8bc3d69fdb020eb3d4e036d8a17d3799cc645996
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 fetch();
313 chdir $SVN_WC or croak $!;
314 my $info = svn_info('.');
315 read_uuid($info);
316 my $svn_current_rev = $info->{'Last Changed Rev'};
317 foreach my $c (@revs) {
318 my $mods = svn_checkout_tree($svn_current_rev, $c);
319 if (scalar @$mods == 0) {
320 print "Skipping, no changes detected\n";
321 next;
322 }
323 $svn_current_rev = svn_commit_tree($svn_current_rev, $c);
324 }
325 print "Done committing ",scalar @revs," revisions to SVN\n";
326}
327sub show_ignore {
329 require File::Find or die $!;
330 my $exclude_file = "$GIT_DIR/info/exclude";
331 open my $fh, '<', $exclude_file or croak $!;
332 chomp(my @excludes = (<$fh>));
333 close $fh or croak $!;
334 $SVN_URL ||= file_to_s("$GIT_DIR/$GIT_SVN/info/url");
336 chdir $SVN_WC or croak $!;
337 my %ign;
338 File::Find::find({wanted=>sub{if(lstat $_ && -d _ && -d "$_/.svn"){
339 s#^\./##;
340 @{$ign{$_}} = svn_propget_base('svn:ignore', $_);
341 }}, no_chdir=>1},'.');
342 print "\n# /\n";
344 foreach (@{$ign{'.'}}) { print '/',$_ if /\S/ }
345 delete $ign{'.'};
346 foreach my $i (sort keys %ign) {
347 print "\n# ",$i,"\n";
348 foreach (@{$ign{$i}}) { print '/',$i,'/',$_ if /\S/ }
349 }
350}
351########################### utility functions #########################
353sub read_uuid {
355 return if $SVN_UUID;
356 my $info = shift || svn_info('.');
357 $SVN_UUID = $info->{'Repository UUID'} or
358 croak "Repository UUID unreadable\n";
359 s_to_file($SVN_UUID,"$GIT_DIR/$GIT_SVN/info/uuid");
360}
361sub setup_git_svn {
363 defined $SVN_URL or croak "SVN repository location required\n";
364 unless (-d $GIT_DIR) {
365 croak "GIT_DIR=$GIT_DIR does not exist!\n";
366 }
367 mkpath(["$GIT_DIR/$GIT_SVN"]);
368 mkpath(["$GIT_DIR/$GIT_SVN/info"]);
369 mkpath([$REV_DIR]);
370 s_to_file($SVN_URL,"$GIT_DIR/$GIT_SVN/info/url");
371 open my $fd, '>>', "$GIT_DIR/$GIT_SVN/info/exclude" or croak $!;
373 print $fd '.svn',"\n";
374 close $fd or croak $!;
375}
376sub assert_svn_wc_clean {
378 my ($svn_rev) = @_;
379 croak "$svn_rev is not an integer!\n" unless ($svn_rev =~ /^\d+$/);
380 my $lcr = svn_info('.')->{'Last Changed Rev'};
381 if ($svn_rev != $lcr) {
382 print STDERR "Checking for copy-tree ... ";
383 my @diff = grep(/^Index: /,(safe_qx(qw(svn diff),
384 "-r$lcr:$svn_rev")));
385 if (@diff) {
386 croak "Nope! Expected r$svn_rev, got r$lcr\n";
387 } else {
388 print STDERR "OK!\n";
389 }
390 }
391 my @status = grep(!/^Performing status on external/,(`svn status`));
392 @status = grep(!/^\s*$/,@status);
393 if (scalar @status) {
394 print STDERR "Tree ($SVN_WC) is not clean:\n";
395 print STDERR $_ foreach @status;
396 croak;
397 }
398}
399sub assert_tree {
401 my ($treeish) = @_;
402 croak "Not a sha1: $treeish\n" unless $treeish =~ /^$sha1$/o;
403 chomp(my $type = `git-cat-file -t $treeish`);
404 my $expected;
405 while ($type eq 'tag') {
406 chomp(($treeish, $type) = `git-cat-file tag $treeish`);
407 }
408 if ($type eq 'commit') {
409 $expected = (grep /^tree /,`git-cat-file commit $treeish`)[0];
410 ($expected) = ($expected =~ /^tree ($sha1)$/);
411 die "Unable to get tree from $treeish\n" unless $expected;
412 } elsif ($type eq 'tree') {
413 $expected = $treeish;
414 } else {
415 die "$treeish is a $type, expected tree, tag or commit\n";
416 }
417 my $old_index = $ENV{GIT_INDEX_FILE};
419 my $tmpindex = $GIT_SVN_INDEX.'.assert-tmp';
420 if (-e $tmpindex) {
421 unlink $tmpindex or croak $!;
