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);
69# convert GetOpt::Long specs for use by git-repo-config
71foreach my $o (keys %opts) {
72 my $v = $opts{$o};
73 my ($key) = ($o =~ /^([a-z\-]+)/);
74 $key =~ s/-//g;
75 my $arg = 'git-repo-config';
76 $arg .= ' --int' if ($o =~ /=i$/);
77 $arg .= ' --bool' if ($o !~ /=[sfi]$/);
78 $arg .= " svn.$key"; # $key only matches [a-z\-], always shell-safe
79 if (ref $v eq 'ARRAY') {
80 chomp(my @tmp = `$arg`);
81 @$v = @tmp if @tmp;
82 } else {
83 chomp(my $tmp = `$arg`);
84 if ($tmp && !($arg =~ / --bool / && $tmp eq 'false')) {
85 $$v = $tmp;
86 }
87 }
88}
89GetOptions(%opts, 'help|H|h' => \$_help,
91 'version|V' => \$_version,
92 'id|i=s' => \$GIT_SVN) or exit 1;
93$GIT_SVN ||= $ENV{GIT_SVN_ID} || 'git-svn';
95$GIT_SVN_INDEX = "$GIT_DIR/$GIT_SVN/index";
96$SVN_URL = undef;
97$REV_DIR = "$GIT_DIR/$GIT_SVN/revs";
98$SVN_WC = "$GIT_DIR/$GIT_SVN/tree";
99usage(0) if $_help;
101version() if $_version;
102usage(1) unless defined $cmd;
103load_authors() if $_authors;
104svn_compat_check();
105$cmd{$cmd}->[0]->(@ARGV);
106exit 0;
107####################### primary functions ######################
109sub usage {
110 my $exit = shift || 0;
111 my $fd = $exit ? \*STDERR : \*STDOUT;
112 print $fd <<"";
113git-svn - bidirectional operations between a single Subversion tree and git
114Usage: $0 <command> [options] [arguments]\n
115 print $fd "Available commands:\n" unless $cmd;
117 foreach (sort keys %cmd) {
119 next if $cmd && $cmd ne $_;
120 print $fd ' ',pack('A13',$_),$cmd{$_}->[1],"\n";
121 foreach (keys %{$cmd{$_}->[2]}) {
122 # prints out arguments as they should be passed:
123 my $x = s#=s$## ? '<arg>' : s#=i$## ? '<num>' : '';
124 print $fd ' ' x 17, join(', ', map { length $_ > 1 ?
125 "--$_" : "-$_" }
126 split /\|/,$_)," $x\n";
127 }
128 }
129 print $fd <<"";
130\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
131arbitrary identifier if you're tracking multiple SVN branches/repositories in
132one git repository and want to keep them separate. See git-svn(1) for more
133information.
134 exit $exit;
136}
137sub version {
139 print "git-svn version $VERSION\n";
140 exit 0;
141}
142sub rebuild {
144 $SVN_URL = shift or undef;
145 my $newest_rev = 0;
146 if ($_upgrade) {
147 sys('git-update-ref',"refs/remotes/$GIT_SVN","$GIT_SVN-HEAD");
148 } else {
149 check_upgrade_needed();
150 }
151 my $pid = open(my $rev_list,'-|');
153 defined $pid or croak $!;
154 if ($pid == 0) {
155 exec("git-rev-list","refs/remotes/$GIT_SVN") or croak $!;
156 }
157 my $latest;
158 while (<$rev_list>) {
159 chomp;
160 my $c = $_;
161 croak "Non-SHA1: $c\n" unless $c =~ /^$sha1$/o;
162 my @commit = grep(/^git-svn-id: /,`git-cat-file commit $c`);
163 next if (!@commit); # skip merges
164 my $id = $commit[$#commit];
165 my ($url, $rev, $uuid) = ($id =~ /^git-svn-id:\s(\S+?)\@(\d+)
166 \s([a-f\d\-]+)$/x);
167 if (!$rev || !$uuid || !$url) {
168 # some of the original repositories I made had
169 # indentifiers like this:
170 ($rev, $uuid) = ($id =~/^git-svn-id:\s(\d+)
171 \@([a-f\d\-]+)/x);
172 if (!$rev || !$uuid) {
173 croak "Unable to extract revision or UUID from ",
174 "$c, $id\n";
175 }
176 }
177 # if we merged or otherwise started elsewhere, this is
179 # how we break out of it
180 next if (defined $SVN_UUID && ($uuid ne $SVN_UUID));
181 next if (defined $SVN_URL && defined $url && ($url ne $SVN_URL));
182 print "r$rev = $c\n";
184 unless (defined $latest) {
185 if (!$SVN_URL && !$url) {
186 croak "SVN repository location required: $url\n";
187 }
188 $SVN_URL ||= $url;
189 $SVN_UUID ||= $uuid;
190 setup_git_svn();
191 $latest = $rev;
192 }
193 assert_revision_eq_or_unknown($rev, $c);
194 sys('git-update-ref',"$GIT_SVN/revs/$rev",$c);
195 $newest_rev = $rev if ($rev > $newest_rev);
