72129de6a3a88021a27f4aaf3be3b4b46ce8c630
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$VERSION = '1.1.0-pre';
12use Cwd qw/abs_path/;
14$GIT_DIR = abs_path($ENV{GIT_DIR} || '.git');
15$ENV{GIT_DIR} = $GIT_DIR;
16my $LC_ALL = $ENV{LC_ALL};
18# make sure the svn binary gives consistent output between locales and TZs:
19$ENV{TZ} = 'UTC';
20$ENV{LC_ALL} = 'C';
21# If SVN:: library support is added, please make the dependencies
23# optional and preserve the capability to use the command-line client.
24# use eval { require SVN::... } to make it lazy load
25# We don't use any modules not in the standard Perl distribution:
26use Carp qw/croak/;
27use IO::File qw//;
28use File::Basename qw/dirname basename/;
29use File::Path qw/mkpath/;
30use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
31use File::Spec qw//;
32use POSIX qw/strftime/;
33my $sha1 = qr/[a-f\d]{40}/;
34my $sha1_short = qr/[a-f\d]{4,40}/;
35my ($_revision,$_stdin,$_no_ignore_ext,$_no_stop_copy,$_help,$_rmdir,$_edit,
36 $_find_copies_harder, $_l, $_version, $_upgrade, $_authors);
37my (@_branch_from, %tree_map, %users);
38my ($_svn_co_url_revs, $_svn_pg_peg_revs);
39my %fc_opts = ( 'no-ignore-externals' => \$_no_ignore_ext,
41 'branch|b=s' => \@_branch_from,
42 'authors-file|A=s' => \$_authors );
43# yes, 'native' sets "\n". Patches to fix this for non-*nix systems welcome:
45my %EOL = ( CR => "\015", LF => "\012", CRLF => "\015\012", native => "\012" );
46my %cmd = (
48 fetch => [ \&fetch, "Download new revisions from SVN",
49 { 'revision|r=s' => \$_revision, %fc_opts } ],
50 init => [ \&init, "Initialize a repo for tracking" .
51 " (requires URL argument)", { } ],
52 commit => [ \&commit, "Commit git revisions to SVN",
53 { 'stdin|' => \$_stdin,
54 'edit|e' => \$_edit,
55 'rmdir' => \$_rmdir,
56 'find-copies-harder' => \$_find_copies_harder,
57 'l=i' => \$_l,
58 %fc_opts,
59 } ],
60 'show-ignore' => [ \&show_ignore, "Show svn:ignore listings", { } ],
61 rebuild => [ \&rebuild, "Rebuild git-svn metadata (after git clone)",
62 { 'no-ignore-externals' => \$_no_ignore_ext,
63 'upgrade' => \$_upgrade } ],
64);
65my $cmd;
66for (my $i = 0; $i < @ARGV; $i++) {
67 if (defined $cmd{$ARGV[$i]}) {
68 $cmd = $ARGV[$i];
69 splice @ARGV, $i, 1;
70 last;
71 }
72};
73my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
75# convert GetOpt::Long specs for use by git-repo-config
77foreach my $o (keys %opts) {
78 my $v = $opts{$o};
79 my ($key) = ($o =~ /^([a-z\-]+)/);
80 $key =~ s/-//g;
81 my $arg = 'git-repo-config';
82 $arg .= ' --int' if ($o =~ /=i$/);
83 $arg .= ' --bool' if ($o !~ /=[sfi]$/);
84 if (ref $v eq 'ARRAY') {
85 chomp(my @tmp = `$arg --get-all svn.$key`);
86 @$v = @tmp if @tmp;
87 } else {
88 chomp(my $tmp = `$arg --get svn.$key`);
89 if ($tmp && !($arg =~ / --bool / && $tmp eq 'false')) {
90 $$v = $tmp;
91 }
92 }
93}
94GetOptions(%opts, 'help|H|h' => \$_help,
96 'version|V' => \$_version,
97 'id|i=s' => \$GIT_SVN) or exit 1;
98$GIT_SVN ||= $ENV{GIT_SVN_ID} || 'git-svn';
100$GIT_SVN_INDEX = "$GIT_DIR/$GIT_SVN/index";
101$SVN_URL = undef;
102$REV_DIR = "$GIT_DIR/$GIT_SVN/revs";
103$SVN_WC = "$GIT_DIR/$GIT_SVN/tree";
104usage(0) if $_help;
106version() if $_version;
107usage(1) unless defined $cmd;
108load_authors() if $_authors;
109svn_compat_check();
110$cmd{$cmd}->[0]->(@ARGV);
111exit 0;
112####################### primary functions ######################
114sub usage {
115 my $exit = shift || 0;
116 my $fd = $exit ? \*STDERR : \*STDOUT;
117 print $fd <<"";
118git-svn - bidirectional operations between a single Subversion tree and git
119Usage: $0 <command> [options] [arguments]\n
120 print $fd "Available commands:\n" unless $cmd;
122 foreach (sort keys %cmd) {
124 next if $cmd && $cmd ne $_;
125 print $fd ' ',pack('A13',$_),$cmd{$_}->[1],"\n";
126 foreach (keys %{$cmd{$_}->[2]}) {
127 # prints out arguments as they should be passed:
128 my $x = s#=s$## ? '<arg>' : s#=i$## ? '<num>' : '';
129 print $fd ' ' x 17, join(', ', map { length $_ > 1 ?
130 "--$_" : "-$_" }
131 split /\|/,$_)," $x\n";
132 }
133 }
134 print $fd <<"";
135\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
136arbitrary identifier if you're tracking multiple SVN branches/repositories in
137one git repository and want to keep them separate. See git-svn(1) for more
138information.