422 }
423 $ENV{GIT_INDEX_FILE} = $tmpindex;
424 index_changes(1);
425 chomp(my $tree = `git-write-tree`);
426 if ($old_index) {
427 $ENV{GIT_INDEX_FILE} = $old_index;
428 } else {
429 delete $ENV{GIT_INDEX_FILE};
430 }
431 if ($tree ne $expected) {
432 croak "Tree mismatch, Got: $tree, Expected: $expected\n";
433 }
434 unlink $tmpindex;
435}
436sub parse_diff_tree {
438 my $diff_fh = shift;
439 local $/ = "\0";
440 my $state = 'meta';
441 my @mods;
442 while (<$diff_fh>) {
443 chomp $_; # this gets rid of the trailing "\0"
444 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
445 $sha1\s($sha1)\s([MTCRAD])\d*$/xo) {
446 push @mods, { mode_a => $1, mode_b => $2,
447 sha1_b => $3, chg => $4 };
448 if ($4 =~ /^(?:C|R)$/) {
449 $state = 'file_a';
450 } else {
451 $state = 'file_b';
452 }
453 } elsif ($state eq 'file_a') {
454 my $x = $mods[$#mods] or croak "Empty array\n";
455 if ($x->{chg} !~ /^(?:C|R)$/) {
456 croak "Error parsing $_, $x->{chg}\n";
457 }
458 $x->{file_a} = $_;
459 $state = 'file_b';
460 } elsif ($state eq 'file_b') {
461 my $x = $mods[$#mods] or croak "Empty array\n";
462 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
463 croak "Error parsing $_, $x->{chg}\n";
464 }
465 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
466 croak "Error parsing $_, $x->{chg}\n";
467 }
468 $x->{file_b} = $_;
469 $state = 'meta';
470 } else {
471 croak "Error parsing $_\n";
472 }
473 }
474 close $diff_fh or croak $!;
475 return \@mods;
477}
478sub svn_check_prop_executable {
480 my $m = shift;
481 return if -l $m->{file_b};
482 if ($m->{mode_b} =~ /755$/) {
483 chmod((0755 &~ umask),$m->{file_b}) or croak $!;
484 if ($m->{mode_a} !~ /755$/) {
485 sys(qw(svn propset svn:executable 1), $m->{file_b});
486 }
487 -x $m->{file_b} or croak "$m->{file_b} is not executable!\n";
488 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
489 sys(qw(svn propdel svn:executable), $m->{file_b});
490 chmod((0644 &~ umask),$m->{file_b}) or croak $!;
491 -x $m->{file_b} and croak "$m->{file_b} is executable!\n";
492 }
493}
494sub svn_ensure_parent_path {
496 my $dir_b = dirname(shift);
497 svn_ensure_parent_path($dir_b) if ($dir_b ne File::Spec->curdir);
498 mkpath([$dir_b]) unless (-d $dir_b);
499 sys(qw(svn add -N), $dir_b) unless (-d "$dir_b/.svn");
500}
501sub precommit_check {
503 my $mods = shift;
504 my (%rm_file, %rmdir_check, %added_check);
505 my %o = ( D => 0, R => 1, C => 2, A => 3, M => 3, T => 3 );
507 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
508 if ($m->{chg} eq 'R') {
509 if (-d $m->{file_b}) {
510 err_dir_to_file("$m->{file_a} => $m->{file_b}");
511 }
512 # dir/$file => dir/file/$file
513 my $dirname = dirname($m->{file_b});
514 while ($dirname ne File::Spec->curdir) {
515 if ($dirname ne $m->{file_a}) {
516 $dirname = dirname($dirname);
517 next;
518 }
519 err_file_to_dir("$m->{file_a} => $m->{file_b}");
520 }
521 # baz/zzz => baz (baz is a file)
522 $dirname = dirname($m->{file_a});
523 while ($dirname ne File::Spec->curdir) {
524 if ($dirname ne $m->{file_b}) {
525 $dirname = dirname($dirname);
526 next;
527 }
528 err_dir_to_file("$m->{file_a} => $m->{file_b}");
529 }
530 }
531 if ($m->{chg} =~ /^(D|R)$/) {
532 my $t = $1 eq 'D' ? 'file_b' : 'file_a';
533 $rm_file{ $m->{$t} } = 1;
534 my $dirname = dirname( $m->{$t} );
535 my $basename = basename( $m->{$t} );
536 $rmdir_check{$dirname}->{$basename} = 1;
537 } elsif ($m->{chg} =~ /^(?:A|C)$/) {
538 if (-d $m->{file_b}) {