196 }
197 close $rev_list or croak $?;
198 if (!chdir $SVN_WC) {
199 svn_cmd_checkout($SVN_URL, $latest, $SVN_WC);
200 chdir $SVN_WC or croak $!;
201 }
202 $pid = fork;
204 defined $pid or croak $!;
205 if ($pid == 0) {
206 my @svn_up = qw(svn up);
207 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
208 sys(@svn_up,"-r$newest_rev");
209 $ENV{GIT_INDEX_FILE} = $GIT_SVN_INDEX;
210 git_addremove();
211 exec('git-write-tree');
212 }
213 waitpid $pid, 0;
214 if ($_upgrade) {
216 print STDERR <<"";
217Keeping deprecated refs/head/$GIT_SVN-HEAD for now. Please remove it
218when you have upgraded your tools and habits to use refs/remotes/$GIT_SVN
219 }
221}
222sub init {
224 $SVN_URL = shift or croak "SVN repository location required\n";
225 unless (-d $GIT_DIR) {
226 sys('git-init-db');
227 }
228 setup_git_svn();
229}
230sub fetch {
232 my (@parents) = @_;
233 check_upgrade_needed();
234 $SVN_URL ||= file_to_s("$GIT_DIR/$GIT_SVN/info/url");
235 my @log_args = -d $SVN_WC ? ($SVN_WC) : ($SVN_URL);
236 unless ($_revision) {
237 $_revision = -d $SVN_WC ? 'BASE:HEAD' : '0:HEAD';
238 }
239 push @log_args, "-r$_revision";
240 push @log_args, '--stop-on-copy' unless $_no_stop_copy;
241 my $svn_log = svn_log_raw(@log_args);
243 my $base = next_log_entry($svn_log) or croak "No base revision!\n";
245 my $last_commit = undef;
246 unless (-d $SVN_WC) {
247 svn_cmd_checkout($SVN_URL,$base->{revision},$SVN_WC);
248 chdir $SVN_WC or croak $!;
249 read_uuid();
250 $last_commit = git_commit($base, @parents);
251 assert_svn_wc_clean($base->{revision}, $last_commit);
252 } else {
253 chdir $SVN_WC or croak $!;
254 read_uuid();
255 $last_commit = file_to_s("$REV_DIR/$base->{revision}");
256 }
257 my @svn_up = qw(svn up);
258 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
259 my $last = $base;
260 while (my $log_msg = next_log_entry($svn_log)) {
261 assert_svn_wc_clean($last->{revision}, $last_commit);
262 if ($last->{revision} >= $log_msg->{revision}) {
263 croak "Out of order: last >= current: ",
264 "$last->{revision} >= $log_msg->{revision}\n";
265 }
266 sys(@svn_up,"-r$log_msg->{revision}");
267 $last_commit = git_commit($log_msg, $last_commit, @parents);
268 $last = $log_msg;
269 }
270 assert_svn_wc_clean($last->{revision}, $last_commit);
271 unless (-e "$GIT_DIR/refs/heads/master") {
272 sys(qw(git-update-ref refs/heads/master),$last_commit);
273 }
274 return $last;
275}
276sub commit {
278 my (@commits) = @_;
279 check_upgrade_needed();
280 if ($_stdin || !@commits) {
281 print "Reading from stdin...\n";
282 @commits = ();
283 while (<STDIN>) {
284 if (/\b($sha1_short)\b/o) {
285 unshift @commits, $1;
286 }
287 }
288 }
289 my @revs;
290 foreach my $c (@commits) {
291 chomp(my @tmp = safe_qx('git-rev-parse',$c));
292 if (scalar @tmp == 1) {
293 push @revs, $tmp[0];
294 } elsif (scalar @tmp > 1) {
295 push @revs, reverse (safe_qx('git-rev-list',@tmp));
296 } else {
297 die "Failed to rev-parse $c\n";
298 }
299 }
300 chomp @revs;
301 fetch();
303 chdir $SVN_WC or croak $!;
304 my $info = svn_info('.');
305 read_uuid($info);
306 my $svn_current_rev = $info->{'Last Changed Rev'};
307 foreach my $c (@revs) {
308 my $mods = svn_checkout_tree($svn_current_rev, $c);
309 if (scalar @$mods == 0) {
310 print "Skipping, no changes detected\n";
311 next;
312 }
313 $svn_current_rev = svn_commit_tree($svn_current_rev, $c);
314 }
315 print "Done committing ",scalar @revs," revisions to SVN\n";
316}
318sub show_ignore {
320 require File::Find or die $!;
321 my $exclude_file = "$GIT_DIR/info/exclude";
322 open my $fh, '<', $exclude_file or croak $!;
323 chomp(my @excludes = (<$fh>));
324 close $fh or croak $!;
325 $SVN_URL ||= file_to_s("$GIT_DIR/$GIT_SVN/info/url");
327 chdir $SVN_WC or croak $!;
328 my %ign;