139 exit $exit;
141}
142sub version {
144 print "git-svn version $VERSION\n";
145 exit 0;
146}
147sub rebuild {
149 $SVN_URL = shift or undef;
150 my $newest_rev = 0;
151 if ($_upgrade) {
152 sys('git-update-ref',"refs/remotes/$GIT_SVN","$GIT_SVN-HEAD");
153 } else {
154 check_upgrade_needed();
155 }
156 my $pid = open(my $rev_list,'-|');
158 defined $pid or croak $!;
159 if ($pid == 0) {
160 exec("git-rev-list","refs/remotes/$GIT_SVN") or croak $!;
161 }
162 my $latest;
163 while (<$rev_list>) {
164 chomp;
165 my $c = $_;
166 croak "Non-SHA1: $c\n" unless $c =~ /^$sha1$/o;
167 my @commit = grep(/^git-svn-id: /,`git-cat-file commit $c`);
168 next if (!@commit); # skip merges
169 my $id = $commit[$#commit];
170 my ($url, $rev, $uuid) = ($id =~ /^git-svn-id:\s(\S+?)\@(\d+)
171 \s([a-f\d\-]+)$/x);
172 if (!$rev || !$uuid || !$url) {
173 # some of the original repositories I made had
174 # indentifiers like this:
175 ($rev, $uuid) = ($id =~/^git-svn-id:\s(\d+)
176 \@([a-f\d\-]+)/x);
177 if (!$rev || !$uuid) {
178 croak "Unable to extract revision or UUID from ",
179 "$c, $id\n";
180 }
181 }
182 # if we merged or otherwise started elsewhere, this is
184 # how we break out of it
185 next if (defined $SVN_UUID && ($uuid ne $SVN_UUID));
186 next if (defined $SVN_URL && defined $url && ($url ne $SVN_URL));
187 print "r$rev = $c\n";
189 unless (defined $latest) {
190 if (!$SVN_URL && !$url) {
191 croak "SVN repository location required: $url\n";
192 }
193 $SVN_URL ||= $url;
194 $SVN_UUID ||= $uuid;
195 setup_git_svn();
196 $latest = $rev;
197 }
198 assert_revision_eq_or_unknown($rev, $c);
199 sys('git-update-ref',"$GIT_SVN/revs/$rev",$c);
200 $newest_rev = $rev if ($rev > $newest_rev);
201 }
202 close $rev_list or croak $?;
203 if (!chdir $SVN_WC) {
204 svn_cmd_checkout($SVN_URL, $latest, $SVN_WC);
205 chdir $SVN_WC or croak $!;
206 }
207 $pid = fork;
209 defined $pid or croak $!;
210 if ($pid == 0) {
211 my @svn_up = qw(svn up);
212 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
213 sys(@svn_up,"-r$newest_rev");
214 $ENV{GIT_INDEX_FILE} = $GIT_SVN_INDEX;
215 index_changes();
216 exec('git-write-tree');
217 }
218 waitpid $pid, 0;
219 if ($_upgrade) {
221 print STDERR <<"";
222Keeping deprecated refs/head/$GIT_SVN-HEAD for now. Please remove it
223when you have upgraded your tools and habits to use refs/remotes/$GIT_SVN
224 }
226}
227sub init {
229 $SVN_URL = shift or die "SVN repository location required " .
230 "as a command-line argument\n";
231 unless (-d $GIT_DIR) {
232 sys('git-init-db');
233 }
234 setup_git_svn();
235}
236sub fetch {
238 my (@parents) = @_;
239 check_upgrade_needed();
240 $SVN_URL ||= file_to_s("$GIT_DIR/$GIT_SVN/info/url");
241 my @log_args = -d $SVN_WC ? ($SVN_WC) : ($SVN_URL);
242 unless ($_revision) {
243 $_revision = -d $SVN_WC ? 'BASE:HEAD' : '0:HEAD';
244 }
245 push @log_args, "-r$_revision";
246 push @log_args, '--stop-on-copy' unless $_no_stop_copy;
247 my $svn_log = svn_log_raw(@log_args);
249 my $base = next_log_entry($svn_log) or croak "No base revision!\n";
251 my $last_commit = undef;
252 unless (-d $SVN_WC) {
253 svn_cmd_checkout($SVN_URL,$base->{revision},$SVN_WC);
254 chdir $SVN_WC or croak $!;
255 read_uuid();
256 $last_commit = git_commit($base, @parents);
257 assert_tree($last_commit);
258 } else {
259 chdir $SVN_WC or croak $!;
260 read_uuid();
261 $last_commit = file_to_s("$REV_DIR/$base->{revision}");
262 }
263 my @svn_up = qw(svn up);
264 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
265 my $last = $base;
266 while (my $log_msg = next_log_entry($svn_log)) {
267 assert_tree($last_commit);
268 if ($last->{revision} >= $log_msg->{revision}) {
269 croak "Out of order: last >= current: ",
270 "$last->{revision} >= $log_msg->{revision}\n";
271 }
272 # Revert is needed for cases like:
273 # https://svn.musicpd.org/Jamming/trunk (r166:167), but
274 # I can't seem to reproduce something like that on a test...
