\n" or croak $!;
847 close $fd or croak $!;
848 chmod ($mode &~ umask, $file) if (defined $mode);
849}
850sub file_to_s {
852 my $file = shift;
853 open my $fd,'<',$file or croak "$!: file: $file\n";
854 local $/;
855 my $ret = <$fd>;
856 close $fd or croak $!;
857 $ret =~ s/\s*$//s;
858 return $ret;
859}
860sub assert_revision_unknown {
862 my $revno = shift;
863 if (-f "$REV_DIR/$revno") {
864 croak "$REV_DIR/$revno already exists! ",
865 "Why are we refetching it?";
866 }
867}
868sub trees_eq {
870 my ($x, $y) = @_;
871 my @x = safe_qx('git-cat-file','commit',$x);
872 my @y = safe_qx('git-cat-file','commit',$y);
873 if (($y[0] ne $x[0]) || $x[0] !~ /^tree $sha1\n$/
874 || $y[0] !~ /^tree $sha1\n$/) {
875 print STDERR "Trees not equal: $y[0] != $x[0]\n";
876 return 0
877 }
878 return 1;
879}
880sub assert_revision_eq_or_unknown {
882 my ($revno, $commit) = @_;
883 if (-f "$REV_DIR/$revno") {
884 my $current = file_to_s("$REV_DIR/$revno");
885 if (($commit ne $current) && !trees_eq($commit, $current)) {
886 croak "$REV_DIR/$revno already exists!\n",
887 "current: $current\nexpected: $commit\n";
888 }
889 return;
890 }
891}
892sub git_commit {
894 my ($log_msg, @parents) = @_;
895 assert_revision_unknown($log_msg->{revision});
896 my $out_fh = IO::File->new_tmpfile or croak $!;
897 map_tree_joins() if (@_branch_from && !%tree_map);
899 # commit parents can be conditionally bound to a particular
901 # svn revision via: "svn_revno=commit_sha1", filter them out here:
902 my @exec_parents;
903 foreach my $p (@parents) {
904 next unless defined $p;
905 if ($p =~ /^(\d+)=($sha1_short)$/o) {
906 if ($1 == $log_msg->{revision}) {
907 push @exec_parents, $2;
908 }
909 } else {
910 push @exec_parents, $p if $p =~ /$sha1_short/o;
911 }
912 }
913 my $pid = fork;
915 defined $pid or croak $!;
916 if ($pid == 0) {
917 $ENV{GIT_INDEX_FILE} = $GIT_SVN_INDEX;
918 git_addremove();
919 chomp(my $tree = `git-write-tree`);
920 croak if $?;
921 if (exists $tree_map{$tree}) {
922 my %seen_parent = map { $_ => 1 } @exec_parents;
923 foreach (@{$tree_map{$tree}}) {
924 # MAXPARENT is defined to 16 in commit-tree.c:
925 if ($seen_parent{$_} || @exec_parents > 16) {
926 next;
927 }
928 push @exec_parents, $_;
929 $seen_parent{$_} = 1;
930 }
931 }
932 my $msg_fh = IO::File->new_tmpfile or croak $!;
933 print $msg_fh $log_msg->{msg}, "\ngit-svn-id: ",
934 "$SVN_URL\@$log_msg->{revision}",
935 " $SVN_UUID\n" or croak $!;
936 $msg_fh->flush == 0 or croak $!;
937 seek $msg_fh, 0, 0 or croak $!;
938 set_commit_env($log_msg);
940 my @exec = ('git-commit-tree',$tree);
942 push @exec, '-p', $_ foreach @exec_parents;
943 open STDIN, '<&', $msg_fh or croak $!;
944 open STDOUT, '>&', $out_fh or croak $!;
945 exec @exec or croak $!;
946 }
947 waitpid($pid,0);
948 croak if $?;
949 $out_fh->flush == 0 or croak $!;
951 seek $out_fh, 0, 0 or croak $!;
952 chomp(my $commit = do { local $/; <$out_fh> });
953 if ($commit !~ /^$sha1$/o) {
954 croak "Failed to commit, invalid sha1: $commit\n";
955 }
956 my @update_ref = ('git-update-ref',"refs/remotes/$GIT_SVN",$commit);
957 if (my $primary_parent = shift @exec_parents) {
958 $pid = fork;
959 defined $pid or croak $!;
960 if (!$pid) {
961 close STDERR;
962 close STDOUT;
963 exec 'git-rev-parse','--verify',
964 "refs/remotes/$GIT_SVN^0";
965 }
966 waitpid $pid, 0;
967 push @update_ref, $primary_parent unless $?;
968 }
969 sys(@update_ref);
970 sys('git-update-ref',"$GIT_SVN/revs/$log_msg->{revision}",$commit);
971 print "r$log_msg->{revision} = $commit\n";
972 return $commit;
973}
974sub set_commit_env {
976 my ($log_msg) = @_;
977 my $author = $log_msg->{author};
978 my ($name,$email) = defined $users{$author} ? @{$users{$author}}
979 : ($author,"$author\@$SVN_UUID");
980 $ENV{GIT_AUTHOR_NAME} = $ENV{GIT_COMMITTER_NAME} = $name;
981 $ENV{GIT_AUTHOR_EMAIL} = $ENV{GIT_COMMITTER_EMAIL} = $email;
982 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_msg->{date};
983}
984sub apply_mod_line_blob {
986 my $m = shift;
987 if ($m->{mode_b} =~ /^120/) {
988 blob_to_symlink($m->{sha1_b}, $m->{file_b});
989 } else {
990 blob_to_file($m->{sha1_b}, $m->{file_b});
991 }
992}
993sub blob_to_symlink {
995 my ($blob, $link) = @_;
996 defined $link or croak "\$link not defined!\n";
997 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
998 if (-l $link || -f _) {
999 unlink $link or croak $!;
1000 }
1001 my $dest = `git-cat-file blob $blob`; # no newline, so no chomp
1003 symlink $dest, $link or croak $!;
1004}
1005sub blob_to_file {