329 File::Find::find({wanted=>sub{if(lstat $_ && -d _ && -d "$_/.svn"){
330 s#^\./##;
331 @{$ign{$_}} = safe_qx(qw(svn propget svn:ignore),$_);
332 }}, no_chdir=>1},'.');
333 print "\n# /\n";
335 foreach (@{$ign{'.'}}) { print '/',$_ if /\S/ }
336 delete $ign{'.'};
337 foreach my $i (sort keys %ign) {
338 print "\n# ",$i,"\n";
339 foreach (@{$ign{$i}}) { print '/',$i,'/',$_ if /\S/ }
340 }
341}
342########################### utility functions #########################
344sub read_uuid {
346 return if $SVN_UUID;
347 my $info = shift || svn_info('.');
348 $SVN_UUID = $info->{'Repository UUID'} or
349 croak "Repository UUID unreadable\n";
350 s_to_file($SVN_UUID,"$GIT_DIR/$GIT_SVN/info/uuid");
351}
352sub setup_git_svn {
354 defined $SVN_URL or croak "SVN repository location required\n";
355 unless (-d $GIT_DIR) {
356 croak "GIT_DIR=$GIT_DIR does not exist!\n";
357 }
358 mkpath(["$GIT_DIR/$GIT_SVN"]);
359 mkpath(["$GIT_DIR/$GIT_SVN/info"]);
360 mkpath([$REV_DIR]);
361 s_to_file($SVN_URL,"$GIT_DIR/$GIT_SVN/info/url");
362 open my $fd, '>>', "$GIT_DIR/$GIT_SVN/info/exclude" or croak $!;
364 print $fd '.svn',"\n";
365 close $fd or croak $!;
366}
367sub assert_svn_wc_clean {
369 my ($svn_rev, $treeish) = @_;
370 croak "$svn_rev is not an integer!\n" unless ($svn_rev =~ /^\d+$/);
371 croak "$treeish is not a sha1!\n" unless ($treeish =~ /^$sha1$/o);
372 my $lcr = svn_info('.')->{'Last Changed Rev'};
373 if ($svn_rev != $lcr) {
374 print STDERR "Checking for copy-tree ... ";
375 # use
376 my @diff = grep(/^Index: /,(safe_qx(qw(svn diff),
377 "-r$lcr:$svn_rev")));
378 if (@diff) {
379 croak "Nope! Expected r$svn_rev, got r$lcr\n";
380 } else {
381 print STDERR "OK!\n";
382 }
383 }
384 my @status = grep(!/^Performing status on external/,(`svn status`));
385 @status = grep(!/^\s*$/,@status);
386 if (scalar @status) {
387 print STDERR "Tree ($SVN_WC) is not clean:\n";
388 print STDERR $_ foreach @status;
389 croak;
390 }
391 assert_tree($treeish);
392}
393sub assert_tree {
395 my ($treeish) = @_;
396 croak "Not a sha1: $treeish\n" unless $treeish =~ /^$sha1$/o;
397 chomp(my $type = `git-cat-file -t $treeish`);
398 my $expected;
399 while ($type eq 'tag') {
400 chomp(($treeish, $type) = `git-cat-file tag $treeish`);
401 }
402 if ($type eq 'commit') {
403 $expected = (grep /^tree /,`git-cat-file commit $treeish`)[0];
404 ($expected) = ($expected =~ /^tree ($sha1)$/);
405 die "Unable to get tree from $treeish\n" unless $expected;
406 } elsif ($type eq 'tree') {
407 $expected = $treeish;
408 } else {
409 die "$treeish is a $type, expected tree, tag or commit\n";
410 }
411 my $old_index = $ENV{GIT_INDEX_FILE};
413 my $tmpindex = $GIT_SVN_INDEX.'.assert-tmp';
414 if (-e $tmpindex) {
415 unlink $tmpindex or croak $!;
416 }
417 $ENV{GIT_INDEX_FILE} = $tmpindex;
418 git_addremove();
419 chomp(my $tree = `git-write-tree`);
420 if ($old_index) {
421 $ENV{GIT_INDEX_FILE} = $old_index;
422 } else {
423 delete $ENV{GIT_INDEX_FILE};
424 }
425 if ($tree ne $expected) {
426 croak "Tree mismatch, Got: $tree, Expected: $expected\n";
427 }
428}
429sub parse_diff_tree {
431 my $diff_fh = shift;
432 local $/ = "\0";
433 my $state = 'meta';
434 my @mods;
435 while (<$diff_fh>) {
436 chomp $_; # this gets rid of the trailing "\0"
437 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
438 $sha1\s($sha1)\s([MTCRAD])\d*$/xo) {
439 push @mods, { mode_a => $1, mode_b => $2,
440 sha1_b => $3, chg => $4 };
441 if ($4 =~ /^(?:C|R)$/) {
442 $state = 'file_a';
443 } else {
444 $state = 'file_b';
445 }
446 } elsif ($state eq 'file_a') {
447 my $x = $mods[$#mods] or croak "Empty array\n";
448 if ($x->{chg} !~ /^(?:C|R)$/) {
449 croak "Error parsing $_, $x->{chg}\n";
450 }
451 $x->{file_a} = $_;
452 $state = 'file_b';
453 } elsif ($state eq 'file_b') {