275 sys(qw/svn revert -R ./);
276 assert_svn_wc_clean($last->{revision});
277 sys(@svn_up,"-r$log_msg->{revision}");
278 $last_commit = git_commit($log_msg, $last_commit, @parents);
279 $last = $log_msg;
280 }
281 unless (-e "$GIT_DIR/refs/heads/master") {
282 sys(qw(git-update-ref refs/heads/master),$last_commit);
283 }
284 return $last;
285}
286sub commit {
288 my (@commits) = @_;
289 check_upgrade_needed();
290 if ($_stdin || !@commits) {
291 print "Reading from stdin...\n";
292 @commits = ();
293 while (<STDIN>) {
294 if (/\b($sha1_short)\b/o) {
295 unshift @commits, $1;
296 }
297 }
298 }
299 my @revs;
300 foreach my $c (@commits) {
301 chomp(my @tmp = safe_qx('git-rev-parse',$c));
302 if (scalar @tmp == 1) {
303 push @revs, $tmp[0];
304 } elsif (scalar @tmp > 1) {
305 push @revs, reverse (safe_qx('git-rev-list',@tmp));
306 } else {
307 die "Failed to rev-parse $c\n";
308 }
309 }
310 chomp @revs;
311 chdir $SVN_WC or croak "Unable to chdir $SVN_WC: $!\n";
313 my $info = svn_info('.');
314 my $fetched = fetch();
315 if ($info->{Revision} != $fetched->{revision}) {
316 print STDERR "There are new revisions that were fetched ",
317 "and need to be merged (or acknowledged) ",
318 "before committing.\n";
319 exit 1;
320 }
321 $info = svn_info('.');
322 read_uuid($info);
323 my $svn_current_rev = $info->{'Last Changed Rev'};
324 foreach my $c (@revs) {
325 my $mods = svn_checkout_tree($svn_current_rev, $c);
326 if (scalar @$mods == 0) {
327 print "Skipping, no changes detected\n";
328 next;
329 }
330 $svn_current_rev = svn_commit_tree($svn_current_rev, $c);
331 }
332 print "Done committing ",scalar @revs," revisions to SVN\n";
333}
334sub show_ignore {
336 require File::Find or die $!;
337 my $exclude_file = "$GIT_DIR/info/exclude";
338 open my $fh, '<', $exclude_file or croak $!;
339 chomp(my @excludes = (<$fh>));
340 close $fh or croak $!;
341 $SVN_URL ||= file_to_s("$GIT_DIR/$GIT_SVN/info/url");
343 chdir $SVN_WC or croak $!;
344 my %ign;
345 File::Find::find({wanted=>sub{if(lstat $_ && -d _ && -d "$_/.svn"){
346 s#^\./##;
347 @{$ign{$_}} = svn_propget_base('svn:ignore', $_);
348 }}, no_chdir=>1},'.');
349 print "\n# /\n";
351 foreach (@{$ign{'.'}}) { print '/',$_ if /\S/ }
352 delete $ign{'.'};
353 foreach my $i (sort keys %ign) {
354 print "\n# ",$i,"\n";
355 foreach (@{$ign{$i}}) { print '/',$i,'/',$_ if /\S/ }
356 }
357}
358########################### utility functions #########################
360sub read_uuid {
362 return if $SVN_UUID;
363 my $info = shift || svn_info('.');
364 $SVN_UUID = $info->{'Repository UUID'} or
365 croak "Repository UUID unreadable\n";
366 s_to_file($SVN_UUID,"$GIT_DIR/$GIT_SVN/info/uuid");
367}
368sub setup_git_svn {
370 defined $SVN_URL or croak "SVN repository location required\n";
371 unless (-d $GIT_DIR) {
372 croak "GIT_DIR=$GIT_DIR does not exist!\n";
373 }
374 mkpath(["$GIT_DIR/$GIT_SVN"]);
375 mkpath(["$GIT_DIR/$GIT_SVN/info"]);
376 mkpath([$REV_DIR]);
377 s_to_file($SVN_URL,"$GIT_DIR/$GIT_SVN/info/url");
378 open my $fd, '>>', "$GIT_DIR/$GIT_SVN/info/exclude" or croak $!;
380 print $fd '.svn',"\n";
381 close $fd or croak $!;
382}
383sub assert_svn_wc_clean {
385 my ($svn_rev) = @_;
386 croak "$svn_rev is not an integer!\n" unless ($svn_rev =~ /^\d+$/);
387 my $lcr = svn_info('.')->{'Last Changed Rev'};
388 if ($svn_rev != $lcr) {
389 print STDERR "Checking for copy-tree ... ";
390 my @diff = grep(/^Index: /,(safe_qx(qw(svn diff),
391 "-r$lcr:$svn_rev")));
392 if (@diff) {
393 croak "Nope! Expected r$svn_rev, got r$lcr\n";
394 } else {
395 print STDERR "OK!\n";
396 }
397 }
398 my @status = grep(!/^Performing status on external/,(`svn status`));
399 @status = grep(!/^\s*$/,@status);
400 if (scalar @status) {
401 print STDERR "Tree ($SVN_WC) is not clean:\n";
402 print STDERR $_ foreach @status;
403 croak;
404 }
405}
406sub assert_tree {
408 my ($treeish) = @_;
409 croak "Not a sha1: $treeish\n" unless $treeish =~ /^$sha1$/o;
410 chomp(my $type = `git-cat-file -t $treeish`);
411 my $expected;
412 while ($type eq 'tag') {
413 chomp(($treeish, $type) = `git-cat-file tag $treeish`);
414 }
415 if ($type eq 'commit') {
416 $expected = (grep /^tree /,`git-cat-file commit $treeish`)[0];
417 ($expected) = ($expected =~ /^tree ($sha1)$/);
418 die "Unable to get tree from $treeish\n" unless $expected;
419 } elsif ($type eq 'tree') {
420 $expected = $treeish;
421 } else {
422 die "$treeish is a $type, expected tree, tag or commit\n";
423 }
424 my $old_index = $ENV{GIT_INDEX_FILE};
426 my $tmpindex = $GIT_SVN_INDEX.'.assert-tmp';
427 if (-e $tmpindex) {
428 unlink $tmpindex or croak $!;
429 }
430 $ENV{GIT_INDEX_FILE} = $tmpindex;
431 index_changes(1);
432 chomp(my $tree = `git-write-tree`);