539 err_dir_to_file($m->{file_b});
540 }
541 my $dirname = dirname( $m->{file_b} );
542 my $basename = basename( $m->{file_b} );
543 $added_check{$dirname}->{$basename} = 1;
544 while ($dirname ne File::Spec->curdir) {
545 if ($rm_file{$dirname}) {
546 err_file_to_dir($m->{file_b});
547 }
548 $dirname = dirname $dirname;
549 }
550 }
551 }
552 return (\%rmdir_check, \%added_check);
553 sub err_dir_to_file {
555 my $file = shift;
556 print STDERR "Node change from directory to file ",
557 "is not supported by Subversion: ",$file,"\n";
558 exit 1;
559 }
560 sub err_file_to_dir {
561 my $file = shift;
562 print STDERR "Node change from file to directory ",
563 "is not supported by Subversion: ",$file,"\n";
564 exit 1;
565 }
566}
567sub svn_checkout_tree {
569 my ($svn_rev, $treeish) = @_;
570 my $from = file_to_s("$REV_DIR/$svn_rev");
571 assert_tree($from);
572 print "diff-tree $from $treeish\n";
573 my $pid = open my $diff_fh, '-|';
574 defined $pid or croak $!;
575 if ($pid == 0) {
576 my @diff_tree = qw(git-diff-tree -z -r -C);
577 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
578 push @diff_tree, "-l$_l" if defined $_l;
579 exec(@diff_tree, $from, $treeish) or croak $!;
580 }
581 my $mods = parse_diff_tree($diff_fh);
582 unless (@$mods) {
583 # git can do empty commits, but SVN doesn't allow it...
584 return $mods;
585 }
586 my ($rm, $add) = precommit_check($mods);
587 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
589 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
590 if ($m->{chg} eq 'C') {
591 svn_ensure_parent_path( $m->{file_b} );
592 sys(qw(svn cp), $m->{file_a}, $m->{file_b});
593 apply_mod_line_blob($m);
594 svn_check_prop_executable($m);
595 } elsif ($m->{chg} eq 'D') {
596 sys(qw(svn rm --force), $m->{file_b});
597 } elsif ($m->{chg} eq 'R') {
598 svn_ensure_parent_path( $m->{file_b} );
599 sys(qw(svn mv --force), $m->{file_a}, $m->{file_b});
600 apply_mod_line_blob($m);
601 svn_check_prop_executable($m);
602 } elsif ($m->{chg} eq 'M') {
603 apply_mod_line_blob($m);
604 svn_check_prop_executable($m);
605 } elsif ($m->{chg} eq 'T') {
606 sys(qw(svn rm --force),$m->{file_b});
607 apply_mod_line_blob($m);
608 sys(qw(svn add --force), $m->{file_b});
609 svn_check_prop_executable($m);
610 } elsif ($m->{chg} eq 'A') {
611 svn_ensure_parent_path( $m->{file_b} );
612 apply_mod_line_blob($m);
613 sys(qw(svn add --force), $m->{file_b});
614 svn_check_prop_executable($m);
615 } else {
616 croak "Invalid chg: $m->{chg}\n";
617 }
618 }
619 assert_tree($treeish);
621 if ($_rmdir) { # remove empty directories
622 handle_rmdir($rm, $add);
623 }
624 assert_tree($treeish);
625 return $mods;
626}
627# svn ls doesn't work with respect to the current working tree, but what's
629# in the repository. There's not even an option for it... *sigh*
630# (added files don't show up and removed files remain in the ls listing)
631sub svn_ls_current {
632 my ($dir, $rm, $add) = @_;
633 chomp(my @ls = safe_qx('svn','ls',$dir));
634 my @ret = ();
635 foreach (@ls) {
636 s#/$##; # trailing slashes are evil
637 push @ret, $_ unless $rm->{$dir}->{$_};
638 }
639 if (exists $add->{$dir}) {
640 push @ret, keys %{$add->{$dir}};
641 }
642 return \@ret;
643}
644sub handle_rmdir {
646 my ($rm, $add) = @_;
647 foreach my $dir (sort {length $b <=> length $a} keys %$rm) {
649 my $ls = svn_ls_current($dir, $rm, $add);
650 next if (scalar @$ls);
651 sys(qw(svn rm --force),$dir);
652 my $dn = dirname $dir;
654 $rm->{ $dn }->{ basename $dir } = 1;