454 my $x = $mods[$#mods] or croak "Empty array\n";
455 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
456 croak "Error parsing $_, $x->{chg}\n";
457 }
458 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
459 croak "Error parsing $_, $x->{chg}\n";
460 }
461 $x->{file_b} = $_;
462 $state = 'meta';
463 } else {
464 croak "Error parsing $_\n";
465 }
466 }
467 close $diff_fh or croak $!;
468 return \@mods;
470}
471sub svn_check_prop_executable {
473 my $m = shift;
474 return if -l $m->{file_b};
475 if ($m->{mode_b} =~ /755$/) {
476 chmod((0755 &~ umask),$m->{file_b}) or croak $!;
477 if ($m->{mode_a} !~ /755$/) {
478 sys(qw(svn propset svn:executable 1), $m->{file_b});
479 }
480 -x $m->{file_b} or croak "$m->{file_b} is not executable!\n";
481 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
482 sys(qw(svn propdel svn:executable), $m->{file_b});
483 chmod((0644 &~ umask),$m->{file_b}) or croak $!;
484 -x $m->{file_b} and croak "$m->{file_b} is executable!\n";
485 }
486}
487sub svn_ensure_parent_path {
489 my $dir_b = dirname(shift);
490 svn_ensure_parent_path($dir_b) if ($dir_b ne File::Spec->curdir);
491 mkpath([$dir_b]) unless (-d $dir_b);
492 sys(qw(svn add -N), $dir_b) unless (-d "$dir_b/.svn");
493}
494sub precommit_check {
496 my $mods = shift;
497 my (%rm_file, %rmdir_check, %added_check);
498 my %o = ( D => 0, R => 1, C => 2, A => 3, M => 3, T => 3 );
500 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
501 if ($m->{chg} eq 'R') {
502 if (-d $m->{file_b}) {
503 err_dir_to_file("$m->{file_a} => $m->{file_b}");
504 }
505 # dir/$file => dir/file/$file
506 my $dirname = dirname($m->{file_b});
507 while ($dirname ne File::Spec->curdir) {
508 if ($dirname ne $m->{file_a}) {
509 $dirname = dirname($dirname);
510 next;
511 }
512 err_file_to_dir("$m->{file_a} => $m->{file_b}");
513 }
514 # baz/zzz => baz (baz is a file)
515 $dirname = dirname($m->{file_a});
516 while ($dirname ne File::Spec->curdir) {
517 if ($dirname ne $m->{file_b}) {
518 $dirname = dirname($dirname);
519 next;
520 }
521 err_dir_to_file("$m->{file_a} => $m->{file_b}");
522 }
523 }
524 if ($m->{chg} =~ /^(D|R)$/) {
525 my $t = $1 eq 'D' ? 'file_b' : 'file_a';
526 $rm_file{ $m->{$t} } = 1;
527 my $dirname = dirname( $m->{$t} );
528 my $basename = basename( $m->{$t} );
529 $rmdir_check{$dirname}->{$basename} = 1;
530 } elsif ($m->{chg} =~ /^(?:A|C)$/) {
531 if (-d $m->{file_b}) {
532 err_dir_to_file($m->{file_b});
533 }
534 my $dirname = dirname( $m->{file_b} );
535 my $basename = basename( $m->{file_b} );
536 $added_check{$dirname}->{$basename} = 1;
537 while ($dirname ne File::Spec->curdir) {
538 if ($rm_file{$dirname}) {
539 err_file_to_dir($m->{file_b});
540 }
541 $dirname = dirname $dirname;
542 }
543 }
544 }
545 return (\%rmdir_check, \%added_check);
546 sub err_dir_to_file {
548 my $file = shift;
549 print STDERR "Node change from directory to file ",
550 "is not supported by Subversion: ",$file,"\n";
551 exit 1;
552 }
553 sub err_file_to_dir {
554 my $file = shift;
555 print STDERR "Node change from file to directory ",
556 "is not supported by Subversion: ",$file,"\n";
557 exit 1;
558 }
559}
560sub svn_checkout_tree {
562 my ($svn_rev, $treeish) = @_;
563 my $from = file_to_s("$REV_DIR/$svn_rev");
564 assert_svn_wc_clean($svn_rev,$from);
565 print "diff-tree $from $treeish\n";
566 my $pid = open my $diff_fh, '-|';
567 defined $pid or croak $!;
568 if ($pid == 0) {
569 my @diff_tree = qw(git-diff-tree -z -r -C);
570 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
571 push @diff_tree, "-l$_l" if defined $_l;
572 exec(@diff_tree, $from, $treeish) or croak $!;
573 }
574 my $mods = parse_diff_tree($diff_fh);
575 unless (@$mods) {
576 # git can do empty commits, but SVN doesn't allow it...