433 if ($old_index) {
434 $ENV{GIT_INDEX_FILE} = $old_index;
435 } else {
436 delete $ENV{GIT_INDEX_FILE};
437 }
438 if ($tree ne $expected) {
439 croak "Tree mismatch, Got: $tree, Expected: $expected\n";
440 }
441 unlink $tmpindex;
442}
443sub parse_diff_tree {
445 my $diff_fh = shift;
446 local $/ = "\0";
447 my $state = 'meta';
448 my @mods;
449 while (<$diff_fh>) {
450 chomp $_; # this gets rid of the trailing "\0"
451 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
452 $sha1\s($sha1)\s([MTCRAD])\d*$/xo) {
453 push @mods, { mode_a => $1, mode_b => $2,
454 sha1_b => $3, chg => $4 };
455 if ($4 =~ /^(?:C|R)$/) {
456 $state = 'file_a';
457 } else {
458 $state = 'file_b';
459 }
460 } elsif ($state eq 'file_a') {
461 my $x = $mods[$#mods] or croak "Empty array\n";
462 if ($x->{chg} !~ /^(?:C|R)$/) {
463 croak "Error parsing $_, $x->{chg}\n";
464 }
465 $x->{file_a} = $_;
466 $state = 'file_b';
467 } elsif ($state eq 'file_b') {
468 my $x = $mods[$#mods] or croak "Empty array\n";
469 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
470 croak "Error parsing $_, $x->{chg}\n";
471 }
472 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
473 croak "Error parsing $_, $x->{chg}\n";
474 }
475 $x->{file_b} = $_;
476 $state = 'meta';
477 } else {
478 croak "Error parsing $_\n";
479 }
480 }
481 close $diff_fh or croak $!;
482 return \@mods;
484}
485sub svn_check_prop_executable {
487 my $m = shift;
488 return if -l $m->{file_b};
489 if ($m->{mode_b} =~ /755$/) {
490 chmod((0755 &~ umask),$m->{file_b}) or croak $!;
491 if ($m->{mode_a} !~ /755$/) {
492 sys(qw(svn propset svn:executable 1), $m->{file_b});
493 }
494 -x $m->{file_b} or croak "$m->{file_b} is not executable!\n";
495 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
496 sys(qw(svn propdel svn:executable), $m->{file_b});
497 chmod((0644 &~ umask),$m->{file_b}) or croak $!;
498 -x $m->{file_b} and croak "$m->{file_b} is executable!\n";
499 }
500}
501sub svn_ensure_parent_path {
503 my $dir_b = dirname(shift);
504 svn_ensure_parent_path($dir_b) if ($dir_b ne File::Spec->curdir);
505 mkpath([$dir_b]) unless (-d $dir_b);
506 sys(qw(svn add -N), $dir_b) unless (-d "$dir_b/.svn");
507}
508sub precommit_check {
510 my $mods = shift;
511 my (%rm_file, %rmdir_check, %added_check);
512 my %o = ( D => 0, R => 1, C => 2, A => 3, M => 3, T => 3 );
514 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
515 if ($m->{chg} eq 'R') {
516 if (-d $m->{file_b}) {
517 err_dir_to_file("$m->{file_a} => $m->{file_b}");
518 }
519 # dir/$file => dir/file/$file
520 my $dirname = dirname($m->{file_b});
521 while ($dirname ne File::Spec->curdir) {
522 if ($dirname ne $m->{file_a}) {
523 $dirname = dirname($dirname);
524 next;
525 }
526 err_file_to_dir("$m->{file_a} => $m->{file_b}");
527 }
528 # baz/zzz => baz (baz is a file)
529 $dirname = dirname($m->{file_a});
530 while ($dirname ne File::Spec->curdir) {
531 if ($dirname ne $m->{file_b}) {
532 $dirname = dirname($dirname);
533 next;
534 }
535 err_dir_to_file("$m->{file_a} => $m->{file_b}");
536 }
537 }
538 if ($m->{chg} =~ /^(D|R)$/) {
539 my $t = $1 eq 'D' ? 'file_b' : 'file_a';
540 $rm_file{ $m->{$t} } = 1;
541 my $dirname = dirname( $m->{$t} );
542 my $basename = basename( $m->{$t} );
543 $rmdir_check{$dirname}->{$basename} = 1;
544 } elsif ($m->{chg} =~ /^(?:A|C)$/) {
545 if (-d $m->{file_b}) {
546 err_dir_to_file($m->{file_b});
547 }
548 my $dirname = dirname( $m->{file_b} );
549 my $basename = basename( $m->{file_b} );
550 $added_check{$dirname}->{$basename} = 1;
551 while ($dirname ne File::Spec->curdir) {
552 if ($rm_file{$dirname}) {
553 err_file_to_dir($m->{file_b});
554 }
555 $dirname = dirname $dirname;
556 }
557 }
558 }
559 return (\%rmdir_check, \%added_check);
560 sub err_dir_to_file {
562 my $file = shift;
563 print STDERR "Node change from directory to file ",
564 "is not supported by Subversion: ",$file,"\n";
565 exit 1;
566 }
567 sub err_file_to_dir {
568 my $file = shift;
569 print STDERR "Node change from file to directory ",
570 "is not supported by Subversion: ",$file,"\n";
571 exit 1;
572 }
573}
574sub svn_checkout_tree {
576 my ($svn_rev, $treeish) = @_;
577 my $from = file_to_s("$REV_DIR/$svn_rev");
578 assert_tree($from);
579 print "diff-tree $from $treeish\n";
580 my $pid = open my $diff_fh, '-|';
581 defined $pid or croak $!;
582 if ($pid == 0) {
583 my @diff_tree = qw(git-diff-tree -z -r -C);
584 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
585 push @diff_tree, "-l$_l" if defined $_l;
586 exec(@diff_tree, $from, $treeish) or croak $!;
587 }
588 my $mods = parse_diff_tree($diff_fh);
589 unless (@$mods) {
590 # git can do empty commits, but SVN doesn't allow it...