655 $ls = svn_ls_current($dn, $rm, $add);
656 while (scalar @$ls == 0 && $dn ne File::Spec->curdir) {
657 sys(qw(svn rm --force),$dn);
658 $dir = basename $dn;
659 $dn = dirname $dn;
660 $rm->{ $dn }->{ $dir } = 1;
661 $ls = svn_ls_current($dn, $rm, $add);
662 }
663 }
664}
665sub svn_commit_tree {
667 my ($svn_rev, $commit) = @_;
668 my $commit_msg = "$GIT_DIR/$GIT_SVN/.svn-commit.tmp.$$";
669 my %log_msg = ( msg => '' );
670 open my $msg, '>', $commit_msg or croak $!;
671 chomp(my $type = `git-cat-file -t $commit`);
673 if ($type eq 'commit') {
674 my $pid = open my $msg_fh, '-|';
675 defined $pid or croak $!;
676 if ($pid == 0) {
678 exec(qw(git-cat-file commit), $commit) or croak $!;
679 }
680 my $in_msg = 0;
681 while (<$msg_fh>) {
682 if (!$in_msg) {
683 $in_msg = 1 if (/^\s*$/);
684 } elsif (/^git-svn-id: /) {
685 # skip this, we regenerate the correct one
686 # on re-fetch anyways
687 } else {
688 print $msg $_ or croak $!;
689 }
690 }
691 close $msg_fh or croak $!;
692 }
693 close $msg or croak $!;
694 if ($_edit || ($type eq 'tree')) {
696 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
697 system($editor, $commit_msg);
698 }
699 # file_to_s removes all trailing newlines, so just use chomp() here:
701 open $msg, '<', $commit_msg or croak $!;
702 { local $/; chomp($log_msg{msg} = <$msg>); }
703 close $msg or croak $!;
704 my ($oneline) = ($log_msg{msg} =~ /([^\n\r]+)/);
706 print "Committing $commit: $oneline\n";
707 if (defined $LC_ALL) {
709 $ENV{LC_ALL} = $LC_ALL;
710 } else {
711 delete $ENV{LC_ALL};
712 }
713 my @ci_output = safe_qx(qw(svn commit -F),$commit_msg);
714 $ENV{LC_ALL} = 'C';
715 unlink $commit_msg;
716 my ($committed) = ($ci_output[$#ci_output] =~ /(\d+)/);
717 if (!defined $committed) {
718 my $out = join("\n",@ci_output);
719 print STDERR "W: Trouble parsing \`svn commit' output:\n\n",
720 $out, "\n\nAssuming English locale...";
721 ($committed) = ($out =~ /^Committed revision \d+\./sm);
722 defined $committed or die " FAILED!\n",
723 "Commit output failed to parse committed revision!\n",
724 print STDERR " OK\n";
725 }
726 my @svn_up = qw(svn up);
728 push @svn_up, '--ignore-externals' unless $_no_ignore_ext;
729 if ($committed == ($svn_rev + 1)) {
730 push @svn_up, "-r$committed";
731 sys(@svn_up);
732 my $info = svn_info('.');
733 my $date = $info->{'Last Changed Date'} or die "Missing date\n";
734 if ($info->{'Last Changed Rev'} != $committed) {
735 croak "$info->{'Last Changed Rev'} != $committed\n"
736 }
737 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
738 /(\d{4})\-(\d\d)\-(\d\d)\s
739 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
740 or croak "Failed to parse date: $date\n";
741 $log_msg{date} = "$tz $Y-$m-$d $H:$M:$S";
742 $log_msg{author} = $info->{'Last Changed Author'};
743 $log_msg{revision} = $committed;
744 $log_msg{msg} .= "\n";
745 my $parent = file_to_s("$REV_DIR/$svn_rev");
746 git_commit(\%log_msg, $parent, $commit);
747 return $committed;
748 }
749 # resync immediately
750 push @svn_up, "-r$svn_rev";
751 sys(@svn_up);
752 return fetch("$committed=$commit")->{revision};
753}
754# read the entire log into a temporary file (which is removed ASAP)
756# and store the file handle + parser state
757sub svn_log_raw {
758 my (@log_args) = @_;
759 my $log_fh = IO::File->new_tmpfile or croak $!;
760 my $pid = fork;
761 defined $pid or croak $!;
762 if (!$pid) {
763 open STDOUT, '>&', $log_fh or croak $!;
764 exec (qw(svn log), @log_args) or croak $!