577 return $mods;
578 }
579 my ($rm, $add) = precommit_check($mods);
580 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
582 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
583 if ($m->{chg} eq 'C') {
584 svn_ensure_parent_path( $m->{file_b} );
585 sys(qw(svn cp), $m->{file_a}, $m->{file_b});
586 apply_mod_line_blob($m);
587 svn_check_prop_executable($m);
588 } elsif ($m->{chg} eq 'D') {
589 sys(qw(svn rm --force), $m->{file_b});
590 } elsif ($m->{chg} eq 'R') {
591 svn_ensure_parent_path( $m->{file_b} );
592 sys(qw(svn mv --force), $m->{file_a}, $m->{file_b});
593 apply_mod_line_blob($m);
594 svn_check_prop_executable($m);
595 } elsif ($m->{chg} eq 'M') {
596 apply_mod_line_blob($m);
597 svn_check_prop_executable($m);
598 } elsif ($m->{chg} eq 'T') {
599 sys(qw(svn rm --force),$m->{file_b});
600 apply_mod_line_blob($m);
601 sys(qw(svn add --force), $m->{file_b});
602 svn_check_prop_executable($m);
603 } elsif ($m->{chg} eq 'A') {
604 svn_ensure_parent_path( $m->{file_b} );
605 apply_mod_line_blob($m);
606 sys(qw(svn add --force), $m->{file_b});
607 svn_check_prop_executable($m);
608 } else {
609 croak "Invalid chg: $m->{chg}\n";
610 }
611 }
612 assert_tree($treeish);
614 if ($_rmdir) { # remove empty directories
615 handle_rmdir($rm, $add);
616 }
617 assert_tree($treeish);
618 return $mods;
619}
620# svn ls doesn't work with respect to the current working tree, but what's
622# in the repository. There's not even an option for it... *sigh*
623# (added files don't show up and removed files remain in the ls listing)
624sub svn_ls_current {
625 my ($dir, $rm, $add) = @_;
626 chomp(my @ls = safe_qx('svn','ls',$dir));
627 my @ret = ();
628 foreach (@ls) {
629 s#/$##; # trailing slashes are evil
630 push @ret, $_ unless $rm->{$dir}->{$_};
631 }
632 if (exists $add->{$dir}) {
633 push @ret, keys %{$add->{$dir}};
634 }
635 return \@ret;
636}
637sub handle_rmdir {
639 my ($rm, $add) = @_;
640 foreach my $dir (sort {length $b <=> length $a} keys %$rm) {
642 my $ls = svn_ls_current($dir, $rm, $add);
643 next if (scalar @$ls);
644 sys(qw(svn rm --force),$dir);
645 my $dn = dirname $dir;
647 $rm->{ $dn }->{ basename $dir } = 1;
648 $ls = svn_ls_current($dn, $rm, $add);
649 while (scalar @$ls == 0 && $dn ne File::Spec->curdir) {
650 sys(qw(svn rm --force),$dn);
651 $dir = basename $dn;
652 $dn = dirname $dn;
653 $rm->{ $dn }->{ $dir } = 1;
654 $ls = svn_ls_current($dn, $rm, $add);
655 }
656 }
657}
658sub svn_commit_tree {
660 my ($svn_rev, $commit) = @_;
661 my $commit_msg = "$GIT_DIR/$GIT_SVN/.svn-commit.tmp.$$";
662 my %log_msg = ( msg => '' );
663 open my $msg, '>', $commit_msg or croak $!;
664 chomp(my $type = `git-cat-file -t $commit`);
666 if ($type eq 'commit') {
667 my $pid = open my $msg_fh, '-|';
668 defined $pid or croak $!;
669 if ($pid == 0) {
671 exec(qw(git-cat-file commit), $commit) or croak $!;
672 }
673 my $in_msg = 0;
674 while (<$msg_fh>) {
675 if (!$in_msg) {
676 $in_msg = 1 if (/^\s*$/);
677 } elsif (/^git-svn-id: /) {
678 # skip this, we regenerate the correct one
679 # on re-fetch anyways
680 } else {
681 print $msg $_ or croak $!;
682 }
683 }
684 close $msg_fh or croak $!;
685 }
686 close $msg or croak $!;
687 if ($_edit || ($type eq 'tree')) {
689 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
690 system($editor, $commit_msg);
691 }
692 # file_to_s removes all trailing newlines, so just use chomp() here:
694 open $msg, '<', $commit_msg or croak $!;
695 { local $/; chomp($log_msg{msg} = <$msg>); }
696 close $msg or croak $!;
697 my ($oneline) = ($log_msg{msg} =~ /([^\n\r]+)/);
699 print "Committing $commit: $oneline\n";
700 my @ci_output = safe_qx(qw(svn commit -F),$commit_msg);
702 my ($committed) = grep(/^Committed revision \d+\./,@ci_output);
703 unlink $commit_msg;
704 defined $committed or croak
705 "Commit output failed to parse committed revision!\n",
706 join("\n",@ci_output),"\n";
707 my ($rev_committed) = ($committed =~ /^Committed revision (\d+)\./);
708 my @svn_up = qw(svn up);
710 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
711 if ($rev_committed == ($svn_rev + 1)) {
712 push @svn_up, "-r$rev_committed";
713 sys(@svn_up);
714 my $info = svn_info('.');
715 my $date = $info->{'Last Changed Date'} or die "Missing date\n";
716 if ($info->{'Last Changed Rev'} != $rev_committed) {
717 croak "$info->{'Last Changed Rev'} != $rev_committed\n"
718 }
719 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
720 /(\d{4})\-(\d\d)\-(\d\d)\s
721 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
722 or croak "Failed to parse date: $date\n";
723 $log_msg{date} = "$tz $Y-$m-$d $H:$M:$S";
724 $log_msg{author} = $info->{'Last Changed Author'};
725 $log_msg{revision} = $rev_committed;
726 $log_msg{msg} .= "\n";
727 my $parent = file_to_s("$REV_DIR/$svn_rev");
728 git_commit(\%log_msg, $parent, $commit);
729 return $rev_committed;
730 }
731 # resync immediately
732 push @svn_up, "-r$svn_rev";
733 sys(@svn_up);
734 return fetch("$rev_committed=$commit")->{revision};
735}
736# read the entire log into a temporary file (which is removed ASAP)
738# and store the file handle + parser state
739sub svn_log_raw {
740 my (@log_args) = @_;
741 my $log_fh = IO::File->new_tmpfile or croak $!;
742 my $pid = fork;
743 defined $pid or croak $!;
744 if (!$pid) {
745 open STDOUT, '>&', $log_fh or croak $!;
746 exec (qw(svn log), @log_args) or croak $!