591 return $mods;
592 }
593 my ($rm, $add) = precommit_check($mods);
594 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
596 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
597 if ($m->{chg} eq 'C') {
598 svn_ensure_parent_path( $m->{file_b} );
599 sys(qw(svn cp), $m->{file_a}, $m->{file_b});
600 apply_mod_line_blob($m);
601 svn_check_prop_executable($m);
602 } elsif ($m->{chg} eq 'D') {
603 sys(qw(svn rm --force), $m->{file_b});
604 } elsif ($m->{chg} eq 'R') {
605 svn_ensure_parent_path( $m->{file_b} );
606 sys(qw(svn mv --force), $m->{file_a}, $m->{file_b});
607 apply_mod_line_blob($m);
608 svn_check_prop_executable($m);
609 } elsif ($m->{chg} eq 'M') {
610 apply_mod_line_blob($m);
611 svn_check_prop_executable($m);
612 } elsif ($m->{chg} eq 'T') {
613 sys(qw(svn rm --force),$m->{file_b});
614 apply_mod_line_blob($m);
615 sys(qw(svn add --force), $m->{file_b});
616 svn_check_prop_executable($m);
617 } elsif ($m->{chg} eq 'A') {
618 svn_ensure_parent_path( $m->{file_b} );
619 apply_mod_line_blob($m);
620 sys(qw(svn add --force), $m->{file_b});
621 svn_check_prop_executable($m);
622 } else {
623 croak "Invalid chg: $m->{chg}\n";
624 }
625 }
626 assert_tree($treeish);
628 if ($_rmdir) { # remove empty directories
629 handle_rmdir($rm, $add);
630 }
631 assert_tree($treeish);
632 return $mods;
633}
634# svn ls doesn't work with respect to the current working tree, but what's
636# in the repository. There's not even an option for it... *sigh*
637# (added files don't show up and removed files remain in the ls listing)
638sub svn_ls_current {
639 my ($dir, $rm, $add) = @_;
640 chomp(my @ls = safe_qx('svn','ls',$dir));
641 my @ret = ();
642 foreach (@ls) {
643 s#/$##; # trailing slashes are evil
644 push @ret, $_ unless $rm->{$dir}->{$_};
645 }
646 if (exists $add->{$dir}) {
647 push @ret, keys %{$add->{$dir}};
648 }
649 return \@ret;
650}
651sub handle_rmdir {
653 my ($rm, $add) = @_;
654 foreach my $dir (sort {length $b <=> length $a} keys %$rm) {
656 my $ls = svn_ls_current($dir, $rm, $add);
657 next if (scalar @$ls);
658 sys(qw(svn rm --force),$dir);
659 my $dn = dirname $dir;
661 $rm->{ $dn }->{ basename $dir } = 1;
662 $ls = svn_ls_current($dn, $rm, $add);
663 while (scalar @$ls == 0 && $dn ne File::Spec->curdir) {
664 sys(qw(svn rm --force),$dn);
665 $dir = basename $dn;
666 $dn = dirname $dn;
667 $rm->{ $dn }->{ $dir } = 1;
668 $ls = svn_ls_current($dn, $rm, $add);
669 }
670 }
671}
672sub svn_commit_tree {
674 my ($svn_rev, $commit) = @_;
675 my $commit_msg = "$GIT_DIR/$GIT_SVN/.svn-commit.tmp.$$";
676 my %log_msg = ( msg => '' );
677 open my $msg, '>', $commit_msg or croak $!;
678 chomp(my $type = `git-cat-file -t $commit`);
680 if ($type eq 'commit') {
681 my $pid = open my $msg_fh, '-|';
682 defined $pid or croak $!;
683 if ($pid == 0) {
685 exec(qw(git-cat-file commit), $commit) or croak $!;
686 }
687 my $in_msg = 0;
688 while (<$msg_fh>) {
689 if (!$in_msg) {
690 $in_msg = 1 if (/^\s*$/);
691 } elsif (/^git-svn-id: /) {
692 # skip this, we regenerate the correct one
693 # on re-fetch anyways
694 } else {
695 print $msg $_ or croak $!;
696 }
697 }
698 close $msg_fh or croak $!;
699 }
700 close $msg or croak $!;
701 if ($_edit || ($type eq 'tree')) {
703 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
704 system($editor, $commit_msg);
705 }
706 # file_to_s removes all trailing newlines, so just use chomp() here:
708 open $msg, '<', $commit_msg or croak $!;
709 { local $/; chomp($log_msg{msg} = <$msg>); }
710 close $msg or croak $!;
711 my ($oneline) = ($log_msg{msg} =~ /([^\n\r]+)/);
713 print "Committing $commit: $oneline\n";
714 if (defined $LC_ALL) {
716 $ENV{LC_ALL} = $LC_ALL;
717 } else {
718 delete $ENV{LC_ALL};
719 }
720 my @ci_output = safe_qx(qw(svn commit -F),$commit_msg);
721 $ENV{LC_ALL} = 'C';
722 unlink $commit_msg;
723 my ($committed) = ($ci_output[$#ci_output] =~ /(\d+)/);
724 if (!defined $committed) {
725 my $out = join("\n",@ci_output);
726 print STDERR "W: Trouble parsing \`svn commit' output:\n\n",
727 $out, "\n\nAssuming English locale...";
728 ($committed) = ($out =~ /^Committed revision \d+\./sm);
729 defined $committed or die " FAILED!\n",
730 "Commit output failed to parse committed revision!\n",
731 print STDERR " OK\n";
732 }
733 my @svn_up = qw(svn up);
735 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
736 if ($committed == ($svn_rev + 1)) {
737 push @svn_up, "-r$committed";
738 sys(@svn_up);
739 my $info = svn_info('.');
740 my $date = $info->{'Last Changed Date'} or die "Missing date\n";
741 if ($info->{'Last Changed Rev'} != $committed) {
742 croak "$info->{'Last Changed Rev'} != $committed\n"
743 }
744 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
745 /(\d{4})\-(\d\d)\-(\d\d)\s
746 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
747 or croak "Failed to parse date: $date\n";
748 $log_msg{date} = "$tz $Y-$m-$d $H:$M:$S";
749 $log_msg{author} = $info->{'Last Changed Author'};
750 $log_msg{revision} = $committed;
751 $log_msg{msg} .= "\n";
752 my $parent = file_to_s("$REV_DIR/$svn_rev");
753 git_commit(\%log_msg, $parent, $commit);
754 return $committed;
755 }
756 # resync immediately
757 push @svn_up, "-r$svn_rev";
758 sys(@svn_up);
759 return fetch("$committed=$commit")->{revision};
760}
761# read the entire log into a temporary file (which is removed ASAP)
763# and store the file handle + parser state
764sub svn_log_raw {
765 my (@log_args) = @_;
766 my $log_fh = IO::File->new_tmpfile or croak $!;
767 my $pid = fork;
768 defined $pid or croak $!;
769 if (!$pid) {
770 open STDOUT, '>&', $log_fh or croak $!;
771 exec (qw(svn log), @log_args) or croak $!