765 }
766 waitpid $pid, 0;
767 croak if $?;
768 seek $log_fh, 0, 0 or croak $!;
769 return { state => 'sep', fh => $log_fh };
770}
771sub next_log_entry {
773 my $log = shift; # retval of svn_log_raw()
774 my $ret = undef;
775 my $fh = $log->{fh};
776 while (<$fh>) {
778 chomp;
779 if (/^\-{72}$/) {
780 if ($log->{state} eq 'msg') {
781 if ($ret->{lines}) {
782 $ret->{msg} .= $_."\n";
783 unless(--$ret->{lines}) {
784 $log->{state} = 'sep';
785 }
786 } else {
787 croak "Log parse error at: $_\n",
788 $ret->{revision},
789 "\n";
790 }
791 next;
792 }
793 if ($log->{state} ne 'sep') {
794 croak "Log parse error at: $_\n",
795 "state: $log->{state}\n",
796 $ret->{revision},
797 "\n";
798 }
799 $log->{state} = 'rev';
800 # if we have an empty log message, put something there:
802 if ($ret) {
803 $ret->{msg} ||= "\n";
804 delete $ret->{lines};
805 return $ret;
806 }
807 next;
808 }
809 if ($log->{state} eq 'rev' && s/^r(\d+)\s*\|\s*//) {
810 my $rev = $1;
811 my ($author, $date, $lines) = split(/\s*\|\s*/, $_, 3);
812 ($lines) = ($lines =~ /(\d+)/);
813 my ($Y,$m,$d,$H,$M,$S,$tz) = ($date =~
814 /(\d{4})\-(\d\d)\-(\d\d)\s
815 (\d\d)\:(\d\d)\:(\d\d)\s([\-\+]\d+)/x)
816 or croak "Failed to parse date: $date\n";
817 $ret = { revision => $rev,
818 date => "$tz $Y-$m-$d $H:$M:$S",
819 author => $author,
820 lines => $lines,
821 msg => '' };
822 if (defined $_authors && ! defined $users{$author}) {
823 die "Author: $author not defined in ",
824 "$_authors file\n";
825 }
826 $log->{state} = 'msg_start';
827 next;
828 }
829 # skip the first blank line of the message:
830 if ($log->{state} eq 'msg_start' && /^$/) {
831 $log->{state} = 'msg';
832 } elsif ($log->{state} eq 'msg') {
833 if ($ret->{lines}) {
834 $ret->{msg} .= $_."\n";
835 unless (--$ret->{lines}) {
836 $log->{state} = 'sep';
837 }
838 } else {
839 croak "Log parse error at: $_\n",
840 $ret->{revision},"\n";
841 }
842 }
843 }
844 return $ret;
845}
846sub svn_info {
848 my $url = shift || $SVN_URL;
849 my $pid = open my $info_fh, '-|';
851 defined $pid or croak $!;
852 if ($pid == 0) {
854 exec(qw(svn info),$url) or croak $!;
855 }
856 my $ret = {};
858 # only single-lines seem to exist in svn info output
859 while (<$info_fh>) {
860 chomp $_;
861 if (m#^([^:]+)\s*:\s*(\S.*)$#) {
862 $ret->{$1} = $2;
863 push @{$ret->{-order}}, $1;
864 }
865 }
866 close $info_fh or croak $!;
867 return $ret;
868}
869sub sys { system(@_) == 0 or croak $? }
871sub eol_cp {
873 my ($from, $to) = @_;
874 my $es = svn_propget_base('svn:eol-style', $to);
875 open my $rfd, '<', $from or croak $!;
876 binmode $rfd or croak $!;
877 open my $wfd, '>', $to or croak $!;
878 binmode $wfd or croak $!;
879 my $eol = $EOL{$es} or undef;
881 my $buf;
882 use bytes;
883 while (1) {
884 my ($r, $w, $t);
885 defined($r = sysread($rfd, $buf, 4096)) or croak $!;
886 return unless $r;
887 if ($eol) {
888 if ($buf =~ /\015$/) {
889 my $c;
890 defined($r = sysread($rfd,$c,1)) or croak $!;
891 $buf .= $c if $r > 0;
892 }
893 $buf =~ s/(?:\015\012|\015|\012)/$eol/gs;
894 $r = length($buf);
895 }
896 for ($w = 0; $w < $r; $w += $t) {
897 $t = syswrite($wfd, $buf, $r - $w, $w) or croak $!;
898 }
899 }
900 no bytes;
901}
902sub do_update_index {
904 my ($z_cmd, $cmd, $no_text_base) = @_;
905 my $z = open my $p, '-|';
907 defined $z or croak $!;
908 unless ($z) { exec @$z_cmd or croak $! }
909 my $pid = open my $ui, '|-';
911 defined $pid or croak $!;
912 unless ($pid) {
913 exec('git-update-index',"--$cmd",'-z','--stdin') or croak $!;