1007 my ($blob, $file) = @_;
1008 defined $file or croak "\$file not defined!\n";
1009 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
1010 if (-l $file || -f _) {
1011 unlink $file or croak $!;
1012 }
1013 open my $blob_fh, '>', $file or croak "$!: $file\n";
1015 my $pid = fork;
1016 defined $pid or croak $!;
1017 if ($pid == 0) {
1019 open STDOUT, '>&', $blob_fh or croak $!;
1020 exec('git-cat-file','blob',$blob);
1021 }
1022 waitpid $pid, 0;
1023 croak $? if $?;
1024 close $blob_fh or croak $!;
1026}
1027sub safe_qx {
1029 my $pid = open my $child, '-|';
1030 defined $pid or croak $!;
1031 if ($pid == 0) {
1032 exec(@_) or croak $?;
1033 }
1034 my @ret = (<$child>);
1035 close $child or croak $?;
1036 die $? if $?; # just in case close didn't error out
1037 return wantarray ? @ret : join('',@ret);
1038}
1039sub svn_compat_check {
1041 my @co_help = safe_qx(qw(svn co -h));
1042 unless (grep /ignore-externals/,@co_help) {
1043 print STDERR "W: Installed svn version does not support ",
1044 "--ignore-externals\n";
1045 $_no_ignore_ext = 1;
1046 }
1047 if (grep /usage: checkout URL\[\@REV\]/,@co_help) {
1048 $_svn_co_url_revs = 1;
1049 }
1050 # I really, really hope nobody hits this...
1052 unless (grep /stop-on-copy/, (safe_qx(qw(svn log -h)))) {
1053 print STDERR <<'';
1054W: The installed svn version does not support the --stop-on-copy flag in
1055 the log command.
1056 Lets hope the directory you're tracking is not a branch or tag
1057 and was never moved within the repository...
1058 $_no_stop_copy = 1;
1060 }
1061}
1062# *sigh*, new versions of svn won't honor -r<rev> without URL@<rev>,
1064# (and they won't honor URL@<rev> without -r<rev>, too!)
1065sub svn_cmd_checkout {
1066 my ($url, $rev, $dir) = @_;
1067 my @cmd = ('svn','co', "-r$rev");
1068 push @cmd, '--ignore-externals' unless $_no_ignore_ext;
1069 $url .= "\@$rev" if $_svn_co_url_revs;
1070 sys(@cmd, $url, $dir);
1071}
1072sub check_upgrade_needed {
1074 my $old = eval {
1075 my $pid = open my $child, '-|';
1076 defined $pid or croak $!;
1077 if ($pid == 0) {
1078 close STDERR;
1079 exec('git-rev-parse',"$GIT_SVN-HEAD") or croak $?;
1080 }
1081 my @ret = (<$child>);
1082 close $child or croak $?;
1083 die $? if $?; # just in case close didn't error out
1084 return wantarray ? @ret : join('',@ret);
1085 };
1086 return unless $old;
1087 my $head = eval { safe_qx('git-rev-parse',"refs/remotes/$GIT_SVN") };
1088 if ($@ || !$head) {
1089 print STDERR "Please run: $0 rebuild --upgrade\n";
1090 exit 1;
1091 }
1092}
1093# fills %tree_map with a reverse mapping of trees to commits. Useful
1095# for finding parents to commit on.
1096sub map_tree_joins {
1097 foreach my $br (@_branch_from) {
1098 my $pid = open my $pipe, '-|';
1099 defined $pid or croak $!;
1100 if ($pid == 0) {
1101 exec(qw(git-rev-list --pretty=raw), $br) or croak $?;
1102 }
1103 while (<$pipe>) {
1104 if (/^commit ($sha1)$/o) {
1105 my $commit = $1;
1106 my ($tree) = (<$pipe> =~ /^tree ($sha1)$/o);
1107 unless (defined $tree) {
1108 die "Failed to parse commit $commit\n";
1109 }
1110 push @{$tree_map{$tree}}, $commit;
1111 }
1112 }
1113 close $pipe or croak $?;
1114 }
1115}
1116# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1118sub load_authors {
1119 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1120 while (<$authors>) {
1121 chomp;
1122 next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/;
1123 my ($user, $name, $email) = ($1, $2, $3);
1124 $users{$user} = [$name, $email];
1125 }
1126 close $authors or croak $!;
1127}
1128__END__
1130Data structures:
1132$svn_log hashref (as returned by svn_log_raw)
1134{
1135 fh => file handle of the log file,
1136 state => state of the log file parser (sep/msg/rev/msg_start...)
1137}
1138$log_msg hashref as returned by next_log_entry($svn_log)
1140{
1141 msg => 'whitespace-formatted log entry
1142', # trailing newline is preserved
1143 revision => '8', # integer
1144 date => '2004-02-24T17:01:44.108345Z', # commit date
1145 author => 'committer name'
1146};
1147@mods = array of diff-index line hashes, each element represents one line
1150 of diff-index output
1151diff-index line ($m hash)
1153{
1154 mode_a => first column of diff-index output, no leading ':',
1155 mode_b => second column of diff-index output,
1156 sha1_b => sha1sum of the final blob,
1157 chg => change type [MCRADT],
1158 file_a => original file name of a file (iff chg is 'C' or 'R')
1159 file_b => new/current file name of a file (any chg)
1160}
1161;