772 }
773 waitpid $pid, 0;
774 croak if $?;
775 seek $log_fh, 0, 0 or croak $!;
776 return { state => 'sep', fh => $log_fh };
777}
778sub next_log_entry {
780 my $log = shift; # retval of svn_log_raw()
781 my $ret = undef;
782 my $fh = $log->{fh};
783 while (<$fh>) {
785 chomp;
786 if (/^\-{72}$/) {
787 if ($log->{state} eq 'msg') {
788 if ($ret->{lines}) {
789 $ret->{msg} .= $_."\n";
790 unless(--$ret->{lines}) {
791 $log->{state} = 'sep';
792 }
793 } else {
794 croak "Log parse error at: $_\n",
795 $ret->{revision},
796 "\n";
797 }
798 next;
799 }
800 if ($log->{state} ne 'sep') {
801 croak "Log parse error at: $_\n",
802 "state: $log->{state}\n",
803 $ret->{revision},
804 "\n";
805 }
806 $log->{state} = 'rev';
807 # if we have an empty log message, put something there:
809 if ($ret) {
810 $ret->{msg} ||= "\n";
811 delete $ret->{lines};
812 return $ret;
813 }
814 next;
815 }
816 if ($log->{state} eq 'rev' && s/^r(\d+)\s*\|\s*//) {
817 my $rev = $1;
818 my ($author, $date, $lines) = split(/\s*\|\s*/, $_, 3);
819 ($lines) = ($lines =~ /(\d+)/);
820 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
821 /(\d{4})\-(\d\d)\-(\d\d)\s
822 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
823 or croak "Failed to parse date: $date\n";
824 $ret = { revision => $rev,
825 date => "$tz $Y-$m-$d $H:$M:$S",
826 author => $author,
827 lines => $lines,
828 msg => '' };
829 if (defined $_authors && ! defined $users{$author}) {
830 die "Author: $author not defined in ",
831 "$_authors file\n";
832 }
833 $log->{state} = 'msg_start';
834 next;
835 }
836 # skip the first blank line of the message:
837 if ($log->{state} eq 'msg_start' && /^$/) {
838 $log->{state} = 'msg';
839 } elsif ($log->{state} eq 'msg') {
840 if ($ret->{lines}) {
841 $ret->{msg} .= $_."\n";
842 unless (--$ret->{lines}) {
843 $log->{state} = 'sep';
844 }
845 } else {
846 croak "Log parse error at: $_\n",
847 $ret->{revision},"\n";
848 }
849 }
850 }
851 return $ret;
852}
853sub svn_info {
855 my $url = shift || $SVN_URL;
856 my $pid = open my $info_fh, '-|';
858 defined $pid or croak $!;
859 if ($pid == 0) {
861 exec(qw(svn info),$url) or croak $!;
862 }
863 my $ret = {};
865 # only single-lines seem to exist in svn info output
866 while (<$info_fh>) {
867 chomp $_;
868 if (m#^([^:]+)\s*:\s*(\S.*)$#) {
869 $ret->{$1} = $2;
870 push @{$ret->{-order}}, $1;
871 }
872 }
873 close $info_fh or croak $!;
874 return $ret;
875}
876sub sys { system(@_) == 0 or croak $? }
878sub eol_cp {
880 my ($from, $to) = @_;
881 my $es = svn_propget_base('svn:eol-style', $to);
882 open my $rfd, '<', $from or croak $!;
883 binmode $rfd or croak $!;
884 open my $wfd, '>', $to or croak $!;
885 binmode $wfd or croak $!;
886 my $eol = $EOL{$es} or undef;
888 my $buf;
889 use bytes;
890 while (1) {
891 my ($r, $w, $t);
892 defined($r = sysread($rfd, $buf, 4096)) or croak $!;
893 return unless $r;
894 if ($eol) {
895 if ($buf =~ /\015$/) {
896 my $c;
897 defined($r = sysread($rfd,$c,1)) or croak $!;
898 $buf .= $c if $r > 0;
899 }
900 $buf =~ s/(?:\015\012|\015|\012)/$eol/gs;
901 $r = length($buf);
902 }
903 for ($w = 0; $w < $r; $w += $t) {
904 $t = syswrite($wfd, $buf, $r - $w, $w) or croak $!;
905 }
906 }
907 no bytes;
908}
909sub do_update_index {
911 my ($z_cmd, $cmd, $no_text_base) = @_;
912 my $z = open my $p, '-|';
914 defined $z or croak $!;
915 unless ($z) { exec @$z_cmd or croak $! }
916 my $pid = open my $ui, '|-';
918 defined $pid or croak $!;
919 unless ($pid) {
920 exec('git-update-index',"--$cmd",'-z','--stdin') or croak $!;
921 }
922 local $/ = "\0";
923 while (my $x = <$p>) {
924 chomp $x;
925 if (!$no_text_base && lstat $x && ! -l _ &&
926 svn_propget_base('svn:keywords', $x)) {
927 my $mode = -x _ ? 0755 : 0644;
928 my ($v,$d,$f) = File::Spec->splitpath($x);
929 my $tb = File::Spec->catfile($d, '.svn', 'tmp',
930 'text-base',"$f.svn-base");
931 $tb =~ s#^/##;
932 unless (-f $tb) {
933 $tb = File::Spec->catfile($d, '.svn',
934 'text-base',"$f.svn-base");
935 $tb =~ s#^/##;
936 }
937 unlink $x or croak $!;
938 eol_cp($tb, $x);
939 chmod(($mode &~ umask), $x) or croak $!;
940 }
941 print $ui $x,"\0";
942 }
943 close $ui or croak $!;
944}
945sub index_changes {
947 my $no_text_base = shift;
948 do_update_index([qw/git-diff-files --name-only -z/],
949 'remove',
950 $no_text_base);
951 do_update_index([qw/git-ls-files -z --others/,
952 "--exclude-from=$GIT_DIR/$GIT_SVN/info/exclude"],
953 'add',
954 $no_text_base);
955}
956sub s_to_file {
958 my ($str, $file, $mode) = @_;
959 open my $fd,'>',$file or croak $!;
960 print $fd $str,"\n" or croak $!;
961 close $fd or croak $!;
962 chmod ($mode &~ umask, $file) if (defined $mode);