914 }
915 local $/ = "\0";
916 while (my $x = <$p>) {
917 chomp $x;
918 if (!$no_text_base && lstat $x && ! -l _ &&
919 svn_propget_base('svn:keywords', $x)) {
920 my $mode = -x _ ? 0755 : 0644;
921 my ($v,$d,$f) = File::Spec->splitpath($x);
922 my $tb = File::Spec->catfile($d, '.svn', 'tmp',
923 'text-base',"$f.svn-base");
924 $tb =~ s#^/##;
925 unless (-f $tb) {
926 $tb = File::Spec->catfile($d, '.svn',
927 'text-base',"$f.svn-base");
928 $tb =~ s#^/##;
929 }
930 unlink $x or croak $!;
931 eol_cp($tb, $x);
932 chmod(($mode &~ umask), $x) or croak $!;
933 }
934 print $ui $x,"\0";
935 }
936 close $ui or croak $!;
937}
938sub index_changes {
940 my $no_text_base = shift;
941 do_update_index([qw/git-diff-files --name-only -z/],
942 'remove',
943 $no_text_base);
944 do_update_index([qw/git-ls-files -z --others/,
945 "--exclude-from=$GIT_DIR/$GIT_SVN/info/exclude"],
946 'add',
947 $no_text_base);
948}
949sub s_to_file {
951 my ($str, $file, $mode) = @_;
952 open my $fd,'>',$file or croak $!;
953 print $fd $str,"\n" or croak $!;
954 close $fd or croak $!;
955 chmod ($mode &~ umask, $file) if (defined $mode);
956}
957sub file_to_s {
959 my $file = shift;
960 open my $fd,'<',$file or croak "$!: file: $file\n";
961 local $/;
962 my $ret = <$fd>;
963 close $fd or croak $!;
964 $ret =~ s/\s*$//s;
965 return $ret;
966}
967sub assert_revision_unknown {
969 my $revno = shift;
970 if (-f "$REV_DIR/$revno") {
971 croak "$REV_DIR/$revno already exists! ",
972 "Why are we refetching it?";
973 }
974}
975sub trees_eq {
977 my ($x, $y) = @_;
978 my @x = safe_qx('git-cat-file','commit',$x);
979 my @y = safe_qx('git-cat-file','commit',$y);
980 if (($y[0] ne $x[0]) || $x[0] !~ /^tree $sha1\n$/
981 || $y[0] !~ /^tree $sha1\n$/) {
982 print STDERR "Trees not equal: $y[0] != $x[0]\n";
983 return 0
984 }
985 return 1;
986}
987sub assert_revision_eq_or_unknown {
989 my ($revno, $commit) = @_;
990 if (-f "$REV_DIR/$revno") {
991 my $current = file_to_s("$REV_DIR/$revno");
992 if (($commit ne $current) && !trees_eq($commit, $current)) {
993 croak "$REV_DIR/$revno already exists!\n",
994 "current: $current\nexpected: $commit\n";
995 }
996 return;
997 }
998}
999sub git_commit {
1001 my ($log_msg, @parents) = @_;
1002 assert_revision_unknown($log_msg->{revision});
1003 my $out_fh = IO::File->new_tmpfile or croak $!;
1004 map_tree_joins() if (@_branch_from && !%tree_map);
1006 # commit parents can be conditionally bound to a particular
1008 # svn revision via: "svn_revno=commit_sha1", filter them out here:
1009 my @exec_parents;
1010 foreach my $p (@parents) {
1011 next unless defined $p;
1012 if ($p =~ /^(\d+)=($sha1_short)$/o) {
1013 if ($1 == $log_msg->{revision}) {
1014 push @exec_parents, $2;
1015 }
1016 } else {
1017 push @exec_parents, $p if $p =~ /$sha1_short/o;
1018 }
1019 }
1020 my $pid = fork;
1022 defined $pid or croak $!;
1023 if ($pid == 0) {
1024 $ENV{GIT_INDEX_FILE} = $GIT_SVN_INDEX;
1025 index_changes();
1026 chomp(my $tree = `git-write-tree`);
1027 croak if $?;
1028 if (exists $tree_map{$tree}) {
1029 my %seen_parent = map { $_ => 1 } @exec_parents;
1030 foreach (@{$tree_map{$tree}}) {
1031 # MAXPARENT is defined to 16 in commit-tree.c:
1032 if ($seen_parent{$_} || @exec_parents > 16) {
1033 next;
1034 }
1035 push @exec_parents, $_;
1036 $seen_parent{$_} = 1;
1037 }
1038 }
1039 my $msg_fh = IO::File->new_tmpfile or croak $!;
1040 print $msg_fh $log_msg->{msg}, "\ngit-svn-id: ",
1041 "$SVN_URL\@$log_msg->{revision}",