747 }
748 waitpid $pid, 0;
749 croak if $?;
750 seek $log_fh, 0, 0 or croak $!;
751 return { state => 'sep', fh => $log_fh };
752}
753sub next_log_entry {
755 my $log = shift; # retval of svn_log_raw()
756 my $ret = undef;
757 my $fh = $log->{fh};
758 while (<$fh>) {
760 chomp;
761 if (/^\-{72}$/) {
762 if ($log->{state} eq 'msg') {
763 if ($ret->{lines}) {
764 $ret->{msg} .= $_."\n";
765 unless(--$ret->{lines}) {
766 $log->{state} = 'sep';
767 }
768 } else {
769 croak "Log parse error at: $_\n",
770 $ret->{revision},
771 "\n";
772 }
773 next;
774 }
775 if ($log->{state} ne 'sep') {
776 croak "Log parse error at: $_\n",
777 "state: $log->{state}\n",
778 $ret->{revision},
779 "\n";
780 }
781 $log->{state} = 'rev';
782 # if we have an empty log message, put something there:
784 if ($ret) {
785 $ret->{msg} ||= "\n";
786 delete $ret->{lines};
787 return $ret;
788 }
789 next;
790 }
791 if ($log->{state} eq 'rev' && s/^r(\d+)\s*\|\s*//) {
792 my $rev = $1;
793 my ($author, $date, $lines) = split(/\s*\|\s*/, $_, 3);
794 ($lines) = ($lines =~ /(\d+)/);
795 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
796 /(\d{4})\-(\d\d)\-(\d\d)\s
797 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
798 or croak "Failed to parse date: $date\n";
799 $ret = { revision => $rev,
800 date => "$tz $Y-$m-$d $H:$M:$S",
801 author => $author,
802 lines => $lines,
803 msg => '' };
804 if (defined $_authors && ! defined $users{$author}) {
805 die "Author: $author not defined in ",
806 "$_authors file\n";
807 }
808 $log->{state} = 'msg_start';
809 next;
810 }
811 # skip the first blank line of the message:
812 if ($log->{state} eq 'msg_start' && /^$/) {
813 $log->{state} = 'msg';
814 } elsif ($log->{state} eq 'msg') {
815 if ($ret->{lines}) {
816 $ret->{msg} .= $_."\n";
817 unless (--$ret->{lines}) {
818 $log->{state} = 'sep';
819 }
820 } else {
821 croak "Log parse error at: $_\n",
822 $ret->{revision},"\n";
823 }
824 }
825 }
826 return $ret;
827}
828sub svn_info {
830 my $url = shift || $SVN_URL;
831 my $pid = open my $info_fh, '-|';
833 defined $pid or croak $!;
834 if ($pid == 0) {
836 exec(qw(svn info),$url) or croak $!;
837 }
838 my $ret = {};
840 # only single-lines seem to exist in svn info output
841 while (<$info_fh>) {
842 chomp $_;
843 if (m#^([^:]+)\s*:\s*(\S.*)$#) {
844 $ret->{$1} = $2;
845 push @{$ret->{-order}}, $1;
846 }
847 }
848 close $info_fh or croak $!;
849 return $ret;
850}
851sub sys { system(@_) == 0 or croak $? }
853sub git_addremove {
855 system( "git-diff-files --name-only -z ".
856 " | git-update-index --remove -z --stdin && ".
857 "git-ls-files -z --others ".
858 "'--exclude-from=$GIT_DIR/$GIT_SVN/info/exclude'".
859 " | git-update-index --add -z --stdin"
860 ) == 0 or croak $?