963}
964sub file_to_s {
966 my $file = shift;
967 open my $fd,'<',$file or croak "$!: file: $file\n";
968 local $/;
969 my $ret = <$fd>;
970 close $fd or croak $!;
971 $ret =~ s/\s*$//s;
972 return $ret;
973}
974sub assert_revision_unknown {
976 my $revno = shift;
977 if (-f "$REV_DIR/$revno") {
978 croak "$REV_DIR/$revno already exists! ",
979 "Why are we refetching it?";
980 }
981}
982sub trees_eq {
984 my ($x, $y) = @_;
985 my @x = safe_qx('git-cat-file','commit',$x);
986 my @y = safe_qx('git-cat-file','commit',$y);
987 if (($y[0] ne $x[0]) || $x[0] !~ /^tree $sha1\n$/
988 || $y[0] !~ /^tree $sha1\n$/) {
989 print STDERR "Trees not equal: $y[0] != $x[0]\n";
990 return 0
991 }
992 return 1;
993}
994sub assert_revision_eq_or_unknown {
996 my ($revno, $commit) = @_;
997 if (-f "$REV_DIR/$revno") {
998 my $current = file_to_s("$REV_DIR/$revno");
999 if (($commit ne $current) && !trees_eq($commit, $current)) {
1000 croak "$REV_DIR/$revno already exists!\n",
1001 "current: $current\nexpected: $commit\n";
1002 }
1003 return;
1004 }
1005}
1006sub git_commit {
1008 my ($log_msg, @parents) = @_;
1009 assert_revision_unknown($log_msg->{revision});
1010 my $out_fh = IO::File->new_tmpfile or croak $!;
1011 map_tree_joins() if (@_branch_from && !%tree_map);
1013 # commit parents can be conditionally bound to a particular
1015 # svn revision via: "svn_revno=commit_sha1", filter them out here:
1016 my @exec_parents;
1017 foreach my $p (@parents) {
1018 next unless defined $p;
1019 if ($p =~ /^(\d+)=($sha1_short)$/o) {
1020 if ($1 == $log_msg->{revision}) {
1021 push @exec_parents, $2;
1022 }
1023 } else {
1024 push @exec_parents, $p if $p =~ /$sha1_short/o;
1025 }
1026 }
1027 my $pid = fork;
1029 defined $pid or croak $!;
1030 if ($pid == 0) {
1031 $ENV{GIT_INDEX_FILE} = $GIT_SVN_INDEX;
1032 index_changes();
1033 chomp(my $tree = `git-write-tree`);
1034 croak if $?;
1035 if (exists $tree_map{$tree}) {
1036 my %seen_parent = map { $_ => 1 } @exec_parents;
1037 foreach (@{$tree_map{$tree}}) {
1038 # MAXPARENT is defined to 16 in commit-tree.c:
1039 if ($seen_parent{$_} || @exec_parents > 16) {
1040 next;
1041 }
1042 push @exec_parents, $_;
1043 $seen_parent{$_} = 1;
1044 }
1045 }
1046 my $msg_fh = IO::File->new_tmpfile or croak $!;
1047 print $msg_fh $log_msg->{msg}, "\ngit-svn-id: ",
1048 "$SVN_URL\@$log_msg->{revision}",
1049 " $SVN_UUID\n" or croak $!;
1050 $msg_fh->flush == 0 or croak $!;
1051 seek $msg_fh, 0, 0 or croak $!;
1052 set_commit_env($log_msg);
1054 my @exec = ('git-commit-tree',$tree);
1056 push @exec, '-p', $_ foreach @exec_parents;
1057 open STDIN, '<&', $msg_fh or croak $!;
1058 open STDOUT, '>&', $out_fh or croak $!;
1059 exec @exec or croak $!;
1060 }
1061 waitpid($pid,0);
1062 croak if $?;
1063 $out_fh->flush == 0 or croak $!;
1065 seek $out_fh, 0, 0 or croak $!;
1066 chomp(my $commit = do { local $/; <$out_fh> });
1067 if ($commit !~ /^$sha1$/o) {
1068 croak "Failed to commit, invalid sha1: $commit\n";
1069 }
1070 my @update_ref = ('git-update-ref',"refs/remotes/$GIT_SVN",$commit);
1071 if (my $primary_parent = shift @exec_parents) {
1072 $pid = fork;
1073 defined $pid or croak $!;
1074 if (!$pid) {
1075 close STDERR;
1076 close STDOUT;
1077 exec 'git-rev-parse','--verify',
1078 "refs/remotes/$GIT_SVN^0";
1079 }
1080 waitpid $pid, 0;
1081 push @update_ref, $primary_parent unless $?;
1082 }
1083 sys(@update_ref);
1084 sys('git-update-ref',"$GIT_SVN/revs/$log_msg->{revision}",$commit);
1085 print "r$log_msg->{revision} = $commit\n";
1086 return $commit;
1087}
1088sub set_commit_env {
1090 my ($log_msg) = @_;
1091 my $author = $log_msg->{author};
1092 my ($name,$email) = defined $users{$author} ? @{$users{$author}}
1093 : ($author,"$author\@$SVN_UUID");
1094 $ENV{GIT_AUTHOR_NAME} = $ENV{GIT_COMMITTER_NAME} = $name;
1095 $ENV{GIT_AUTHOR_EMAIL} = $ENV{GIT_COMMITTER_EMAIL} = $email;
1096 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_msg->{date};
1097}
1098sub apply_mod_line_blob {
1100 my $m = shift;
1101 if ($m->{mode_b} =~ /^120/) {
1102 blob_to_symlink($m->{sha1_b}, $m->{file_b});
1103 } else {
1104 blob_to_file($m->{sha1_b}, $m->{file_b});
1105 }
1106}
1107sub blob_to_symlink {
1109 my ($blob, $link) = @_;
1110 defined $link or croak "\$link not defined!\n";
1111 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
1112 if (-l $link || -f _) {
1113 unlink $link or croak $!;
1114 }
1115 my $dest = `git-cat-file blob $blob`; # no newline, so no chomp
1117 symlink $dest, $link or croak $!;
1118}
1119sub blob_to_file {
1121 my ($blob, $file) = @_;
1122 defined $file or croak "\$file not defined!\n";