1042 " $SVN_UUID\n" or croak $!;
1043 $msg_fh->flush == 0 or croak $!;
1044 seek $msg_fh, 0, 0 or croak $!;
1045 set_commit_env($log_msg);
1047 my @exec = ('git-commit-tree',$tree);
1049 push @exec, '-p', $_ foreach @exec_parents;
1050 open STDIN, '<&', $msg_fh or croak $!;
1051 open STDOUT, '>&', $out_fh or croak $!;
1052 exec @exec or croak $!;
1053 }
1054 waitpid($pid,0);
1055 croak if $?;
1056 $out_fh->flush == 0 or croak $!;
1058 seek $out_fh, 0, 0 or croak $!;
1059 chomp(my $commit = do { local $/; <$out_fh> });
1060 if ($commit !~ /^$sha1$/o) {
1061 croak "Failed to commit, invalid sha1: $commit\n";
1062 }
1063 my @update_ref = ('git-update-ref',"refs/remotes/$GIT_SVN",$commit);
1064 if (my $primary_parent = shift @exec_parents) {
1065 $pid = fork;
1066 defined $pid or croak $!;
1067 if (!$pid) {
1068 close STDERR;
1069 close STDOUT;
1070 exec 'git-rev-parse','--verify',
1071 "refs/remotes/$GIT_SVN^0";
1072 }
1073 waitpid $pid, 0;
1074 push @update_ref, $primary_parent unless $?;
1075 }
1076 sys(@update_ref);
1077 sys('git-update-ref',"$GIT_SVN/revs/$log_msg->{revision}",$commit);
1078 print "r$log_msg->{revision} = $commit\n";
1079 return $commit;
1080}
1081sub set_commit_env {
1083 my ($log_msg) = @_;
1084 my $author = $log_msg->{author};
1085 my ($name,$email) = defined $users{$author} ? @{$users{$author}}
1086 : ($author,"$author\@$SVN_UUID");
1087 $ENV{GIT_AUTHOR_NAME} = $ENV{GIT_COMMITTER_NAME} = $name;
1088 $ENV{GIT_AUTHOR_EMAIL} = $ENV{GIT_COMMITTER_EMAIL} = $email;
1089 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_msg->{date};
1090}
1091sub apply_mod_line_blob {
1093 my $m = shift;
1094 if ($m->{mode_b} =~ /^120/) {
1095 blob_to_symlink($m->{sha1_b}, $m->{file_b});
1096 } else {
1097 blob_to_file($m->{sha1_b}, $m->{file_b});
1098 }
1099}
1100sub blob_to_symlink {
1102 my ($blob, $link) = @_;
1103 defined $link or croak "\$link not defined!\n";
1104 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
1105 if (-l $link || -f _) {
1106 unlink $link or croak $!;
1107 }
1108 my $dest = `git-cat-file blob $blob`; # no newline, so no chomp
1110 symlink $dest, $link or croak $!;
1111}
1112sub blob_to_file {
1114 my ($blob, $file) = @_;
1115 defined $file or croak "\$file not defined!\n";
1116 croak "Not a sha1: $blob\n" unless $blob =~ /^$sha1$/o;
1117 if (-l $file || -f _) {
1118 unlink $file or croak $!;
1119 }
1120 open my $blob_fh, '>', $file or croak "$!: $file\n";
1122 my $pid = fork;
1123 defined $pid or croak $!;
1124 if ($pid == 0) {
1126 open STDOUT, '>&', $blob_fh or croak $!;
1127 exec('git-cat-file','blob',$blob);
1128 }
1129 waitpid $pid, 0;
1130 croak $? if $?;
1131 close $blob_fh or croak $!;
1133}
1134sub safe_qx {
1136 my $pid = open my $child, '-|';
1137 defined $pid or croak $!;
1138 if ($pid == 0) {
1139 exec(@_) or croak $?;
1140 }
1141 my @ret = (<$child>);
1142 close $child or croak $?;
1143 die $? if $?; # just in case close didn't error out
1144 return wantarray ? @ret : join('',@ret);
1145}
1146sub svn_compat_check {
1148 my @co_help = safe_qx(qw(svn co -h));
1149 unless (grep /ignore-externals/,@co_help) {
1150 print STDERR "W: Installed svn version does not support ",
1151 "--ignore-externals\n";
1152 $_no_ignore_ext = 1;
1153 }
1154 if (grep /usage: checkout URL\[\@REV\]/,@co_help) {
1155 $_svn_co_url_revs = 1;
1156 }
1157 if (grep /\[TARGET\[\@REV\]\.\.\.\]/, `svn propget -h`) {
1158 $_svn_pg_peg_revs = 1;
1159 }
1160 # I really, really hope nobody hits this...