861}
862sub s_to_file {
864 my ($str, $file, $mode) = @_;
865 open my $fd,'>',$file or croak $!;
866 print $fd $str,"\n" or croak $!;
867 close $fd or croak $!;
868 chmod ($mode &~ umask, $file) if (defined $mode);
869}
870sub file_to_s {
872 my $file = shift;
873 open my $fd,'<',$file or croak "$!: file: $file\n";
874 local $/;
875 my $ret = <$fd>;
876 close $fd or croak $!;
877 $ret =~ s/\s*$//s;
878 return $ret;
879}
880sub assert_revision_unknown {
882 my $revno = shift;
883 if (-f "$REV_DIR/$revno") {
884 croak "$REV_DIR/$revno already exists! ",
885 "Why are we refetching it?";
886 }
887}
888sub trees_eq {
890 my ($x, $y) = @_;
891 my @x = safe_qx('git-cat-file','commit',$x);
892 my @y = safe_qx('git-cat-file','commit',$y);
893 if (($y[0] ne $x[0]) || $x[0] !~ /^tree $sha1\n$/
894 || $y[0] !~ /^tree $sha1\n$/) {
895 print STDERR "Trees not equal: $y[0] != $x[0]\n";
896 return 0
897 }
898 return 1;
899}
900sub assert_revision_eq_or_unknown {
902 my ($revno, $commit) = @_;
903 if (-f "$REV_DIR/$revno") {
904 my $current = file_to_s("$REV_DIR/$revno");
905 if (($commit ne $current) && !trees_eq($commit, $current)) {
906 croak "$REV_DIR/$revno already exists!\n",
907 "current: $current\nexpected: $commit\n";
908 }
909 return;
910 }
911}
912sub git_commit {
914 my ($log_msg, @parents) = @_;
915 assert_revision_unknown($log_msg->{revision});
916 my $out_fh = IO::File->new_tmpfile or croak $!;
917 map_tree_joins() if (@_branch_from && !%tree_map);
919 # commit parents can be conditionally bound to a particular
921 # svn revision via: "svn_revno=commit_sha1", filter them out here:
922 my @exec_parents;
923 foreach my $p (@parents) {
924 next unless defined $p;
925 if ($p =~ /^(\d+)=($sha1_short)$/o) {
926 if ($1 == $log_msg->{revision}) {
927 push @exec_parents, $2;
928 }
929 } else {
930 push @exec_parents, $p if $p =~ /$sha1_short/o;
931 }
932 }
933 my $pid = fork;
935 defined $pid or croak $!;
936 if ($pid == 0) {
937 $ENV{GIT_INDEX_FILE} = $GIT_SVN_INDEX;
938 git_addremove();
939 chomp(my $tree = `git-write-tree`);
940 croak if $?;
941 if (exists $tree_map{$tree}) {
942 my %seen_parent = map { $_ => 1 } @exec_parents;
943 foreach (@{$tree_map{$tree}}) {
944 # MAXPARENT is defined to 16 in commit-tree.c:
945 if ($seen_parent{$_} || @exec_parents > 16) {
946 next;
947 }
948 push @exec_parents, $_;
949 $seen_parent{$_} = 1;
950 }
951 }
952 my $msg_fh = IO::File->new_tmpfile or croak $!;
953 print $msg_fh $log_msg->{msg}, "\ngit-svn-id: ",
954 "$SVN_URL\@$log_msg->{revision}",
955 " $SVN_UUID\n" or croak $!;
956 $msg_fh->flush == 0 or croak $!;
957 seek $msg_fh, 0, 0 or croak $!;
958 set_commit_env($log_msg);
960 my @exec = ('git-commit-tree',$tree);
962 push @exec, '-p', $_ foreach @exec_parents;
963 open STDIN, '<&', $msg_fh or croak $!;
964 open STDOUT, '>&', $out_fh or croak $!;
965 exec @exec or croak $!;
966 }
967 waitpid($pid,0);
968 croak if $?;
969 $out_fh->flush == 0 or croak $!;
971 seek $out_fh, 0, 0 or croak $!;
972 chomp(my $commit = do { local $/; <$out_fh> });
973 if ($commit !~ /^$sha1$/o) {
974 croak "Failed to commit, invalid sha1: $commit\n";
975 }
976 my @update_ref = ('git-update-ref',"refs/remotes/$GIT_SVN",$commit);
977 if (my $primary_parent = shift @exec_parents) {
978 $pid = fork;
979 defined $pid or croak $!;
980 if (!$pid) {
981 close STDERR;
982 close STDOUT;
983 exec 'git-rev-parse','--verify',
984 "refs/remotes/$GIT_SVN^0";
985 }
986 waitpid $pid, 0;
987 push @update_ref, $primary_parent unless $?;
988 }
989 sys(@update_ref);
990 sys('git-update-ref',"$GIT_SVN/revs/$log_msg->{revision}",$commit);
991 print "r$log_msg->{revision} = $commit\n";
992 return $commit;
993}
994sub set_commit_env {
996 my ($log_msg) = @_;
997 my $author = $log_msg->{author};
998 my ($name,$email) = defined $users{$author} ? @{$users{$author}}
999 : ($author,"$author\@$SVN_UUID");
1000 $ENV{GIT_AUTHOR_NAME} = $ENV{GIT_COMMITTER_NAME} = $name;
1001 $ENV{GIT_AUTHOR_EMAIL} = $ENV{GIT_COMMITTER_EMAIL} = $email;
1002 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_msg->{date};
1003}
1004sub apply_mod_line_blob {
1006 my $m = shift;
1007 if ($m->{mode_b} =~ /^120/) {
1008 blob_to_symlink($m->{sha1_b}, $m->{file_b});
1009 } else {
1010 blob_to_file($m->{sha1_b}, $m->{file_b});
1011 }
1012}
1013sub blob_to_symlink {
1015 my ($blob, $link) = @_;
1016 defined $link or croak "\$link not defined!\n";
1017 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
1018 if (-l $link || -f _) {
1019 unlink $link or croak $!;
1020 }
1021 my $dest = `git-cat-file blob $blob`; # no newline, so no chomp
1023 symlink $dest, $link or croak $!;
1024}
1025sub blob_to_file {
1027 my ($blob, $file) = @_;
1028 defined $file or croak "\$file not defined!\n";
1029 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
1030 if (-l $file || -f _) {
1031 unlink $file or croak $!;
1032 }
1033 open my $blob_fh, '>', $file or croak "$!: $file\n";
1035 my $pid = fork;
1036 defined $pid or croak $!;
1037 if ($pid == 0) {
1039 open STDOUT, '>&', $blob_fh or croak $!;
1040 exec('git-cat-file','blob',$blob);
1041 }
1042 waitpid $pid, 0;
1043 croak $? if $?;
1044 close $blob_fh or croak $!;
1046}
1047sub safe_qx {
1049 my $pid = open my $child, '-|';
1050 defined $pid or croak $!;
1051 if ($pid == 0) {
1052 exec(@_) or croak $?;
1053 }
1054 my @ret = (<$child>);
1055 close $child or croak $?;
1056 die $? if $?; # just in case close didn't error out
1057 return wantarray ? @ret : join('',@ret);
1058}
1059sub svn_compat_check {
1061 my @co_help = safe_qx(qw(svn co -h));
1062 unless (grep /ignore-externals/,@co_help) {
1063 print STDERR "W: Installed svn version does not support ",
1064 "--ignore-externals\n";
1065 $_no_ignore_ext = 1;
1066 }
1067 if (grep /usage: checkout URL\[\@REV\]/,@co_help) {
1068 $_svn_co_url_revs = 1;
1069 }
1070 # I really, really hope nobody hits this...