1123 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
1124 if (-l $file || -f _) {
1125 unlink $file or croak $!;
1126 }
1127 open my $blob_fh, '>', $file or croak "$!: $file\n";
1129 my $pid = fork;
1130 defined $pid or croak $!;
1131 if ($pid == 0) {
1133 open STDOUT, '>&', $blob_fh or croak $!;
1134 exec('git-cat-file','blob',$blob);
1135 }
1136 waitpid $pid, 0;
1137 croak $? if $?;
1138 close $blob_fh or croak $!;
1140}
1141sub safe_qx {
1143 my $pid = open my $child, '-|';
1144 defined $pid or croak $!;
1145 if ($pid == 0) {
1146 exec(@_) or croak $?;
1147 }
1148 my @ret = (<$child>);
1149 close $child or croak $?;
1150 die $? if $?; # just in case close didn't error out
1151 return wantarray ? @ret : join('',@ret);
1152}
1153sub svn_compat_check {
1155 my @co_help = safe_qx(qw(svn co -h));
1156 unless (grep /ignore-externals/,@co_help) {
1157 print STDERR "W: Installed svn version does not support ",
1158 "--ignore-externals\n";
1159 $_no_ignore_ext = 1;
1160 }
1161 if (grep /usage: checkout URL\[\@REV\]/,@co_help) {
1162 $_svn_co_url_revs = 1;
1163 }
1164 if (grep /\[TARGET\[\@REV\]\.\.\.\]/, `svn propget -h`) {
1165 $_svn_pg_peg_revs = 1;
1166 }
1167 # I really, really hope nobody hits this...
1169 unless (grep /stop-on-copy/, (safe_qx(qw(svn log -h)))) {
1170 print STDERR <<'';
1171W: The installed svn version does not support the --stop-on-copy flag in
1172 the log command.
1173 Lets hope the directory you're tracking is not a branch or tag
1174 and was never moved within the repository...
1175 $_no_stop_copy = 1;
1177 }
1178}
1179# *sigh*, new versions of svn won't honor -r<rev> without URL@<rev>,
1181# (and they won't honor URL@<rev> without -r<rev>, too!)
1182sub svn_cmd_checkout {
1183 my ($url, $rev, $dir) = @_;
1184 my @cmd = ('svn','co', "-r$rev");
1185 push @cmd, '--ignore-externals' unless $_no_ignore_ext;
1186 $url .= "\@$rev" if $_svn_co_url_revs;
1187 sys(@cmd, $url, $dir);
1188}
1189sub check_upgrade_needed {
1191 my $old = eval {
1192 my $pid = open my $child, '-|';
1193 defined $pid or croak $!;
1194 if ($pid == 0) {
1195 close STDERR;
1196 exec('git-rev-parse',"$GIT_SVN-HEAD") or croak $?;
1197 }
1198 my @ret = (<$child>);
1199 close $child or croak $?;
1200 die $? if $?; # just in case close didn't error out
1201 return wantarray ? @ret : join('',@ret);
1202 };
1203 return unless $old;
1204 my $head = eval { safe_qx('git-rev-parse',"refs/remotes/$GIT_SVN") };
1205 if ($@ || !$head) {
1206 print STDERR "Please run: $0 rebuild --upgrade\n";
1207 exit 1;
1208 }
1209}
1210# fills %tree_map with a reverse mapping of trees to commits. Useful
1212# for finding parents to commit on.
1213sub map_tree_joins {
1214 foreach my $br (@_branch_from) {
1215 my $pid = open my $pipe, '-|';
1216 defined $pid or croak $!;
1217 if ($pid == 0) {
1218 exec(qw(git-rev-list --pretty=raw), $br) or croak $?;
1219 }
1220 while (<$pipe>) {
1221 if (/^commit ($sha1)$/o) {
1222 my $commit = $1;
1223 my ($tree) = (<$pipe> =~ /^tree ($sha1)$/o);
1224 unless (defined $tree) {
1225 die "Failed to parse commit $commit\n";
1226 }
1227 push @{$tree_map{$tree}}, $commit;
1228 }
1229 }
1230 close $pipe or croak $?;
1231 }
1232}
1233# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1235sub load_authors {
1236 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1237 while (<$authors>) {
1238 chomp;
1239 next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/;
1240 my ($user, $name, $email) = ($1, $2, $3);
1241 $users{$user} = [$name, $email];
1242 }
1243 close $authors or croak $!;
1244}
1245sub svn_propget_base {
1247 my ($p, $f) = @_;
1248 $f .= '@BASE' if $_svn_pg_peg_revs;
1249 return safe_qx(qw/svn propget/, $p, $f);
1250}
1251__END__
1253Data structures:
1255$svn_log hashref (as returned by svn_log_raw)
1257{
1258 fh => file handle of the log file,
1259 state => state of the log file parser (sep/msg/rev/msg_start...)
1260}
1261$log_msg hashref as returned by next_log_entry($svn_log)
1263{
1264 msg => 'whitespace-formatted log entry
1265', # trailing newline is preserved
1266 revision => '8', # integer
1267 date => '2004-02-24T17:01:44.108345Z', # commit date
1268 author => 'committer name'
1269};
1270@mods = array of diff-index line hashes, each element represents one line
1273 of diff-index output
1274diff-index line ($m hash)
1276{
1277 mode_a => first column of diff-index output, no leading ':',
1278 mode_b => second column of diff-index output,
1279 sha1_b => sha1sum of the final blob,
1280 chg => change type [MCRADT],
1281 file_a => original file name of a file (iff chg is 'C' or 'R')
1282 file_b => new/current file name of a file (any chg)
1283}
1284;