1162 unless (grep /stop-on-copy/, (safe_qx(qw(svn log -h)))) {
1163 print STDERR <<'';
1164W: The installed svn version does not support the --stop-on-copy flag in
1165 the log command.
1166 Lets hope the directory you're tracking is not a branch or tag
1167 and was never moved within the repository...
1168 $_no_stop_copy = 1;
1170 }
1171}
1172# *sigh*, new versions of svn won't honor -r<rev> without URL@<rev>,
1174# (and they won't honor URL@<rev> without -r<rev>, too!)
1175sub svn_cmd_checkout {
1176 my ($url, $rev, $dir) = @_;
1177 my @cmd = ('svn','co', "-r$rev");
1178 push @cmd, '--ignore-externals' unless $_no_ignore_ext;
1179 $url .= "\@$rev" if $_svn_co_url_revs;
1180 sys(@cmd, $url, $dir);
1181}
1182sub check_upgrade_needed {
1184 my $old = eval {
1185 my $pid = open my $child, '-|';
1186 defined $pid or croak $!;
1187 if ($pid == 0) {
1188 close STDERR;
1189 exec('git-rev-parse',"$GIT_SVN-HEAD") or croak $?;
1190 }
1191 my @ret = (<$child>);
1192 close $child or croak $?;
1193 die $? if $?; # just in case close didn't error out
1194 return wantarray ? @ret : join('',@ret);
1195 };
1196 return unless $old;
1197 my $head = eval { safe_qx('git-rev-parse',"refs/remotes/$GIT_SVN") };
1198 if ($@ || !$head) {
1199 print STDERR "Please run: $0 rebuild --upgrade\n";
1200 exit 1;
1201 }
1202}
1203# fills %tree_map with a reverse mapping of trees to commits. Useful
1205# for finding parents to commit on.
1206sub map_tree_joins {
1207 foreach my $br (@_branch_from) {
1208 my $pid = open my $pipe, '-|';
1209 defined $pid or croak $!;
1210 if ($pid == 0) {
1211 exec(qw(git-rev-list --pretty=raw), $br) or croak $?;
1212 }
1213 while (<$pipe>) {
1214 if (/^commit ($sha1)$/o) {
1215 my $commit = $1;
1216 my ($tree) = (<$pipe> =~ /^tree ($sha1)$/o);
1217 unless (defined $tree) {
1218 die "Failed to parse commit $commit\n";
1219 }
1220 push @{$tree_map{$tree}}, $commit;
1221 }
1222 }
1223 close $pipe or croak $?;
1224 }
1225}
1226# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1228sub load_authors {
1229 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1230 while (<$authors>) {
1231 chomp;
1232 next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/;
1233 my ($user, $name, $email) = ($1, $2, $3);
1234 $users{$user} = [$name, $email];
1235 }
1236 close $authors or croak $!;
1237}
1238sub svn_propget_base {
1240 my ($p, $f) = @_;
1241 $f .= '@BASE' if $_svn_pg_peg_revs;
1242 return safe_qx(qw/svn propget/, $p, $f);
1243}
1244__END__
1246Data structures:
1248$svn_log hashref (as returned by svn_log_raw)
1250{
1251 fh => file handle of the log file,
1252 state => state of the log file parser (sep/msg/rev/msg_start...)
1253}
1254$log_msg hashref as returned by next_log_entry($svn_log)
1256{
1257 msg => 'whitespace-formatted log entry
1258', # trailing newline is preserved
1259 revision => '8', # integer
1260 date => '2004-02-24T17:01:44.108345Z', # commit date
1261 author => 'committer name'
1262};
1263@mods = array of diff-index line hashes, each element represents one line
1266 of diff-index output
1267diff-index line ($m hash)
1269{
1270 mode_a => first column of diff-index output, no leading ':',
1271 mode_b => second column of diff-index output,
1272 sha1_b => sha1sum of the final blob,
1273 chg => change type [MCRADT],
1274 file_a => original file name of a file (iff chg is 'C' or 'R')
1275 file_b => new/current file name of a file (any chg)
1276}
1277;