1072 unless (grep /stop-on-copy/, (safe_qx(qw(svn log -h)))) {
1073 print STDERR <<'';
1074W: The installed svn version does not support the --stop-on-copy flag in
1075 the log command.
1076 Lets hope the directory you're tracking is not a branch or tag
1077 and was never moved within the repository...
1078 $_no_stop_copy = 1;
1080 }
1081}
1082# *sigh*, new versions of svn won't honor -r<rev> without URL@<rev>,
1084# (and they won't honor URL@<rev> without -r<rev>, too!)
1085sub svn_cmd_checkout {
1086 my ($url, $rev, $dir) = @_;
1087 my @cmd = ('svn','co', "-r$rev");
1088 push @cmd, '--ignore-externals' unless $_no_ignore_ext;
1089 $url .= "\@$rev" if $_svn_co_url_revs;
1090 sys(@cmd, $url, $dir);
1091}
1092sub check_upgrade_needed {
1094 my $old = eval {
1095 my $pid = open my $child, '-|';
1096 defined $pid or croak $!;
1097 if ($pid == 0) {
1098 close STDERR;
1099 exec('git-rev-parse',"$GIT_SVN-HEAD") or croak $?;
1100 }
1101 my @ret = (<$child>);
1102 close $child or croak $?;
1103 die $? if $?; # just in case close didn't error out
1104 return wantarray ? @ret : join('',@ret);
1105 };
1106 return unless $old;
1107 my $head = eval { safe_qx('git-rev-parse',"refs/remotes/$GIT_SVN") };
1108 if ($@ || !$head) {
1109 print STDERR "Please run: $0 rebuild --upgrade\n";
1110 exit 1;
1111 }
1112}
1113# fills %tree_map with a reverse mapping of trees to commits. Useful
1115# for finding parents to commit on.
1116sub map_tree_joins {
1117 foreach my $br (@_branch_from) {
1118 my $pid = open my $pipe, '-|';
1119 defined $pid or croak $!;
1120 if ($pid == 0) {
1121 exec(qw(git-rev-list --pretty=raw), $br) or croak $?;
1122 }
1123 while (<$pipe>) {
1124 if (/^commit ($sha1)$/o) {
1125 my $commit = $1;
1126 my ($tree) = (<$pipe> =~ /^tree ($sha1)$/o);
1127 unless (defined $tree) {
1128 die "Failed to parse commit $commit\n";
1129 }
1130 push @{$tree_map{$tree}}, $commit;
1131 }
1132 }
1133 close $pipe or croak $?;
1134 }
1135}
1136# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1138sub load_authors {
1139 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1140 while (<$authors>) {
1141 chomp;
1142 next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/;
1143 my ($user, $name, $email) = ($1, $2, $3);
1144 $users{$user} = [$name, $email];
1145 }
1146 close $authors or croak $!;
1147}
1148__END__
1150Data structures:
1152$svn_log hashref (as returned by svn_log_raw)
1154{
1155 fh => file handle of the log file,
1156 state => state of the log file parser (sep/msg/rev/msg_start...)
1157}
1158$log_msg hashref as returned by next_log_entry($svn_log)
1160{
1161 msg => 'whitespace-formatted log entry
1162', # trailing newline is preserved
1163 revision => '8', # integer
1164 date => '2004-02-24T17:01:44.108345Z', # commit date
1165 author => 'committer name'
1166};
1167@mods = array of diff-index line hashes, each element represents one line
1170 of diff-index output
1171diff-index line ($m hash)
1173{
1174 mode_a => first column of diff-index output, no leading ':',
1175 mode_b => second column of diff-index output,
1176 sha1_b => sha1sum of the final blob,
1177 chg => change type [MCRADT],
1178 file_a => original file name of a file (iff chg is 'C' or 'R')
1179 file_b => new/current file name of a file (any chg)
1180}
1181;